diff --git a/assets/assets.sgp b/assets/assets.sgp new file mode 100644 index 0000000..2a7ab6e Binary files /dev/null and b/assets/assets.sgp differ diff --git a/build-vma.sh b/build-vma.sh index 1fa8e03..e9dd872 100755 --- a/build-vma.sh +++ b/build-vma.sh @@ -34,3 +34,27 @@ if ! [ -f build/libxxhash.a ]; then ar rcs $lib $obj rm $obj fi + +# STB_IMAGE +src="external/stb/stb_image.c" +flags="-std=c99 -Wno-everything -Iexternal/stb -c -static" +obj="build/stb_image.o" +lib="build/libstb_image.a" + +if ! [ -f build/libstb_image.a ]; then + $c_compiler $flags $src $out $obj + ar rcs $lib $obj + rm $obj +fi + +# M3D +src="external/m3d/m3d.c" +flags="-std=c99 -Wno-everything -Iexternal/m3d -c -static" +obj="build/m3d.o" +lib="build/libm3d.a" + +if ! [ -f build/libm3d.a ]; then + $c_compiler $flags $src $out $obj + ar rcs $lib $obj + rm $obj +fi diff --git a/dub.json b/dub.json index e925884..c27b067 100644 --- a/dub.json +++ b/dub.json @@ -7,10 +7,10 @@ "targetType": "executable", "targetName": "Gears", "targetPath": "build", - "sourceFiles-linux": ["build/libvma.a", "build/libxxhash.a"], + "sourceFiles-linux": ["build/libvma.a", "build/libxxhash.a", "build/libstb_image.a"], "sourceFiles-windows": [], - "importPaths": ["src/gears", "src/shared"], - "sourcePaths": ["src/gears", "src/shared"], + "importPaths": ["src/gears", "src/shared", "src/generated"], + "sourcePaths": ["src/gears", "src/shared", "src/generated"], "libs-linux": ["xcb", "X11", "X11-xcb", "vulkan", "stdc++"], "libs-windows": [], "preGenerateCommands-linux": ["./build-vma.sh"], @@ -21,9 +21,9 @@ "targetType": "executable", "targetPath": "build", "targetName": "Packer", - "importPaths": ["src/packer", "src/shared"], - "sourcePaths": ["src/packer", "src/shared"], - "sourceFiles-linux": ["build/libxxhash.a"], + "importPaths": ["src/packer", "src/shared", "src/generated"], + "sourcePaths": ["src/packer", "src/shared", "src/generated"], + "sourceFiles-linux": ["build/libxxhash.a", "build/libstb_image.a", "build/libm3d.a"], "preGenerateCommands-linux": ["./build-vma.sh"], "preGenerateCommands-windows": [], }, @@ -34,7 +34,7 @@ "targetName": "Codegen", "importPaths": ["src/codegen", "src/shared"], "sourcePaths": ["src/codegen", "src/shared"], - "sourceFiles-linux": ["build/libxxhash.a"], + "sourceFiles-linux": ["build/libxxhash.a", "build/libstb_image.a"], "preGenerateCommands-linux": ["./build-vma.sh"], "preGenerateCommands-windows": [], } diff --git a/external/m3d/m3d.c b/external/m3d/m3d.c new file mode 100644 index 0000000..769d9a4 --- /dev/null +++ b/external/m3d/m3d.c @@ -0,0 +1,3 @@ +#define M3D_IMPLEMENTATION + +#include "m3d.h" diff --git a/external/m3d/m3d.h b/external/m3d/m3d.h new file mode 100644 index 0000000..8bba183 --- /dev/null +++ b/external/m3d/m3d.h @@ -0,0 +1,6574 @@ +/* + * m3d.h + * https://gitlab.com/bztsrc/model3d + * + * Copyright (C) 2020 bzt (bztsrc@gitlab) + * + * Permission is hereby granted, free of charge, to any person + * obtaining a copy of this software and associated documentation + * files (the "Software"), to deal in the Software without + * restriction, including without limitation the rights to use, copy, + * modify, merge, publish, distribute, sublicense, and/or sell copies + * of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be + * included in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT + * HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, + * WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + * + * @brief ANSI C89 / C++11 single header importer / exporter SDK for the Model 3D (.M3D) format + * https://gitlab.com/bztsrc/model3d + * + * PNG decompressor included from (with minor modifications to make it C89 valid): + * stb_image - v2.13 - public domain image loader - http://nothings.org/stb_image.h + * + * @version: 1.0.0 + */ + +#ifndef _M3D_H_ +#define _M3D_H_ + +#ifdef __cplusplus +extern "C" { +#endif + +#include + +/*** configuration ***/ +#ifndef M3D_MALLOC +# define M3D_MALLOC(sz) malloc(sz) +#endif +#ifndef M3D_REALLOC +# define M3D_REALLOC(p,nsz) realloc(p,nsz) +#endif +#ifndef M3D_FREE +# define M3D_FREE(p) free(p) +#endif +#ifndef M3D_LOG +# define M3D_LOG(x) +#endif +#ifndef M3D_APIVERSION +#define M3D_APIVERSION 0x0100 +#ifndef M3D_DOUBLE +typedef float M3D_FLOAT; +#ifndef M3D_EPSILON +/* carefully choosen for IEEE 754 don't change */ +#define M3D_EPSILON ((M3D_FLOAT)1e-7) +#endif +#else +typedef double M3D_FLOAT; +#ifndef M3D_EPSILON +#define M3D_EPSILON ((M3D_FLOAT)1e-14) +#endif +#endif +#if !defined(M3D_SMALLINDEX) +typedef uint32_t M3D_INDEX; +typedef uint16_t M3D_VOXEL; +#define M3D_UNDEF 0xffffffff +#define M3D_INDEXMAX 0xfffffffe +#define M3D_VOXUNDEF 0xffff +#define M3D_VOXCLEAR 0xfffe +#else +typedef uint16_t M3D_INDEX; +typedef uint8_t M3D_VOXEL; +#define M3D_UNDEF 0xffff +#define M3D_INDEXMAX 0xfffe +#define M3D_VOXUNDEF 0xff +#define M3D_VOXCLEAR 0xfe +#endif +#define M3D_NOTDEFINED 0xffffffff +#ifndef M3D_NUMBONE +#define M3D_NUMBONE 4 +#endif +#ifndef M3D_BONEMAXLEVEL +#define M3D_BONEMAXLEVEL 64 +#endif +#ifndef _MSC_VER +#ifndef _inline +#define _inline __inline__ +#endif +#define _pack __attribute__((packed)) +#define _unused __attribute__((unused)) +#else +#define _inline +#define _pack +#define _unused __pragma(warning(suppress:4100)) +#endif +#ifndef __cplusplus +#define _register register +#else +#define _register +#endif + +/*** File format structures ***/ + +/** + * M3D file format structure + * 3DMO m3dchunk_t file header chunk, may followed by compressed data + * PRVW preview chunk (optional) + * HEAD m3dhdr_t model header chunk + * n x m3dchunk_t more chunks follow + * CMAP color map chunk (optional) + * TMAP texture map chunk (optional) + * VRTS vertex data chunk (optional if it's a material library) + * BONE bind-pose skeleton, bone hierarchy chunk (optional) + * n x m3db_t contains propably more, but at least one bone + * n x m3ds_t skin group records + * MTRL* material chunk(s), can be more (optional) + * n x m3dp_t each material contains propapbly more, but at least one property + * the properties are configurable with a static array, see m3d_propertytypes + * n x m3dchunk_t at least one, but maybe more face chunks + * PROC* procedural face, or + * MESH* triangle mesh (vertex index list) or + * VOXT, VOXD* voxel image (converted to mesh) or + * SHPE* mathematical shapes like parameterized surfaces + * LBLS* annotation label chunks, can be more (optional) + * ACTN* action chunk(s), animation-pose skeletons, can be more (optional) + * n x m3dfr_t each action contains probably more, but at least one frame + * n x m3dtr_t each frame contains probably more, but at least one transformation + * ASET* inlined asset chunk(s), can be more (optional) + * OMD3 end chunk + * + * Typical chunks for a game engine: 3DMO, HEAD, CMAP, TMAP, VRTS, BONE, MTRL, MESH, ACTN, OMD3 + * Typical chunks for distibution: 3DMO, PRVW, HEAD, CMAP, TMAP, VRTS, BONE, MTRL, MESH, ACTN, ASET, OMD3 + * Typical chunks for voxel image: 3DMO, HEAD, CMAP, MTRL, VOXT, VOXD, VOXD, VOXD, OMD3 + * Typical chunks for CAD software: 3DMO, PRVW, HEAD, CMAP, TMAP, VRTS, MTRL, SHPE, LBLS, OMD3 + */ +#ifdef _MSC_VER +#pragma pack(push) +#pragma pack(1) +#endif + +typedef struct { + char magic[4]; + uint32_t length; + float scale; /* deliberately not M3D_FLOAT */ + uint32_t types; +} _pack m3dhdr_t; + +typedef struct { + char magic[4]; + uint32_t length; +} _pack m3dchunk_t; + +#ifdef _MSC_VER +#pragma pack(pop) +#endif + +/*** in-memory model structure ***/ + +/* textmap entry */ +typedef struct { + M3D_FLOAT u; + M3D_FLOAT v; +} m3dti_t; +#define m3d_textureindex_t m3dti_t + +/* texture */ +typedef struct { + char *name; /* texture name */ + uint8_t *d; /* pixels data */ + uint16_t w; /* width */ + uint16_t h; /* height */ + uint8_t f; /* format, 1 = grayscale, 2 = grayscale+alpha, 3 = rgb, 4 = rgba */ +} m3dtx_t; +#define m3d_texturedata_t m3dtx_t + +typedef struct { + M3D_INDEX vertexid; + M3D_FLOAT weight; +} m3dw_t; +#define m3d_weight_t m3dw_t + +/* bone entry */ +typedef struct { + M3D_INDEX parent; /* parent bone index */ + char *name; /* name for this bone */ + M3D_INDEX pos; /* vertex index position */ + M3D_INDEX ori; /* vertex index orientation (quaternion) */ + M3D_INDEX numweight; /* number of controlled vertices */ + m3dw_t *weight; /* weights for those vertices */ + M3D_FLOAT mat4[16]; /* transformation matrix */ +} m3db_t; +#define m3d_bone_t m3db_t + +/* skin: bone per vertex entry */ +typedef struct { + M3D_INDEX boneid[M3D_NUMBONE]; + M3D_FLOAT weight[M3D_NUMBONE]; +} m3ds_t; +#define m3d_skin_t m3ds_t + +/* vertex entry */ +typedef struct { + M3D_FLOAT x; /* 3D coordinates and weight */ + M3D_FLOAT y; + M3D_FLOAT z; + M3D_FLOAT w; + uint32_t color; /* default vertex color */ + M3D_INDEX skinid; /* skin index */ +#ifdef M3D_VERTEXTYPE + uint8_t type; +#endif +} m3dv_t; +#define m3d_vertex_t m3dv_t + +/* material property formats */ +enum { + m3dpf_color, + m3dpf_uint8, + m3dpf_uint16, + m3dpf_uint32, + m3dpf_float, + m3dpf_map +}; +typedef struct { + uint8_t format; + uint8_t id; +#ifdef M3D_ASCII +#define M3D_PROPERTYDEF(f,i,n) { (f), (i), (char*)(n) } + char *key; +#endif +#ifndef M3D_ASCII +#define M3D_PROPERTYDEF(f,i,n) { (f), (i) } +#endif +} m3dpd_t; + +/* material property types */ +/* You shouldn't change the first 8 display and first 4 physical property. Assign the rest as you like. */ +enum { + m3dp_Kd = 0, /* scalar display properties */ + m3dp_Ka, + m3dp_Ks, + m3dp_Ns, + m3dp_Ke, + m3dp_Tf, + m3dp_Km, + m3dp_d, + m3dp_il, + + m3dp_Pr = 64, /* scalar physical properties */ + m3dp_Pm, + m3dp_Ps, + m3dp_Ni, + m3dp_Nt, + + m3dp_map_Kd = 128, /* textured display map properties */ + m3dp_map_Ka, + m3dp_map_Ks, + m3dp_map_Ns, + m3dp_map_Ke, + m3dp_map_Tf, + m3dp_map_Km, /* bump map */ + m3dp_map_D, + m3dp_map_N, /* normal map */ + + m3dp_map_Pr = 192, /* textured physical map properties */ + m3dp_map_Pm, + m3dp_map_Ps, + m3dp_map_Ni, + m3dp_map_Nt +}; +enum { /* aliases */ + m3dp_bump = m3dp_map_Km, + m3dp_map_il = m3dp_map_N, + m3dp_refl = m3dp_map_Pm +}; + +/* material property */ +typedef struct { + uint8_t type; /* property type, see "m3dp_*" enumeration */ + union { + uint32_t color; /* if value is a color, m3dpf_color */ + uint32_t num; /* if value is a number, m3dpf_uint8, m3pf_uint16, m3dpf_uint32 */ + float fnum; /* if value is a floating point number, m3dpf_float */ + M3D_INDEX textureid; /* if value is a texture, m3dpf_map */ + } value; +} m3dp_t; +#define m3d_property_t m3dp_t + +/* material entry */ +typedef struct { + char *name; /* name of the material */ + uint8_t numprop; /* number of properties */ + m3dp_t *prop; /* properties array */ +} m3dm_t; +#define m3d_material_t m3dm_t + +/* face entry */ +typedef struct { + M3D_INDEX materialid; /* material index */ + M3D_INDEX vertex[3]; /* 3D points of the triangle in CCW order */ + M3D_INDEX normal[3]; /* normal vectors */ + M3D_INDEX texcoord[3]; /* UV coordinates */ +#ifdef M3D_VERTEXMAX + M3D_INDEX paramid; /* parameter index */ + M3D_INDEX vertmax[3]; /* maximum 3D points of the triangle in CCW order */ +#endif +} m3df_t; +#define m3d_face_t m3df_t + +typedef struct { + uint16_t count; + char *name; +} m3dvi_t; +#define m3d_voxelitem_t m3dvi_t +#define m3d_parameter_t m3dvi_t + +/* voxel types (voxel palette) */ +typedef struct { + char *name; /* technical name of the voxel */ + uint8_t rotation; /* rotation info */ + uint16_t voxshape; /* voxel shape */ + M3D_INDEX materialid; /* material index */ + uint32_t color; /* default voxel color */ + M3D_INDEX skinid; /* skin index */ + uint8_t numitem; /* number of sub-voxels */ + m3dvi_t *item; /* list of sub-voxels */ +} m3dvt_t; +#define m3d_voxeltype_t m3dvt_t + +/* voxel data blocks */ +typedef struct { + char *name; /* name of the block */ + int32_t x, y, z; /* position */ + uint32_t w, h, d; /* dimension */ + uint8_t uncertain; /* probability */ + uint8_t groupid; /* block group id */ + M3D_VOXEL *data; /* voxel data, indices to voxel type */ +} m3dvx_t; +#define m3d_voxel_t m3dvx_t + +/* shape command types. must match the row in m3d_commandtypes */ +enum { + /* special commands */ + m3dc_use = 0, /* use material */ + m3dc_inc, /* include another shape */ + m3dc_mesh, /* include part of polygon mesh */ + /* approximations */ + m3dc_div, /* subdivision by constant resolution for both u, v */ + m3dc_sub, /* subdivision by constant, different for u and v */ + m3dc_len, /* spacial subdivision by maxlength */ + m3dc_dist, /* subdivision by maxdistance and maxangle */ + /* modifiers */ + m3dc_degu, /* degree for both u, v */ + m3dc_deg, /* separate degree for u and v */ + m3dc_rangeu, /* range for u */ + m3dc_range, /* range for u and v */ + m3dc_paru, /* u parameters (knots) */ + m3dc_parv, /* v parameters */ + m3dc_trim, /* outer trimming curve */ + m3dc_hole, /* inner trimming curve */ + m3dc_scrv, /* spacial curve */ + m3dc_sp, /* special points */ + /* helper curves */ + m3dc_bez1, /* Bezier 1D */ + m3dc_bsp1, /* B-spline 1D */ + m3dc_bez2, /* bezier 2D */ + m3dc_bsp2, /* B-spline 2D */ + /* surfaces */ + m3dc_bezun, /* Bezier 3D with control, UV, normal */ + m3dc_bezu, /* with control and UV */ + m3dc_bezn, /* with control and normal */ + m3dc_bez, /* control points only */ + m3dc_nurbsun, /* B-spline 3D */ + m3dc_nurbsu, + m3dc_nurbsn, + m3dc_nurbs, + m3dc_conn, /* connect surfaces */ + /* geometrical */ + m3dc_line, + m3dc_polygon, + m3dc_circle, + m3dc_cylinder, + m3dc_shpere, + m3dc_torus, + m3dc_cone, + m3dc_cube +}; + +/* shape command argument types */ +enum { + m3dcp_mi_t = 1, /* material index */ + m3dcp_hi_t, /* shape index */ + m3dcp_fi_t, /* face index */ + m3dcp_ti_t, /* texture map index */ + m3dcp_vi_t, /* vertex index */ + m3dcp_qi_t, /* vertex index for quaternions */ + m3dcp_vc_t, /* coordinate or radius, float scalar */ + m3dcp_i1_t, /* int8 scalar */ + m3dcp_i2_t, /* int16 scalar */ + m3dcp_i4_t, /* int32 scalar */ + m3dcp_va_t /* variadic arguments */ +}; + +#define M3D_CMDMAXARG 8 /* if you increase this, add more arguments to the macro below */ +typedef struct { +#ifdef M3D_ASCII +#define M3D_CMDDEF(t,n,p,a,b,c,d,e,f,g,h) { (char*)(n), (p), { (a), (b), (c), (d), (e), (f), (g), (h) } } + char *key; +#endif +#ifndef M3D_ASCII +#define M3D_CMDDEF(t,n,p,a,b,c,d,e,f,g,h) { (p), { (a), (b), (c), (d), (e), (f), (g), (h) } } +#endif + uint8_t p; + uint8_t a[M3D_CMDMAXARG]; +} m3dcd_t; + +/* shape command */ +typedef struct { + uint16_t type; /* shape type */ + uint32_t *arg; /* arguments array */ +} m3dc_t; +#define m3d_shapecommand_t m3dc_t + +/* shape entry */ +typedef struct { + char *name; /* name of the mathematical shape */ + M3D_INDEX group; /* group this shape belongs to or -1 */ + uint32_t numcmd; /* number of commands */ + m3dc_t *cmd; /* commands array */ +} m3dh_t; +#define m3d_shape_t m3dh_t + +/* label entry */ +typedef struct { + char *name; /* name of the annotation layer or NULL */ + char *lang; /* language code or NULL */ + char *text; /* the label text */ + uint32_t color; /* color */ + M3D_INDEX vertexid; /* the vertex the label refers to */ +} m3dl_t; +#define m3d_label_t m3dl_t + +/* frame transformations / working copy skeleton entry */ +typedef struct { + M3D_INDEX boneid; /* selects a node in bone hierarchy */ + M3D_INDEX pos; /* vertex index new position */ + M3D_INDEX ori; /* vertex index new orientation (quaternion) */ +} m3dtr_t; +#define m3d_transform_t m3dtr_t + +/* animation frame entry */ +typedef struct { + uint32_t msec; /* frame's position on the timeline, timestamp */ + M3D_INDEX numtransform; /* number of transformations in this frame */ + m3dtr_t *transform; /* transformations */ +} m3dfr_t; +#define m3d_frame_t m3dfr_t + +/* model action entry */ +typedef struct { + char *name; /* name of the action */ + uint32_t durationmsec; /* duration in millisec (1/1000 sec) */ + M3D_INDEX numframe; /* number of frames in this animation */ + m3dfr_t *frame; /* frames array */ +} m3da_t; +#define m3d_action_t m3da_t + +/* inlined asset */ +typedef struct { + char *name; /* asset name (same pointer as in texture[].name) */ + uint8_t *data; /* compressed asset data */ + uint32_t length; /* compressed data length */ +} m3di_t; +#define m3d_inlinedasset_t m3di_t + +/*** in-memory model structure ***/ +#define M3D_FLG_FREERAW (1<<0) +#define M3D_FLG_FREESTR (1<<1) +#define M3D_FLG_MTLLIB (1<<2) +#define M3D_FLG_GENNORM (1<<3) + +typedef struct { + m3dhdr_t *raw; /* pointer to raw data */ + char flags; /* internal flags */ + signed char errcode; /* returned error code */ + char vc_s, vi_s, si_s, ci_s, ti_s, bi_s, nb_s, sk_s, fc_s, hi_s, fi_s, vd_s, vp_s; /* decoded sizes for types */ + char *name; /* name of the model, like "Utah teapot" */ + char *license; /* usage condition or license, like "MIT", "LGPL" or "BSD-3clause" */ + char *author; /* nickname, email, homepage or github URL etc. */ + char *desc; /* comments, descriptions. May contain '\n' newline character */ + M3D_FLOAT scale; /* the model's bounding cube's size in SI meters */ + M3D_INDEX numcmap; + uint32_t *cmap; /* color map */ + M3D_INDEX numtmap; + m3dti_t *tmap; /* texture map indices */ + M3D_INDEX numtexture; + m3dtx_t *texture; /* uncompressed textures */ + M3D_INDEX numbone; + m3db_t *bone; /* bone hierarchy */ + M3D_INDEX numvertex; + m3dv_t *vertex; /* vertex data */ + M3D_INDEX numskin; + m3ds_t *skin; /* skin data */ + M3D_INDEX nummaterial; + m3dm_t *material; /* material list */ +#ifdef M3D_VERTEXMAX + M3D_INDEX numparam; + m3dvi_t *param; /* parameters and their values list */ +#endif + M3D_INDEX numface; + m3df_t *face; /* model face, polygon (triangle) mesh */ + M3D_INDEX numvoxtype; + m3dvt_t *voxtype; /* model face, voxel types */ + M3D_INDEX numvoxel; + m3dvx_t *voxel; /* model face, cubes compressed into voxels */ + M3D_INDEX numshape; + m3dh_t *shape; /* model face, shape commands */ + M3D_INDEX numlabel; + m3dl_t *label; /* annotation labels */ + M3D_INDEX numaction; + m3da_t *action; /* action animations */ + M3D_INDEX numinlined; + m3di_t *inlined; /* inlined assets */ + M3D_INDEX numextra; + m3dchunk_t **extra; /* unknown chunks, application / engine specific data probably */ + m3di_t preview; /* preview chunk */ +} m3d_t; + +/*** export parameters ***/ +#define M3D_EXP_INT8 0 +#define M3D_EXP_INT16 1 +#define M3D_EXP_FLOAT 2 +#define M3D_EXP_DOUBLE 3 + +#define M3D_EXP_NOCMAP (1<<0) +#define M3D_EXP_NOMATERIAL (1<<1) +#define M3D_EXP_NOFACE (1<<2) +#define M3D_EXP_NONORMAL (1<<3) +#define M3D_EXP_NOTXTCRD (1<<4) +#define M3D_EXP_FLIPTXTCRD (1<<5) +#define M3D_EXP_NORECALC (1<<6) +#define M3D_EXP_IDOSUCK (1<<7) +#define M3D_EXP_NOBONE (1<<8) +#define M3D_EXP_NOACTION (1<<9) +#define M3D_EXP_INLINE (1<<10) +#define M3D_EXP_EXTRA (1<<11) +#define M3D_EXP_NOZLIB (1<<14) +#define M3D_EXP_ASCII (1<<15) +#define M3D_EXP_NOVRTMAX (1<<16) + +/*** error codes ***/ +#define M3D_SUCCESS 0 +#define M3D_ERR_ALLOC -1 +#define M3D_ERR_BADFILE -2 +#define M3D_ERR_UNIMPL -65 +#define M3D_ERR_UNKPROP -66 +#define M3D_ERR_UNKMESH -67 +#define M3D_ERR_UNKIMG -68 +#define M3D_ERR_UNKFRAME -69 +#define M3D_ERR_UNKCMD -70 +#define M3D_ERR_UNKVOX -71 +#define M3D_ERR_TRUNC -72 +#define M3D_ERR_CMAP -73 +#define M3D_ERR_TMAP -74 +#define M3D_ERR_VRTS -75 +#define M3D_ERR_BONE -76 +#define M3D_ERR_MTRL -77 +#define M3D_ERR_SHPE -78 +#define M3D_ERR_VOXT -79 + +#define M3D_ERR_ISFATAL(x) ((x) < 0 && (x) > -65) + +/* callbacks */ +typedef unsigned char *(*m3dread_t)(char *filename, unsigned int *size); /* read file contents into buffer */ +typedef void (*m3dfree_t)(void *buffer); /* free file contents buffer */ +typedef int (*m3dtxsc_t)(const char *name, const void *script, uint32_t len, m3dtx_t *output); /* interpret texture script */ +typedef int (*m3dprsc_t)(const char *name, const void *script, uint32_t len, m3d_t *model); /* interpret surface script */ +#endif /* ifndef M3D_APIVERSION */ + +/*** C prototypes ***/ +/* import / export */ +m3d_t *m3d_load(unsigned char *data, m3dread_t readfilecb, m3dfree_t freecb, m3d_t *mtllib); +unsigned char *m3d_save(m3d_t *model, int quality, int flags, unsigned int *size); +void m3d_free(m3d_t *model); +/* generate animation pose skeleton */ +m3dtr_t *m3d_frame(m3d_t *model, M3D_INDEX actionid, M3D_INDEX frameid, m3dtr_t *skeleton); +m3db_t *m3d_pose(m3d_t *model, M3D_INDEX actionid, uint32_t msec); + +/* private prototypes used by both importer and exporter */ +char *_m3d_safestr(char *in, int morelines); + +/*** C implementation ***/ +#ifdef M3D_IMPLEMENTATION +#if !defined(M3D_NOIMPORTER) || defined(M3D_EXPORTER) +/* material property definitions */ +static m3dpd_t m3d_propertytypes[] = { + M3D_PROPERTYDEF(m3dpf_color, m3dp_Kd, "Kd"), /* diffuse color */ + M3D_PROPERTYDEF(m3dpf_color, m3dp_Ka, "Ka"), /* ambient color */ + M3D_PROPERTYDEF(m3dpf_color, m3dp_Ks, "Ks"), /* specular color */ + M3D_PROPERTYDEF(m3dpf_float, m3dp_Ns, "Ns"), /* specular exponent */ + M3D_PROPERTYDEF(m3dpf_color, m3dp_Ke, "Ke"), /* emissive (emitting light of this color) */ + M3D_PROPERTYDEF(m3dpf_color, m3dp_Tf, "Tf"), /* transmission color */ + M3D_PROPERTYDEF(m3dpf_float, m3dp_Km, "Km"), /* bump strength */ + M3D_PROPERTYDEF(m3dpf_float, m3dp_d, "d"), /* dissolve (transparency) */ + M3D_PROPERTYDEF(m3dpf_uint8, m3dp_il, "il"), /* illumination model (informational, ignored by PBR-shaders) */ + + M3D_PROPERTYDEF(m3dpf_float, m3dp_Pr, "Pr"), /* roughness */ + M3D_PROPERTYDEF(m3dpf_float, m3dp_Pm, "Pm"), /* metallic, also reflection */ + M3D_PROPERTYDEF(m3dpf_float, m3dp_Ps, "Ps"), /* sheen */ + M3D_PROPERTYDEF(m3dpf_float, m3dp_Ni, "Ni"), /* index of refraction (optical density) */ + M3D_PROPERTYDEF(m3dpf_float, m3dp_Nt, "Nt"), /* thickness of face in millimeter, for printing */ + + /* aliases, note that "map_*" aliases are handled automatically */ + M3D_PROPERTYDEF(m3dpf_map, m3dp_map_Km, "bump"), + M3D_PROPERTYDEF(m3dpf_map, m3dp_map_N, "map_N"),/* as normal map has no scalar version, it's counterpart is 'il' */ + M3D_PROPERTYDEF(m3dpf_map, m3dp_map_Pm, "refl") +}; +/* shape command definitions. if more commands start with the same string, the longer must come first */ +static m3dcd_t m3d_commandtypes[] = { + /* technical */ + M3D_CMDDEF(m3dc_use, "use", 1, m3dcp_mi_t, 0, 0, 0, 0, 0, 0, 0), + M3D_CMDDEF(m3dc_inc, "inc", 3, m3dcp_hi_t, m3dcp_vi_t, m3dcp_qi_t, m3dcp_vi_t, 0, 0, 0, 0), + M3D_CMDDEF(m3dc_mesh, "mesh", 1, m3dcp_fi_t, m3dcp_fi_t, m3dcp_vi_t, m3dcp_qi_t, m3dcp_vi_t, 0, 0, 0), + /* approximations */ + M3D_CMDDEF(m3dc_div, "div", 1, m3dcp_vc_t, 0, 0, 0, 0, 0, 0, 0), + M3D_CMDDEF(m3dc_sub, "sub", 2, m3dcp_vc_t, m3dcp_vc_t, 0, 0, 0, 0, 0, 0), + M3D_CMDDEF(m3dc_len, "len", 1, m3dcp_vc_t, 0, 0, 0, 0, 0, 0, 0), + M3D_CMDDEF(m3dc_dist, "dist", 2, m3dcp_vc_t, m3dcp_vc_t, 0, 0, 0, 0, 0, 0), + /* modifiers */ + M3D_CMDDEF(m3dc_degu, "degu", 1, m3dcp_i1_t, 0, 0, 0, 0, 0, 0, 0), + M3D_CMDDEF(m3dc_deg, "deg", 2, m3dcp_i1_t, m3dcp_i1_t, 0, 0, 0, 0, 0, 0), + M3D_CMDDEF(m3dc_rangeu, "rangeu", 1, m3dcp_ti_t, 0, 0, 0, 0, 0, 0, 0), + M3D_CMDDEF(m3dc_range, "range", 2, m3dcp_ti_t, m3dcp_ti_t, 0, 0, 0, 0, 0, 0), + M3D_CMDDEF(m3dc_paru, "paru", 2, m3dcp_va_t, m3dcp_vc_t, 0, 0, 0, 0, 0, 0), + M3D_CMDDEF(m3dc_parv, "parv", 2, m3dcp_va_t, m3dcp_vc_t, 0, 0, 0, 0, 0, 0), + M3D_CMDDEF(m3dc_trim, "trim", 3, m3dcp_va_t, m3dcp_ti_t, m3dcp_i2_t, 0, 0, 0, 0, 0), + M3D_CMDDEF(m3dc_hole, "hole", 3, m3dcp_va_t, m3dcp_ti_t, m3dcp_i2_t, 0, 0, 0, 0, 0), + M3D_CMDDEF(m3dc_scrv, "scrv", 3, m3dcp_va_t, m3dcp_ti_t, m3dcp_i2_t, 0, 0, 0, 0, 0), + M3D_CMDDEF(m3dc_sp, "sp", 2, m3dcp_va_t, m3dcp_vi_t, 0, 0, 0, 0, 0, 0), + /* helper curves */ + M3D_CMDDEF(m3dc_bez1, "bez1", 2, m3dcp_va_t, m3dcp_vi_t, 0, 0, 0, 0, 0, 0), + M3D_CMDDEF(m3dc_bsp1, "bsp1", 2, m3dcp_va_t, m3dcp_vi_t, 0, 0, 0, 0, 0, 0), + M3D_CMDDEF(m3dc_bez2, "bez2", 2, m3dcp_va_t, m3dcp_vi_t, 0, 0, 0, 0, 0, 0), + M3D_CMDDEF(m3dc_bsp2, "bsp2", 2, m3dcp_va_t, m3dcp_vi_t, 0, 0, 0, 0, 0, 0), + /* surfaces */ + M3D_CMDDEF(m3dc_bezun, "bezun", 4, m3dcp_va_t, m3dcp_vi_t, m3dcp_ti_t, m3dcp_vi_t, 0, 0, 0, 0), + M3D_CMDDEF(m3dc_bezu, "bezu", 3, m3dcp_va_t, m3dcp_vi_t, m3dcp_ti_t, 0, 0, 0, 0, 0), + M3D_CMDDEF(m3dc_bezn, "bezn", 3, m3dcp_va_t, m3dcp_vi_t, m3dcp_vi_t, 0, 0, 0, 0, 0), + M3D_CMDDEF(m3dc_bez, "bez", 2, m3dcp_va_t, m3dcp_vi_t, 0, 0, 0, 0, 0, 0), + M3D_CMDDEF(m3dc_nurbsun, "nurbsun", 4, m3dcp_va_t, m3dcp_vi_t, m3dcp_ti_t, m3dcp_vi_t, 0, 0, 0, 0), + M3D_CMDDEF(m3dc_nurbsu, "nurbsu", 3, m3dcp_va_t, m3dcp_vi_t, m3dcp_ti_t, 0, 0, 0, 0, 0), + M3D_CMDDEF(m3dc_nurbsn, "nurbsn", 3, m3dcp_va_t, m3dcp_vi_t, m3dcp_vi_t, 0, 0, 0, 0, 0), + M3D_CMDDEF(m3dc_nurbs, "nurbs", 2, m3dcp_va_t, m3dcp_vi_t, 0, 0, 0, 0, 0, 0), + M3D_CMDDEF(m3dc_conn, "conn", 6, m3dcp_i2_t, m3dcp_ti_t, m3dcp_i2_t, m3dcp_i2_t, m3dcp_ti_t, m3dcp_i2_t, 0, 0), + /* geometrical */ + M3D_CMDDEF(m3dc_line, "line", 2, m3dcp_va_t, m3dcp_vi_t, 0, 0, 0, 0, 0, 0), + M3D_CMDDEF(m3dc_polygon, "polygon", 2, m3dcp_va_t, m3dcp_vi_t, 0, 0, 0, 0, 0, 0), + M3D_CMDDEF(m3dc_circle, "circle", 3, m3dcp_vi_t, m3dcp_qi_t, m3dcp_vc_t, 0, 0, 0, 0, 0), + M3D_CMDDEF(m3dc_cylinder,"cylinder",6, m3dcp_vi_t, m3dcp_qi_t, m3dcp_vc_t, m3dcp_vi_t, m3dcp_qi_t, m3dcp_vc_t, 0, 0), + M3D_CMDDEF(m3dc_shpere, "shpere", 2, m3dcp_vi_t, m3dcp_vc_t, 0, 0, 0, 0, 0, 0), + M3D_CMDDEF(m3dc_torus, "torus", 4, m3dcp_vi_t, m3dcp_qi_t, m3dcp_vc_t, m3dcp_vc_t, 0, 0, 0, 0), + M3D_CMDDEF(m3dc_cone, "cone", 3, m3dcp_vi_t, m3dcp_vi_t, m3dcp_vi_t, 0, 0, 0, 0, 0), + M3D_CMDDEF(m3dc_cube, "cube", 3, m3dcp_vi_t, m3dcp_vi_t, m3dcp_vi_t, 0, 0, 0, 0, 0) +}; +#endif + +#include +#include + +/* we'll need this with M3D_NOTEXTURE */ +char *stbi_zlib_decode_malloc_guesssize_headerflag(const char *buffer, int len, int initial_size, int *outlen, int parse_header); + +#ifndef M3D_NOTEXTURE +#if !defined(M3D_NOIMPORTER) && !defined(STBI_INCLUDE_STB_IMAGE_H) +/* PNG decompressor from + + stb_image - v2.23 - public domain image loader - http://nothings.org/stb_image.h +*/ +static const char *_m3dstbi__g_failure_reason; + +enum +{ + STBI_default = 0, + + STBI_grey = 1, + STBI_grey_alpha = 2, + STBI_rgb = 3, + STBI_rgb_alpha = 4 +}; + +enum +{ + STBI__SCAN_load=0, + STBI__SCAN_type, + STBI__SCAN_header +}; + +typedef unsigned short _m3dstbi_us; + +typedef uint16_t _m3dstbi__uint16; +typedef int16_t _m3dstbi__int16; +typedef uint32_t _m3dstbi__uint32; +typedef int32_t _m3dstbi__int32; + +typedef struct +{ + _m3dstbi__uint32 img_x, img_y; + int img_n, img_out_n; + + void *io_user_data; + + int read_from_callbacks; + int buflen; + unsigned char buffer_start[128]; + + unsigned char *img_buffer, *img_buffer_end; + unsigned char *img_buffer_original, *img_buffer_original_end; +} _m3dstbi__context; + +typedef struct +{ + int bits_per_channel; + int num_channels; + int channel_order; +} _m3dstbi__result_info; + +#define STBI_ASSERT(v) +#define STBI_NOTUSED(v) (void)sizeof(v) +#define STBI__BYTECAST(x) ((unsigned char) ((x) & 255)) +#define STBI_MALLOC(sz) M3D_MALLOC(sz) +#define STBI_REALLOC(p,newsz) M3D_REALLOC(p,newsz) +#define STBI_FREE(p) M3D_FREE(p) +#define STBI_REALLOC_SIZED(p,oldsz,newsz) STBI_REALLOC(p,newsz) + +_inline static unsigned char _m3dstbi__get8(_m3dstbi__context *s) +{ + if (s->img_buffer < s->img_buffer_end) + return *s->img_buffer++; + return 0; +} + +_inline static int _m3dstbi__at_eof(_m3dstbi__context *s) +{ + return s->img_buffer >= s->img_buffer_end; +} + +static void _m3dstbi__skip(_m3dstbi__context *s, int n) +{ + if (n < 0) { + s->img_buffer = s->img_buffer_end; + return; + } + s->img_buffer += n; +} + +static int _m3dstbi__getn(_m3dstbi__context *s, unsigned char *buffer, int n) +{ + if (s->img_buffer+n <= s->img_buffer_end) { + memcpy(buffer, s->img_buffer, n); + s->img_buffer += n; + return 1; + } else + return 0; +} + +static int _m3dstbi__get16be(_m3dstbi__context *s) +{ + int z = _m3dstbi__get8(s); + return (z << 8) + _m3dstbi__get8(s); +} + +static _m3dstbi__uint32 _m3dstbi__get32be(_m3dstbi__context *s) +{ + _m3dstbi__uint32 z = _m3dstbi__get16be(s); + return (z << 16) + _m3dstbi__get16be(s); +} + +#define _m3dstbi__err(x,y) _m3dstbi__errstr(y) +static int _m3dstbi__errstr(const char *str) +{ + _m3dstbi__g_failure_reason = str; + return 0; +} + +_inline static void *_m3dstbi__malloc(size_t size) +{ + return STBI_MALLOC(size); +} + +static int _m3dstbi__addsizes_valid(int a, int b) +{ + if (b < 0) return 0; + return a <= 2147483647 - b; +} + +static int _m3dstbi__mul2sizes_valid(int a, int b) +{ + if (a < 0 || b < 0) return 0; + if (b == 0) return 1; + return a <= 2147483647/b; +} + +static int _m3dstbi__mad2sizes_valid(int a, int b, int add) +{ + return _m3dstbi__mul2sizes_valid(a, b) && _m3dstbi__addsizes_valid(a*b, add); +} + +static int _m3dstbi__mad3sizes_valid(int a, int b, int c, int add) +{ + return _m3dstbi__mul2sizes_valid(a, b) && _m3dstbi__mul2sizes_valid(a*b, c) && + _m3dstbi__addsizes_valid(a*b*c, add); +} + +static void *_m3dstbi__malloc_mad2(int a, int b, int add) +{ + if (!_m3dstbi__mad2sizes_valid(a, b, add)) return NULL; + return _m3dstbi__malloc(a*b + add); +} + +static void *_m3dstbi__malloc_mad3(int a, int b, int c, int add) +{ + if (!_m3dstbi__mad3sizes_valid(a, b, c, add)) return NULL; + return _m3dstbi__malloc(a*b*c + add); +} + +static unsigned char _m3dstbi__compute_y(int r, int g, int b) +{ + return (unsigned char) (((r*77) + (g*150) + (29*b)) >> 8); +} + +static unsigned char *_m3dstbi__convert_format(unsigned char *data, int img_n, int req_comp, unsigned int x, unsigned int y) +{ + int i,j; + unsigned char *good; + + if (req_comp == img_n) return data; + STBI_ASSERT(req_comp >= 1 && req_comp <= 4); + + good = (unsigned char *) _m3dstbi__malloc_mad3(req_comp, x, y, 0); + if (good == NULL) { + STBI_FREE(data); + _m3dstbi__err("outofmem", "Out of memory"); + return NULL; + } + + for (j=0; j < (int) y; ++j) { + unsigned char *src = data + j * x * img_n ; + unsigned char *dest = good + j * x * req_comp; + + #define STBI__COMBO(a,b) ((a)*8+(b)) + #define STBI__CASE(a,b) case STBI__COMBO(a,b): for(i=x-1; i >= 0; --i, src += a, dest += b) + switch (STBI__COMBO(img_n, req_comp)) { + STBI__CASE(1,2) { dest[0]=src[0], dest[1]=255; } break; + STBI__CASE(1,3) { dest[0]=dest[1]=dest[2]=src[0]; } break; + STBI__CASE(1,4) { dest[0]=dest[1]=dest[2]=src[0], dest[3]=255; } break; + STBI__CASE(2,1) { dest[0]=src[0]; } break; + STBI__CASE(2,3) { dest[0]=dest[1]=dest[2]=src[0]; } break; + STBI__CASE(2,4) { dest[0]=dest[1]=dest[2]=src[0], dest[3]=src[1]; } break; + STBI__CASE(3,4) { dest[0]=src[0],dest[1]=src[1],dest[2]=src[2],dest[3]=255; } break; + STBI__CASE(3,1) { dest[0]=_m3dstbi__compute_y(src[0],src[1],src[2]); } break; + STBI__CASE(3,2) { dest[0]=_m3dstbi__compute_y(src[0],src[1],src[2]), dest[1] = 255; } break; + STBI__CASE(4,1) { dest[0]=_m3dstbi__compute_y(src[0],src[1],src[2]); } break; + STBI__CASE(4,2) { dest[0]=_m3dstbi__compute_y(src[0],src[1],src[2]), dest[1] = src[3]; } break; + STBI__CASE(4,3) { dest[0]=src[0],dest[1]=src[1],dest[2]=src[2]; } break; + default: STBI_ASSERT(0); + } + #undef STBI__CASE + } + + STBI_FREE(data); + return good; +} + +static _m3dstbi__uint16 _m3dstbi__compute_y_16(int r, int g, int b) +{ + return (_m3dstbi__uint16) (((r*77) + (g*150) + (29*b)) >> 8); +} + +static _m3dstbi__uint16 *_m3dstbi__convert_format16(_m3dstbi__uint16 *data, int img_n, int req_comp, unsigned int x, unsigned int y) +{ + int i,j; + _m3dstbi__uint16 *good; + + if (req_comp == img_n) return data; + STBI_ASSERT(req_comp >= 1 && req_comp <= 4); + + good = (_m3dstbi__uint16 *) _m3dstbi__malloc(req_comp * x * y * 2); + if (good == NULL) { + STBI_FREE(data); + _m3dstbi__err("outofmem", "Out of memory"); + return NULL; + } + + for (j=0; j < (int) y; ++j) { + _m3dstbi__uint16 *src = data + j * x * img_n ; + _m3dstbi__uint16 *dest = good + j * x * req_comp; + + #define STBI__COMBO(a,b) ((a)*8+(b)) + #define STBI__CASE(a,b) case STBI__COMBO(a,b): for(i=x-1; i >= 0; --i, src += a, dest += b) + switch (STBI__COMBO(img_n, req_comp)) { + STBI__CASE(1,2) { dest[0]=src[0], dest[1]=0xffff; } break; + STBI__CASE(1,3) { dest[0]=dest[1]=dest[2]=src[0]; } break; + STBI__CASE(1,4) { dest[0]=dest[1]=dest[2]=src[0], dest[3]=0xffff; } break; + STBI__CASE(2,1) { dest[0]=src[0]; } break; + STBI__CASE(2,3) { dest[0]=dest[1]=dest[2]=src[0]; } break; + STBI__CASE(2,4) { dest[0]=dest[1]=dest[2]=src[0], dest[3]=src[1]; } break; + STBI__CASE(3,4) { dest[0]=src[0],dest[1]=src[1],dest[2]=src[2],dest[3]=0xffff; } break; + STBI__CASE(3,1) { dest[0]=_m3dstbi__compute_y_16(src[0],src[1],src[2]); } break; + STBI__CASE(3,2) { dest[0]=_m3dstbi__compute_y_16(src[0],src[1],src[2]), dest[1] = 0xffff; } break; + STBI__CASE(4,1) { dest[0]=_m3dstbi__compute_y_16(src[0],src[1],src[2]); } break; + STBI__CASE(4,2) { dest[0]=_m3dstbi__compute_y_16(src[0],src[1],src[2]), dest[1] = src[3]; } break; + STBI__CASE(4,3) { dest[0]=src[0],dest[1]=src[1],dest[2]=src[2]; } break; + default: STBI_ASSERT(0); + } + #undef STBI__CASE + } + + STBI_FREE(data); + return good; +} + +#define STBI__ZFAST_BITS 9 +#define STBI__ZFAST_MASK ((1 << STBI__ZFAST_BITS) - 1) + +typedef struct +{ + _m3dstbi__uint16 fast[1 << STBI__ZFAST_BITS]; + _m3dstbi__uint16 firstcode[16]; + int maxcode[17]; + _m3dstbi__uint16 firstsymbol[16]; + unsigned char size[288]; + _m3dstbi__uint16 value[288]; +} _m3dstbi__zhuffman; + +_inline static int _m3dstbi__bitreverse16(int n) +{ + n = ((n & 0xAAAA) >> 1) | ((n & 0x5555) << 1); + n = ((n & 0xCCCC) >> 2) | ((n & 0x3333) << 2); + n = ((n & 0xF0F0) >> 4) | ((n & 0x0F0F) << 4); + n = ((n & 0xFF00) >> 8) | ((n & 0x00FF) << 8); + return n; +} + +_inline static int _m3dstbi__bit_reverse(int v, int bits) +{ + STBI_ASSERT(bits <= 16); + return _m3dstbi__bitreverse16(v) >> (16-bits); +} + +static int _m3dstbi__zbuild_huffman(_m3dstbi__zhuffman *z, unsigned char *sizelist, int num) +{ + int i,k=0; + int code, next_code[16], sizes[17]; + + memset(sizes, 0, sizeof(sizes)); + memset(z->fast, 0, sizeof(z->fast)); + for (i=0; i < num; ++i) + ++sizes[sizelist[i]]; + sizes[0] = 0; + for (i=1; i < 16; ++i) + if (sizes[i] > (1 << i)) + return _m3dstbi__err("bad sizes", "Corrupt PNG"); + code = 0; + for (i=1; i < 16; ++i) { + next_code[i] = code; + z->firstcode[i] = (_m3dstbi__uint16) code; + z->firstsymbol[i] = (_m3dstbi__uint16) k; + code = (code + sizes[i]); + if (sizes[i]) + if (code-1 >= (1 << i)) return _m3dstbi__err("bad codelengths","Corrupt PNG"); + z->maxcode[i] = code << (16-i); + code <<= 1; + k += sizes[i]; + } + z->maxcode[16] = 0x10000; + for (i=0; i < num; ++i) { + int s = sizelist[i]; + if (s) { + int c = next_code[s] - z->firstcode[s] + z->firstsymbol[s]; + _m3dstbi__uint16 fastv = (_m3dstbi__uint16) ((s << 9) | i); + z->size [c] = (unsigned char ) s; + z->value[c] = (_m3dstbi__uint16) i; + if (s <= STBI__ZFAST_BITS) { + int j = _m3dstbi__bit_reverse(next_code[s],s); + while (j < (1 << STBI__ZFAST_BITS)) { + z->fast[j] = fastv; + j += (1 << s); + } + } + ++next_code[s]; + } + } + return 1; +} + +typedef struct +{ + unsigned char *zbuffer, *zbuffer_end; + int num_bits; + _m3dstbi__uint32 code_buffer; + + char *zout; + char *zout_start; + char *zout_end; + int z_expandable; + + _m3dstbi__zhuffman z_length, z_distance; +} _m3dstbi__zbuf; + +_inline static unsigned char _m3dstbi__zget8(_m3dstbi__zbuf *z) +{ + if (z->zbuffer >= z->zbuffer_end) return 0; + return *z->zbuffer++; +} + +static void _m3dstbi__fill_bits(_m3dstbi__zbuf *z) +{ + do { + STBI_ASSERT(z->code_buffer < (1U << z->num_bits)); + z->code_buffer |= (unsigned int) _m3dstbi__zget8(z) << z->num_bits; + z->num_bits += 8; + } while (z->num_bits <= 24); +} + +_inline static unsigned int _m3dstbi__zreceive(_m3dstbi__zbuf *z, int n) +{ + unsigned int k; + if (z->num_bits < n) _m3dstbi__fill_bits(z); + k = z->code_buffer & ((1 << n) - 1); + z->code_buffer >>= n; + z->num_bits -= n; + return k; +} + +static int _m3dstbi__zhuffman_decode_slowpath(_m3dstbi__zbuf *a, _m3dstbi__zhuffman *z) +{ + int b,s,k; + k = _m3dstbi__bit_reverse(a->code_buffer, 16); + for (s=STBI__ZFAST_BITS+1; ; ++s) + if (k < z->maxcode[s]) + break; + if (s == 16) return -1; + b = (k >> (16-s)) - z->firstcode[s] + z->firstsymbol[s]; + STBI_ASSERT(z->size[b] == s); + a->code_buffer >>= s; + a->num_bits -= s; + return z->value[b]; +} + +_inline static int _m3dstbi__zhuffman_decode(_m3dstbi__zbuf *a, _m3dstbi__zhuffman *z) +{ + int b,s; + if (a->num_bits < 16) _m3dstbi__fill_bits(a); + b = z->fast[a->code_buffer & STBI__ZFAST_MASK]; + if (b) { + s = b >> 9; + a->code_buffer >>= s; + a->num_bits -= s; + return b & 511; + } + return _m3dstbi__zhuffman_decode_slowpath(a, z); +} + +static int _m3dstbi__zexpand(_m3dstbi__zbuf *z, char *zout, int n) +{ + char *q; + int cur, limit, old_limit; + z->zout = zout; + if (!z->z_expandable) return _m3dstbi__err("output buffer limit","Corrupt PNG"); + cur = (int) (z->zout - z->zout_start); + limit = old_limit = (int) (z->zout_end - z->zout_start); + while (cur + n > limit) + limit *= 2; + q = (char *) STBI_REALLOC_SIZED(z->zout_start, old_limit, limit); + STBI_NOTUSED(old_limit); + if (q == NULL) return _m3dstbi__err("outofmem", "Out of memory"); + z->zout_start = q; + z->zout = q + cur; + z->zout_end = q + limit; + return 1; +} + +static int _m3dstbi__zlength_base[31] = { + 3,4,5,6,7,8,9,10,11,13, + 15,17,19,23,27,31,35,43,51,59, + 67,83,99,115,131,163,195,227,258,0,0 }; + +static int _m3dstbi__zlength_extra[31]= +{ 0,0,0,0,0,0,0,0,1,1,1,1,2,2,2,2,3,3,3,3,4,4,4,4,5,5,5,5,0,0,0 }; + +static int _m3dstbi__zdist_base[32] = { 1,2,3,4,5,7,9,13,17,25,33,49,65,97,129,193, +257,385,513,769,1025,1537,2049,3073,4097,6145,8193,12289,16385,24577,0,0}; + +static int _m3dstbi__zdist_extra[32] = +{ 0,0,0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,11,11,12,12,13,13}; + +static int _m3dstbi__parse_huffman_block(_m3dstbi__zbuf *a) +{ + char *zout = a->zout; + for(;;) { + int z = _m3dstbi__zhuffman_decode(a, &a->z_length); + if (z < 256) { + if (z < 0) return _m3dstbi__err("bad huffman code","Corrupt PNG"); + if (zout >= a->zout_end) { + if (!_m3dstbi__zexpand(a, zout, 1)) return 0; + zout = a->zout; + } + *zout++ = (char) z; + } else { + unsigned char *p; + int len,dist; + if (z == 256) { + a->zout = zout; + return 1; + } + z -= 257; + len = _m3dstbi__zlength_base[z]; + if (_m3dstbi__zlength_extra[z]) len += _m3dstbi__zreceive(a, _m3dstbi__zlength_extra[z]); + z = _m3dstbi__zhuffman_decode(a, &a->z_distance); + if (z < 0) return _m3dstbi__err("bad huffman code","Corrupt PNG"); + dist = _m3dstbi__zdist_base[z]; + if (_m3dstbi__zdist_extra[z]) dist += _m3dstbi__zreceive(a, _m3dstbi__zdist_extra[z]); + if (zout - a->zout_start < dist) return _m3dstbi__err("bad dist","Corrupt PNG"); + if (zout + len > a->zout_end) { + if (!_m3dstbi__zexpand(a, zout, len)) return 0; + zout = a->zout; + } + p = (unsigned char *) (zout - dist); + if (dist == 1) { + unsigned char v = *p; + if (len) { do *zout++ = v; while (--len); } + } else { + if (len) { do *zout++ = *p++; while (--len); } + } + } + } +} + +static int _m3dstbi__compute_huffman_codes(_m3dstbi__zbuf *a) +{ + static unsigned char length_dezigzag[19] = { 16,17,18,0,8,7,9,6,10,5,11,4,12,3,13,2,14,1,15 }; + _m3dstbi__zhuffman z_codelength; + unsigned char lencodes[286+32+137]; + unsigned char codelength_sizes[19]; + int i,n; + + int hlit = _m3dstbi__zreceive(a,5) + 257; + int hdist = _m3dstbi__zreceive(a,5) + 1; + int hclen = _m3dstbi__zreceive(a,4) + 4; + int ntot = hlit + hdist; + + memset(codelength_sizes, 0, sizeof(codelength_sizes)); + for (i=0; i < hclen; ++i) { + int s = _m3dstbi__zreceive(a,3); + codelength_sizes[length_dezigzag[i]] = (unsigned char) s; + } + if (!_m3dstbi__zbuild_huffman(&z_codelength, codelength_sizes, 19)) return 0; + + n = 0; + while (n < ntot) { + int c = _m3dstbi__zhuffman_decode(a, &z_codelength); + if (c < 0 || c >= 19) return _m3dstbi__err("bad codelengths", "Corrupt PNG"); + if (c < 16) + lencodes[n++] = (unsigned char) c; + else { + unsigned char fill = 0; + if (c == 16) { + c = _m3dstbi__zreceive(a,2)+3; + if (n == 0) return _m3dstbi__err("bad codelengths", "Corrupt PNG"); + fill = lencodes[n-1]; + } else if (c == 17) + c = _m3dstbi__zreceive(a,3)+3; + else { + STBI_ASSERT(c == 18); + c = _m3dstbi__zreceive(a,7)+11; + } + if (ntot - n < c) return _m3dstbi__err("bad codelengths", "Corrupt PNG"); + memset(lencodes+n, fill, c); + n += c; + } + } + if (n != ntot) return _m3dstbi__err("bad codelengths","Corrupt PNG"); + if (!_m3dstbi__zbuild_huffman(&a->z_length, lencodes, hlit)) return 0; + if (!_m3dstbi__zbuild_huffman(&a->z_distance, lencodes+hlit, hdist)) return 0; + return 1; +} + +_inline static int _m3dstbi__parse_uncompressed_block(_m3dstbi__zbuf *a) +{ + unsigned char header[4]; + int len,nlen,k; + if (a->num_bits & 7) + _m3dstbi__zreceive(a, a->num_bits & 7); + k = 0; + while (a->num_bits > 0) { + header[k++] = (unsigned char) (a->code_buffer & 255); + a->code_buffer >>= 8; + a->num_bits -= 8; + } + STBI_ASSERT(a->num_bits == 0); + while (k < 4) + header[k++] = _m3dstbi__zget8(a); + len = header[1] * 256 + header[0]; + nlen = header[3] * 256 + header[2]; + if (nlen != (len ^ 0xffff)) return _m3dstbi__err("zlib corrupt","Corrupt PNG"); + if (a->zbuffer + len > a->zbuffer_end) return _m3dstbi__err("read past buffer","Corrupt PNG"); + if (a->zout + len > a->zout_end) + if (!_m3dstbi__zexpand(a, a->zout, len)) return 0; + memcpy(a->zout, a->zbuffer, len); + a->zbuffer += len; + a->zout += len; + return 1; +} + +static int _m3dstbi__parse_zlib_header(_m3dstbi__zbuf *a) +{ + int cmf = _m3dstbi__zget8(a); + int cm = cmf & 15; + /* int cinfo = cmf >> 4; */ + int flg = _m3dstbi__zget8(a); + if ((cmf*256+flg) % 31 != 0) return _m3dstbi__err("bad zlib header","Corrupt PNG"); + if (flg & 32) return _m3dstbi__err("no preset dict","Corrupt PNG"); + if (cm != 8) return _m3dstbi__err("bad compression","Corrupt PNG"); + return 1; +} + +static unsigned char _m3dstbi__zdefault_length[288], _m3dstbi__zdefault_distance[32]; +static void _m3dstbi__init_zdefaults(void) +{ + int i; + for (i=0; i <= 143; ++i) _m3dstbi__zdefault_length[i] = 8; + for ( ; i <= 255; ++i) _m3dstbi__zdefault_length[i] = 9; + for ( ; i <= 279; ++i) _m3dstbi__zdefault_length[i] = 7; + for ( ; i <= 287; ++i) _m3dstbi__zdefault_length[i] = 8; + + for (i=0; i <= 31; ++i) _m3dstbi__zdefault_distance[i] = 5; +} + +static int _m3dstbi__parse_zlib(_m3dstbi__zbuf *a, int parse_header) +{ + int final, type; + if (parse_header) + if (!_m3dstbi__parse_zlib_header(a)) return 0; + a->num_bits = 0; + a->code_buffer = 0; + do { + final = _m3dstbi__zreceive(a,1); + type = _m3dstbi__zreceive(a,2); + if (type == 0) { + if (!_m3dstbi__parse_uncompressed_block(a)) return 0; + } else if (type == 3) { + return 0; + } else { + if (type == 1) { + if (!_m3dstbi__zbuild_huffman(&a->z_length , _m3dstbi__zdefault_length , 288)) return 0; + if (!_m3dstbi__zbuild_huffman(&a->z_distance, _m3dstbi__zdefault_distance, 32)) return 0; + } else { + if (!_m3dstbi__compute_huffman_codes(a)) return 0; + } + if (!_m3dstbi__parse_huffman_block(a)) return 0; + } + } while (!final); + return 1; +} + +static int _m3dstbi__do_zlib(_m3dstbi__zbuf *a, char *obuf, int olen, int exp, int parse_header) +{ + a->zout_start = obuf; + a->zout = obuf; + a->zout_end = obuf + olen; + a->z_expandable = exp; + _m3dstbi__init_zdefaults(); + return _m3dstbi__parse_zlib(a, parse_header); +} + +char *_m3dstbi_zlib_decode_malloc_guesssize_headerflag(const char *buffer, int len, int initial_size, int *outlen, int parse_header) +{ + _m3dstbi__zbuf a; + char *p = (char *) _m3dstbi__malloc(initial_size); + if (p == NULL) return NULL; + a.zbuffer = (unsigned char *) buffer; + a.zbuffer_end = (unsigned char *) buffer + len; + if (_m3dstbi__do_zlib(&a, p, initial_size, 1, parse_header)) { + if (outlen) *outlen = (int) (a.zout - a.zout_start); + return a.zout_start; + } else { + STBI_FREE(a.zout_start); + return NULL; + } +} + +typedef struct +{ + _m3dstbi__uint32 length; + _m3dstbi__uint32 type; +} _m3dstbi__pngchunk; + +static _m3dstbi__pngchunk _m3dstbi__get_chunk_header(_m3dstbi__context *s) +{ + _m3dstbi__pngchunk c; + c.length = _m3dstbi__get32be(s); + c.type = _m3dstbi__get32be(s); + return c; +} + +_inline static int _m3dstbi__check_png_header(_m3dstbi__context *s) +{ + static unsigned char png_sig[8] = { 137,80,78,71,13,10,26,10 }; + int i; + for (i=0; i < 8; ++i) + if (_m3dstbi__get8(s) != png_sig[i]) return _m3dstbi__err("bad png sig","Not a PNG"); + return 1; +} + +typedef struct +{ + _m3dstbi__context *s; + unsigned char *idata, *expanded, *out; + int depth; +} _m3dstbi__png; + + +enum { + STBI__F_none=0, + STBI__F_sub=1, + STBI__F_up=2, + STBI__F_avg=3, + STBI__F_paeth=4, + STBI__F_avg_first, + STBI__F_paeth_first +}; + +static unsigned char first_row_filter[5] = +{ + STBI__F_none, + STBI__F_sub, + STBI__F_none, + STBI__F_avg_first, + STBI__F_paeth_first +}; + +static int _m3dstbi__paeth(int a, int b, int c) +{ + int p = a + b - c; + int pa = abs(p-a); + int pb = abs(p-b); + int pc = abs(p-c); + if (pa <= pb && pa <= pc) return a; + if (pb <= pc) return b; + return c; +} + +static unsigned char _m3dstbi__depth_scale_table[9] = { 0, 0xff, 0x55, 0, 0x11, 0,0,0, 0x01 }; + +static int _m3dstbi__create_png_image_raw(_m3dstbi__png *a, unsigned char *raw, _m3dstbi__uint32 raw_len, int out_n, _m3dstbi__uint32 x, _m3dstbi__uint32 y, int depth, int color) +{ + int bytes = (depth == 16? 2 : 1); + _m3dstbi__context *s = a->s; + _m3dstbi__uint32 i,j,stride = x*out_n*bytes; + _m3dstbi__uint32 img_len, img_width_bytes; + int k; + int img_n = s->img_n; + + int output_bytes = out_n*bytes; + int filter_bytes = img_n*bytes; + int width = x; + + STBI_ASSERT(out_n == s->img_n || out_n == s->img_n+1); + a->out = (unsigned char *) _m3dstbi__malloc_mad3(x, y, output_bytes, 0); + if (!a->out) return _m3dstbi__err("outofmem", "Out of memory"); + + if (!_m3dstbi__mad3sizes_valid(img_n, x, depth, 7)) return _m3dstbi__err("too large", "Corrupt PNG"); + img_width_bytes = (((img_n * x * depth) + 7) >> 3); + img_len = (img_width_bytes + 1) * y; + if (s->img_x == x && s->img_y == y) { + if (raw_len != img_len) return _m3dstbi__err("not enough pixels","Corrupt PNG"); + } else { + if (raw_len < img_len) return _m3dstbi__err("not enough pixels","Corrupt PNG"); + } + + for (j=0; j < y; ++j) { + unsigned char *cur = a->out + stride*j; + unsigned char *prior = cur - stride; + int filter = *raw++; + + if (filter > 4) + return _m3dstbi__err("invalid filter","Corrupt PNG"); + + if (depth < 8) { + STBI_ASSERT(img_width_bytes <= x); + cur += x*out_n - img_width_bytes; + filter_bytes = 1; + width = img_width_bytes; + } + prior = cur - stride; + + if (j == 0) filter = first_row_filter[filter]; + + for (k=0; k < filter_bytes; ++k) { + switch (filter) { + case STBI__F_none : cur[k] = raw[k]; break; + case STBI__F_sub : cur[k] = raw[k]; break; + case STBI__F_up : cur[k] = STBI__BYTECAST(raw[k] + prior[k]); break; + case STBI__F_avg : cur[k] = STBI__BYTECAST(raw[k] + (prior[k]>>1)); break; + case STBI__F_paeth : cur[k] = STBI__BYTECAST(raw[k] + _m3dstbi__paeth(0,prior[k],0)); break; + case STBI__F_avg_first : cur[k] = raw[k]; break; + case STBI__F_paeth_first: cur[k] = raw[k]; break; + } + } + + if (depth == 8) { + if (img_n != out_n) + cur[img_n] = 255; + raw += img_n; + cur += out_n; + prior += out_n; + } else if (depth == 16) { + if (img_n != out_n) { + cur[filter_bytes] = 255; + cur[filter_bytes+1] = 255; + } + raw += filter_bytes; + cur += output_bytes; + prior += output_bytes; + } else { + raw += 1; + cur += 1; + prior += 1; + } + + if (depth < 8 || img_n == out_n) { + int nk = (width - 1)*filter_bytes; + #define STBI__CASE(f) \ + case f: \ + for (k=0; k < nk; ++k) + switch (filter) { + case STBI__F_none: memcpy(cur, raw, nk); break; + STBI__CASE(STBI__F_sub) { cur[k] = STBI__BYTECAST(raw[k] + cur[k-filter_bytes]); } break; + STBI__CASE(STBI__F_up) { cur[k] = STBI__BYTECAST(raw[k] + prior[k]); } break; + STBI__CASE(STBI__F_avg) { cur[k] = STBI__BYTECAST(raw[k] + ((prior[k] + cur[k-filter_bytes])>>1)); } break; + STBI__CASE(STBI__F_paeth) { cur[k] = STBI__BYTECAST(raw[k] + _m3dstbi__paeth(cur[k-filter_bytes],prior[k],prior[k-filter_bytes])); } break; + STBI__CASE(STBI__F_avg_first) { cur[k] = STBI__BYTECAST(raw[k] + (cur[k-filter_bytes] >> 1)); } break; + STBI__CASE(STBI__F_paeth_first) { cur[k] = STBI__BYTECAST(raw[k] + _m3dstbi__paeth(cur[k-filter_bytes],0,0)); } break; + } + #undef STBI__CASE + raw += nk; + } else { + STBI_ASSERT(img_n+1 == out_n); + #define STBI__CASE(f) \ + case f: \ + for (i=x-1; i >= 1; --i, cur[filter_bytes]=255,raw+=filter_bytes,cur+=output_bytes,prior+=output_bytes) \ + for (k=0; k < filter_bytes; ++k) + switch (filter) { + STBI__CASE(STBI__F_none) { cur[k] = raw[k]; } break; + STBI__CASE(STBI__F_sub) { cur[k] = STBI__BYTECAST(raw[k] + cur[k- output_bytes]); } break; + STBI__CASE(STBI__F_up) { cur[k] = STBI__BYTECAST(raw[k] + prior[k]); } break; + STBI__CASE(STBI__F_avg) { cur[k] = STBI__BYTECAST(raw[k] + ((prior[k] + cur[k- output_bytes])>>1)); } break; + STBI__CASE(STBI__F_paeth) { cur[k] = STBI__BYTECAST(raw[k] + _m3dstbi__paeth(cur[k- output_bytes],prior[k],prior[k- output_bytes])); } break; + STBI__CASE(STBI__F_avg_first) { cur[k] = STBI__BYTECAST(raw[k] + (cur[k- output_bytes] >> 1)); } break; + STBI__CASE(STBI__F_paeth_first) { cur[k] = STBI__BYTECAST(raw[k] + _m3dstbi__paeth(cur[k- output_bytes],0,0)); } break; + } + #undef STBI__CASE + + if (depth == 16) { + cur = a->out + stride*j; + for (i=0; i < x; ++i,cur+=output_bytes) { + cur[filter_bytes+1] = 255; + } + } + } + } + + if (depth < 8) { + for (j=0; j < y; ++j) { + unsigned char *cur = a->out + stride*j; + unsigned char *in = a->out + stride*j + x*out_n - img_width_bytes; + unsigned char scale = (color == 0) ? _m3dstbi__depth_scale_table[depth] : 1; + + if (depth == 4) { + for (k=x*img_n; k >= 2; k-=2, ++in) { + *cur++ = scale * ((*in >> 4) ); + *cur++ = scale * ((*in ) & 0x0f); + } + if (k > 0) *cur++ = scale * ((*in >> 4) ); + } else if (depth == 2) { + for (k=x*img_n; k >= 4; k-=4, ++in) { + *cur++ = scale * ((*in >> 6) ); + *cur++ = scale * ((*in >> 4) & 0x03); + *cur++ = scale * ((*in >> 2) & 0x03); + *cur++ = scale * ((*in ) & 0x03); + } + if (k > 0) *cur++ = scale * ((*in >> 6) ); + if (k > 1) *cur++ = scale * ((*in >> 4) & 0x03); + if (k > 2) *cur++ = scale * ((*in >> 2) & 0x03); + } else if (depth == 1) { + for (k=x*img_n; k >= 8; k-=8, ++in) { + *cur++ = scale * ((*in >> 7) ); + *cur++ = scale * ((*in >> 6) & 0x01); + *cur++ = scale * ((*in >> 5) & 0x01); + *cur++ = scale * ((*in >> 4) & 0x01); + *cur++ = scale * ((*in >> 3) & 0x01); + *cur++ = scale * ((*in >> 2) & 0x01); + *cur++ = scale * ((*in >> 1) & 0x01); + *cur++ = scale * ((*in ) & 0x01); + } + if (k > 0) *cur++ = scale * ((*in >> 7) ); + if (k > 1) *cur++ = scale * ((*in >> 6) & 0x01); + if (k > 2) *cur++ = scale * ((*in >> 5) & 0x01); + if (k > 3) *cur++ = scale * ((*in >> 4) & 0x01); + if (k > 4) *cur++ = scale * ((*in >> 3) & 0x01); + if (k > 5) *cur++ = scale * ((*in >> 2) & 0x01); + if (k > 6) *cur++ = scale * ((*in >> 1) & 0x01); + } + if (img_n != out_n) { + int q; + cur = a->out + stride*j; + if (img_n == 1) { + for (q=x-1; q >= 0; --q) { + cur[q*2+1] = 255; + cur[q*2+0] = cur[q]; + } + } else { + STBI_ASSERT(img_n == 3); + for (q=x-1; q >= 0; --q) { + cur[q*4+3] = 255; + cur[q*4+2] = cur[q*3+2]; + cur[q*4+1] = cur[q*3+1]; + cur[q*4+0] = cur[q*3+0]; + } + } + } + } + } else if (depth == 16) { + unsigned char *cur = a->out; + _m3dstbi__uint16 *cur16 = (_m3dstbi__uint16*)cur; + + for(i=0; i < x*y*out_n; ++i,cur16++,cur+=2) { + *cur16 = (cur[0] << 8) | cur[1]; + } + } + + return 1; +} + +static int _m3dstbi__create_png_image(_m3dstbi__png *a, unsigned char *image_data, _m3dstbi__uint32 image_data_len, int out_n, int depth, int color, int interlaced) +{ + int bytes = (depth == 16 ? 2 : 1); + int out_bytes = out_n * bytes; + unsigned char *final; + int p; + if (!interlaced) + return _m3dstbi__create_png_image_raw(a, image_data, image_data_len, out_n, a->s->img_x, a->s->img_y, depth, color); + + final = (unsigned char *) _m3dstbi__malloc_mad3(a->s->img_x, a->s->img_y, out_bytes, 0); + for (p=0; p < 7; ++p) { + int xorig[] = { 0,4,0,2,0,1,0 }; + int yorig[] = { 0,0,4,0,2,0,1 }; + int xspc[] = { 8,8,4,4,2,2,1 }; + int yspc[] = { 8,8,8,4,4,2,2 }; + int i,j,x,y; + x = (a->s->img_x - xorig[p] + xspc[p]-1) / xspc[p]; + y = (a->s->img_y - yorig[p] + yspc[p]-1) / yspc[p]; + if (x && y) { + _m3dstbi__uint32 img_len = ((((a->s->img_n * x * depth) + 7) >> 3) + 1) * y; + if (!_m3dstbi__create_png_image_raw(a, image_data, image_data_len, out_n, x, y, depth, color)) { + STBI_FREE(final); + return 0; + } + for (j=0; j < y; ++j) { + for (i=0; i < x; ++i) { + int out_y = j*yspc[p]+yorig[p]; + int out_x = i*xspc[p]+xorig[p]; + memcpy(final + out_y*a->s->img_x*out_bytes + out_x*out_bytes, + a->out + (j*x+i)*out_bytes, out_bytes); + } + } + STBI_FREE(a->out); + image_data += img_len; + image_data_len -= img_len; + } + } + a->out = final; + + return 1; +} + +static int _m3dstbi__compute_transparency(_m3dstbi__png *z, unsigned char tc[3], int out_n) +{ + _m3dstbi__context *s = z->s; + _m3dstbi__uint32 i, pixel_count = s->img_x * s->img_y; + unsigned char *p = z->out; + + STBI_ASSERT(out_n == 2 || out_n == 4); + + if (out_n == 2) { + for (i=0; i < pixel_count; ++i) { + p[1] = (p[0] == tc[0] ? 0 : 255); + p += 2; + } + } else { + for (i=0; i < pixel_count; ++i) { + if (p[0] == tc[0] && p[1] == tc[1] && p[2] == tc[2]) + p[3] = 0; + p += 4; + } + } + return 1; +} + +static int _m3dstbi__compute_transparency16(_m3dstbi__png *z, _m3dstbi__uint16 tc[3], int out_n) +{ + _m3dstbi__context *s = z->s; + _m3dstbi__uint32 i, pixel_count = s->img_x * s->img_y; + _m3dstbi__uint16 *p = (_m3dstbi__uint16*) z->out; + + STBI_ASSERT(out_n == 2 || out_n == 4); + + if (out_n == 2) { + for (i = 0; i < pixel_count; ++i) { + p[1] = (p[0] == tc[0] ? 0 : 65535); + p += 2; + } + } else { + for (i = 0; i < pixel_count; ++i) { + if (p[0] == tc[0] && p[1] == tc[1] && p[2] == tc[2]) + p[3] = 0; + p += 4; + } + } + return 1; +} + +static int _m3dstbi__expand_png_palette(_m3dstbi__png *a, unsigned char *palette, int len, int pal_img_n) +{ + _m3dstbi__uint32 i, pixel_count = a->s->img_x * a->s->img_y; + unsigned char *p, *temp_out, *orig = a->out; + + p = (unsigned char *) _m3dstbi__malloc_mad2(pixel_count, pal_img_n, 0); + if (p == NULL) return _m3dstbi__err("outofmem", "Out of memory"); + + temp_out = p; + + if (pal_img_n == 3) { + for (i=0; i < pixel_count; ++i) { + int n = orig[i]*4; + p[0] = palette[n ]; + p[1] = palette[n+1]; + p[2] = palette[n+2]; + p += 3; + } + } else { + for (i=0; i < pixel_count; ++i) { + int n = orig[i]*4; + p[0] = palette[n ]; + p[1] = palette[n+1]; + p[2] = palette[n+2]; + p[3] = palette[n+3]; + p += 4; + } + } + STBI_FREE(a->out); + a->out = temp_out; + + STBI_NOTUSED(len); + + return 1; +} + +#define STBI__PNG_TYPE(a,b,c,d) (((unsigned) (a) << 24) + ((unsigned) (b) << 16) + ((unsigned) (c) << 8) + (unsigned) (d)) + +static int _m3dstbi__parse_png_file(_m3dstbi__png *z, int scan, int req_comp) +{ + unsigned char palette[1024], pal_img_n=0; + unsigned char has_trans=0, tc[3]; + _m3dstbi__uint16 tc16[3]; + _m3dstbi__uint32 ioff=0, idata_limit=0, i, pal_len=0; + int first=1,k,interlace=0, color=0; + _m3dstbi__context *s = z->s; + + z->expanded = NULL; + z->idata = NULL; + z->out = NULL; + + if (!_m3dstbi__check_png_header(s)) return 0; + + if (scan == STBI__SCAN_type) return 1; + + for (;;) { + _m3dstbi__pngchunk c = _m3dstbi__get_chunk_header(s); + switch (c.type) { + case STBI__PNG_TYPE('C','g','B','I'): + _m3dstbi__skip(s, c.length); + break; + case STBI__PNG_TYPE('I','H','D','R'): { + int comp,filter; + if (!first) return _m3dstbi__err("multiple IHDR","Corrupt PNG"); + first = 0; + if (c.length != 13) return _m3dstbi__err("bad IHDR len","Corrupt PNG"); + s->img_x = _m3dstbi__get32be(s); if (s->img_x > (1 << 24)) return _m3dstbi__err("too large","Very large image (corrupt?)"); + s->img_y = _m3dstbi__get32be(s); if (s->img_y > (1 << 24)) return _m3dstbi__err("too large","Very large image (corrupt?)"); + z->depth = _m3dstbi__get8(s); if (z->depth != 1 && z->depth != 2 && z->depth != 4 && z->depth != 8 && z->depth != 16) return _m3dstbi__err("1/2/4/8/16-bit only","PNG not supported: 1/2/4/8/16-bit only"); + color = _m3dstbi__get8(s); if (color > 6) return _m3dstbi__err("bad ctype","Corrupt PNG"); + if (color == 3 && z->depth == 16) return _m3dstbi__err("bad ctype","Corrupt PNG"); + if (color == 3) pal_img_n = 3; else if (color & 1) return _m3dstbi__err("bad ctype","Corrupt PNG"); + comp = _m3dstbi__get8(s); if (comp) return _m3dstbi__err("bad comp method","Corrupt PNG"); + filter= _m3dstbi__get8(s); if (filter) return _m3dstbi__err("bad filter method","Corrupt PNG"); + interlace = _m3dstbi__get8(s); if (interlace>1) return _m3dstbi__err("bad interlace method","Corrupt PNG"); + if (!s->img_x || !s->img_y) return _m3dstbi__err("0-pixel image","Corrupt PNG"); + if (!pal_img_n) { + s->img_n = (color & 2 ? 3 : 1) + (color & 4 ? 1 : 0); + if ((1 << 30) / s->img_x / s->img_n < s->img_y) return _m3dstbi__err("too large", "Image too large to decode"); + if (scan == STBI__SCAN_header) return 1; + } else { + s->img_n = 1; + if ((1 << 30) / s->img_x / 4 < s->img_y) return _m3dstbi__err("too large","Corrupt PNG"); + } + break; + } + + case STBI__PNG_TYPE('P','L','T','E'): { + if (first) return _m3dstbi__err("first not IHDR", "Corrupt PNG"); + if (c.length > 256*3) return _m3dstbi__err("invalid PLTE","Corrupt PNG"); + pal_len = c.length / 3; + if (pal_len * 3 != c.length) return _m3dstbi__err("invalid PLTE","Corrupt PNG"); + for (i=0; i < pal_len; ++i) { + palette[i*4+0] = _m3dstbi__get8(s); + palette[i*4+1] = _m3dstbi__get8(s); + palette[i*4+2] = _m3dstbi__get8(s); + palette[i*4+3] = 255; + } + break; + } + + case STBI__PNG_TYPE('t','R','N','S'): { + if (first) return _m3dstbi__err("first not IHDR", "Corrupt PNG"); + if (z->idata) return _m3dstbi__err("tRNS after IDAT","Corrupt PNG"); + if (pal_img_n) { + if (scan == STBI__SCAN_header) { s->img_n = 4; return 1; } + if (pal_len == 0) return _m3dstbi__err("tRNS before PLTE","Corrupt PNG"); + if (c.length > pal_len) return _m3dstbi__err("bad tRNS len","Corrupt PNG"); + pal_img_n = 4; + for (i=0; i < c.length; ++i) + palette[i*4+3] = _m3dstbi__get8(s); + } else { + if (!(s->img_n & 1)) return _m3dstbi__err("tRNS with alpha","Corrupt PNG"); + if (c.length != (_m3dstbi__uint32) s->img_n*2) return _m3dstbi__err("bad tRNS len","Corrupt PNG"); + has_trans = 1; + if (z->depth == 16) { + for (k = 0; k < s->img_n; ++k) tc16[k] = (_m3dstbi__uint16)_m3dstbi__get16be(s); + } else { + for (k = 0; k < s->img_n; ++k) tc[k] = (unsigned char)(_m3dstbi__get16be(s) & 255) * _m3dstbi__depth_scale_table[z->depth]; + } + } + break; + } + + case STBI__PNG_TYPE('I','D','A','T'): { + if (first) return _m3dstbi__err("first not IHDR", "Corrupt PNG"); + if (pal_img_n && !pal_len) return _m3dstbi__err("no PLTE","Corrupt PNG"); + if (scan == STBI__SCAN_header) { s->img_n = pal_img_n; return 1; } + if ((int)(ioff + c.length) < (int)ioff) return 0; + if (ioff + c.length > idata_limit) { + _m3dstbi__uint32 idata_limit_old = idata_limit; + unsigned char *p; + if (idata_limit == 0) idata_limit = c.length > 4096 ? c.length : 4096; + while (ioff + c.length > idata_limit) + idata_limit *= 2; + STBI_NOTUSED(idata_limit_old); + p = (unsigned char *) STBI_REALLOC_SIZED(z->idata, idata_limit_old, idata_limit); if (p == NULL) return _m3dstbi__err("outofmem", "Out of memory"); + z->idata = p; + } + if (!_m3dstbi__getn(s, z->idata+ioff,c.length)) return _m3dstbi__err("outofdata","Corrupt PNG"); + ioff += c.length; + break; + } + + case STBI__PNG_TYPE('I','E','N','D'): { + _m3dstbi__uint32 raw_len, bpl; + if (first) return _m3dstbi__err("first not IHDR", "Corrupt PNG"); + if (scan != STBI__SCAN_load) return 1; + if (z->idata == NULL) return _m3dstbi__err("no IDAT","Corrupt PNG"); + bpl = (s->img_x * z->depth + 7) / 8; + raw_len = bpl * s->img_y * s->img_n /* pixels */ + s->img_y /* filter mode per row */; + z->expanded = (unsigned char *) _m3dstbi_zlib_decode_malloc_guesssize_headerflag((char *) z->idata, ioff, raw_len, (int *) &raw_len, 1); + if (z->expanded == NULL) return 0; + STBI_FREE(z->idata); z->idata = NULL; + if ((req_comp == s->img_n+1 && req_comp != 3 && !pal_img_n) || has_trans) + s->img_out_n = s->img_n+1; + else + s->img_out_n = s->img_n; + if (!_m3dstbi__create_png_image(z, z->expanded, raw_len, s->img_out_n, z->depth, color, interlace)) return 0; + if (has_trans) { + if (z->depth == 16) { + if (!_m3dstbi__compute_transparency16(z, tc16, s->img_out_n)) return 0; + } else { + if (!_m3dstbi__compute_transparency(z, tc, s->img_out_n)) return 0; + } + } + if (pal_img_n) { + s->img_n = pal_img_n; + s->img_out_n = pal_img_n; + if (req_comp >= 3) s->img_out_n = req_comp; + if (!_m3dstbi__expand_png_palette(z, palette, pal_len, s->img_out_n)) + return 0; + } else if (has_trans) { + ++s->img_n; + } + STBI_FREE(z->expanded); z->expanded = NULL; + return 1; + } + + default: + if (first) return _m3dstbi__err("first not IHDR", "Corrupt PNG"); + if ((c.type & (1 << 29)) == 0) { + return _m3dstbi__err("invalid_chunk", "PNG not supported: unknown PNG chunk type"); + } + _m3dstbi__skip(s, c.length); + break; + } + _m3dstbi__get32be(s); + } +} + +static void *_m3dstbi__do_png(_m3dstbi__png *p, int *x, int *y, int *n, int req_comp, _m3dstbi__result_info *ri) +{ + void *result=NULL; + if (req_comp < 0 || req_comp > 4) { _m3dstbi__err("bad req_comp", "Internal error"); return NULL; } + if (_m3dstbi__parse_png_file(p, STBI__SCAN_load, req_comp)) { + if (p->depth < 8) + ri->bits_per_channel = 8; + else + ri->bits_per_channel = p->depth; + result = p->out; + p->out = NULL; + if (req_comp && req_comp != p->s->img_out_n) { + if (ri->bits_per_channel == 8) + result = _m3dstbi__convert_format((unsigned char *) result, p->s->img_out_n, req_comp, p->s->img_x, p->s->img_y); + else + result = _m3dstbi__convert_format16((_m3dstbi__uint16 *) result, p->s->img_out_n, req_comp, p->s->img_x, p->s->img_y); + p->s->img_out_n = req_comp; + if (result == NULL) return result; + } + *x = p->s->img_x; + *y = p->s->img_y; + if (n) *n = p->s->img_n; + } + STBI_FREE(p->out); p->out = NULL; + STBI_FREE(p->expanded); p->expanded = NULL; + STBI_FREE(p->idata); p->idata = NULL; + + return result; +} + +static void *_m3dstbi__png_load(_m3dstbi__context *s, int *x, int *y, int *comp, int req_comp, _m3dstbi__result_info *ri) +{ + _m3dstbi__png p; + p.s = s; + return _m3dstbi__do_png(&p, x,y,comp,req_comp, ri); +} +#define stbi__context _m3dstbi__context +#define stbi__result_info _m3dstbi__result_info +#define stbi__png_load _m3dstbi__png_load +#define stbi_zlib_decode_malloc_guesssize_headerflag _m3dstbi_zlib_decode_malloc_guesssize_headerflag +#endif +#if !defined(M3D_NOIMPORTER) && defined(STBI_INCLUDE_STB_IMAGE_H) && !defined(STB_IMAGE_IMPLEMENTATION) +#error "stb_image.h included without STB_IMAGE_IMPLEMENTATION. Sorry, we need some stuff defined inside the ifguard for proper integration" +#endif +#else +#if !defined(STBI_INCLUDE_STB_IMAGE_H) || defined(STBI_NO_ZLIB) +#error "stb_image.h not included or STBI_NO_ZLIB defined. Sorry, we need its zlib implementation for proper integration" +#endif +#endif /* M3D_NOTEXTURE */ + +#if defined(M3D_EXPORTER) && !defined(INCLUDE_STB_IMAGE_WRITE_H) +/* zlib_compressor from + + stb_image_write - v1.13 - public domain - http://nothings.org/stb/stb_image_write.h +*/ +typedef unsigned char _m3dstbiw__uc; +typedef unsigned short _m3dstbiw__us; + +typedef uint16_t _m3dstbiw__uint16; +typedef int16_t _m3dstbiw__int16; +typedef uint32_t _m3dstbiw__uint32; +typedef int32_t _m3dstbiw__int32; + +#define STBIW_MALLOC(s) M3D_MALLOC(s) +#define STBIW_REALLOC(p,ns) M3D_REALLOC(p,ns) +#define STBIW_REALLOC_SIZED(p,oldsz,newsz) STBIW_REALLOC(p,newsz) +#define STBIW_FREE M3D_FREE +#define STBIW_MEMMOVE memmove +#define STBIW_UCHAR (uint8_t) +#define STBIW_ASSERT(x) +#define _m3dstbiw___sbraw(a) ((int *) (a) - 2) +#define _m3dstbiw___sbm(a) _m3dstbiw___sbraw(a)[0] +#define _m3dstbiw___sbn(a) _m3dstbiw___sbraw(a)[1] + +#define _m3dstbiw___sbneedgrow(a,n) ((a)==0 || _m3dstbiw___sbn(a)+n >= _m3dstbiw___sbm(a)) +#define _m3dstbiw___sbmaybegrow(a,n) (_m3dstbiw___sbneedgrow(a,(n)) ? _m3dstbiw___sbgrow(a,n) : 0) +#define _m3dstbiw___sbgrow(a,n) _m3dstbiw___sbgrowf((void **) &(a), (n), sizeof(*(a))) + +#define _m3dstbiw___sbpush(a, v) (_m3dstbiw___sbmaybegrow(a,1), (a)[_m3dstbiw___sbn(a)++] = (v)) +#define _m3dstbiw___sbcount(a) ((a) ? _m3dstbiw___sbn(a) : 0) +#define _m3dstbiw___sbfree(a) ((a) ? STBIW_FREE(_m3dstbiw___sbraw(a)),0 : 0) + +static void *_m3dstbiw___sbgrowf(void **arr, int increment, int itemsize) +{ + int m = *arr ? 2*_m3dstbiw___sbm(*arr)+increment : increment+1; + void *p = STBIW_REALLOC_SIZED(*arr ? _m3dstbiw___sbraw(*arr) : 0, *arr ? (_m3dstbiw___sbm(*arr)*itemsize + sizeof(int)*2) : 0, itemsize * m + sizeof(int)*2); + STBIW_ASSERT(p); + if (p) { + if (!*arr) ((int *) p)[1] = 0; + *arr = (void *) ((int *) p + 2); + _m3dstbiw___sbm(*arr) = m; + } + return *arr; +} + +static unsigned char *_m3dstbiw___zlib_flushf(unsigned char *data, unsigned int *bitbuffer, int *bitcount) +{ + while (*bitcount >= 8) { + _m3dstbiw___sbpush(data, STBIW_UCHAR(*bitbuffer)); + *bitbuffer >>= 8; + *bitcount -= 8; + } + return data; +} + +static int _m3dstbiw___zlib_bitrev(int code, int codebits) +{ + int res=0; + while (codebits--) { + res = (res << 1) | (code & 1); + code >>= 1; + } + return res; +} + +static unsigned int _m3dstbiw___zlib_countm(unsigned char *a, unsigned char *b, int limit) +{ + int i; + for (i=0; i < limit && i < 258; ++i) + if (a[i] != b[i]) break; + return i; +} + +static unsigned int _m3dstbiw___zhash(unsigned char *data) +{ + _m3dstbiw__uint32 hash = data[0] + (data[1] << 8) + (data[2] << 16); + hash ^= hash << 3; + hash += hash >> 5; + hash ^= hash << 4; + hash += hash >> 17; + hash ^= hash << 25; + hash += hash >> 6; + return hash; +} + +#define _m3dstbiw___zlib_flush() (out = _m3dstbiw___zlib_flushf(out, &bitbuf, &bitcount)) +#define _m3dstbiw___zlib_add(code,codebits) \ + (bitbuf |= (code) << bitcount, bitcount += (codebits), _m3dstbiw___zlib_flush()) +#define _m3dstbiw___zlib_huffa(b,c) _m3dstbiw___zlib_add(_m3dstbiw___zlib_bitrev(b,c),c) +#define _m3dstbiw___zlib_huff1(n) _m3dstbiw___zlib_huffa(0x30 + (n), 8) +#define _m3dstbiw___zlib_huff2(n) _m3dstbiw___zlib_huffa(0x190 + (n)-144, 9) +#define _m3dstbiw___zlib_huff3(n) _m3dstbiw___zlib_huffa(0 + (n)-256,7) +#define _m3dstbiw___zlib_huff4(n) _m3dstbiw___zlib_huffa(0xc0 + (n)-280,8) +#define _m3dstbiw___zlib_huff(n) ((n) <= 143 ? _m3dstbiw___zlib_huff1(n) : (n) <= 255 ? _m3dstbiw___zlib_huff2(n) : (n) <= 279 ? _m3dstbiw___zlib_huff3(n) : _m3dstbiw___zlib_huff4(n)) +#define _m3dstbiw___zlib_huffb(n) ((n) <= 143 ? _m3dstbiw___zlib_huff1(n) : _m3dstbiw___zlib_huff2(n)) + +#define _m3dstbiw___ZHASH 16384 + +unsigned char * _m3dstbi_zlib_compress(unsigned char *data, int data_len, int *out_len, int quality) +{ + static unsigned short lengthc[] = { 3,4,5,6,7,8,9,10,11,13,15,17,19,23,27,31,35,43,51,59,67,83,99,115,131,163,195,227,258, 259 }; + static unsigned char lengtheb[]= { 0,0,0,0,0,0,0, 0, 1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4, 4, 4, 5, 5, 5, 5, 0 }; + static unsigned short distc[] = { 1,2,3,4,5,7,9,13,17,25,33,49,65,97,129,193,257,385,513,769,1025,1537,2049,3073,4097,6145,8193,12289,16385,24577, 32768 }; + static unsigned char disteb[] = { 0,0,0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,11,11,12,12,13,13 }; + unsigned int bitbuf=0; + int i,j, bitcount=0; + unsigned char *out = NULL; + unsigned char ***hash_table = (unsigned char***) STBIW_MALLOC(_m3dstbiw___ZHASH * sizeof(char**)); + if (hash_table == NULL) + return NULL; + if (quality < 5) quality = 5; + + _m3dstbiw___sbpush(out, 0x78); + _m3dstbiw___sbpush(out, 0x5e); + _m3dstbiw___zlib_add(1,1); + _m3dstbiw___zlib_add(1,2); + + for (i=0; i < _m3dstbiw___ZHASH; ++i) + hash_table[i] = NULL; + + i=0; + while (i < data_len-3) { + int h = _m3dstbiw___zhash(data+i)&(_m3dstbiw___ZHASH-1), best=3; + unsigned char *bestloc = 0; + unsigned char **hlist = hash_table[h]; + int n = _m3dstbiw___sbcount(hlist); + for (j=0; j < n; ++j) { + if (hlist[j]-data > i-32768) { + int d = _m3dstbiw___zlib_countm(hlist[j], data+i, data_len-i); + if (d >= best) best=d,bestloc=hlist[j]; + } + } + if (hash_table[h] && _m3dstbiw___sbn(hash_table[h]) == 2*quality) { + STBIW_MEMMOVE(hash_table[h], hash_table[h]+quality, sizeof(hash_table[h][0])*quality); + _m3dstbiw___sbn(hash_table[h]) = quality; + } + _m3dstbiw___sbpush(hash_table[h],data+i); + + if (bestloc) { + h = _m3dstbiw___zhash(data+i+1)&(_m3dstbiw___ZHASH-1); + hlist = hash_table[h]; + n = _m3dstbiw___sbcount(hlist); + for (j=0; j < n; ++j) { + if (hlist[j]-data > i-32767) { + int e = _m3dstbiw___zlib_countm(hlist[j], data+i+1, data_len-i-1); + if (e > best) { + bestloc = NULL; + break; + } + } + } + } + + if (bestloc) { + int d = (int) (data+i - bestloc); + STBIW_ASSERT(d <= 32767 && best <= 258); + for (j=0; best > lengthc[j+1]-1; ++j); + _m3dstbiw___zlib_huff(j+257); + if (lengtheb[j]) _m3dstbiw___zlib_add(best - lengthc[j], lengtheb[j]); + for (j=0; d > distc[j+1]-1; ++j); + _m3dstbiw___zlib_add(_m3dstbiw___zlib_bitrev(j,5),5); + if (disteb[j]) _m3dstbiw___zlib_add(d - distc[j], disteb[j]); + i += best; + } else { + _m3dstbiw___zlib_huffb(data[i]); + ++i; + } + } + for (;i < data_len; ++i) + _m3dstbiw___zlib_huffb(data[i]); + _m3dstbiw___zlib_huff(256); + while (bitcount) + _m3dstbiw___zlib_add(0,1); + + for (i=0; i < _m3dstbiw___ZHASH; ++i) + (void) _m3dstbiw___sbfree(hash_table[i]); + STBIW_FREE(hash_table); + + { + unsigned int s1=1, s2=0; + int blocklen = (int) (data_len % 5552); + j=0; + while (j < data_len) { + for (i=0; i < blocklen; ++i) s1 += data[j+i], s2 += s1; + s1 %= 65521, s2 %= 65521; + j += blocklen; + blocklen = 5552; + } + _m3dstbiw___sbpush(out, STBIW_UCHAR(s2 >> 8)); + _m3dstbiw___sbpush(out, STBIW_UCHAR(s2)); + _m3dstbiw___sbpush(out, STBIW_UCHAR(s1 >> 8)); + _m3dstbiw___sbpush(out, STBIW_UCHAR(s1)); + } + *out_len = _m3dstbiw___sbn(out); + STBIW_MEMMOVE(_m3dstbiw___sbraw(out), out, *out_len); + return (unsigned char *) _m3dstbiw___sbraw(out); +} +#define stbi_zlib_compress _m3dstbi_zlib_compress +#else +unsigned char * _m3dstbi_zlib_compress(unsigned char *data, int data_len, int *out_len, int quality); +#endif + +#define M3D_CHUNKMAGIC(m, a,b,c,d) ((m)[0]==(a) && (m)[1]==(b) && (m)[2]==(c) && (m)[3]==(d)) + +#ifdef M3D_ASCII +#include /* get sprintf */ +#include /* sprintf and strtod cares about number locale */ +#endif +#ifdef M3D_PROFILING +#include +#endif + +#if !defined(M3D_NOIMPORTER) && defined(M3D_ASCII) +/* helper functions for the ASCII parser */ +static char *_m3d_findarg(char *s) { + while(s && *s && *s != ' ' && *s != '\t' && *s != '\r' && *s != '\n') s++; + while(s && *s && (*s == ' ' || *s == '\t')) s++; + return s; +} +static char *_m3d_findnl(char *s) { + while(s && *s && *s != '\r' && *s != '\n') s++; + if(*s == '\r') s++; + if(*s == '\n') s++; + return s; +} +static char *_m3d_gethex(char *s, uint32_t *ret) +{ + if(*s == '#') s++; + *ret = 0; + for(; *s; s++) { + if(*s >= '0' && *s <= '9') { *ret <<= 4; *ret += (uint32_t)(*s-'0'); } + else if(*s >= 'a' && *s <= 'f') { *ret <<= 4; *ret += (uint32_t)(*s-'a'+10); } + else if(*s >= 'A' && *s <= 'F') { *ret <<= 4; *ret += (uint32_t)(*s-'A'+10); } + else break; + } + return _m3d_findarg(s); +} +static char *_m3d_getint(char *s, uint32_t *ret) +{ + char *e = s; + if(!s || !*s || *s == '\r' || *s == '\n') return s; + for(; *e >= '0' && *e <= '9'; e++); + *ret = atoi(s); + return e; +} +static char *_m3d_getfloat(char *s, M3D_FLOAT *ret) +{ + char *e = s; + if(!s || !*s || *s == '\r' || *s == '\n') return s; + for(; *e == '-' || *e == '+' || *e == '.' || (*e >= '0' && *e <= '9') || *e == 'e' || *e == 'E'; e++); + *ret = (M3D_FLOAT)strtod(s, NULL); + return _m3d_findarg(e); +} +#endif +#if !defined(M3D_NODUP) && (!defined(M3D_NOIMPORTER) || defined(M3D_ASCII) || defined(M3D_EXPORTER)) +/* helper function to create safe strings */ +char *_m3d_safestr(char *in, int morelines) +{ + char *out, *o, *i = in; + int l; + if(!in || !*in) { + out = (char*)M3D_MALLOC(1); + if(!out) return NULL; + out[0] =0; + } else { + for(o = in, l = 0; *o && ((morelines & 1) || (*o != '\r' && *o != '\n')) && l < 256; o++, l++); + out = o = (char*)M3D_MALLOC(l+1); + if(!out) return NULL; + while(*i == ' ' || *i == '\t' || *i == '\r' || (morelines && *i == '\n')) i++; + for(; *i && (morelines || (*i != '\r' && *i != '\n')); i++) { + if(*i == '\r') continue; + if(*i == '\n') { + if(morelines >= 3 && o > out && *(o-1) == '\n') break; + if(i > in && *(i-1) == '\n') continue; + if(morelines & 1) { + if(morelines == 1) *o++ = '\r'; + *o++ = '\n'; + } else + break; + } else + if(*i == ' ' || *i == '\t') { + *o++ = morelines? ' ' : '_'; + } else + *o++ = !morelines && (*i == '/' || *i == '\\') ? '_' : *i; + } + for(; o > out && (*(o-1) == ' ' || *(o-1) == '\t' || *(o-1) == '\r' || *(o-1) == '\n'); o--); + *o = 0; + out = (char*)M3D_REALLOC(out, (uintptr_t)o - (uintptr_t)out + 1); + } + return out; +} +#endif +#ifndef M3D_NOIMPORTER +/* helper function to load and decode/generate a texture */ +M3D_INDEX _m3d_gettx(m3d_t *model, m3dread_t readfilecb, m3dfree_t freecb, char *fn) +{ + unsigned int i, len = 0; + unsigned char *buff = NULL; + char *fn2; +#ifndef M3D_NOTEXTURE + unsigned int w, h; + stbi__context s; + stbi__result_info ri; +#endif + + /* failsafe */ + if(!fn || !*fn) return M3D_UNDEF; + /* do we have loaded this texture already? */ + for(i = 0; i < model->numtexture; i++) + if(!strcmp(fn, model->texture[i].name)) return i; + /* see if it's inlined in the model */ + if(model->inlined) { + for(i = 0; i < model->numinlined; i++) + if(!strcmp(fn, model->inlined[i].name)) { + buff = model->inlined[i].data; + len = model->inlined[i].length; + freecb = NULL; + break; + } + } + /* try to load from external source */ + if(!buff && readfilecb) { + i = (unsigned int)strlen(fn); + if(i < 5 || fn[i - 4] != '.') { + fn2 = (char*)M3D_MALLOC(i + 5); + if(!fn2) { model->errcode = M3D_ERR_ALLOC; return M3D_UNDEF; } + memcpy(fn2, fn, i); + memcpy(fn2+i, ".png", 5); + buff = (*readfilecb)(fn2, &len); + M3D_FREE(fn2); + } + if(!buff) { + buff = (*readfilecb)(fn, &len); + if(!buff) return M3D_UNDEF; + } + } + /* add to textures array */ + i = model->numtexture++; + model->texture = (m3dtx_t*)M3D_REALLOC(model->texture, model->numtexture * sizeof(m3dtx_t)); + if(!model->texture) { + if(buff && freecb) (*freecb)(buff); + model->errcode = M3D_ERR_ALLOC; + model->numtexture = 0; + return M3D_UNDEF; + } + memset(&model->texture[i], 0, sizeof(m3dtx_t)); + model->texture[i].name = fn; + if(buff) { + if(buff[0] == 0x89 && buff[1] == 'P' && buff[2] == 'N' && buff[3] == 'G') { +#ifndef M3D_NOTEXTURE + /* return pixel buffer of the decoded texture */ + memset(&s, 0, sizeof(s)); + memset(&ri, 0, sizeof(ri)); + s.img_buffer = s.img_buffer_original = (unsigned char *) buff; + s.img_buffer_end = s.img_buffer_original_end = (unsigned char *) buff+len; + /* don't use model->texture[i].w directly, it's a uint16_t */ + w = h = len = 0; + ri.bits_per_channel = 8; + model->texture[i].d = (uint8_t*)stbi__png_load(&s, (int*)&w, (int*)&h, (int*)&len, 0, &ri); + model->texture[i].w = w; + model->texture[i].h = h; + model->texture[i].f = (uint8_t)len; +#else + /* return only the raw undecoded texture */ + if((model->texture[i].d = (uint8_t*)M3D_MALLOC(len))) { + memcpy(model->texture[i].d, buff, len); + model->texture[i].w = len & 0xffff; + model->texture[i].h = (len >> 16) & 0xffff; + model->texture[i].f = 0; + } else + model->errcode = M3D_ERR_ALLOC; +#endif + } else { +#ifdef M3D_TX_INTERP + if((model->errcode = M3D_TX_INTERP(fn, buff, len, &model->texture[i])) != M3D_SUCCESS) { + M3D_LOG("Unable to generate texture"); + M3D_LOG(fn); + } +#else + M3D_LOG("Unimplemented interpreter"); + M3D_LOG(fn); +#endif + } + if(freecb) (*freecb)(buff); + } + if(!model->texture[i].d) + model->errcode = M3D_ERR_UNKIMG; + return i; +} + +/* helper function to load and generate a procedural surface */ +void _m3d_getpr(m3d_t *model, _unused m3dread_t readfilecb, _unused m3dfree_t freecb, _unused char *fn) +{ +#ifdef M3D_PR_INTERP + unsigned int i, len = 0; + unsigned char *buff = readfilecb && fn && *fn ? (*readfilecb)(fn, &len) : NULL; + + if(!buff && fn && *fn && model->inlined) { + for(i = 0; i < model->numinlined; i++) + if(!strcmp(fn, model->inlined[i].name)) { + buff = model->inlined[i].data; + len = model->inlined[i].length; + freecb = NULL; + break; + } + } + if(!buff || !len || (model->errcode = M3D_PR_INTERP(fn, buff, len, model)) != M3D_SUCCESS) { + M3D_LOG("Unable to generate procedural surface"); + M3D_LOG(fn); + model->errcode = M3D_ERR_UNKIMG; + } + if(freecb && buff) (*freecb)(buff); +#else + (void)readfilecb; + (void)freecb; + (void)fn; + M3D_LOG("Unimplemented interpreter"); + M3D_LOG(fn); + model->errcode = M3D_ERR_UNIMPL; +#endif +} +/* helpers to read indices from data stream */ +#define M3D_GETSTR(x) do{offs=0;data=_m3d_getidx(data,model->si_s,&offs);x=offs?((char*)model->raw+16+offs):NULL;}while(0) +_inline static unsigned char *_m3d_getidx(unsigned char *data, char type, M3D_INDEX *idx) +{ + switch(type) { + case 1: *idx = data[0] > 253 ? (int8_t)data[0] : data[0]; data++; break; + case 2: *idx = *((uint16_t*)data) > 65533 ? *((int16_t*)data) : *((uint16_t*)data); data += 2; break; + case 4: *idx = *((int32_t*)data); data += 4; break; + } + return data; +} + +#ifndef M3D_NOANIMATION +/* multiply 4 x 4 matrices. Do not use float *r[16] as argument, because some compilers misinterpret that as + * 16 pointers each pointing to a float, but we need a single pointer to 16 floats. */ +void _m3d_mul(M3D_FLOAT *r, M3D_FLOAT *a, M3D_FLOAT *b) +{ + r[ 0] = b[ 0] * a[ 0] + b[ 4] * a[ 1] + b[ 8] * a[ 2] + b[12] * a[ 3]; + r[ 1] = b[ 1] * a[ 0] + b[ 5] * a[ 1] + b[ 9] * a[ 2] + b[13] * a[ 3]; + r[ 2] = b[ 2] * a[ 0] + b[ 6] * a[ 1] + b[10] * a[ 2] + b[14] * a[ 3]; + r[ 3] = b[ 3] * a[ 0] + b[ 7] * a[ 1] + b[11] * a[ 2] + b[15] * a[ 3]; + r[ 4] = b[ 0] * a[ 4] + b[ 4] * a[ 5] + b[ 8] * a[ 6] + b[12] * a[ 7]; + r[ 5] = b[ 1] * a[ 4] + b[ 5] * a[ 5] + b[ 9] * a[ 6] + b[13] * a[ 7]; + r[ 6] = b[ 2] * a[ 4] + b[ 6] * a[ 5] + b[10] * a[ 6] + b[14] * a[ 7]; + r[ 7] = b[ 3] * a[ 4] + b[ 7] * a[ 5] + b[11] * a[ 6] + b[15] * a[ 7]; + r[ 8] = b[ 0] * a[ 8] + b[ 4] * a[ 9] + b[ 8] * a[10] + b[12] * a[11]; + r[ 9] = b[ 1] * a[ 8] + b[ 5] * a[ 9] + b[ 9] * a[10] + b[13] * a[11]; + r[10] = b[ 2] * a[ 8] + b[ 6] * a[ 9] + b[10] * a[10] + b[14] * a[11]; + r[11] = b[ 3] * a[ 8] + b[ 7] * a[ 9] + b[11] * a[10] + b[15] * a[11]; + r[12] = b[ 0] * a[12] + b[ 4] * a[13] + b[ 8] * a[14] + b[12] * a[15]; + r[13] = b[ 1] * a[12] + b[ 5] * a[13] + b[ 9] * a[14] + b[13] * a[15]; + r[14] = b[ 2] * a[12] + b[ 6] * a[13] + b[10] * a[14] + b[14] * a[15]; + r[15] = b[ 3] * a[12] + b[ 7] * a[13] + b[11] * a[14] + b[15] * a[15]; +} +/* calculate 4 x 4 matrix inverse */ +void _m3d_inv(M3D_FLOAT *m) +{ + M3D_FLOAT r[16]; + M3D_FLOAT det = + m[ 0]*m[ 5]*m[10]*m[15] - m[ 0]*m[ 5]*m[11]*m[14] + m[ 0]*m[ 6]*m[11]*m[13] - m[ 0]*m[ 6]*m[ 9]*m[15] + + m[ 0]*m[ 7]*m[ 9]*m[14] - m[ 0]*m[ 7]*m[10]*m[13] - m[ 1]*m[ 6]*m[11]*m[12] + m[ 1]*m[ 6]*m[ 8]*m[15] + - m[ 1]*m[ 7]*m[ 8]*m[14] + m[ 1]*m[ 7]*m[10]*m[12] - m[ 1]*m[ 4]*m[10]*m[15] + m[ 1]*m[ 4]*m[11]*m[14] + + m[ 2]*m[ 7]*m[ 8]*m[13] - m[ 2]*m[ 7]*m[ 9]*m[12] + m[ 2]*m[ 4]*m[ 9]*m[15] - m[ 2]*m[ 4]*m[11]*m[13] + + m[ 2]*m[ 5]*m[11]*m[12] - m[ 2]*m[ 5]*m[ 8]*m[15] - m[ 3]*m[ 4]*m[ 9]*m[14] + m[ 3]*m[ 4]*m[10]*m[13] + - m[ 3]*m[ 5]*m[10]*m[12] + m[ 3]*m[ 5]*m[ 8]*m[14] - m[ 3]*m[ 6]*m[ 8]*m[13] + m[ 3]*m[ 6]*m[ 9]*m[12]; + if(det == (M3D_FLOAT)0.0 || det == (M3D_FLOAT)-0.0) det = (M3D_FLOAT)1.0; else det = (M3D_FLOAT)1.0 / det; + r[ 0] = det *(m[ 5]*(m[10]*m[15] - m[11]*m[14]) + m[ 6]*(m[11]*m[13] - m[ 9]*m[15]) + m[ 7]*(m[ 9]*m[14] - m[10]*m[13])); + r[ 1] = -det*(m[ 1]*(m[10]*m[15] - m[11]*m[14]) + m[ 2]*(m[11]*m[13] - m[ 9]*m[15]) + m[ 3]*(m[ 9]*m[14] - m[10]*m[13])); + r[ 2] = det *(m[ 1]*(m[ 6]*m[15] - m[ 7]*m[14]) + m[ 2]*(m[ 7]*m[13] - m[ 5]*m[15]) + m[ 3]*(m[ 5]*m[14] - m[ 6]*m[13])); + r[ 3] = -det*(m[ 1]*(m[ 6]*m[11] - m[ 7]*m[10]) + m[ 2]*(m[ 7]*m[ 9] - m[ 5]*m[11]) + m[ 3]*(m[ 5]*m[10] - m[ 6]*m[ 9])); + r[ 4] = -det*(m[ 4]*(m[10]*m[15] - m[11]*m[14]) + m[ 6]*(m[11]*m[12] - m[ 8]*m[15]) + m[ 7]*(m[ 8]*m[14] - m[10]*m[12])); + r[ 5] = det *(m[ 0]*(m[10]*m[15] - m[11]*m[14]) + m[ 2]*(m[11]*m[12] - m[ 8]*m[15]) + m[ 3]*(m[ 8]*m[14] - m[10]*m[12])); + r[ 6] = -det*(m[ 0]*(m[ 6]*m[15] - m[ 7]*m[14]) + m[ 2]*(m[ 7]*m[12] - m[ 4]*m[15]) + m[ 3]*(m[ 4]*m[14] - m[ 6]*m[12])); + r[ 7] = det *(m[ 0]*(m[ 6]*m[11] - m[ 7]*m[10]) + m[ 2]*(m[ 7]*m[ 8] - m[ 4]*m[11]) + m[ 3]*(m[ 4]*m[10] - m[ 6]*m[ 8])); + r[ 8] = det *(m[ 4]*(m[ 9]*m[15] - m[11]*m[13]) + m[ 5]*(m[11]*m[12] - m[ 8]*m[15]) + m[ 7]*(m[ 8]*m[13] - m[ 9]*m[12])); + r[ 9] = -det*(m[ 0]*(m[ 9]*m[15] - m[11]*m[13]) + m[ 1]*(m[11]*m[12] - m[ 8]*m[15]) + m[ 3]*(m[ 8]*m[13] - m[ 9]*m[12])); + r[10] = det *(m[ 0]*(m[ 5]*m[15] - m[ 7]*m[13]) + m[ 1]*(m[ 7]*m[12] - m[ 4]*m[15]) + m[ 3]*(m[ 4]*m[13] - m[ 5]*m[12])); + r[11] = -det*(m[ 0]*(m[ 5]*m[11] - m[ 7]*m[ 9]) + m[ 1]*(m[ 7]*m[ 8] - m[ 4]*m[11]) + m[ 3]*(m[ 4]*m[ 9] - m[ 5]*m[ 8])); + r[12] = -det*(m[ 4]*(m[ 9]*m[14] - m[10]*m[13]) + m[ 5]*(m[10]*m[12] - m[ 8]*m[14]) + m[ 6]*(m[ 8]*m[13] - m[ 9]*m[12])); + r[13] = det *(m[ 0]*(m[ 9]*m[14] - m[10]*m[13]) + m[ 1]*(m[10]*m[12] - m[ 8]*m[14]) + m[ 2]*(m[ 8]*m[13] - m[ 9]*m[12])); + r[14] = -det*(m[ 0]*(m[ 5]*m[14] - m[ 6]*m[13]) + m[ 1]*(m[ 6]*m[12] - m[ 4]*m[14]) + m[ 2]*(m[ 4]*m[13] - m[ 5]*m[12])); + r[15] = det *(m[ 0]*(m[ 5]*m[10] - m[ 6]*m[ 9]) + m[ 1]*(m[ 6]*m[ 8] - m[ 4]*m[10]) + m[ 2]*(m[ 4]*m[ 9] - m[ 5]*m[ 8])); + memcpy(m, &r, sizeof(r)); +} +/* compose a coloumn major 4 x 4 matrix from vec3 position and vec4 orientation/rotation quaternion */ +void _m3d_mat(M3D_FLOAT *r, m3dv_t *p, m3dv_t *q) +{ + if(q->x == (M3D_FLOAT)0.0 && q->y == (M3D_FLOAT)0.0 && q->z >=(M3D_FLOAT) 0.7071065 && q->z <= (M3D_FLOAT)0.7071075 && + q->w == (M3D_FLOAT)0.0) { + r[ 1] = r[ 2] = r[ 4] = r[ 6] = r[ 8] = r[ 9] = (M3D_FLOAT)0.0; + r[ 0] = r[ 5] = r[10] = (M3D_FLOAT)-1.0; + } else { + r[ 0] = 1 - 2 * (q->y * q->y + q->z * q->z); if(r[ 0]>-M3D_EPSILON && r[ 0]x * q->y - q->z * q->w); if(r[ 1]>-M3D_EPSILON && r[ 1]x * q->z + q->y * q->w); if(r[ 2]>-M3D_EPSILON && r[ 2]x * q->y + q->z * q->w); if(r[ 4]>-M3D_EPSILON && r[ 4]x * q->x + q->z * q->z); if(r[ 5]>-M3D_EPSILON && r[ 5]y * q->z - q->x * q->w); if(r[ 6]>-M3D_EPSILON && r[ 6]x * q->z - q->y * q->w); if(r[ 8]>-M3D_EPSILON && r[ 8]y * q->z + q->x * q->w); if(r[ 9]>-M3D_EPSILON && r[ 9]x * q->x + q->y * q->y); if(r[10]>-M3D_EPSILON && r[10]x; r[ 7] = p->y; r[11] = p->z; + r[12] = 0; r[13] = 0; r[14] = 0; r[15] = 1; +} +#endif +#if !defined(M3D_NOANIMATION) || !defined(M3D_NONORMALS) +/* portable fast inverse square root calculation. returns 1/sqrt(x) */ +static M3D_FLOAT _m3d_rsq(M3D_FLOAT x) +{ +#ifdef M3D_DOUBLE + return ((M3D_FLOAT)15.0/(M3D_FLOAT)8.0) + ((M3D_FLOAT)-5.0/(M3D_FLOAT)4.0)*x + ((M3D_FLOAT)3.0/(M3D_FLOAT)8.0)*x*x; +#else + /* John Carmack's */ + float x2 = x * 0.5f; + uint32_t *i = (uint32_t*)&x; + *i = (0x5f3759df - (*i >> 1)); + return x * (1.5f - (x2 * x * x)); +#endif +} +#endif + +/** + * Function to decode a Model 3D into in-memory format + */ +m3d_t *m3d_load(unsigned char *data, m3dread_t readfilecb, m3dfree_t freecb, m3d_t *mtllib) +{ + unsigned char *end, *chunk, *buff, weights[8]; + unsigned int i, j, k, l, n, am, len = 0, reclen, offs; +#ifndef M3D_NOVOXELS + int32_t min_x, min_y, min_z, max_x, max_y, max_z, sx, sy, sz, x, y, z; + M3D_INDEX edge[8], enorm; +#endif + char *name, *lang; + float f; + m3d_t *model; + M3D_INDEX mi; +#ifdef M3D_VERTEXMAX + M3D_INDEX pi; +#endif + M3D_FLOAT w; + m3dcd_t *cd; + m3dtx_t *tx; + m3dh_t *h; + m3dm_t *m; + m3da_t *a; + m3di_t *t; +#ifndef M3D_NONORMALS + char neednorm = 0; + m3dv_t *norm = NULL, *v0, *v1, *v2, va, vb; +#endif +#ifndef M3D_NOANIMATION + M3D_FLOAT r[16]; +#endif +#if !defined(M3D_NOWEIGHTS) || !defined(M3D_NOANIMATION) + m3db_t *b; +#endif +#ifndef M3D_NOWEIGHTS + m3ds_t *sk; +#endif +#ifdef M3D_ASCII + m3ds_t s; + M3D_INDEX bi[M3D_BONEMAXLEVEL+1], level; + const char *ol; + char *ptr, *pe, *fn; +#endif +#ifdef M3D_PROFILING + struct timeval tv0, tv1, tvd; + gettimeofday(&tv0, NULL); +#endif + + if(!data || (!M3D_CHUNKMAGIC(data, '3','D','M','O') +#ifdef M3D_ASCII + && !M3D_CHUNKMAGIC(data, '3','d','m','o') +#endif + )) return NULL; + model = (m3d_t*)M3D_MALLOC(sizeof(m3d_t)); + if(!model) { + M3D_LOG("Out of memory"); + return NULL; + } + memset(model, 0, sizeof(m3d_t)); + + if(mtllib) { + model->nummaterial = mtllib->nummaterial; + model->material = mtllib->material; + model->numtexture = mtllib->numtexture; + model->texture = mtllib->texture; + model->flags |= M3D_FLG_MTLLIB; + } +#ifdef M3D_ASCII + /* ASCII variant? */ + if(M3D_CHUNKMAGIC(data, '3','d','m','o')) { + model->errcode = M3D_ERR_BADFILE; + model->flags |= M3D_FLG_FREESTR; + model->raw = (m3dhdr_t*)data; + ptr = (char*)data; + ol = setlocale(LC_NUMERIC, NULL); + setlocale(LC_NUMERIC, "C"); + /* parse header. Don't use sscanf, that's incredibly slow */ + ptr = _m3d_findarg(ptr); + if(!*ptr || *ptr == '\r' || *ptr == '\n') goto asciiend; + pe = _m3d_findnl(ptr); + model->scale = (float)strtod(ptr, NULL); ptr = pe; + if(model->scale <= (M3D_FLOAT)0.0) model->scale = (M3D_FLOAT)1.0; + model->name = _m3d_safestr(ptr, 2); ptr = _m3d_findnl(ptr); + if(!*ptr) goto asciiend; + model->license = _m3d_safestr(ptr, 2); ptr = _m3d_findnl(ptr); + if(!*ptr) goto asciiend; + model->author = _m3d_safestr(ptr, 2); ptr = _m3d_findnl(ptr); + if(!*ptr) goto asciiend; + if(*ptr != '\r' && *ptr != '\n') + model->desc = _m3d_safestr(ptr, 3); + while(*ptr) { + while(*ptr && *ptr!='\n') ptr++; + ptr++; if(*ptr=='\r') ptr++; + if(*ptr == '\n') break; + } + + /* the main chunk reader loop */ + while(*ptr) { + while(*ptr && (*ptr == '\r' || *ptr == '\n')) ptr++; + if(!*ptr || (ptr[0]=='E' && ptr[1]=='n' && ptr[2]=='d')) break; + /* make sure there's at least one data row */ + pe = ptr; ptr = _m3d_findnl(ptr); + if(!*ptr || *ptr == '\r' || *ptr == '\n') goto asciiend; + /* Preview chunk */ + if(!memcmp(pe, "Preview", 7)) { + if(readfilecb) { + pe = _m3d_safestr(ptr, 0); + if(!pe || !*pe) goto asciiend; + model->preview.data = (*readfilecb)(pe, &model->preview.length); + M3D_FREE(pe); + } + while(*ptr && *ptr != '\r' && *ptr != '\n') + ptr = _m3d_findnl(ptr); + } else + /* texture map chunk */ + if(!memcmp(pe, "Textmap", 7)) { + if(model->tmap) { M3D_LOG("More texture map chunks, should be unique"); goto asciiend; } + while(*ptr && *ptr != '\r' && *ptr != '\n') { + i = model->numtmap++; + model->tmap = (m3dti_t*)M3D_REALLOC(model->tmap, model->numtmap * sizeof(m3dti_t)); + if(!model->tmap) goto memerr; + ptr = _m3d_getfloat(ptr, &model->tmap[i].u); + if(!*ptr || *ptr == '\r' || *ptr == '\n') goto asciiend; + _m3d_getfloat(ptr, &model->tmap[i].v); + ptr = _m3d_findnl(ptr); + } + } else + /* vertex chunk */ + if(!memcmp(pe, "Vertex", 6)) { + if(model->vertex) { M3D_LOG("More vertex chunks, should be unique"); goto asciiend; } + while(*ptr && *ptr != '\r' && *ptr != '\n') { + i = model->numvertex++; + model->vertex = (m3dv_t*)M3D_REALLOC(model->vertex, model->numvertex * sizeof(m3dv_t)); + if(!model->vertex) goto memerr; + memset(&model->vertex[i], 0, sizeof(m3dv_t)); + model->vertex[i].skinid = M3D_UNDEF; + model->vertex[i].color = 0; + model->vertex[i].w = (M3D_FLOAT)1.0; + ptr = _m3d_getfloat(ptr, &model->vertex[i].x); + if(!*ptr || *ptr == '\r' || *ptr == '\n') goto asciiend; + ptr = _m3d_getfloat(ptr, &model->vertex[i].y); + if(!*ptr || *ptr == '\r' || *ptr == '\n') goto asciiend; + ptr = _m3d_getfloat(ptr, &model->vertex[i].z); + if(!*ptr || *ptr == '\r' || *ptr == '\n') goto asciiend; + ptr = _m3d_getfloat(ptr, &model->vertex[i].w); + if(!*ptr) goto asciiend; + if(*ptr == '#') { + ptr = _m3d_gethex(ptr, &model->vertex[i].color); + if(!*ptr) goto asciiend; + } + /* parse skin */ + memset(&s, 0, sizeof(m3ds_t)); + for(j = 0, w = (M3D_FLOAT)0.0; j < M3D_NUMBONE && *ptr && *ptr != '\r' && *ptr != '\n'; j++) { + ptr = _m3d_findarg(ptr); + if(!*ptr || *ptr == '\r' || *ptr == '\n') goto asciiend; + ptr = _m3d_getint(ptr, &k); + s.boneid[j] = (M3D_INDEX)k; + if(*ptr == ':') { + ptr++; + ptr = _m3d_getfloat(ptr, &s.weight[j]); + w += s.weight[j]; + } else if(!j) + s.weight[j] = (M3D_FLOAT)1.0; + if(!*ptr) goto asciiend; + } + if(s.boneid[0] != M3D_UNDEF && s.weight[0] > (M3D_FLOAT)0.0) { + if(w != (M3D_FLOAT)1.0 && w != (M3D_FLOAT)0.0) + for(j = 0; j < M3D_NUMBONE && s.weight[j] > (M3D_FLOAT)0.0; j++) + s.weight[j] /= w; + k = M3D_NOTDEFINED; + if(model->skin) { + for(j = 0; j < model->numskin; j++) + if(!memcmp(&model->skin[j], &s, sizeof(m3ds_t))) { k = j; break; } + } + if(k == M3D_NOTDEFINED) { + k = model->numskin++; + model->skin = (m3ds_t*)M3D_REALLOC(model->skin, model->numskin * sizeof(m3ds_t)); + if(!model->skin) goto memerr; + memcpy(&model->skin[k], &s, sizeof(m3ds_t)); + } + model->vertex[i].skinid = (M3D_INDEX)k; + } + ptr = _m3d_findnl(ptr); + } + } else + /* Skeleton, bone hierarchy */ + if(!memcmp(pe, "Bones", 5)) { + if(model->bone) { M3D_LOG("More bones chunks, should be unique"); goto asciiend; } + bi[0] = M3D_UNDEF; + while(*ptr && *ptr != '\r' && *ptr != '\n') { + i = model->numbone++; + model->bone = (m3db_t*)M3D_REALLOC(model->bone, model->numbone * sizeof(m3db_t)); + if(!model->bone) goto memerr; + for(level = 0; *ptr == '/'; ptr++, level++); + if(level > M3D_BONEMAXLEVEL || !*ptr || *ptr == '\r' || *ptr == '\n') goto asciiend; + bi[level+1] = i; + model->bone[i].numweight = 0; + model->bone[i].weight = NULL; + model->bone[i].parent = bi[level]; + ptr = _m3d_getint(ptr, &k); + ptr = _m3d_findarg(ptr); + if(!*ptr || *ptr == '\r' || *ptr == '\n') goto asciiend; + model->bone[i].pos = (M3D_INDEX)k; + ptr = _m3d_getint(ptr, &k); + ptr = _m3d_findarg(ptr); + if(!*ptr || *ptr == '\r' || *ptr == '\n') goto asciiend; + model->bone[i].ori = (M3D_INDEX)k; + model->vertex[k].skinid = M3D_INDEXMAX; + pe = _m3d_safestr(ptr, 0); + if(!pe || !*pe) goto asciiend; + model->bone[i].name = pe; + ptr = _m3d_findnl(ptr); + } + } else + /* material chunk */ + if(!memcmp(pe, "Material", 8)) { + pe = _m3d_findarg(pe); + if(!*pe || *pe == '\r' || *pe == '\n') goto asciiend; + pe = _m3d_safestr(pe, 0); + if(!pe || !*pe) goto asciiend; + for(i = 0; i < model->nummaterial; i++) + if(!strcmp(pe, model->material[i].name)) { + M3D_LOG("Multiple definitions for material"); + M3D_LOG(pe); + M3D_FREE(pe); + pe = NULL; + while(*ptr && *ptr != '\r' && *ptr != '\n') ptr = _m3d_findnl(ptr); + break; + } + if(!pe) continue; + i = model->nummaterial++; + if(model->flags & M3D_FLG_MTLLIB) { + m = model->material; + model->material = (m3dm_t*)M3D_MALLOC(model->nummaterial * sizeof(m3dm_t)); + if(!model->material) goto memerr; + memcpy(model->material, m, (model->nummaterial - 1) * sizeof(m3dm_t)); + if(model->texture) { + tx = model->texture; + model->texture = (m3dtx_t*)M3D_MALLOC(model->numtexture * sizeof(m3dtx_t)); + if(!model->texture) goto memerr; + memcpy(model->texture, tx, model->numtexture * sizeof(m3dm_t)); + } + model->flags &= ~M3D_FLG_MTLLIB; + } else { + model->material = (m3dm_t*)M3D_REALLOC(model->material, model->nummaterial * sizeof(m3dm_t)); + if(!model->material) goto memerr; + } + m = &model->material[i]; + m->name = pe; + m->numprop = 0; + m->prop = NULL; + while(*ptr && *ptr != '\r' && *ptr != '\n') { + k = n = 256; + if(*ptr == 'm' && *(ptr+1) == 'a' && *(ptr+2) == 'p' && *(ptr+3) == '_') { + k = m3dpf_map; + ptr += 4; + } + for(j = 0; j < sizeof(m3d_propertytypes)/sizeof(m3d_propertytypes[0]); j++) + if(!memcmp(ptr, m3d_propertytypes[j].key, strlen(m3d_propertytypes[j].key))) { + n = m3d_propertytypes[j].id; + if(k != m3dpf_map) k = m3d_propertytypes[j].format; + break; + } + if(n != 256 && k != 256) { + ptr = _m3d_findarg(ptr); + if(!*ptr || *ptr == '\r' || *ptr == '\n') goto asciiend; + j = m->numprop++; + m->prop = (m3dp_t*)M3D_REALLOC(m->prop, m->numprop * sizeof(m3dp_t)); + if(!m->prop) goto memerr; + m->prop[j].type = n + (k == m3dpf_map && n < 128 ? 128 : 0); + switch(k) { + case m3dpf_color: ptr = _m3d_gethex(ptr, &m->prop[j].value.color); break; + case m3dpf_uint8: + case m3dpf_uint16: + case m3dpf_uint32: ptr = _m3d_getint(ptr, &m->prop[j].value.num); break; + case m3dpf_float: ptr = _m3d_getfloat(ptr, &m->prop[j].value.fnum); break; + case m3dpf_map: + pe = _m3d_safestr(ptr, 0); + if(!pe || !*pe) goto asciiend; + m->prop[j].value.textureid = _m3d_gettx(model, readfilecb, freecb, pe); + if(model->errcode == M3D_ERR_ALLOC) { M3D_FREE(pe); goto memerr; } + /* this error code only returned if readfilecb was specified */ + if(m->prop[j].value.textureid == M3D_UNDEF) { + M3D_LOG("Texture not found"); + M3D_LOG(pe); + m->numprop--; + } + M3D_FREE(pe); + break; + } + } else { + M3D_LOG("Unknown material property in"); + M3D_LOG(m->name); + model->errcode = M3D_ERR_UNKPROP; + } + ptr = _m3d_findnl(ptr); + } + if(!m->numprop) model->nummaterial--; + } else + /* procedural */ + if(!memcmp(pe, "Procedural", 10)) { + pe = _m3d_safestr(ptr, 0); + _m3d_getpr(model, readfilecb, freecb, pe); + M3D_FREE(pe); + while(*ptr && *ptr != '\r' && *ptr != '\n') ptr = _m3d_findnl(ptr); + } else + /* mesh */ + if(!memcmp(pe, "Mesh", 4)) { + mi = M3D_UNDEF; +#ifdef M3D_VERTEXMAX + pi = M3D_UNDEF; +#endif + while(*ptr && *ptr != '\r' && *ptr != '\n') { + if(*ptr == 'u') { + ptr = _m3d_findarg(ptr); + if(!*ptr) goto asciiend; + mi = M3D_UNDEF; + if(*ptr != '\r' && *ptr != '\n') { + pe = _m3d_safestr(ptr, 0); + if(!pe || !*pe) goto asciiend; + for(j = 0; j < model->nummaterial; j++) + if(!strcmp(pe, model->material[j].name)) { mi = (M3D_INDEX)j; break; } + if(mi == M3D_UNDEF && !(model->flags & M3D_FLG_MTLLIB)) { + mi = model->nummaterial++; + model->material = (m3dm_t*)M3D_REALLOC(model->material, model->nummaterial * sizeof(m3dm_t)); + if(!model->material) goto memerr; + model->material[mi].name = pe; + model->material[mi].numprop = 1; + model->material[mi].prop = NULL; + } else + M3D_FREE(pe); + } + } else + if(*ptr == 'p') { + ptr = _m3d_findarg(ptr); + if(!*ptr) goto asciiend; +#ifdef M3D_VERTEXMAX + pi = M3D_UNDEF; + if(*ptr != '\r' && *ptr != '\n') { + pe = _m3d_safestr(ptr, 0); + if(!pe || !*pe) goto asciiend; + for(j = 0; j < model->numparam; j++) + if(!strcmp(pe, model->param[j].name)) { pi = (M3D_INDEX)j; break; } + if(pi == M3D_UNDEF) { + pi = model->numparam++; + model->param = (m3dvi_t*)M3D_REALLOC(model->param, model->numparam * sizeof(m3dvi_t)); + if(!model->param) goto memerr; + model->param[pi].name = pe; + model->param[pi].count = 0; + } else + M3D_FREE(pe); + } +#endif + } else { + i = model->numface++; + model->face = (m3df_t*)M3D_REALLOC(model->face, model->numface * sizeof(m3df_t)); + if(!model->face) goto memerr; + memset(&model->face[i], 255, sizeof(m3df_t)); /* set all index to -1 by default */ + model->face[i].materialid = mi; +#ifdef M3D_VERTEXMAX + model->face[i].paramid = pi; +#endif + /* hardcoded triangles. */ + for(j = 0; j < 3; j++) { + /* vertex */ + ptr = _m3d_getint(ptr, &k); + model->face[i].vertex[j] = (M3D_INDEX)k; + if(!*ptr) goto asciiend; + if(*ptr == '/') { + ptr++; + if(*ptr != '/') { + /* texcoord */ + ptr = _m3d_getint(ptr, &k); + model->face[i].texcoord[j] = (M3D_INDEX)k; + if(!*ptr) goto asciiend; + } + if(*ptr == '/') { + ptr++; + /* normal */ + ptr = _m3d_getint(ptr, &k); + model->face[i].normal[j] = (M3D_INDEX)k; + if(!*ptr) goto asciiend; + } + if(*ptr == '/') { + ptr++; + /* maximum */ + ptr = _m3d_getint(ptr, &k); +#ifdef M3D_VERTEXMAX + model->face[i].vertmax[j] = (M3D_INDEX)k; +#endif + if(!*ptr) goto asciiend; + } + } +#ifndef M3D_NONORMALS + if(model->face[i].normal[j] == M3D_UNDEF) neednorm = 1; +#endif + ptr = _m3d_findarg(ptr); + } + } + ptr = _m3d_findnl(ptr); + } + } else + /* voxel types chunk */ + if(!memcmp(pe, "VoxTypes", 8) || !memcmp(pe, "Voxtypes", 8)) { + if(model->voxtype) { M3D_LOG("More voxel types chunks, should be unique"); goto asciiend; } + while(*ptr && *ptr != '\r' && *ptr != '\n') { + i = model->numvoxtype++; + model->voxtype = (m3dvt_t*)M3D_REALLOC(model->voxtype, model->numvoxtype * sizeof(m3dvt_t)); + if(!model->voxtype) goto memerr; + memset(&model->voxtype[i], 0, sizeof(m3dvt_t)); + model->voxtype[i].materialid = M3D_UNDEF; + model->voxtype[i].skinid = M3D_UNDEF; + ptr = _m3d_gethex(ptr, &model->voxtype[i].color); + if(!*ptr) goto asciiend; + if(*ptr == '/') { + ptr = _m3d_gethex(ptr, &k); + model->voxtype[i].rotation = k; + if(!*ptr) goto asciiend; + if(*ptr == '/') { + ptr = _m3d_gethex(ptr, &k); + model->voxtype[i].voxshape = k; + if(!*ptr) goto asciiend; + } + } + while(*ptr == ' ' || *ptr == '\t') ptr++; + if(*ptr == '\r' || *ptr == '\n') { ptr = _m3d_findnl(ptr); continue; } + /* name */ + if(*ptr != '-') { + pe = _m3d_safestr(ptr, 0); + if(!pe || !*pe) goto asciiend; + model->voxtype[i].name = pe; + for(j = 0; j < model->nummaterial; j++) + if(!strcmp(pe, model->material[j].name)) { model->voxtype[i].materialid = (M3D_INDEX)j; break; } + } + ptr = _m3d_findarg(ptr); + /* parse skin */ + memset(&s, 0, sizeof(m3ds_t)); + for(j = 0, w = (M3D_FLOAT)0.0; j < M3D_NUMBONE && *ptr && *ptr != '{' && *ptr != '\r' && *ptr != '\n'; j++) { + ptr = _m3d_getint(ptr, &k); + s.boneid[j] = (M3D_INDEX)k; + if(*ptr == ':') { + ptr++; + ptr = _m3d_getfloat(ptr, &s.weight[j]); + w += s.weight[j]; + } else if(!j) + s.weight[j] = (M3D_FLOAT)1.0; + if(!*ptr) goto asciiend; + ptr = _m3d_findarg(ptr); + } + if(s.boneid[0] != M3D_UNDEF && s.weight[0] > (M3D_FLOAT)0.0) { + if(w != (M3D_FLOAT)1.0 && w != (M3D_FLOAT)0.0) + for(j = 0; j < M3D_NUMBONE && s.weight[j] > (M3D_FLOAT)0.0; j++) + s.weight[j] /= w; + k = M3D_NOTDEFINED; + if(model->skin) { + for(j = 0; j < model->numskin; j++) + if(!memcmp(&model->skin[j], &s, sizeof(m3ds_t))) { k = j; break; } + } + if(k == M3D_NOTDEFINED) { + k = model->numskin++; + model->skin = (m3ds_t*)M3D_REALLOC(model->skin, model->numskin * sizeof(m3ds_t)); + if(!model->skin) goto memerr; + memcpy(&model->skin[k], &s, sizeof(m3ds_t)); + } + model->voxtype[i].skinid = (M3D_INDEX)k; + } + /* parse item list */ + if(*ptr == '{') { + while(*ptr == '{' || *ptr == ' ' || *ptr == '\t') ptr++; + while(*ptr && *ptr != '}' && *ptr != '\r' && *ptr != '\n') { + ptr = _m3d_getint(ptr, &k); + ptr = _m3d_findarg(ptr); + if(!*ptr || *ptr == '}' || *ptr == '\r' || *ptr == '\n') goto asciiend; + pe = _m3d_safestr(ptr, 0); + if(!pe || !*pe) goto asciiend; + ptr = _m3d_findarg(ptr); + j = model->voxtype[i].numitem++; + model->voxtype[i].item = (m3dvi_t*)M3D_REALLOC(model->voxtype[i].item, + model->voxtype[i].numitem * sizeof(m3dvi_t)); + if(!model->voxtype[i].item) goto memerr; + model->voxtype[i].item[j].count = k; + model->voxtype[i].item[j].name = pe; + } + if(*ptr != '}') goto asciiend; + } + ptr = _m3d_findnl(ptr); + } + } else + /* voxel data */ + if(!memcmp(pe, "Voxel", 5)) { + if(!model->voxtype) { M3D_LOG("No voxel type chunk before voxel data"); goto asciiend; } + pe = _m3d_findarg(pe); + if(!*pe) goto asciiend; + if(*pe == '\r' || *pe == '\n') pe = NULL; + else pe = _m3d_safestr(pe, 0); + i = model->numvoxel++; + model->voxel = (m3dvx_t*)M3D_REALLOC(model->voxel, model->numvoxel * sizeof(m3dvx_t)); + if(!model->voxel) goto memerr; + memset(&model->voxel[i], 0, sizeof(m3dvx_t)); + model->voxel[i].name = pe; + k = l = 0; + while(*ptr && *ptr != '\r' && *ptr != '\n') { + switch(*ptr) { + case 'u': + ptr = _m3d_findarg(ptr); + if(!*ptr || *ptr == '\r' || *ptr == '\n') goto asciiend; + ptr = _m3d_getint(ptr, &n); + model->voxel[i].uncertain = ((n > 0 && n < 256 ? n : 0) * 255) / 100; + ptr = _m3d_findarg(ptr); + if(*ptr && *ptr != '\r' && *ptr != '\n') { + ptr = _m3d_getint(ptr, &n); + model->voxel[i].groupid = n > 0 && n < 256 ? n : 0; + } + break; + case 'p': + ptr = _m3d_findarg(ptr); + if(!*ptr || *ptr == '\r' || *ptr == '\n') goto asciiend; + ptr = _m3d_getint(ptr, &n); + model->voxel[i].x = n; + ptr = _m3d_findarg(ptr); + if(!*ptr || *ptr == '\r' || *ptr == '\n') goto asciiend; + ptr = _m3d_getint(ptr, &n); + model->voxel[i].y = n; + ptr = _m3d_findarg(ptr); + if(!*ptr || *ptr == '\r' || *ptr == '\n') goto asciiend; + ptr = _m3d_getint(ptr, &n); + model->voxel[i].z = n; + break; + case 'd': + ptr = _m3d_findarg(ptr); + if(!*ptr || *ptr == '\r' || *ptr == '\n') goto asciiend; + ptr = _m3d_getint(ptr, &n); + model->voxel[i].w = n; + ptr = _m3d_findarg(ptr); + if(!*ptr || *ptr == '\r' || *ptr == '\n') goto asciiend; + ptr = _m3d_getint(ptr, &n); + model->voxel[i].h = n; + ptr = _m3d_findarg(ptr); + if(!*ptr || *ptr == '\r' || *ptr == '\n') goto asciiend; + ptr = _m3d_getint(ptr, &n); + model->voxel[i].d = n; + break; + case 'l': + if(model->voxel[i].data) { l++; k = 0; } + else { + if(!model->voxel[i].w || !model->voxel[i].h || !model->voxel[i].d) { + M3D_LOG("No voxel dimension before layer data"); + goto asciiend; + } + model->voxel[i].data = (M3D_VOXEL*)M3D_MALLOC( + model->voxel[i].w * model->voxel[i].h * model->voxel[i].d * sizeof(M3D_VOXEL)); + if(!model->voxel[i].data) goto memerr; + } + break; + default: + if(!model->voxel[i].data || l >= model->voxel[i].h || k >= model->voxel[i].d) { + M3D_LOG("Missing voxel attributes or out of bound data"); + goto asciiend; + } + for(n = l * model->voxel[i].w * model->voxel[i].d + k * model->voxel[i].w; + j < model->voxel[i].w && *ptr && *ptr != '\r' && *ptr != '\n'; j++) { + ptr = _m3d_getint(ptr, &am); + if(am >= model->numvoxtype) goto asciiend; + model->voxel[i].data[n + j] = am; + } + k++; + break; + } + ptr = _m3d_findnl(ptr); + } + } else + /* mathematical shape */ + if(!memcmp(pe, "Shape", 5)) { + pe = _m3d_findarg(pe); + if(!*pe || *pe == '\r' || *pe == '\n') goto asciiend; + pe = _m3d_safestr(pe, 0); + if(!pe || !*pe) goto asciiend; + i = model->numshape++; + model->shape = (m3dh_t*)M3D_REALLOC(model->shape, model->numshape * sizeof(m3ds_t)); + if(!model->shape) goto memerr; + h = &model->shape[i]; + h->name = pe; + h->group = M3D_UNDEF; + h->numcmd = 0; + h->cmd = NULL; + while(*ptr && *ptr != '\r' && *ptr != '\n') { + if(!memcmp(ptr, "group", 5)) { + ptr = _m3d_findarg(ptr); + ptr = _m3d_getint(ptr, &h->group); + ptr = _m3d_findnl(ptr); + if(h->group != M3D_UNDEF && h->group >= model->numbone) { + M3D_LOG("Unknown bone id as shape group in shape"); + M3D_LOG(pe); + h->group = M3D_UNDEF; + model->errcode = M3D_ERR_SHPE; + } + continue; + } + for(cd = NULL, k = 0; k < (unsigned int)(sizeof(m3d_commandtypes)/sizeof(m3d_commandtypes[0])); k++) { + j = (unsigned int)strlen(m3d_commandtypes[k].key); + if(!memcmp(ptr, m3d_commandtypes[k].key, j) && (ptr[j] == ' ' || ptr[j] == '\r' || ptr[j] == '\n')) + { cd = &m3d_commandtypes[k]; break; } + } + if(cd) { + j = h->numcmd++; + h->cmd = (m3dc_t*)M3D_REALLOC(h->cmd, h->numcmd * sizeof(m3dc_t)); + if(!h->cmd) goto memerr; + h->cmd[j].type = k; + h->cmd[j].arg = (uint32_t*)M3D_MALLOC(cd->p * sizeof(uint32_t)); + if(!h->cmd[j].arg) goto memerr; + memset(h->cmd[j].arg, 0, cd->p * sizeof(uint32_t)); + for(k = n = 0, l = cd->p; k < l; k++) { + ptr = _m3d_findarg(ptr); + if(!*ptr) goto asciiend; + if(*ptr == '[') { + ptr = _m3d_findarg(ptr + 1); + if(!*ptr) goto asciiend; + } + if(*ptr == ']' || *ptr == '\r' || *ptr == '\n') break; + switch(cd->a[((k - n) % (cd->p - n)) + n]) { + case m3dcp_mi_t: + mi = M3D_UNDEF; + if(*ptr != '\r' && *ptr != '\n') { + pe = _m3d_safestr(ptr, 0); + if(!pe || !*pe) goto asciiend; + for(n = 0; n < model->nummaterial; n++) + if(!strcmp(pe, model->material[n].name)) { mi = (M3D_INDEX)n; break; } + if(mi == M3D_UNDEF && !(model->flags & M3D_FLG_MTLLIB)) { + mi = model->nummaterial++; + model->material = (m3dm_t*)M3D_REALLOC(model->material, + model->nummaterial * sizeof(m3dm_t)); + if(!model->material) goto memerr; + model->material[mi].name = pe; + model->material[mi].numprop = 1; + model->material[mi].prop = NULL; + } else + M3D_FREE(pe); + } + h->cmd[j].arg[k] = mi; + break; + case m3dcp_vc_t: +#ifdef M3D_DOUBLE + _m3d_getfloat(ptr, &w); f = w; + memcpy(&h->cmd[j].arg[k], &f, 4); +#else + _m3d_getfloat(ptr, (float*)&h->cmd[j].arg[k]); +#endif + break; + case m3dcp_va_t: + ptr = _m3d_getint(ptr, &h->cmd[j].arg[k]); + n = k + 1; l += (h->cmd[j].arg[k] - 1) * (cd->p - k - 1); + h->cmd[j].arg = (uint32_t*)M3D_REALLOC(h->cmd[j].arg, l * sizeof(uint32_t)); + if(!h->cmd[j].arg) goto memerr; + memset(&h->cmd[j].arg[k + 1], 0, (l - k - 1) * sizeof(uint32_t)); + break; + case m3dcp_qi_t: + ptr = _m3d_getint(ptr, &h->cmd[j].arg[k]); + model->vertex[h->cmd[i].arg[k]].skinid = M3D_INDEXMAX; + break; + default: + ptr = _m3d_getint(ptr, &h->cmd[j].arg[k]); + break; + } + } + } else { + M3D_LOG("Unknown shape command in"); + M3D_LOG(h->name); + model->errcode = M3D_ERR_UNKCMD; + } + ptr = _m3d_findnl(ptr); + } + if(!h->numcmd) model->numshape--; + } else + /* annotation labels */ + if(!memcmp(pe, "Labels", 6)) { + pe = _m3d_findarg(pe); + if(!*pe) goto asciiend; + if(*pe == '\r' || *pe == '\n') pe = NULL; + else pe = _m3d_safestr(pe, 0); + k = 0; fn = NULL; + while(*ptr && *ptr != '\r' && *ptr != '\n') { + if(*ptr == 'c') { + ptr = _m3d_findarg(ptr); + if(!*pe || *pe == '\r' || *pe == '\n') goto asciiend; + ptr = _m3d_gethex(ptr, &k); + } else + if(*ptr == 'l') { + ptr = _m3d_findarg(ptr); + if(!*pe || *pe == '\r' || *pe == '\n') goto asciiend; + fn = _m3d_safestr(ptr, 2); + } else { + i = model->numlabel++; + model->label = (m3dl_t*)M3D_REALLOC(model->label, model->numlabel * sizeof(m3dl_t)); + if(!model->label) goto memerr; + model->label[i].name = pe; + model->label[i].lang = fn; + model->label[i].color = k; + ptr = _m3d_getint(ptr, &j); + model->label[i].vertexid = (M3D_INDEX)j; + ptr = _m3d_findarg(ptr); + if(!*pe || *pe == '\r' || *pe == '\n') goto asciiend; + model->label[i].text = _m3d_safestr(ptr, 2); + } + ptr = _m3d_findnl(ptr); + } + } else + /* action */ + if(!memcmp(pe, "Action", 6)) { + pe = _m3d_findarg(pe); + if(!*pe || *pe == '\r' || *pe == '\n') goto asciiend; + pe = _m3d_getint(pe, &k); + pe = _m3d_findarg(pe); + if(!*pe || *pe == '\r' || *pe == '\n') goto asciiend; + pe = _m3d_safestr(pe, 0); + if(!pe || !*pe) goto asciiend; + i = model->numaction++; + model->action = (m3da_t*)M3D_REALLOC(model->action, model->numaction * sizeof(m3da_t)); + if(!model->action) goto memerr; + a = &model->action[i]; + a->name = pe; + a->durationmsec = k; + /* skip the first frame marker as there's always at least one frame */ + a->numframe = 1; + a->frame = (m3dfr_t*)M3D_MALLOC(sizeof(m3dfr_t)); + if(!a->frame) goto memerr; + a->frame[0].msec = 0; + a->frame[0].numtransform = 0; + a->frame[0].transform = NULL; + i = 0; + if(*ptr == 'f') + ptr = _m3d_findnl(ptr); + while(*ptr && *ptr != '\r' && *ptr != '\n') { + if(*ptr == 'f') { + i = a->numframe++; + a->frame = (m3dfr_t*)M3D_REALLOC(a->frame, a->numframe * sizeof(m3dfr_t)); + if(!a->frame) goto memerr; + ptr = _m3d_findarg(ptr); + ptr = _m3d_getint(ptr, &a->frame[i].msec); + a->frame[i].numtransform = 0; + a->frame[i].transform = NULL; + } else { + j = a->frame[i].numtransform++; + a->frame[i].transform = (m3dtr_t*)M3D_REALLOC(a->frame[i].transform, + a->frame[i].numtransform * sizeof(m3dtr_t)); + if(!a->frame[i].transform) goto memerr; + ptr = _m3d_getint(ptr, &k); + ptr = _m3d_findarg(ptr); + if(!*ptr || *ptr == '\r' || *ptr == '\n') goto asciiend; + a->frame[i].transform[j].boneid = (M3D_INDEX)k; + ptr = _m3d_getint(ptr, &k); + ptr = _m3d_findarg(ptr); + if(!*ptr || *ptr == '\r' || *ptr == '\n') goto asciiend; + a->frame[i].transform[j].pos = (M3D_INDEX)k; + ptr = _m3d_getint(ptr, &k); + if(!*ptr || *ptr == '\r' || *ptr == '\n') goto asciiend; + a->frame[i].transform[j].ori = (M3D_INDEX)k; + model->vertex[k].skinid = M3D_INDEXMAX; + } + ptr = _m3d_findnl(ptr); + } + } else + /* inlined assets chunk */ + if(!memcmp(pe, "Assets", 6)) { + while(*ptr && *ptr != '\r' && *ptr != '\n') { + if(readfilecb) { + pe = _m3d_safestr(ptr, 2); + if(!pe || !*pe) goto asciiend; + i = model->numinlined++; + model->inlined = (m3di_t*)M3D_REALLOC(model->inlined, model->numinlined * sizeof(m3di_t)); + if(!model->inlined) goto memerr; + t = &model->inlined[i]; + model->inlined[i].data = (*readfilecb)(pe, &model->inlined[i].length); + if(model->inlined[i].data) { + fn = strrchr(pe, '.'); + if(fn && (fn[1] == 'p' || fn[1] == 'P') && (fn[2] == 'n' || fn[2] == 'N') && + (fn[3] == 'g' || fn[3] == 'G')) *fn = 0; + fn = strrchr(pe, '/'); + if(!fn) fn = strrchr(pe, '\\'); + if(!fn) fn = pe; else fn++; + model->inlined[i].name = _m3d_safestr(fn, 0); + } else + model->numinlined--; + M3D_FREE(pe); + } + ptr = _m3d_findnl(ptr); + } + } else + /* extra chunks */ + if(!memcmp(pe, "Extra", 5)) { + pe = _m3d_findarg(pe); + if(!*pe || *pe == '\r' || *pe == '\n') goto asciiend; + buff = (unsigned char*)_m3d_findnl(ptr); + k = ((uint32_t)((uintptr_t)buff - (uintptr_t)ptr) / 3) + 1; + i = model->numextra++; + model->extra = (m3dchunk_t**)M3D_REALLOC(model->extra, model->numextra * sizeof(m3dchunk_t*)); + if(!model->extra) goto memerr; + model->extra[i] = (m3dchunk_t*)M3D_MALLOC(k + sizeof(m3dchunk_t)); + if(!model->extra[i]) goto memerr; + memcpy(&model->extra[i]->magic, pe, 4); + model->extra[i]->length = sizeof(m3dchunk_t); + pe = (char*)model->extra[i] + sizeof(m3dchunk_t); + while(*ptr && *ptr != '\r' && *ptr != '\n') { + ptr = _m3d_gethex(ptr, &k); + *pe++ = (uint8_t)k; + model->extra[i]->length++; + } + } else + goto asciiend; + } + model->errcode = M3D_SUCCESS; +asciiend: + setlocale(LC_NUMERIC, ol); + goto postprocess; + } +#endif + /* Binary variant */ + len = ((m3dhdr_t*)data)->length - 8; + data += 8; + if(M3D_CHUNKMAGIC(data, 'P','R','V','W')) { + /* optional preview chunk */ + model->preview.length = ((m3dchunk_t*)data)->length; + model->preview.data = data + sizeof(m3dchunk_t); + data += model->preview.length; + len -= model->preview.length; + } + if(!M3D_CHUNKMAGIC(data, 'H','E','A','D')) { + buff = (unsigned char *)stbi_zlib_decode_malloc_guesssize_headerflag((const char*)data, len, 4096, (int*)&len, 1); + if(!buff || !len || !M3D_CHUNKMAGIC(buff, 'H','E','A','D')) { + if(buff) M3D_FREE(buff); + M3D_FREE(model); + return NULL; + } + buff = (unsigned char*)M3D_REALLOC(buff, len); + model->flags |= M3D_FLG_FREERAW; /* mark that we have to free the raw buffer */ + data = buff; +#ifdef M3D_PROFILING + gettimeofday(&tv1, NULL); + tvd.tv_sec = tv1.tv_sec - tv0.tv_sec; + tvd.tv_usec = tv1.tv_usec - tv0.tv_usec; + if(tvd.tv_usec < 0) { tvd.tv_sec--; tvd.tv_usec += 1000000L; } + printf(" Deflate model %ld.%06ld sec\n", tvd.tv_sec, tvd.tv_usec); + memcpy(&tv0, &tv1, sizeof(struct timeval)); +#endif + } + model->raw = (m3dhdr_t*)data; + end = data + len; + + /* parse header */ + data += sizeof(m3dhdr_t); + M3D_LOG((char*)data); + model->name = (char*)data; + for(; data < end && *data; data++) {}; data++; + model->license = (char*)data; + for(; data < end && *data; data++) {}; data++; + model->author = (char*)data; + for(; data < end && *data; data++) {}; data++; + model->desc = (char*)data; + chunk = (unsigned char*)model->raw + model->raw->length; + model->scale = (M3D_FLOAT)model->raw->scale; + if(model->scale <= (M3D_FLOAT)0.0) model->scale = (M3D_FLOAT)1.0; + model->vc_s = 1 << ((model->raw->types >> 0) & 3); /* vertex coordinate size */ + model->vi_s = 1 << ((model->raw->types >> 2) & 3); /* vertex index size */ + model->si_s = 1 << ((model->raw->types >> 4) & 3); /* string offset size */ + model->ci_s = 1 << ((model->raw->types >> 6) & 3); /* color index size */ + model->ti_s = 1 << ((model->raw->types >> 8) & 3); /* tmap index size */ + model->bi_s = 1 << ((model->raw->types >>10) & 3); /* bone index size */ + model->nb_s = 1 << ((model->raw->types >>12) & 3); /* number of bones per vertex */ + model->sk_s = 1 << ((model->raw->types >>14) & 3); /* skin index size */ + model->fc_s = 1 << ((model->raw->types >>16) & 3); /* frame counter size */ + model->hi_s = 1 << ((model->raw->types >>18) & 3); /* shape index size */ + model->fi_s = 1 << ((model->raw->types >>20) & 3); /* face index size */ + model->vd_s = 1 << ((model->raw->types >>22) & 3); /* voxel dimension size */ + model->vp_s = 1 << ((model->raw->types >>24) & 3); /* voxel pixel size */ + if(model->ci_s == 8) model->ci_s = 0; /* optional indices */ + if(model->ti_s == 8) model->ti_s = 0; + if(model->bi_s == 8) model->bi_s = 0; + if(model->sk_s == 8) model->sk_s = 0; + if(model->fc_s == 8) model->fc_s = 0; + if(model->hi_s == 8) model->hi_s = 0; + if(model->fi_s == 8) model->fi_s = 0; + + /* variable limit checks */ + if(sizeof(M3D_FLOAT) == 4 && model->vc_s > 4) { + M3D_LOG("Double precision coordinates not supported, truncating to float..."); + model->errcode = M3D_ERR_TRUNC; + } + if((sizeof(M3D_INDEX) == 2 && (model->vi_s > 2 || model->si_s > 2 || model->ci_s > 2 || model->ti_s > 2 || + model->bi_s > 2 || model->sk_s > 2 || model->fc_s > 2 || model->hi_s > 2 || model->fi_s > 2)) || + (sizeof(M3D_VOXEL) < (size_t)model->vp_s && model->vp_s != 8)) { + M3D_LOG("32 bit indices not supported, unable to load model"); + M3D_FREE(model); + return NULL; + } + if(model->vi_s > 4 || model->si_s > 4 || model->vp_s == 4) { + M3D_LOG("Invalid index size, unable to load model"); + M3D_FREE(model); + return NULL; + } + if(!M3D_CHUNKMAGIC(end - 4, 'O','M','D','3')) { + M3D_LOG("Missing end chunk"); + M3D_FREE(model); + return NULL; + } + if(model->nb_s > M3D_NUMBONE) { + M3D_LOG("Model has more bones per vertex than what importer was configured to support"); + model->errcode = M3D_ERR_TRUNC; + } + + /* look for inlined assets in advance, material and procedural chunks may need them */ + buff = chunk; + while(buff < end && !M3D_CHUNKMAGIC(buff, 'O','M','D','3')) { + data = buff; + len = ((m3dchunk_t*)data)->length; + buff += len; + if(len < sizeof(m3dchunk_t) || buff >= end) { + M3D_LOG("Invalid chunk size"); + break; + } + len -= sizeof(m3dchunk_t) + model->si_s; + + /* inlined assets */ + if(M3D_CHUNKMAGIC(data, 'A','S','E','T') && len > 0) { + M3D_LOG("Inlined asset"); + i = model->numinlined++; + model->inlined = (m3di_t*)M3D_REALLOC(model->inlined, model->numinlined * sizeof(m3di_t)); + if(!model->inlined) { +memerr: M3D_LOG("Out of memory"); + model->errcode = M3D_ERR_ALLOC; + return model; + } + data += sizeof(m3dchunk_t); + t = &model->inlined[i]; + M3D_GETSTR(t->name); + M3D_LOG(t->name); + t->data = (uint8_t*)data; + t->length = len; + } + } + + /* parse chunks */ + while(chunk < end && !M3D_CHUNKMAGIC(chunk, 'O','M','D','3')) { + data = chunk; + len = ((m3dchunk_t*)chunk)->length; + chunk += len; + if(len < sizeof(m3dchunk_t) || chunk >= end) { + M3D_LOG("Invalid chunk size"); + break; + } + len -= sizeof(m3dchunk_t); + + /* color map */ + if(M3D_CHUNKMAGIC(data, 'C','M','A','P')) { + M3D_LOG("Color map"); + if(model->cmap) { M3D_LOG("More color map chunks, should be unique"); model->errcode = M3D_ERR_CMAP; continue; } + if(!model->ci_s) { M3D_LOG("Color map chunk, shouldn't be any"); model->errcode = M3D_ERR_CMAP; continue; } + model->numcmap = len / sizeof(uint32_t); + model->cmap = (uint32_t*)(data + sizeof(m3dchunk_t)); + } else + /* texture map */ + if(M3D_CHUNKMAGIC(data, 'T','M','A','P')) { + M3D_LOG("Texture map"); + if(model->tmap) { M3D_LOG("More texture map chunks, should be unique"); model->errcode = M3D_ERR_TMAP; continue; } + if(!model->ti_s) { M3D_LOG("Texture map chunk, shouldn't be any"); model->errcode = M3D_ERR_TMAP; continue; } + reclen = model->vc_s + model->vc_s; + model->numtmap = len / reclen; + model->tmap = (m3dti_t*)M3D_MALLOC(model->numtmap * sizeof(m3dti_t)); + if(!model->tmap) goto memerr; + for(i = 0, data += sizeof(m3dchunk_t); data < chunk; i++) { + switch(model->vc_s) { + case 1: + model->tmap[i].u = (M3D_FLOAT)((uint8_t)data[0]) / (M3D_FLOAT)255.0; + model->tmap[i].v = (M3D_FLOAT)((uint8_t)data[1]) / (M3D_FLOAT)255.0; + break; + case 2: + model->tmap[i].u = (M3D_FLOAT)(*((uint16_t*)(data+0))) / (M3D_FLOAT)65535.0; + model->tmap[i].v = (M3D_FLOAT)(*((uint16_t*)(data+2))) / (M3D_FLOAT)65535.0; + break; + case 4: + model->tmap[i].u = (M3D_FLOAT)(*((float*)(data+0))); + model->tmap[i].v = (M3D_FLOAT)(*((float*)(data+4))); + break; + case 8: + model->tmap[i].u = (M3D_FLOAT)(*((double*)(data+0))); + model->tmap[i].v = (M3D_FLOAT)(*((double*)(data+8))); + break; + } + data += reclen; + } + } else + /* vertex list */ + if(M3D_CHUNKMAGIC(data, 'V','R','T','S')) { + M3D_LOG("Vertex list"); + if(model->vertex) { M3D_LOG("More vertex chunks, should be unique"); model->errcode = M3D_ERR_VRTS; continue; } + if(model->ci_s && model->ci_s < 4 && !model->cmap) model->errcode = M3D_ERR_CMAP; + reclen = model->ci_s + model->sk_s + 4 * model->vc_s; + model->numvertex = len / reclen; + model->vertex = (m3dv_t*)M3D_MALLOC(model->numvertex * sizeof(m3dv_t)); + if(!model->vertex) goto memerr; + memset(model->vertex, 0, model->numvertex * sizeof(m3dv_t)); + for(i = 0, data += sizeof(m3dchunk_t); data < chunk && i < model->numvertex; i++) { + switch(model->vc_s) { + case 1: + model->vertex[i].x = (M3D_FLOAT)((int8_t)data[0]) / (M3D_FLOAT)127.0; + model->vertex[i].y = (M3D_FLOAT)((int8_t)data[1]) / (M3D_FLOAT)127.0; + model->vertex[i].z = (M3D_FLOAT)((int8_t)data[2]) / (M3D_FLOAT)127.0; + model->vertex[i].w = (M3D_FLOAT)((int8_t)data[3]) / (M3D_FLOAT)127.0; + data += 4; + break; + case 2: + model->vertex[i].x = (M3D_FLOAT)(*((int16_t*)(data+0))) / (M3D_FLOAT)32767.0; + model->vertex[i].y = (M3D_FLOAT)(*((int16_t*)(data+2))) / (M3D_FLOAT)32767.0; + model->vertex[i].z = (M3D_FLOAT)(*((int16_t*)(data+4))) / (M3D_FLOAT)32767.0; + model->vertex[i].w = (M3D_FLOAT)(*((int16_t*)(data+6))) / (M3D_FLOAT)32767.0; + data += 8; + break; + case 4: + model->vertex[i].x = (M3D_FLOAT)(*((float*)(data+0))); + model->vertex[i].y = (M3D_FLOAT)(*((float*)(data+4))); + model->vertex[i].z = (M3D_FLOAT)(*((float*)(data+8))); + model->vertex[i].w = (M3D_FLOAT)(*((float*)(data+12))); + data += 16; + break; + case 8: + model->vertex[i].x = (M3D_FLOAT)(*((double*)(data+0))); + model->vertex[i].y = (M3D_FLOAT)(*((double*)(data+8))); + model->vertex[i].z = (M3D_FLOAT)(*((double*)(data+16))); + model->vertex[i].w = (M3D_FLOAT)(*((double*)(data+24))); + data += 32; + break; + } + switch(model->ci_s) { + case 1: model->vertex[i].color = model->cmap ? model->cmap[data[0]] : 0; data++; break; + case 2: model->vertex[i].color = model->cmap ? model->cmap[*((uint16_t*)data)] : 0; data += 2; break; + case 4: model->vertex[i].color = *((uint32_t*)data); data += 4; break; + /* case 8: break; */ + } + model->vertex[i].skinid = M3D_UNDEF; + data = _m3d_getidx(data, model->sk_s, &model->vertex[i].skinid); + } + } else + /* skeleton: bone hierarchy and skin */ + if(M3D_CHUNKMAGIC(data, 'B','O','N','E')) { + M3D_LOG("Skeleton"); + if(model->bone) { M3D_LOG("More bone chunks, should be unique"); model->errcode = M3D_ERR_BONE; continue; } + if(!model->bi_s) { M3D_LOG("Bone chunk, shouldn't be any"); model->errcode=M3D_ERR_BONE; continue; } + if(!model->vertex) { M3D_LOG("No vertex chunk before bones"); model->errcode = M3D_ERR_VRTS; break; } + data += sizeof(m3dchunk_t); + model->numbone = 0; + data = _m3d_getidx(data, model->bi_s, &model->numbone); + if(model->numbone) { + model->bone = (m3db_t*)M3D_MALLOC(model->numbone * sizeof(m3db_t)); + if(!model->bone) goto memerr; + } + model->numskin = 0; + data = _m3d_getidx(data, model->sk_s, &model->numskin); + /* read bone hierarchy */ + for(i = 0; data < chunk && i < model->numbone; i++) { + data = _m3d_getidx(data, model->bi_s, &model->bone[i].parent); + M3D_GETSTR(model->bone[i].name); + data = _m3d_getidx(data, model->vi_s, &model->bone[i].pos); + data = _m3d_getidx(data, model->vi_s, &model->bone[i].ori); + model->bone[i].numweight = 0; + model->bone[i].weight = NULL; + } + if(i != model->numbone) { M3D_LOG("Truncated bone chunk"); model->numbone = i; model->numskin = 0; model->errcode = M3D_ERR_BONE; } + /* read skin definitions */ + if(model->numskin) { + model->skin = (m3ds_t*)M3D_MALLOC(model->numskin * sizeof(m3ds_t)); + if(!model->skin) goto memerr; + for(i = 0; data < chunk && i < model->numskin; i++) { + for(j = 0; j < M3D_NUMBONE; j++) { + model->skin[i].boneid[j] = M3D_UNDEF; + model->skin[i].weight[j] = (M3D_FLOAT)0.0; + } + memset(&weights, 0, sizeof(weights)); + if(model->nb_s == 1) weights[0] = 255; + else { + memcpy(&weights, data, model->nb_s); + data += model->nb_s; + } + for(j = 0, w = (M3D_FLOAT)0.0; j < (unsigned int)model->nb_s; j++) { + if(weights[j]) { + if(j >= M3D_NUMBONE) + data += model->bi_s; + else { + model->skin[i].weight[j] = (M3D_FLOAT)(weights[j]) / (M3D_FLOAT)255.0; + w += model->skin[i].weight[j]; + data = _m3d_getidx(data, model->bi_s, &model->skin[i].boneid[j]); + } + } + } + /* this can occur if model has more bones than what the importer is configured to handle */ + if(w != (M3D_FLOAT)1.0 && w != (M3D_FLOAT)0.0) { + for(j = 0; j < M3D_NUMBONE; j++) + model->skin[i].weight[j] /= w; + } + } + if(i != model->numskin) { M3D_LOG("Truncated skin in bone chunk"); model->numskin = i; model->errcode = M3D_ERR_BONE; } + } + } else + /* material */ + if(M3D_CHUNKMAGIC(data, 'M','T','R','L')) { + data += sizeof(m3dchunk_t); + M3D_GETSTR(name); + M3D_LOG("Material"); + M3D_LOG(name); + if(model->ci_s < 4 && !model->numcmap) model->errcode = M3D_ERR_CMAP; + for(i = 0; i < model->nummaterial; i++) + if(!strcmp(name, model->material[i].name)) { + model->errcode = M3D_ERR_MTRL; + M3D_LOG("Multiple definitions for material"); + M3D_LOG(name); + name = NULL; + break; + } + if(name) { + i = model->nummaterial++; + if(model->flags & M3D_FLG_MTLLIB) { + m = model->material; + model->material = (m3dm_t*)M3D_MALLOC(model->nummaterial * sizeof(m3dm_t)); + if(!model->material) goto memerr; + memcpy(model->material, m, (model->nummaterial - 1) * sizeof(m3dm_t)); + if(model->texture) { + tx = model->texture; + model->texture = (m3dtx_t*)M3D_MALLOC(model->numtexture * sizeof(m3dtx_t)); + if(!model->texture) goto memerr; + memcpy(model->texture, tx, model->numtexture * sizeof(m3dm_t)); + } + model->flags &= ~M3D_FLG_MTLLIB; + } else { + model->material = (m3dm_t*)M3D_REALLOC(model->material, model->nummaterial * sizeof(m3dm_t)); + if(!model->material) goto memerr; + } + m = &model->material[i]; + m->numprop = 0; + m->name = name; + m->prop = (m3dp_t*)M3D_MALLOC((len / 2) * sizeof(m3dp_t)); + if(!m->prop) goto memerr; + while(data < chunk) { + i = m->numprop++; + m->prop[i].type = *data++; + m->prop[i].value.num = 0; + if(m->prop[i].type >= 128) + k = m3dpf_map; + else { + for(k = 256, j = 0; j < sizeof(m3d_propertytypes)/sizeof(m3d_propertytypes[0]); j++) + if(m->prop[i].type == m3d_propertytypes[j].id) { k = m3d_propertytypes[j].format; break; } + } + switch(k) { + case m3dpf_color: + switch(model->ci_s) { + case 1: m->prop[i].value.color = model->cmap ? model->cmap[data[0]] : 0; data++; break; + case 2: m->prop[i].value.color = model->cmap ? model->cmap[*((uint16_t*)data)] : 0; data += 2; break; + case 4: m->prop[i].value.color = *((uint32_t*)data); data += 4; break; + } + break; + + case m3dpf_uint8: m->prop[i].value.num = *data++; break; + case m3dpf_uint16:m->prop[i].value.num = *((uint16_t*)data); data += 2; break; + case m3dpf_uint32:m->prop[i].value.num = *((uint32_t*)data); data += 4; break; + case m3dpf_float: m->prop[i].value.fnum = *((float*)data); data += 4; break; + + case m3dpf_map: + M3D_GETSTR(name); + m->prop[i].value.textureid = _m3d_gettx(model, readfilecb, freecb, name); + if(model->errcode == M3D_ERR_ALLOC) goto memerr; + /* this error code only returned if readfilecb was specified */ + if(m->prop[i].value.textureid == M3D_UNDEF) { + M3D_LOG("Texture not found"); + M3D_LOG(m->name); + m->numprop--; + } + break; + + default: + M3D_LOG("Unknown material property in"); + M3D_LOG(m->name); + model->errcode = M3D_ERR_UNKPROP; + data = chunk; + break; + } + } + m->prop = (m3dp_t*)M3D_REALLOC(m->prop, m->numprop * sizeof(m3dp_t)); + if(!m->prop) goto memerr; + } + } else + /* face */ + if(M3D_CHUNKMAGIC(data, 'P','R','O','C')) { + /* procedural surface */ + M3D_GETSTR(name); + M3D_LOG("Procedural surface"); + M3D_LOG(name); + _m3d_getpr(model, readfilecb, freecb, name); + } else + if(M3D_CHUNKMAGIC(data, 'M','E','S','H')) { + M3D_LOG("Mesh data"); + if(!model->vertex) { M3D_LOG("No vertex chunk before mesh"); model->errcode = M3D_ERR_VRTS; } + /* mesh */ + data += sizeof(m3dchunk_t); + mi = M3D_UNDEF; +#ifdef M3D_VERTEXMAX + pi = M3D_UNDEF; +#endif + am = model->numface; + while(data < chunk) { + k = *data++; + n = k >> 4; + k &= 15; + if(!n) { + if(!k) { + /* use material */ + mi = M3D_UNDEF; + M3D_GETSTR(name); + if(name) { + for(j = 0; j < model->nummaterial; j++) + if(!strcmp(name, model->material[j].name)) { + mi = (M3D_INDEX)j; + break; + } + if(mi == M3D_UNDEF) model->errcode = M3D_ERR_MTRL; + } + } else { + /* use parameter */ + M3D_GETSTR(name); +#ifdef M3D_VERTEXMAX + pi = M3D_UNDEF; + if(name) { + for(j = 0; j < model->numparam; j++) + if(!strcmp(name, model->param[j].name)) { + pi = (M3D_INDEX)j; + break; + } + if(pi == M3D_UNDEF) { + pi = model->numparam++; + model->param = (m3dvi_t*)M3D_REALLOC(model->param, model->numparam * sizeof(m3dvi_t)); + if(!model->param) goto memerr; + model->param[pi].name = name; + model->param[pi].count = 0; + } + } +#endif + } + continue; + } + if(n != 3) { M3D_LOG("Only triangle mesh supported for now"); model->errcode = M3D_ERR_UNKMESH; return model; } + i = model->numface++; + if(model->numface > am) { + am = model->numface + 4095; + model->face = (m3df_t*)M3D_REALLOC(model->face, am * sizeof(m3df_t)); + if(!model->face) goto memerr; + } + memset(&model->face[i], 255, sizeof(m3df_t)); /* set all index to -1 by default */ + model->face[i].materialid = mi; +#ifdef M3D_VERTEXMAX + model->face[i].paramid = pi; +#endif + for(j = 0; data < chunk && j < n; j++) { + /* vertex */ + data = _m3d_getidx(data, model->vi_s, &model->face[i].vertex[j]); + /* texcoord */ + if(k & 1) + data = _m3d_getidx(data, model->ti_s, &model->face[i].texcoord[j]); + /* normal */ + if(k & 2) + data = _m3d_getidx(data, model->vi_s, &model->face[i].normal[j]); +#ifndef M3D_NONORMALS + if(model->face[i].normal[j] == M3D_UNDEF) neednorm = 1; +#endif + /* maximum */ + if(k & 4) +#ifdef M3D_VERTEXMAX + data = _m3d_getidx(data, model->vi_s, &model->face[i].vertmax[j]); +#else + data += model->vi_s; +#endif + } + if(j != n) { M3D_LOG("Invalid mesh"); model->numface = 0; model->errcode = M3D_ERR_UNKMESH; return model; } + } + model->face = (m3df_t*)M3D_REALLOC(model->face, model->numface * sizeof(m3df_t)); + } else + if(M3D_CHUNKMAGIC(data, 'V','O','X','T')) { + /* voxel types */ + M3D_LOG("Voxel types list"); + if(model->voxtype) { M3D_LOG("More voxel type chunks, should be unique"); model->errcode = M3D_ERR_VOXT; continue; } + if(model->ci_s && model->ci_s < 4 && !model->cmap) model->errcode = M3D_ERR_CMAP; + reclen = model->ci_s + model->si_s + 3 + model->sk_s; + k = len / reclen; + model->voxtype = (m3dvt_t*)M3D_MALLOC(k * sizeof(m3dvt_t)); + if(!model->voxtype) goto memerr; + memset(model->voxtype, 0, k * sizeof(m3dvt_t)); + model->numvoxtype = 0; + for(i = 0, data += sizeof(m3dchunk_t); data < chunk && i < k; i++) { + switch(model->ci_s) { + case 1: model->voxtype[i].color = model->cmap ? model->cmap[data[0]] : 0; data++; break; + case 2: model->voxtype[i].color = model->cmap ? model->cmap[*((uint16_t*)data)] : 0; data += 2; break; + case 4: model->voxtype[i].color = *((uint32_t*)data); data += 4; break; + /* case 8: break; */ + } + M3D_GETSTR(name); + model->voxtype[i].materialid = M3D_UNDEF; + if(name) { + model->voxtype[i].name = name; +/* + for(j = 0; j < model->nummaterial; j++) + if(!strcmp(name, model->material[j].name)) { + model->voxtype[i].materialid = (M3D_INDEX)j; + break; + } +*/ + } + j = *data++; + model->voxtype[i].rotation = j & 0xBF; + model->voxtype[i].voxshape = ((j & 0x40) << 2) | *data++; + model->voxtype[i].numitem = *data++; + model->voxtype[i].skinid = M3D_UNDEF; + data = _m3d_getidx(data, model->sk_s, &model->voxtype[i].skinid); + if(model->voxtype[i].numitem) { + model->voxtype[i].item = (m3dvi_t*)M3D_MALLOC(model->voxtype[i].numitem * sizeof(m3dvi_t)); + if(!model->voxtype[i].item) goto memerr; + memset(model->voxtype[i].item, 0, model->voxtype[i].numitem * sizeof(m3dvi_t)); + for(j = 0; j < model->voxtype[i].numitem; j++) { + model->voxtype[i].item[j].count = *data++; + model->voxtype[i].item[j].count |= (*data++) << 8; + M3D_GETSTR(model->voxtype[i].item[j].name); + } + } + } + model->numvoxtype = i; + if(k != model->numvoxtype) { + model->voxtype = (m3dvt_t*)M3D_REALLOC(model->voxtype, model->numvoxtype * sizeof(m3dvt_t)); + if(!model->voxtype) goto memerr; + } + } else + if(M3D_CHUNKMAGIC(data, 'V','O','X','D')) { + /* voxel data */ + data += sizeof(m3dchunk_t); + M3D_GETSTR(name); + M3D_LOG("Voxel Data Layer"); + M3D_LOG(name); + if(model->vd_s > 4 || model->vp_s > 2) { M3D_LOG("No voxel index size"); model->errcode = M3D_ERR_UNKVOX; continue; } + if(!model->voxtype) { M3D_LOG("No voxel type chunk before voxel data"); model->errcode = M3D_ERR_VOXT; } + i = model->numvoxel++; + model->voxel = (m3dvx_t*)M3D_REALLOC(model->voxel, model->numvoxel * sizeof(m3dvx_t)); + if(!model->voxel) goto memerr; + memset(&model->voxel[i], 0, sizeof(m3dvx_t)); + model->voxel[i].name = name; + switch(model->vd_s) { + case 1: + model->voxel[i].x = (int32_t)((int8_t)data[0]); + model->voxel[i].y = (int32_t)((int8_t)data[1]); + model->voxel[i].z = (int32_t)((int8_t)data[2]); + model->voxel[i].w = (uint32_t)(data[3]); + model->voxel[i].h = (uint32_t)(data[4]); + model->voxel[i].d = (uint32_t)(data[5]); + data += 6; + break; + case 2: + model->voxel[i].x = (int32_t)(*((int16_t*)(data+0))); + model->voxel[i].y = (int32_t)(*((int16_t*)(data+2))); + model->voxel[i].z = (int32_t)(*((int16_t*)(data+4))); + model->voxel[i].w = (uint32_t)(*((uint16_t*)(data+6))); + model->voxel[i].h = (uint32_t)(*((uint16_t*)(data+8))); + model->voxel[i].d = (uint32_t)(*((uint16_t*)(data+10))); + data += 12; + break; + case 4: + model->voxel[i].x = *((int32_t*)(data+0)); + model->voxel[i].y = *((int32_t*)(data+4)); + model->voxel[i].z = *((int32_t*)(data+8)); + model->voxel[i].w = *((uint32_t*)(data+12)); + model->voxel[i].h = *((uint32_t*)(data+16)); + model->voxel[i].d = *((uint32_t*)(data+20)); + data += 24; + break; + } + model->voxel[i].uncertain = *data++; + model->voxel[i].groupid = *data++; + k = model->voxel[i].w * model->voxel[i].h * model->voxel[i].d; + model->voxel[i].data = (M3D_VOXEL*)M3D_MALLOC(k * sizeof(M3D_VOXEL)); + if(!model->voxel[i].data) goto memerr; + memset(model->voxel[i].data, 0xff, k * sizeof(M3D_VOXEL)); + for(j = 0; data < chunk && j < k;) { + l = ((*data++) & 0x7F) + 1; + if(data[-1] & 0x80) { + data = _m3d_getidx(data, model->vp_s, &mi); + while(l-- && j < k) model->voxel[i].data[j++] = (M3D_VOXEL)mi; + } else + while(l-- && j < k) { + data = _m3d_getidx(data, model->vp_s, &mi); + model->voxel[i].data[j++] = (M3D_VOXEL)mi; + } + } + } else + if(M3D_CHUNKMAGIC(data, 'S','H','P','E')) { + /* mathematical shape */ + data += sizeof(m3dchunk_t); + M3D_GETSTR(name); + M3D_LOG("Mathematical Shape"); + M3D_LOG(name); + i = model->numshape++; + model->shape = (m3dh_t*)M3D_REALLOC(model->shape, model->numshape * sizeof(m3dh_t)); + if(!model->shape) goto memerr; + h = &model->shape[i]; + h->numcmd = 0; + h->cmd = NULL; + h->name = name; + h->group = M3D_UNDEF; + data = _m3d_getidx(data, model->bi_s, &h->group); + if(h->group != M3D_UNDEF && h->group >= model->numbone) { + M3D_LOG("Unknown bone id as shape group in shape"); + M3D_LOG(name); + h->group = M3D_UNDEF; + model->errcode = M3D_ERR_SHPE; + } + while(data < chunk) { + i = h->numcmd++; + h->cmd = (m3dc_t*)M3D_REALLOC(h->cmd, h->numcmd * sizeof(m3dc_t)); + if(!h->cmd) goto memerr; + h->cmd[i].type = *data++; + if(h->cmd[i].type & 0x80) { + h->cmd[i].type &= 0x7F; + h->cmd[i].type |= (*data++ << 7); + } + if(h->cmd[i].type >= (unsigned int)(sizeof(m3d_commandtypes)/sizeof(m3d_commandtypes[0]))) { + M3D_LOG("Unknown shape command in"); + M3D_LOG(h->name); + model->errcode = M3D_ERR_UNKCMD; + break; + } + cd = &m3d_commandtypes[h->cmd[i].type]; + h->cmd[i].arg = (uint32_t*)M3D_MALLOC(cd->p * sizeof(uint32_t)); + if(!h->cmd[i].arg) goto memerr; + memset(h->cmd[i].arg, 0, cd->p * sizeof(uint32_t)); + for(k = n = 0, l = cd->p; k < l; k++) + switch(cd->a[((k - n) % (cd->p - n)) + n]) { + case m3dcp_mi_t: + h->cmd[i].arg[k] = M3D_NOTDEFINED; + M3D_GETSTR(name); + if(name) { + for(n = 0; n < model->nummaterial; n++) + if(!strcmp(name, model->material[n].name)) { + h->cmd[i].arg[k] = n; + break; + } + if(h->cmd[i].arg[k] == M3D_NOTDEFINED) model->errcode = M3D_ERR_MTRL; + } + break; + case m3dcp_vc_t: + f = 0.0f; + switch(model->vc_s) { + case 1: f = (float)((int8_t)data[0]) / 127; break; + case 2: f = (float)(*((int16_t*)(data+0))) / 32767; break; + case 4: f = (float)(*((float*)(data+0))); break; + case 8: f = (float)(*((double*)(data+0))); break; + } + memcpy(&h->cmd[i].arg[k], &f, 4); + data += model->vc_s; + break; + case m3dcp_hi_t: data = _m3d_getidx(data, model->hi_s, &h->cmd[i].arg[k]); break; + case m3dcp_fi_t: data = _m3d_getidx(data, model->fi_s, &h->cmd[i].arg[k]); break; + case m3dcp_ti_t: data = _m3d_getidx(data, model->ti_s, &h->cmd[i].arg[k]); break; + case m3dcp_qi_t: + case m3dcp_vi_t: data = _m3d_getidx(data, model->vi_s, &h->cmd[i].arg[k]); break; + case m3dcp_i1_t: data = _m3d_getidx(data, 1, &h->cmd[i].arg[k]); break; + case m3dcp_i2_t: data = _m3d_getidx(data, 2, &h->cmd[i].arg[k]); break; + case m3dcp_i4_t: data = _m3d_getidx(data, 4, &h->cmd[i].arg[k]); break; + case m3dcp_va_t: data = _m3d_getidx(data, 4, &h->cmd[i].arg[k]); + n = k + 1; l += (h->cmd[i].arg[k] - 1) * (cd->p - k - 1); + h->cmd[i].arg = (uint32_t*)M3D_REALLOC(h->cmd[i].arg, l * sizeof(uint32_t)); + if(!h->cmd[i].arg) goto memerr; + memset(&h->cmd[i].arg[k + 1], 0, (l - k - 1) * sizeof(uint32_t)); + break; + } + } + } else + /* annotation label list */ + if(M3D_CHUNKMAGIC(data, 'L','B','L','S')) { + data += sizeof(m3dchunk_t); + M3D_GETSTR(name); + M3D_GETSTR(lang); + M3D_LOG("Label list"); + if(name) { M3D_LOG(name); } + if(lang) { M3D_LOG(lang); } + if(model->ci_s && model->ci_s < 4 && !model->cmap) model->errcode = M3D_ERR_CMAP; + k = 0; + switch(model->ci_s) { + case 1: k = model->cmap ? model->cmap[data[0]] : 0; data++; break; + case 2: k = model->cmap ? model->cmap[*((uint16_t*)data)] : 0; data += 2; break; + case 4: k = *((uint32_t*)data); data += 4; break; + /* case 8: break; */ + } + reclen = model->vi_s + model->si_s; + i = model->numlabel; model->numlabel += len / reclen; + model->label = (m3dl_t*)M3D_REALLOC(model->label, model->numlabel * sizeof(m3dl_t)); + if(!model->label) goto memerr; + memset(&model->label[i], 0, (model->numlabel - i) * sizeof(m3dl_t)); + for(; data < chunk && i < model->numlabel; i++) { + model->label[i].name = name; + model->label[i].lang = lang; + model->label[i].color = k; + data = _m3d_getidx(data, model->vi_s, &model->label[i].vertexid); + M3D_GETSTR(model->label[i].text); + } + } else + /* action */ + if(M3D_CHUNKMAGIC(data, 'A','C','T','N')) { + M3D_LOG("Action"); + i = model->numaction++; + model->action = (m3da_t*)M3D_REALLOC(model->action, model->numaction * sizeof(m3da_t)); + if(!model->action) goto memerr; + a = &model->action[i]; + data += sizeof(m3dchunk_t); + M3D_GETSTR(a->name); + M3D_LOG(a->name); + a->numframe = *((uint16_t*)data); data += 2; + if(a->numframe < 1) { + model->numaction--; + } else { + a->durationmsec = *((uint32_t*)data); data += 4; + a->frame = (m3dfr_t*)M3D_MALLOC(a->numframe * sizeof(m3dfr_t)); + if(!a->frame) goto memerr; + for(i = 0; data < chunk && i < a->numframe; i++) { + a->frame[i].msec = *((uint32_t*)data); data += 4; + a->frame[i].numtransform = 0; a->frame[i].transform = NULL; + data = _m3d_getidx(data, model->fc_s, &a->frame[i].numtransform); + if(a->frame[i].numtransform > 0) { + a->frame[i].transform = (m3dtr_t*)M3D_MALLOC(a->frame[i].numtransform * sizeof(m3dtr_t)); + for(j = 0; j < a->frame[i].numtransform; j++) { + data = _m3d_getidx(data, model->bi_s, &a->frame[i].transform[j].boneid); + data = _m3d_getidx(data, model->vi_s, &a->frame[i].transform[j].pos); + data = _m3d_getidx(data, model->vi_s, &a->frame[i].transform[j].ori); + } + } + } + } + } else { + i = model->numextra++; + model->extra = (m3dchunk_t**)M3D_REALLOC(model->extra, model->numextra * sizeof(m3dchunk_t*)); + if(!model->extra) goto memerr; + model->extra[i] = (m3dchunk_t*)data; + } + } + /* calculate normals, normalize skin weights, create bone/vertex cross-references and calculate transform matrices */ +#ifdef M3D_ASCII +postprocess: +#endif + if(model) { + M3D_LOG("Post-process"); +#ifdef M3D_PROFILING + gettimeofday(&tv1, NULL); + tvd.tv_sec = tv1.tv_sec - tv0.tv_sec; + tvd.tv_usec = tv1.tv_usec - tv0.tv_usec; + if(tvd.tv_usec < 0) { tvd.tv_sec--; tvd.tv_usec += 1000000L; } + printf(" Parsing chunks %ld.%06ld sec\n", tvd.tv_sec, tvd.tv_usec); +#endif +#ifndef M3D_NOVOXELS + if(model->numvoxel && model->voxel) { + M3D_LOG("Converting voxels into vertices and mesh"); + /* add normals */ + enorm = model->numvertex; model->numvertex += 6; + model->vertex = (m3dv_t*)M3D_REALLOC(model->vertex, model->numvertex * sizeof(m3dv_t)); + if(!model->vertex) goto memerr; + memset(&model->vertex[enorm], 0, 6 * sizeof(m3dv_t)); + for(l = 0; l < 6; l++) + model->vertex[enorm+l].skinid = M3D_UNDEF; + model->vertex[enorm+0].y = (M3D_FLOAT)-1.0; + model->vertex[enorm+1].z = (M3D_FLOAT)-1.0; + model->vertex[enorm+2].x = (M3D_FLOAT)-1.0; + model->vertex[enorm+3].y = (M3D_FLOAT)1.0; + model->vertex[enorm+4].z = (M3D_FLOAT)1.0; + model->vertex[enorm+5].x = (M3D_FLOAT)1.0; + /* this is a fast, not so memory efficient version, only basic face culling used */ + min_x = min_y = min_z = 2147483647L; + max_x = max_y = max_z = -2147483647L; + for(i = 0; i < model->numvoxel; i++) { + if(model->voxel[i].x + (int32_t)model->voxel[i].w > max_x) max_x = model->voxel[i].x + (int32_t)model->voxel[i].w; + if(model->voxel[i].x < min_x) min_x = model->voxel[i].x; + if(model->voxel[i].y + (int32_t)model->voxel[i].h > max_y) max_y = model->voxel[i].y + (int32_t)model->voxel[i].h; + if(model->voxel[i].y < min_y) min_y = model->voxel[i].y; + if(model->voxel[i].z + (int32_t)model->voxel[i].d > max_z) max_z = model->voxel[i].z + (int32_t)model->voxel[i].d; + if(model->voxel[i].z < min_z) min_z = model->voxel[i].z; + } + i = (-min_x > max_x ? -min_x : max_x); + j = (-min_y > max_y ? -min_y : max_y); + k = (-min_z > max_z ? -min_z : max_z); + if(j > i) i = j; + if(k > i) i = k; + if(i <= 1) i = 1; + w = (M3D_FLOAT)1.0 / (M3D_FLOAT)i; + if(i >= 254) model->vc_s = 2; + if(i >= 65534) model->vc_s = 4; + for(i = 0; i < model->numvoxel; i++) { + sx = model->voxel[i].w; sz = model->voxel[i].d; sy = model->voxel[i].h; + for(y = 0, j = 0; y < sy; y++) + for(z = 0; z < sz; z++) + for(x = 0; x < sx; x++, j++) + if(model->voxel[i].data[j] < model->numvoxtype) { + k = 0; + /* 16__32 ____ + * /| /| /|2 /| + *64_128 | /_8_/ 32 + * | 1_|_2 |4|_|_| + * |/ |/ |/ 1|/ + * 4___8 |16_| */ + k = n = am = 0; + if(!y || model->voxel[i].data[j - sx*sz] >= model->numvoxtype) { n++; am |= 1; k |= 1|2|4|8; } + if(!z || model->voxel[i].data[j - sx] >= model->numvoxtype) { n++; am |= 2; k |= 1|2|16|32; } + if(!x || model->voxel[i].data[j - 1] >= model->numvoxtype) { n++; am |= 4; k |= 1|4|16|64; } + if(y == sy-1 || model->voxel[i].data[j + sx*sz] >= model->numvoxtype) { n++; am |= 8; k |= 16|32|64|128; } + if(z == sz-1 || model->voxel[i].data[j + sx] >= model->numvoxtype) { n++; am |= 16; k |= 4|8|64|128; } + if(x == sx-1 || model->voxel[i].data[j + 1] >= model->numvoxtype) { n++; am |= 32; k |= 2|8|32|128; } + if(k) { + memset(edge, 255, sizeof(edge)); + for(l = 0, len = 1, reclen = model->numvertex; l < 8; l++, len <<= 1) + if(k & len) edge[l] = model->numvertex++; + model->vertex = (m3dv_t*)M3D_REALLOC(model->vertex, model->numvertex * sizeof(m3dv_t)); + if(!model->vertex) goto memerr; + memset(&model->vertex[reclen], 0, (model->numvertex-reclen) * sizeof(m3dv_t)); + for(l = reclen; l < model->numvertex; l++) { + model->vertex[l].skinid = model->voxtype[model->voxel[i].data[j]].skinid; + model->vertex[l].color = model->voxtype[model->voxel[i].data[j]].color; + } + l = reclen; + if(k & 1) { + model->vertex[l].x = (model->voxel[i].x + x) * w; + model->vertex[l].y = (model->voxel[i].y + y) * w; + model->vertex[l].z = (model->voxel[i].z + z) * w; + l++; + } + if(k & 2) { + model->vertex[l].x = (model->voxel[i].x + x + 1) * w; + model->vertex[l].y = (model->voxel[i].y + y) * w; + model->vertex[l].z = (model->voxel[i].z + z) * w; + l++; + } + if(k & 4) { + model->vertex[l].x = (model->voxel[i].x + x) * w; + model->vertex[l].y = (model->voxel[i].y + y) * w; + model->vertex[l].z = (model->voxel[i].z + z + 1) * w; + l++; + } + if(k & 8) { + model->vertex[l].x = (model->voxel[i].x + x + 1) * w; + model->vertex[l].y = (model->voxel[i].y + y) * w; + model->vertex[l].z = (model->voxel[i].z + z + 1) * w; + l++; + } + if(k & 16) { + model->vertex[l].x = (model->voxel[i].x + x) * w; + model->vertex[l].y = (model->voxel[i].y + y + 1) * w; + model->vertex[l].z = (model->voxel[i].z + z) * w; + l++; + } + if(k & 32) { + model->vertex[l].x = (model->voxel[i].x + x + 1) * w; + model->vertex[l].y = (model->voxel[i].y + y + 1) * w; + model->vertex[l].z = (model->voxel[i].z + z) * w; + l++; + } + if(k & 64) { + model->vertex[l].x = (model->voxel[i].x + x) * w; + model->vertex[l].y = (model->voxel[i].y + y + 1) * w; + model->vertex[l].z = (model->voxel[i].z + z + 1) * w; + l++; + } + if(k & 128) { + model->vertex[l].x = (model->voxel[i].x + x + 1) * w; + model->vertex[l].y = (model->voxel[i].y + y + 1) * w; + model->vertex[l].z = (model->voxel[i].z + z + 1) * w; + l++; + } + n <<= 1; + l = model->numface; model->numface += n; + model->face = (m3df_t*)M3D_REALLOC(model->face, model->numface * sizeof(m3df_t)); + if(!model->face) goto memerr; + memset(&model->face[l], 255, n * sizeof(m3df_t)); + for(reclen = l; reclen < model->numface; reclen++) + model->face[reclen].materialid = model->voxtype[model->voxel[i].data[j]].materialid; + if(am & 1) { /* bottom */ + model->face[l].vertex[0] = edge[0]; model->face[l].vertex[1] = edge[1]; model->face[l].vertex[2] = edge[2]; + model->face[l+1].vertex[0] = edge[2]; model->face[l+1].vertex[1] = edge[1]; model->face[l+1].vertex[2] = edge[3]; + model->face[l].normal[0] = model->face[l].normal[1] = model->face[l].normal[2] = + model->face[l+1].normal[0] = model->face[l+1].normal[1] = model->face[l+1].normal[2] = enorm; + l += 2; + } + if(am & 2) { /* north */ + model->face[l].vertex[0] = edge[0]; model->face[l].vertex[1] = edge[4]; model->face[l].vertex[2] = edge[1]; + model->face[l+1].vertex[0] = edge[1]; model->face[l+1].vertex[1] = edge[4]; model->face[l+1].vertex[2] = edge[5]; + model->face[l].normal[0] = model->face[l].normal[1] = model->face[l].normal[2] = + model->face[l+1].normal[0] = model->face[l+1].normal[1] = model->face[l+1].normal[2] = enorm+1; + l += 2; + } + if(am & 4) { /* west */ + model->face[l].vertex[0] = edge[0]; model->face[l].vertex[1] = edge[2]; model->face[l].vertex[2] = edge[4]; + model->face[l+1].vertex[0] = edge[2]; model->face[l+1].vertex[1] = edge[6]; model->face[l+1].vertex[2] = edge[4]; + model->face[l].normal[0] = model->face[l].normal[1] = model->face[l].normal[2] = + model->face[l+1].normal[0] = model->face[l+1].normal[1] = model->face[l+1].normal[2] = enorm+2; + l += 2; + } + if(am & 8) { /* top */ + model->face[l].vertex[0] = edge[4]; model->face[l].vertex[1] = edge[6]; model->face[l].vertex[2] = edge[5]; + model->face[l+1].vertex[0] = edge[5]; model->face[l+1].vertex[1] = edge[6]; model->face[l+1].vertex[2] = edge[7]; + model->face[l].normal[0] = model->face[l].normal[1] = model->face[l].normal[2] = + model->face[l+1].normal[0] = model->face[l+1].normal[1] = model->face[l+1].normal[2] = enorm+3; + l += 2; + } + if(am & 16) { /* south */ + model->face[l].vertex[0] = edge[2]; model->face[l].vertex[1] = edge[7]; model->face[l].vertex[2] = edge[6]; + model->face[l+1].vertex[0] = edge[7]; model->face[l+1].vertex[1] = edge[2]; model->face[l+1].vertex[2] = edge[3]; + model->face[l].normal[0] = model->face[l].normal[1] = model->face[l].normal[2] = + model->face[l+1].normal[0] = model->face[l+1].normal[1] = model->face[l+1].normal[2] = enorm+4; + l += 2; + } + if(am & 32) { /* east */ + model->face[l].vertex[0] = edge[1]; model->face[l].vertex[1] = edge[5]; model->face[l].vertex[2] = edge[7]; + model->face[l+1].vertex[0] = edge[1]; model->face[l+1].vertex[1] = edge[7]; model->face[l+1].vertex[2] = edge[3]; + model->face[l].normal[0] = model->face[l].normal[1] = model->face[l].normal[2] = + model->face[l+1].normal[0] = model->face[l+1].normal[1] = model->face[l+1].normal[2] = enorm+5; + l += 2; + } + } + } + } + } +#endif +#ifndef M3D_NONORMALS + if(model->numface && model->face && neednorm) { + /* if they are missing, calculate triangle normals into a temporary buffer */ + norm = (m3dv_t*)M3D_MALLOC(model->numface * sizeof(m3dv_t)); + if(!norm) goto memerr; + for(i = 0, n = model->numvertex; i < model->numface; i++) + if(model->face[i].normal[0] == M3D_UNDEF) { + v0 = &model->vertex[model->face[i].vertex[0]]; + v1 = &model->vertex[model->face[i].vertex[1]]; + v2 = &model->vertex[model->face[i].vertex[2]]; + va.x = v1->x - v0->x; va.y = v1->y - v0->y; va.z = v1->z - v0->z; + vb.x = v2->x - v0->x; vb.y = v2->y - v0->y; vb.z = v2->z - v0->z; + v0 = &norm[i]; + v0->x = (va.y * vb.z) - (va.z * vb.y); + v0->y = (va.z * vb.x) - (va.x * vb.z); + v0->z = (va.x * vb.y) - (va.y * vb.x); + w = _m3d_rsq((v0->x * v0->x) + (v0->y * v0->y) + (v0->z * v0->z)); + v0->x *= w; v0->y *= w; v0->z *= w; + model->face[i].normal[0] = model->face[i].vertex[0] + n; + model->face[i].normal[1] = model->face[i].vertex[1] + n; + model->face[i].normal[2] = model->face[i].vertex[2] + n; + } + /* this is the fast way, we don't care if a normal is repeated in model->vertex */ + M3D_LOG("Generating normals"); + model->flags |= M3D_FLG_GENNORM; + model->numvertex <<= 1; + model->vertex = (m3dv_t*)M3D_REALLOC(model->vertex, model->numvertex * sizeof(m3dv_t)); + if(!model->vertex) goto memerr; + memset(&model->vertex[n], 0, n * sizeof(m3dv_t)); + for(i = 0; i < model->numface; i++) + for(j = 0; j < 3; j++) { + v0 = &model->vertex[model->face[i].vertex[j] + n]; + v0->x += norm[i].x; + v0->y += norm[i].y; + v0->z += norm[i].z; + } + /* for each vertex, take the average of the temporary normals and use that */ + for(i = 0, v0 = &model->vertex[n]; i < n; i++, v0++) { + w = _m3d_rsq((v0->x * v0->x) + (v0->y * v0->y) + (v0->z * v0->z)); + v0->x *= w; v0->y *= w; v0->z *= w; + v0->skinid = M3D_UNDEF; + } + M3D_FREE(norm); + } +#endif + if(model->numbone && model->bone && model->numskin && model->skin && model->numvertex && model->vertex) { +#ifndef M3D_NOWEIGHTS + M3D_LOG("Generating weight cross-reference"); + for(i = 0; i < model->numvertex; i++) { + if(model->vertex[i].skinid < model->numskin) { + sk = &model->skin[model->vertex[i].skinid]; + w = (M3D_FLOAT)0.0; + for(j = 0; j < M3D_NUMBONE && sk->boneid[j] != M3D_UNDEF && sk->weight[j] > (M3D_FLOAT)0.0; j++) + w += sk->weight[j]; + for(j = 0; j < M3D_NUMBONE && sk->boneid[j] != M3D_UNDEF && sk->weight[j] > (M3D_FLOAT)0.0; j++) { + sk->weight[j] /= w; + b = &model->bone[sk->boneid[j]]; + k = b->numweight++; + b->weight = (m3dw_t*)M3D_REALLOC(b->weight, b->numweight * sizeof(m3da_t)); + if(!b->weight) goto memerr; + b->weight[k].vertexid = i; + b->weight[k].weight = sk->weight[j]; + } + } + } +#endif +#ifndef M3D_NOANIMATION + M3D_LOG("Calculating bone transformation matrices"); + for(i = 0; i < model->numbone; i++) { + b = &model->bone[i]; + if(model->bone[i].parent == M3D_UNDEF) { + _m3d_mat((M3D_FLOAT*)&b->mat4, &model->vertex[b->pos], &model->vertex[b->ori]); + } else { + _m3d_mat((M3D_FLOAT*)&r, &model->vertex[b->pos], &model->vertex[b->ori]); + _m3d_mul((M3D_FLOAT*)&b->mat4, (M3D_FLOAT*)&model->bone[b->parent].mat4, (M3D_FLOAT*)&r); + } + } + for(i = 0; i < model->numbone; i++) + _m3d_inv((M3D_FLOAT*)&model->bone[i].mat4); +#endif + } +#ifdef M3D_PROFILING + gettimeofday(&tv0, NULL); + tvd.tv_sec = tv0.tv_sec - tv1.tv_sec; + tvd.tv_usec = tv0.tv_usec - tv1.tv_usec; + if(tvd.tv_usec < 0) { tvd.tv_sec--; tvd.tv_usec += 1000000L; } + printf(" Post-process %ld.%06ld sec\n", tvd.tv_sec, tvd.tv_usec); +#endif + } + return model; +} + +/** + * Calculates skeletons for animation frames, returns a working copy (should be freed after use) + */ +m3dtr_t *m3d_frame(m3d_t *model, M3D_INDEX actionid, M3D_INDEX frameid, m3dtr_t *skeleton) +{ + unsigned int i; + M3D_INDEX s = frameid; + m3dfr_t *fr; + + if(!model || !model->numbone || !model->bone || (actionid != M3D_UNDEF && (!model->action || + actionid >= model->numaction || frameid >= model->action[actionid].numframe))) { + model->errcode = M3D_ERR_UNKFRAME; + return skeleton; + } + model->errcode = M3D_SUCCESS; + if(!skeleton) { + skeleton = (m3dtr_t*)M3D_MALLOC(model->numbone * sizeof(m3dtr_t)); + if(!skeleton) { + model->errcode = M3D_ERR_ALLOC; + return NULL; + } + goto gen; + } + if(actionid == M3D_UNDEF || !frameid) { +gen: s = 0; + for(i = 0; i < model->numbone; i++) { + skeleton[i].boneid = i; + skeleton[i].pos = model->bone[i].pos; + skeleton[i].ori = model->bone[i].ori; + } + } + if(actionid < model->numaction && (frameid || !model->action[actionid].frame[0].msec)) { + for(; s <= frameid; s++) { + fr = &model->action[actionid].frame[s]; + for(i = 0; i < fr->numtransform; i++) { + skeleton[fr->transform[i].boneid].pos = fr->transform[i].pos; + skeleton[fr->transform[i].boneid].ori = fr->transform[i].ori; + } + } + } + return skeleton; +} + +#ifndef M3D_NOANIMATION +/** + * Returns interpolated animation-pose, a working copy (should be freed after use) + */ +m3db_t *m3d_pose(m3d_t *model, M3D_INDEX actionid, uint32_t msec) +{ + unsigned int i, j, l; + M3D_FLOAT r[16], t, c, d, s; + m3db_t *ret; + m3dv_t *v, *p, *f; + m3dtr_t *tmp; + m3dfr_t *fr; + + if(!model || !model->numbone || !model->bone) { + model->errcode = M3D_ERR_UNKFRAME; + return NULL; + } + ret = (m3db_t*)M3D_MALLOC(model->numbone * sizeof(m3db_t)); + if(!ret) { + model->errcode = M3D_ERR_ALLOC; + return NULL; + } + memcpy(ret, model->bone, model->numbone * sizeof(m3db_t)); + for(i = 0; i < model->numbone; i++) + _m3d_inv((M3D_FLOAT*)&ret[i].mat4); + if(!model->action || actionid >= model->numaction) { + model->errcode = M3D_ERR_UNKFRAME; + return ret; + } + msec %= model->action[actionid].durationmsec; + model->errcode = M3D_SUCCESS; + fr = &model->action[actionid].frame[0]; + for(j = l = 0; j < model->action[actionid].numframe && model->action[actionid].frame[j].msec <= msec; j++) { + fr = &model->action[actionid].frame[j]; + l = fr->msec; + for(i = 0; i < fr->numtransform; i++) { + ret[fr->transform[i].boneid].pos = fr->transform[i].pos; + ret[fr->transform[i].boneid].ori = fr->transform[i].ori; + } + } + if(l != msec) { + model->vertex = (m3dv_t*)M3D_REALLOC(model->vertex, (model->numvertex + 2 * model->numbone) * sizeof(m3dv_t)); + if(!model->vertex) { + M3D_FREE(ret); + model->errcode = M3D_ERR_ALLOC; + return NULL; + } + tmp = (m3dtr_t*)M3D_MALLOC(model->numbone * sizeof(m3dtr_t)); + if(tmp) { + for(i = 0; i < model->numbone; i++) { + tmp[i].pos = ret[i].pos; + tmp[i].ori = ret[i].ori; + } + fr = &model->action[actionid].frame[j % model->action[actionid].numframe]; + t = l >= fr->msec ? (M3D_FLOAT)1.0 : (M3D_FLOAT)(msec - l) / (M3D_FLOAT)(fr->msec - l); + for(i = 0; i < fr->numtransform; i++) { + tmp[fr->transform[i].boneid].pos = fr->transform[i].pos; + tmp[fr->transform[i].boneid].ori = fr->transform[i].ori; + } + for(i = 0, j = model->numvertex; i < model->numbone; i++) { + /* interpolation of position */ + if(ret[i].pos != tmp[i].pos) { + p = &model->vertex[ret[i].pos]; + f = &model->vertex[tmp[i].pos]; + v = &model->vertex[j]; + v->x = p->x + t * (f->x - p->x); + v->y = p->y + t * (f->y - p->y); + v->z = p->z + t * (f->z - p->z); + ret[i].pos = j++; + } + /* interpolation of orientation */ + if(ret[i].ori != tmp[i].ori) { + p = &model->vertex[ret[i].ori]; + f = &model->vertex[tmp[i].ori]; + v = &model->vertex[j]; + d = p->w * f->w + p->x * f->x + p->y * f->y + p->z * f->z; + if(d < 0) { d = -d; s = (M3D_FLOAT)-1.0; } else s = (M3D_FLOAT)1.0; +#if 0 + /* don't use SLERP, requires two more variables, libm linkage and it is slow (but nice) */ + a = (M3D_FLOAT)1.0 - t; b = t; + if(d < (M3D_FLOAT)0.999999) { c = acosf(d); b = 1 / sinf(c); a = sinf(a * c) * b; b *= sinf(t * c) * s; } + v->x = p->x * a + f->x * b; + v->y = p->y * a + f->y * b; + v->z = p->z * a + f->z * b; + v->w = p->w * a + f->w * b; +#else + /* approximated NLERP, original approximation by Arseny Kapoulkine, heavily optimized by me */ + c = t - (M3D_FLOAT)0.5; t += t * c * (t - (M3D_FLOAT)1.0) * (((M3D_FLOAT)1.0904 + d * ((M3D_FLOAT)-3.2452 + + d * ((M3D_FLOAT)3.55645 - d * (M3D_FLOAT)1.43519))) * c * c + ((M3D_FLOAT)0.848013 + d * + ((M3D_FLOAT)-1.06021 + d * (M3D_FLOAT)0.215638))); + v->x = p->x + t * (s * f->x - p->x); + v->y = p->y + t * (s * f->y - p->y); + v->z = p->z + t * (s * f->z - p->z); + v->w = p->w + t * (s * f->w - p->w); + d = _m3d_rsq(v->w * v->w + v->x * v->x + v->y * v->y + v->z * v->z); + v->x *= d; v->y *= d; v->z *= d; v->w *= d; +#endif + ret[i].ori = j++; + } + } + M3D_FREE(tmp); + } + } + for(i = 0; i < model->numbone; i++) { + if(ret[i].parent == M3D_UNDEF) { + _m3d_mat((M3D_FLOAT*)&ret[i].mat4, &model->vertex[ret[i].pos], &model->vertex[ret[i].ori]); + } else { + _m3d_mat((M3D_FLOAT*)&r, &model->vertex[ret[i].pos], &model->vertex[ret[i].ori]); + _m3d_mul((M3D_FLOAT*)&ret[i].mat4, (M3D_FLOAT*)&ret[ret[i].parent].mat4, (M3D_FLOAT*)&r); + } + } + return ret; +} + +#endif /* M3D_NOANIMATION */ + +#endif /* M3D_IMPLEMENTATION */ + +#if !defined(M3D_NODUP) && (!defined(M3D_NOIMPORTER) || defined(M3D_EXPORTER)) +/** + * Free the in-memory model + */ +void m3d_free(m3d_t *model) +{ + unsigned int i, j; + + if(!model) return; +#ifdef M3D_ASCII + /* if model imported from ASCII, we have to free all strings as well */ + if(model->flags & M3D_FLG_FREESTR) { + if(model->name) M3D_FREE(model->name); + if(model->license) M3D_FREE(model->license); + if(model->author) M3D_FREE(model->author); + if(model->desc) M3D_FREE(model->desc); + if(model->bone) + for(i = 0; i < model->numbone; i++) + if(model->bone[i].name) + M3D_FREE(model->bone[i].name); + if(model->shape) + for(i = 0; i < model->numshape; i++) + if(model->shape[i].name) + M3D_FREE(model->shape[i].name); + if(model->numvoxtype) + for(i = 0; i < model->numvoxtype; i++) { + if(model->voxtype[i].name) + M3D_FREE(model->voxtype[i].name); + for(j = 0; j < model->voxtype[i].numitem; j++) + if(model->voxtype[i].item[j].name) + M3D_FREE(model->voxtype[i].item[j].name); + } + if(model->numvoxel) + for(i = 0; i < model->numvoxel; i++) + if(model->voxel[i].name) + M3D_FREE(model->voxel[i].name); + if(model->material) + for(i = 0; i < model->nummaterial; i++) + if(model->material[i].name) + M3D_FREE(model->material[i].name); + if(model->action) + for(i = 0; i < model->numaction; i++) + if(model->action[i].name) + M3D_FREE(model->action[i].name); + if(model->texture) + for(i = 0; i < model->numtexture; i++) + if(model->texture[i].name) + M3D_FREE(model->texture[i].name); + if(model->inlined) + for(i = 0; i < model->numinlined; i++) { + if(model->inlined[i].name) + M3D_FREE(model->inlined[i].name); + if(model->inlined[i].data) + M3D_FREE(model->inlined[i].data); + } + if(model->extra) + for(i = 0; i < model->numextra; i++) + if(model->extra[i]) + M3D_FREE(model->extra[i]); + if(model->label) + for(i = 0; i < model->numlabel; i++) { + if(model->label[i].name) { + for(j = i + 1; j < model->numlabel; j++) + if(model->label[j].name == model->label[i].name) + model->label[j].name = NULL; + M3D_FREE(model->label[i].name); + } + if(model->label[i].lang) { + for(j = i + 1; j < model->numlabel; j++) + if(model->label[j].lang == model->label[i].lang) + model->label[j].lang = NULL; + M3D_FREE(model->label[i].lang); + } + if(model->label[i].text) + M3D_FREE(model->label[i].text); + } + if(model->preview.data) + M3D_FREE(model->preview.data); + } +#endif + if(model->flags & M3D_FLG_FREERAW) M3D_FREE(model->raw); + + if(model->tmap) M3D_FREE(model->tmap); + if(model->bone) { + for(i = 0; i < model->numbone; i++) + if(model->bone[i].weight) + M3D_FREE(model->bone[i].weight); + M3D_FREE(model->bone); + } + if(model->skin) M3D_FREE(model->skin); + if(model->vertex) M3D_FREE(model->vertex); + if(model->face) M3D_FREE(model->face); + if(model->voxtype) { + for(i = 0; i < model->numvoxtype; i++) + if(model->voxtype[i].item) + M3D_FREE(model->voxtype[i].item); + M3D_FREE(model->voxtype); + } + if(model->voxel) { + for(i = 0; i < model->numvoxel; i++) + if(model->voxel[i].data) + M3D_FREE(model->voxel[i].data); + M3D_FREE(model->voxel); + } + if(model->shape) { + for(i = 0; i < model->numshape; i++) { + if(model->shape[i].cmd) { + for(j = 0; j < model->shape[i].numcmd; j++) + if(model->shape[i].cmd[j].arg) M3D_FREE(model->shape[i].cmd[j].arg); + M3D_FREE(model->shape[i].cmd); + } + } + M3D_FREE(model->shape); + } + if(model->material && !(model->flags & M3D_FLG_MTLLIB)) { + for(i = 0; i < model->nummaterial; i++) + if(model->material[i].prop) M3D_FREE(model->material[i].prop); + M3D_FREE(model->material); + } + if(model->texture) { + for(i = 0; i < model->numtexture; i++) + if(model->texture[i].d) M3D_FREE(model->texture[i].d); + M3D_FREE(model->texture); + } + if(model->action) { + for(i = 0; i < model->numaction; i++) { + if(model->action[i].frame) { + for(j = 0; j < model->action[i].numframe; j++) + if(model->action[i].frame[j].transform) M3D_FREE(model->action[i].frame[j].transform); + M3D_FREE(model->action[i].frame); + } + } + M3D_FREE(model->action); + } + if(model->label) M3D_FREE(model->label); + if(model->inlined) M3D_FREE(model->inlined); + if(model->extra) M3D_FREE(model->extra); + M3D_FREE(model); +} +#endif + +#ifdef M3D_EXPORTER +typedef struct { + char *str; + uint32_t offs; +} m3dstr_t; + +typedef struct { + m3dti_t data; + M3D_INDEX oldidx; + M3D_INDEX newidx; +} m3dtisave_t; + +typedef struct { + m3dv_t data; + M3D_INDEX oldidx; + M3D_INDEX newidx; + unsigned char norm; +} m3dvsave_t; + +typedef struct { + m3ds_t data; + M3D_INDEX oldidx; + M3D_INDEX newidx; +} m3dssave_t; + +typedef struct { + m3df_t data; + int group; + uint8_t opacity; +} m3dfsave_t; + +/* create unique list of strings */ +static m3dstr_t *_m3d_addstr(m3dstr_t *str, uint32_t *numstr, char *s) +{ + uint32_t i; + if(!s || !*s) return str; + if(str) { + for(i = 0; i < *numstr; i++) + if(str[i].str == s || !strcmp(str[i].str, s)) return str; + } + str = (m3dstr_t*)M3D_REALLOC(str, ((*numstr) + 1) * sizeof(m3dstr_t)); + str[*numstr].str = s; + str[*numstr].offs = 0; + (*numstr)++; + return str; +} + +/* add strings to header */ +m3dhdr_t *_m3d_addhdr(m3dhdr_t *h, m3dstr_t *s) +{ + int i; + char *safe = _m3d_safestr(s->str, 0); + i = (int)strlen(safe); + h = (m3dhdr_t*)M3D_REALLOC(h, h->length + i+1); + if(!h) { M3D_FREE(safe); return NULL; } + memcpy((uint8_t*)h + h->length, safe, i+1); + s->offs = h->length - 16; + h->length += i+1; + M3D_FREE(safe); + return h; +} + +/* return offset of string */ +static uint32_t _m3d_stridx(m3dstr_t *str, uint32_t numstr, char *s) +{ + uint32_t i; + char *safe; + if(!s || !*s) return 0; + if(str) { + safe = _m3d_safestr(s, 0); + if(!safe) return 0; + if(!*safe) { + M3D_FREE(safe); + return 0; + } + for(i = 0; i < numstr; i++) + if(!strcmp(str[i].str, s)) { + M3D_FREE(safe); + return str[i].offs; + } + M3D_FREE(safe); + } + return 0; +} + +/* compare to faces by their material */ +static int _m3d_facecmp(const void *a, const void *b) { + const m3dfsave_t *A = (const m3dfsave_t*)a, *B = (const m3dfsave_t*)b; + return A->group != B->group ? A->group - B->group : (A->opacity != B->opacity ? (int)B->opacity - (int)A->opacity : + (int)A->data.materialid - (int)B->data.materialid); +} +/* compare face groups */ +static int _m3d_grpcmp(const void *a, const void *b) { return *((uint32_t*)a) - *((uint32_t*)b); } +/* compare UVs */ +static int _m3d_ticmp(const void *a, const void *b) { return memcmp(a, b, sizeof(m3dti_t)); } +/* compare skin groups */ +static int _m3d_skincmp(const void *a, const void *b) { return memcmp(a, b, sizeof(m3ds_t)); } +/* compare vertices */ +static int _m3d_vrtxcmp(const void *a, const void *b) { + int c = memcmp(a, b, 3 * sizeof(M3D_FLOAT)); + if(c) return c; + c = ((m3dvsave_t*)a)->norm - ((m3dvsave_t*)b)->norm; + if(c) return c; + return memcmp(a, b, sizeof(m3dv_t)); +} +/* compare labels */ +static _inline int _m3d_strcmp(char *a, char *b) +{ + if(a == NULL && b != NULL) return -1; + if(a != NULL && b == NULL) return 1; + if(a == NULL && b == NULL) return 0; + return strcmp(a, b); +} +static int _m3d_lblcmp(const void *a, const void *b) { + const m3dl_t *A = (const m3dl_t*)a, *B = (const m3dl_t*)b; + int c = _m3d_strcmp(A->lang, B->lang); + if(!c) c = _m3d_strcmp(A->name, B->name); + if(!c) c = _m3d_strcmp(A->text, B->text); + return c; +} +/* compare two colors by HSV value */ +_inline static int _m3d_cmapcmp(const void *a, const void *b) +{ + uint8_t *A = (uint8_t*)a, *B = (uint8_t*)b; + _register int m, vA, vB; + /* get HSV value for A */ + m = A[2] < A[1]? A[2] : A[1]; if(A[0] < m) m = A[0]; + vA = A[2] > A[1]? A[2] : A[1]; if(A[0] > vA) vA = A[0]; + /* get HSV value for B */ + m = B[2] < B[1]? B[2] : B[1]; if(B[0] < m) m = B[0]; + vB = B[2] > B[1]? B[2] : B[1]; if(B[0] > vB) vB = B[0]; + return vA - vB; +} + +/* create sorted list of colors */ +static uint32_t *_m3d_addcmap(uint32_t *cmap, uint32_t *numcmap, uint32_t color) +{ + uint32_t i; + if(cmap) { + for(i = 0; i < *numcmap; i++) + if(cmap[i] == color) return cmap; + } + cmap = (uint32_t*)M3D_REALLOC(cmap, ((*numcmap) + 1) * sizeof(uint32_t)); + for(i = 0; i < *numcmap && _m3d_cmapcmp(&color, &cmap[i]) > 0; i++); + if(i < *numcmap) memmove(&cmap[i+1], &cmap[i], ((*numcmap) - i)*sizeof(uint32_t)); + cmap[i] = color; + (*numcmap)++; + return cmap; +} + +/* look up a color and return its index */ +static uint32_t _m3d_cmapidx(uint32_t *cmap, uint32_t numcmap, uint32_t color) +{ + uint32_t i; + if(numcmap >= 65536) + return color; + for(i = 0; i < numcmap; i++) + if(cmap[i] == color) return i; + return 0; +} + +/* add index to output */ +static unsigned char *_m3d_addidx(unsigned char *out, char type, uint32_t idx) { + switch(type) { + case 1: *out++ = (uint8_t)(idx); break; + case 2: *((uint16_t*)out) = (uint16_t)(idx); out += 2; break; + case 4: *((uint32_t*)out) = (uint32_t)(idx); out += 4; break; + /* case 0: case 8: break; */ + } + return out; +} + +/* round a vertex position */ +static void _m3d_round(int quality, m3dv_t *src, m3dv_t *dst) +{ + _register int t; + /* copy additional attributes */ + if(src != dst) memcpy(dst, src, sizeof(m3dv_t)); + /* round according to quality */ + switch(quality) { + case M3D_EXP_INT8: + t = (int)(src->x * 127 + (src->x >= 0 ? (M3D_FLOAT)0.5 : (M3D_FLOAT)-0.5)); dst->x = (M3D_FLOAT)t / (M3D_FLOAT)127.0; + t = (int)(src->y * 127 + (src->y >= 0 ? (M3D_FLOAT)0.5 : (M3D_FLOAT)-0.5)); dst->y = (M3D_FLOAT)t / (M3D_FLOAT)127.0; + t = (int)(src->z * 127 + (src->z >= 0 ? (M3D_FLOAT)0.5 : (M3D_FLOAT)-0.5)); dst->z = (M3D_FLOAT)t / (M3D_FLOAT)127.0; + t = (int)(src->w * 127 + (src->w >= 0 ? (M3D_FLOAT)0.5 : (M3D_FLOAT)-0.5)); dst->w = (M3D_FLOAT)t / (M3D_FLOAT)127.0; + break; + case M3D_EXP_INT16: + t = (int)(src->x * 32767 + (src->x >= 0 ? (M3D_FLOAT)0.5 : (M3D_FLOAT)-0.5)); dst->x = (M3D_FLOAT)t / (M3D_FLOAT)32767.0; + t = (int)(src->y * 32767 + (src->y >= 0 ? (M3D_FLOAT)0.5 : (M3D_FLOAT)-0.5)); dst->y = (M3D_FLOAT)t / (M3D_FLOAT)32767.0; + t = (int)(src->z * 32767 + (src->z >= 0 ? (M3D_FLOAT)0.5 : (M3D_FLOAT)-0.5)); dst->z = (M3D_FLOAT)t / (M3D_FLOAT)32767.0; + t = (int)(src->w * 32767 + (src->w >= 0 ? (M3D_FLOAT)0.5 : (M3D_FLOAT)-0.5)); dst->w = (M3D_FLOAT)t / (M3D_FLOAT)32767.0; + break; + } + if(dst->x == (M3D_FLOAT)-0.0) dst->x = (M3D_FLOAT)0.0; + if(dst->y == (M3D_FLOAT)-0.0) dst->y = (M3D_FLOAT)0.0; + if(dst->z == (M3D_FLOAT)-0.0) dst->z = (M3D_FLOAT)0.0; + if(dst->w == (M3D_FLOAT)-0.0) dst->w = (M3D_FLOAT)0.0; +} + +#ifdef M3D_ASCII +/* add a bone to ascii output */ +static char *_m3d_prtbone(char *ptr, m3db_t *bone, M3D_INDEX numbone, M3D_INDEX parent, uint32_t level, M3D_INDEX *vrtxidx) +{ + uint32_t i, j; + char *sn; + + if(level > M3D_BONEMAXLEVEL || !bone) return ptr; + for(i = 0; i < numbone; i++) { + if(bone[i].parent == parent) { + for(j = 0; j < level; j++) *ptr++ = '/'; + sn = _m3d_safestr(bone[i].name, 0); + ptr += sprintf(ptr, "%d %d %s\r\n", vrtxidx[bone[i].pos], vrtxidx[bone[i].ori], sn); + M3D_FREE(sn); + ptr = _m3d_prtbone(ptr, bone, numbone, i, level + 1, vrtxidx); + } + } + return ptr; +} +#endif + +/** + * Function to encode an in-memory model into on storage Model 3D format + */ +unsigned char *m3d_save(m3d_t *model, int quality, int flags, unsigned int *size) +{ +#ifdef M3D_ASCII + const char *ol; + char *ptr; +#endif + char vc_s, vi_s, si_s, ci_s, ti_s, bi_s, nb_s, sk_s, fc_s, hi_s, fi_s, vd_s, vp_s; + char *sn = NULL, *sl = NULL, *sa = NULL, *sd = NULL; + unsigned char *out = NULL, *z = NULL, weights[M3D_NUMBONE < 8 ? 8 : M3D_NUMBONE], *norm = NULL; + unsigned int i, j, k, l, n, o, len, chunklen, *length; + int maxvox = 0, minvox = 0; + M3D_FLOAT scale = (M3D_FLOAT)0.0, min_x, max_x, min_y, max_y, min_z, max_z, mw; + M3D_INDEX last, *vrtxidx = NULL, *mtrlidx = NULL, *tmapidx = NULL, *skinidx = NULL; +#ifdef M3D_VERTEXMAX + M3D_INDEX lastp; +#endif + uint32_t idx, numcmap = 0, *cmap = NULL, numvrtx = 0, maxvrtx = 0, numtmap = 0, maxtmap = 0, numproc = 0; + uint32_t numskin = 0, maxskin = 0, numstr = 0, maxt = 0, maxbone = 0, numgrp = 0, maxgrp = 0, *grpidx = NULL; + uint8_t *opa = NULL; + m3dcd_t *cd; + m3dc_t *cmd; + m3dstr_t *str = NULL; + m3dvsave_t *vrtx = NULL, vertex; + m3dtisave_t *tmap = NULL, tcoord; + m3dssave_t *skin = NULL, sk; + m3dfsave_t *face = NULL; + m3dhdr_t *h = NULL; + m3dm_t *m; + m3da_t *a; + + if(!model) { + if(size) *size = 0; + return NULL; + } + model->errcode = M3D_SUCCESS; +#ifdef M3D_ASCII + if(flags & M3D_EXP_ASCII) quality = M3D_EXP_DOUBLE; +#endif + vrtxidx = (M3D_INDEX*)M3D_MALLOC(model->numvertex * sizeof(M3D_INDEX)); + if(!vrtxidx) goto memerr; + memset(vrtxidx, 255, model->numvertex * sizeof(M3D_INDEX)); + if(model->numvertex && !(flags & M3D_EXP_NONORMAL)){ + norm = (unsigned char*)M3D_MALLOC(model->numvertex * sizeof(unsigned char)); + if(!norm) goto memerr; + memset(norm, 0, model->numvertex * sizeof(unsigned char)); + } + if(model->nummaterial && !(flags & M3D_EXP_NOMATERIAL)) { + mtrlidx = (M3D_INDEX*)M3D_MALLOC(model->nummaterial * sizeof(M3D_INDEX)); + if(!mtrlidx) goto memerr; + memset(mtrlidx, 255, model->nummaterial * sizeof(M3D_INDEX)); + opa = (uint8_t*)M3D_MALLOC(model->nummaterial * 2 * sizeof(M3D_INDEX)); + if(!opa) goto memerr; + memset(opa, 255, model->nummaterial * 2 * sizeof(M3D_INDEX)); + } + if(model->numtmap && !(flags & M3D_EXP_NOTXTCRD)) { + tmapidx = (M3D_INDEX*)M3D_MALLOC(model->numtmap * sizeof(M3D_INDEX)); + if(!tmapidx) goto memerr; + memset(tmapidx, 255, model->numtmap * sizeof(M3D_INDEX)); + } + /** collect array elements that are actually referenced **/ + if(!(flags & M3D_EXP_NOFACE)) { + /* face */ + if(model->numface && model->face) { + M3D_LOG("Processing mesh face"); + face = (m3dfsave_t*)M3D_MALLOC(model->numface * sizeof(m3dfsave_t)); + if(!face) goto memerr; + for(i = 0; i < model->numface; i++) { + memcpy(&face[i].data, &model->face[i], sizeof(m3df_t)); + face[i].group = 0; + face[i].opacity = 255; + if(!(flags & M3D_EXP_NOMATERIAL) && model->face[i].materialid < model->nummaterial) { + if(model->material[model->face[i].materialid].numprop) { + mtrlidx[model->face[i].materialid] = 0; + if(opa[model->face[i].materialid * 2]) { + m = &model->material[model->face[i].materialid]; + for(j = 0; j < m->numprop; j++) + if(m->prop[j].type == m3dp_Kd) { + opa[model->face[i].materialid * 2 + 1] = ((uint8_t*)&m->prop[j].value.color)[3]; + break; + } + for(j = 0; j < m->numprop; j++) + if(m->prop[j].type == m3dp_d) { + opa[model->face[i].materialid * 2 + 1] = (uint8_t)(m->prop[j].value.fnum * 255); + break; + } + opa[model->face[i].materialid * 2] = 0; + } + face[i].opacity = opa[model->face[i].materialid * 2 + 1]; + } else + face[i].data.materialid = M3D_UNDEF; + } + for(j = 0; j < 3; j++) { + k = model->face[i].vertex[j]; + if(k < model->numvertex) + vrtxidx[k] = 0; + if(!(flags & M3D_EXP_NOCMAP)) { + cmap = _m3d_addcmap(cmap, &numcmap, model->vertex[k].color); + if(!cmap) goto memerr; + } + k = model->face[i].normal[j]; + if(k < model->numvertex && !(flags & M3D_EXP_NONORMAL)) { + vrtxidx[k] = 0; + norm[k] = 1; + } + k = model->face[i].texcoord[j]; + if(k < model->numtmap && !(flags & M3D_EXP_NOTXTCRD)) + tmapidx[k] = 0; +#ifdef M3D_VERTEXMAX + k = model->face[i].vertmax[j]; + if(k < model->numvertex && !(flags & M3D_EXP_NOVRTMAX)) + vrtxidx[k] = 0; +#endif + } + /* convert from CW to CCW */ + if(flags & M3D_EXP_IDOSUCK) { + j = face[i].data.vertex[1]; + face[i].data.vertex[1] = face[i].data.vertex[2]; + face[i].data.vertex[2] = j; + j = face[i].data.normal[1]; + face[i].data.normal[1] = face[i].data.normal[2]; + face[i].data.normal[2] = j; + j = face[i].data.texcoord[1]; + face[i].data.texcoord[1] = face[i].data.texcoord[2]; + face[i].data.texcoord[2] = j; +#ifdef M3D_VERTEXMAX + j = face[i].data.vertmax[1]; + face[i].data.vertmax[1] = face[i].data.vertmax[2]; + face[i].data.vertmax[2] = j; +#endif + } + } + } + if((model->numvoxtype && model->voxtype) || (model->numvoxel && model->voxel)) { + M3D_LOG("Processing voxel face"); + for(i = 0; i < model->numvoxtype; i++) { + str = _m3d_addstr(str, &numstr, model->voxtype[i].name); + if(model->voxtype[i].name && !str) goto memerr; + if(!(flags & M3D_EXP_NOCMAP)) { + cmap = _m3d_addcmap(cmap, &numcmap, model->voxtype[i].color); + if(!cmap) goto memerr; + } + for(j = 0; j < model->voxtype[i].numitem; j++) { + str = _m3d_addstr(str, &numstr, model->voxtype[i].item[j].name); + if(model->voxtype[i].item[j].name && !str) goto memerr; + } + } + for(i = 0; i < model->numvoxel; i++) { + str = _m3d_addstr(str, &numstr, model->voxel[i].name); + if(model->voxel[i].name && !str) goto memerr; + if(model->voxel[i].x < minvox) minvox = model->voxel[i].x; + if(model->voxel[i].x + (int)model->voxel[i].w > maxvox) maxvox = model->voxel[i].x + model->voxel[i].w; + if(model->voxel[i].y < minvox) minvox = model->voxel[i].y; + if(model->voxel[i].y + (int)model->voxel[i].h > maxvox) maxvox = model->voxel[i].y + model->voxel[i].h; + if(model->voxel[i].z < minvox) minvox = model->voxel[i].z; + if(model->voxel[i].z + (int)model->voxel[i].d > maxvox) maxvox = model->voxel[i].z + model->voxel[i].d; + } + } + if(model->numshape && model->shape) { + M3D_LOG("Processing shape face"); + for(i = 0; i < model->numshape; i++) { + if(!model->shape[i].numcmd) continue; + str = _m3d_addstr(str, &numstr, model->shape[i].name); + if(model->shape[i].name && !str) goto memerr; + for(j = 0; j < model->shape[i].numcmd; j++) { + cmd = &model->shape[i].cmd[j]; + if(cmd->type >= (unsigned int)(sizeof(m3d_commandtypes)/sizeof(m3d_commandtypes[0])) || !cmd->arg) + continue; + if(cmd->type == m3dc_mesh) { + if(numgrp + 2 < maxgrp) { + maxgrp += 1024; + grpidx = (uint32_t*)M3D_REALLOC(grpidx, maxgrp * sizeof(uint32_t)); + if(!grpidx) goto memerr; + if(!numgrp) { + grpidx[0] = 0; + grpidx[1] = model->numface; + numgrp += 2; + } + } + grpidx[numgrp + 0] = cmd->arg[0]; + grpidx[numgrp + 1] = cmd->arg[0] + cmd->arg[1]; + numgrp += 2; + } + cd = &m3d_commandtypes[cmd->type]; + for(k = n = 0, l = cd->p; k < l; k++) + switch(cd->a[((k - n) % (cd->p - n)) + n]) { + case m3dcp_mi_t: + if(!(flags & M3D_EXP_NOMATERIAL) && cmd->arg[k] < model->nummaterial) + mtrlidx[cmd->arg[k]] = 0; + break; + case m3dcp_ti_t: + if(!(flags & M3D_EXP_NOTXTCRD) && cmd->arg[k] < model->numtmap) + tmapidx[cmd->arg[k]] = 0; + break; + case m3dcp_qi_t: + case m3dcp_vi_t: + if(cmd->arg[k] < model->numvertex) + vrtxidx[cmd->arg[k]] = 0; + break; + case m3dcp_va_t: + n = k + 1; l += (cmd->arg[k] - 1) * (cd->p - k - 1); + break; + } + } + } + } + if(model->numface && face) { + if(numgrp && grpidx) { + qsort(grpidx, numgrp, sizeof(uint32_t), _m3d_grpcmp); + for(i = j = 0; i < model->numface && j < numgrp; i++) { + while(j < numgrp && grpidx[j] < i) j++; + face[i].group = j; + } + } + qsort(face, model->numface, sizeof(m3dfsave_t), _m3d_facecmp); + } + if(grpidx) { M3D_FREE(grpidx); grpidx = NULL; } + if(model->numlabel && model->label) { + M3D_LOG("Processing annotation labels"); + for(i = 0; i < model->numlabel; i++) { + str = _m3d_addstr(str, &numstr, model->label[i].name); + str = _m3d_addstr(str, &numstr, model->label[i].lang); + str = _m3d_addstr(str, &numstr, model->label[i].text); + if(!(flags & M3D_EXP_NOCMAP)) { + cmap = _m3d_addcmap(cmap, &numcmap, model->label[i].color); + if(!cmap) goto memerr; + } + if(model->label[i].vertexid < model->numvertex) + vrtxidx[model->label[i].vertexid] = 0; + } + qsort(model->label, model->numlabel, sizeof(m3dl_t), _m3d_lblcmp); + } + } else if(!(flags & M3D_EXP_NOMATERIAL)) { + /* without a face, simply add all materials, because it can be an mtllib */ + for(i = 0; i < model->nummaterial; i++) + mtrlidx[i] = i; + } + /* bind-pose skeleton */ + if(model->numbone && model->bone && !(flags & M3D_EXP_NOBONE)) { + M3D_LOG("Processing bones"); + for(i = 0; i < model->numbone; i++) { + str = _m3d_addstr(str, &numstr, model->bone[i].name); + if(!str) goto memerr; + k = model->bone[i].pos; + if(k < model->numvertex) + vrtxidx[k] = 0; + k = model->bone[i].ori; + if(k < model->numvertex) + vrtxidx[k] = 0; + } + } + /* actions, animated skeleton poses */ + if(model->numaction && model->action && !(flags & M3D_EXP_NOACTION)) { + M3D_LOG("Processing action list"); + for(j = 0; j < model->numaction; j++) { + a = &model->action[j]; + str = _m3d_addstr(str, &numstr, a->name); + if(!str) goto memerr; + if(a->numframe > 65535) a->numframe = 65535; + for(i = 0; i < a->numframe; i++) { + for(l = 0; l < a->frame[i].numtransform; l++) { + k = a->frame[i].transform[l].pos; + if(k < model->numvertex) + vrtxidx[k] = 0; + k = a->frame[i].transform[l].ori; + if(k < model->numvertex) + vrtxidx[k] = 0; + } + if(l > maxt) maxt = l; + } + } + } + /* add colors to color map and texture names to string table */ + if(!(flags & M3D_EXP_NOMATERIAL)) { + M3D_LOG("Processing materials"); + for(i = k = 0; i < model->nummaterial; i++) { + if(mtrlidx[i] == M3D_UNDEF || !model->material[i].numprop) continue; + mtrlidx[i] = k++; + m = &model->material[i]; + str = _m3d_addstr(str, &numstr, m->name); + if(!str) goto memerr; + if(m->prop) + for(j = 0; j < m->numprop; j++) { + if(!(flags & M3D_EXP_NOCMAP) && m->prop[j].type < 128) { + for(l = 0; l < sizeof(m3d_propertytypes)/sizeof(m3d_propertytypes[0]); l++) { + if(m->prop[j].type == m3d_propertytypes[l].id && m3d_propertytypes[l].format == m3dpf_color) { + ((uint8_t*)&m->prop[j].value.color)[3] = opa[i * 2 + 1]; + cmap = _m3d_addcmap(cmap, &numcmap, m->prop[j].value.color); + if(!cmap) goto memerr; + break; + } + } + } + if(m->prop[j].type >= 128 && m->prop[j].value.textureid < model->numtexture && + model->texture[m->prop[j].value.textureid].name) { + str = _m3d_addstr(str, &numstr, model->texture[m->prop[j].value.textureid].name); + if(!str) goto memerr; + } + } + } + } + /* if there's only one black color, don't store it */ + if(numcmap == 1 && cmap && !cmap[0]) numcmap = 0; + + /** compress lists **/ + if(model->numtmap && !(flags & M3D_EXP_NOTXTCRD)) { + M3D_LOG("Compressing tmap"); + tmap = (m3dtisave_t*)M3D_MALLOC(model->numtmap * sizeof(m3dtisave_t)); + if(!tmap) goto memerr; + for(i = 0; i < model->numtmap; i++) { + if(tmapidx[i] == M3D_UNDEF) continue; + switch(quality) { + case M3D_EXP_INT8: + l = (unsigned int)(model->tmap[i].u * 255); tcoord.data.u = (M3D_FLOAT)l / (M3D_FLOAT)255.0; + l = (unsigned int)(model->tmap[i].v * 255); tcoord.data.v = (M3D_FLOAT)l / (M3D_FLOAT)255.0; + break; + case M3D_EXP_INT16: + l = (unsigned int)(model->tmap[i].u * 65535); tcoord.data.u = (M3D_FLOAT)l / (M3D_FLOAT)65535.0; + l = (unsigned int)(model->tmap[i].v * 65535); tcoord.data.v = (M3D_FLOAT)l / (M3D_FLOAT)65535.0; + break; + default: + tcoord.data.u = model->tmap[i].u; + tcoord.data.v = model->tmap[i].v; + break; + } + if(flags & M3D_EXP_FLIPTXTCRD) + tcoord.data.v = (M3D_FLOAT)1.0 - tcoord.data.v; + tcoord.oldidx = i; + memcpy(&tmap[numtmap++], &tcoord, sizeof(m3dtisave_t)); + } + if(numtmap) { + qsort(tmap, numtmap, sizeof(m3dtisave_t), _m3d_ticmp); + memcpy(&tcoord.data, &tmap[0], sizeof(m3dti_t)); + for(i = 0; i < numtmap; i++) { + if(memcmp(&tcoord.data, &tmap[i].data, sizeof(m3dti_t))) { + memcpy(&tcoord.data, &tmap[i].data, sizeof(m3dti_t)); + maxtmap++; + } + tmap[i].newidx = maxtmap; + tmapidx[tmap[i].oldidx] = maxtmap; + } + maxtmap++; + } + } + if(model->numskin && model->skin && !(flags & M3D_EXP_NOBONE)) { + M3D_LOG("Compressing skin"); + skinidx = (M3D_INDEX*)M3D_MALLOC(model->numskin * sizeof(M3D_INDEX)); + if(!skinidx) goto memerr; + skin = (m3dssave_t*)M3D_MALLOC(model->numskin * sizeof(m3dssave_t)); + if(!skin) goto memerr; + memset(skinidx, 255, model->numskin * sizeof(M3D_INDEX)); + for(i = 0; i < model->numvertex; i++) { + if(vrtxidx[i] != M3D_UNDEF && model->vertex[i].skinid < model->numskin) + skinidx[model->vertex[i].skinid] = 0; + } + for(i = 0; i < model->numskin; i++) { + if(skinidx[i] == M3D_UNDEF) continue; + memset(&sk, 0, sizeof(m3dssave_t)); + for(j = 0, min_x = (M3D_FLOAT)0.0; j < M3D_NUMBONE && model->skin[i].boneid[j] != M3D_UNDEF; j++) { + sk.data.boneid[j] = model->skin[i].boneid[j]; + sk.data.weight[j] = model->skin[i].weight[j] > (M3D_FLOAT)0.0 ? model->skin[i].weight[j] : (M3D_FLOAT)0.01; + min_x += sk.data.weight[j]; + } + if(j > maxbone) maxbone = j; + if(min_x != (M3D_FLOAT)1.0 && min_x != (M3D_FLOAT)0.0) + for(j = 0; j < M3D_NUMBONE && sk.data.weight[j] > (M3D_FLOAT)0.0; j++) + sk.data.weight[j] /= min_x; + sk.oldidx = i; + memcpy(&skin[numskin++], &sk, sizeof(m3dssave_t)); + } + if(numskin) { + qsort(skin, numskin, sizeof(m3dssave_t), _m3d_skincmp); + memcpy(&sk.data, &skin[0].data, sizeof(m3ds_t)); + for(i = 0; i < numskin; i++) { + if(memcmp(&sk.data, &skin[i].data, sizeof(m3ds_t))) { + memcpy(&sk.data, &skin[i].data, sizeof(m3ds_t)); + maxskin++; + } + skin[i].newidx = maxskin; + skinidx[skin[i].oldidx] = maxskin; + } + maxskin++; + } + } + + M3D_LOG("Compressing vertex list"); + min_x = min_y = min_z = (M3D_FLOAT)1e10; + max_x = max_y = max_z = (M3D_FLOAT)-1e10; + if(vrtxidx) { + vrtx = (m3dvsave_t*)M3D_MALLOC(model->numvertex * sizeof(m3dvsave_t)); + if(!vrtx) goto memerr; + for(i = numvrtx = 0; i < model->numvertex; i++) { + if(vrtxidx[i] == M3D_UNDEF) continue; + _m3d_round(quality, &model->vertex[i], &vertex.data); + vertex.norm = norm ? norm[i] : 0; + if(vertex.data.skinid != M3D_INDEXMAX && !vertex.norm) { + vertex.data.skinid = vertex.data.skinid != M3D_UNDEF && skinidx ? skinidx[vertex.data.skinid] : M3D_UNDEF; + if(vertex.data.x > max_x) max_x = vertex.data.x; + if(vertex.data.x < min_x) min_x = vertex.data.x; + if(vertex.data.y > max_y) max_y = vertex.data.y; + if(vertex.data.y < min_y) min_y = vertex.data.y; + if(vertex.data.z > max_z) max_z = vertex.data.z; + if(vertex.data.z < min_z) min_z = vertex.data.z; + } +#ifdef M3D_VERTEXTYPE + vertex.data.type = 0; +#endif + vertex.oldidx = i; + memcpy(&vrtx[numvrtx++], &vertex, sizeof(m3dvsave_t)); + } + if(numvrtx) { + qsort(vrtx, numvrtx, sizeof(m3dvsave_t), _m3d_vrtxcmp); + memcpy(&vertex.data, &vrtx[0].data, sizeof(m3dv_t)); + for(i = 0; i < numvrtx; i++) { + if(memcmp(&vertex.data, &vrtx[i].data, vrtx[i].norm ? 3 * sizeof(M3D_FLOAT) : sizeof(m3dv_t))) { + memcpy(&vertex.data, &vrtx[i].data, sizeof(m3dv_t)); + maxvrtx++; + } + vrtx[i].newidx = maxvrtx; + vrtxidx[vrtx[i].oldidx] = maxvrtx; + } + maxvrtx++; + } + } + if(norm) { M3D_FREE(norm); norm = NULL; } + + /* normalize to bounding cube */ + if(numvrtx && !(flags & M3D_EXP_NORECALC)) { + M3D_LOG("Normalizing coordinates"); + if(min_x < (M3D_FLOAT)0.0) min_x = -min_x; + if(max_x < (M3D_FLOAT)0.0) max_x = -max_x; + if(min_y < (M3D_FLOAT)0.0) min_y = -min_y; + if(max_y < (M3D_FLOAT)0.0) max_y = -max_y; + if(min_z < (M3D_FLOAT)0.0) min_z = -min_z; + if(max_z < (M3D_FLOAT)0.0) max_z = -max_z; + scale = min_x; + if(max_x > scale) scale = max_x; + if(min_y > scale) scale = min_y; + if(max_y > scale) scale = max_y; + if(min_z > scale) scale = min_z; + if(max_z > scale) scale = max_z; + if(scale <= (M3D_FLOAT)0.0) scale = (M3D_FLOAT)1.0; + if(scale != (M3D_FLOAT)1.0) { + for(i = 0; i < numvrtx; i++) { + if(vrtx[i].data.skinid == M3D_INDEXMAX) continue; + vrtx[i].data.x /= scale; + vrtx[i].data.y /= scale; + vrtx[i].data.z /= scale; + } + } + } + if(model->scale > (M3D_FLOAT)0.0) scale = model->scale; + if(scale <= (M3D_FLOAT)0.0) scale = (M3D_FLOAT)1.0; + + /* meta info */ + sn = _m3d_safestr(model->name && *model->name ? model->name : (char*)"(noname)", 2); + sl = _m3d_safestr(model->license ? model->license : (char*)"MIT", 2); + sa = _m3d_safestr(model->author ? model->author : getenv("LOGNAME"), 2); + if(!sn || !sl || !sa) { +memerr: if(vrtxidx) M3D_FREE(vrtxidx); + if(mtrlidx) M3D_FREE(mtrlidx); + if(tmapidx) M3D_FREE(tmapidx); + if(skinidx) M3D_FREE(skinidx); + if(grpidx) M3D_FREE(grpidx); + if(norm) M3D_FREE(norm); + if(face) M3D_FREE(face); + if(cmap) M3D_FREE(cmap); + if(tmap) M3D_FREE(tmap); + if(skin) M3D_FREE(skin); + if(str) M3D_FREE(str); + if(vrtx) M3D_FREE(vrtx); + if(sn) M3D_FREE(sn); + if(sl) M3D_FREE(sl); + if(sa) M3D_FREE(sa); + if(sd) M3D_FREE(sd); + if(out) M3D_FREE(out); + if(opa) M3D_FREE(opa); + if(h) M3D_FREE(h); + M3D_LOG("Out of memory"); + model->errcode = M3D_ERR_ALLOC; + return NULL; + } + + M3D_LOG("Serializing model"); +#ifdef M3D_ASCII + if(flags & M3D_EXP_ASCII) { + /* use CRLF to make model creators on Win happy... */ + sd = _m3d_safestr(model->desc, 1); + if(!sd) goto memerr; + ol = setlocale(LC_NUMERIC, NULL); + setlocale(LC_NUMERIC, "C"); + /* header */ + len = 64 + (unsigned int)(strlen(sn) + strlen(sl) + strlen(sa) + strlen(sd)); + out = (unsigned char*)M3D_MALLOC(len); + if(!out) { setlocale(LC_NUMERIC, ol); goto memerr; } + ptr = (char*)out; + ptr += sprintf(ptr, "3dmodel %g\r\n%s\r\n%s\r\n%s\r\n%s\r\n\r\n", scale, + sn, sl, sa, sd); + M3D_FREE(sl); M3D_FREE(sa); M3D_FREE(sd); + sl = sa = sd = NULL; + /* preview chunk */ + if(model->preview.data && model->preview.length) { + sl = _m3d_safestr(sn, 0); + if(sl) { +/* gcc thinks that "ptr is used after free", well, gcc is simply wrong. */ +#ifdef __GNUC__ +#pragma GCC diagnostic push +#pragma GCC diagnostic ignored "-Wuse-after-free" +#endif + ptr -= (uintptr_t)out; len = (unsigned int)((uintptr_t)ptr + (uintptr_t)20 + strlen(sl)); + out = (unsigned char*)M3D_REALLOC(out, len); ptr += (uintptr_t)out; +#ifdef __GNUC__ +#pragma GCC diagnostic pop +#endif + if(!out) { setlocale(LC_NUMERIC, ol); goto memerr; } + ptr += sprintf(ptr, "Preview\r\n%s.png\r\n\r\n", sl); + M3D_FREE(sl); sl = NULL; + } + } + M3D_FREE(sn); sn = NULL; + /* texture map */ + if(numtmap && tmap && !(flags & M3D_EXP_NOTXTCRD) && !(flags & M3D_EXP_NOFACE)) { +/* interestingly gcc does not complain about "ptr is used after free" here, although the code is 100% the same */ + ptr -= (uintptr_t)out; len = (unsigned int)((uintptr_t)ptr + (uintptr_t)(maxtmap * 32) + (uintptr_t)12); + out = (unsigned char*)M3D_REALLOC(out, len); ptr += (uintptr_t)out; + if(!out) { setlocale(LC_NUMERIC, ol); goto memerr; } + ptr += sprintf(ptr, "Textmap\r\n"); + last = M3D_UNDEF; + for(i = 0; i < numtmap; i++) { + if(tmap[i].newidx == last) continue; + last = tmap[i].newidx; + ptr += sprintf(ptr, "%g %g\r\n", tmap[i].data.u, tmap[i].data.v); + } + ptr += sprintf(ptr, "\r\n"); + } + /* vertex chunk */ + if(numvrtx && vrtx && !(flags & M3D_EXP_NOFACE)) { + ptr -= (uintptr_t)out; len = (unsigned int)((uintptr_t)ptr + (uintptr_t)(maxvrtx * 128) + (uintptr_t)10); + out = (unsigned char*)M3D_REALLOC(out, len); ptr += (uintptr_t)out; + if(!out) { setlocale(LC_NUMERIC, ol); goto memerr; } + ptr += sprintf(ptr, "Vertex\r\n"); + last = M3D_UNDEF; + for(i = 0; i < numvrtx; i++) { + if(vrtx[i].newidx == last) continue; + last = vrtx[i].newidx; + ptr += sprintf(ptr, "%g %g %g %g", vrtx[i].data.x, vrtx[i].data.y, vrtx[i].data.z, vrtx[i].data.w); + if(!(flags & M3D_EXP_NOCMAP) && vrtx[i].data.color) + ptr += sprintf(ptr, " #%08x", vrtx[i].data.color); + if(!(flags & M3D_EXP_NOBONE) && model->numbone && maxskin && vrtx[i].data.skinid < M3D_INDEXMAX) { + if(skin[vrtx[i].data.skinid].data.weight[0] == (M3D_FLOAT)1.0) + ptr += sprintf(ptr, " %d", skin[vrtx[i].data.skinid].data.boneid[0]); + else + for(j = 0; j < M3D_NUMBONE && skin[vrtx[i].data.skinid].data.boneid[j] != M3D_UNDEF && + skin[vrtx[i].data.skinid].data.weight[j] > (M3D_FLOAT)0.0; j++) + ptr += sprintf(ptr, " %d:%g", skin[vrtx[i].data.skinid].data.boneid[j], + skin[vrtx[i].data.skinid].data.weight[j]); + } + ptr += sprintf(ptr, "\r\n"); + } + ptr += sprintf(ptr, "\r\n"); + } + /* bones chunk */ + if(model->numbone && model->bone && !(flags & M3D_EXP_NOBONE)) { + ptr -= (uintptr_t)out; len = (unsigned int)((uintptr_t)ptr + (uintptr_t)9); + for(i = 0; i < model->numbone; i++) { + len += (unsigned int)strlen(model->bone[i].name) + 128; + } + out = (unsigned char*)M3D_REALLOC(out, len); ptr += (uintptr_t)out; + if(!out) { setlocale(LC_NUMERIC, ol); goto memerr; } + ptr += sprintf(ptr, "Bones\r\n"); + ptr = _m3d_prtbone(ptr, model->bone, model->numbone, M3D_UNDEF, 0, vrtxidx); + ptr += sprintf(ptr, "\r\n"); + } + /* materials */ + if(model->nummaterial && !(flags & M3D_EXP_NOMATERIAL)) { + for(j = 0; j < model->nummaterial; j++) { + if(mtrlidx[j] == M3D_UNDEF || !model->material[j].numprop || !model->material[j].prop) continue; + m = &model->material[j]; + sn = _m3d_safestr(m->name, 0); + if(!sn) { setlocale(LC_NUMERIC, ol); goto memerr; } + ptr -= (uintptr_t)out; len = (unsigned int)((uintptr_t)ptr + (uintptr_t)strlen(sn) + (uintptr_t)12); + for(i = 0; i < m->numprop; i++) { + if(m->prop[i].type < 128) + len += 32; + else if(m->prop[i].value.textureid < model->numtexture && model->texture[m->prop[i].value.textureid].name) + len += (unsigned int)strlen(model->texture[m->prop[i].value.textureid].name) + 16; + } + out = (unsigned char*)M3D_REALLOC(out, len); ptr += (uintptr_t)out; + if(!out) { setlocale(LC_NUMERIC, ol); goto memerr; } + ptr += sprintf(ptr, "Material %s\r\n", sn); + M3D_FREE(sn); sn = NULL; + for(i = 0; i < m->numprop; i++) { + k = 256; + if(m->prop[i].type >= 128) { + for(l = 0; l < sizeof(m3d_propertytypes)/sizeof(m3d_propertytypes[0]); l++) + if(m->prop[i].type == m3d_propertytypes[l].id) { + sn = m3d_propertytypes[l].key; + break; + } + if(!sn) + for(l = 0; l < sizeof(m3d_propertytypes)/sizeof(m3d_propertytypes[0]); l++) + if(m->prop[i].type - 128 == m3d_propertytypes[l].id) { + sn = m3d_propertytypes[l].key; + break; + } + k = sn ? m3dpf_map : 256; + } else { + for(l = 0; l < sizeof(m3d_propertytypes)/sizeof(m3d_propertytypes[0]); l++) + if(m->prop[i].type == m3d_propertytypes[l].id) { + sn = m3d_propertytypes[l].key; + k = m3d_propertytypes[l].format; + break; + } + } + switch(k) { + case m3dpf_color: ptr += sprintf(ptr, "%s #%08x\r\n", sn, m->prop[i].value.color); break; + case m3dpf_uint8: + case m3dpf_uint16: + case m3dpf_uint32: ptr += sprintf(ptr, "%s %d\r\n", sn, m->prop[i].value.num); break; + case m3dpf_float: ptr += sprintf(ptr, "%s %g\r\n", sn, m->prop[i].value.fnum); break; + case m3dpf_map: + if(m->prop[i].value.textureid < model->numtexture && + model->texture[m->prop[i].value.textureid].name) { + sl = _m3d_safestr(model->texture[m->prop[i].value.textureid].name, 0); + if(!sl) { setlocale(LC_NUMERIC, ol); goto memerr; } + if(*sl) + ptr += sprintf(ptr, "map_%s %s\r\n", sn, sl); + M3D_FREE(sn); M3D_FREE(sl); sl = NULL; + } + break; + } + sn = NULL; + } + ptr += sprintf(ptr, "\r\n"); + } + } + /* procedural face */ + if(model->numinlined && model->inlined && !(flags & M3D_EXP_NOFACE)) { + /* all inlined assets which are not textures should be procedural surfaces */ + for(j = 0; j < model->numinlined; j++) { + if(!model->inlined[j].name || !*model->inlined[j].name || !model->inlined[j].length || !model->inlined[j].data || + (model->inlined[j].data[1] == 'P' && model->inlined[j].data[2] == 'N' && model->inlined[j].data[3] == 'G')) + continue; + for(i = k = 0; i < model->numtexture; i++) { + if(!strcmp(model->inlined[j].name, model->texture[i].name)) { k = 1; break; } + } + if(k) continue; + sn = _m3d_safestr(model->inlined[j].name, 0); + if(!sn) { setlocale(LC_NUMERIC, ol); goto memerr; } + ptr -= (uintptr_t)out; len = (unsigned int)((uintptr_t)ptr + (uintptr_t)strlen(sn) + (uintptr_t)18); + out = (unsigned char*)M3D_REALLOC(out, len); ptr += (uintptr_t)out; + if(!out) { setlocale(LC_NUMERIC, ol); goto memerr; } + ptr += sprintf(ptr, "Procedural\r\n%s\r\n\r\n", sn); + M3D_FREE(sn); sn = NULL; + } + } + /* mesh face */ + if(model->numface && face && !(flags & M3D_EXP_NOFACE)) { + ptr -= (uintptr_t)out; len = (unsigned int)((uintptr_t)ptr + (uintptr_t)(model->numface * 128) + (uintptr_t)6); + last = M3D_UNDEF; +#ifdef M3D_VERTEXMAX + lastp = M3D_UNDEF; +#endif + if(!(flags & M3D_EXP_NOMATERIAL)) + for(i = 0; i < model->numface; i++) { + j = face[i].data.materialid < model->nummaterial ? face[i].data.materialid : M3D_UNDEF; + if(j != last) { + last = j; + if(last < model->nummaterial) + len += (unsigned int)strlen(model->material[last].name); + len += 6; + } +#ifdef M3D_VERTEXMAX + j = face[i].data.paramid < model->numparam ? face[i].data.paramid : M3D_UNDEF; + if(j != lastp) { + lastp = j; + if(lastp < model->numparam) + len += (unsigned int)strlen(model->param[lastp].name); + len += 6; + } +#endif + } + out = (unsigned char*)M3D_REALLOC(out, len); ptr += (uintptr_t)out; + if(!out) { setlocale(LC_NUMERIC, ol); goto memerr; } + ptr += sprintf(ptr, "Mesh\r\n"); + last = M3D_UNDEF; +#ifdef M3D_VERTEXMAX + lastp = M3D_UNDEF; +#endif + for(i = 0; i < model->numface; i++) { + j = face[i].data.materialid < model->nummaterial ? face[i].data.materialid : M3D_UNDEF; + if(!(flags & M3D_EXP_NOMATERIAL) && j != last) { + last = j; + if(last < model->nummaterial) { + sn = _m3d_safestr(model->material[last].name, 0); + if(!sn) { setlocale(LC_NUMERIC, ol); goto memerr; } + ptr += sprintf(ptr, "use %s\r\n", sn); + M3D_FREE(sn); sn = NULL; + } else + ptr += sprintf(ptr, "use\r\n"); + } +#ifdef M3D_VERTEXMAX + j = face[i].data.paramid < model->numparam ? face[i].data.paramid : M3D_UNDEF; + if(!(flags & M3D_EXP_NOVRTMAX) && j != lastp) { + lastp = j; + if(lastp < model->numparam) { + sn = _m3d_safestr(model->param[lastp].name, 0); + if(!sn) { setlocale(LC_NUMERIC, ol); goto memerr; } + ptr += sprintf(ptr, "par %s\r\n", sn); + M3D_FREE(sn); sn = NULL; + } else + ptr += sprintf(ptr, "par\r\n"); + } +#endif + /* hardcoded triangles. Should be repeated as many times as the number of edges in polygon */ + for(j = 0; j < 3; j++) { + ptr += sprintf(ptr, "%s%d", j?" ":"", vrtxidx[face[i].data.vertex[j]]); + k = l = M3D_NOTDEFINED; + if(!(flags & M3D_EXP_NOTXTCRD) && (face[i].data.texcoord[j] != M3D_UNDEF) && + (tmapidx[face[i].data.texcoord[j]] != M3D_UNDEF)) { + k = tmapidx[face[i].data.texcoord[j]]; + ptr += sprintf(ptr, "/%d", k); + } + if(!(flags & M3D_EXP_NONORMAL) && (face[i].data.normal[j] != M3D_UNDEF)) { + l = vrtxidx[face[i].data.normal[j]]; + ptr += sprintf(ptr, "%s/%d", k == M3D_NOTDEFINED? "/" : "", l); + } +#ifdef M3D_VERTEXMAX + if(!(flags & M3D_EXP_NOVRTMAX) && (face[i].data.vertmax[j] != M3D_UNDEF)) { + ptr += sprintf(ptr, "%s%s/%d", k == M3D_NOTDEFINED? "/" : "", l == M3D_NOTDEFINED? "/" : "", + vrtxidx[face[i].data.vertmax[j]]); + } +#endif + } + ptr += sprintf(ptr, "\r\n"); + } + ptr += sprintf(ptr, "\r\n"); + } + /* voxel face */ + if(model->numvoxtype && model->voxtype && !(flags & M3D_EXP_NOFACE)) { + ptr -= (uintptr_t)out; len = (unsigned int)((uintptr_t)ptr + (uintptr_t)(model->numvoxtype * 128) + (uintptr_t)10); + for(i = 0; i < model->numvoxtype; i++) { + if(model->voxtype[i].name) len += (unsigned int)strlen(model->voxtype[i].name); + for(j = 0; j < model->voxtype[i].numitem; j++) + if(model->voxtype[i].item[j].name) + len += (unsigned int)strlen(model->voxtype[i].item[j].name) + 6; + } + out = (unsigned char*)M3D_REALLOC(out, len); ptr += (uintptr_t)out; + if(!out) { setlocale(LC_NUMERIC, ol); goto memerr; } + ptr += sprintf(ptr, "VoxTypes\r\n"); + for(i = 0; i < model->numvoxtype; i++) { + ptr += sprintf(ptr, "#%08x", model->voxtype[i].color); + if(model->voxtype[i].rotation) + ptr += sprintf(ptr, "/%02x", model->voxtype[i].rotation); + if(model->voxtype[i].voxshape) + ptr += sprintf(ptr, "%s/%03x", model->voxtype[i].rotation ? "" : "/", model->voxtype[i].voxshape); + sn = _m3d_safestr(model->voxtype[i].name, 0); + if(!sn) { setlocale(LC_NUMERIC, ol); goto memerr; } + ptr += sprintf(ptr, " %s", sn && sn[0] ? sn : "-"); + M3D_FREE(sn); sn = NULL; + if(!(flags & M3D_EXP_NOBONE) && model->numbone && maxskin && model->voxtype[i].skinid < M3D_INDEXMAX) { + if(skin[skinidx[model->voxtype[i].skinid]].data.weight[0] == (M3D_FLOAT)1.0) + ptr += sprintf(ptr, " %d", skin[skinidx[model->voxtype[i].skinid]].data.boneid[0]); + else + for(j = 0; j < M3D_NUMBONE && skin[skinidx[model->voxtype[i].skinid]].data.boneid[j] != M3D_UNDEF && + skin[skinidx[model->voxtype[i].skinid]].data.weight[j] > (M3D_FLOAT)0.0; j++) + ptr += sprintf(ptr, " %d:%g", skin[skinidx[model->voxtype[i].skinid]].data.boneid[j], + skin[skinidx[model->voxtype[i].skinid]].data.weight[j]); + } + if(model->voxtype[i].numitem && model->voxtype[i].item) { + for(j = k = 0; j < model->voxtype[i].numitem; j++) { + if(!model->voxtype[i].item[j].count || !model->voxtype[i].item[j].name || + !model->voxtype[i].item[j].name[0]) continue; + if(!k) { ptr += sprintf(ptr, " {"); k = 1; } + sn = _m3d_safestr(model->voxtype[i].item[j].name, 0); + if(!sn) { setlocale(LC_NUMERIC, ol); goto memerr; } + ptr += sprintf(ptr, " %d %s", model->voxtype[i].item[j].count, sn); + M3D_FREE(sn); sn = NULL; + } + if(k) ptr += sprintf(ptr, " }"); + } + while(ptr[-1] == '-' || ptr[-1] == ' ') ptr--; + ptr += sprintf(ptr, "\r\n"); + } + ptr += sprintf(ptr, "\r\n"); + } + if(model->numvoxel && model->voxel && !(flags & M3D_EXP_NOFACE)) { + for(i = 0; i < model->numvoxel; i++) { + ptr -= (uintptr_t)out; len = (unsigned int)((uintptr_t)ptr + (uintptr_t)128); + if(model->voxel[i].name) len += (unsigned int)strlen(model->voxel[i].name); + len += model->voxel[i].h * ((model->voxel[i].w * 6 + 2) * model->voxel[i].d + 9); + out = (unsigned char*)M3D_REALLOC(out, len); ptr += (uintptr_t)out; + if(!out) { setlocale(LC_NUMERIC, ol); goto memerr; } + ptr += sprintf(ptr, "Voxel"); + sn = _m3d_safestr(model->voxel[i].name, 0); + if(!sn) { setlocale(LC_NUMERIC, ol); goto memerr; } + if(sn && sn[0]) + ptr += sprintf(ptr, " %s", sn); + M3D_FREE(sn); sn = NULL; + ptr += sprintf(ptr, "\r\n"); + if(model->voxel[i].uncertain) + ptr += sprintf(ptr, "uncertain %d %d\r\n", (model->voxel[i].uncertain * 100) / 255, model->voxel[i].groupid); + if(model->voxel[i].x || model->voxel[i].y || model->voxel[i].z) + ptr += sprintf(ptr, "pos %d %d %d\r\n", model->voxel[i].x, model->voxel[i].y, model->voxel[i].z); + ptr += sprintf(ptr, "dim %d %d %d\r\n", model->voxel[i].w, model->voxel[i].h, model->voxel[i].d); + for(j = n = 0; j < model->voxel[i].h; j++) { + ptr += sprintf(ptr, "layer\r\n"); + for(k = 0; k < model->voxel[i].d; k++) { + for(l = 0; l < model->voxel[i].w; l++, n++) { + switch(model->voxel[i].data[n]) { + case M3D_VOXCLEAR: *ptr++ = '-'; break; + case M3D_VOXUNDEF: *ptr++ = '.'; break; + default: ptr += sprintf(ptr, "%d", model->voxel[i].data[n]); break; + } + *ptr++ = ' '; + } + ptr--; + ptr += sprintf(ptr, "\r\n"); + } + } + ptr += sprintf(ptr, "\r\n"); + } + } + /* mathematical shapes face */ + if(model->numshape && model->numshape && !(flags & M3D_EXP_NOFACE)) { + for(j = 0; j < model->numshape; j++) { + sn = _m3d_safestr(model->shape[j].name, 0); + if(!sn) { setlocale(LC_NUMERIC, ol); goto memerr; } + ptr -= (uintptr_t)out; len = (unsigned int)((uintptr_t)ptr + (uintptr_t)strlen(sn) + (uintptr_t)33); + out = (unsigned char*)M3D_REALLOC(out, len); ptr += (uintptr_t)out; + if(!out) { setlocale(LC_NUMERIC, ol); goto memerr; } + ptr += sprintf(ptr, "Shape %s\r\n", sn); + M3D_FREE(sn); sn = NULL; + if(model->shape[j].group != M3D_UNDEF && !(flags & M3D_EXP_NOBONE)) + ptr += sprintf(ptr, "group %d\r\n", model->shape[j].group); + for(i = 0; i < model->shape[j].numcmd; i++) { + cmd = &model->shape[j].cmd[i]; + if(cmd->type >= (unsigned int)(sizeof(m3d_commandtypes)/sizeof(m3d_commandtypes[0])) || !cmd->arg) + continue; + cd = &m3d_commandtypes[cmd->type]; + ptr -= (uintptr_t)out; len = (unsigned int)((uintptr_t)ptr + (uintptr_t)strlen(cd->key) + (uintptr_t)3); + for(k = 0; k < cd->p; k++) + switch(cd->a[k]) { + case m3dcp_mi_t: if(cmd->arg[k] != M3D_NOTDEFINED) { len += (unsigned int)strlen(model->material[cmd->arg[k]].name) + 1; } break; + case m3dcp_va_t: len += cmd->arg[k] * (cd->p - k - 1) * 16; k = cd->p; break; + default: len += 16; break; + } + out = (unsigned char*)M3D_REALLOC(out, len); ptr += (uintptr_t)out; + if(!out) { setlocale(LC_NUMERIC, ol); goto memerr; } + ptr += sprintf(ptr, "%s", cd->key); + for(k = n = 0, l = cd->p; k < l; k++) { + switch(cd->a[((k - n) % (cd->p - n)) + n]) { + case m3dcp_mi_t: + if(cmd->arg[k] != M3D_NOTDEFINED) { + sn = _m3d_safestr(model->material[cmd->arg[k]].name, 0); + if(!sn) { setlocale(LC_NUMERIC, ol); goto memerr; } + ptr += sprintf(ptr, " %s", sn); + M3D_FREE(sn); sn = NULL; + } + break; + case m3dcp_vc_t: ptr += sprintf(ptr, " %g", *((float*)&cmd->arg[k])); break; + case m3dcp_va_t: ptr += sprintf(ptr, " %d[", cmd->arg[k]); + n = k + 1; l += (cmd->arg[k] - 1) * (cd->p - k - 1); + break; + default: ptr += sprintf(ptr, " %d", cmd->arg[k]); break; + } + } + ptr += sprintf(ptr, "%s\r\n", l > cd->p ? " ]" : ""); + } + ptr += sprintf(ptr, "\r\n"); + } + } + /* annotation labels */ + if(model->numlabel && model->label && !(flags & M3D_EXP_NOFACE)) { + for(i = 0, j = 3, length = NULL; i < model->numlabel; i++) { + if(model->label[i].name) j += (unsigned int)strlen(model->label[i].name); + if(model->label[i].lang) j += (unsigned int)strlen(model->label[i].lang); + if(model->label[i].text) j += (unsigned int)strlen(model->label[i].text); + j += 40; + } + ptr -= (uintptr_t)out; len = (unsigned int)((uintptr_t)ptr + (uintptr_t)j); + out = (unsigned char*)M3D_REALLOC(out, len); ptr += (uintptr_t)out; + if(!out) { setlocale(LC_NUMERIC, ol); goto memerr; } + for(i = 0; i < model->numlabel; i++) { + if(!i || _m3d_strcmp(sl, model->label[i].lang) || _m3d_strcmp(sn, model->label[i].name)) { + sl = model->label[i].lang; + sn = model->label[i].name; + sd = _m3d_safestr(sn, 0); + if(!sd) { setlocale(LC_NUMERIC, ol); sn = sl = NULL; goto memerr; } + if(i) ptr += sprintf(ptr, "\r\n"); + ptr += sprintf(ptr, "Labels %s\r\n", sd); + M3D_FREE(sd); sd = NULL; + if(model->label[i].color) + ptr += sprintf(ptr, "color #0x%08x\r\n", model->label[i].color); + if(sl && *sl) { + sd = _m3d_safestr(sl, 0); + if(!sd) { setlocale(LC_NUMERIC, ol); sn = sl = NULL; goto memerr; } + ptr += sprintf(ptr, "lang %s\r\n", sd); + M3D_FREE(sd); sd = NULL; + } + } + sd = _m3d_safestr(model->label[i].text, 2); + if(!sd) { setlocale(LC_NUMERIC, ol); sn = sl = NULL; goto memerr; } + ptr += sprintf(ptr, "%d %s\r\n", model->label[i].vertexid, sd); + M3D_FREE(sd); sd = NULL; + } + ptr += sprintf(ptr, "\r\n"); + sn = sl = NULL; + } + /* actions */ + if(model->numaction && model->action && !(flags & M3D_EXP_NOACTION)) { + for(j = 0; j < model->numaction; j++) { + a = &model->action[j]; + sn = _m3d_safestr(a->name, 0); + if(!sn) { setlocale(LC_NUMERIC, ol); goto memerr; } + ptr -= (uintptr_t)out; len = (unsigned int)((uintptr_t)ptr + (uintptr_t)strlen(sn) + (uintptr_t)48); + for(i = 0; i < a->numframe; i++) + len += a->frame[i].numtransform * 128 + 8; + out = (unsigned char*)M3D_REALLOC(out, len); ptr += (uintptr_t)out; + if(!out) { setlocale(LC_NUMERIC, ol); goto memerr; } + ptr += sprintf(ptr, "Action %d %s\r\n", a->durationmsec, sn); + M3D_FREE(sn); sn = NULL; + for(i = 0; i < a->numframe; i++) { + ptr += sprintf(ptr, "frame %d\r\n", a->frame[i].msec); + for(k = 0; k < a->frame[i].numtransform; k++) { + ptr += sprintf(ptr, "%d %d %d\r\n", a->frame[i].transform[k].boneid, + vrtxidx[a->frame[i].transform[k].pos], vrtxidx[a->frame[i].transform[k].ori]); + } + } + ptr += sprintf(ptr, "\r\n"); + } + } + /* inlined assets */ + if(model->numinlined && model->inlined) { + for(i = j = 0; i < model->numinlined; i++) + if(model->inlined[i].name) + j += (unsigned int)strlen(model->inlined[i].name) + 6; + if(j > 0) { + ptr -= (uintptr_t)out; len = (unsigned int)((uintptr_t)ptr + (uintptr_t)j + (uintptr_t)16); + out = (unsigned char*)M3D_REALLOC(out, len); ptr += (uintptr_t)out; + if(!out) { setlocale(LC_NUMERIC, ol); goto memerr; } + ptr += sprintf(ptr, "Assets\r\n"); + for(i = 0; i < model->numinlined; i++) + if(model->inlined[i].name) + ptr += sprintf(ptr, "%s%s\r\n", model->inlined[i].name, strrchr(model->inlined[i].name, '.') ? "" : ".png"); + ptr += sprintf(ptr, "\r\n"); + } + } + /* extra info */ + if(model->numextra && (flags & M3D_EXP_EXTRA)) { + for(i = 0; i < model->numextra; i++) { + if(model->extra[i]->length < 9) continue; + ptr -= (uintptr_t)out; len = (unsigned int)((uintptr_t)ptr + (uintptr_t)17 + (uintptr_t)(model->extra[i]->length * 3)); + out = (unsigned char*)M3D_REALLOC(out, len); ptr += (uintptr_t)out; + if(!out) { setlocale(LC_NUMERIC, ol); goto memerr; } + ptr += sprintf(ptr, "Extra %c%c%c%c\r\n", + model->extra[i]->magic[0] > ' ' ? model->extra[i]->magic[0] : '_', + model->extra[i]->magic[1] > ' ' ? model->extra[i]->magic[1] : '_', + model->extra[i]->magic[2] > ' ' ? model->extra[i]->magic[2] : '_', + model->extra[i]->magic[3] > ' ' ? model->extra[i]->magic[3] : '_'); + for(j = 0; j < model->extra[i]->length; j++) + ptr += sprintf(ptr, "%02x ", *((unsigned char *)model->extra + sizeof(m3dchunk_t) + j)); + ptr--; + ptr += sprintf(ptr, "\r\n\r\n"); + } + } + setlocale(LC_NUMERIC, ol); + len = (unsigned int)((uintptr_t)ptr - (uintptr_t)out); + out = (unsigned char*)M3D_REALLOC(out, len + 1); + if(!out) goto memerr; + out[len] = 0; + } else +#endif + { + /* stricly only use LF (newline) in binary */ + sd = _m3d_safestr(model->desc, 3); + if(!sd) goto memerr; + /* header */ + h = (m3dhdr_t*)M3D_MALLOC(sizeof(m3dhdr_t) + strlen(sn) + strlen(sl) + strlen(sa) + strlen(sd) + 4); + if(!h) goto memerr; + memcpy((uint8_t*)h, "HEAD", 4); + h->length = sizeof(m3dhdr_t); + h->scale = scale; + i = (unsigned int)strlen(sn); memcpy((uint8_t*)h + h->length, sn, i+1); h->length += i+1; M3D_FREE(sn); + i = (unsigned int)strlen(sl); memcpy((uint8_t*)h + h->length, sl, i+1); h->length += i+1; M3D_FREE(sl); + i = (unsigned int)strlen(sa); memcpy((uint8_t*)h + h->length, sa, i+1); h->length += i+1; M3D_FREE(sa); + i = (unsigned int)strlen(sd); memcpy((uint8_t*)h + h->length, sd, i+1); h->length += i+1; M3D_FREE(sd); + sn = sl = sa = sd = NULL; + if(model->inlined) + for(i = 0; i < model->numinlined; i++) { + if(model->inlined[i].name && *model->inlined[i].name && model->inlined[i].length > 0) { + str = _m3d_addstr(str, &numstr, model->inlined[i].name); + if(!str) goto memerr; + } + } + if(str) + for(i = 0; i < numstr; i++) { + h = _m3d_addhdr(h, &str[i]); + if(!h) goto memerr; + } + vc_s = quality == M3D_EXP_INT8? 1 : (quality == M3D_EXP_INT16? 2 : (quality == M3D_EXP_DOUBLE? 8 : 4)); + vi_s = maxvrtx < 254 ? 1 : (maxvrtx < 65534 ? 2 : 4); + si_s = h->length - 16 < 254 ? 1 : (h->length - 16 < 65534 ? 2 : 4); + ci_s = !numcmap || !cmap ? 0 : (numcmap < 254 ? 1 : (numcmap < 65534 ? 2 : 4)); + ti_s = !maxtmap || !tmap ? 0 : (maxtmap < 254 ? 1 : (maxtmap < 65534 ? 2 : 4)); + bi_s = !model->numbone || !model->bone || (flags & M3D_EXP_NOBONE)? 0 : (model->numbone < 254 ? 1 : + (model->numbone < 65534 ? 2 : 4)); + nb_s = maxbone < 2 ? 1 : (maxbone == 2 ? 2 : (maxbone <= 4 ? 4 : 8)); + sk_s = !bi_s || !maxskin || !skin ? 0 : (maxskin < 254 ? 1 : (maxskin < 65534 ? 2 : 4)); + fc_s = maxt < 254 ? 1 : (maxt < 65534 ? 2 : 4); + hi_s = !model->numshape || !model->shape || (flags & M3D_EXP_NOFACE)? 0 : (model->numshape < 254 ? 1 : + (model->numshape < 65534 ? 2 : 4)); + fi_s = !model->numface || !model->face || (flags & M3D_EXP_NOFACE)? 0 : (model->numface < 254 ? 1 : + (model->numface < 65534 ? 2 : 4)); + vd_s = !model->numvoxel || !model->voxel || (flags & M3D_EXP_NOFACE)? 0 : (minvox >= -128 && maxvox <= 127 ? 1 : + (minvox >= -32768 && maxvox <= 32767 ? 2 : 4)); + vp_s = !model->numvoxtype || !model->voxtype || (flags & M3D_EXP_NOFACE)? 0 : (model->numvoxtype < 254 ? 1 : + (model->numvoxtype < 65534 ? 2 : 4)); + h->types = (vc_s == 8 ? (3<<0) : (vc_s == 2 ? (1<<0) : (vc_s == 1 ? (0<<0) : (2<<0)))) | + (vi_s == 2 ? (1<<2) : (vi_s == 1 ? (0<<2) : (2<<2))) | + (si_s == 2 ? (1<<4) : (si_s == 1 ? (0<<4) : (2<<4))) | + (ci_s == 2 ? (1<<6) : (ci_s == 1 ? (0<<6) : (ci_s == 4 ? (2<<6) : (3<<6)))) | + (ti_s == 2 ? (1<<8) : (ti_s == 1 ? (0<<8) : (ti_s == 4 ? (2<<8) : (3<<8)))) | + (bi_s == 2 ? (1<<10): (bi_s == 1 ? (0<<10): (bi_s == 4 ? (2<<10) : (3<<10)))) | + (nb_s == 2 ? (1<<12): (nb_s == 1 ? (0<<12): (2<<12))) | + (sk_s == 2 ? (1<<14): (sk_s == 1 ? (0<<14): (sk_s == 4 ? (2<<14) : (3<<14)))) | + (fc_s == 2 ? (1<<16): (fc_s == 1 ? (0<<16): (2<<16))) | + (hi_s == 2 ? (1<<18): (hi_s == 1 ? (0<<18): (hi_s == 4 ? (2<<18) : (3<<18)))) | + (fi_s == 2 ? (1<<20): (fi_s == 1 ? (0<<20): (fi_s == 4 ? (2<<20) : (3<<20)))) | + (vd_s == 2 ? (1<<22): (vd_s == 1 ? (0<<22): (vd_s == 4 ? (2<<22) : (3<<22)))) | + (vp_s == 2 ? (1<<24): (vp_s == 1 ? (0<<24): (vp_s == 4 ? (2<<24) : (3<<24)))); + len = h->length; + /* color map */ + if(numcmap && cmap && ci_s < 4 && !(flags & M3D_EXP_NOCMAP)) { + chunklen = 8 + numcmap * sizeof(uint32_t); + h = (m3dhdr_t*)M3D_REALLOC(h, len + chunklen); + if(!h) goto memerr; + memcpy((uint8_t*)h + len, "CMAP", 4); + *((uint32_t*)((uint8_t*)h + len + 4)) = chunklen; + memcpy((uint8_t*)h + len + 8, cmap, chunklen - 8); + len += chunklen; + } else numcmap = 0; + /* texture map */ + if(numtmap && tmap && !(flags & M3D_EXP_NOTXTCRD) && !(flags & M3D_EXP_NOFACE)) { + chunklen = 8 + maxtmap * vc_s * 2; + h = (m3dhdr_t*)M3D_REALLOC(h, len + chunklen); + if(!h) goto memerr; + memcpy((uint8_t*)h + len, "TMAP", 4); + length = (uint32_t*)((uint8_t*)h + len + 4); + out = (uint8_t*)h + len + 8; + last = M3D_UNDEF; + for(i = 0; i < numtmap; i++) { + if(tmap[i].newidx == last) continue; + last = tmap[i].newidx; + switch(vc_s) { + case 1: *out++ = (uint8_t)(tmap[i].data.u * 255); *out++ = (uint8_t)(tmap[i].data.v * 255); break; + case 2: + *((uint16_t*)out) = (uint16_t)(tmap[i].data.u * 65535); out += 2; + *((uint16_t*)out) = (uint16_t)(tmap[i].data.v * 65535); out += 2; + break; + case 4: *((float*)out) = tmap[i].data.u; out += 4; *((float*)out) = tmap[i].data.v; out += 4; break; + case 8: *((double*)out) = tmap[i].data.u; out += 8; *((double*)out) = tmap[i].data.v; out += 8; break; + } + } + *length = (uint32_t)((uintptr_t)out - (uintptr_t)((uint8_t*)h + len)); + out = NULL; + len += *length; + } + /* vertex */ + if(numvrtx && vrtx) { + chunklen = 8 + maxvrtx * (ci_s + sk_s + 4 * vc_s); + h = (m3dhdr_t*)M3D_REALLOC(h, len + chunklen); + if(!h) goto memerr; + memcpy((uint8_t*)h + len, "VRTS", 4); + length = (uint32_t*)((uint8_t*)h + len + 4); + out = (uint8_t*)h + len + 8; + last = M3D_UNDEF; + for(i = 0; i < numvrtx; i++) { + if(vrtx[i].newidx == last) continue; + last = vrtx[i].newidx; + switch(vc_s) { + case 1: + *out++ = (int8_t)(vrtx[i].data.x * 127); + *out++ = (int8_t)(vrtx[i].data.y * 127); + *out++ = (int8_t)(vrtx[i].data.z * 127); + *out++ = (int8_t)(vrtx[i].data.w * 127); + break; + case 2: + *((int16_t*)out) = (int16_t)(vrtx[i].data.x * 32767); out += 2; + *((int16_t*)out) = (int16_t)(vrtx[i].data.y * 32767); out += 2; + *((int16_t*)out) = (int16_t)(vrtx[i].data.z * 32767); out += 2; + *((int16_t*)out) = (int16_t)(vrtx[i].data.w * 32767); out += 2; + break; + case 4: + *((float*)out) = vrtx[i].data.x; out += 4; + *((float*)out) = vrtx[i].data.y; out += 4; + *((float*)out) = vrtx[i].data.z; out += 4; + *((float*)out) = vrtx[i].data.w; out += 4; + break; + case 8: + *((double*)out) = vrtx[i].data.x; out += 8; + *((double*)out) = vrtx[i].data.y; out += 8; + *((double*)out) = vrtx[i].data.z; out += 8; + *((double*)out) = vrtx[i].data.w; out += 8; + break; + } + idx = _m3d_cmapidx(cmap, numcmap, vrtx[i].data.color); + switch(ci_s) { + case 1: *out++ = (uint8_t)(idx); break; + case 2: *((uint16_t*)out) = (uint16_t)(idx); out += 2; break; + case 4: *((uint32_t*)out) = vrtx[i].data.color; out += 4; break; + } + out = _m3d_addidx(out, sk_s, vrtx[i].data.skinid); + } + *length = (uint32_t)((uintptr_t)out - (uintptr_t)((uint8_t*)h + len)); + out = NULL; + len += *length; + } + /* bones chunk */ + if(model->numbone && model->bone && !(flags & M3D_EXP_NOBONE)) { + i = 8 + bi_s + sk_s + model->numbone * (bi_s + si_s + 2*vi_s); + chunklen = i + numskin * nb_s * (bi_s + 1); + h = (m3dhdr_t*)M3D_REALLOC(h, len + chunklen); + if(!h) goto memerr; + memcpy((uint8_t*)h + len, "BONE", 4); + length = (uint32_t*)((uint8_t*)h + len + 4); + out = (uint8_t*)h + len + 8; + out = _m3d_addidx(out, bi_s, model->numbone); + out = _m3d_addidx(out, sk_s, maxskin); + for(i = 0; i < model->numbone; i++) { + out = _m3d_addidx(out, bi_s, model->bone[i].parent); + out = _m3d_addidx(out, si_s, _m3d_stridx(str, numstr, model->bone[i].name)); + out = _m3d_addidx(out, vi_s, vrtxidx[model->bone[i].pos]); + out = _m3d_addidx(out, vi_s, vrtxidx[model->bone[i].ori]); + } + if(numskin && skin && sk_s) { + last = M3D_UNDEF; + for(i = 0; i < numskin; i++) { + if(skin[i].newidx == last) continue; + last = skin[i].newidx; + memset(&weights, 0, nb_s); + for(j = k = l = 0, mw = 0.0; j < (uint32_t)nb_s && skin[i].data.boneid[j] != M3D_UNDEF && + skin[i].data.weight[j] > (M3D_FLOAT)0.0; j++) { + if(mw < skin[i].data.weight[j]) { mw = skin[i].data.weight[j]; k = j; } + weights[j] = (uint8_t)(skin[i].data.weight[j] * 255); + if(!weights[j]) { weights[j]++; l--; } + } + weights[k] += l; + switch(nb_s) { + case 1: weights[0] = 255; break; + case 2: memcpy(out, weights, 2); out += 2; break; + case 4: memcpy(out, weights, 4); out += 4; break; + case 8: memcpy(out, weights, 8); out += 8; break; + } + for(j = 0; j < (uint32_t)nb_s && skin[i].data.boneid[j] != M3D_UNDEF && weights[j]; j++) { + out = _m3d_addidx(out, bi_s, skin[i].data.boneid[j]); + *length += bi_s; + } + } + } + *length = (uint32_t)((uintptr_t)out - (uintptr_t)((uint8_t*)h + len)); + out = NULL; + len += *length; + } + /* materials */ + if(model->nummaterial && !(flags & M3D_EXP_NOMATERIAL)) { + for(j = 0; j < model->nummaterial; j++) { + if(mtrlidx[j] == M3D_UNDEF || !model->material[j].numprop || !model->material[j].prop) continue; + m = &model->material[j]; + chunklen = 12 + si_s + m->numprop * 5; + h = (m3dhdr_t*)M3D_REALLOC(h, len + chunklen); + if(!h) goto memerr; + memcpy((uint8_t*)h + len, "MTRL", 4); + length = (uint32_t*)((uint8_t*)h + len + 4); + out = (uint8_t*)h + len + 8; + out = _m3d_addidx(out, si_s, _m3d_stridx(str, numstr, m->name)); + for(i = 0; i < m->numprop; i++) { + if(m->prop[i].type >= 128) { + if(m->prop[i].value.textureid >= model->numtexture || + !model->texture[m->prop[i].value.textureid].name) continue; + k = m3dpf_map; + } else { + for(k = 256, l = 0; l < sizeof(m3d_propertytypes)/sizeof(m3d_propertytypes[0]); l++) + if(m->prop[i].type == m3d_propertytypes[l].id) { k = m3d_propertytypes[l].format; break; } + } + if(k == 256) continue; + *out++ = m->prop[i].type; + switch(k) { + case m3dpf_color: + if(!(flags & M3D_EXP_NOCMAP)) { + idx = _m3d_cmapidx(cmap, numcmap, m->prop[i].value.color); + switch(ci_s) { + case 1: *out++ = (uint8_t)(idx); break; + case 2: *((uint16_t*)out) = (uint16_t)(idx); out += 2; break; + case 4: *((uint32_t*)out) = (uint32_t)(m->prop[i].value.color); out += 4; break; + } + } else out--; + break; + case m3dpf_uint8: *out++ = m->prop[i].value.num; break; + case m3dpf_uint16: *((uint16_t*)out) = m->prop[i].value.num; out += 2; break; + case m3dpf_uint32: *((uint32_t*)out) = m->prop[i].value.num; out += 4; break; + case m3dpf_float: *((float*)out) = m->prop[i].value.fnum; out += 4; break; + + case m3dpf_map: + idx = _m3d_stridx(str, numstr, model->texture[m->prop[i].value.textureid].name); + out = _m3d_addidx(out, si_s, idx); + break; + } + } + *length = (uint32_t)((uintptr_t)out - (uintptr_t)((uint8_t*)h + len)); + len += *length; + out = NULL; + } + } + /* procedural face */ + if(model->numinlined && model->inlined && !(flags & M3D_EXP_NOFACE)) { + /* all inlined assets which are not textures should be procedural surfaces */ + for(j = 0; j < model->numinlined; j++) { + if(!model->inlined[j].name || !model->inlined[j].name[0] || model->inlined[j].length < 4 || + !model->inlined[j].data || (model->inlined[j].data[1] == 'P' && model->inlined[j].data[2] == 'N' && + model->inlined[j].data[3] == 'G')) + continue; + for(i = k = 0; i < model->numtexture; i++) { + if(!strcmp(model->inlined[j].name, model->texture[i].name)) { k = 1; break; } + } + if(k) continue; + numproc++; + chunklen = 8 + si_s; + h = (m3dhdr_t*)M3D_REALLOC(h, len + chunklen); + if(!h) goto memerr; + memcpy((uint8_t*)h + len, "PROC", 4); + *((uint32_t*)((uint8_t*)h + len + 4)) = chunklen; + out = (uint8_t*)h + len + 8; + out = _m3d_addidx(out, si_s, _m3d_stridx(str, numstr, model->inlined[j].name)); + out = NULL; + len += chunklen; + } + } + /* mesh face */ + if(model->numface && face && !(flags & M3D_EXP_NOFACE)) { + chunklen = 8 + si_s + model->numface * (9 * vi_s + 3 * ti_s + si_s + 1); + h = (m3dhdr_t*)M3D_REALLOC(h, len + chunklen); + if(!h) goto memerr; + memcpy((uint8_t*)h + len, "MESH", 4); + length = (uint32_t*)((uint8_t*)h + len + 4); + out = (uint8_t*)h + len + 8; + last = M3D_UNDEF; +#ifdef M3D_VERTEXMAX + lastp = M3D_UNDEF; +#endif + for(i = 0; i < model->numface; i++) { + if(!(flags & M3D_EXP_NOMATERIAL) && face[i].data.materialid != last) { + last = face[i].data.materialid; + idx = last < model->nummaterial ? _m3d_stridx(str, numstr, model->material[last].name) : 0; + *out++ = 0; + out = _m3d_addidx(out, si_s, idx); + } +#ifdef M3D_VERTEXMAX + if(!(flags & M3D_EXP_NOVRTMAX) && face[i].data.paramid != lastp) { + lastp = face[i].data.paramid; + idx = lastp < model->numparam ? _m3d_stridx(str, numstr, model->param[lastp].name) : 0; + *out++ = 0; + out = _m3d_addidx(out, si_s, idx); + } +#endif + /* hardcoded triangles. */ + k = (3 << 4) | + (((flags & M3D_EXP_NOTXTCRD) || !ti_s || face[i].data.texcoord[0] == M3D_UNDEF || + face[i].data.texcoord[1] == M3D_UNDEF || face[i].data.texcoord[2] == M3D_UNDEF) ? 0 : 1) | + (((flags & M3D_EXP_NONORMAL) || face[i].data.normal[0] == M3D_UNDEF || + face[i].data.normal[1] == M3D_UNDEF || face[i].data.normal[2] == M3D_UNDEF) ? 0 : 2) +#ifdef M3D_VERTEXMAX + | (((flags & M3D_EXP_NOVRTMAX) || face[i].data.vertmax[0] == M3D_UNDEF || + face[i].data.vertmax[1] == M3D_UNDEF || face[i].data.vertmax[2] == M3D_UNDEF) ? 0 : 4) +#endif + ; + *out++ = k; + for(j = 0; j < 3; j++) { + out = _m3d_addidx(out, vi_s, vrtxidx[face[i].data.vertex[j]]); + if(k & 1) + out = _m3d_addidx(out, ti_s, tmapidx[face[i].data.texcoord[j]]); + if(k & 2) + out = _m3d_addidx(out, vi_s, vrtxidx[face[i].data.normal[j]]); +#ifdef M3D_VERTEXMAX + if(k & 4) + out = _m3d_addidx(out, vi_s, vrtxidx[face[i].data.vertmax[j]]); +#endif + } + } + *length = (uint32_t)((uintptr_t)out - (uintptr_t)((uint8_t*)h + len)); + len += *length; + out = NULL; + } + /* voxel face */ + if(model->numvoxtype && model->voxtype && !(flags & M3D_EXP_NOFACE)) { + chunklen = 8 + si_s + model->numvoxtype * (ci_s + si_s + 3 + sk_s); + for(i = 0; i < model->numvoxtype; i++) + chunklen += model->voxtype[i].numitem * (2 + si_s); + h = (m3dhdr_t*)M3D_REALLOC(h, len + chunklen); + if(!h) goto memerr; + memcpy((uint8_t*)h + len, "VOXT", 4); + length = (uint32_t*)((uint8_t*)h + len + 4); + out = (uint8_t*)h + len + 8; + for(i = 0; i < model->numvoxtype; i++) { + if(!(flags & M3D_EXP_NOCMAP)) { + idx = _m3d_cmapidx(cmap, numcmap, model->voxtype[i].color); + switch(ci_s) { + case 1: *out++ = (uint8_t)(idx); break; + case 2: *((uint16_t*)out) = (uint16_t)(idx); out += 2; break; + case 4: *((uint32_t*)out) = (uint32_t)(model->voxtype[i].color); out += 4; break; + } + } + out = _m3d_addidx(out, si_s, _m3d_stridx(str, numstr, model->voxtype[i].name)); + *out++ = (model->voxtype[i].rotation & 0xBF) | (((model->voxtype[i].voxshape >> 8) & 1) << 6); + *out++ = model->voxtype[i].voxshape; + *out++ = model->voxtype[i].numitem; + if(!(flags & M3D_EXP_NOBONE) && model->numbone && maxskin) + out = _m3d_addidx(out, sk_s, skinidx[model->voxtype[i].skinid]); + for(j = 0; j < model->voxtype[i].numitem; j++) { + out = _m3d_addidx(out, 2, model->voxtype[i].item[j].count); + out = _m3d_addidx(out, si_s, _m3d_stridx(str, numstr, model->voxtype[i].item[j].name)); + } + } + *length = (uint32_t)((uintptr_t)out - (uintptr_t)((uint8_t*)h + len)); + len += *length; + out = NULL; + } + if(model->numvoxel && model->voxel && !(flags & M3D_EXP_NOFACE)) { + for(j = 0; j < model->numvoxel; j++) { + chunklen = 8 + si_s + 6 * vd_s + 2 + model->voxel[j].w * model->voxel[j].h * model->voxel[j].d * 3; + h = (m3dhdr_t*)M3D_REALLOC(h, len + chunklen); + if(!h) goto memerr; + memcpy((uint8_t*)h + len, "VOXD", 4); + length = (uint32_t*)((uint8_t*)h + len + 4); + out = (uint8_t*)h + len + 8; + out = _m3d_addidx(out, si_s, _m3d_stridx(str, numstr, model->voxel[j].name)); + out = _m3d_addidx(out, vd_s, model->voxel[j].x); + out = _m3d_addidx(out, vd_s, model->voxel[j].y); + out = _m3d_addidx(out, vd_s, model->voxel[j].z); + out = _m3d_addidx(out, vd_s, model->voxel[j].w); + out = _m3d_addidx(out, vd_s, model->voxel[j].h); + out = _m3d_addidx(out, vd_s, model->voxel[j].d); + *out++ = model->voxel[j].uncertain; + *out++ = model->voxel[j].groupid; + /* RLE compress voxel data */ + n = model->voxel[j].w * model->voxel[j].h * model->voxel[j].d; + k = o = 0; out[o++] = 0; + for(i = 0; i < n; i++) { + for(l = 1; l < 128 && i + l < n && model->voxel[j].data[i] == model->voxel[j].data[i + l]; l++); + if(l > 1) { + l--; + if(out[k]) { out[k]--; out[o++] = 0x80 | l; } + else out[k] = 0x80 | l; + switch(vp_s) { + case 1: out[o++] = model->voxel[j].data[i]; break; + default: *((uint16_t*)(out + o)) = model->voxel[j].data[i]; o += 2; break; + } + k = o; out[o++] = 0; + i += l; + continue; + } + out[k]++; + switch(vp_s) { + case 1: out[o++] = model->voxel[j].data[i]; break; + default: *((uint16_t*)(out + o)) = model->voxel[j].data[i]; o += 2; break; + } + if(out[k] > 127) { out[k]--; k = o; out[o++] = 0; } + } + if(!(out[k] & 0x80)) { if(out[k]) out[k]--; else o--; } + *length = (uint32_t)((uintptr_t)out + (uintptr_t)o - (uintptr_t)((uint8_t*)h + len)); + len += *length; + out = NULL; + } + } + /* mathematical shapes face */ + if(model->numshape && model->shape && !(flags & M3D_EXP_NOFACE)) { + for(j = 0; j < model->numshape; j++) { + chunklen = 12 + si_s + model->shape[j].numcmd * (M3D_CMDMAXARG + 1) * 4; + h = (m3dhdr_t*)M3D_REALLOC(h, len + chunklen); + if(!h) goto memerr; + memcpy((uint8_t*)h + len, "SHPE", 4); + length = (uint32_t*)((uint8_t*)h + len + 4); + out = (uint8_t*)h + len + 8; + out = _m3d_addidx(out, si_s, _m3d_stridx(str, numstr, model->shape[j].name)); + out = _m3d_addidx(out, bi_s, model->shape[j].group); + for(i = 0; i < model->shape[j].numcmd; i++) { + cmd = &model->shape[j].cmd[i]; + if(cmd->type >= (unsigned int)(sizeof(m3d_commandtypes)/sizeof(m3d_commandtypes[0])) || !cmd->arg) + continue; + cd = &m3d_commandtypes[cmd->type]; + *out++ = (cmd->type & 0x7F) | (cmd->type > 127 ? 0x80 : 0); + if(cmd->type > 127) *out++ = (cmd->type >> 7) & 0xff; + for(k = n = 0, l = cd->p; k < l; k++) { + switch(cd->a[((k - n) % (cd->p - n)) + n]) { + case m3dcp_mi_t: + out = _m3d_addidx(out, si_s, cmd->arg[k] < model->nummaterial ? + _m3d_stridx(str, numstr, model->material[cmd->arg[k]].name) : 0); + break; + case m3dcp_vc_t: + min_x = *((float*)&cmd->arg[k]); + switch(vc_s) { + case 1: *out++ = (int8_t)(min_x * 127); break; + case 2: *((int16_t*)out) = (int16_t)(min_x * 32767); out += 2; break; + case 4: *((float*)out) = min_x; out += 4; break; + case 8: *((double*)out) = min_x; out += 8; break; + } + break; + case m3dcp_hi_t: out = _m3d_addidx(out, hi_s, cmd->arg[k]); break; + case m3dcp_fi_t: out = _m3d_addidx(out, fi_s, cmd->arg[k]); break; + case m3dcp_ti_t: out = _m3d_addidx(out, ti_s, cmd->arg[k]); break; + case m3dcp_qi_t: + case m3dcp_vi_t: out = _m3d_addidx(out, vi_s, cmd->arg[k]); break; + case m3dcp_i1_t: out = _m3d_addidx(out, 1, cmd->arg[k]); break; + case m3dcp_i2_t: out = _m3d_addidx(out, 2, cmd->arg[k]); break; + case m3dcp_i4_t: out = _m3d_addidx(out, 4, cmd->arg[k]); break; + case m3dcp_va_t: out = _m3d_addidx(out, 4, cmd->arg[k]); + n = k + 1; l += (cmd->arg[k] - 1) * (cd->p - k - 1); + break; + } + } + } + *length = (uint32_t)((uintptr_t)out - (uintptr_t)((uint8_t*)h + len)); + len += *length; + out = NULL; + } + } + /* annotation labels */ + if(model->numlabel && model->label) { + for(i = 0, length = NULL; i < model->numlabel; i++) { + if(!i || _m3d_strcmp(sl, model->label[i].lang) || _m3d_strcmp(sn, model->label[i].name)) { + sl = model->label[i].lang; + sn = model->label[i].name; + if(length) { + *length = (uint32_t)((uintptr_t)out - (uintptr_t)((uint8_t*)h + len)); + len += *length; + } + chunklen = 8 + 2 * si_s + ci_s + model->numlabel * (vi_s + si_s); + h = (m3dhdr_t*)M3D_REALLOC(h, len + chunklen); + if(!h) { sn = NULL; sl = NULL; goto memerr; } + memcpy((uint8_t*)h + len, "LBLS", 4); + length = (uint32_t*)((uint8_t*)h + len + 4); + out = (uint8_t*)h + len + 8; + out = _m3d_addidx(out, si_s, _m3d_stridx(str, numstr, model->label[l].name)); + out = _m3d_addidx(out, si_s, _m3d_stridx(str, numstr, model->label[l].lang)); + idx = _m3d_cmapidx(cmap, numcmap, model->label[i].color); + switch(ci_s) { + case 1: *out++ = (uint8_t)(idx); break; + case 2: *((uint16_t*)out) = (uint16_t)(idx); out += 2; break; + case 4: *((uint32_t*)out) = model->label[i].color; out += 4; break; + } + } + out = _m3d_addidx(out, vi_s, vrtxidx[model->label[i].vertexid]); + out = _m3d_addidx(out, si_s, _m3d_stridx(str, numstr, model->label[l].text)); + } + if(length) { + *length = (uint32_t)((uintptr_t)out - (uintptr_t)((uint8_t*)h + len)); + len += *length; + } + out = NULL; + sn = sl = NULL; + } + /* actions */ + if(model->numaction && model->action && model->numbone && model->bone && !(flags & M3D_EXP_NOACTION)) { + for(j = 0; j < model->numaction; j++) { + a = &model->action[j]; + chunklen = 14 + si_s + a->numframe * (4 + fc_s + maxt * (bi_s + 2 * vi_s)); + h = (m3dhdr_t*)M3D_REALLOC(h, len + chunklen); + if(!h) goto memerr; + memcpy((uint8_t*)h + len, "ACTN", 4); + length = (uint32_t*)((uint8_t*)h + len + 4); + out = (uint8_t*)h + len + 8; + out = _m3d_addidx(out, si_s, _m3d_stridx(str, numstr, a->name)); + *((uint16_t*)out) = (uint16_t)(a->numframe); out += 2; + *((uint32_t*)out) = (uint32_t)(a->durationmsec); out += 4; + for(i = 0; i < a->numframe; i++) { + *((uint32_t*)out) = (uint32_t)(a->frame[i].msec); out += 4; + out = _m3d_addidx(out, fc_s, a->frame[i].numtransform); + for(k = 0; k < a->frame[i].numtransform; k++) { + out = _m3d_addidx(out, bi_s, a->frame[i].transform[k].boneid); + out = _m3d_addidx(out, vi_s, vrtxidx[a->frame[i].transform[k].pos]); + out = _m3d_addidx(out, vi_s, vrtxidx[a->frame[i].transform[k].ori]); + } + } + *length = (uint32_t)((uintptr_t)out - (uintptr_t)((uint8_t*)h + len)); + len += *length; + out = NULL; + } + } + /* inlined assets */ + if(model->numinlined && model->inlined && (numproc || (flags & M3D_EXP_INLINE))) { + for(j = 0; j < model->numinlined; j++) { + if(!model->inlined[j].name || !model->inlined[j].name[0] || model->inlined[j].length<4 || !model->inlined[j].data) + continue; + if(!(flags & M3D_EXP_INLINE)) { + if(model->inlined[j].data[1] == 'P' && model->inlined[j].data[2] == 'N' && model->inlined[j].data[3] == 'G') + continue; + for(i = k = 0; i < model->numtexture; i++) { + if(!strcmp(model->inlined[j].name, model->texture[i].name)) { k = 1; break; } + } + if(k) continue; + } + chunklen = 8 + si_s + model->inlined[j].length; + h = (m3dhdr_t*)M3D_REALLOC(h, len + chunklen); + if(!h) goto memerr; + memcpy((uint8_t*)h + len, "ASET", 4); + *((uint32_t*)((uint8_t*)h + len + 4)) = chunklen; + out = (uint8_t*)h + len + 8; + out = _m3d_addidx(out, si_s, _m3d_stridx(str, numstr, model->inlined[j].name)); + memcpy(out, model->inlined[j].data, model->inlined[j].length); + out = NULL; + len += chunklen; + } + } + /* extra chunks */ + if(model->numextra && model->extra && (flags & M3D_EXP_EXTRA)) { + for(j = 0; j < model->numextra; j++) { + if(!model->extra[j] || model->extra[j]->length < 8) + continue; + chunklen = model->extra[j]->length; + h = (m3dhdr_t*)M3D_REALLOC(h, len + chunklen); + if(!h) goto memerr; + memcpy((uint8_t*)h + len, model->extra[j], chunklen); + len += chunklen; + } + } + /* add end chunk */ + h = (m3dhdr_t*)M3D_REALLOC(h, len + 4); + if(!h) goto memerr; + memcpy((uint8_t*)h + len, "OMD3", 4); + len += 4; + /* zlib compress */ + if(!(flags & M3D_EXP_NOZLIB)) { + M3D_LOG("Deflating chunks"); + z = stbi_zlib_compress((unsigned char *)h, len, (int*)&l, 9); + if(z && l > 0 && l < len) { len = l; M3D_FREE(h); h = (m3dhdr_t*)z; } + } + /* add file header at the begining */ + len += 8; + out = (unsigned char*)M3D_MALLOC(len); + if(!out) goto memerr; + memcpy(out, "3DMO", 4); + *((uint32_t*)(out + 4)) = len; + /* preview image chunk, must be the first if exists */ + if(model->preview.data && model->preview.length) { + chunklen = 8 + model->preview.length; + out = (unsigned char*)M3D_REALLOC(out, len + chunklen); + if(!out) goto memerr; + memcpy((uint8_t*)out + 8, "PRVW", 4); + *((uint32_t*)((uint8_t*)out + 8 + 4)) = chunklen; + memcpy((uint8_t*)out + 8 + 8, model->preview.data, model->preview.length); + *((uint32_t*)(out + 4)) += chunklen; + } else + chunklen = 0; + memcpy(out + 8 + chunklen, h, len - 8); + } + if(size) *size = out ? len : 0; + if(vrtxidx) M3D_FREE(vrtxidx); + if(mtrlidx) M3D_FREE(mtrlidx); + if(tmapidx) M3D_FREE(tmapidx); + if(skinidx) M3D_FREE(skinidx); + if(norm) M3D_FREE(norm); + if(face) M3D_FREE(face); + if(cmap) M3D_FREE(cmap); + if(tmap) M3D_FREE(tmap); + if(skin) M3D_FREE(skin); + if(str) M3D_FREE(str); + if(vrtx) M3D_FREE(vrtx); + if(opa) M3D_FREE(opa); + if(h) M3D_FREE(h); + return out; +} +#endif + +#endif + +#ifdef __cplusplus +} +#ifdef M3D_CPPWRAPPER +#include +#include +#include + +/*** C++ wrapper class ***/ +namespace M3D { +#ifdef M3D_IMPLEMENTATION + + class Model { + public: + m3d_t *model; + + public: + Model() { + this->model = (m3d_t*)M3D_MALLOC(sizeof(m3d_t)); memset(this->model, 0, sizeof(m3d_t)); + } + Model(_unused const std::string &data, _unused m3dread_t ReadFileCB, + _unused m3dfree_t FreeCB, _unused M3D::Model mtllib) { +#ifndef M3D_NOIMPORTER + this->model = m3d_load((unsigned char *)data.data(), ReadFileCB, FreeCB, mtllib.model); +#else + Model(); +#endif + } + Model(_unused const std::vector data, _unused m3dread_t ReadFileCB, + _unused m3dfree_t FreeCB, _unused M3D::Model mtllib) { +#ifndef M3D_NOIMPORTER + this->model = m3d_load((unsigned char *)&data[0], ReadFileCB, FreeCB, mtllib.model); +#else + Model(); +#endif + } + Model(_unused const unsigned char *data, _unused m3dread_t ReadFileCB, + _unused m3dfree_t FreeCB, _unused M3D::Model mtllib) { +#ifndef M3D_NOIMPORTER + this->model = m3d_load((unsigned char*)data, ReadFileCB, FreeCB, mtllib.model); +#else + Model(); +#endif + } + ~Model() { m3d_free(this->model); } + + public: + m3d_t *getCStruct() { return this->model; } + std::string getName() { return std::string(this->model->name); } + void setName(std::string name) { this->model->name = (char*)name.c_str(); } + std::string getLicense() { return std::string(this->model->license); } + void setLicense(std::string license) { this->model->license = (char*)license.c_str(); } + std::string getAuthor() { return std::string(this->model->author); } + void setAuthor(std::string author) { this->model->author = (char*)author.c_str(); } + std::string getDescription() { return std::string(this->model->desc); } + void setDescription(std::string desc) { this->model->desc = (char*)desc.c_str(); } + float getScale() { return this->model->scale; } + void setScale(float scale) { this->model->scale = scale; } + std::vector getPreview() { return this->model->preview.data ? + std::vector(this->model->preview.data, this->model->preview.data + this->model->preview.length) : + std::vector(); } + std::vector getColorMap() { return this->model->cmap ? std::vector(this->model->cmap, + this->model->cmap + this->model->numcmap) : std::vector(); } + std::vector getTextureMap() { return this->model->tmap ? std::vector(this->model->tmap, + this->model->tmap + this->model->numtmap) : std::vector(); } + std::vector getTextures() { return this->model->texture ? std::vector(this->model->texture, + this->model->texture + this->model->numtexture) : std::vector(); } + std::string getTextureName(int idx) { return idx >= 0 && (unsigned int)idx < this->model->numtexture ? + std::string(this->model->texture[idx].name) : nullptr; } + std::vector getBones() { return this->model->bone ? std::vector(this->model->bone, this->model->bone + + this->model->numbone) : std::vector(); } + std::string getBoneName(int idx) { return idx >= 0 && (unsigned int)idx < this->model->numbone ? + std::string(this->model->bone[idx].name) : nullptr; } + std::vector getMaterials() { return this->model->material ? std::vector(this->model->material, + this->model->material + this->model->nummaterial) : std::vector(); } + std::string getMaterialName(int idx) { return idx >= 0 && (unsigned int)idx < this->model->nummaterial ? + std::string(this->model->material[idx].name) : nullptr; } + int getMaterialPropertyInt(int idx, int type) { + if (idx < 0 || (unsigned int)idx >= this->model->nummaterial || type < 0 || type >= 127 || + !this->model->material[idx].prop) return -1; + for (int i = 0; i < this->model->material[idx].numprop; i++) { + if (this->model->material[idx].prop[i].type == type) + return this->model->material[idx].prop[i].value.num; + } + return -1; + } + uint32_t getMaterialPropertyColor(int idx, int type) { return this->getMaterialPropertyInt(idx, type); } + float getMaterialPropertyFloat(int idx, int type) { + if (idx < 0 || (unsigned int)idx >= this->model->nummaterial || type < 0 || type >= 127 || + !this->model->material[idx].prop) return -1.0f; + for (int i = 0; i < this->model->material[idx].numprop; i++) { + if (this->model->material[idx].prop[i].type == type) + return this->model->material[idx].prop[i].value.fnum; + } + return -1.0f; + } + m3dtx_t* getMaterialPropertyMap(int idx, int type) { + if (idx < 0 || (unsigned int)idx >= this->model->nummaterial || type < 128 || type > 255 || + !this->model->material[idx].prop) return nullptr; + for (int i = 0; i < this->model->material[idx].numprop; i++) { + if (this->model->material[idx].prop[i].type == type) + return this->model->material[idx].prop[i].value.textureid < this->model->numtexture ? + &this->model->texture[this->model->material[idx].prop[i].value.textureid] : nullptr; + } + return nullptr; + } + std::vector getVertices() { return this->model->vertex ? std::vector(this->model->vertex, + this->model->vertex + this->model->numvertex) : std::vector(); } + std::vector getFace() { return this->model->face ? std::vector(this->model->face, this->model->face + + this->model->numface) : std::vector(); } + std::vector getVoxelTypes() { return this->model->voxtype ? std::vector(this->model->voxtype, + this->model->voxtype + this->model->numvoxtype) : std::vector(); } + std::string getVoxelTypeName(int idx) { return idx >= 0 && (unsigned int)idx < this->model->numvoxtype && + this->model->voxtype[idx].name && this->model->voxtype[idx].name[0] ? + std::string(this->model->voxtype[idx].name) : nullptr; } + std::vector getVoxelTypeItems(int idx) { return idx >= 0 && (unsigned int)idx < this->model->numvoxtype && + this->model->voxtype[idx].item ? std::vector(this->model->voxtype[idx].item, + this->model->voxtype[idx].item + this->model->voxtype[idx].numitem) : std::vector(); } + std::vector getVoxelBlocks() { return this->model->voxel ? std::vector(this->model->voxel, + this->model->voxel + this->model->numvoxel) : std::vector(); } + std::string getVoxelBlockName(int idx) { return idx >= 0 && (unsigned int)idx < this->model->numvoxel && + this->model->voxel[idx].name && this->model->voxel[idx].name[0] ? + std::string(this->model->voxel[idx].name) : nullptr; } + std::vector getVoxelBlockData(int idx) { return idx >= 0 && (unsigned int)idx < this->model->numvoxel && + this->model->voxel[idx].data ? std::vector(this->model->voxel[idx].data, + this->model->voxel[idx].data + this->model->voxel[idx].w*this->model->voxel[idx].h*this->model->voxel[idx].d) : + std::vector(); } + std::vector getShape() { return this->model->shape ? std::vector(this->model->shape, + this->model->shape + this->model->numshape) : std::vector(); } + std::string getShapeName(int idx) { return idx >= 0 && (unsigned int)idx < this->model->numshape && + this->model->shape[idx].name && this->model->shape[idx].name[0] ? + std::string(this->model->shape[idx].name) : nullptr; } + unsigned int getShapeGroup(int idx) { return idx >= 0 && (unsigned int)idx < this->model->numshape ? + this->model->shape[idx].group : 0xFFFFFFFF; } + std::vector getShapeCommands(int idx) { return idx >= 0 && (unsigned int)idx < this->model->numshape && + this->model->shape[idx].cmd ? std::vector(this->model->shape[idx].cmd, this->model->shape[idx].cmd + + this->model->shape[idx].numcmd) : std::vector(); } + std::vector getAnnotationLabels() { return this->model->label ? std::vector(this->model->label, + this->model->label + this->model->numlabel) : std::vector(); } + std::vector getSkin() { return this->model->skin ? std::vector(this->model->skin, this->model->skin + + this->model->numskin) : std::vector(); } + std::vector getActions() { return this->model->action ? std::vector(this->model->action, + this->model->action + this->model->numaction) : std::vector(); } + std::string getActionName(int aidx) { return aidx >= 0 && (unsigned int)aidx < this->model->numaction ? + std::string(this->model->action[aidx].name) : nullptr; } + unsigned int getActionDuration(int aidx) { return aidx >= 0 && (unsigned int)aidx < this->model->numaction ? + this->model->action[aidx].durationmsec : 0; } + std::vector getActionFrames(int aidx) { return aidx >= 0 && (unsigned int)aidx < this->model->numaction ? + std::vector(this->model->action[aidx].frame, this->model->action[aidx].frame + + this->model->action[aidx].numframe) : std::vector(); } + unsigned int getActionFrameTimestamp(int aidx, int fidx) { return aidx >= 0 && (unsigned int)aidx < this->model->numaction? + (fidx >= 0 && (unsigned int)fidx < this->model->action[aidx].numframe ? + this->model->action[aidx].frame[fidx].msec : 0) : 0; } + std::vector getActionFrameTransforms(int aidx, int fidx) { + return aidx >= 0 && (unsigned int)aidx < this->model->numaction ? ( + fidx >= 0 && (unsigned int)fidx < this->model->action[aidx].numframe ? + std::vector(this->model->action[aidx].frame[fidx].transform, + this->model->action[aidx].frame[fidx].transform + this->model->action[aidx].frame[fidx].numtransform) : + std::vector()) : std::vector(); } + std::vector getActionFrame(int aidx, int fidx, std::vector skeleton) { + m3dtr_t *pose = m3d_frame(this->model, (unsigned int)aidx, (unsigned int)fidx, + skeleton.size() ? &skeleton[0] : nullptr); + return std::vector(pose, pose + this->model->numbone); } + std::vector getActionPose(int aidx, unsigned int msec) { + m3db_t *pose = m3d_pose(this->model, (unsigned int)aidx, (unsigned int)msec); + return std::vector(pose, pose + this->model->numbone); } + std::vector getInlinedAssets() { return this->model->inlined ? std::vector(this->model->inlined, + this->model->inlined + this->model->numinlined) : std::vector(); } + std::vector> getExtras() { return this->model->extra ? + std::vector>(this->model->extra, + this->model->extra + this->model->numextra) : std::vector>(); } + std::vector Save(_unused int quality, _unused int flags) { +#ifdef M3D_EXPORTER + unsigned int size; + unsigned char *ptr = m3d_save(this->model, quality, flags, &size); + return ptr && size ? std::vector(ptr, ptr + size) : std::vector(); +#else + return std::vector(); +#endif + } + }; + +#else + class Model { + private: + m3d_t *model; + + public: + Model(const std::string &data, m3dread_t ReadFileCB, m3dfree_t FreeCB); + Model(const std::vector data, m3dread_t ReadFileCB, m3dfree_t FreeCB); + Model(const unsigned char *data, m3dread_t ReadFileCB, m3dfree_t FreeCB); + Model(); + ~Model(); + + public: + m3d_t *getCStruct(); + std::string getName(); + void setName(std::string name); + std::string getLicense(); + void setLicense(std::string license); + std::string getAuthor(); + void setAuthor(std::string author); + std::string getDescription(); + void setDescription(std::string desc); + float getScale(); + void setScale(float scale); + std::vector getPreview(); + std::vector getColorMap(); + std::vector getTextureMap(); + std::vector getTextures(); + std::string getTextureName(int idx); + std::vector getBones(); + std::string getBoneName(int idx); + std::vector getMaterials(); + std::string getMaterialName(int idx); + int getMaterialPropertyInt(int idx, int type); + uint32_t getMaterialPropertyColor(int idx, int type); + float getMaterialPropertyFloat(int idx, int type); + m3dtx_t* getMaterialPropertyMap(int idx, int type); + std::vector getVertices(); + std::vector getFace(); + std::vector getVoxelTypes(); + std::string getVoxelTypeName(int idx); + std::vector getVoxelTypeItems(int idx); + std::vector getVoxelBlocks(); + std::string getVoxelBlockName(int idx); + std::vector getVoxelBlockData(int idx); + std::vector getShape(); + std::string getShapeName(int idx); + unsigned int getShapeGroup(int idx); + std::vector getShapeCommands(int idx); + std::vector getAnnotationLabels(); + std::vector getSkin(); + std::vector getActions(); + std::string getActionName(int aidx); + unsigned int getActionDuration(int aidx); + std::vector getActionFrames(int aidx); + unsigned int getActionFrameTimestamp(int aidx, int fidx); + std::vector getActionFrameTransforms(int aidx, int fidx); + std::vector getActionFrame(int aidx, int fidx, std::vector skeleton); + std::vector getActionPose(int aidx, unsigned int msec); + std::vector getInlinedAssets(); + std::vector> getExtras(); + std::vector Save(int quality, int flags); + }; + +#endif /* impl */ +} +#endif + +#endif /* __cplusplus */ + +#endif diff --git a/external/stb/stb_image.c b/external/stb/stb_image.c new file mode 100644 index 0000000..5b04e48 --- /dev/null +++ b/external/stb/stb_image.c @@ -0,0 +1,3 @@ +#define STB_IMAGE_IMPLEMENTATION + +#include "stb_image.h" diff --git a/external/stb/stb_image.h b/external/stb/stb_image.h new file mode 100644 index 0000000..9eedabe --- /dev/null +++ b/external/stb/stb_image.h @@ -0,0 +1,7988 @@ +/* stb_image - v2.30 - public domain image loader - http://nothings.org/stb + no warranty implied; use at your own risk + + Do this: + #define STB_IMAGE_IMPLEMENTATION + before you include this file in *one* C or C++ file to create the implementation. + + // i.e. it should look like this: + #include ... + #include ... + #include ... + #define STB_IMAGE_IMPLEMENTATION + #include "stb_image.h" + + You can #define STBI_ASSERT(x) before the #include to avoid using assert.h. + And #define STBI_MALLOC, STBI_REALLOC, and STBI_FREE to avoid using malloc,realloc,free + + + QUICK NOTES: + Primarily of interest to game developers and other people who can + avoid problematic images and only need the trivial interface + + JPEG baseline & progressive (12 bpc/arithmetic not supported, same as stock IJG lib) + PNG 1/2/4/8/16-bit-per-channel + + TGA (not sure what subset, if a subset) + BMP non-1bpp, non-RLE + PSD (composited view only, no extra channels, 8/16 bit-per-channel) + + GIF (*comp always reports as 4-channel) + HDR (radiance rgbE format) + PIC (Softimage PIC) + PNM (PPM and PGM binary only) + + Animated GIF still needs a proper API, but here's one way to do it: + http://gist.github.com/urraka/685d9a6340b26b830d49 + + - decode from memory or through FILE (define STBI_NO_STDIO to remove code) + - decode from arbitrary I/O callbacks + - SIMD acceleration on x86/x64 (SSE2) and ARM (NEON) + + Full documentation under "DOCUMENTATION" below. + + +LICENSE + + See end of file for license information. + +RECENT REVISION HISTORY: + + 2.30 (2024-05-31) avoid erroneous gcc warning + 2.29 (2023-05-xx) optimizations + 2.28 (2023-01-29) many error fixes, security errors, just tons of stuff + 2.27 (2021-07-11) document stbi_info better, 16-bit PNM support, bug fixes + 2.26 (2020-07-13) many minor fixes + 2.25 (2020-02-02) fix warnings + 2.24 (2020-02-02) fix warnings; thread-local failure_reason and flip_vertically + 2.23 (2019-08-11) fix clang static analysis warning + 2.22 (2019-03-04) gif fixes, fix warnings + 2.21 (2019-02-25) fix typo in comment + 2.20 (2019-02-07) support utf8 filenames in Windows; fix warnings and platform ifdefs + 2.19 (2018-02-11) fix warning + 2.18 (2018-01-30) fix warnings + 2.17 (2018-01-29) bugfix, 1-bit BMP, 16-bitness query, fix warnings + 2.16 (2017-07-23) all functions have 16-bit variants; optimizations; bugfixes + 2.15 (2017-03-18) fix png-1,2,4; all Imagenet JPGs; no runtime SSE detection on GCC + 2.14 (2017-03-03) remove deprecated STBI_JPEG_OLD; fixes for Imagenet JPGs + 2.13 (2016-12-04) experimental 16-bit API, only for PNG so far; fixes + 2.12 (2016-04-02) fix typo in 2.11 PSD fix that caused crashes + 2.11 (2016-04-02) 16-bit PNGS; enable SSE2 in non-gcc x64 + RGB-format JPEG; remove white matting in PSD; + allocate large structures on the stack; + correct channel count for PNG & BMP + 2.10 (2016-01-22) avoid warning introduced in 2.09 + 2.09 (2016-01-16) 16-bit TGA; comments in PNM files; STBI_REALLOC_SIZED + + See end of file for full revision history. + + + ============================ Contributors ========================= + + Image formats Extensions, features + Sean Barrett (jpeg, png, bmp) Jetro Lauha (stbi_info) + Nicolas Schulz (hdr, psd) Martin "SpartanJ" Golini (stbi_info) + Jonathan Dummer (tga) James "moose2000" Brown (iPhone PNG) + Jean-Marc Lienher (gif) Ben "Disch" Wenger (io callbacks) + Tom Seddon (pic) Omar Cornut (1/2/4-bit PNG) + Thatcher Ulrich (psd) Nicolas Guillemot (vertical flip) + Ken Miller (pgm, ppm) Richard Mitton (16-bit PSD) + github:urraka (animated gif) Junggon Kim (PNM comments) + Christopher Forseth (animated gif) Daniel Gibson (16-bit TGA) + socks-the-fox (16-bit PNG) + Jeremy Sawicki (handle all ImageNet JPGs) + Optimizations & bugfixes Mikhail Morozov (1-bit BMP) + Fabian "ryg" Giesen Anael Seghezzi (is-16-bit query) + Arseny Kapoulkine Simon Breuss (16-bit PNM) + John-Mark Allen + Carmelo J Fdez-Aguera + + Bug & warning fixes + Marc LeBlanc David Woo Guillaume George Martins Mozeiko + Christpher Lloyd Jerry Jansson Joseph Thomson Blazej Dariusz Roszkowski + Phil Jordan Dave Moore Roy Eltham + Hayaki Saito Nathan Reed Won Chun + Luke Graham Johan Duparc Nick Verigakis the Horde3D community + Thomas Ruf Ronny Chevalier github:rlyeh + Janez Zemva John Bartholomew Michal Cichon github:romigrou + Jonathan Blow Ken Hamada Tero Hanninen github:svdijk + Eugene Golushkov Laurent Gomila Cort Stratton github:snagar + Aruelien Pocheville Sergio Gonzalez Thibault Reuille github:Zelex + Cass Everitt Ryamond Barbiero github:grim210 + Paul Du Bois Engin Manap Aldo Culquicondor github:sammyhw + Philipp Wiesemann Dale Weiler Oriol Ferrer Mesia github:phprus + Josh Tobin Neil Bickford Matthew Gregan github:poppolopoppo + Julian Raschke Gregory Mullen Christian Floisand github:darealshinji + Baldur Karlsson Kevin Schmidt JR Smith github:Michaelangel007 + Brad Weinberger Matvey Cherevko github:mosra + Luca Sas Alexander Veselov Zack Middleton [reserved] + Ryan C. Gordon [reserved] [reserved] + DO NOT ADD YOUR NAME HERE + + Jacko Dirks + + To add your name to the credits, pick a random blank space in the middle and fill it. + 80% of merge conflicts on stb PRs are due to people adding their name at the end + of the credits. +*/ + +#ifndef STBI_INCLUDE_STB_IMAGE_H +#define STBI_INCLUDE_STB_IMAGE_H + +// DOCUMENTATION +// +// Limitations: +// - no 12-bit-per-channel JPEG +// - no JPEGs with arithmetic coding +// - GIF always returns *comp=4 +// +// Basic usage (see HDR discussion below for HDR usage): +// int x,y,n; +// unsigned char *data = stbi_load(filename, &x, &y, &n, 0); +// // ... process data if not NULL ... +// // ... x = width, y = height, n = # 8-bit components per pixel ... +// // ... replace '0' with '1'..'4' to force that many components per pixel +// // ... but 'n' will always be the number that it would have been if you said 0 +// stbi_image_free(data); +// +// Standard parameters: +// int *x -- outputs image width in pixels +// int *y -- outputs image height in pixels +// int *channels_in_file -- outputs # of image components in image file +// int desired_channels -- if non-zero, # of image components requested in result +// +// The return value from an image loader is an 'unsigned char *' which points +// to the pixel data, or NULL on an allocation failure or if the image is +// corrupt or invalid. The pixel data consists of *y scanlines of *x pixels, +// with each pixel consisting of N interleaved 8-bit components; the first +// pixel pointed to is top-left-most in the image. There is no padding between +// image scanlines or between pixels, regardless of format. The number of +// components N is 'desired_channels' if desired_channels is non-zero, or +// *channels_in_file otherwise. If desired_channels is non-zero, +// *channels_in_file has the number of components that _would_ have been +// output otherwise. E.g. if you set desired_channels to 4, you will always +// get RGBA output, but you can check *channels_in_file to see if it's trivially +// opaque because e.g. there were only 3 channels in the source image. +// +// An output image with N components has the following components interleaved +// in this order in each pixel: +// +// N=#comp components +// 1 grey +// 2 grey, alpha +// 3 red, green, blue +// 4 red, green, blue, alpha +// +// If image loading fails for any reason, the return value will be NULL, +// and *x, *y, *channels_in_file will be unchanged. The function +// stbi_failure_reason() can be queried for an extremely brief, end-user +// unfriendly explanation of why the load failed. Define STBI_NO_FAILURE_STRINGS +// to avoid compiling these strings at all, and STBI_FAILURE_USERMSG to get slightly +// more user-friendly ones. +// +// Paletted PNG, BMP, GIF, and PIC images are automatically depalettized. +// +// To query the width, height and component count of an image without having to +// decode the full file, you can use the stbi_info family of functions: +// +// int x,y,n,ok; +// ok = stbi_info(filename, &x, &y, &n); +// // returns ok=1 and sets x, y, n if image is a supported format, +// // 0 otherwise. +// +// Note that stb_image pervasively uses ints in its public API for sizes, +// including sizes of memory buffers. This is now part of the API and thus +// hard to change without causing breakage. As a result, the various image +// loaders all have certain limits on image size; these differ somewhat +// by format but generally boil down to either just under 2GB or just under +// 1GB. When the decoded image would be larger than this, stb_image decoding +// will fail. +// +// Additionally, stb_image will reject image files that have any of their +// dimensions set to a larger value than the configurable STBI_MAX_DIMENSIONS, +// which defaults to 2**24 = 16777216 pixels. Due to the above memory limit, +// the only way to have an image with such dimensions load correctly +// is for it to have a rather extreme aspect ratio. Either way, the +// assumption here is that such larger images are likely to be malformed +// or malicious. If you do need to load an image with individual dimensions +// larger than that, and it still fits in the overall size limit, you can +// #define STBI_MAX_DIMENSIONS on your own to be something larger. +// +// =========================================================================== +// +// UNICODE: +// +// If compiling for Windows and you wish to use Unicode filenames, compile +// with +// #define STBI_WINDOWS_UTF8 +// and pass utf8-encoded filenames. Call stbi_convert_wchar_to_utf8 to convert +// Windows wchar_t filenames to utf8. +// +// =========================================================================== +// +// Philosophy +// +// stb libraries are designed with the following priorities: +// +// 1. easy to use +// 2. easy to maintain +// 3. good performance +// +// Sometimes I let "good performance" creep up in priority over "easy to maintain", +// and for best performance I may provide less-easy-to-use APIs that give higher +// performance, in addition to the easy-to-use ones. Nevertheless, it's important +// to keep in mind that from the standpoint of you, a client of this library, +// all you care about is #1 and #3, and stb libraries DO NOT emphasize #3 above all. +// +// Some secondary priorities arise directly from the first two, some of which +// provide more explicit reasons why performance can't be emphasized. +// +// - Portable ("ease of use") +// - Small source code footprint ("easy to maintain") +// - No dependencies ("ease of use") +// +// =========================================================================== +// +// I/O callbacks +// +// I/O callbacks allow you to read from arbitrary sources, like packaged +// files or some other source. Data read from callbacks are processed +// through a small internal buffer (currently 128 bytes) to try to reduce +// overhead. +// +// The three functions you must define are "read" (reads some bytes of data), +// "skip" (skips some bytes of data), "eof" (reports if the stream is at the end). +// +// =========================================================================== +// +// SIMD support +// +// The JPEG decoder will try to automatically use SIMD kernels on x86 when +// supported by the compiler. For ARM Neon support, you must explicitly +// request it. +// +// (The old do-it-yourself SIMD API is no longer supported in the current +// code.) +// +// On x86, SSE2 will automatically be used when available based on a run-time +// test; if not, the generic C versions are used as a fall-back. On ARM targets, +// the typical path is to have separate builds for NEON and non-NEON devices +// (at least this is true for iOS and Android). Therefore, the NEON support is +// toggled by a build flag: define STBI_NEON to get NEON loops. +// +// If for some reason you do not want to use any of SIMD code, or if +// you have issues compiling it, you can disable it entirely by +// defining STBI_NO_SIMD. +// +// =========================================================================== +// +// HDR image support (disable by defining STBI_NO_HDR) +// +// stb_image supports loading HDR images in general, and currently the Radiance +// .HDR file format specifically. You can still load any file through the existing +// interface; if you attempt to load an HDR file, it will be automatically remapped +// to LDR, assuming gamma 2.2 and an arbitrary scale factor defaulting to 1; +// both of these constants can be reconfigured through this interface: +// +// stbi_hdr_to_ldr_gamma(2.2f); +// stbi_hdr_to_ldr_scale(1.0f); +// +// (note, do not use _inverse_ constants; stbi_image will invert them +// appropriately). +// +// Additionally, there is a new, parallel interface for loading files as +// (linear) floats to preserve the full dynamic range: +// +// float *data = stbi_loadf(filename, &x, &y, &n, 0); +// +// If you load LDR images through this interface, those images will +// be promoted to floating point values, run through the inverse of +// constants corresponding to the above: +// +// stbi_ldr_to_hdr_scale(1.0f); +// stbi_ldr_to_hdr_gamma(2.2f); +// +// Finally, given a filename (or an open file or memory block--see header +// file for details) containing image data, you can query for the "most +// appropriate" interface to use (that is, whether the image is HDR or +// not), using: +// +// stbi_is_hdr(char *filename); +// +// =========================================================================== +// +// iPhone PNG support: +// +// We optionally support converting iPhone-formatted PNGs (which store +// premultiplied BGRA) back to RGB, even though they're internally encoded +// differently. To enable this conversion, call +// stbi_convert_iphone_png_to_rgb(1). +// +// Call stbi_set_unpremultiply_on_load(1) as well to force a divide per +// pixel to remove any premultiplied alpha *only* if the image file explicitly +// says there's premultiplied data (currently only happens in iPhone images, +// and only if iPhone convert-to-rgb processing is on). +// +// =========================================================================== +// +// ADDITIONAL CONFIGURATION +// +// - You can suppress implementation of any of the decoders to reduce +// your code footprint by #defining one or more of the following +// symbols before creating the implementation. +// +// STBI_NO_JPEG +// STBI_NO_PNG +// STBI_NO_BMP +// STBI_NO_PSD +// STBI_NO_TGA +// STBI_NO_GIF +// STBI_NO_HDR +// STBI_NO_PIC +// STBI_NO_PNM (.ppm and .pgm) +// +// - You can request *only* certain decoders and suppress all other ones +// (this will be more forward-compatible, as addition of new decoders +// doesn't require you to disable them explicitly): +// +// STBI_ONLY_JPEG +// STBI_ONLY_PNG +// STBI_ONLY_BMP +// STBI_ONLY_PSD +// STBI_ONLY_TGA +// STBI_ONLY_GIF +// STBI_ONLY_HDR +// STBI_ONLY_PIC +// STBI_ONLY_PNM (.ppm and .pgm) +// +// - If you use STBI_NO_PNG (or _ONLY_ without PNG), and you still +// want the zlib decoder to be available, #define STBI_SUPPORT_ZLIB +// +// - If you define STBI_MAX_DIMENSIONS, stb_image will reject images greater +// than that size (in either width or height) without further processing. +// This is to let programs in the wild set an upper bound to prevent +// denial-of-service attacks on untrusted data, as one could generate a +// valid image of gigantic dimensions and force stb_image to allocate a +// huge block of memory and spend disproportionate time decoding it. By +// default this is set to (1 << 24), which is 16777216, but that's still +// very big. + +#ifndef STBI_NO_STDIO +#include +#endif // STBI_NO_STDIO + +#define STBI_VERSION 1 + +enum +{ + STBI_default = 0, // only used for desired_channels + + STBI_grey = 1, + STBI_grey_alpha = 2, + STBI_rgb = 3, + STBI_rgb_alpha = 4 +}; + +#include +typedef unsigned char stbi_uc; +typedef unsigned short stbi_us; + +#ifdef __cplusplus +extern "C" { +#endif + +#ifndef STBIDEF +#ifdef STB_IMAGE_STATIC +#define STBIDEF static +#else +#define STBIDEF extern +#endif +#endif + +////////////////////////////////////////////////////////////////////////////// +// +// PRIMARY API - works on images of any type +// + +// +// load image by filename, open file, or memory buffer +// + +typedef struct +{ + int (*read) (void *user,char *data,int size); // fill 'data' with 'size' bytes. return number of bytes actually read + void (*skip) (void *user,int n); // skip the next 'n' bytes, or 'unget' the last -n bytes if negative + int (*eof) (void *user); // returns nonzero if we are at end of file/data +} stbi_io_callbacks; + +//////////////////////////////////// +// +// 8-bits-per-channel interface +// + +STBIDEF stbi_uc *stbi_load_from_memory (stbi_uc const *buffer, int len , int *x, int *y, int *channels_in_file, int desired_channels); +STBIDEF stbi_uc *stbi_load_from_callbacks(stbi_io_callbacks const *clbk , void *user, int *x, int *y, int *channels_in_file, int desired_channels); + +#ifndef STBI_NO_STDIO +STBIDEF stbi_uc *stbi_load (char const *filename, int *x, int *y, int *channels_in_file, int desired_channels); +STBIDEF stbi_uc *stbi_load_from_file (FILE *f, int *x, int *y, int *channels_in_file, int desired_channels); +// for stbi_load_from_file, file pointer is left pointing immediately after image +#endif + +#ifndef STBI_NO_GIF +STBIDEF stbi_uc *stbi_load_gif_from_memory(stbi_uc const *buffer, int len, int **delays, int *x, int *y, int *z, int *comp, int req_comp); +#endif + +#ifdef STBI_WINDOWS_UTF8 +STBIDEF int stbi_convert_wchar_to_utf8(char *buffer, size_t bufferlen, const wchar_t* input); +#endif + +//////////////////////////////////// +// +// 16-bits-per-channel interface +// + +STBIDEF stbi_us *stbi_load_16_from_memory (stbi_uc const *buffer, int len, int *x, int *y, int *channels_in_file, int desired_channels); +STBIDEF stbi_us *stbi_load_16_from_callbacks(stbi_io_callbacks const *clbk, void *user, int *x, int *y, int *channels_in_file, int desired_channels); + +#ifndef STBI_NO_STDIO +STBIDEF stbi_us *stbi_load_16 (char const *filename, int *x, int *y, int *channels_in_file, int desired_channels); +STBIDEF stbi_us *stbi_load_from_file_16(FILE *f, int *x, int *y, int *channels_in_file, int desired_channels); +#endif + +//////////////////////////////////// +// +// float-per-channel interface +// +#ifndef STBI_NO_LINEAR + STBIDEF float *stbi_loadf_from_memory (stbi_uc const *buffer, int len, int *x, int *y, int *channels_in_file, int desired_channels); + STBIDEF float *stbi_loadf_from_callbacks (stbi_io_callbacks const *clbk, void *user, int *x, int *y, int *channels_in_file, int desired_channels); + + #ifndef STBI_NO_STDIO + STBIDEF float *stbi_loadf (char const *filename, int *x, int *y, int *channels_in_file, int desired_channels); + STBIDEF float *stbi_loadf_from_file (FILE *f, int *x, int *y, int *channels_in_file, int desired_channels); + #endif +#endif + +#ifndef STBI_NO_HDR + STBIDEF void stbi_hdr_to_ldr_gamma(float gamma); + STBIDEF void stbi_hdr_to_ldr_scale(float scale); +#endif // STBI_NO_HDR + +#ifndef STBI_NO_LINEAR + STBIDEF void stbi_ldr_to_hdr_gamma(float gamma); + STBIDEF void stbi_ldr_to_hdr_scale(float scale); +#endif // STBI_NO_LINEAR + +// stbi_is_hdr is always defined, but always returns false if STBI_NO_HDR +STBIDEF int stbi_is_hdr_from_callbacks(stbi_io_callbacks const *clbk, void *user); +STBIDEF int stbi_is_hdr_from_memory(stbi_uc const *buffer, int len); +#ifndef STBI_NO_STDIO +STBIDEF int stbi_is_hdr (char const *filename); +STBIDEF int stbi_is_hdr_from_file(FILE *f); +#endif // STBI_NO_STDIO + + +// get a VERY brief reason for failure +// on most compilers (and ALL modern mainstream compilers) this is threadsafe +STBIDEF const char *stbi_failure_reason (void); + +// free the loaded image -- this is just free() +STBIDEF void stbi_image_free (void *retval_from_stbi_load); + +// get image dimensions & components without fully decoding +STBIDEF int stbi_info_from_memory(stbi_uc const *buffer, int len, int *x, int *y, int *comp); +STBIDEF int stbi_info_from_callbacks(stbi_io_callbacks const *clbk, void *user, int *x, int *y, int *comp); +STBIDEF int stbi_is_16_bit_from_memory(stbi_uc const *buffer, int len); +STBIDEF int stbi_is_16_bit_from_callbacks(stbi_io_callbacks const *clbk, void *user); + +#ifndef STBI_NO_STDIO +STBIDEF int stbi_info (char const *filename, int *x, int *y, int *comp); +STBIDEF int stbi_info_from_file (FILE *f, int *x, int *y, int *comp); +STBIDEF int stbi_is_16_bit (char const *filename); +STBIDEF int stbi_is_16_bit_from_file(FILE *f); +#endif + + + +// for image formats that explicitly notate that they have premultiplied alpha, +// we just return the colors as stored in the file. set this flag to force +// unpremultiplication. results are undefined if the unpremultiply overflow. +STBIDEF void stbi_set_unpremultiply_on_load(int flag_true_if_should_unpremultiply); + +// indicate whether we should process iphone images back to canonical format, +// or just pass them through "as-is" +STBIDEF void stbi_convert_iphone_png_to_rgb(int flag_true_if_should_convert); + +// flip the image vertically, so the first pixel in the output array is the bottom left +STBIDEF void stbi_set_flip_vertically_on_load(int flag_true_if_should_flip); + +// as above, but only applies to images loaded on the thread that calls the function +// this function is only available if your compiler supports thread-local variables; +// calling it will fail to link if your compiler doesn't +STBIDEF void stbi_set_unpremultiply_on_load_thread(int flag_true_if_should_unpremultiply); +STBIDEF void stbi_convert_iphone_png_to_rgb_thread(int flag_true_if_should_convert); +STBIDEF void stbi_set_flip_vertically_on_load_thread(int flag_true_if_should_flip); + +// ZLIB client - used by PNG, available for other purposes + +STBIDEF char *stbi_zlib_decode_malloc_guesssize(const char *buffer, int len, int initial_size, int *outlen); +STBIDEF char *stbi_zlib_decode_malloc_guesssize_headerflag(const char *buffer, int len, int initial_size, int *outlen, int parse_header); +STBIDEF char *stbi_zlib_decode_malloc(const char *buffer, int len, int *outlen); +STBIDEF int stbi_zlib_decode_buffer(char *obuffer, int olen, const char *ibuffer, int ilen); + +STBIDEF char *stbi_zlib_decode_noheader_malloc(const char *buffer, int len, int *outlen); +STBIDEF int stbi_zlib_decode_noheader_buffer(char *obuffer, int olen, const char *ibuffer, int ilen); + + +#ifdef __cplusplus +} +#endif + +// +// +//// end header file ///////////////////////////////////////////////////// +#endif // STBI_INCLUDE_STB_IMAGE_H + +#ifdef STB_IMAGE_IMPLEMENTATION + +#if defined(STBI_ONLY_JPEG) || defined(STBI_ONLY_PNG) || defined(STBI_ONLY_BMP) \ + || defined(STBI_ONLY_TGA) || defined(STBI_ONLY_GIF) || defined(STBI_ONLY_PSD) \ + || defined(STBI_ONLY_HDR) || defined(STBI_ONLY_PIC) || defined(STBI_ONLY_PNM) \ + || defined(STBI_ONLY_ZLIB) + #ifndef STBI_ONLY_JPEG + #define STBI_NO_JPEG + #endif + #ifndef STBI_ONLY_PNG + #define STBI_NO_PNG + #endif + #ifndef STBI_ONLY_BMP + #define STBI_NO_BMP + #endif + #ifndef STBI_ONLY_PSD + #define STBI_NO_PSD + #endif + #ifndef STBI_ONLY_TGA + #define STBI_NO_TGA + #endif + #ifndef STBI_ONLY_GIF + #define STBI_NO_GIF + #endif + #ifndef STBI_ONLY_HDR + #define STBI_NO_HDR + #endif + #ifndef STBI_ONLY_PIC + #define STBI_NO_PIC + #endif + #ifndef STBI_ONLY_PNM + #define STBI_NO_PNM + #endif +#endif + +#if defined(STBI_NO_PNG) && !defined(STBI_SUPPORT_ZLIB) && !defined(STBI_NO_ZLIB) +#define STBI_NO_ZLIB +#endif + + +#include +#include // ptrdiff_t on osx +#include +#include +#include + +#if !defined(STBI_NO_LINEAR) || !defined(STBI_NO_HDR) +#include // ldexp, pow +#endif + +#ifndef STBI_NO_STDIO +#include +#endif + +#ifndef STBI_ASSERT +#include +#define STBI_ASSERT(x) assert(x) +#endif + +#ifdef __cplusplus +#define STBI_EXTERN extern "C" +#else +#define STBI_EXTERN extern +#endif + + +#ifndef _MSC_VER + #ifdef __cplusplus + #define stbi_inline inline + #else + #define stbi_inline + #endif +#else + #define stbi_inline __forceinline +#endif + +#ifndef STBI_NO_THREAD_LOCALS + #if defined(__cplusplus) && __cplusplus >= 201103L + #define STBI_THREAD_LOCAL thread_local + #elif defined(__GNUC__) && __GNUC__ < 5 + #define STBI_THREAD_LOCAL __thread + #elif defined(_MSC_VER) + #define STBI_THREAD_LOCAL __declspec(thread) + #elif defined (__STDC_VERSION__) && __STDC_VERSION__ >= 201112L && !defined(__STDC_NO_THREADS__) + #define STBI_THREAD_LOCAL _Thread_local + #endif + + #ifndef STBI_THREAD_LOCAL + #if defined(__GNUC__) + #define STBI_THREAD_LOCAL __thread + #endif + #endif +#endif + +#if defined(_MSC_VER) || defined(__SYMBIAN32__) +typedef unsigned short stbi__uint16; +typedef signed short stbi__int16; +typedef unsigned int stbi__uint32; +typedef signed int stbi__int32; +#else +#include +typedef uint16_t stbi__uint16; +typedef int16_t stbi__int16; +typedef uint32_t stbi__uint32; +typedef int32_t stbi__int32; +#endif + +// should produce compiler error if size is wrong +typedef unsigned char validate_uint32[sizeof(stbi__uint32)==4 ? 1 : -1]; + +#ifdef _MSC_VER +#define STBI_NOTUSED(v) (void)(v) +#else +#define STBI_NOTUSED(v) (void)sizeof(v) +#endif + +#ifdef _MSC_VER +#define STBI_HAS_LROTL +#endif + +#ifdef STBI_HAS_LROTL + #define stbi_lrot(x,y) _lrotl(x,y) +#else + #define stbi_lrot(x,y) (((x) << (y)) | ((x) >> (-(y) & 31))) +#endif + +#if defined(STBI_MALLOC) && defined(STBI_FREE) && (defined(STBI_REALLOC) || defined(STBI_REALLOC_SIZED)) +// ok +#elif !defined(STBI_MALLOC) && !defined(STBI_FREE) && !defined(STBI_REALLOC) && !defined(STBI_REALLOC_SIZED) +// ok +#else +#error "Must define all or none of STBI_MALLOC, STBI_FREE, and STBI_REALLOC (or STBI_REALLOC_SIZED)." +#endif + +#ifndef STBI_MALLOC +#define STBI_MALLOC(sz) malloc(sz) +#define STBI_REALLOC(p,newsz) realloc(p,newsz) +#define STBI_FREE(p) free(p) +#endif + +#ifndef STBI_REALLOC_SIZED +#define STBI_REALLOC_SIZED(p,oldsz,newsz) STBI_REALLOC(p,newsz) +#endif + +// x86/x64 detection +#if defined(__x86_64__) || defined(_M_X64) +#define STBI__X64_TARGET +#elif defined(__i386) || defined(_M_IX86) +#define STBI__X86_TARGET +#endif + +#if defined(__GNUC__) && defined(STBI__X86_TARGET) && !defined(__SSE2__) && !defined(STBI_NO_SIMD) +// gcc doesn't support sse2 intrinsics unless you compile with -msse2, +// which in turn means it gets to use SSE2 everywhere. This is unfortunate, +// but previous attempts to provide the SSE2 functions with runtime +// detection caused numerous issues. The way architecture extensions are +// exposed in GCC/Clang is, sadly, not really suited for one-file libs. +// New behavior: if compiled with -msse2, we use SSE2 without any +// detection; if not, we don't use it at all. +#define STBI_NO_SIMD +#endif + +#if defined(__MINGW32__) && defined(STBI__X86_TARGET) && !defined(STBI_MINGW_ENABLE_SSE2) && !defined(STBI_NO_SIMD) +// Note that __MINGW32__ doesn't actually mean 32-bit, so we have to avoid STBI__X64_TARGET +// +// 32-bit MinGW wants ESP to be 16-byte aligned, but this is not in the +// Windows ABI and VC++ as well as Windows DLLs don't maintain that invariant. +// As a result, enabling SSE2 on 32-bit MinGW is dangerous when not +// simultaneously enabling "-mstackrealign". +// +// See https://github.com/nothings/stb/issues/81 for more information. +// +// So default to no SSE2 on 32-bit MinGW. If you've read this far and added +// -mstackrealign to your build settings, feel free to #define STBI_MINGW_ENABLE_SSE2. +#define STBI_NO_SIMD +#endif + +#if !defined(STBI_NO_SIMD) && (defined(STBI__X86_TARGET) || defined(STBI__X64_TARGET)) +#define STBI_SSE2 +#include + +#ifdef _MSC_VER + +#if _MSC_VER >= 1400 // not VC6 +#include // __cpuid +static int stbi__cpuid3(void) +{ + int info[4]; + __cpuid(info,1); + return info[3]; +} +#else +static int stbi__cpuid3(void) +{ + int res; + __asm { + mov eax,1 + cpuid + mov res,edx + } + return res; +} +#endif + +#define STBI_SIMD_ALIGN(type, name) __declspec(align(16)) type name + +#if !defined(STBI_NO_JPEG) && defined(STBI_SSE2) +static int stbi__sse2_available(void) +{ + int info3 = stbi__cpuid3(); + return ((info3 >> 26) & 1) != 0; +} +#endif + +#else // assume GCC-style if not VC++ +#define STBI_SIMD_ALIGN(type, name) type name __attribute__((aligned(16))) + +#if !defined(STBI_NO_JPEG) && defined(STBI_SSE2) +static int stbi__sse2_available(void) +{ + // If we're even attempting to compile this on GCC/Clang, that means + // -msse2 is on, which means the compiler is allowed to use SSE2 + // instructions at will, and so are we. + return 1; +} +#endif + +#endif +#endif + +// ARM NEON +#if defined(STBI_NO_SIMD) && defined(STBI_NEON) +#undef STBI_NEON +#endif + +#ifdef STBI_NEON +#include +#ifdef _MSC_VER +#define STBI_SIMD_ALIGN(type, name) __declspec(align(16)) type name +#else +#define STBI_SIMD_ALIGN(type, name) type name __attribute__((aligned(16))) +#endif +#endif + +#ifndef STBI_SIMD_ALIGN +#define STBI_SIMD_ALIGN(type, name) type name +#endif + +#ifndef STBI_MAX_DIMENSIONS +#define STBI_MAX_DIMENSIONS (1 << 24) +#endif + +/////////////////////////////////////////////// +// +// stbi__context struct and start_xxx functions + +// stbi__context structure is our basic context used by all images, so it +// contains all the IO context, plus some basic image information +typedef struct +{ + stbi__uint32 img_x, img_y; + int img_n, img_out_n; + + stbi_io_callbacks io; + void *io_user_data; + + int read_from_callbacks; + int buflen; + stbi_uc buffer_start[128]; + int callback_already_read; + + stbi_uc *img_buffer, *img_buffer_end; + stbi_uc *img_buffer_original, *img_buffer_original_end; +} stbi__context; + + +static void stbi__refill_buffer(stbi__context *s); + +// initialize a memory-decode context +static void stbi__start_mem(stbi__context *s, stbi_uc const *buffer, int len) +{ + s->io.read = NULL; + s->read_from_callbacks = 0; + s->callback_already_read = 0; + s->img_buffer = s->img_buffer_original = (stbi_uc *) buffer; + s->img_buffer_end = s->img_buffer_original_end = (stbi_uc *) buffer+len; +} + +// initialize a callback-based context +static void stbi__start_callbacks(stbi__context *s, stbi_io_callbacks *c, void *user) +{ + s->io = *c; + s->io_user_data = user; + s->buflen = sizeof(s->buffer_start); + s->read_from_callbacks = 1; + s->callback_already_read = 0; + s->img_buffer = s->img_buffer_original = s->buffer_start; + stbi__refill_buffer(s); + s->img_buffer_original_end = s->img_buffer_end; +} + +#ifndef STBI_NO_STDIO + +static int stbi__stdio_read(void *user, char *data, int size) +{ + return (int) fread(data,1,size,(FILE*) user); +} + +static void stbi__stdio_skip(void *user, int n) +{ + int ch; + fseek((FILE*) user, n, SEEK_CUR); + ch = fgetc((FILE*) user); /* have to read a byte to reset feof()'s flag */ + if (ch != EOF) { + ungetc(ch, (FILE *) user); /* push byte back onto stream if valid. */ + } +} + +static int stbi__stdio_eof(void *user) +{ + return feof((FILE*) user) || ferror((FILE *) user); +} + +static stbi_io_callbacks stbi__stdio_callbacks = +{ + stbi__stdio_read, + stbi__stdio_skip, + stbi__stdio_eof, +}; + +static void stbi__start_file(stbi__context *s, FILE *f) +{ + stbi__start_callbacks(s, &stbi__stdio_callbacks, (void *) f); +} + +//static void stop_file(stbi__context *s) { } + +#endif // !STBI_NO_STDIO + +static void stbi__rewind(stbi__context *s) +{ + // conceptually rewind SHOULD rewind to the beginning of the stream, + // but we just rewind to the beginning of the initial buffer, because + // we only use it after doing 'test', which only ever looks at at most 92 bytes + s->img_buffer = s->img_buffer_original; + s->img_buffer_end = s->img_buffer_original_end; +} + +enum +{ + STBI_ORDER_RGB, + STBI_ORDER_BGR +}; + +typedef struct +{ + int bits_per_channel; + int num_channels; + int channel_order; +} stbi__result_info; + +#ifndef STBI_NO_JPEG +static int stbi__jpeg_test(stbi__context *s); +static void *stbi__jpeg_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri); +static int stbi__jpeg_info(stbi__context *s, int *x, int *y, int *comp); +#endif + +#ifndef STBI_NO_PNG +static int stbi__png_test(stbi__context *s); +static void *stbi__png_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri); +static int stbi__png_info(stbi__context *s, int *x, int *y, int *comp); +static int stbi__png_is16(stbi__context *s); +#endif + +#ifndef STBI_NO_BMP +static int stbi__bmp_test(stbi__context *s); +static void *stbi__bmp_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri); +static int stbi__bmp_info(stbi__context *s, int *x, int *y, int *comp); +#endif + +#ifndef STBI_NO_TGA +static int stbi__tga_test(stbi__context *s); +static void *stbi__tga_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri); +static int stbi__tga_info(stbi__context *s, int *x, int *y, int *comp); +#endif + +#ifndef STBI_NO_PSD +static int stbi__psd_test(stbi__context *s); +static void *stbi__psd_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri, int bpc); +static int stbi__psd_info(stbi__context *s, int *x, int *y, int *comp); +static int stbi__psd_is16(stbi__context *s); +#endif + +#ifndef STBI_NO_HDR +static int stbi__hdr_test(stbi__context *s); +static float *stbi__hdr_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri); +static int stbi__hdr_info(stbi__context *s, int *x, int *y, int *comp); +#endif + +#ifndef STBI_NO_PIC +static int stbi__pic_test(stbi__context *s); +static void *stbi__pic_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri); +static int stbi__pic_info(stbi__context *s, int *x, int *y, int *comp); +#endif + +#ifndef STBI_NO_GIF +static int stbi__gif_test(stbi__context *s); +static void *stbi__gif_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri); +static void *stbi__load_gif_main(stbi__context *s, int **delays, int *x, int *y, int *z, int *comp, int req_comp); +static int stbi__gif_info(stbi__context *s, int *x, int *y, int *comp); +#endif + +#ifndef STBI_NO_PNM +static int stbi__pnm_test(stbi__context *s); +static void *stbi__pnm_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri); +static int stbi__pnm_info(stbi__context *s, int *x, int *y, int *comp); +static int stbi__pnm_is16(stbi__context *s); +#endif + +static +#ifdef STBI_THREAD_LOCAL +STBI_THREAD_LOCAL +#endif +const char *stbi__g_failure_reason; + +STBIDEF const char *stbi_failure_reason(void) +{ + return stbi__g_failure_reason; +} + +#ifndef STBI_NO_FAILURE_STRINGS +static int stbi__err(const char *str) +{ + stbi__g_failure_reason = str; + return 0; +} +#endif + +static void *stbi__malloc(size_t size) +{ + return STBI_MALLOC(size); +} + +// stb_image uses ints pervasively, including for offset calculations. +// therefore the largest decoded image size we can support with the +// current code, even on 64-bit targets, is INT_MAX. this is not a +// significant limitation for the intended use case. +// +// we do, however, need to make sure our size calculations don't +// overflow. hence a few helper functions for size calculations that +// multiply integers together, making sure that they're non-negative +// and no overflow occurs. + +// return 1 if the sum is valid, 0 on overflow. +// negative terms are considered invalid. +static int stbi__addsizes_valid(int a, int b) +{ + if (b < 0) return 0; + // now 0 <= b <= INT_MAX, hence also + // 0 <= INT_MAX - b <= INTMAX. + // And "a + b <= INT_MAX" (which might overflow) is the + // same as a <= INT_MAX - b (no overflow) + return a <= INT_MAX - b; +} + +// returns 1 if the product is valid, 0 on overflow. +// negative factors are considered invalid. +static int stbi__mul2sizes_valid(int a, int b) +{ + if (a < 0 || b < 0) return 0; + if (b == 0) return 1; // mul-by-0 is always safe + // portable way to check for no overflows in a*b + return a <= INT_MAX/b; +} + +#if !defined(STBI_NO_JPEG) || !defined(STBI_NO_PNG) || !defined(STBI_NO_TGA) || !defined(STBI_NO_HDR) +// returns 1 if "a*b + add" has no negative terms/factors and doesn't overflow +static int stbi__mad2sizes_valid(int a, int b, int add) +{ + return stbi__mul2sizes_valid(a, b) && stbi__addsizes_valid(a*b, add); +} +#endif + +// returns 1 if "a*b*c + add" has no negative terms/factors and doesn't overflow +static int stbi__mad3sizes_valid(int a, int b, int c, int add) +{ + return stbi__mul2sizes_valid(a, b) && stbi__mul2sizes_valid(a*b, c) && + stbi__addsizes_valid(a*b*c, add); +} + +// returns 1 if "a*b*c*d + add" has no negative terms/factors and doesn't overflow +#if !defined(STBI_NO_LINEAR) || !defined(STBI_NO_HDR) || !defined(STBI_NO_PNM) +static int stbi__mad4sizes_valid(int a, int b, int c, int d, int add) +{ + return stbi__mul2sizes_valid(a, b) && stbi__mul2sizes_valid(a*b, c) && + stbi__mul2sizes_valid(a*b*c, d) && stbi__addsizes_valid(a*b*c*d, add); +} +#endif + +#if !defined(STBI_NO_JPEG) || !defined(STBI_NO_PNG) || !defined(STBI_NO_TGA) || !defined(STBI_NO_HDR) +// mallocs with size overflow checking +static void *stbi__malloc_mad2(int a, int b, int add) +{ + if (!stbi__mad2sizes_valid(a, b, add)) return NULL; + return stbi__malloc(a*b + add); +} +#endif + +static void *stbi__malloc_mad3(int a, int b, int c, int add) +{ + if (!stbi__mad3sizes_valid(a, b, c, add)) return NULL; + return stbi__malloc(a*b*c + add); +} + +#if !defined(STBI_NO_LINEAR) || !defined(STBI_NO_HDR) || !defined(STBI_NO_PNM) +static void *stbi__malloc_mad4(int a, int b, int c, int d, int add) +{ + if (!stbi__mad4sizes_valid(a, b, c, d, add)) return NULL; + return stbi__malloc(a*b*c*d + add); +} +#endif + +// returns 1 if the sum of two signed ints is valid (between -2^31 and 2^31-1 inclusive), 0 on overflow. +static int stbi__addints_valid(int a, int b) +{ + if ((a >= 0) != (b >= 0)) return 1; // a and b have different signs, so no overflow + if (a < 0 && b < 0) return a >= INT_MIN - b; // same as a + b >= INT_MIN; INT_MIN - b cannot overflow since b < 0. + return a <= INT_MAX - b; +} + +// returns 1 if the product of two ints fits in a signed short, 0 on overflow. +static int stbi__mul2shorts_valid(int a, int b) +{ + if (b == 0 || b == -1) return 1; // multiplication by 0 is always 0; check for -1 so SHRT_MIN/b doesn't overflow + if ((a >= 0) == (b >= 0)) return a <= SHRT_MAX/b; // product is positive, so similar to mul2sizes_valid + if (b < 0) return a <= SHRT_MIN / b; // same as a * b >= SHRT_MIN + return a >= SHRT_MIN / b; +} + +// stbi__err - error +// stbi__errpf - error returning pointer to float +// stbi__errpuc - error returning pointer to unsigned char + +#ifdef STBI_NO_FAILURE_STRINGS + #define stbi__err(x,y) 0 +#elif defined(STBI_FAILURE_USERMSG) + #define stbi__err(x,y) stbi__err(y) +#else + #define stbi__err(x,y) stbi__err(x) +#endif + +#define stbi__errpf(x,y) ((float *)(size_t) (stbi__err(x,y)?NULL:NULL)) +#define stbi__errpuc(x,y) ((unsigned char *)(size_t) (stbi__err(x,y)?NULL:NULL)) + +STBIDEF void stbi_image_free(void *retval_from_stbi_load) +{ + STBI_FREE(retval_from_stbi_load); +} + +#ifndef STBI_NO_LINEAR +static float *stbi__ldr_to_hdr(stbi_uc *data, int x, int y, int comp); +#endif + +#ifndef STBI_NO_HDR +static stbi_uc *stbi__hdr_to_ldr(float *data, int x, int y, int comp); +#endif + +static int stbi__vertically_flip_on_load_global = 0; + +STBIDEF void stbi_set_flip_vertically_on_load(int flag_true_if_should_flip) +{ + stbi__vertically_flip_on_load_global = flag_true_if_should_flip; +} + +#ifndef STBI_THREAD_LOCAL +#define stbi__vertically_flip_on_load stbi__vertically_flip_on_load_global +#else +static STBI_THREAD_LOCAL int stbi__vertically_flip_on_load_local, stbi__vertically_flip_on_load_set; + +STBIDEF void stbi_set_flip_vertically_on_load_thread(int flag_true_if_should_flip) +{ + stbi__vertically_flip_on_load_local = flag_true_if_should_flip; + stbi__vertically_flip_on_load_set = 1; +} + +#define stbi__vertically_flip_on_load (stbi__vertically_flip_on_load_set \ + ? stbi__vertically_flip_on_load_local \ + : stbi__vertically_flip_on_load_global) +#endif // STBI_THREAD_LOCAL + +static void *stbi__load_main(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri, int bpc) +{ + memset(ri, 0, sizeof(*ri)); // make sure it's initialized if we add new fields + ri->bits_per_channel = 8; // default is 8 so most paths don't have to be changed + ri->channel_order = STBI_ORDER_RGB; // all current input & output are this, but this is here so we can add BGR order + ri->num_channels = 0; + + // test the formats with a very explicit header first (at least a FOURCC + // or distinctive magic number first) + #ifndef STBI_NO_PNG + if (stbi__png_test(s)) return stbi__png_load(s,x,y,comp,req_comp, ri); + #endif + #ifndef STBI_NO_BMP + if (stbi__bmp_test(s)) return stbi__bmp_load(s,x,y,comp,req_comp, ri); + #endif + #ifndef STBI_NO_GIF + if (stbi__gif_test(s)) return stbi__gif_load(s,x,y,comp,req_comp, ri); + #endif + #ifndef STBI_NO_PSD + if (stbi__psd_test(s)) return stbi__psd_load(s,x,y,comp,req_comp, ri, bpc); + #else + STBI_NOTUSED(bpc); + #endif + #ifndef STBI_NO_PIC + if (stbi__pic_test(s)) return stbi__pic_load(s,x,y,comp,req_comp, ri); + #endif + + // then the formats that can end up attempting to load with just 1 or 2 + // bytes matching expectations; these are prone to false positives, so + // try them later + #ifndef STBI_NO_JPEG + if (stbi__jpeg_test(s)) return stbi__jpeg_load(s,x,y,comp,req_comp, ri); + #endif + #ifndef STBI_NO_PNM + if (stbi__pnm_test(s)) return stbi__pnm_load(s,x,y,comp,req_comp, ri); + #endif + + #ifndef STBI_NO_HDR + if (stbi__hdr_test(s)) { + float *hdr = stbi__hdr_load(s, x,y,comp,req_comp, ri); + return stbi__hdr_to_ldr(hdr, *x, *y, req_comp ? req_comp : *comp); + } + #endif + + #ifndef STBI_NO_TGA + // test tga last because it's a crappy test! + if (stbi__tga_test(s)) + return stbi__tga_load(s,x,y,comp,req_comp, ri); + #endif + + return stbi__errpuc("unknown image type", "Image not of any known type, or corrupt"); +} + +static stbi_uc *stbi__convert_16_to_8(stbi__uint16 *orig, int w, int h, int channels) +{ + int i; + int img_len = w * h * channels; + stbi_uc *reduced; + + reduced = (stbi_uc *) stbi__malloc(img_len); + if (reduced == NULL) return stbi__errpuc("outofmem", "Out of memory"); + + for (i = 0; i < img_len; ++i) + reduced[i] = (stbi_uc)((orig[i] >> 8) & 0xFF); // top half of each byte is sufficient approx of 16->8 bit scaling + + STBI_FREE(orig); + return reduced; +} + +static stbi__uint16 *stbi__convert_8_to_16(stbi_uc *orig, int w, int h, int channels) +{ + int i; + int img_len = w * h * channels; + stbi__uint16 *enlarged; + + enlarged = (stbi__uint16 *) stbi__malloc(img_len*2); + if (enlarged == NULL) return (stbi__uint16 *) stbi__errpuc("outofmem", "Out of memory"); + + for (i = 0; i < img_len; ++i) + enlarged[i] = (stbi__uint16)((orig[i] << 8) + orig[i]); // replicate to high and low byte, maps 0->0, 255->0xffff + + STBI_FREE(orig); + return enlarged; +} + +static void stbi__vertical_flip(void *image, int w, int h, int bytes_per_pixel) +{ + int row; + size_t bytes_per_row = (size_t)w * bytes_per_pixel; + stbi_uc temp[2048]; + stbi_uc *bytes = (stbi_uc *)image; + + for (row = 0; row < (h>>1); row++) { + stbi_uc *row0 = bytes + row*bytes_per_row; + stbi_uc *row1 = bytes + (h - row - 1)*bytes_per_row; + // swap row0 with row1 + size_t bytes_left = bytes_per_row; + while (bytes_left) { + size_t bytes_copy = (bytes_left < sizeof(temp)) ? bytes_left : sizeof(temp); + memcpy(temp, row0, bytes_copy); + memcpy(row0, row1, bytes_copy); + memcpy(row1, temp, bytes_copy); + row0 += bytes_copy; + row1 += bytes_copy; + bytes_left -= bytes_copy; + } + } +} + +#ifndef STBI_NO_GIF +static void stbi__vertical_flip_slices(void *image, int w, int h, int z, int bytes_per_pixel) +{ + int slice; + int slice_size = w * h * bytes_per_pixel; + + stbi_uc *bytes = (stbi_uc *)image; + for (slice = 0; slice < z; ++slice) { + stbi__vertical_flip(bytes, w, h, bytes_per_pixel); + bytes += slice_size; + } +} +#endif + +static unsigned char *stbi__load_and_postprocess_8bit(stbi__context *s, int *x, int *y, int *comp, int req_comp) +{ + stbi__result_info ri; + void *result = stbi__load_main(s, x, y, comp, req_comp, &ri, 8); + + if (result == NULL) + return NULL; + + // it is the responsibility of the loaders to make sure we get either 8 or 16 bit. + STBI_ASSERT(ri.bits_per_channel == 8 || ri.bits_per_channel == 16); + + if (ri.bits_per_channel != 8) { + result = stbi__convert_16_to_8((stbi__uint16 *) result, *x, *y, req_comp == 0 ? *comp : req_comp); + ri.bits_per_channel = 8; + } + + // @TODO: move stbi__convert_format to here + + if (stbi__vertically_flip_on_load) { + int channels = req_comp ? req_comp : *comp; + stbi__vertical_flip(result, *x, *y, channels * sizeof(stbi_uc)); + } + + return (unsigned char *) result; +} + +static stbi__uint16 *stbi__load_and_postprocess_16bit(stbi__context *s, int *x, int *y, int *comp, int req_comp) +{ + stbi__result_info ri; + void *result = stbi__load_main(s, x, y, comp, req_comp, &ri, 16); + + if (result == NULL) + return NULL; + + // it is the responsibility of the loaders to make sure we get either 8 or 16 bit. + STBI_ASSERT(ri.bits_per_channel == 8 || ri.bits_per_channel == 16); + + if (ri.bits_per_channel != 16) { + result = stbi__convert_8_to_16((stbi_uc *) result, *x, *y, req_comp == 0 ? *comp : req_comp); + ri.bits_per_channel = 16; + } + + // @TODO: move stbi__convert_format16 to here + // @TODO: special case RGB-to-Y (and RGBA-to-YA) for 8-bit-to-16-bit case to keep more precision + + if (stbi__vertically_flip_on_load) { + int channels = req_comp ? req_comp : *comp; + stbi__vertical_flip(result, *x, *y, channels * sizeof(stbi__uint16)); + } + + return (stbi__uint16 *) result; +} + +#if !defined(STBI_NO_HDR) && !defined(STBI_NO_LINEAR) +static void stbi__float_postprocess(float *result, int *x, int *y, int *comp, int req_comp) +{ + if (stbi__vertically_flip_on_load && result != NULL) { + int channels = req_comp ? req_comp : *comp; + stbi__vertical_flip(result, *x, *y, channels * sizeof(float)); + } +} +#endif + +#ifndef STBI_NO_STDIO + +#if defined(_WIN32) && defined(STBI_WINDOWS_UTF8) +STBI_EXTERN __declspec(dllimport) int __stdcall MultiByteToWideChar(unsigned int cp, unsigned long flags, const char *str, int cbmb, wchar_t *widestr, int cchwide); +STBI_EXTERN __declspec(dllimport) int __stdcall WideCharToMultiByte(unsigned int cp, unsigned long flags, const wchar_t *widestr, int cchwide, char *str, int cbmb, const char *defchar, int *used_default); +#endif + +#if defined(_WIN32) && defined(STBI_WINDOWS_UTF8) +STBIDEF int stbi_convert_wchar_to_utf8(char *buffer, size_t bufferlen, const wchar_t* input) +{ + return WideCharToMultiByte(65001 /* UTF8 */, 0, input, -1, buffer, (int) bufferlen, NULL, NULL); +} +#endif + +static FILE *stbi__fopen(char const *filename, char const *mode) +{ + FILE *f; +#if defined(_WIN32) && defined(STBI_WINDOWS_UTF8) + wchar_t wMode[64]; + wchar_t wFilename[1024]; + if (0 == MultiByteToWideChar(65001 /* UTF8 */, 0, filename, -1, wFilename, sizeof(wFilename)/sizeof(*wFilename))) + return 0; + + if (0 == MultiByteToWideChar(65001 /* UTF8 */, 0, mode, -1, wMode, sizeof(wMode)/sizeof(*wMode))) + return 0; + +#if defined(_MSC_VER) && _MSC_VER >= 1400 + if (0 != _wfopen_s(&f, wFilename, wMode)) + f = 0; +#else + f = _wfopen(wFilename, wMode); +#endif + +#elif defined(_MSC_VER) && _MSC_VER >= 1400 + if (0 != fopen_s(&f, filename, mode)) + f=0; +#else + f = fopen(filename, mode); +#endif + return f; +} + + +STBIDEF stbi_uc *stbi_load(char const *filename, int *x, int *y, int *comp, int req_comp) +{ + FILE *f = stbi__fopen(filename, "rb"); + unsigned char *result; + if (!f) return stbi__errpuc("can't fopen", "Unable to open file"); + result = stbi_load_from_file(f,x,y,comp,req_comp); + fclose(f); + return result; +} + +STBIDEF stbi_uc *stbi_load_from_file(FILE *f, int *x, int *y, int *comp, int req_comp) +{ + unsigned char *result; + stbi__context s; + stbi__start_file(&s,f); + result = stbi__load_and_postprocess_8bit(&s,x,y,comp,req_comp); + if (result) { + // need to 'unget' all the characters in the IO buffer + fseek(f, - (int) (s.img_buffer_end - s.img_buffer), SEEK_CUR); + } + return result; +} + +STBIDEF stbi__uint16 *stbi_load_from_file_16(FILE *f, int *x, int *y, int *comp, int req_comp) +{ + stbi__uint16 *result; + stbi__context s; + stbi__start_file(&s,f); + result = stbi__load_and_postprocess_16bit(&s,x,y,comp,req_comp); + if (result) { + // need to 'unget' all the characters in the IO buffer + fseek(f, - (int) (s.img_buffer_end - s.img_buffer), SEEK_CUR); + } + return result; +} + +STBIDEF stbi_us *stbi_load_16(char const *filename, int *x, int *y, int *comp, int req_comp) +{ + FILE *f = stbi__fopen(filename, "rb"); + stbi__uint16 *result; + if (!f) return (stbi_us *) stbi__errpuc("can't fopen", "Unable to open file"); + result = stbi_load_from_file_16(f,x,y,comp,req_comp); + fclose(f); + return result; +} + + +#endif //!STBI_NO_STDIO + +STBIDEF stbi_us *stbi_load_16_from_memory(stbi_uc const *buffer, int len, int *x, int *y, int *channels_in_file, int desired_channels) +{ + stbi__context s; + stbi__start_mem(&s,buffer,len); + return stbi__load_and_postprocess_16bit(&s,x,y,channels_in_file,desired_channels); +} + +STBIDEF stbi_us *stbi_load_16_from_callbacks(stbi_io_callbacks const *clbk, void *user, int *x, int *y, int *channels_in_file, int desired_channels) +{ + stbi__context s; + stbi__start_callbacks(&s, (stbi_io_callbacks *)clbk, user); + return stbi__load_and_postprocess_16bit(&s,x,y,channels_in_file,desired_channels); +} + +STBIDEF stbi_uc *stbi_load_from_memory(stbi_uc const *buffer, int len, int *x, int *y, int *comp, int req_comp) +{ + stbi__context s; + stbi__start_mem(&s,buffer,len); + return stbi__load_and_postprocess_8bit(&s,x,y,comp,req_comp); +} + +STBIDEF stbi_uc *stbi_load_from_callbacks(stbi_io_callbacks const *clbk, void *user, int *x, int *y, int *comp, int req_comp) +{ + stbi__context s; + stbi__start_callbacks(&s, (stbi_io_callbacks *) clbk, user); + return stbi__load_and_postprocess_8bit(&s,x,y,comp,req_comp); +} + +#ifndef STBI_NO_GIF +STBIDEF stbi_uc *stbi_load_gif_from_memory(stbi_uc const *buffer, int len, int **delays, int *x, int *y, int *z, int *comp, int req_comp) +{ + unsigned char *result; + stbi__context s; + stbi__start_mem(&s,buffer,len); + + result = (unsigned char*) stbi__load_gif_main(&s, delays, x, y, z, comp, req_comp); + if (stbi__vertically_flip_on_load) { + stbi__vertical_flip_slices( result, *x, *y, *z, *comp ); + } + + return result; +} +#endif + +#ifndef STBI_NO_LINEAR +static float *stbi__loadf_main(stbi__context *s, int *x, int *y, int *comp, int req_comp) +{ + unsigned char *data; + #ifndef STBI_NO_HDR + if (stbi__hdr_test(s)) { + stbi__result_info ri; + float *hdr_data = stbi__hdr_load(s,x,y,comp,req_comp, &ri); + if (hdr_data) + stbi__float_postprocess(hdr_data,x,y,comp,req_comp); + return hdr_data; + } + #endif + data = stbi__load_and_postprocess_8bit(s, x, y, comp, req_comp); + if (data) + return stbi__ldr_to_hdr(data, *x, *y, req_comp ? req_comp : *comp); + return stbi__errpf("unknown image type", "Image not of any known type, or corrupt"); +} + +STBIDEF float *stbi_loadf_from_memory(stbi_uc const *buffer, int len, int *x, int *y, int *comp, int req_comp) +{ + stbi__context s; + stbi__start_mem(&s,buffer,len); + return stbi__loadf_main(&s,x,y,comp,req_comp); +} + +STBIDEF float *stbi_loadf_from_callbacks(stbi_io_callbacks const *clbk, void *user, int *x, int *y, int *comp, int req_comp) +{ + stbi__context s; + stbi__start_callbacks(&s, (stbi_io_callbacks *) clbk, user); + return stbi__loadf_main(&s,x,y,comp,req_comp); +} + +#ifndef STBI_NO_STDIO +STBIDEF float *stbi_loadf(char const *filename, int *x, int *y, int *comp, int req_comp) +{ + float *result; + FILE *f = stbi__fopen(filename, "rb"); + if (!f) return stbi__errpf("can't fopen", "Unable to open file"); + result = stbi_loadf_from_file(f,x,y,comp,req_comp); + fclose(f); + return result; +} + +STBIDEF float *stbi_loadf_from_file(FILE *f, int *x, int *y, int *comp, int req_comp) +{ + stbi__context s; + stbi__start_file(&s,f); + return stbi__loadf_main(&s,x,y,comp,req_comp); +} +#endif // !STBI_NO_STDIO + +#endif // !STBI_NO_LINEAR + +// these is-hdr-or-not is defined independent of whether STBI_NO_LINEAR is +// defined, for API simplicity; if STBI_NO_LINEAR is defined, it always +// reports false! + +STBIDEF int stbi_is_hdr_from_memory(stbi_uc const *buffer, int len) +{ + #ifndef STBI_NO_HDR + stbi__context s; + stbi__start_mem(&s,buffer,len); + return stbi__hdr_test(&s); + #else + STBI_NOTUSED(buffer); + STBI_NOTUSED(len); + return 0; + #endif +} + +#ifndef STBI_NO_STDIO +STBIDEF int stbi_is_hdr (char const *filename) +{ + FILE *f = stbi__fopen(filename, "rb"); + int result=0; + if (f) { + result = stbi_is_hdr_from_file(f); + fclose(f); + } + return result; +} + +STBIDEF int stbi_is_hdr_from_file(FILE *f) +{ + #ifndef STBI_NO_HDR + long pos = ftell(f); + int res; + stbi__context s; + stbi__start_file(&s,f); + res = stbi__hdr_test(&s); + fseek(f, pos, SEEK_SET); + return res; + #else + STBI_NOTUSED(f); + return 0; + #endif +} +#endif // !STBI_NO_STDIO + +STBIDEF int stbi_is_hdr_from_callbacks(stbi_io_callbacks const *clbk, void *user) +{ + #ifndef STBI_NO_HDR + stbi__context s; + stbi__start_callbacks(&s, (stbi_io_callbacks *) clbk, user); + return stbi__hdr_test(&s); + #else + STBI_NOTUSED(clbk); + STBI_NOTUSED(user); + return 0; + #endif +} + +#ifndef STBI_NO_LINEAR +static float stbi__l2h_gamma=2.2f, stbi__l2h_scale=1.0f; + +STBIDEF void stbi_ldr_to_hdr_gamma(float gamma) { stbi__l2h_gamma = gamma; } +STBIDEF void stbi_ldr_to_hdr_scale(float scale) { stbi__l2h_scale = scale; } +#endif + +static float stbi__h2l_gamma_i=1.0f/2.2f, stbi__h2l_scale_i=1.0f; + +STBIDEF void stbi_hdr_to_ldr_gamma(float gamma) { stbi__h2l_gamma_i = 1/gamma; } +STBIDEF void stbi_hdr_to_ldr_scale(float scale) { stbi__h2l_scale_i = 1/scale; } + + +////////////////////////////////////////////////////////////////////////////// +// +// Common code used by all image loaders +// + +enum +{ + STBI__SCAN_load=0, + STBI__SCAN_type, + STBI__SCAN_header +}; + +static void stbi__refill_buffer(stbi__context *s) +{ + int n = (s->io.read)(s->io_user_data,(char*)s->buffer_start,s->buflen); + s->callback_already_read += (int) (s->img_buffer - s->img_buffer_original); + if (n == 0) { + // at end of file, treat same as if from memory, but need to handle case + // where s->img_buffer isn't pointing to safe memory, e.g. 0-byte file + s->read_from_callbacks = 0; + s->img_buffer = s->buffer_start; + s->img_buffer_end = s->buffer_start+1; + *s->img_buffer = 0; + } else { + s->img_buffer = s->buffer_start; + s->img_buffer_end = s->buffer_start + n; + } +} + +stbi_inline static stbi_uc stbi__get8(stbi__context *s) +{ + if (s->img_buffer < s->img_buffer_end) + return *s->img_buffer++; + if (s->read_from_callbacks) { + stbi__refill_buffer(s); + return *s->img_buffer++; + } + return 0; +} + +#if defined(STBI_NO_JPEG) && defined(STBI_NO_HDR) && defined(STBI_NO_PIC) && defined(STBI_NO_PNM) +// nothing +#else +stbi_inline static int stbi__at_eof(stbi__context *s) +{ + if (s->io.read) { + if (!(s->io.eof)(s->io_user_data)) return 0; + // if feof() is true, check if buffer = end + // special case: we've only got the special 0 character at the end + if (s->read_from_callbacks == 0) return 1; + } + + return s->img_buffer >= s->img_buffer_end; +} +#endif + +#if defined(STBI_NO_JPEG) && defined(STBI_NO_PNG) && defined(STBI_NO_BMP) && defined(STBI_NO_PSD) && defined(STBI_NO_TGA) && defined(STBI_NO_GIF) && defined(STBI_NO_PIC) +// nothing +#else +static void stbi__skip(stbi__context *s, int n) +{ + if (n == 0) return; // already there! + if (n < 0) { + s->img_buffer = s->img_buffer_end; + return; + } + if (s->io.read) { + int blen = (int) (s->img_buffer_end - s->img_buffer); + if (blen < n) { + s->img_buffer = s->img_buffer_end; + (s->io.skip)(s->io_user_data, n - blen); + return; + } + } + s->img_buffer += n; +} +#endif + +#if defined(STBI_NO_PNG) && defined(STBI_NO_TGA) && defined(STBI_NO_HDR) && defined(STBI_NO_PNM) +// nothing +#else +static int stbi__getn(stbi__context *s, stbi_uc *buffer, int n) +{ + if (s->io.read) { + int blen = (int) (s->img_buffer_end - s->img_buffer); + if (blen < n) { + int res, count; + + memcpy(buffer, s->img_buffer, blen); + + count = (s->io.read)(s->io_user_data, (char*) buffer + blen, n - blen); + res = (count == (n-blen)); + s->img_buffer = s->img_buffer_end; + return res; + } + } + + if (s->img_buffer+n <= s->img_buffer_end) { + memcpy(buffer, s->img_buffer, n); + s->img_buffer += n; + return 1; + } else + return 0; +} +#endif + +#if defined(STBI_NO_JPEG) && defined(STBI_NO_PNG) && defined(STBI_NO_PSD) && defined(STBI_NO_PIC) +// nothing +#else +static int stbi__get16be(stbi__context *s) +{ + int z = stbi__get8(s); + return (z << 8) + stbi__get8(s); +} +#endif + +#if defined(STBI_NO_PNG) && defined(STBI_NO_PSD) && defined(STBI_NO_PIC) +// nothing +#else +static stbi__uint32 stbi__get32be(stbi__context *s) +{ + stbi__uint32 z = stbi__get16be(s); + return (z << 16) + stbi__get16be(s); +} +#endif + +#if defined(STBI_NO_BMP) && defined(STBI_NO_TGA) && defined(STBI_NO_GIF) +// nothing +#else +static int stbi__get16le(stbi__context *s) +{ + int z = stbi__get8(s); + return z + (stbi__get8(s) << 8); +} +#endif + +#ifndef STBI_NO_BMP +static stbi__uint32 stbi__get32le(stbi__context *s) +{ + stbi__uint32 z = stbi__get16le(s); + z += (stbi__uint32)stbi__get16le(s) << 16; + return z; +} +#endif + +#define STBI__BYTECAST(x) ((stbi_uc) ((x) & 255)) // truncate int to byte without warnings + +#if defined(STBI_NO_JPEG) && defined(STBI_NO_PNG) && defined(STBI_NO_BMP) && defined(STBI_NO_PSD) && defined(STBI_NO_TGA) && defined(STBI_NO_GIF) && defined(STBI_NO_PIC) && defined(STBI_NO_PNM) +// nothing +#else +////////////////////////////////////////////////////////////////////////////// +// +// generic converter from built-in img_n to req_comp +// individual types do this automatically as much as possible (e.g. jpeg +// does all cases internally since it needs to colorspace convert anyway, +// and it never has alpha, so very few cases ). png can automatically +// interleave an alpha=255 channel, but falls back to this for other cases +// +// assume data buffer is malloced, so malloc a new one and free that one +// only failure mode is malloc failing + +static stbi_uc stbi__compute_y(int r, int g, int b) +{ + return (stbi_uc) (((r*77) + (g*150) + (29*b)) >> 8); +} +#endif + +#if defined(STBI_NO_PNG) && defined(STBI_NO_BMP) && defined(STBI_NO_PSD) && defined(STBI_NO_TGA) && defined(STBI_NO_GIF) && defined(STBI_NO_PIC) && defined(STBI_NO_PNM) +// nothing +#else +static unsigned char *stbi__convert_format(unsigned char *data, int img_n, int req_comp, unsigned int x, unsigned int y) +{ + int i,j; + unsigned char *good; + + if (req_comp == img_n) return data; + STBI_ASSERT(req_comp >= 1 && req_comp <= 4); + + good = (unsigned char *) stbi__malloc_mad3(req_comp, x, y, 0); + if (good == NULL) { + STBI_FREE(data); + return stbi__errpuc("outofmem", "Out of memory"); + } + + for (j=0; j < (int) y; ++j) { + unsigned char *src = data + j * x * img_n ; + unsigned char *dest = good + j * x * req_comp; + + #define STBI__COMBO(a,b) ((a)*8+(b)) + #define STBI__CASE(a,b) case STBI__COMBO(a,b): for(i=x-1; i >= 0; --i, src += a, dest += b) + // convert source image with img_n components to one with req_comp components; + // avoid switch per pixel, so use switch per scanline and massive macros + switch (STBI__COMBO(img_n, req_comp)) { + STBI__CASE(1,2) { dest[0]=src[0]; dest[1]=255; } break; + STBI__CASE(1,3) { dest[0]=dest[1]=dest[2]=src[0]; } break; + STBI__CASE(1,4) { dest[0]=dest[1]=dest[2]=src[0]; dest[3]=255; } break; + STBI__CASE(2,1) { dest[0]=src[0]; } break; + STBI__CASE(2,3) { dest[0]=dest[1]=dest[2]=src[0]; } break; + STBI__CASE(2,4) { dest[0]=dest[1]=dest[2]=src[0]; dest[3]=src[1]; } break; + STBI__CASE(3,4) { dest[0]=src[0];dest[1]=src[1];dest[2]=src[2];dest[3]=255; } break; + STBI__CASE(3,1) { dest[0]=stbi__compute_y(src[0],src[1],src[2]); } break; + STBI__CASE(3,2) { dest[0]=stbi__compute_y(src[0],src[1],src[2]); dest[1] = 255; } break; + STBI__CASE(4,1) { dest[0]=stbi__compute_y(src[0],src[1],src[2]); } break; + STBI__CASE(4,2) { dest[0]=stbi__compute_y(src[0],src[1],src[2]); dest[1] = src[3]; } break; + STBI__CASE(4,3) { dest[0]=src[0];dest[1]=src[1];dest[2]=src[2]; } break; + default: STBI_ASSERT(0); STBI_FREE(data); STBI_FREE(good); return stbi__errpuc("unsupported", "Unsupported format conversion"); + } + #undef STBI__CASE + } + + STBI_FREE(data); + return good; +} +#endif + +#if defined(STBI_NO_PNG) && defined(STBI_NO_PSD) +// nothing +#else +static stbi__uint16 stbi__compute_y_16(int r, int g, int b) +{ + return (stbi__uint16) (((r*77) + (g*150) + (29*b)) >> 8); +} +#endif + +#if defined(STBI_NO_PNG) && defined(STBI_NO_PSD) +// nothing +#else +static stbi__uint16 *stbi__convert_format16(stbi__uint16 *data, int img_n, int req_comp, unsigned int x, unsigned int y) +{ + int i,j; + stbi__uint16 *good; + + if (req_comp == img_n) return data; + STBI_ASSERT(req_comp >= 1 && req_comp <= 4); + + good = (stbi__uint16 *) stbi__malloc(req_comp * x * y * 2); + if (good == NULL) { + STBI_FREE(data); + return (stbi__uint16 *) stbi__errpuc("outofmem", "Out of memory"); + } + + for (j=0; j < (int) y; ++j) { + stbi__uint16 *src = data + j * x * img_n ; + stbi__uint16 *dest = good + j * x * req_comp; + + #define STBI__COMBO(a,b) ((a)*8+(b)) + #define STBI__CASE(a,b) case STBI__COMBO(a,b): for(i=x-1; i >= 0; --i, src += a, dest += b) + // convert source image with img_n components to one with req_comp components; + // avoid switch per pixel, so use switch per scanline and massive macros + switch (STBI__COMBO(img_n, req_comp)) { + STBI__CASE(1,2) { dest[0]=src[0]; dest[1]=0xffff; } break; + STBI__CASE(1,3) { dest[0]=dest[1]=dest[2]=src[0]; } break; + STBI__CASE(1,4) { dest[0]=dest[1]=dest[2]=src[0]; dest[3]=0xffff; } break; + STBI__CASE(2,1) { dest[0]=src[0]; } break; + STBI__CASE(2,3) { dest[0]=dest[1]=dest[2]=src[0]; } break; + STBI__CASE(2,4) { dest[0]=dest[1]=dest[2]=src[0]; dest[3]=src[1]; } break; + STBI__CASE(3,4) { dest[0]=src[0];dest[1]=src[1];dest[2]=src[2];dest[3]=0xffff; } break; + STBI__CASE(3,1) { dest[0]=stbi__compute_y_16(src[0],src[1],src[2]); } break; + STBI__CASE(3,2) { dest[0]=stbi__compute_y_16(src[0],src[1],src[2]); dest[1] = 0xffff; } break; + STBI__CASE(4,1) { dest[0]=stbi__compute_y_16(src[0],src[1],src[2]); } break; + STBI__CASE(4,2) { dest[0]=stbi__compute_y_16(src[0],src[1],src[2]); dest[1] = src[3]; } break; + STBI__CASE(4,3) { dest[0]=src[0];dest[1]=src[1];dest[2]=src[2]; } break; + default: STBI_ASSERT(0); STBI_FREE(data); STBI_FREE(good); return (stbi__uint16*) stbi__errpuc("unsupported", "Unsupported format conversion"); + } + #undef STBI__CASE + } + + STBI_FREE(data); + return good; +} +#endif + +#ifndef STBI_NO_LINEAR +static float *stbi__ldr_to_hdr(stbi_uc *data, int x, int y, int comp) +{ + int i,k,n; + float *output; + if (!data) return NULL; + output = (float *) stbi__malloc_mad4(x, y, comp, sizeof(float), 0); + if (output == NULL) { STBI_FREE(data); return stbi__errpf("outofmem", "Out of memory"); } + // compute number of non-alpha components + if (comp & 1) n = comp; else n = comp-1; + for (i=0; i < x*y; ++i) { + for (k=0; k < n; ++k) { + output[i*comp + k] = (float) (pow(data[i*comp+k]/255.0f, stbi__l2h_gamma) * stbi__l2h_scale); + } + } + if (n < comp) { + for (i=0; i < x*y; ++i) { + output[i*comp + n] = data[i*comp + n]/255.0f; + } + } + STBI_FREE(data); + return output; +} +#endif + +#ifndef STBI_NO_HDR +#define stbi__float2int(x) ((int) (x)) +static stbi_uc *stbi__hdr_to_ldr(float *data, int x, int y, int comp) +{ + int i,k,n; + stbi_uc *output; + if (!data) return NULL; + output = (stbi_uc *) stbi__malloc_mad3(x, y, comp, 0); + if (output == NULL) { STBI_FREE(data); return stbi__errpuc("outofmem", "Out of memory"); } + // compute number of non-alpha components + if (comp & 1) n = comp; else n = comp-1; + for (i=0; i < x*y; ++i) { + for (k=0; k < n; ++k) { + float z = (float) pow(data[i*comp+k]*stbi__h2l_scale_i, stbi__h2l_gamma_i) * 255 + 0.5f; + if (z < 0) z = 0; + if (z > 255) z = 255; + output[i*comp + k] = (stbi_uc) stbi__float2int(z); + } + if (k < comp) { + float z = data[i*comp+k] * 255 + 0.5f; + if (z < 0) z = 0; + if (z > 255) z = 255; + output[i*comp + k] = (stbi_uc) stbi__float2int(z); + } + } + STBI_FREE(data); + return output; +} +#endif + +////////////////////////////////////////////////////////////////////////////// +// +// "baseline" JPEG/JFIF decoder +// +// simple implementation +// - doesn't support delayed output of y-dimension +// - simple interface (only one output format: 8-bit interleaved RGB) +// - doesn't try to recover corrupt jpegs +// - doesn't allow partial loading, loading multiple at once +// - still fast on x86 (copying globals into locals doesn't help x86) +// - allocates lots of intermediate memory (full size of all components) +// - non-interleaved case requires this anyway +// - allows good upsampling (see next) +// high-quality +// - upsampled channels are bilinearly interpolated, even across blocks +// - quality integer IDCT derived from IJG's 'slow' +// performance +// - fast huffman; reasonable integer IDCT +// - some SIMD kernels for common paths on targets with SSE2/NEON +// - uses a lot of intermediate memory, could cache poorly + +#ifndef STBI_NO_JPEG + +// huffman decoding acceleration +#define FAST_BITS 9 // larger handles more cases; smaller stomps less cache + +typedef struct +{ + stbi_uc fast[1 << FAST_BITS]; + // weirdly, repacking this into AoS is a 10% speed loss, instead of a win + stbi__uint16 code[256]; + stbi_uc values[256]; + stbi_uc size[257]; + unsigned int maxcode[18]; + int delta[17]; // old 'firstsymbol' - old 'firstcode' +} stbi__huffman; + +typedef struct +{ + stbi__context *s; + stbi__huffman huff_dc[4]; + stbi__huffman huff_ac[4]; + stbi__uint16 dequant[4][64]; + stbi__int16 fast_ac[4][1 << FAST_BITS]; + +// sizes for components, interleaved MCUs + int img_h_max, img_v_max; + int img_mcu_x, img_mcu_y; + int img_mcu_w, img_mcu_h; + +// definition of jpeg image component + struct + { + int id; + int h,v; + int tq; + int hd,ha; + int dc_pred; + + int x,y,w2,h2; + stbi_uc *data; + void *raw_data, *raw_coeff; + stbi_uc *linebuf; + short *coeff; // progressive only + int coeff_w, coeff_h; // number of 8x8 coefficient blocks + } img_comp[4]; + + stbi__uint32 code_buffer; // jpeg entropy-coded buffer + int code_bits; // number of valid bits + unsigned char marker; // marker seen while filling entropy buffer + int nomore; // flag if we saw a marker so must stop + + int progressive; + int spec_start; + int spec_end; + int succ_high; + int succ_low; + int eob_run; + int jfif; + int app14_color_transform; // Adobe APP14 tag + int rgb; + + int scan_n, order[4]; + int restart_interval, todo; + +// kernels + void (*idct_block_kernel)(stbi_uc *out, int out_stride, short data[64]); + void (*YCbCr_to_RGB_kernel)(stbi_uc *out, const stbi_uc *y, const stbi_uc *pcb, const stbi_uc *pcr, int count, int step); + stbi_uc *(*resample_row_hv_2_kernel)(stbi_uc *out, stbi_uc *in_near, stbi_uc *in_far, int w, int hs); +} stbi__jpeg; + +static int stbi__build_huffman(stbi__huffman *h, int *count) +{ + int i,j,k=0; + unsigned int code; + // build size list for each symbol (from JPEG spec) + for (i=0; i < 16; ++i) { + for (j=0; j < count[i]; ++j) { + h->size[k++] = (stbi_uc) (i+1); + if(k >= 257) return stbi__err("bad size list","Corrupt JPEG"); + } + } + h->size[k] = 0; + + // compute actual symbols (from jpeg spec) + code = 0; + k = 0; + for(j=1; j <= 16; ++j) { + // compute delta to add to code to compute symbol id + h->delta[j] = k - code; + if (h->size[k] == j) { + while (h->size[k] == j) + h->code[k++] = (stbi__uint16) (code++); + if (code-1 >= (1u << j)) return stbi__err("bad code lengths","Corrupt JPEG"); + } + // compute largest code + 1 for this size, preshifted as needed later + h->maxcode[j] = code << (16-j); + code <<= 1; + } + h->maxcode[j] = 0xffffffff; + + // build non-spec acceleration table; 255 is flag for not-accelerated + memset(h->fast, 255, 1 << FAST_BITS); + for (i=0; i < k; ++i) { + int s = h->size[i]; + if (s <= FAST_BITS) { + int c = h->code[i] << (FAST_BITS-s); + int m = 1 << (FAST_BITS-s); + for (j=0; j < m; ++j) { + h->fast[c+j] = (stbi_uc) i; + } + } + } + return 1; +} + +// build a table that decodes both magnitude and value of small ACs in +// one go. +static void stbi__build_fast_ac(stbi__int16 *fast_ac, stbi__huffman *h) +{ + int i; + for (i=0; i < (1 << FAST_BITS); ++i) { + stbi_uc fast = h->fast[i]; + fast_ac[i] = 0; + if (fast < 255) { + int rs = h->values[fast]; + int run = (rs >> 4) & 15; + int magbits = rs & 15; + int len = h->size[fast]; + + if (magbits && len + magbits <= FAST_BITS) { + // magnitude code followed by receive_extend code + int k = ((i << len) & ((1 << FAST_BITS) - 1)) >> (FAST_BITS - magbits); + int m = 1 << (magbits - 1); + if (k < m) k += (~0U << magbits) + 1; + // if the result is small enough, we can fit it in fast_ac table + if (k >= -128 && k <= 127) + fast_ac[i] = (stbi__int16) ((k * 256) + (run * 16) + (len + magbits)); + } + } + } +} + +static void stbi__grow_buffer_unsafe(stbi__jpeg *j) +{ + do { + unsigned int b = j->nomore ? 0 : stbi__get8(j->s); + if (b == 0xff) { + int c = stbi__get8(j->s); + while (c == 0xff) c = stbi__get8(j->s); // consume fill bytes + if (c != 0) { + j->marker = (unsigned char) c; + j->nomore = 1; + return; + } + } + j->code_buffer |= b << (24 - j->code_bits); + j->code_bits += 8; + } while (j->code_bits <= 24); +} + +// (1 << n) - 1 +static const stbi__uint32 stbi__bmask[17]={0,1,3,7,15,31,63,127,255,511,1023,2047,4095,8191,16383,32767,65535}; + +// decode a jpeg huffman value from the bitstream +stbi_inline static int stbi__jpeg_huff_decode(stbi__jpeg *j, stbi__huffman *h) +{ + unsigned int temp; + int c,k; + + if (j->code_bits < 16) stbi__grow_buffer_unsafe(j); + + // look at the top FAST_BITS and determine what symbol ID it is, + // if the code is <= FAST_BITS + c = (j->code_buffer >> (32 - FAST_BITS)) & ((1 << FAST_BITS)-1); + k = h->fast[c]; + if (k < 255) { + int s = h->size[k]; + if (s > j->code_bits) + return -1; + j->code_buffer <<= s; + j->code_bits -= s; + return h->values[k]; + } + + // naive test is to shift the code_buffer down so k bits are + // valid, then test against maxcode. To speed this up, we've + // preshifted maxcode left so that it has (16-k) 0s at the + // end; in other words, regardless of the number of bits, it + // wants to be compared against something shifted to have 16; + // that way we don't need to shift inside the loop. + temp = j->code_buffer >> 16; + for (k=FAST_BITS+1 ; ; ++k) + if (temp < h->maxcode[k]) + break; + if (k == 17) { + // error! code not found + j->code_bits -= 16; + return -1; + } + + if (k > j->code_bits) + return -1; + + // convert the huffman code to the symbol id + c = ((j->code_buffer >> (32 - k)) & stbi__bmask[k]) + h->delta[k]; + if(c < 0 || c >= 256) // symbol id out of bounds! + return -1; + STBI_ASSERT((((j->code_buffer) >> (32 - h->size[c])) & stbi__bmask[h->size[c]]) == h->code[c]); + + // convert the id to a symbol + j->code_bits -= k; + j->code_buffer <<= k; + return h->values[c]; +} + +// bias[n] = (-1<code_bits < n) stbi__grow_buffer_unsafe(j); + if (j->code_bits < n) return 0; // ran out of bits from stream, return 0s intead of continuing + + sgn = j->code_buffer >> 31; // sign bit always in MSB; 0 if MSB clear (positive), 1 if MSB set (negative) + k = stbi_lrot(j->code_buffer, n); + j->code_buffer = k & ~stbi__bmask[n]; + k &= stbi__bmask[n]; + j->code_bits -= n; + return k + (stbi__jbias[n] & (sgn - 1)); +} + +// get some unsigned bits +stbi_inline static int stbi__jpeg_get_bits(stbi__jpeg *j, int n) +{ + unsigned int k; + if (j->code_bits < n) stbi__grow_buffer_unsafe(j); + if (j->code_bits < n) return 0; // ran out of bits from stream, return 0s intead of continuing + k = stbi_lrot(j->code_buffer, n); + j->code_buffer = k & ~stbi__bmask[n]; + k &= stbi__bmask[n]; + j->code_bits -= n; + return k; +} + +stbi_inline static int stbi__jpeg_get_bit(stbi__jpeg *j) +{ + unsigned int k; + if (j->code_bits < 1) stbi__grow_buffer_unsafe(j); + if (j->code_bits < 1) return 0; // ran out of bits from stream, return 0s intead of continuing + k = j->code_buffer; + j->code_buffer <<= 1; + --j->code_bits; + return k & 0x80000000; +} + +// given a value that's at position X in the zigzag stream, +// where does it appear in the 8x8 matrix coded as row-major? +static const stbi_uc stbi__jpeg_dezigzag[64+15] = +{ + 0, 1, 8, 16, 9, 2, 3, 10, + 17, 24, 32, 25, 18, 11, 4, 5, + 12, 19, 26, 33, 40, 48, 41, 34, + 27, 20, 13, 6, 7, 14, 21, 28, + 35, 42, 49, 56, 57, 50, 43, 36, + 29, 22, 15, 23, 30, 37, 44, 51, + 58, 59, 52, 45, 38, 31, 39, 46, + 53, 60, 61, 54, 47, 55, 62, 63, + // let corrupt input sample past end + 63, 63, 63, 63, 63, 63, 63, 63, + 63, 63, 63, 63, 63, 63, 63 +}; + +// decode one 64-entry block-- +static int stbi__jpeg_decode_block(stbi__jpeg *j, short data[64], stbi__huffman *hdc, stbi__huffman *hac, stbi__int16 *fac, int b, stbi__uint16 *dequant) +{ + int diff,dc,k; + int t; + + if (j->code_bits < 16) stbi__grow_buffer_unsafe(j); + t = stbi__jpeg_huff_decode(j, hdc); + if (t < 0 || t > 15) return stbi__err("bad huffman code","Corrupt JPEG"); + + // 0 all the ac values now so we can do it 32-bits at a time + memset(data,0,64*sizeof(data[0])); + + diff = t ? stbi__extend_receive(j, t) : 0; + if (!stbi__addints_valid(j->img_comp[b].dc_pred, diff)) return stbi__err("bad delta","Corrupt JPEG"); + dc = j->img_comp[b].dc_pred + diff; + j->img_comp[b].dc_pred = dc; + if (!stbi__mul2shorts_valid(dc, dequant[0])) return stbi__err("can't merge dc and ac", "Corrupt JPEG"); + data[0] = (short) (dc * dequant[0]); + + // decode AC components, see JPEG spec + k = 1; + do { + unsigned int zig; + int c,r,s; + if (j->code_bits < 16) stbi__grow_buffer_unsafe(j); + c = (j->code_buffer >> (32 - FAST_BITS)) & ((1 << FAST_BITS)-1); + r = fac[c]; + if (r) { // fast-AC path + k += (r >> 4) & 15; // run + s = r & 15; // combined length + if (s > j->code_bits) return stbi__err("bad huffman code", "Combined length longer than code bits available"); + j->code_buffer <<= s; + j->code_bits -= s; + // decode into unzigzag'd location + zig = stbi__jpeg_dezigzag[k++]; + data[zig] = (short) ((r >> 8) * dequant[zig]); + } else { + int rs = stbi__jpeg_huff_decode(j, hac); + if (rs < 0) return stbi__err("bad huffman code","Corrupt JPEG"); + s = rs & 15; + r = rs >> 4; + if (s == 0) { + if (rs != 0xf0) break; // end block + k += 16; + } else { + k += r; + // decode into unzigzag'd location + zig = stbi__jpeg_dezigzag[k++]; + data[zig] = (short) (stbi__extend_receive(j,s) * dequant[zig]); + } + } + } while (k < 64); + return 1; +} + +static int stbi__jpeg_decode_block_prog_dc(stbi__jpeg *j, short data[64], stbi__huffman *hdc, int b) +{ + int diff,dc; + int t; + if (j->spec_end != 0) return stbi__err("can't merge dc and ac", "Corrupt JPEG"); + + if (j->code_bits < 16) stbi__grow_buffer_unsafe(j); + + if (j->succ_high == 0) { + // first scan for DC coefficient, must be first + memset(data,0,64*sizeof(data[0])); // 0 all the ac values now + t = stbi__jpeg_huff_decode(j, hdc); + if (t < 0 || t > 15) return stbi__err("can't merge dc and ac", "Corrupt JPEG"); + diff = t ? stbi__extend_receive(j, t) : 0; + + if (!stbi__addints_valid(j->img_comp[b].dc_pred, diff)) return stbi__err("bad delta", "Corrupt JPEG"); + dc = j->img_comp[b].dc_pred + diff; + j->img_comp[b].dc_pred = dc; + if (!stbi__mul2shorts_valid(dc, 1 << j->succ_low)) return stbi__err("can't merge dc and ac", "Corrupt JPEG"); + data[0] = (short) (dc * (1 << j->succ_low)); + } else { + // refinement scan for DC coefficient + if (stbi__jpeg_get_bit(j)) + data[0] += (short) (1 << j->succ_low); + } + return 1; +} + +// @OPTIMIZE: store non-zigzagged during the decode passes, +// and only de-zigzag when dequantizing +static int stbi__jpeg_decode_block_prog_ac(stbi__jpeg *j, short data[64], stbi__huffman *hac, stbi__int16 *fac) +{ + int k; + if (j->spec_start == 0) return stbi__err("can't merge dc and ac", "Corrupt JPEG"); + + if (j->succ_high == 0) { + int shift = j->succ_low; + + if (j->eob_run) { + --j->eob_run; + return 1; + } + + k = j->spec_start; + do { + unsigned int zig; + int c,r,s; + if (j->code_bits < 16) stbi__grow_buffer_unsafe(j); + c = (j->code_buffer >> (32 - FAST_BITS)) & ((1 << FAST_BITS)-1); + r = fac[c]; + if (r) { // fast-AC path + k += (r >> 4) & 15; // run + s = r & 15; // combined length + if (s > j->code_bits) return stbi__err("bad huffman code", "Combined length longer than code bits available"); + j->code_buffer <<= s; + j->code_bits -= s; + zig = stbi__jpeg_dezigzag[k++]; + data[zig] = (short) ((r >> 8) * (1 << shift)); + } else { + int rs = stbi__jpeg_huff_decode(j, hac); + if (rs < 0) return stbi__err("bad huffman code","Corrupt JPEG"); + s = rs & 15; + r = rs >> 4; + if (s == 0) { + if (r < 15) { + j->eob_run = (1 << r); + if (r) + j->eob_run += stbi__jpeg_get_bits(j, r); + --j->eob_run; + break; + } + k += 16; + } else { + k += r; + zig = stbi__jpeg_dezigzag[k++]; + data[zig] = (short) (stbi__extend_receive(j,s) * (1 << shift)); + } + } + } while (k <= j->spec_end); + } else { + // refinement scan for these AC coefficients + + short bit = (short) (1 << j->succ_low); + + if (j->eob_run) { + --j->eob_run; + for (k = j->spec_start; k <= j->spec_end; ++k) { + short *p = &data[stbi__jpeg_dezigzag[k]]; + if (*p != 0) + if (stbi__jpeg_get_bit(j)) + if ((*p & bit)==0) { + if (*p > 0) + *p += bit; + else + *p -= bit; + } + } + } else { + k = j->spec_start; + do { + int r,s; + int rs = stbi__jpeg_huff_decode(j, hac); // @OPTIMIZE see if we can use the fast path here, advance-by-r is so slow, eh + if (rs < 0) return stbi__err("bad huffman code","Corrupt JPEG"); + s = rs & 15; + r = rs >> 4; + if (s == 0) { + if (r < 15) { + j->eob_run = (1 << r) - 1; + if (r) + j->eob_run += stbi__jpeg_get_bits(j, r); + r = 64; // force end of block + } else { + // r=15 s=0 should write 16 0s, so we just do + // a run of 15 0s and then write s (which is 0), + // so we don't have to do anything special here + } + } else { + if (s != 1) return stbi__err("bad huffman code", "Corrupt JPEG"); + // sign bit + if (stbi__jpeg_get_bit(j)) + s = bit; + else + s = -bit; + } + + // advance by r + while (k <= j->spec_end) { + short *p = &data[stbi__jpeg_dezigzag[k++]]; + if (*p != 0) { + if (stbi__jpeg_get_bit(j)) + if ((*p & bit)==0) { + if (*p > 0) + *p += bit; + else + *p -= bit; + } + } else { + if (r == 0) { + *p = (short) s; + break; + } + --r; + } + } + } while (k <= j->spec_end); + } + } + return 1; +} + +// take a -128..127 value and stbi__clamp it and convert to 0..255 +stbi_inline static stbi_uc stbi__clamp(int x) +{ + // trick to use a single test to catch both cases + if ((unsigned int) x > 255) { + if (x < 0) return 0; + if (x > 255) return 255; + } + return (stbi_uc) x; +} + +#define stbi__f2f(x) ((int) (((x) * 4096 + 0.5))) +#define stbi__fsh(x) ((x) * 4096) + +// derived from jidctint -- DCT_ISLOW +#define STBI__IDCT_1D(s0,s1,s2,s3,s4,s5,s6,s7) \ + int t0,t1,t2,t3,p1,p2,p3,p4,p5,x0,x1,x2,x3; \ + p2 = s2; \ + p3 = s6; \ + p1 = (p2+p3) * stbi__f2f(0.5411961f); \ + t2 = p1 + p3*stbi__f2f(-1.847759065f); \ + t3 = p1 + p2*stbi__f2f( 0.765366865f); \ + p2 = s0; \ + p3 = s4; \ + t0 = stbi__fsh(p2+p3); \ + t1 = stbi__fsh(p2-p3); \ + x0 = t0+t3; \ + x3 = t0-t3; \ + x1 = t1+t2; \ + x2 = t1-t2; \ + t0 = s7; \ + t1 = s5; \ + t2 = s3; \ + t3 = s1; \ + p3 = t0+t2; \ + p4 = t1+t3; \ + p1 = t0+t3; \ + p2 = t1+t2; \ + p5 = (p3+p4)*stbi__f2f( 1.175875602f); \ + t0 = t0*stbi__f2f( 0.298631336f); \ + t1 = t1*stbi__f2f( 2.053119869f); \ + t2 = t2*stbi__f2f( 3.072711026f); \ + t3 = t3*stbi__f2f( 1.501321110f); \ + p1 = p5 + p1*stbi__f2f(-0.899976223f); \ + p2 = p5 + p2*stbi__f2f(-2.562915447f); \ + p3 = p3*stbi__f2f(-1.961570560f); \ + p4 = p4*stbi__f2f(-0.390180644f); \ + t3 += p1+p4; \ + t2 += p2+p3; \ + t1 += p2+p4; \ + t0 += p1+p3; + +static void stbi__idct_block(stbi_uc *out, int out_stride, short data[64]) +{ + int i,val[64],*v=val; + stbi_uc *o; + short *d = data; + + // columns + for (i=0; i < 8; ++i,++d, ++v) { + // if all zeroes, shortcut -- this avoids dequantizing 0s and IDCTing + if (d[ 8]==0 && d[16]==0 && d[24]==0 && d[32]==0 + && d[40]==0 && d[48]==0 && d[56]==0) { + // no shortcut 0 seconds + // (1|2|3|4|5|6|7)==0 0 seconds + // all separate -0.047 seconds + // 1 && 2|3 && 4|5 && 6|7: -0.047 seconds + int dcterm = d[0]*4; + v[0] = v[8] = v[16] = v[24] = v[32] = v[40] = v[48] = v[56] = dcterm; + } else { + STBI__IDCT_1D(d[ 0],d[ 8],d[16],d[24],d[32],d[40],d[48],d[56]) + // constants scaled things up by 1<<12; let's bring them back + // down, but keep 2 extra bits of precision + x0 += 512; x1 += 512; x2 += 512; x3 += 512; + v[ 0] = (x0+t3) >> 10; + v[56] = (x0-t3) >> 10; + v[ 8] = (x1+t2) >> 10; + v[48] = (x1-t2) >> 10; + v[16] = (x2+t1) >> 10; + v[40] = (x2-t1) >> 10; + v[24] = (x3+t0) >> 10; + v[32] = (x3-t0) >> 10; + } + } + + for (i=0, v=val, o=out; i < 8; ++i,v+=8,o+=out_stride) { + // no fast case since the first 1D IDCT spread components out + STBI__IDCT_1D(v[0],v[1],v[2],v[3],v[4],v[5],v[6],v[7]) + // constants scaled things up by 1<<12, plus we had 1<<2 from first + // loop, plus horizontal and vertical each scale by sqrt(8) so together + // we've got an extra 1<<3, so 1<<17 total we need to remove. + // so we want to round that, which means adding 0.5 * 1<<17, + // aka 65536. Also, we'll end up with -128 to 127 that we want + // to encode as 0..255 by adding 128, so we'll add that before the shift + x0 += 65536 + (128<<17); + x1 += 65536 + (128<<17); + x2 += 65536 + (128<<17); + x3 += 65536 + (128<<17); + // tried computing the shifts into temps, or'ing the temps to see + // if any were out of range, but that was slower + o[0] = stbi__clamp((x0+t3) >> 17); + o[7] = stbi__clamp((x0-t3) >> 17); + o[1] = stbi__clamp((x1+t2) >> 17); + o[6] = stbi__clamp((x1-t2) >> 17); + o[2] = stbi__clamp((x2+t1) >> 17); + o[5] = stbi__clamp((x2-t1) >> 17); + o[3] = stbi__clamp((x3+t0) >> 17); + o[4] = stbi__clamp((x3-t0) >> 17); + } +} + +#ifdef STBI_SSE2 +// sse2 integer IDCT. not the fastest possible implementation but it +// produces bit-identical results to the generic C version so it's +// fully "transparent". +static void stbi__idct_simd(stbi_uc *out, int out_stride, short data[64]) +{ + // This is constructed to match our regular (generic) integer IDCT exactly. + __m128i row0, row1, row2, row3, row4, row5, row6, row7; + __m128i tmp; + + // dot product constant: even elems=x, odd elems=y + #define dct_const(x,y) _mm_setr_epi16((x),(y),(x),(y),(x),(y),(x),(y)) + + // out(0) = c0[even]*x + c0[odd]*y (c0, x, y 16-bit, out 32-bit) + // out(1) = c1[even]*x + c1[odd]*y + #define dct_rot(out0,out1, x,y,c0,c1) \ + __m128i c0##lo = _mm_unpacklo_epi16((x),(y)); \ + __m128i c0##hi = _mm_unpackhi_epi16((x),(y)); \ + __m128i out0##_l = _mm_madd_epi16(c0##lo, c0); \ + __m128i out0##_h = _mm_madd_epi16(c0##hi, c0); \ + __m128i out1##_l = _mm_madd_epi16(c0##lo, c1); \ + __m128i out1##_h = _mm_madd_epi16(c0##hi, c1) + + // out = in << 12 (in 16-bit, out 32-bit) + #define dct_widen(out, in) \ + __m128i out##_l = _mm_srai_epi32(_mm_unpacklo_epi16(_mm_setzero_si128(), (in)), 4); \ + __m128i out##_h = _mm_srai_epi32(_mm_unpackhi_epi16(_mm_setzero_si128(), (in)), 4) + + // wide add + #define dct_wadd(out, a, b) \ + __m128i out##_l = _mm_add_epi32(a##_l, b##_l); \ + __m128i out##_h = _mm_add_epi32(a##_h, b##_h) + + // wide sub + #define dct_wsub(out, a, b) \ + __m128i out##_l = _mm_sub_epi32(a##_l, b##_l); \ + __m128i out##_h = _mm_sub_epi32(a##_h, b##_h) + + // butterfly a/b, add bias, then shift by "s" and pack + #define dct_bfly32o(out0, out1, a,b,bias,s) \ + { \ + __m128i abiased_l = _mm_add_epi32(a##_l, bias); \ + __m128i abiased_h = _mm_add_epi32(a##_h, bias); \ + dct_wadd(sum, abiased, b); \ + dct_wsub(dif, abiased, b); \ + out0 = _mm_packs_epi32(_mm_srai_epi32(sum_l, s), _mm_srai_epi32(sum_h, s)); \ + out1 = _mm_packs_epi32(_mm_srai_epi32(dif_l, s), _mm_srai_epi32(dif_h, s)); \ + } + + // 8-bit interleave step (for transposes) + #define dct_interleave8(a, b) \ + tmp = a; \ + a = _mm_unpacklo_epi8(a, b); \ + b = _mm_unpackhi_epi8(tmp, b) + + // 16-bit interleave step (for transposes) + #define dct_interleave16(a, b) \ + tmp = a; \ + a = _mm_unpacklo_epi16(a, b); \ + b = _mm_unpackhi_epi16(tmp, b) + + #define dct_pass(bias,shift) \ + { \ + /* even part */ \ + dct_rot(t2e,t3e, row2,row6, rot0_0,rot0_1); \ + __m128i sum04 = _mm_add_epi16(row0, row4); \ + __m128i dif04 = _mm_sub_epi16(row0, row4); \ + dct_widen(t0e, sum04); \ + dct_widen(t1e, dif04); \ + dct_wadd(x0, t0e, t3e); \ + dct_wsub(x3, t0e, t3e); \ + dct_wadd(x1, t1e, t2e); \ + dct_wsub(x2, t1e, t2e); \ + /* odd part */ \ + dct_rot(y0o,y2o, row7,row3, rot2_0,rot2_1); \ + dct_rot(y1o,y3o, row5,row1, rot3_0,rot3_1); \ + __m128i sum17 = _mm_add_epi16(row1, row7); \ + __m128i sum35 = _mm_add_epi16(row3, row5); \ + dct_rot(y4o,y5o, sum17,sum35, rot1_0,rot1_1); \ + dct_wadd(x4, y0o, y4o); \ + dct_wadd(x5, y1o, y5o); \ + dct_wadd(x6, y2o, y5o); \ + dct_wadd(x7, y3o, y4o); \ + dct_bfly32o(row0,row7, x0,x7,bias,shift); \ + dct_bfly32o(row1,row6, x1,x6,bias,shift); \ + dct_bfly32o(row2,row5, x2,x5,bias,shift); \ + dct_bfly32o(row3,row4, x3,x4,bias,shift); \ + } + + __m128i rot0_0 = dct_const(stbi__f2f(0.5411961f), stbi__f2f(0.5411961f) + stbi__f2f(-1.847759065f)); + __m128i rot0_1 = dct_const(stbi__f2f(0.5411961f) + stbi__f2f( 0.765366865f), stbi__f2f(0.5411961f)); + __m128i rot1_0 = dct_const(stbi__f2f(1.175875602f) + stbi__f2f(-0.899976223f), stbi__f2f(1.175875602f)); + __m128i rot1_1 = dct_const(stbi__f2f(1.175875602f), stbi__f2f(1.175875602f) + stbi__f2f(-2.562915447f)); + __m128i rot2_0 = dct_const(stbi__f2f(-1.961570560f) + stbi__f2f( 0.298631336f), stbi__f2f(-1.961570560f)); + __m128i rot2_1 = dct_const(stbi__f2f(-1.961570560f), stbi__f2f(-1.961570560f) + stbi__f2f( 3.072711026f)); + __m128i rot3_0 = dct_const(stbi__f2f(-0.390180644f) + stbi__f2f( 2.053119869f), stbi__f2f(-0.390180644f)); + __m128i rot3_1 = dct_const(stbi__f2f(-0.390180644f), stbi__f2f(-0.390180644f) + stbi__f2f( 1.501321110f)); + + // rounding biases in column/row passes, see stbi__idct_block for explanation. + __m128i bias_0 = _mm_set1_epi32(512); + __m128i bias_1 = _mm_set1_epi32(65536 + (128<<17)); + + // load + row0 = _mm_load_si128((const __m128i *) (data + 0*8)); + row1 = _mm_load_si128((const __m128i *) (data + 1*8)); + row2 = _mm_load_si128((const __m128i *) (data + 2*8)); + row3 = _mm_load_si128((const __m128i *) (data + 3*8)); + row4 = _mm_load_si128((const __m128i *) (data + 4*8)); + row5 = _mm_load_si128((const __m128i *) (data + 5*8)); + row6 = _mm_load_si128((const __m128i *) (data + 6*8)); + row7 = _mm_load_si128((const __m128i *) (data + 7*8)); + + // column pass + dct_pass(bias_0, 10); + + { + // 16bit 8x8 transpose pass 1 + dct_interleave16(row0, row4); + dct_interleave16(row1, row5); + dct_interleave16(row2, row6); + dct_interleave16(row3, row7); + + // transpose pass 2 + dct_interleave16(row0, row2); + dct_interleave16(row1, row3); + dct_interleave16(row4, row6); + dct_interleave16(row5, row7); + + // transpose pass 3 + dct_interleave16(row0, row1); + dct_interleave16(row2, row3); + dct_interleave16(row4, row5); + dct_interleave16(row6, row7); + } + + // row pass + dct_pass(bias_1, 17); + + { + // pack + __m128i p0 = _mm_packus_epi16(row0, row1); // a0a1a2a3...a7b0b1b2b3...b7 + __m128i p1 = _mm_packus_epi16(row2, row3); + __m128i p2 = _mm_packus_epi16(row4, row5); + __m128i p3 = _mm_packus_epi16(row6, row7); + + // 8bit 8x8 transpose pass 1 + dct_interleave8(p0, p2); // a0e0a1e1... + dct_interleave8(p1, p3); // c0g0c1g1... + + // transpose pass 2 + dct_interleave8(p0, p1); // a0c0e0g0... + dct_interleave8(p2, p3); // b0d0f0h0... + + // transpose pass 3 + dct_interleave8(p0, p2); // a0b0c0d0... + dct_interleave8(p1, p3); // a4b4c4d4... + + // store + _mm_storel_epi64((__m128i *) out, p0); out += out_stride; + _mm_storel_epi64((__m128i *) out, _mm_shuffle_epi32(p0, 0x4e)); out += out_stride; + _mm_storel_epi64((__m128i *) out, p2); out += out_stride; + _mm_storel_epi64((__m128i *) out, _mm_shuffle_epi32(p2, 0x4e)); out += out_stride; + _mm_storel_epi64((__m128i *) out, p1); out += out_stride; + _mm_storel_epi64((__m128i *) out, _mm_shuffle_epi32(p1, 0x4e)); out += out_stride; + _mm_storel_epi64((__m128i *) out, p3); out += out_stride; + _mm_storel_epi64((__m128i *) out, _mm_shuffle_epi32(p3, 0x4e)); + } + +#undef dct_const +#undef dct_rot +#undef dct_widen +#undef dct_wadd +#undef dct_wsub +#undef dct_bfly32o +#undef dct_interleave8 +#undef dct_interleave16 +#undef dct_pass +} + +#endif // STBI_SSE2 + +#ifdef STBI_NEON + +// NEON integer IDCT. should produce bit-identical +// results to the generic C version. +static void stbi__idct_simd(stbi_uc *out, int out_stride, short data[64]) +{ + int16x8_t row0, row1, row2, row3, row4, row5, row6, row7; + + int16x4_t rot0_0 = vdup_n_s16(stbi__f2f(0.5411961f)); + int16x4_t rot0_1 = vdup_n_s16(stbi__f2f(-1.847759065f)); + int16x4_t rot0_2 = vdup_n_s16(stbi__f2f( 0.765366865f)); + int16x4_t rot1_0 = vdup_n_s16(stbi__f2f( 1.175875602f)); + int16x4_t rot1_1 = vdup_n_s16(stbi__f2f(-0.899976223f)); + int16x4_t rot1_2 = vdup_n_s16(stbi__f2f(-2.562915447f)); + int16x4_t rot2_0 = vdup_n_s16(stbi__f2f(-1.961570560f)); + int16x4_t rot2_1 = vdup_n_s16(stbi__f2f(-0.390180644f)); + int16x4_t rot3_0 = vdup_n_s16(stbi__f2f( 0.298631336f)); + int16x4_t rot3_1 = vdup_n_s16(stbi__f2f( 2.053119869f)); + int16x4_t rot3_2 = vdup_n_s16(stbi__f2f( 3.072711026f)); + int16x4_t rot3_3 = vdup_n_s16(stbi__f2f( 1.501321110f)); + +#define dct_long_mul(out, inq, coeff) \ + int32x4_t out##_l = vmull_s16(vget_low_s16(inq), coeff); \ + int32x4_t out##_h = vmull_s16(vget_high_s16(inq), coeff) + +#define dct_long_mac(out, acc, inq, coeff) \ + int32x4_t out##_l = vmlal_s16(acc##_l, vget_low_s16(inq), coeff); \ + int32x4_t out##_h = vmlal_s16(acc##_h, vget_high_s16(inq), coeff) + +#define dct_widen(out, inq) \ + int32x4_t out##_l = vshll_n_s16(vget_low_s16(inq), 12); \ + int32x4_t out##_h = vshll_n_s16(vget_high_s16(inq), 12) + +// wide add +#define dct_wadd(out, a, b) \ + int32x4_t out##_l = vaddq_s32(a##_l, b##_l); \ + int32x4_t out##_h = vaddq_s32(a##_h, b##_h) + +// wide sub +#define dct_wsub(out, a, b) \ + int32x4_t out##_l = vsubq_s32(a##_l, b##_l); \ + int32x4_t out##_h = vsubq_s32(a##_h, b##_h) + +// butterfly a/b, then shift using "shiftop" by "s" and pack +#define dct_bfly32o(out0,out1, a,b,shiftop,s) \ + { \ + dct_wadd(sum, a, b); \ + dct_wsub(dif, a, b); \ + out0 = vcombine_s16(shiftop(sum_l, s), shiftop(sum_h, s)); \ + out1 = vcombine_s16(shiftop(dif_l, s), shiftop(dif_h, s)); \ + } + +#define dct_pass(shiftop, shift) \ + { \ + /* even part */ \ + int16x8_t sum26 = vaddq_s16(row2, row6); \ + dct_long_mul(p1e, sum26, rot0_0); \ + dct_long_mac(t2e, p1e, row6, rot0_1); \ + dct_long_mac(t3e, p1e, row2, rot0_2); \ + int16x8_t sum04 = vaddq_s16(row0, row4); \ + int16x8_t dif04 = vsubq_s16(row0, row4); \ + dct_widen(t0e, sum04); \ + dct_widen(t1e, dif04); \ + dct_wadd(x0, t0e, t3e); \ + dct_wsub(x3, t0e, t3e); \ + dct_wadd(x1, t1e, t2e); \ + dct_wsub(x2, t1e, t2e); \ + /* odd part */ \ + int16x8_t sum15 = vaddq_s16(row1, row5); \ + int16x8_t sum17 = vaddq_s16(row1, row7); \ + int16x8_t sum35 = vaddq_s16(row3, row5); \ + int16x8_t sum37 = vaddq_s16(row3, row7); \ + int16x8_t sumodd = vaddq_s16(sum17, sum35); \ + dct_long_mul(p5o, sumodd, rot1_0); \ + dct_long_mac(p1o, p5o, sum17, rot1_1); \ + dct_long_mac(p2o, p5o, sum35, rot1_2); \ + dct_long_mul(p3o, sum37, rot2_0); \ + dct_long_mul(p4o, sum15, rot2_1); \ + dct_wadd(sump13o, p1o, p3o); \ + dct_wadd(sump24o, p2o, p4o); \ + dct_wadd(sump23o, p2o, p3o); \ + dct_wadd(sump14o, p1o, p4o); \ + dct_long_mac(x4, sump13o, row7, rot3_0); \ + dct_long_mac(x5, sump24o, row5, rot3_1); \ + dct_long_mac(x6, sump23o, row3, rot3_2); \ + dct_long_mac(x7, sump14o, row1, rot3_3); \ + dct_bfly32o(row0,row7, x0,x7,shiftop,shift); \ + dct_bfly32o(row1,row6, x1,x6,shiftop,shift); \ + dct_bfly32o(row2,row5, x2,x5,shiftop,shift); \ + dct_bfly32o(row3,row4, x3,x4,shiftop,shift); \ + } + + // load + row0 = vld1q_s16(data + 0*8); + row1 = vld1q_s16(data + 1*8); + row2 = vld1q_s16(data + 2*8); + row3 = vld1q_s16(data + 3*8); + row4 = vld1q_s16(data + 4*8); + row5 = vld1q_s16(data + 5*8); + row6 = vld1q_s16(data + 6*8); + row7 = vld1q_s16(data + 7*8); + + // add DC bias + row0 = vaddq_s16(row0, vsetq_lane_s16(1024, vdupq_n_s16(0), 0)); + + // column pass + dct_pass(vrshrn_n_s32, 10); + + // 16bit 8x8 transpose + { +// these three map to a single VTRN.16, VTRN.32, and VSWP, respectively. +// whether compilers actually get this is another story, sadly. +#define dct_trn16(x, y) { int16x8x2_t t = vtrnq_s16(x, y); x = t.val[0]; y = t.val[1]; } +#define dct_trn32(x, y) { int32x4x2_t t = vtrnq_s32(vreinterpretq_s32_s16(x), vreinterpretq_s32_s16(y)); x = vreinterpretq_s16_s32(t.val[0]); y = vreinterpretq_s16_s32(t.val[1]); } +#define dct_trn64(x, y) { int16x8_t x0 = x; int16x8_t y0 = y; x = vcombine_s16(vget_low_s16(x0), vget_low_s16(y0)); y = vcombine_s16(vget_high_s16(x0), vget_high_s16(y0)); } + + // pass 1 + dct_trn16(row0, row1); // a0b0a2b2a4b4a6b6 + dct_trn16(row2, row3); + dct_trn16(row4, row5); + dct_trn16(row6, row7); + + // pass 2 + dct_trn32(row0, row2); // a0b0c0d0a4b4c4d4 + dct_trn32(row1, row3); + dct_trn32(row4, row6); + dct_trn32(row5, row7); + + // pass 3 + dct_trn64(row0, row4); // a0b0c0d0e0f0g0h0 + dct_trn64(row1, row5); + dct_trn64(row2, row6); + dct_trn64(row3, row7); + +#undef dct_trn16 +#undef dct_trn32 +#undef dct_trn64 + } + + // row pass + // vrshrn_n_s32 only supports shifts up to 16, we need + // 17. so do a non-rounding shift of 16 first then follow + // up with a rounding shift by 1. + dct_pass(vshrn_n_s32, 16); + + { + // pack and round + uint8x8_t p0 = vqrshrun_n_s16(row0, 1); + uint8x8_t p1 = vqrshrun_n_s16(row1, 1); + uint8x8_t p2 = vqrshrun_n_s16(row2, 1); + uint8x8_t p3 = vqrshrun_n_s16(row3, 1); + uint8x8_t p4 = vqrshrun_n_s16(row4, 1); + uint8x8_t p5 = vqrshrun_n_s16(row5, 1); + uint8x8_t p6 = vqrshrun_n_s16(row6, 1); + uint8x8_t p7 = vqrshrun_n_s16(row7, 1); + + // again, these can translate into one instruction, but often don't. +#define dct_trn8_8(x, y) { uint8x8x2_t t = vtrn_u8(x, y); x = t.val[0]; y = t.val[1]; } +#define dct_trn8_16(x, y) { uint16x4x2_t t = vtrn_u16(vreinterpret_u16_u8(x), vreinterpret_u16_u8(y)); x = vreinterpret_u8_u16(t.val[0]); y = vreinterpret_u8_u16(t.val[1]); } +#define dct_trn8_32(x, y) { uint32x2x2_t t = vtrn_u32(vreinterpret_u32_u8(x), vreinterpret_u32_u8(y)); x = vreinterpret_u8_u32(t.val[0]); y = vreinterpret_u8_u32(t.val[1]); } + + // sadly can't use interleaved stores here since we only write + // 8 bytes to each scan line! + + // 8x8 8-bit transpose pass 1 + dct_trn8_8(p0, p1); + dct_trn8_8(p2, p3); + dct_trn8_8(p4, p5); + dct_trn8_8(p6, p7); + + // pass 2 + dct_trn8_16(p0, p2); + dct_trn8_16(p1, p3); + dct_trn8_16(p4, p6); + dct_trn8_16(p5, p7); + + // pass 3 + dct_trn8_32(p0, p4); + dct_trn8_32(p1, p5); + dct_trn8_32(p2, p6); + dct_trn8_32(p3, p7); + + // store + vst1_u8(out, p0); out += out_stride; + vst1_u8(out, p1); out += out_stride; + vst1_u8(out, p2); out += out_stride; + vst1_u8(out, p3); out += out_stride; + vst1_u8(out, p4); out += out_stride; + vst1_u8(out, p5); out += out_stride; + vst1_u8(out, p6); out += out_stride; + vst1_u8(out, p7); + +#undef dct_trn8_8 +#undef dct_trn8_16 +#undef dct_trn8_32 + } + +#undef dct_long_mul +#undef dct_long_mac +#undef dct_widen +#undef dct_wadd +#undef dct_wsub +#undef dct_bfly32o +#undef dct_pass +} + +#endif // STBI_NEON + +#define STBI__MARKER_none 0xff +// if there's a pending marker from the entropy stream, return that +// otherwise, fetch from the stream and get a marker. if there's no +// marker, return 0xff, which is never a valid marker value +static stbi_uc stbi__get_marker(stbi__jpeg *j) +{ + stbi_uc x; + if (j->marker != STBI__MARKER_none) { x = j->marker; j->marker = STBI__MARKER_none; return x; } + x = stbi__get8(j->s); + if (x != 0xff) return STBI__MARKER_none; + while (x == 0xff) + x = stbi__get8(j->s); // consume repeated 0xff fill bytes + return x; +} + +// in each scan, we'll have scan_n components, and the order +// of the components is specified by order[] +#define STBI__RESTART(x) ((x) >= 0xd0 && (x) <= 0xd7) + +// after a restart interval, stbi__jpeg_reset the entropy decoder and +// the dc prediction +static void stbi__jpeg_reset(stbi__jpeg *j) +{ + j->code_bits = 0; + j->code_buffer = 0; + j->nomore = 0; + j->img_comp[0].dc_pred = j->img_comp[1].dc_pred = j->img_comp[2].dc_pred = j->img_comp[3].dc_pred = 0; + j->marker = STBI__MARKER_none; + j->todo = j->restart_interval ? j->restart_interval : 0x7fffffff; + j->eob_run = 0; + // no more than 1<<31 MCUs if no restart_interal? that's plenty safe, + // since we don't even allow 1<<30 pixels +} + +static int stbi__parse_entropy_coded_data(stbi__jpeg *z) +{ + stbi__jpeg_reset(z); + if (!z->progressive) { + if (z->scan_n == 1) { + int i,j; + STBI_SIMD_ALIGN(short, data[64]); + int n = z->order[0]; + // non-interleaved data, we just need to process one block at a time, + // in trivial scanline order + // number of blocks to do just depends on how many actual "pixels" this + // component has, independent of interleaved MCU blocking and such + int w = (z->img_comp[n].x+7) >> 3; + int h = (z->img_comp[n].y+7) >> 3; + for (j=0; j < h; ++j) { + for (i=0; i < w; ++i) { + int ha = z->img_comp[n].ha; + if (!stbi__jpeg_decode_block(z, data, z->huff_dc+z->img_comp[n].hd, z->huff_ac+ha, z->fast_ac[ha], n, z->dequant[z->img_comp[n].tq])) return 0; + z->idct_block_kernel(z->img_comp[n].data+z->img_comp[n].w2*j*8+i*8, z->img_comp[n].w2, data); + // every data block is an MCU, so countdown the restart interval + if (--z->todo <= 0) { + if (z->code_bits < 24) stbi__grow_buffer_unsafe(z); + // if it's NOT a restart, then just bail, so we get corrupt data + // rather than no data + if (!STBI__RESTART(z->marker)) return 1; + stbi__jpeg_reset(z); + } + } + } + return 1; + } else { // interleaved + int i,j,k,x,y; + STBI_SIMD_ALIGN(short, data[64]); + for (j=0; j < z->img_mcu_y; ++j) { + for (i=0; i < z->img_mcu_x; ++i) { + // scan an interleaved mcu... process scan_n components in order + for (k=0; k < z->scan_n; ++k) { + int n = z->order[k]; + // scan out an mcu's worth of this component; that's just determined + // by the basic H and V specified for the component + for (y=0; y < z->img_comp[n].v; ++y) { + for (x=0; x < z->img_comp[n].h; ++x) { + int x2 = (i*z->img_comp[n].h + x)*8; + int y2 = (j*z->img_comp[n].v + y)*8; + int ha = z->img_comp[n].ha; + if (!stbi__jpeg_decode_block(z, data, z->huff_dc+z->img_comp[n].hd, z->huff_ac+ha, z->fast_ac[ha], n, z->dequant[z->img_comp[n].tq])) return 0; + z->idct_block_kernel(z->img_comp[n].data+z->img_comp[n].w2*y2+x2, z->img_comp[n].w2, data); + } + } + } + // after all interleaved components, that's an interleaved MCU, + // so now count down the restart interval + if (--z->todo <= 0) { + if (z->code_bits < 24) stbi__grow_buffer_unsafe(z); + if (!STBI__RESTART(z->marker)) return 1; + stbi__jpeg_reset(z); + } + } + } + return 1; + } + } else { + if (z->scan_n == 1) { + int i,j; + int n = z->order[0]; + // non-interleaved data, we just need to process one block at a time, + // in trivial scanline order + // number of blocks to do just depends on how many actual "pixels" this + // component has, independent of interleaved MCU blocking and such + int w = (z->img_comp[n].x+7) >> 3; + int h = (z->img_comp[n].y+7) >> 3; + for (j=0; j < h; ++j) { + for (i=0; i < w; ++i) { + short *data = z->img_comp[n].coeff + 64 * (i + j * z->img_comp[n].coeff_w); + if (z->spec_start == 0) { + if (!stbi__jpeg_decode_block_prog_dc(z, data, &z->huff_dc[z->img_comp[n].hd], n)) + return 0; + } else { + int ha = z->img_comp[n].ha; + if (!stbi__jpeg_decode_block_prog_ac(z, data, &z->huff_ac[ha], z->fast_ac[ha])) + return 0; + } + // every data block is an MCU, so countdown the restart interval + if (--z->todo <= 0) { + if (z->code_bits < 24) stbi__grow_buffer_unsafe(z); + if (!STBI__RESTART(z->marker)) return 1; + stbi__jpeg_reset(z); + } + } + } + return 1; + } else { // interleaved + int i,j,k,x,y; + for (j=0; j < z->img_mcu_y; ++j) { + for (i=0; i < z->img_mcu_x; ++i) { + // scan an interleaved mcu... process scan_n components in order + for (k=0; k < z->scan_n; ++k) { + int n = z->order[k]; + // scan out an mcu's worth of this component; that's just determined + // by the basic H and V specified for the component + for (y=0; y < z->img_comp[n].v; ++y) { + for (x=0; x < z->img_comp[n].h; ++x) { + int x2 = (i*z->img_comp[n].h + x); + int y2 = (j*z->img_comp[n].v + y); + short *data = z->img_comp[n].coeff + 64 * (x2 + y2 * z->img_comp[n].coeff_w); + if (!stbi__jpeg_decode_block_prog_dc(z, data, &z->huff_dc[z->img_comp[n].hd], n)) + return 0; + } + } + } + // after all interleaved components, that's an interleaved MCU, + // so now count down the restart interval + if (--z->todo <= 0) { + if (z->code_bits < 24) stbi__grow_buffer_unsafe(z); + if (!STBI__RESTART(z->marker)) return 1; + stbi__jpeg_reset(z); + } + } + } + return 1; + } + } +} + +static void stbi__jpeg_dequantize(short *data, stbi__uint16 *dequant) +{ + int i; + for (i=0; i < 64; ++i) + data[i] *= dequant[i]; +} + +static void stbi__jpeg_finish(stbi__jpeg *z) +{ + if (z->progressive) { + // dequantize and idct the data + int i,j,n; + for (n=0; n < z->s->img_n; ++n) { + int w = (z->img_comp[n].x+7) >> 3; + int h = (z->img_comp[n].y+7) >> 3; + for (j=0; j < h; ++j) { + for (i=0; i < w; ++i) { + short *data = z->img_comp[n].coeff + 64 * (i + j * z->img_comp[n].coeff_w); + stbi__jpeg_dequantize(data, z->dequant[z->img_comp[n].tq]); + z->idct_block_kernel(z->img_comp[n].data+z->img_comp[n].w2*j*8+i*8, z->img_comp[n].w2, data); + } + } + } + } +} + +static int stbi__process_marker(stbi__jpeg *z, int m) +{ + int L; + switch (m) { + case STBI__MARKER_none: // no marker found + return stbi__err("expected marker","Corrupt JPEG"); + + case 0xDD: // DRI - specify restart interval + if (stbi__get16be(z->s) != 4) return stbi__err("bad DRI len","Corrupt JPEG"); + z->restart_interval = stbi__get16be(z->s); + return 1; + + case 0xDB: // DQT - define quantization table + L = stbi__get16be(z->s)-2; + while (L > 0) { + int q = stbi__get8(z->s); + int p = q >> 4, sixteen = (p != 0); + int t = q & 15,i; + if (p != 0 && p != 1) return stbi__err("bad DQT type","Corrupt JPEG"); + if (t > 3) return stbi__err("bad DQT table","Corrupt JPEG"); + + for (i=0; i < 64; ++i) + z->dequant[t][stbi__jpeg_dezigzag[i]] = (stbi__uint16)(sixteen ? stbi__get16be(z->s) : stbi__get8(z->s)); + L -= (sixteen ? 129 : 65); + } + return L==0; + + case 0xC4: // DHT - define huffman table + L = stbi__get16be(z->s)-2; + while (L > 0) { + stbi_uc *v; + int sizes[16],i,n=0; + int q = stbi__get8(z->s); + int tc = q >> 4; + int th = q & 15; + if (tc > 1 || th > 3) return stbi__err("bad DHT header","Corrupt JPEG"); + for (i=0; i < 16; ++i) { + sizes[i] = stbi__get8(z->s); + n += sizes[i]; + } + if(n > 256) return stbi__err("bad DHT header","Corrupt JPEG"); // Loop over i < n would write past end of values! + L -= 17; + if (tc == 0) { + if (!stbi__build_huffman(z->huff_dc+th, sizes)) return 0; + v = z->huff_dc[th].values; + } else { + if (!stbi__build_huffman(z->huff_ac+th, sizes)) return 0; + v = z->huff_ac[th].values; + } + for (i=0; i < n; ++i) + v[i] = stbi__get8(z->s); + if (tc != 0) + stbi__build_fast_ac(z->fast_ac[th], z->huff_ac + th); + L -= n; + } + return L==0; + } + + // check for comment block or APP blocks + if ((m >= 0xE0 && m <= 0xEF) || m == 0xFE) { + L = stbi__get16be(z->s); + if (L < 2) { + if (m == 0xFE) + return stbi__err("bad COM len","Corrupt JPEG"); + else + return stbi__err("bad APP len","Corrupt JPEG"); + } + L -= 2; + + if (m == 0xE0 && L >= 5) { // JFIF APP0 segment + static const unsigned char tag[5] = {'J','F','I','F','\0'}; + int ok = 1; + int i; + for (i=0; i < 5; ++i) + if (stbi__get8(z->s) != tag[i]) + ok = 0; + L -= 5; + if (ok) + z->jfif = 1; + } else if (m == 0xEE && L >= 12) { // Adobe APP14 segment + static const unsigned char tag[6] = {'A','d','o','b','e','\0'}; + int ok = 1; + int i; + for (i=0; i < 6; ++i) + if (stbi__get8(z->s) != tag[i]) + ok = 0; + L -= 6; + if (ok) { + stbi__get8(z->s); // version + stbi__get16be(z->s); // flags0 + stbi__get16be(z->s); // flags1 + z->app14_color_transform = stbi__get8(z->s); // color transform + L -= 6; + } + } + + stbi__skip(z->s, L); + return 1; + } + + return stbi__err("unknown marker","Corrupt JPEG"); +} + +// after we see SOS +static int stbi__process_scan_header(stbi__jpeg *z) +{ + int i; + int Ls = stbi__get16be(z->s); + z->scan_n = stbi__get8(z->s); + if (z->scan_n < 1 || z->scan_n > 4 || z->scan_n > (int) z->s->img_n) return stbi__err("bad SOS component count","Corrupt JPEG"); + if (Ls != 6+2*z->scan_n) return stbi__err("bad SOS len","Corrupt JPEG"); + for (i=0; i < z->scan_n; ++i) { + int id = stbi__get8(z->s), which; + int q = stbi__get8(z->s); + for (which = 0; which < z->s->img_n; ++which) + if (z->img_comp[which].id == id) + break; + if (which == z->s->img_n) return 0; // no match + z->img_comp[which].hd = q >> 4; if (z->img_comp[which].hd > 3) return stbi__err("bad DC huff","Corrupt JPEG"); + z->img_comp[which].ha = q & 15; if (z->img_comp[which].ha > 3) return stbi__err("bad AC huff","Corrupt JPEG"); + z->order[i] = which; + } + + { + int aa; + z->spec_start = stbi__get8(z->s); + z->spec_end = stbi__get8(z->s); // should be 63, but might be 0 + aa = stbi__get8(z->s); + z->succ_high = (aa >> 4); + z->succ_low = (aa & 15); + if (z->progressive) { + if (z->spec_start > 63 || z->spec_end > 63 || z->spec_start > z->spec_end || z->succ_high > 13 || z->succ_low > 13) + return stbi__err("bad SOS", "Corrupt JPEG"); + } else { + if (z->spec_start != 0) return stbi__err("bad SOS","Corrupt JPEG"); + if (z->succ_high != 0 || z->succ_low != 0) return stbi__err("bad SOS","Corrupt JPEG"); + z->spec_end = 63; + } + } + + return 1; +} + +static int stbi__free_jpeg_components(stbi__jpeg *z, int ncomp, int why) +{ + int i; + for (i=0; i < ncomp; ++i) { + if (z->img_comp[i].raw_data) { + STBI_FREE(z->img_comp[i].raw_data); + z->img_comp[i].raw_data = NULL; + z->img_comp[i].data = NULL; + } + if (z->img_comp[i].raw_coeff) { + STBI_FREE(z->img_comp[i].raw_coeff); + z->img_comp[i].raw_coeff = 0; + z->img_comp[i].coeff = 0; + } + if (z->img_comp[i].linebuf) { + STBI_FREE(z->img_comp[i].linebuf); + z->img_comp[i].linebuf = NULL; + } + } + return why; +} + +static int stbi__process_frame_header(stbi__jpeg *z, int scan) +{ + stbi__context *s = z->s; + int Lf,p,i,q, h_max=1,v_max=1,c; + Lf = stbi__get16be(s); if (Lf < 11) return stbi__err("bad SOF len","Corrupt JPEG"); // JPEG + p = stbi__get8(s); if (p != 8) return stbi__err("only 8-bit","JPEG format not supported: 8-bit only"); // JPEG baseline + s->img_y = stbi__get16be(s); if (s->img_y == 0) return stbi__err("no header height", "JPEG format not supported: delayed height"); // Legal, but we don't handle it--but neither does IJG + s->img_x = stbi__get16be(s); if (s->img_x == 0) return stbi__err("0 width","Corrupt JPEG"); // JPEG requires + if (s->img_y > STBI_MAX_DIMENSIONS) return stbi__err("too large","Very large image (corrupt?)"); + if (s->img_x > STBI_MAX_DIMENSIONS) return stbi__err("too large","Very large image (corrupt?)"); + c = stbi__get8(s); + if (c != 3 && c != 1 && c != 4) return stbi__err("bad component count","Corrupt JPEG"); + s->img_n = c; + for (i=0; i < c; ++i) { + z->img_comp[i].data = NULL; + z->img_comp[i].linebuf = NULL; + } + + if (Lf != 8+3*s->img_n) return stbi__err("bad SOF len","Corrupt JPEG"); + + z->rgb = 0; + for (i=0; i < s->img_n; ++i) { + static const unsigned char rgb[3] = { 'R', 'G', 'B' }; + z->img_comp[i].id = stbi__get8(s); + if (s->img_n == 3 && z->img_comp[i].id == rgb[i]) + ++z->rgb; + q = stbi__get8(s); + z->img_comp[i].h = (q >> 4); if (!z->img_comp[i].h || z->img_comp[i].h > 4) return stbi__err("bad H","Corrupt JPEG"); + z->img_comp[i].v = q & 15; if (!z->img_comp[i].v || z->img_comp[i].v > 4) return stbi__err("bad V","Corrupt JPEG"); + z->img_comp[i].tq = stbi__get8(s); if (z->img_comp[i].tq > 3) return stbi__err("bad TQ","Corrupt JPEG"); + } + + if (scan != STBI__SCAN_load) return 1; + + if (!stbi__mad3sizes_valid(s->img_x, s->img_y, s->img_n, 0)) return stbi__err("too large", "Image too large to decode"); + + for (i=0; i < s->img_n; ++i) { + if (z->img_comp[i].h > h_max) h_max = z->img_comp[i].h; + if (z->img_comp[i].v > v_max) v_max = z->img_comp[i].v; + } + + // check that plane subsampling factors are integer ratios; our resamplers can't deal with fractional ratios + // and I've never seen a non-corrupted JPEG file actually use them + for (i=0; i < s->img_n; ++i) { + if (h_max % z->img_comp[i].h != 0) return stbi__err("bad H","Corrupt JPEG"); + if (v_max % z->img_comp[i].v != 0) return stbi__err("bad V","Corrupt JPEG"); + } + + // compute interleaved mcu info + z->img_h_max = h_max; + z->img_v_max = v_max; + z->img_mcu_w = h_max * 8; + z->img_mcu_h = v_max * 8; + // these sizes can't be more than 17 bits + z->img_mcu_x = (s->img_x + z->img_mcu_w-1) / z->img_mcu_w; + z->img_mcu_y = (s->img_y + z->img_mcu_h-1) / z->img_mcu_h; + + for (i=0; i < s->img_n; ++i) { + // number of effective pixels (e.g. for non-interleaved MCU) + z->img_comp[i].x = (s->img_x * z->img_comp[i].h + h_max-1) / h_max; + z->img_comp[i].y = (s->img_y * z->img_comp[i].v + v_max-1) / v_max; + // to simplify generation, we'll allocate enough memory to decode + // the bogus oversized data from using interleaved MCUs and their + // big blocks (e.g. a 16x16 iMCU on an image of width 33); we won't + // discard the extra data until colorspace conversion + // + // img_mcu_x, img_mcu_y: <=17 bits; comp[i].h and .v are <=4 (checked earlier) + // so these muls can't overflow with 32-bit ints (which we require) + z->img_comp[i].w2 = z->img_mcu_x * z->img_comp[i].h * 8; + z->img_comp[i].h2 = z->img_mcu_y * z->img_comp[i].v * 8; + z->img_comp[i].coeff = 0; + z->img_comp[i].raw_coeff = 0; + z->img_comp[i].linebuf = NULL; + z->img_comp[i].raw_data = stbi__malloc_mad2(z->img_comp[i].w2, z->img_comp[i].h2, 15); + if (z->img_comp[i].raw_data == NULL) + return stbi__free_jpeg_components(z, i+1, stbi__err("outofmem", "Out of memory")); + // align blocks for idct using mmx/sse + z->img_comp[i].data = (stbi_uc*) (((size_t) z->img_comp[i].raw_data + 15) & ~15); + if (z->progressive) { + // w2, h2 are multiples of 8 (see above) + z->img_comp[i].coeff_w = z->img_comp[i].w2 / 8; + z->img_comp[i].coeff_h = z->img_comp[i].h2 / 8; + z->img_comp[i].raw_coeff = stbi__malloc_mad3(z->img_comp[i].w2, z->img_comp[i].h2, sizeof(short), 15); + if (z->img_comp[i].raw_coeff == NULL) + return stbi__free_jpeg_components(z, i+1, stbi__err("outofmem", "Out of memory")); + z->img_comp[i].coeff = (short*) (((size_t) z->img_comp[i].raw_coeff + 15) & ~15); + } + } + + return 1; +} + +// use comparisons since in some cases we handle more than one case (e.g. SOF) +#define stbi__DNL(x) ((x) == 0xdc) +#define stbi__SOI(x) ((x) == 0xd8) +#define stbi__EOI(x) ((x) == 0xd9) +#define stbi__SOF(x) ((x) == 0xc0 || (x) == 0xc1 || (x) == 0xc2) +#define stbi__SOS(x) ((x) == 0xda) + +#define stbi__SOF_progressive(x) ((x) == 0xc2) + +static int stbi__decode_jpeg_header(stbi__jpeg *z, int scan) +{ + int m; + z->jfif = 0; + z->app14_color_transform = -1; // valid values are 0,1,2 + z->marker = STBI__MARKER_none; // initialize cached marker to empty + m = stbi__get_marker(z); + if (!stbi__SOI(m)) return stbi__err("no SOI","Corrupt JPEG"); + if (scan == STBI__SCAN_type) return 1; + m = stbi__get_marker(z); + while (!stbi__SOF(m)) { + if (!stbi__process_marker(z,m)) return 0; + m = stbi__get_marker(z); + while (m == STBI__MARKER_none) { + // some files have extra padding after their blocks, so ok, we'll scan + if (stbi__at_eof(z->s)) return stbi__err("no SOF", "Corrupt JPEG"); + m = stbi__get_marker(z); + } + } + z->progressive = stbi__SOF_progressive(m); + if (!stbi__process_frame_header(z, scan)) return 0; + return 1; +} + +static stbi_uc stbi__skip_jpeg_junk_at_end(stbi__jpeg *j) +{ + // some JPEGs have junk at end, skip over it but if we find what looks + // like a valid marker, resume there + while (!stbi__at_eof(j->s)) { + stbi_uc x = stbi__get8(j->s); + while (x == 0xff) { // might be a marker + if (stbi__at_eof(j->s)) return STBI__MARKER_none; + x = stbi__get8(j->s); + if (x != 0x00 && x != 0xff) { + // not a stuffed zero or lead-in to another marker, looks + // like an actual marker, return it + return x; + } + // stuffed zero has x=0 now which ends the loop, meaning we go + // back to regular scan loop. + // repeated 0xff keeps trying to read the next byte of the marker. + } + } + return STBI__MARKER_none; +} + +// decode image to YCbCr format +static int stbi__decode_jpeg_image(stbi__jpeg *j) +{ + int m; + for (m = 0; m < 4; m++) { + j->img_comp[m].raw_data = NULL; + j->img_comp[m].raw_coeff = NULL; + } + j->restart_interval = 0; + if (!stbi__decode_jpeg_header(j, STBI__SCAN_load)) return 0; + m = stbi__get_marker(j); + while (!stbi__EOI(m)) { + if (stbi__SOS(m)) { + if (!stbi__process_scan_header(j)) return 0; + if (!stbi__parse_entropy_coded_data(j)) return 0; + if (j->marker == STBI__MARKER_none ) { + j->marker = stbi__skip_jpeg_junk_at_end(j); + // if we reach eof without hitting a marker, stbi__get_marker() below will fail and we'll eventually return 0 + } + m = stbi__get_marker(j); + if (STBI__RESTART(m)) + m = stbi__get_marker(j); + } else if (stbi__DNL(m)) { + int Ld = stbi__get16be(j->s); + stbi__uint32 NL = stbi__get16be(j->s); + if (Ld != 4) return stbi__err("bad DNL len", "Corrupt JPEG"); + if (NL != j->s->img_y) return stbi__err("bad DNL height", "Corrupt JPEG"); + m = stbi__get_marker(j); + } else { + if (!stbi__process_marker(j, m)) return 1; + m = stbi__get_marker(j); + } + } + if (j->progressive) + stbi__jpeg_finish(j); + return 1; +} + +// static jfif-centered resampling (across block boundaries) + +typedef stbi_uc *(*resample_row_func)(stbi_uc *out, stbi_uc *in0, stbi_uc *in1, + int w, int hs); + +#define stbi__div4(x) ((stbi_uc) ((x) >> 2)) + +static stbi_uc *resample_row_1(stbi_uc *out, stbi_uc *in_near, stbi_uc *in_far, int w, int hs) +{ + STBI_NOTUSED(out); + STBI_NOTUSED(in_far); + STBI_NOTUSED(w); + STBI_NOTUSED(hs); + return in_near; +} + +static stbi_uc* stbi__resample_row_v_2(stbi_uc *out, stbi_uc *in_near, stbi_uc *in_far, int w, int hs) +{ + // need to generate two samples vertically for every one in input + int i; + STBI_NOTUSED(hs); + for (i=0; i < w; ++i) + out[i] = stbi__div4(3*in_near[i] + in_far[i] + 2); + return out; +} + +static stbi_uc* stbi__resample_row_h_2(stbi_uc *out, stbi_uc *in_near, stbi_uc *in_far, int w, int hs) +{ + // need to generate two samples horizontally for every one in input + int i; + stbi_uc *input = in_near; + + if (w == 1) { + // if only one sample, can't do any interpolation + out[0] = out[1] = input[0]; + return out; + } + + out[0] = input[0]; + out[1] = stbi__div4(input[0]*3 + input[1] + 2); + for (i=1; i < w-1; ++i) { + int n = 3*input[i]+2; + out[i*2+0] = stbi__div4(n+input[i-1]); + out[i*2+1] = stbi__div4(n+input[i+1]); + } + out[i*2+0] = stbi__div4(input[w-2]*3 + input[w-1] + 2); + out[i*2+1] = input[w-1]; + + STBI_NOTUSED(in_far); + STBI_NOTUSED(hs); + + return out; +} + +#define stbi__div16(x) ((stbi_uc) ((x) >> 4)) + +static stbi_uc *stbi__resample_row_hv_2(stbi_uc *out, stbi_uc *in_near, stbi_uc *in_far, int w, int hs) +{ + // need to generate 2x2 samples for every one in input + int i,t0,t1; + if (w == 1) { + out[0] = out[1] = stbi__div4(3*in_near[0] + in_far[0] + 2); + return out; + } + + t1 = 3*in_near[0] + in_far[0]; + out[0] = stbi__div4(t1+2); + for (i=1; i < w; ++i) { + t0 = t1; + t1 = 3*in_near[i]+in_far[i]; + out[i*2-1] = stbi__div16(3*t0 + t1 + 8); + out[i*2 ] = stbi__div16(3*t1 + t0 + 8); + } + out[w*2-1] = stbi__div4(t1+2); + + STBI_NOTUSED(hs); + + return out; +} + +#if defined(STBI_SSE2) || defined(STBI_NEON) +static stbi_uc *stbi__resample_row_hv_2_simd(stbi_uc *out, stbi_uc *in_near, stbi_uc *in_far, int w, int hs) +{ + // need to generate 2x2 samples for every one in input + int i=0,t0,t1; + + if (w == 1) { + out[0] = out[1] = stbi__div4(3*in_near[0] + in_far[0] + 2); + return out; + } + + t1 = 3*in_near[0] + in_far[0]; + // process groups of 8 pixels for as long as we can. + // note we can't handle the last pixel in a row in this loop + // because we need to handle the filter boundary conditions. + for (; i < ((w-1) & ~7); i += 8) { +#if defined(STBI_SSE2) + // load and perform the vertical filtering pass + // this uses 3*x + y = 4*x + (y - x) + __m128i zero = _mm_setzero_si128(); + __m128i farb = _mm_loadl_epi64((__m128i *) (in_far + i)); + __m128i nearb = _mm_loadl_epi64((__m128i *) (in_near + i)); + __m128i farw = _mm_unpacklo_epi8(farb, zero); + __m128i nearw = _mm_unpacklo_epi8(nearb, zero); + __m128i diff = _mm_sub_epi16(farw, nearw); + __m128i nears = _mm_slli_epi16(nearw, 2); + __m128i curr = _mm_add_epi16(nears, diff); // current row + + // horizontal filter works the same based on shifted vers of current + // row. "prev" is current row shifted right by 1 pixel; we need to + // insert the previous pixel value (from t1). + // "next" is current row shifted left by 1 pixel, with first pixel + // of next block of 8 pixels added in. + __m128i prv0 = _mm_slli_si128(curr, 2); + __m128i nxt0 = _mm_srli_si128(curr, 2); + __m128i prev = _mm_insert_epi16(prv0, t1, 0); + __m128i next = _mm_insert_epi16(nxt0, 3*in_near[i+8] + in_far[i+8], 7); + + // horizontal filter, polyphase implementation since it's convenient: + // even pixels = 3*cur + prev = cur*4 + (prev - cur) + // odd pixels = 3*cur + next = cur*4 + (next - cur) + // note the shared term. + __m128i bias = _mm_set1_epi16(8); + __m128i curs = _mm_slli_epi16(curr, 2); + __m128i prvd = _mm_sub_epi16(prev, curr); + __m128i nxtd = _mm_sub_epi16(next, curr); + __m128i curb = _mm_add_epi16(curs, bias); + __m128i even = _mm_add_epi16(prvd, curb); + __m128i odd = _mm_add_epi16(nxtd, curb); + + // interleave even and odd pixels, then undo scaling. + __m128i int0 = _mm_unpacklo_epi16(even, odd); + __m128i int1 = _mm_unpackhi_epi16(even, odd); + __m128i de0 = _mm_srli_epi16(int0, 4); + __m128i de1 = _mm_srli_epi16(int1, 4); + + // pack and write output + __m128i outv = _mm_packus_epi16(de0, de1); + _mm_storeu_si128((__m128i *) (out + i*2), outv); +#elif defined(STBI_NEON) + // load and perform the vertical filtering pass + // this uses 3*x + y = 4*x + (y - x) + uint8x8_t farb = vld1_u8(in_far + i); + uint8x8_t nearb = vld1_u8(in_near + i); + int16x8_t diff = vreinterpretq_s16_u16(vsubl_u8(farb, nearb)); + int16x8_t nears = vreinterpretq_s16_u16(vshll_n_u8(nearb, 2)); + int16x8_t curr = vaddq_s16(nears, diff); // current row + + // horizontal filter works the same based on shifted vers of current + // row. "prev" is current row shifted right by 1 pixel; we need to + // insert the previous pixel value (from t1). + // "next" is current row shifted left by 1 pixel, with first pixel + // of next block of 8 pixels added in. + int16x8_t prv0 = vextq_s16(curr, curr, 7); + int16x8_t nxt0 = vextq_s16(curr, curr, 1); + int16x8_t prev = vsetq_lane_s16(t1, prv0, 0); + int16x8_t next = vsetq_lane_s16(3*in_near[i+8] + in_far[i+8], nxt0, 7); + + // horizontal filter, polyphase implementation since it's convenient: + // even pixels = 3*cur + prev = cur*4 + (prev - cur) + // odd pixels = 3*cur + next = cur*4 + (next - cur) + // note the shared term. + int16x8_t curs = vshlq_n_s16(curr, 2); + int16x8_t prvd = vsubq_s16(prev, curr); + int16x8_t nxtd = vsubq_s16(next, curr); + int16x8_t even = vaddq_s16(curs, prvd); + int16x8_t odd = vaddq_s16(curs, nxtd); + + // undo scaling and round, then store with even/odd phases interleaved + uint8x8x2_t o; + o.val[0] = vqrshrun_n_s16(even, 4); + o.val[1] = vqrshrun_n_s16(odd, 4); + vst2_u8(out + i*2, o); +#endif + + // "previous" value for next iter + t1 = 3*in_near[i+7] + in_far[i+7]; + } + + t0 = t1; + t1 = 3*in_near[i] + in_far[i]; + out[i*2] = stbi__div16(3*t1 + t0 + 8); + + for (++i; i < w; ++i) { + t0 = t1; + t1 = 3*in_near[i]+in_far[i]; + out[i*2-1] = stbi__div16(3*t0 + t1 + 8); + out[i*2 ] = stbi__div16(3*t1 + t0 + 8); + } + out[w*2-1] = stbi__div4(t1+2); + + STBI_NOTUSED(hs); + + return out; +} +#endif + +static stbi_uc *stbi__resample_row_generic(stbi_uc *out, stbi_uc *in_near, stbi_uc *in_far, int w, int hs) +{ + // resample with nearest-neighbor + int i,j; + STBI_NOTUSED(in_far); + for (i=0; i < w; ++i) + for (j=0; j < hs; ++j) + out[i*hs+j] = in_near[i]; + return out; +} + +// this is a reduced-precision calculation of YCbCr-to-RGB introduced +// to make sure the code produces the same results in both SIMD and scalar +#define stbi__float2fixed(x) (((int) ((x) * 4096.0f + 0.5f)) << 8) +static void stbi__YCbCr_to_RGB_row(stbi_uc *out, const stbi_uc *y, const stbi_uc *pcb, const stbi_uc *pcr, int count, int step) +{ + int i; + for (i=0; i < count; ++i) { + int y_fixed = (y[i] << 20) + (1<<19); // rounding + int r,g,b; + int cr = pcr[i] - 128; + int cb = pcb[i] - 128; + r = y_fixed + cr* stbi__float2fixed(1.40200f); + g = y_fixed + (cr*-stbi__float2fixed(0.71414f)) + ((cb*-stbi__float2fixed(0.34414f)) & 0xffff0000); + b = y_fixed + cb* stbi__float2fixed(1.77200f); + r >>= 20; + g >>= 20; + b >>= 20; + if ((unsigned) r > 255) { if (r < 0) r = 0; else r = 255; } + if ((unsigned) g > 255) { if (g < 0) g = 0; else g = 255; } + if ((unsigned) b > 255) { if (b < 0) b = 0; else b = 255; } + out[0] = (stbi_uc)r; + out[1] = (stbi_uc)g; + out[2] = (stbi_uc)b; + out[3] = 255; + out += step; + } +} + +#if defined(STBI_SSE2) || defined(STBI_NEON) +static void stbi__YCbCr_to_RGB_simd(stbi_uc *out, stbi_uc const *y, stbi_uc const *pcb, stbi_uc const *pcr, int count, int step) +{ + int i = 0; + +#ifdef STBI_SSE2 + // step == 3 is pretty ugly on the final interleave, and i'm not convinced + // it's useful in practice (you wouldn't use it for textures, for example). + // so just accelerate step == 4 case. + if (step == 4) { + // this is a fairly straightforward implementation and not super-optimized. + __m128i signflip = _mm_set1_epi8(-0x80); + __m128i cr_const0 = _mm_set1_epi16( (short) ( 1.40200f*4096.0f+0.5f)); + __m128i cr_const1 = _mm_set1_epi16( - (short) ( 0.71414f*4096.0f+0.5f)); + __m128i cb_const0 = _mm_set1_epi16( - (short) ( 0.34414f*4096.0f+0.5f)); + __m128i cb_const1 = _mm_set1_epi16( (short) ( 1.77200f*4096.0f+0.5f)); + __m128i y_bias = _mm_set1_epi8((char) (unsigned char) 128); + __m128i xw = _mm_set1_epi16(255); // alpha channel + + for (; i+7 < count; i += 8) { + // load + __m128i y_bytes = _mm_loadl_epi64((__m128i *) (y+i)); + __m128i cr_bytes = _mm_loadl_epi64((__m128i *) (pcr+i)); + __m128i cb_bytes = _mm_loadl_epi64((__m128i *) (pcb+i)); + __m128i cr_biased = _mm_xor_si128(cr_bytes, signflip); // -128 + __m128i cb_biased = _mm_xor_si128(cb_bytes, signflip); // -128 + + // unpack to short (and left-shift cr, cb by 8) + __m128i yw = _mm_unpacklo_epi8(y_bias, y_bytes); + __m128i crw = _mm_unpacklo_epi8(_mm_setzero_si128(), cr_biased); + __m128i cbw = _mm_unpacklo_epi8(_mm_setzero_si128(), cb_biased); + + // color transform + __m128i yws = _mm_srli_epi16(yw, 4); + __m128i cr0 = _mm_mulhi_epi16(cr_const0, crw); + __m128i cb0 = _mm_mulhi_epi16(cb_const0, cbw); + __m128i cb1 = _mm_mulhi_epi16(cbw, cb_const1); + __m128i cr1 = _mm_mulhi_epi16(crw, cr_const1); + __m128i rws = _mm_add_epi16(cr0, yws); + __m128i gwt = _mm_add_epi16(cb0, yws); + __m128i bws = _mm_add_epi16(yws, cb1); + __m128i gws = _mm_add_epi16(gwt, cr1); + + // descale + __m128i rw = _mm_srai_epi16(rws, 4); + __m128i bw = _mm_srai_epi16(bws, 4); + __m128i gw = _mm_srai_epi16(gws, 4); + + // back to byte, set up for transpose + __m128i brb = _mm_packus_epi16(rw, bw); + __m128i gxb = _mm_packus_epi16(gw, xw); + + // transpose to interleave channels + __m128i t0 = _mm_unpacklo_epi8(brb, gxb); + __m128i t1 = _mm_unpackhi_epi8(brb, gxb); + __m128i o0 = _mm_unpacklo_epi16(t0, t1); + __m128i o1 = _mm_unpackhi_epi16(t0, t1); + + // store + _mm_storeu_si128((__m128i *) (out + 0), o0); + _mm_storeu_si128((__m128i *) (out + 16), o1); + out += 32; + } + } +#endif + +#ifdef STBI_NEON + // in this version, step=3 support would be easy to add. but is there demand? + if (step == 4) { + // this is a fairly straightforward implementation and not super-optimized. + uint8x8_t signflip = vdup_n_u8(0x80); + int16x8_t cr_const0 = vdupq_n_s16( (short) ( 1.40200f*4096.0f+0.5f)); + int16x8_t cr_const1 = vdupq_n_s16( - (short) ( 0.71414f*4096.0f+0.5f)); + int16x8_t cb_const0 = vdupq_n_s16( - (short) ( 0.34414f*4096.0f+0.5f)); + int16x8_t cb_const1 = vdupq_n_s16( (short) ( 1.77200f*4096.0f+0.5f)); + + for (; i+7 < count; i += 8) { + // load + uint8x8_t y_bytes = vld1_u8(y + i); + uint8x8_t cr_bytes = vld1_u8(pcr + i); + uint8x8_t cb_bytes = vld1_u8(pcb + i); + int8x8_t cr_biased = vreinterpret_s8_u8(vsub_u8(cr_bytes, signflip)); + int8x8_t cb_biased = vreinterpret_s8_u8(vsub_u8(cb_bytes, signflip)); + + // expand to s16 + int16x8_t yws = vreinterpretq_s16_u16(vshll_n_u8(y_bytes, 4)); + int16x8_t crw = vshll_n_s8(cr_biased, 7); + int16x8_t cbw = vshll_n_s8(cb_biased, 7); + + // color transform + int16x8_t cr0 = vqdmulhq_s16(crw, cr_const0); + int16x8_t cb0 = vqdmulhq_s16(cbw, cb_const0); + int16x8_t cr1 = vqdmulhq_s16(crw, cr_const1); + int16x8_t cb1 = vqdmulhq_s16(cbw, cb_const1); + int16x8_t rws = vaddq_s16(yws, cr0); + int16x8_t gws = vaddq_s16(vaddq_s16(yws, cb0), cr1); + int16x8_t bws = vaddq_s16(yws, cb1); + + // undo scaling, round, convert to byte + uint8x8x4_t o; + o.val[0] = vqrshrun_n_s16(rws, 4); + o.val[1] = vqrshrun_n_s16(gws, 4); + o.val[2] = vqrshrun_n_s16(bws, 4); + o.val[3] = vdup_n_u8(255); + + // store, interleaving r/g/b/a + vst4_u8(out, o); + out += 8*4; + } + } +#endif + + for (; i < count; ++i) { + int y_fixed = (y[i] << 20) + (1<<19); // rounding + int r,g,b; + int cr = pcr[i] - 128; + int cb = pcb[i] - 128; + r = y_fixed + cr* stbi__float2fixed(1.40200f); + g = y_fixed + cr*-stbi__float2fixed(0.71414f) + ((cb*-stbi__float2fixed(0.34414f)) & 0xffff0000); + b = y_fixed + cb* stbi__float2fixed(1.77200f); + r >>= 20; + g >>= 20; + b >>= 20; + if ((unsigned) r > 255) { if (r < 0) r = 0; else r = 255; } + if ((unsigned) g > 255) { if (g < 0) g = 0; else g = 255; } + if ((unsigned) b > 255) { if (b < 0) b = 0; else b = 255; } + out[0] = (stbi_uc)r; + out[1] = (stbi_uc)g; + out[2] = (stbi_uc)b; + out[3] = 255; + out += step; + } +} +#endif + +// set up the kernels +static void stbi__setup_jpeg(stbi__jpeg *j) +{ + j->idct_block_kernel = stbi__idct_block; + j->YCbCr_to_RGB_kernel = stbi__YCbCr_to_RGB_row; + j->resample_row_hv_2_kernel = stbi__resample_row_hv_2; + +#ifdef STBI_SSE2 + if (stbi__sse2_available()) { + j->idct_block_kernel = stbi__idct_simd; + j->YCbCr_to_RGB_kernel = stbi__YCbCr_to_RGB_simd; + j->resample_row_hv_2_kernel = stbi__resample_row_hv_2_simd; + } +#endif + +#ifdef STBI_NEON + j->idct_block_kernel = stbi__idct_simd; + j->YCbCr_to_RGB_kernel = stbi__YCbCr_to_RGB_simd; + j->resample_row_hv_2_kernel = stbi__resample_row_hv_2_simd; +#endif +} + +// clean up the temporary component buffers +static void stbi__cleanup_jpeg(stbi__jpeg *j) +{ + stbi__free_jpeg_components(j, j->s->img_n, 0); +} + +typedef struct +{ + resample_row_func resample; + stbi_uc *line0,*line1; + int hs,vs; // expansion factor in each axis + int w_lores; // horizontal pixels pre-expansion + int ystep; // how far through vertical expansion we are + int ypos; // which pre-expansion row we're on +} stbi__resample; + +// fast 0..255 * 0..255 => 0..255 rounded multiplication +static stbi_uc stbi__blinn_8x8(stbi_uc x, stbi_uc y) +{ + unsigned int t = x*y + 128; + return (stbi_uc) ((t + (t >>8)) >> 8); +} + +static stbi_uc *load_jpeg_image(stbi__jpeg *z, int *out_x, int *out_y, int *comp, int req_comp) +{ + int n, decode_n, is_rgb; + z->s->img_n = 0; // make stbi__cleanup_jpeg safe + + // validate req_comp + if (req_comp < 0 || req_comp > 4) return stbi__errpuc("bad req_comp", "Internal error"); + + // load a jpeg image from whichever source, but leave in YCbCr format + if (!stbi__decode_jpeg_image(z)) { stbi__cleanup_jpeg(z); return NULL; } + + // determine actual number of components to generate + n = req_comp ? req_comp : z->s->img_n >= 3 ? 3 : 1; + + is_rgb = z->s->img_n == 3 && (z->rgb == 3 || (z->app14_color_transform == 0 && !z->jfif)); + + if (z->s->img_n == 3 && n < 3 && !is_rgb) + decode_n = 1; + else + decode_n = z->s->img_n; + + // nothing to do if no components requested; check this now to avoid + // accessing uninitialized coutput[0] later + if (decode_n <= 0) { stbi__cleanup_jpeg(z); return NULL; } + + // resample and color-convert + { + int k; + unsigned int i,j; + stbi_uc *output; + stbi_uc *coutput[4] = { NULL, NULL, NULL, NULL }; + + stbi__resample res_comp[4]; + + for (k=0; k < decode_n; ++k) { + stbi__resample *r = &res_comp[k]; + + // allocate line buffer big enough for upsampling off the edges + // with upsample factor of 4 + z->img_comp[k].linebuf = (stbi_uc *) stbi__malloc(z->s->img_x + 3); + if (!z->img_comp[k].linebuf) { stbi__cleanup_jpeg(z); return stbi__errpuc("outofmem", "Out of memory"); } + + r->hs = z->img_h_max / z->img_comp[k].h; + r->vs = z->img_v_max / z->img_comp[k].v; + r->ystep = r->vs >> 1; + r->w_lores = (z->s->img_x + r->hs-1) / r->hs; + r->ypos = 0; + r->line0 = r->line1 = z->img_comp[k].data; + + if (r->hs == 1 && r->vs == 1) r->resample = resample_row_1; + else if (r->hs == 1 && r->vs == 2) r->resample = stbi__resample_row_v_2; + else if (r->hs == 2 && r->vs == 1) r->resample = stbi__resample_row_h_2; + else if (r->hs == 2 && r->vs == 2) r->resample = z->resample_row_hv_2_kernel; + else r->resample = stbi__resample_row_generic; + } + + // can't error after this so, this is safe + output = (stbi_uc *) stbi__malloc_mad3(n, z->s->img_x, z->s->img_y, 1); + if (!output) { stbi__cleanup_jpeg(z); return stbi__errpuc("outofmem", "Out of memory"); } + + // now go ahead and resample + for (j=0; j < z->s->img_y; ++j) { + stbi_uc *out = output + n * z->s->img_x * j; + for (k=0; k < decode_n; ++k) { + stbi__resample *r = &res_comp[k]; + int y_bot = r->ystep >= (r->vs >> 1); + coutput[k] = r->resample(z->img_comp[k].linebuf, + y_bot ? r->line1 : r->line0, + y_bot ? r->line0 : r->line1, + r->w_lores, r->hs); + if (++r->ystep >= r->vs) { + r->ystep = 0; + r->line0 = r->line1; + if (++r->ypos < z->img_comp[k].y) + r->line1 += z->img_comp[k].w2; + } + } + if (n >= 3) { + stbi_uc *y = coutput[0]; + if (z->s->img_n == 3) { + if (is_rgb) { + for (i=0; i < z->s->img_x; ++i) { + out[0] = y[i]; + out[1] = coutput[1][i]; + out[2] = coutput[2][i]; + out[3] = 255; + out += n; + } + } else { + z->YCbCr_to_RGB_kernel(out, y, coutput[1], coutput[2], z->s->img_x, n); + } + } else if (z->s->img_n == 4) { + if (z->app14_color_transform == 0) { // CMYK + for (i=0; i < z->s->img_x; ++i) { + stbi_uc m = coutput[3][i]; + out[0] = stbi__blinn_8x8(coutput[0][i], m); + out[1] = stbi__blinn_8x8(coutput[1][i], m); + out[2] = stbi__blinn_8x8(coutput[2][i], m); + out[3] = 255; + out += n; + } + } else if (z->app14_color_transform == 2) { // YCCK + z->YCbCr_to_RGB_kernel(out, y, coutput[1], coutput[2], z->s->img_x, n); + for (i=0; i < z->s->img_x; ++i) { + stbi_uc m = coutput[3][i]; + out[0] = stbi__blinn_8x8(255 - out[0], m); + out[1] = stbi__blinn_8x8(255 - out[1], m); + out[2] = stbi__blinn_8x8(255 - out[2], m); + out += n; + } + } else { // YCbCr + alpha? Ignore the fourth channel for now + z->YCbCr_to_RGB_kernel(out, y, coutput[1], coutput[2], z->s->img_x, n); + } + } else + for (i=0; i < z->s->img_x; ++i) { + out[0] = out[1] = out[2] = y[i]; + out[3] = 255; // not used if n==3 + out += n; + } + } else { + if (is_rgb) { + if (n == 1) + for (i=0; i < z->s->img_x; ++i) + *out++ = stbi__compute_y(coutput[0][i], coutput[1][i], coutput[2][i]); + else { + for (i=0; i < z->s->img_x; ++i, out += 2) { + out[0] = stbi__compute_y(coutput[0][i], coutput[1][i], coutput[2][i]); + out[1] = 255; + } + } + } else if (z->s->img_n == 4 && z->app14_color_transform == 0) { + for (i=0; i < z->s->img_x; ++i) { + stbi_uc m = coutput[3][i]; + stbi_uc r = stbi__blinn_8x8(coutput[0][i], m); + stbi_uc g = stbi__blinn_8x8(coutput[1][i], m); + stbi_uc b = stbi__blinn_8x8(coutput[2][i], m); + out[0] = stbi__compute_y(r, g, b); + out[1] = 255; + out += n; + } + } else if (z->s->img_n == 4 && z->app14_color_transform == 2) { + for (i=0; i < z->s->img_x; ++i) { + out[0] = stbi__blinn_8x8(255 - coutput[0][i], coutput[3][i]); + out[1] = 255; + out += n; + } + } else { + stbi_uc *y = coutput[0]; + if (n == 1) + for (i=0; i < z->s->img_x; ++i) out[i] = y[i]; + else + for (i=0; i < z->s->img_x; ++i) { *out++ = y[i]; *out++ = 255; } + } + } + } + stbi__cleanup_jpeg(z); + *out_x = z->s->img_x; + *out_y = z->s->img_y; + if (comp) *comp = z->s->img_n >= 3 ? 3 : 1; // report original components, not output + return output; + } +} + +static void *stbi__jpeg_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri) +{ + unsigned char* result; + stbi__jpeg* j = (stbi__jpeg*) stbi__malloc(sizeof(stbi__jpeg)); + if (!j) return stbi__errpuc("outofmem", "Out of memory"); + memset(j, 0, sizeof(stbi__jpeg)); + STBI_NOTUSED(ri); + j->s = s; + stbi__setup_jpeg(j); + result = load_jpeg_image(j, x,y,comp,req_comp); + STBI_FREE(j); + return result; +} + +static int stbi__jpeg_test(stbi__context *s) +{ + int r; + stbi__jpeg* j = (stbi__jpeg*)stbi__malloc(sizeof(stbi__jpeg)); + if (!j) return stbi__err("outofmem", "Out of memory"); + memset(j, 0, sizeof(stbi__jpeg)); + j->s = s; + stbi__setup_jpeg(j); + r = stbi__decode_jpeg_header(j, STBI__SCAN_type); + stbi__rewind(s); + STBI_FREE(j); + return r; +} + +static int stbi__jpeg_info_raw(stbi__jpeg *j, int *x, int *y, int *comp) +{ + if (!stbi__decode_jpeg_header(j, STBI__SCAN_header)) { + stbi__rewind( j->s ); + return 0; + } + if (x) *x = j->s->img_x; + if (y) *y = j->s->img_y; + if (comp) *comp = j->s->img_n >= 3 ? 3 : 1; + return 1; +} + +static int stbi__jpeg_info(stbi__context *s, int *x, int *y, int *comp) +{ + int result; + stbi__jpeg* j = (stbi__jpeg*) (stbi__malloc(sizeof(stbi__jpeg))); + if (!j) return stbi__err("outofmem", "Out of memory"); + memset(j, 0, sizeof(stbi__jpeg)); + j->s = s; + result = stbi__jpeg_info_raw(j, x, y, comp); + STBI_FREE(j); + return result; +} +#endif + +// public domain zlib decode v0.2 Sean Barrett 2006-11-18 +// simple implementation +// - all input must be provided in an upfront buffer +// - all output is written to a single output buffer (can malloc/realloc) +// performance +// - fast huffman + +#ifndef STBI_NO_ZLIB + +// fast-way is faster to check than jpeg huffman, but slow way is slower +#define STBI__ZFAST_BITS 9 // accelerate all cases in default tables +#define STBI__ZFAST_MASK ((1 << STBI__ZFAST_BITS) - 1) +#define STBI__ZNSYMS 288 // number of symbols in literal/length alphabet + +// zlib-style huffman encoding +// (jpegs packs from left, zlib from right, so can't share code) +typedef struct +{ + stbi__uint16 fast[1 << STBI__ZFAST_BITS]; + stbi__uint16 firstcode[16]; + int maxcode[17]; + stbi__uint16 firstsymbol[16]; + stbi_uc size[STBI__ZNSYMS]; + stbi__uint16 value[STBI__ZNSYMS]; +} stbi__zhuffman; + +stbi_inline static int stbi__bitreverse16(int n) +{ + n = ((n & 0xAAAA) >> 1) | ((n & 0x5555) << 1); + n = ((n & 0xCCCC) >> 2) | ((n & 0x3333) << 2); + n = ((n & 0xF0F0) >> 4) | ((n & 0x0F0F) << 4); + n = ((n & 0xFF00) >> 8) | ((n & 0x00FF) << 8); + return n; +} + +stbi_inline static int stbi__bit_reverse(int v, int bits) +{ + STBI_ASSERT(bits <= 16); + // to bit reverse n bits, reverse 16 and shift + // e.g. 11 bits, bit reverse and shift away 5 + return stbi__bitreverse16(v) >> (16-bits); +} + +static int stbi__zbuild_huffman(stbi__zhuffman *z, const stbi_uc *sizelist, int num) +{ + int i,k=0; + int code, next_code[16], sizes[17]; + + // DEFLATE spec for generating codes + memset(sizes, 0, sizeof(sizes)); + memset(z->fast, 0, sizeof(z->fast)); + for (i=0; i < num; ++i) + ++sizes[sizelist[i]]; + sizes[0] = 0; + for (i=1; i < 16; ++i) + if (sizes[i] > (1 << i)) + return stbi__err("bad sizes", "Corrupt PNG"); + code = 0; + for (i=1; i < 16; ++i) { + next_code[i] = code; + z->firstcode[i] = (stbi__uint16) code; + z->firstsymbol[i] = (stbi__uint16) k; + code = (code + sizes[i]); + if (sizes[i]) + if (code-1 >= (1 << i)) return stbi__err("bad codelengths","Corrupt PNG"); + z->maxcode[i] = code << (16-i); // preshift for inner loop + code <<= 1; + k += sizes[i]; + } + z->maxcode[16] = 0x10000; // sentinel + for (i=0; i < num; ++i) { + int s = sizelist[i]; + if (s) { + int c = next_code[s] - z->firstcode[s] + z->firstsymbol[s]; + stbi__uint16 fastv = (stbi__uint16) ((s << 9) | i); + z->size [c] = (stbi_uc ) s; + z->value[c] = (stbi__uint16) i; + if (s <= STBI__ZFAST_BITS) { + int j = stbi__bit_reverse(next_code[s],s); + while (j < (1 << STBI__ZFAST_BITS)) { + z->fast[j] = fastv; + j += (1 << s); + } + } + ++next_code[s]; + } + } + return 1; +} + +// zlib-from-memory implementation for PNG reading +// because PNG allows splitting the zlib stream arbitrarily, +// and it's annoying structurally to have PNG call ZLIB call PNG, +// we require PNG read all the IDATs and combine them into a single +// memory buffer + +typedef struct +{ + stbi_uc *zbuffer, *zbuffer_end; + int num_bits; + int hit_zeof_once; + stbi__uint32 code_buffer; + + char *zout; + char *zout_start; + char *zout_end; + int z_expandable; + + stbi__zhuffman z_length, z_distance; +} stbi__zbuf; + +stbi_inline static int stbi__zeof(stbi__zbuf *z) +{ + return (z->zbuffer >= z->zbuffer_end); +} + +stbi_inline static stbi_uc stbi__zget8(stbi__zbuf *z) +{ + return stbi__zeof(z) ? 0 : *z->zbuffer++; +} + +static void stbi__fill_bits(stbi__zbuf *z) +{ + do { + if (z->code_buffer >= (1U << z->num_bits)) { + z->zbuffer = z->zbuffer_end; /* treat this as EOF so we fail. */ + return; + } + z->code_buffer |= (unsigned int) stbi__zget8(z) << z->num_bits; + z->num_bits += 8; + } while (z->num_bits <= 24); +} + +stbi_inline static unsigned int stbi__zreceive(stbi__zbuf *z, int n) +{ + unsigned int k; + if (z->num_bits < n) stbi__fill_bits(z); + k = z->code_buffer & ((1 << n) - 1); + z->code_buffer >>= n; + z->num_bits -= n; + return k; +} + +static int stbi__zhuffman_decode_slowpath(stbi__zbuf *a, stbi__zhuffman *z) +{ + int b,s,k; + // not resolved by fast table, so compute it the slow way + // use jpeg approach, which requires MSbits at top + k = stbi__bit_reverse(a->code_buffer, 16); + for (s=STBI__ZFAST_BITS+1; ; ++s) + if (k < z->maxcode[s]) + break; + if (s >= 16) return -1; // invalid code! + // code size is s, so: + b = (k >> (16-s)) - z->firstcode[s] + z->firstsymbol[s]; + if (b >= STBI__ZNSYMS) return -1; // some data was corrupt somewhere! + if (z->size[b] != s) return -1; // was originally an assert, but report failure instead. + a->code_buffer >>= s; + a->num_bits -= s; + return z->value[b]; +} + +stbi_inline static int stbi__zhuffman_decode(stbi__zbuf *a, stbi__zhuffman *z) +{ + int b,s; + if (a->num_bits < 16) { + if (stbi__zeof(a)) { + if (!a->hit_zeof_once) { + // This is the first time we hit eof, insert 16 extra padding btis + // to allow us to keep going; if we actually consume any of them + // though, that is invalid data. This is caught later. + a->hit_zeof_once = 1; + a->num_bits += 16; // add 16 implicit zero bits + } else { + // We already inserted our extra 16 padding bits and are again + // out, this stream is actually prematurely terminated. + return -1; + } + } else { + stbi__fill_bits(a); + } + } + b = z->fast[a->code_buffer & STBI__ZFAST_MASK]; + if (b) { + s = b >> 9; + a->code_buffer >>= s; + a->num_bits -= s; + return b & 511; + } + return stbi__zhuffman_decode_slowpath(a, z); +} + +static int stbi__zexpand(stbi__zbuf *z, char *zout, int n) // need to make room for n bytes +{ + char *q; + unsigned int cur, limit, old_limit; + z->zout = zout; + if (!z->z_expandable) return stbi__err("output buffer limit","Corrupt PNG"); + cur = (unsigned int) (z->zout - z->zout_start); + limit = old_limit = (unsigned) (z->zout_end - z->zout_start); + if (UINT_MAX - cur < (unsigned) n) return stbi__err("outofmem", "Out of memory"); + while (cur + n > limit) { + if(limit > UINT_MAX / 2) return stbi__err("outofmem", "Out of memory"); + limit *= 2; + } + q = (char *) STBI_REALLOC_SIZED(z->zout_start, old_limit, limit); + STBI_NOTUSED(old_limit); + if (q == NULL) return stbi__err("outofmem", "Out of memory"); + z->zout_start = q; + z->zout = q + cur; + z->zout_end = q + limit; + return 1; +} + +static const int stbi__zlength_base[31] = { + 3,4,5,6,7,8,9,10,11,13, + 15,17,19,23,27,31,35,43,51,59, + 67,83,99,115,131,163,195,227,258,0,0 }; + +static const int stbi__zlength_extra[31]= +{ 0,0,0,0,0,0,0,0,1,1,1,1,2,2,2,2,3,3,3,3,4,4,4,4,5,5,5,5,0,0,0 }; + +static const int stbi__zdist_base[32] = { 1,2,3,4,5,7,9,13,17,25,33,49,65,97,129,193, +257,385,513,769,1025,1537,2049,3073,4097,6145,8193,12289,16385,24577,0,0}; + +static const int stbi__zdist_extra[32] = +{ 0,0,0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,11,11,12,12,13,13}; + +static int stbi__parse_huffman_block(stbi__zbuf *a) +{ + char *zout = a->zout; + for(;;) { + int z = stbi__zhuffman_decode(a, &a->z_length); + if (z < 256) { + if (z < 0) return stbi__err("bad huffman code","Corrupt PNG"); // error in huffman codes + if (zout >= a->zout_end) { + if (!stbi__zexpand(a, zout, 1)) return 0; + zout = a->zout; + } + *zout++ = (char) z; + } else { + stbi_uc *p; + int len,dist; + if (z == 256) { + a->zout = zout; + if (a->hit_zeof_once && a->num_bits < 16) { + // The first time we hit zeof, we inserted 16 extra zero bits into our bit + // buffer so the decoder can just do its speculative decoding. But if we + // actually consumed any of those bits (which is the case when num_bits < 16), + // the stream actually read past the end so it is malformed. + return stbi__err("unexpected end","Corrupt PNG"); + } + return 1; + } + if (z >= 286) return stbi__err("bad huffman code","Corrupt PNG"); // per DEFLATE, length codes 286 and 287 must not appear in compressed data + z -= 257; + len = stbi__zlength_base[z]; + if (stbi__zlength_extra[z]) len += stbi__zreceive(a, stbi__zlength_extra[z]); + z = stbi__zhuffman_decode(a, &a->z_distance); + if (z < 0 || z >= 30) return stbi__err("bad huffman code","Corrupt PNG"); // per DEFLATE, distance codes 30 and 31 must not appear in compressed data + dist = stbi__zdist_base[z]; + if (stbi__zdist_extra[z]) dist += stbi__zreceive(a, stbi__zdist_extra[z]); + if (zout - a->zout_start < dist) return stbi__err("bad dist","Corrupt PNG"); + if (len > a->zout_end - zout) { + if (!stbi__zexpand(a, zout, len)) return 0; + zout = a->zout; + } + p = (stbi_uc *) (zout - dist); + if (dist == 1) { // run of one byte; common in images. + stbi_uc v = *p; + if (len) { do *zout++ = v; while (--len); } + } else { + if (len) { do *zout++ = *p++; while (--len); } + } + } + } +} + +static int stbi__compute_huffman_codes(stbi__zbuf *a) +{ + static const stbi_uc length_dezigzag[19] = { 16,17,18,0,8,7,9,6,10,5,11,4,12,3,13,2,14,1,15 }; + stbi__zhuffman z_codelength; + stbi_uc lencodes[286+32+137];//padding for maximum single op + stbi_uc codelength_sizes[19]; + int i,n; + + int hlit = stbi__zreceive(a,5) + 257; + int hdist = stbi__zreceive(a,5) + 1; + int hclen = stbi__zreceive(a,4) + 4; + int ntot = hlit + hdist; + + memset(codelength_sizes, 0, sizeof(codelength_sizes)); + for (i=0; i < hclen; ++i) { + int s = stbi__zreceive(a,3); + codelength_sizes[length_dezigzag[i]] = (stbi_uc) s; + } + if (!stbi__zbuild_huffman(&z_codelength, codelength_sizes, 19)) return 0; + + n = 0; + while (n < ntot) { + int c = stbi__zhuffman_decode(a, &z_codelength); + if (c < 0 || c >= 19) return stbi__err("bad codelengths", "Corrupt PNG"); + if (c < 16) + lencodes[n++] = (stbi_uc) c; + else { + stbi_uc fill = 0; + if (c == 16) { + c = stbi__zreceive(a,2)+3; + if (n == 0) return stbi__err("bad codelengths", "Corrupt PNG"); + fill = lencodes[n-1]; + } else if (c == 17) { + c = stbi__zreceive(a,3)+3; + } else if (c == 18) { + c = stbi__zreceive(a,7)+11; + } else { + return stbi__err("bad codelengths", "Corrupt PNG"); + } + if (ntot - n < c) return stbi__err("bad codelengths", "Corrupt PNG"); + memset(lencodes+n, fill, c); + n += c; + } + } + if (n != ntot) return stbi__err("bad codelengths","Corrupt PNG"); + if (!stbi__zbuild_huffman(&a->z_length, lencodes, hlit)) return 0; + if (!stbi__zbuild_huffman(&a->z_distance, lencodes+hlit, hdist)) return 0; + return 1; +} + +static int stbi__parse_uncompressed_block(stbi__zbuf *a) +{ + stbi_uc header[4]; + int len,nlen,k; + if (a->num_bits & 7) + stbi__zreceive(a, a->num_bits & 7); // discard + // drain the bit-packed data into header + k = 0; + while (a->num_bits > 0) { + header[k++] = (stbi_uc) (a->code_buffer & 255); // suppress MSVC run-time check + a->code_buffer >>= 8; + a->num_bits -= 8; + } + if (a->num_bits < 0) return stbi__err("zlib corrupt","Corrupt PNG"); + // now fill header the normal way + while (k < 4) + header[k++] = stbi__zget8(a); + len = header[1] * 256 + header[0]; + nlen = header[3] * 256 + header[2]; + if (nlen != (len ^ 0xffff)) return stbi__err("zlib corrupt","Corrupt PNG"); + if (a->zbuffer + len > a->zbuffer_end) return stbi__err("read past buffer","Corrupt PNG"); + if (a->zout + len > a->zout_end) + if (!stbi__zexpand(a, a->zout, len)) return 0; + memcpy(a->zout, a->zbuffer, len); + a->zbuffer += len; + a->zout += len; + return 1; +} + +static int stbi__parse_zlib_header(stbi__zbuf *a) +{ + int cmf = stbi__zget8(a); + int cm = cmf & 15; + /* int cinfo = cmf >> 4; */ + int flg = stbi__zget8(a); + if (stbi__zeof(a)) return stbi__err("bad zlib header","Corrupt PNG"); // zlib spec + if ((cmf*256+flg) % 31 != 0) return stbi__err("bad zlib header","Corrupt PNG"); // zlib spec + if (flg & 32) return stbi__err("no preset dict","Corrupt PNG"); // preset dictionary not allowed in png + if (cm != 8) return stbi__err("bad compression","Corrupt PNG"); // DEFLATE required for png + // window = 1 << (8 + cinfo)... but who cares, we fully buffer output + return 1; +} + +static const stbi_uc stbi__zdefault_length[STBI__ZNSYMS] = +{ + 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, + 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, + 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, + 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, + 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, 9,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9, + 9,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9, 9,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9, + 9,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9, 9,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9, + 9,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9, 9,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9, + 7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7, 7,7,7,7,7,7,7,7,8,8,8,8,8,8,8,8 +}; +static const stbi_uc stbi__zdefault_distance[32] = +{ + 5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5 +}; +/* +Init algorithm: +{ + int i; // use <= to match clearly with spec + for (i=0; i <= 143; ++i) stbi__zdefault_length[i] = 8; + for ( ; i <= 255; ++i) stbi__zdefault_length[i] = 9; + for ( ; i <= 279; ++i) stbi__zdefault_length[i] = 7; + for ( ; i <= 287; ++i) stbi__zdefault_length[i] = 8; + + for (i=0; i <= 31; ++i) stbi__zdefault_distance[i] = 5; +} +*/ + +static int stbi__parse_zlib(stbi__zbuf *a, int parse_header) +{ + int final, type; + if (parse_header) + if (!stbi__parse_zlib_header(a)) return 0; + a->num_bits = 0; + a->code_buffer = 0; + a->hit_zeof_once = 0; + do { + final = stbi__zreceive(a,1); + type = stbi__zreceive(a,2); + if (type == 0) { + if (!stbi__parse_uncompressed_block(a)) return 0; + } else if (type == 3) { + return 0; + } else { + if (type == 1) { + // use fixed code lengths + if (!stbi__zbuild_huffman(&a->z_length , stbi__zdefault_length , STBI__ZNSYMS)) return 0; + if (!stbi__zbuild_huffman(&a->z_distance, stbi__zdefault_distance, 32)) return 0; + } else { + if (!stbi__compute_huffman_codes(a)) return 0; + } + if (!stbi__parse_huffman_block(a)) return 0; + } + } while (!final); + return 1; +} + +static int stbi__do_zlib(stbi__zbuf *a, char *obuf, int olen, int exp, int parse_header) +{ + a->zout_start = obuf; + a->zout = obuf; + a->zout_end = obuf + olen; + a->z_expandable = exp; + + return stbi__parse_zlib(a, parse_header); +} + +STBIDEF char *stbi_zlib_decode_malloc_guesssize(const char *buffer, int len, int initial_size, int *outlen) +{ + stbi__zbuf a; + char *p = (char *) stbi__malloc(initial_size); + if (p == NULL) return NULL; + a.zbuffer = (stbi_uc *) buffer; + a.zbuffer_end = (stbi_uc *) buffer + len; + if (stbi__do_zlib(&a, p, initial_size, 1, 1)) { + if (outlen) *outlen = (int) (a.zout - a.zout_start); + return a.zout_start; + } else { + STBI_FREE(a.zout_start); + return NULL; + } +} + +STBIDEF char *stbi_zlib_decode_malloc(char const *buffer, int len, int *outlen) +{ + return stbi_zlib_decode_malloc_guesssize(buffer, len, 16384, outlen); +} + +STBIDEF char *stbi_zlib_decode_malloc_guesssize_headerflag(const char *buffer, int len, int initial_size, int *outlen, int parse_header) +{ + stbi__zbuf a; + char *p = (char *) stbi__malloc(initial_size); + if (p == NULL) return NULL; + a.zbuffer = (stbi_uc *) buffer; + a.zbuffer_end = (stbi_uc *) buffer + len; + if (stbi__do_zlib(&a, p, initial_size, 1, parse_header)) { + if (outlen) *outlen = (int) (a.zout - a.zout_start); + return a.zout_start; + } else { + STBI_FREE(a.zout_start); + return NULL; + } +} + +STBIDEF int stbi_zlib_decode_buffer(char *obuffer, int olen, char const *ibuffer, int ilen) +{ + stbi__zbuf a; + a.zbuffer = (stbi_uc *) ibuffer; + a.zbuffer_end = (stbi_uc *) ibuffer + ilen; + if (stbi__do_zlib(&a, obuffer, olen, 0, 1)) + return (int) (a.zout - a.zout_start); + else + return -1; +} + +STBIDEF char *stbi_zlib_decode_noheader_malloc(char const *buffer, int len, int *outlen) +{ + stbi__zbuf a; + char *p = (char *) stbi__malloc(16384); + if (p == NULL) return NULL; + a.zbuffer = (stbi_uc *) buffer; + a.zbuffer_end = (stbi_uc *) buffer+len; + if (stbi__do_zlib(&a, p, 16384, 1, 0)) { + if (outlen) *outlen = (int) (a.zout - a.zout_start); + return a.zout_start; + } else { + STBI_FREE(a.zout_start); + return NULL; + } +} + +STBIDEF int stbi_zlib_decode_noheader_buffer(char *obuffer, int olen, const char *ibuffer, int ilen) +{ + stbi__zbuf a; + a.zbuffer = (stbi_uc *) ibuffer; + a.zbuffer_end = (stbi_uc *) ibuffer + ilen; + if (stbi__do_zlib(&a, obuffer, olen, 0, 0)) + return (int) (a.zout - a.zout_start); + else + return -1; +} +#endif + +// public domain "baseline" PNG decoder v0.10 Sean Barrett 2006-11-18 +// simple implementation +// - only 8-bit samples +// - no CRC checking +// - allocates lots of intermediate memory +// - avoids problem of streaming data between subsystems +// - avoids explicit window management +// performance +// - uses stb_zlib, a PD zlib implementation with fast huffman decoding + +#ifndef STBI_NO_PNG +typedef struct +{ + stbi__uint32 length; + stbi__uint32 type; +} stbi__pngchunk; + +static stbi__pngchunk stbi__get_chunk_header(stbi__context *s) +{ + stbi__pngchunk c; + c.length = stbi__get32be(s); + c.type = stbi__get32be(s); + return c; +} + +static int stbi__check_png_header(stbi__context *s) +{ + static const stbi_uc png_sig[8] = { 137,80,78,71,13,10,26,10 }; + int i; + for (i=0; i < 8; ++i) + if (stbi__get8(s) != png_sig[i]) return stbi__err("bad png sig","Not a PNG"); + return 1; +} + +typedef struct +{ + stbi__context *s; + stbi_uc *idata, *expanded, *out; + int depth; +} stbi__png; + + +enum { + STBI__F_none=0, + STBI__F_sub=1, + STBI__F_up=2, + STBI__F_avg=3, + STBI__F_paeth=4, + // synthetic filter used for first scanline to avoid needing a dummy row of 0s + STBI__F_avg_first +}; + +static stbi_uc first_row_filter[5] = +{ + STBI__F_none, + STBI__F_sub, + STBI__F_none, + STBI__F_avg_first, + STBI__F_sub // Paeth with b=c=0 turns out to be equivalent to sub +}; + +static int stbi__paeth(int a, int b, int c) +{ + // This formulation looks very different from the reference in the PNG spec, but is + // actually equivalent and has favorable data dependencies and admits straightforward + // generation of branch-free code, which helps performance significantly. + int thresh = c*3 - (a + b); + int lo = a < b ? a : b; + int hi = a < b ? b : a; + int t0 = (hi <= thresh) ? lo : c; + int t1 = (thresh <= lo) ? hi : t0; + return t1; +} + +static const stbi_uc stbi__depth_scale_table[9] = { 0, 0xff, 0x55, 0, 0x11, 0,0,0, 0x01 }; + +// adds an extra all-255 alpha channel +// dest == src is legal +// img_n must be 1 or 3 +static void stbi__create_png_alpha_expand8(stbi_uc *dest, stbi_uc *src, stbi__uint32 x, int img_n) +{ + int i; + // must process data backwards since we allow dest==src + if (img_n == 1) { + for (i=x-1; i >= 0; --i) { + dest[i*2+1] = 255; + dest[i*2+0] = src[i]; + } + } else { + STBI_ASSERT(img_n == 3); + for (i=x-1; i >= 0; --i) { + dest[i*4+3] = 255; + dest[i*4+2] = src[i*3+2]; + dest[i*4+1] = src[i*3+1]; + dest[i*4+0] = src[i*3+0]; + } + } +} + +// create the png data from post-deflated data +static int stbi__create_png_image_raw(stbi__png *a, stbi_uc *raw, stbi__uint32 raw_len, int out_n, stbi__uint32 x, stbi__uint32 y, int depth, int color) +{ + int bytes = (depth == 16 ? 2 : 1); + stbi__context *s = a->s; + stbi__uint32 i,j,stride = x*out_n*bytes; + stbi__uint32 img_len, img_width_bytes; + stbi_uc *filter_buf; + int all_ok = 1; + int k; + int img_n = s->img_n; // copy it into a local for later + + int output_bytes = out_n*bytes; + int filter_bytes = img_n*bytes; + int width = x; + + STBI_ASSERT(out_n == s->img_n || out_n == s->img_n+1); + a->out = (stbi_uc *) stbi__malloc_mad3(x, y, output_bytes, 0); // extra bytes to write off the end into + if (!a->out) return stbi__err("outofmem", "Out of memory"); + + // note: error exits here don't need to clean up a->out individually, + // stbi__do_png always does on error. + if (!stbi__mad3sizes_valid(img_n, x, depth, 7)) return stbi__err("too large", "Corrupt PNG"); + img_width_bytes = (((img_n * x * depth) + 7) >> 3); + if (!stbi__mad2sizes_valid(img_width_bytes, y, img_width_bytes)) return stbi__err("too large", "Corrupt PNG"); + img_len = (img_width_bytes + 1) * y; + + // we used to check for exact match between raw_len and img_len on non-interlaced PNGs, + // but issue #276 reported a PNG in the wild that had extra data at the end (all zeros), + // so just check for raw_len < img_len always. + if (raw_len < img_len) return stbi__err("not enough pixels","Corrupt PNG"); + + // Allocate two scan lines worth of filter workspace buffer. + filter_buf = (stbi_uc *) stbi__malloc_mad2(img_width_bytes, 2, 0); + if (!filter_buf) return stbi__err("outofmem", "Out of memory"); + + // Filtering for low-bit-depth images + if (depth < 8) { + filter_bytes = 1; + width = img_width_bytes; + } + + for (j=0; j < y; ++j) { + // cur/prior filter buffers alternate + stbi_uc *cur = filter_buf + (j & 1)*img_width_bytes; + stbi_uc *prior = filter_buf + (~j & 1)*img_width_bytes; + stbi_uc *dest = a->out + stride*j; + int nk = width * filter_bytes; + int filter = *raw++; + + // check filter type + if (filter > 4) { + all_ok = stbi__err("invalid filter","Corrupt PNG"); + break; + } + + // if first row, use special filter that doesn't sample previous row + if (j == 0) filter = first_row_filter[filter]; + + // perform actual filtering + switch (filter) { + case STBI__F_none: + memcpy(cur, raw, nk); + break; + case STBI__F_sub: + memcpy(cur, raw, filter_bytes); + for (k = filter_bytes; k < nk; ++k) + cur[k] = STBI__BYTECAST(raw[k] + cur[k-filter_bytes]); + break; + case STBI__F_up: + for (k = 0; k < nk; ++k) + cur[k] = STBI__BYTECAST(raw[k] + prior[k]); + break; + case STBI__F_avg: + for (k = 0; k < filter_bytes; ++k) + cur[k] = STBI__BYTECAST(raw[k] + (prior[k]>>1)); + for (k = filter_bytes; k < nk; ++k) + cur[k] = STBI__BYTECAST(raw[k] + ((prior[k] + cur[k-filter_bytes])>>1)); + break; + case STBI__F_paeth: + for (k = 0; k < filter_bytes; ++k) + cur[k] = STBI__BYTECAST(raw[k] + prior[k]); // prior[k] == stbi__paeth(0,prior[k],0) + for (k = filter_bytes; k < nk; ++k) + cur[k] = STBI__BYTECAST(raw[k] + stbi__paeth(cur[k-filter_bytes], prior[k], prior[k-filter_bytes])); + break; + case STBI__F_avg_first: + memcpy(cur, raw, filter_bytes); + for (k = filter_bytes; k < nk; ++k) + cur[k] = STBI__BYTECAST(raw[k] + (cur[k-filter_bytes] >> 1)); + break; + } + + raw += nk; + + // expand decoded bits in cur to dest, also adding an extra alpha channel if desired + if (depth < 8) { + stbi_uc scale = (color == 0) ? stbi__depth_scale_table[depth] : 1; // scale grayscale values to 0..255 range + stbi_uc *in = cur; + stbi_uc *out = dest; + stbi_uc inb = 0; + stbi__uint32 nsmp = x*img_n; + + // expand bits to bytes first + if (depth == 4) { + for (i=0; i < nsmp; ++i) { + if ((i & 1) == 0) inb = *in++; + *out++ = scale * (inb >> 4); + inb <<= 4; + } + } else if (depth == 2) { + for (i=0; i < nsmp; ++i) { + if ((i & 3) == 0) inb = *in++; + *out++ = scale * (inb >> 6); + inb <<= 2; + } + } else { + STBI_ASSERT(depth == 1); + for (i=0; i < nsmp; ++i) { + if ((i & 7) == 0) inb = *in++; + *out++ = scale * (inb >> 7); + inb <<= 1; + } + } + + // insert alpha=255 values if desired + if (img_n != out_n) + stbi__create_png_alpha_expand8(dest, dest, x, img_n); + } else if (depth == 8) { + if (img_n == out_n) + memcpy(dest, cur, x*img_n); + else + stbi__create_png_alpha_expand8(dest, cur, x, img_n); + } else if (depth == 16) { + // convert the image data from big-endian to platform-native + stbi__uint16 *dest16 = (stbi__uint16*)dest; + stbi__uint32 nsmp = x*img_n; + + if (img_n == out_n) { + for (i = 0; i < nsmp; ++i, ++dest16, cur += 2) + *dest16 = (cur[0] << 8) | cur[1]; + } else { + STBI_ASSERT(img_n+1 == out_n); + if (img_n == 1) { + for (i = 0; i < x; ++i, dest16 += 2, cur += 2) { + dest16[0] = (cur[0] << 8) | cur[1]; + dest16[1] = 0xffff; + } + } else { + STBI_ASSERT(img_n == 3); + for (i = 0; i < x; ++i, dest16 += 4, cur += 6) { + dest16[0] = (cur[0] << 8) | cur[1]; + dest16[1] = (cur[2] << 8) | cur[3]; + dest16[2] = (cur[4] << 8) | cur[5]; + dest16[3] = 0xffff; + } + } + } + } + } + + STBI_FREE(filter_buf); + if (!all_ok) return 0; + + return 1; +} + +static int stbi__create_png_image(stbi__png *a, stbi_uc *image_data, stbi__uint32 image_data_len, int out_n, int depth, int color, int interlaced) +{ + int bytes = (depth == 16 ? 2 : 1); + int out_bytes = out_n * bytes; + stbi_uc *final; + int p; + if (!interlaced) + return stbi__create_png_image_raw(a, image_data, image_data_len, out_n, a->s->img_x, a->s->img_y, depth, color); + + // de-interlacing + final = (stbi_uc *) stbi__malloc_mad3(a->s->img_x, a->s->img_y, out_bytes, 0); + if (!final) return stbi__err("outofmem", "Out of memory"); + for (p=0; p < 7; ++p) { + int xorig[] = { 0,4,0,2,0,1,0 }; + int yorig[] = { 0,0,4,0,2,0,1 }; + int xspc[] = { 8,8,4,4,2,2,1 }; + int yspc[] = { 8,8,8,4,4,2,2 }; + int i,j,x,y; + // pass1_x[4] = 0, pass1_x[5] = 1, pass1_x[12] = 1 + x = (a->s->img_x - xorig[p] + xspc[p]-1) / xspc[p]; + y = (a->s->img_y - yorig[p] + yspc[p]-1) / yspc[p]; + if (x && y) { + stbi__uint32 img_len = ((((a->s->img_n * x * depth) + 7) >> 3) + 1) * y; + if (!stbi__create_png_image_raw(a, image_data, image_data_len, out_n, x, y, depth, color)) { + STBI_FREE(final); + return 0; + } + for (j=0; j < y; ++j) { + for (i=0; i < x; ++i) { + int out_y = j*yspc[p]+yorig[p]; + int out_x = i*xspc[p]+xorig[p]; + memcpy(final + out_y*a->s->img_x*out_bytes + out_x*out_bytes, + a->out + (j*x+i)*out_bytes, out_bytes); + } + } + STBI_FREE(a->out); + image_data += img_len; + image_data_len -= img_len; + } + } + a->out = final; + + return 1; +} + +static int stbi__compute_transparency(stbi__png *z, stbi_uc tc[3], int out_n) +{ + stbi__context *s = z->s; + stbi__uint32 i, pixel_count = s->img_x * s->img_y; + stbi_uc *p = z->out; + + // compute color-based transparency, assuming we've + // already got 255 as the alpha value in the output + STBI_ASSERT(out_n == 2 || out_n == 4); + + if (out_n == 2) { + for (i=0; i < pixel_count; ++i) { + p[1] = (p[0] == tc[0] ? 0 : 255); + p += 2; + } + } else { + for (i=0; i < pixel_count; ++i) { + if (p[0] == tc[0] && p[1] == tc[1] && p[2] == tc[2]) + p[3] = 0; + p += 4; + } + } + return 1; +} + +static int stbi__compute_transparency16(stbi__png *z, stbi__uint16 tc[3], int out_n) +{ + stbi__context *s = z->s; + stbi__uint32 i, pixel_count = s->img_x * s->img_y; + stbi__uint16 *p = (stbi__uint16*) z->out; + + // compute color-based transparency, assuming we've + // already got 65535 as the alpha value in the output + STBI_ASSERT(out_n == 2 || out_n == 4); + + if (out_n == 2) { + for (i = 0; i < pixel_count; ++i) { + p[1] = (p[0] == tc[0] ? 0 : 65535); + p += 2; + } + } else { + for (i = 0; i < pixel_count; ++i) { + if (p[0] == tc[0] && p[1] == tc[1] && p[2] == tc[2]) + p[3] = 0; + p += 4; + } + } + return 1; +} + +static int stbi__expand_png_palette(stbi__png *a, stbi_uc *palette, int len, int pal_img_n) +{ + stbi__uint32 i, pixel_count = a->s->img_x * a->s->img_y; + stbi_uc *p, *temp_out, *orig = a->out; + + p = (stbi_uc *) stbi__malloc_mad2(pixel_count, pal_img_n, 0); + if (p == NULL) return stbi__err("outofmem", "Out of memory"); + + // between here and free(out) below, exitting would leak + temp_out = p; + + if (pal_img_n == 3) { + for (i=0; i < pixel_count; ++i) { + int n = orig[i]*4; + p[0] = palette[n ]; + p[1] = palette[n+1]; + p[2] = palette[n+2]; + p += 3; + } + } else { + for (i=0; i < pixel_count; ++i) { + int n = orig[i]*4; + p[0] = palette[n ]; + p[1] = palette[n+1]; + p[2] = palette[n+2]; + p[3] = palette[n+3]; + p += 4; + } + } + STBI_FREE(a->out); + a->out = temp_out; + + STBI_NOTUSED(len); + + return 1; +} + +static int stbi__unpremultiply_on_load_global = 0; +static int stbi__de_iphone_flag_global = 0; + +STBIDEF void stbi_set_unpremultiply_on_load(int flag_true_if_should_unpremultiply) +{ + stbi__unpremultiply_on_load_global = flag_true_if_should_unpremultiply; +} + +STBIDEF void stbi_convert_iphone_png_to_rgb(int flag_true_if_should_convert) +{ + stbi__de_iphone_flag_global = flag_true_if_should_convert; +} + +#ifndef STBI_THREAD_LOCAL +#define stbi__unpremultiply_on_load stbi__unpremultiply_on_load_global +#define stbi__de_iphone_flag stbi__de_iphone_flag_global +#else +static STBI_THREAD_LOCAL int stbi__unpremultiply_on_load_local, stbi__unpremultiply_on_load_set; +static STBI_THREAD_LOCAL int stbi__de_iphone_flag_local, stbi__de_iphone_flag_set; + +STBIDEF void stbi_set_unpremultiply_on_load_thread(int flag_true_if_should_unpremultiply) +{ + stbi__unpremultiply_on_load_local = flag_true_if_should_unpremultiply; + stbi__unpremultiply_on_load_set = 1; +} + +STBIDEF void stbi_convert_iphone_png_to_rgb_thread(int flag_true_if_should_convert) +{ + stbi__de_iphone_flag_local = flag_true_if_should_convert; + stbi__de_iphone_flag_set = 1; +} + +#define stbi__unpremultiply_on_load (stbi__unpremultiply_on_load_set \ + ? stbi__unpremultiply_on_load_local \ + : stbi__unpremultiply_on_load_global) +#define stbi__de_iphone_flag (stbi__de_iphone_flag_set \ + ? stbi__de_iphone_flag_local \ + : stbi__de_iphone_flag_global) +#endif // STBI_THREAD_LOCAL + +static void stbi__de_iphone(stbi__png *z) +{ + stbi__context *s = z->s; + stbi__uint32 i, pixel_count = s->img_x * s->img_y; + stbi_uc *p = z->out; + + if (s->img_out_n == 3) { // convert bgr to rgb + for (i=0; i < pixel_count; ++i) { + stbi_uc t = p[0]; + p[0] = p[2]; + p[2] = t; + p += 3; + } + } else { + STBI_ASSERT(s->img_out_n == 4); + if (stbi__unpremultiply_on_load) { + // convert bgr to rgb and unpremultiply + for (i=0; i < pixel_count; ++i) { + stbi_uc a = p[3]; + stbi_uc t = p[0]; + if (a) { + stbi_uc half = a / 2; + p[0] = (p[2] * 255 + half) / a; + p[1] = (p[1] * 255 + half) / a; + p[2] = ( t * 255 + half) / a; + } else { + p[0] = p[2]; + p[2] = t; + } + p += 4; + } + } else { + // convert bgr to rgb + for (i=0; i < pixel_count; ++i) { + stbi_uc t = p[0]; + p[0] = p[2]; + p[2] = t; + p += 4; + } + } + } +} + +#define STBI__PNG_TYPE(a,b,c,d) (((unsigned) (a) << 24) + ((unsigned) (b) << 16) + ((unsigned) (c) << 8) + (unsigned) (d)) + +static int stbi__parse_png_file(stbi__png *z, int scan, int req_comp) +{ + stbi_uc palette[1024], pal_img_n=0; + stbi_uc has_trans=0, tc[3]={0}; + stbi__uint16 tc16[3]; + stbi__uint32 ioff=0, idata_limit=0, i, pal_len=0; + int first=1,k,interlace=0, color=0, is_iphone=0; + stbi__context *s = z->s; + + z->expanded = NULL; + z->idata = NULL; + z->out = NULL; + + if (!stbi__check_png_header(s)) return 0; + + if (scan == STBI__SCAN_type) return 1; + + for (;;) { + stbi__pngchunk c = stbi__get_chunk_header(s); + switch (c.type) { + case STBI__PNG_TYPE('C','g','B','I'): + is_iphone = 1; + stbi__skip(s, c.length); + break; + case STBI__PNG_TYPE('I','H','D','R'): { + int comp,filter; + if (!first) return stbi__err("multiple IHDR","Corrupt PNG"); + first = 0; + if (c.length != 13) return stbi__err("bad IHDR len","Corrupt PNG"); + s->img_x = stbi__get32be(s); + s->img_y = stbi__get32be(s); + if (s->img_y > STBI_MAX_DIMENSIONS) return stbi__err("too large","Very large image (corrupt?)"); + if (s->img_x > STBI_MAX_DIMENSIONS) return stbi__err("too large","Very large image (corrupt?)"); + z->depth = stbi__get8(s); if (z->depth != 1 && z->depth != 2 && z->depth != 4 && z->depth != 8 && z->depth != 16) return stbi__err("1/2/4/8/16-bit only","PNG not supported: 1/2/4/8/16-bit only"); + color = stbi__get8(s); if (color > 6) return stbi__err("bad ctype","Corrupt PNG"); + if (color == 3 && z->depth == 16) return stbi__err("bad ctype","Corrupt PNG"); + if (color == 3) pal_img_n = 3; else if (color & 1) return stbi__err("bad ctype","Corrupt PNG"); + comp = stbi__get8(s); if (comp) return stbi__err("bad comp method","Corrupt PNG"); + filter= stbi__get8(s); if (filter) return stbi__err("bad filter method","Corrupt PNG"); + interlace = stbi__get8(s); if (interlace>1) return stbi__err("bad interlace method","Corrupt PNG"); + if (!s->img_x || !s->img_y) return stbi__err("0-pixel image","Corrupt PNG"); + if (!pal_img_n) { + s->img_n = (color & 2 ? 3 : 1) + (color & 4 ? 1 : 0); + if ((1 << 30) / s->img_x / s->img_n < s->img_y) return stbi__err("too large", "Image too large to decode"); + } else { + // if paletted, then pal_n is our final components, and + // img_n is # components to decompress/filter. + s->img_n = 1; + if ((1 << 30) / s->img_x / 4 < s->img_y) return stbi__err("too large","Corrupt PNG"); + } + // even with SCAN_header, have to scan to see if we have a tRNS + break; + } + + case STBI__PNG_TYPE('P','L','T','E'): { + if (first) return stbi__err("first not IHDR", "Corrupt PNG"); + if (c.length > 256*3) return stbi__err("invalid PLTE","Corrupt PNG"); + pal_len = c.length / 3; + if (pal_len * 3 != c.length) return stbi__err("invalid PLTE","Corrupt PNG"); + for (i=0; i < pal_len; ++i) { + palette[i*4+0] = stbi__get8(s); + palette[i*4+1] = stbi__get8(s); + palette[i*4+2] = stbi__get8(s); + palette[i*4+3] = 255; + } + break; + } + + case STBI__PNG_TYPE('t','R','N','S'): { + if (first) return stbi__err("first not IHDR", "Corrupt PNG"); + if (z->idata) return stbi__err("tRNS after IDAT","Corrupt PNG"); + if (pal_img_n) { + if (scan == STBI__SCAN_header) { s->img_n = 4; return 1; } + if (pal_len == 0) return stbi__err("tRNS before PLTE","Corrupt PNG"); + if (c.length > pal_len) return stbi__err("bad tRNS len","Corrupt PNG"); + pal_img_n = 4; + for (i=0; i < c.length; ++i) + palette[i*4+3] = stbi__get8(s); + } else { + if (!(s->img_n & 1)) return stbi__err("tRNS with alpha","Corrupt PNG"); + if (c.length != (stbi__uint32) s->img_n*2) return stbi__err("bad tRNS len","Corrupt PNG"); + has_trans = 1; + // non-paletted with tRNS = constant alpha. if header-scanning, we can stop now. + if (scan == STBI__SCAN_header) { ++s->img_n; return 1; } + if (z->depth == 16) { + for (k = 0; k < s->img_n && k < 3; ++k) // extra loop test to suppress false GCC warning + tc16[k] = (stbi__uint16)stbi__get16be(s); // copy the values as-is + } else { + for (k = 0; k < s->img_n && k < 3; ++k) + tc[k] = (stbi_uc)(stbi__get16be(s) & 255) * stbi__depth_scale_table[z->depth]; // non 8-bit images will be larger + } + } + break; + } + + case STBI__PNG_TYPE('I','D','A','T'): { + if (first) return stbi__err("first not IHDR", "Corrupt PNG"); + if (pal_img_n && !pal_len) return stbi__err("no PLTE","Corrupt PNG"); + if (scan == STBI__SCAN_header) { + // header scan definitely stops at first IDAT + if (pal_img_n) + s->img_n = pal_img_n; + return 1; + } + if (c.length > (1u << 30)) return stbi__err("IDAT size limit", "IDAT section larger than 2^30 bytes"); + if ((int)(ioff + c.length) < (int)ioff) return 0; + if (ioff + c.length > idata_limit) { + stbi__uint32 idata_limit_old = idata_limit; + stbi_uc *p; + if (idata_limit == 0) idata_limit = c.length > 4096 ? c.length : 4096; + while (ioff + c.length > idata_limit) + idata_limit *= 2; + STBI_NOTUSED(idata_limit_old); + p = (stbi_uc *) STBI_REALLOC_SIZED(z->idata, idata_limit_old, idata_limit); if (p == NULL) return stbi__err("outofmem", "Out of memory"); + z->idata = p; + } + if (!stbi__getn(s, z->idata+ioff,c.length)) return stbi__err("outofdata","Corrupt PNG"); + ioff += c.length; + break; + } + + case STBI__PNG_TYPE('I','E','N','D'): { + stbi__uint32 raw_len, bpl; + if (first) return stbi__err("first not IHDR", "Corrupt PNG"); + if (scan != STBI__SCAN_load) return 1; + if (z->idata == NULL) return stbi__err("no IDAT","Corrupt PNG"); + // initial guess for decoded data size to avoid unnecessary reallocs + bpl = (s->img_x * z->depth + 7) / 8; // bytes per line, per component + raw_len = bpl * s->img_y * s->img_n /* pixels */ + s->img_y /* filter mode per row */; + z->expanded = (stbi_uc *) stbi_zlib_decode_malloc_guesssize_headerflag((char *) z->idata, ioff, raw_len, (int *) &raw_len, !is_iphone); + if (z->expanded == NULL) return 0; // zlib should set error + STBI_FREE(z->idata); z->idata = NULL; + if ((req_comp == s->img_n+1 && req_comp != 3 && !pal_img_n) || has_trans) + s->img_out_n = s->img_n+1; + else + s->img_out_n = s->img_n; + if (!stbi__create_png_image(z, z->expanded, raw_len, s->img_out_n, z->depth, color, interlace)) return 0; + if (has_trans) { + if (z->depth == 16) { + if (!stbi__compute_transparency16(z, tc16, s->img_out_n)) return 0; + } else { + if (!stbi__compute_transparency(z, tc, s->img_out_n)) return 0; + } + } + if (is_iphone && stbi__de_iphone_flag && s->img_out_n > 2) + stbi__de_iphone(z); + if (pal_img_n) { + // pal_img_n == 3 or 4 + s->img_n = pal_img_n; // record the actual colors we had + s->img_out_n = pal_img_n; + if (req_comp >= 3) s->img_out_n = req_comp; + if (!stbi__expand_png_palette(z, palette, pal_len, s->img_out_n)) + return 0; + } else if (has_trans) { + // non-paletted image with tRNS -> source image has (constant) alpha + ++s->img_n; + } + STBI_FREE(z->expanded); z->expanded = NULL; + // end of PNG chunk, read and skip CRC + stbi__get32be(s); + return 1; + } + + default: + // if critical, fail + if (first) return stbi__err("first not IHDR", "Corrupt PNG"); + if ((c.type & (1 << 29)) == 0) { + #ifndef STBI_NO_FAILURE_STRINGS + // not threadsafe + static char invalid_chunk[] = "XXXX PNG chunk not known"; + invalid_chunk[0] = STBI__BYTECAST(c.type >> 24); + invalid_chunk[1] = STBI__BYTECAST(c.type >> 16); + invalid_chunk[2] = STBI__BYTECAST(c.type >> 8); + invalid_chunk[3] = STBI__BYTECAST(c.type >> 0); + #endif + return stbi__err(invalid_chunk, "PNG not supported: unknown PNG chunk type"); + } + stbi__skip(s, c.length); + break; + } + // end of PNG chunk, read and skip CRC + stbi__get32be(s); + } +} + +static void *stbi__do_png(stbi__png *p, int *x, int *y, int *n, int req_comp, stbi__result_info *ri) +{ + void *result=NULL; + if (req_comp < 0 || req_comp > 4) return stbi__errpuc("bad req_comp", "Internal error"); + if (stbi__parse_png_file(p, STBI__SCAN_load, req_comp)) { + if (p->depth <= 8) + ri->bits_per_channel = 8; + else if (p->depth == 16) + ri->bits_per_channel = 16; + else + return stbi__errpuc("bad bits_per_channel", "PNG not supported: unsupported color depth"); + result = p->out; + p->out = NULL; + if (req_comp && req_comp != p->s->img_out_n) { + if (ri->bits_per_channel == 8) + result = stbi__convert_format((unsigned char *) result, p->s->img_out_n, req_comp, p->s->img_x, p->s->img_y); + else + result = stbi__convert_format16((stbi__uint16 *) result, p->s->img_out_n, req_comp, p->s->img_x, p->s->img_y); + p->s->img_out_n = req_comp; + if (result == NULL) return result; + } + *x = p->s->img_x; + *y = p->s->img_y; + if (n) *n = p->s->img_n; + } + STBI_FREE(p->out); p->out = NULL; + STBI_FREE(p->expanded); p->expanded = NULL; + STBI_FREE(p->idata); p->idata = NULL; + + return result; +} + +static void *stbi__png_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri) +{ + stbi__png p; + p.s = s; + return stbi__do_png(&p, x,y,comp,req_comp, ri); +} + +static int stbi__png_test(stbi__context *s) +{ + int r; + r = stbi__check_png_header(s); + stbi__rewind(s); + return r; +} + +static int stbi__png_info_raw(stbi__png *p, int *x, int *y, int *comp) +{ + if (!stbi__parse_png_file(p, STBI__SCAN_header, 0)) { + stbi__rewind( p->s ); + return 0; + } + if (x) *x = p->s->img_x; + if (y) *y = p->s->img_y; + if (comp) *comp = p->s->img_n; + return 1; +} + +static int stbi__png_info(stbi__context *s, int *x, int *y, int *comp) +{ + stbi__png p; + p.s = s; + return stbi__png_info_raw(&p, x, y, comp); +} + +static int stbi__png_is16(stbi__context *s) +{ + stbi__png p; + p.s = s; + if (!stbi__png_info_raw(&p, NULL, NULL, NULL)) + return 0; + if (p.depth != 16) { + stbi__rewind(p.s); + return 0; + } + return 1; +} +#endif + +// Microsoft/Windows BMP image + +#ifndef STBI_NO_BMP +static int stbi__bmp_test_raw(stbi__context *s) +{ + int r; + int sz; + if (stbi__get8(s) != 'B') return 0; + if (stbi__get8(s) != 'M') return 0; + stbi__get32le(s); // discard filesize + stbi__get16le(s); // discard reserved + stbi__get16le(s); // discard reserved + stbi__get32le(s); // discard data offset + sz = stbi__get32le(s); + r = (sz == 12 || sz == 40 || sz == 56 || sz == 108 || sz == 124); + return r; +} + +static int stbi__bmp_test(stbi__context *s) +{ + int r = stbi__bmp_test_raw(s); + stbi__rewind(s); + return r; +} + + +// returns 0..31 for the highest set bit +static int stbi__high_bit(unsigned int z) +{ + int n=0; + if (z == 0) return -1; + if (z >= 0x10000) { n += 16; z >>= 16; } + if (z >= 0x00100) { n += 8; z >>= 8; } + if (z >= 0x00010) { n += 4; z >>= 4; } + if (z >= 0x00004) { n += 2; z >>= 2; } + if (z >= 0x00002) { n += 1;/* >>= 1;*/ } + return n; +} + +static int stbi__bitcount(unsigned int a) +{ + a = (a & 0x55555555) + ((a >> 1) & 0x55555555); // max 2 + a = (a & 0x33333333) + ((a >> 2) & 0x33333333); // max 4 + a = (a + (a >> 4)) & 0x0f0f0f0f; // max 8 per 4, now 8 bits + a = (a + (a >> 8)); // max 16 per 8 bits + a = (a + (a >> 16)); // max 32 per 8 bits + return a & 0xff; +} + +// extract an arbitrarily-aligned N-bit value (N=bits) +// from v, and then make it 8-bits long and fractionally +// extend it to full full range. +static int stbi__shiftsigned(unsigned int v, int shift, int bits) +{ + static unsigned int mul_table[9] = { + 0, + 0xff/*0b11111111*/, 0x55/*0b01010101*/, 0x49/*0b01001001*/, 0x11/*0b00010001*/, + 0x21/*0b00100001*/, 0x41/*0b01000001*/, 0x81/*0b10000001*/, 0x01/*0b00000001*/, + }; + static unsigned int shift_table[9] = { + 0, 0,0,1,0,2,4,6,0, + }; + if (shift < 0) + v <<= -shift; + else + v >>= shift; + STBI_ASSERT(v < 256); + v >>= (8-bits); + STBI_ASSERT(bits >= 0 && bits <= 8); + return (int) ((unsigned) v * mul_table[bits]) >> shift_table[bits]; +} + +typedef struct +{ + int bpp, offset, hsz; + unsigned int mr,mg,mb,ma, all_a; + int extra_read; +} stbi__bmp_data; + +static int stbi__bmp_set_mask_defaults(stbi__bmp_data *info, int compress) +{ + // BI_BITFIELDS specifies masks explicitly, don't override + if (compress == 3) + return 1; + + if (compress == 0) { + if (info->bpp == 16) { + info->mr = 31u << 10; + info->mg = 31u << 5; + info->mb = 31u << 0; + } else if (info->bpp == 32) { + info->mr = 0xffu << 16; + info->mg = 0xffu << 8; + info->mb = 0xffu << 0; + info->ma = 0xffu << 24; + info->all_a = 0; // if all_a is 0 at end, then we loaded alpha channel but it was all 0 + } else { + // otherwise, use defaults, which is all-0 + info->mr = info->mg = info->mb = info->ma = 0; + } + return 1; + } + return 0; // error +} + +static void *stbi__bmp_parse_header(stbi__context *s, stbi__bmp_data *info) +{ + int hsz; + if (stbi__get8(s) != 'B' || stbi__get8(s) != 'M') return stbi__errpuc("not BMP", "Corrupt BMP"); + stbi__get32le(s); // discard filesize + stbi__get16le(s); // discard reserved + stbi__get16le(s); // discard reserved + info->offset = stbi__get32le(s); + info->hsz = hsz = stbi__get32le(s); + info->mr = info->mg = info->mb = info->ma = 0; + info->extra_read = 14; + + if (info->offset < 0) return stbi__errpuc("bad BMP", "bad BMP"); + + if (hsz != 12 && hsz != 40 && hsz != 56 && hsz != 108 && hsz != 124) return stbi__errpuc("unknown BMP", "BMP type not supported: unknown"); + if (hsz == 12) { + s->img_x = stbi__get16le(s); + s->img_y = stbi__get16le(s); + } else { + s->img_x = stbi__get32le(s); + s->img_y = stbi__get32le(s); + } + if (stbi__get16le(s) != 1) return stbi__errpuc("bad BMP", "bad BMP"); + info->bpp = stbi__get16le(s); + if (hsz != 12) { + int compress = stbi__get32le(s); + if (compress == 1 || compress == 2) return stbi__errpuc("BMP RLE", "BMP type not supported: RLE"); + if (compress >= 4) return stbi__errpuc("BMP JPEG/PNG", "BMP type not supported: unsupported compression"); // this includes PNG/JPEG modes + if (compress == 3 && info->bpp != 16 && info->bpp != 32) return stbi__errpuc("bad BMP", "bad BMP"); // bitfields requires 16 or 32 bits/pixel + stbi__get32le(s); // discard sizeof + stbi__get32le(s); // discard hres + stbi__get32le(s); // discard vres + stbi__get32le(s); // discard colorsused + stbi__get32le(s); // discard max important + if (hsz == 40 || hsz == 56) { + if (hsz == 56) { + stbi__get32le(s); + stbi__get32le(s); + stbi__get32le(s); + stbi__get32le(s); + } + if (info->bpp == 16 || info->bpp == 32) { + if (compress == 0) { + stbi__bmp_set_mask_defaults(info, compress); + } else if (compress == 3) { + info->mr = stbi__get32le(s); + info->mg = stbi__get32le(s); + info->mb = stbi__get32le(s); + info->extra_read += 12; + // not documented, but generated by photoshop and handled by mspaint + if (info->mr == info->mg && info->mg == info->mb) { + // ?!?!? + return stbi__errpuc("bad BMP", "bad BMP"); + } + } else + return stbi__errpuc("bad BMP", "bad BMP"); + } + } else { + // V4/V5 header + int i; + if (hsz != 108 && hsz != 124) + return stbi__errpuc("bad BMP", "bad BMP"); + info->mr = stbi__get32le(s); + info->mg = stbi__get32le(s); + info->mb = stbi__get32le(s); + info->ma = stbi__get32le(s); + if (compress != 3) // override mr/mg/mb unless in BI_BITFIELDS mode, as per docs + stbi__bmp_set_mask_defaults(info, compress); + stbi__get32le(s); // discard color space + for (i=0; i < 12; ++i) + stbi__get32le(s); // discard color space parameters + if (hsz == 124) { + stbi__get32le(s); // discard rendering intent + stbi__get32le(s); // discard offset of profile data + stbi__get32le(s); // discard size of profile data + stbi__get32le(s); // discard reserved + } + } + } + return (void *) 1; +} + + +static void *stbi__bmp_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri) +{ + stbi_uc *out; + unsigned int mr=0,mg=0,mb=0,ma=0, all_a; + stbi_uc pal[256][4]; + int psize=0,i,j,width; + int flip_vertically, pad, target; + stbi__bmp_data info; + STBI_NOTUSED(ri); + + info.all_a = 255; + if (stbi__bmp_parse_header(s, &info) == NULL) + return NULL; // error code already set + + flip_vertically = ((int) s->img_y) > 0; + s->img_y = abs((int) s->img_y); + + if (s->img_y > STBI_MAX_DIMENSIONS) return stbi__errpuc("too large","Very large image (corrupt?)"); + if (s->img_x > STBI_MAX_DIMENSIONS) return stbi__errpuc("too large","Very large image (corrupt?)"); + + mr = info.mr; + mg = info.mg; + mb = info.mb; + ma = info.ma; + all_a = info.all_a; + + if (info.hsz == 12) { + if (info.bpp < 24) + psize = (info.offset - info.extra_read - 24) / 3; + } else { + if (info.bpp < 16) + psize = (info.offset - info.extra_read - info.hsz) >> 2; + } + if (psize == 0) { + // accept some number of extra bytes after the header, but if the offset points either to before + // the header ends or implies a large amount of extra data, reject the file as malformed + int bytes_read_so_far = s->callback_already_read + (int)(s->img_buffer - s->img_buffer_original); + int header_limit = 1024; // max we actually read is below 256 bytes currently. + int extra_data_limit = 256*4; // what ordinarily goes here is a palette; 256 entries*4 bytes is its max size. + if (bytes_read_so_far <= 0 || bytes_read_so_far > header_limit) { + return stbi__errpuc("bad header", "Corrupt BMP"); + } + // we established that bytes_read_so_far is positive and sensible. + // the first half of this test rejects offsets that are either too small positives, or + // negative, and guarantees that info.offset >= bytes_read_so_far > 0. this in turn + // ensures the number computed in the second half of the test can't overflow. + if (info.offset < bytes_read_so_far || info.offset - bytes_read_so_far > extra_data_limit) { + return stbi__errpuc("bad offset", "Corrupt BMP"); + } else { + stbi__skip(s, info.offset - bytes_read_so_far); + } + } + + if (info.bpp == 24 && ma == 0xff000000) + s->img_n = 3; + else + s->img_n = ma ? 4 : 3; + if (req_comp && req_comp >= 3) // we can directly decode 3 or 4 + target = req_comp; + else + target = s->img_n; // if they want monochrome, we'll post-convert + + // sanity-check size + if (!stbi__mad3sizes_valid(target, s->img_x, s->img_y, 0)) + return stbi__errpuc("too large", "Corrupt BMP"); + + out = (stbi_uc *) stbi__malloc_mad3(target, s->img_x, s->img_y, 0); + if (!out) return stbi__errpuc("outofmem", "Out of memory"); + if (info.bpp < 16) { + int z=0; + if (psize == 0 || psize > 256) { STBI_FREE(out); return stbi__errpuc("invalid", "Corrupt BMP"); } + for (i=0; i < psize; ++i) { + pal[i][2] = stbi__get8(s); + pal[i][1] = stbi__get8(s); + pal[i][0] = stbi__get8(s); + if (info.hsz != 12) stbi__get8(s); + pal[i][3] = 255; + } + stbi__skip(s, info.offset - info.extra_read - info.hsz - psize * (info.hsz == 12 ? 3 : 4)); + if (info.bpp == 1) width = (s->img_x + 7) >> 3; + else if (info.bpp == 4) width = (s->img_x + 1) >> 1; + else if (info.bpp == 8) width = s->img_x; + else { STBI_FREE(out); return stbi__errpuc("bad bpp", "Corrupt BMP"); } + pad = (-width)&3; + if (info.bpp == 1) { + for (j=0; j < (int) s->img_y; ++j) { + int bit_offset = 7, v = stbi__get8(s); + for (i=0; i < (int) s->img_x; ++i) { + int color = (v>>bit_offset)&0x1; + out[z++] = pal[color][0]; + out[z++] = pal[color][1]; + out[z++] = pal[color][2]; + if (target == 4) out[z++] = 255; + if (i+1 == (int) s->img_x) break; + if((--bit_offset) < 0) { + bit_offset = 7; + v = stbi__get8(s); + } + } + stbi__skip(s, pad); + } + } else { + for (j=0; j < (int) s->img_y; ++j) { + for (i=0; i < (int) s->img_x; i += 2) { + int v=stbi__get8(s),v2=0; + if (info.bpp == 4) { + v2 = v & 15; + v >>= 4; + } + out[z++] = pal[v][0]; + out[z++] = pal[v][1]; + out[z++] = pal[v][2]; + if (target == 4) out[z++] = 255; + if (i+1 == (int) s->img_x) break; + v = (info.bpp == 8) ? stbi__get8(s) : v2; + out[z++] = pal[v][0]; + out[z++] = pal[v][1]; + out[z++] = pal[v][2]; + if (target == 4) out[z++] = 255; + } + stbi__skip(s, pad); + } + } + } else { + int rshift=0,gshift=0,bshift=0,ashift=0,rcount=0,gcount=0,bcount=0,acount=0; + int z = 0; + int easy=0; + stbi__skip(s, info.offset - info.extra_read - info.hsz); + if (info.bpp == 24) width = 3 * s->img_x; + else if (info.bpp == 16) width = 2*s->img_x; + else /* bpp = 32 and pad = 0 */ width=0; + pad = (-width) & 3; + if (info.bpp == 24) { + easy = 1; + } else if (info.bpp == 32) { + if (mb == 0xff && mg == 0xff00 && mr == 0x00ff0000 && ma == 0xff000000) + easy = 2; + } + if (!easy) { + if (!mr || !mg || !mb) { STBI_FREE(out); return stbi__errpuc("bad masks", "Corrupt BMP"); } + // right shift amt to put high bit in position #7 + rshift = stbi__high_bit(mr)-7; rcount = stbi__bitcount(mr); + gshift = stbi__high_bit(mg)-7; gcount = stbi__bitcount(mg); + bshift = stbi__high_bit(mb)-7; bcount = stbi__bitcount(mb); + ashift = stbi__high_bit(ma)-7; acount = stbi__bitcount(ma); + if (rcount > 8 || gcount > 8 || bcount > 8 || acount > 8) { STBI_FREE(out); return stbi__errpuc("bad masks", "Corrupt BMP"); } + } + for (j=0; j < (int) s->img_y; ++j) { + if (easy) { + for (i=0; i < (int) s->img_x; ++i) { + unsigned char a; + out[z+2] = stbi__get8(s); + out[z+1] = stbi__get8(s); + out[z+0] = stbi__get8(s); + z += 3; + a = (easy == 2 ? stbi__get8(s) : 255); + all_a |= a; + if (target == 4) out[z++] = a; + } + } else { + int bpp = info.bpp; + for (i=0; i < (int) s->img_x; ++i) { + stbi__uint32 v = (bpp == 16 ? (stbi__uint32) stbi__get16le(s) : stbi__get32le(s)); + unsigned int a; + out[z++] = STBI__BYTECAST(stbi__shiftsigned(v & mr, rshift, rcount)); + out[z++] = STBI__BYTECAST(stbi__shiftsigned(v & mg, gshift, gcount)); + out[z++] = STBI__BYTECAST(stbi__shiftsigned(v & mb, bshift, bcount)); + a = (ma ? stbi__shiftsigned(v & ma, ashift, acount) : 255); + all_a |= a; + if (target == 4) out[z++] = STBI__BYTECAST(a); + } + } + stbi__skip(s, pad); + } + } + + // if alpha channel is all 0s, replace with all 255s + if (target == 4 && all_a == 0) + for (i=4*s->img_x*s->img_y-1; i >= 0; i -= 4) + out[i] = 255; + + if (flip_vertically) { + stbi_uc t; + for (j=0; j < (int) s->img_y>>1; ++j) { + stbi_uc *p1 = out + j *s->img_x*target; + stbi_uc *p2 = out + (s->img_y-1-j)*s->img_x*target; + for (i=0; i < (int) s->img_x*target; ++i) { + t = p1[i]; p1[i] = p2[i]; p2[i] = t; + } + } + } + + if (req_comp && req_comp != target) { + out = stbi__convert_format(out, target, req_comp, s->img_x, s->img_y); + if (out == NULL) return out; // stbi__convert_format frees input on failure + } + + *x = s->img_x; + *y = s->img_y; + if (comp) *comp = s->img_n; + return out; +} +#endif + +// Targa Truevision - TGA +// by Jonathan Dummer +#ifndef STBI_NO_TGA +// returns STBI_rgb or whatever, 0 on error +static int stbi__tga_get_comp(int bits_per_pixel, int is_grey, int* is_rgb16) +{ + // only RGB or RGBA (incl. 16bit) or grey allowed + if (is_rgb16) *is_rgb16 = 0; + switch(bits_per_pixel) { + case 8: return STBI_grey; + case 16: if(is_grey) return STBI_grey_alpha; + // fallthrough + case 15: if(is_rgb16) *is_rgb16 = 1; + return STBI_rgb; + case 24: // fallthrough + case 32: return bits_per_pixel/8; + default: return 0; + } +} + +static int stbi__tga_info(stbi__context *s, int *x, int *y, int *comp) +{ + int tga_w, tga_h, tga_comp, tga_image_type, tga_bits_per_pixel, tga_colormap_bpp; + int sz, tga_colormap_type; + stbi__get8(s); // discard Offset + tga_colormap_type = stbi__get8(s); // colormap type + if( tga_colormap_type > 1 ) { + stbi__rewind(s); + return 0; // only RGB or indexed allowed + } + tga_image_type = stbi__get8(s); // image type + if ( tga_colormap_type == 1 ) { // colormapped (paletted) image + if (tga_image_type != 1 && tga_image_type != 9) { + stbi__rewind(s); + return 0; + } + stbi__skip(s,4); // skip index of first colormap entry and number of entries + sz = stbi__get8(s); // check bits per palette color entry + if ( (sz != 8) && (sz != 15) && (sz != 16) && (sz != 24) && (sz != 32) ) { + stbi__rewind(s); + return 0; + } + stbi__skip(s,4); // skip image x and y origin + tga_colormap_bpp = sz; + } else { // "normal" image w/o colormap - only RGB or grey allowed, +/- RLE + if ( (tga_image_type != 2) && (tga_image_type != 3) && (tga_image_type != 10) && (tga_image_type != 11) ) { + stbi__rewind(s); + return 0; // only RGB or grey allowed, +/- RLE + } + stbi__skip(s,9); // skip colormap specification and image x/y origin + tga_colormap_bpp = 0; + } + tga_w = stbi__get16le(s); + if( tga_w < 1 ) { + stbi__rewind(s); + return 0; // test width + } + tga_h = stbi__get16le(s); + if( tga_h < 1 ) { + stbi__rewind(s); + return 0; // test height + } + tga_bits_per_pixel = stbi__get8(s); // bits per pixel + stbi__get8(s); // ignore alpha bits + if (tga_colormap_bpp != 0) { + if((tga_bits_per_pixel != 8) && (tga_bits_per_pixel != 16)) { + // when using a colormap, tga_bits_per_pixel is the size of the indexes + // I don't think anything but 8 or 16bit indexes makes sense + stbi__rewind(s); + return 0; + } + tga_comp = stbi__tga_get_comp(tga_colormap_bpp, 0, NULL); + } else { + tga_comp = stbi__tga_get_comp(tga_bits_per_pixel, (tga_image_type == 3) || (tga_image_type == 11), NULL); + } + if(!tga_comp) { + stbi__rewind(s); + return 0; + } + if (x) *x = tga_w; + if (y) *y = tga_h; + if (comp) *comp = tga_comp; + return 1; // seems to have passed everything +} + +static int stbi__tga_test(stbi__context *s) +{ + int res = 0; + int sz, tga_color_type; + stbi__get8(s); // discard Offset + tga_color_type = stbi__get8(s); // color type + if ( tga_color_type > 1 ) goto errorEnd; // only RGB or indexed allowed + sz = stbi__get8(s); // image type + if ( tga_color_type == 1 ) { // colormapped (paletted) image + if (sz != 1 && sz != 9) goto errorEnd; // colortype 1 demands image type 1 or 9 + stbi__skip(s,4); // skip index of first colormap entry and number of entries + sz = stbi__get8(s); // check bits per palette color entry + if ( (sz != 8) && (sz != 15) && (sz != 16) && (sz != 24) && (sz != 32) ) goto errorEnd; + stbi__skip(s,4); // skip image x and y origin + } else { // "normal" image w/o colormap + if ( (sz != 2) && (sz != 3) && (sz != 10) && (sz != 11) ) goto errorEnd; // only RGB or grey allowed, +/- RLE + stbi__skip(s,9); // skip colormap specification and image x/y origin + } + if ( stbi__get16le(s) < 1 ) goto errorEnd; // test width + if ( stbi__get16le(s) < 1 ) goto errorEnd; // test height + sz = stbi__get8(s); // bits per pixel + if ( (tga_color_type == 1) && (sz != 8) && (sz != 16) ) goto errorEnd; // for colormapped images, bpp is size of an index + if ( (sz != 8) && (sz != 15) && (sz != 16) && (sz != 24) && (sz != 32) ) goto errorEnd; + + res = 1; // if we got this far, everything's good and we can return 1 instead of 0 + +errorEnd: + stbi__rewind(s); + return res; +} + +// read 16bit value and convert to 24bit RGB +static void stbi__tga_read_rgb16(stbi__context *s, stbi_uc* out) +{ + stbi__uint16 px = (stbi__uint16)stbi__get16le(s); + stbi__uint16 fiveBitMask = 31; + // we have 3 channels with 5bits each + int r = (px >> 10) & fiveBitMask; + int g = (px >> 5) & fiveBitMask; + int b = px & fiveBitMask; + // Note that this saves the data in RGB(A) order, so it doesn't need to be swapped later + out[0] = (stbi_uc)((r * 255)/31); + out[1] = (stbi_uc)((g * 255)/31); + out[2] = (stbi_uc)((b * 255)/31); + + // some people claim that the most significant bit might be used for alpha + // (possibly if an alpha-bit is set in the "image descriptor byte") + // but that only made 16bit test images completely translucent.. + // so let's treat all 15 and 16bit TGAs as RGB with no alpha. +} + +static void *stbi__tga_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri) +{ + // read in the TGA header stuff + int tga_offset = stbi__get8(s); + int tga_indexed = stbi__get8(s); + int tga_image_type = stbi__get8(s); + int tga_is_RLE = 0; + int tga_palette_start = stbi__get16le(s); + int tga_palette_len = stbi__get16le(s); + int tga_palette_bits = stbi__get8(s); + int tga_x_origin = stbi__get16le(s); + int tga_y_origin = stbi__get16le(s); + int tga_width = stbi__get16le(s); + int tga_height = stbi__get16le(s); + int tga_bits_per_pixel = stbi__get8(s); + int tga_comp, tga_rgb16=0; + int tga_inverted = stbi__get8(s); + // int tga_alpha_bits = tga_inverted & 15; // the 4 lowest bits - unused (useless?) + // image data + unsigned char *tga_data; + unsigned char *tga_palette = NULL; + int i, j; + unsigned char raw_data[4] = {0}; + int RLE_count = 0; + int RLE_repeating = 0; + int read_next_pixel = 1; + STBI_NOTUSED(ri); + STBI_NOTUSED(tga_x_origin); // @TODO + STBI_NOTUSED(tga_y_origin); // @TODO + + if (tga_height > STBI_MAX_DIMENSIONS) return stbi__errpuc("too large","Very large image (corrupt?)"); + if (tga_width > STBI_MAX_DIMENSIONS) return stbi__errpuc("too large","Very large image (corrupt?)"); + + // do a tiny bit of precessing + if ( tga_image_type >= 8 ) + { + tga_image_type -= 8; + tga_is_RLE = 1; + } + tga_inverted = 1 - ((tga_inverted >> 5) & 1); + + // If I'm paletted, then I'll use the number of bits from the palette + if ( tga_indexed ) tga_comp = stbi__tga_get_comp(tga_palette_bits, 0, &tga_rgb16); + else tga_comp = stbi__tga_get_comp(tga_bits_per_pixel, (tga_image_type == 3), &tga_rgb16); + + if(!tga_comp) // shouldn't really happen, stbi__tga_test() should have ensured basic consistency + return stbi__errpuc("bad format", "Can't find out TGA pixelformat"); + + // tga info + *x = tga_width; + *y = tga_height; + if (comp) *comp = tga_comp; + + if (!stbi__mad3sizes_valid(tga_width, tga_height, tga_comp, 0)) + return stbi__errpuc("too large", "Corrupt TGA"); + + tga_data = (unsigned char*)stbi__malloc_mad3(tga_width, tga_height, tga_comp, 0); + if (!tga_data) return stbi__errpuc("outofmem", "Out of memory"); + + // skip to the data's starting position (offset usually = 0) + stbi__skip(s, tga_offset ); + + if ( !tga_indexed && !tga_is_RLE && !tga_rgb16 ) { + for (i=0; i < tga_height; ++i) { + int row = tga_inverted ? tga_height -i - 1 : i; + stbi_uc *tga_row = tga_data + row*tga_width*tga_comp; + stbi__getn(s, tga_row, tga_width * tga_comp); + } + } else { + // do I need to load a palette? + if ( tga_indexed) + { + if (tga_palette_len == 0) { /* you have to have at least one entry! */ + STBI_FREE(tga_data); + return stbi__errpuc("bad palette", "Corrupt TGA"); + } + + // any data to skip? (offset usually = 0) + stbi__skip(s, tga_palette_start ); + // load the palette + tga_palette = (unsigned char*)stbi__malloc_mad2(tga_palette_len, tga_comp, 0); + if (!tga_palette) { + STBI_FREE(tga_data); + return stbi__errpuc("outofmem", "Out of memory"); + } + if (tga_rgb16) { + stbi_uc *pal_entry = tga_palette; + STBI_ASSERT(tga_comp == STBI_rgb); + for (i=0; i < tga_palette_len; ++i) { + stbi__tga_read_rgb16(s, pal_entry); + pal_entry += tga_comp; + } + } else if (!stbi__getn(s, tga_palette, tga_palette_len * tga_comp)) { + STBI_FREE(tga_data); + STBI_FREE(tga_palette); + return stbi__errpuc("bad palette", "Corrupt TGA"); + } + } + // load the data + for (i=0; i < tga_width * tga_height; ++i) + { + // if I'm in RLE mode, do I need to get a RLE stbi__pngchunk? + if ( tga_is_RLE ) + { + if ( RLE_count == 0 ) + { + // yep, get the next byte as a RLE command + int RLE_cmd = stbi__get8(s); + RLE_count = 1 + (RLE_cmd & 127); + RLE_repeating = RLE_cmd >> 7; + read_next_pixel = 1; + } else if ( !RLE_repeating ) + { + read_next_pixel = 1; + } + } else + { + read_next_pixel = 1; + } + // OK, if I need to read a pixel, do it now + if ( read_next_pixel ) + { + // load however much data we did have + if ( tga_indexed ) + { + // read in index, then perform the lookup + int pal_idx = (tga_bits_per_pixel == 8) ? stbi__get8(s) : stbi__get16le(s); + if ( pal_idx >= tga_palette_len ) { + // invalid index + pal_idx = 0; + } + pal_idx *= tga_comp; + for (j = 0; j < tga_comp; ++j) { + raw_data[j] = tga_palette[pal_idx+j]; + } + } else if(tga_rgb16) { + STBI_ASSERT(tga_comp == STBI_rgb); + stbi__tga_read_rgb16(s, raw_data); + } else { + // read in the data raw + for (j = 0; j < tga_comp; ++j) { + raw_data[j] = stbi__get8(s); + } + } + // clear the reading flag for the next pixel + read_next_pixel = 0; + } // end of reading a pixel + + // copy data + for (j = 0; j < tga_comp; ++j) + tga_data[i*tga_comp+j] = raw_data[j]; + + // in case we're in RLE mode, keep counting down + --RLE_count; + } + // do I need to invert the image? + if ( tga_inverted ) + { + for (j = 0; j*2 < tga_height; ++j) + { + int index1 = j * tga_width * tga_comp; + int index2 = (tga_height - 1 - j) * tga_width * tga_comp; + for (i = tga_width * tga_comp; i > 0; --i) + { + unsigned char temp = tga_data[index1]; + tga_data[index1] = tga_data[index2]; + tga_data[index2] = temp; + ++index1; + ++index2; + } + } + } + // clear my palette, if I had one + if ( tga_palette != NULL ) + { + STBI_FREE( tga_palette ); + } + } + + // swap RGB - if the source data was RGB16, it already is in the right order + if (tga_comp >= 3 && !tga_rgb16) + { + unsigned char* tga_pixel = tga_data; + for (i=0; i < tga_width * tga_height; ++i) + { + unsigned char temp = tga_pixel[0]; + tga_pixel[0] = tga_pixel[2]; + tga_pixel[2] = temp; + tga_pixel += tga_comp; + } + } + + // convert to target component count + if (req_comp && req_comp != tga_comp) + tga_data = stbi__convert_format(tga_data, tga_comp, req_comp, tga_width, tga_height); + + // the things I do to get rid of an error message, and yet keep + // Microsoft's C compilers happy... [8^( + tga_palette_start = tga_palette_len = tga_palette_bits = + tga_x_origin = tga_y_origin = 0; + STBI_NOTUSED(tga_palette_start); + // OK, done + return tga_data; +} +#endif + +// ************************************************************************************************* +// Photoshop PSD loader -- PD by Thatcher Ulrich, integration by Nicolas Schulz, tweaked by STB + +#ifndef STBI_NO_PSD +static int stbi__psd_test(stbi__context *s) +{ + int r = (stbi__get32be(s) == 0x38425053); + stbi__rewind(s); + return r; +} + +static int stbi__psd_decode_rle(stbi__context *s, stbi_uc *p, int pixelCount) +{ + int count, nleft, len; + + count = 0; + while ((nleft = pixelCount - count) > 0) { + len = stbi__get8(s); + if (len == 128) { + // No-op. + } else if (len < 128) { + // Copy next len+1 bytes literally. + len++; + if (len > nleft) return 0; // corrupt data + count += len; + while (len) { + *p = stbi__get8(s); + p += 4; + len--; + } + } else if (len > 128) { + stbi_uc val; + // Next -len+1 bytes in the dest are replicated from next source byte. + // (Interpret len as a negative 8-bit int.) + len = 257 - len; + if (len > nleft) return 0; // corrupt data + val = stbi__get8(s); + count += len; + while (len) { + *p = val; + p += 4; + len--; + } + } + } + + return 1; +} + +static void *stbi__psd_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri, int bpc) +{ + int pixelCount; + int channelCount, compression; + int channel, i; + int bitdepth; + int w,h; + stbi_uc *out; + STBI_NOTUSED(ri); + + // Check identifier + if (stbi__get32be(s) != 0x38425053) // "8BPS" + return stbi__errpuc("not PSD", "Corrupt PSD image"); + + // Check file type version. + if (stbi__get16be(s) != 1) + return stbi__errpuc("wrong version", "Unsupported version of PSD image"); + + // Skip 6 reserved bytes. + stbi__skip(s, 6 ); + + // Read the number of channels (R, G, B, A, etc). + channelCount = stbi__get16be(s); + if (channelCount < 0 || channelCount > 16) + return stbi__errpuc("wrong channel count", "Unsupported number of channels in PSD image"); + + // Read the rows and columns of the image. + h = stbi__get32be(s); + w = stbi__get32be(s); + + if (h > STBI_MAX_DIMENSIONS) return stbi__errpuc("too large","Very large image (corrupt?)"); + if (w > STBI_MAX_DIMENSIONS) return stbi__errpuc("too large","Very large image (corrupt?)"); + + // Make sure the depth is 8 bits. + bitdepth = stbi__get16be(s); + if (bitdepth != 8 && bitdepth != 16) + return stbi__errpuc("unsupported bit depth", "PSD bit depth is not 8 or 16 bit"); + + // Make sure the color mode is RGB. + // Valid options are: + // 0: Bitmap + // 1: Grayscale + // 2: Indexed color + // 3: RGB color + // 4: CMYK color + // 7: Multichannel + // 8: Duotone + // 9: Lab color + if (stbi__get16be(s) != 3) + return stbi__errpuc("wrong color format", "PSD is not in RGB color format"); + + // Skip the Mode Data. (It's the palette for indexed color; other info for other modes.) + stbi__skip(s,stbi__get32be(s) ); + + // Skip the image resources. (resolution, pen tool paths, etc) + stbi__skip(s, stbi__get32be(s) ); + + // Skip the reserved data. + stbi__skip(s, stbi__get32be(s) ); + + // Find out if the data is compressed. + // Known values: + // 0: no compression + // 1: RLE compressed + compression = stbi__get16be(s); + if (compression > 1) + return stbi__errpuc("bad compression", "PSD has an unknown compression format"); + + // Check size + if (!stbi__mad3sizes_valid(4, w, h, 0)) + return stbi__errpuc("too large", "Corrupt PSD"); + + // Create the destination image. + + if (!compression && bitdepth == 16 && bpc == 16) { + out = (stbi_uc *) stbi__malloc_mad3(8, w, h, 0); + ri->bits_per_channel = 16; + } else + out = (stbi_uc *) stbi__malloc(4 * w*h); + + if (!out) return stbi__errpuc("outofmem", "Out of memory"); + pixelCount = w*h; + + // Initialize the data to zero. + //memset( out, 0, pixelCount * 4 ); + + // Finally, the image data. + if (compression) { + // RLE as used by .PSD and .TIFF + // Loop until you get the number of unpacked bytes you are expecting: + // Read the next source byte into n. + // If n is between 0 and 127 inclusive, copy the next n+1 bytes literally. + // Else if n is between -127 and -1 inclusive, copy the next byte -n+1 times. + // Else if n is 128, noop. + // Endloop + + // The RLE-compressed data is preceded by a 2-byte data count for each row in the data, + // which we're going to just skip. + stbi__skip(s, h * channelCount * 2 ); + + // Read the RLE data by channel. + for (channel = 0; channel < 4; channel++) { + stbi_uc *p; + + p = out+channel; + if (channel >= channelCount) { + // Fill this channel with default data. + for (i = 0; i < pixelCount; i++, p += 4) + *p = (channel == 3 ? 255 : 0); + } else { + // Read the RLE data. + if (!stbi__psd_decode_rle(s, p, pixelCount)) { + STBI_FREE(out); + return stbi__errpuc("corrupt", "bad RLE data"); + } + } + } + + } else { + // We're at the raw image data. It's each channel in order (Red, Green, Blue, Alpha, ...) + // where each channel consists of an 8-bit (or 16-bit) value for each pixel in the image. + + // Read the data by channel. + for (channel = 0; channel < 4; channel++) { + if (channel >= channelCount) { + // Fill this channel with default data. + if (bitdepth == 16 && bpc == 16) { + stbi__uint16 *q = ((stbi__uint16 *) out) + channel; + stbi__uint16 val = channel == 3 ? 65535 : 0; + for (i = 0; i < pixelCount; i++, q += 4) + *q = val; + } else { + stbi_uc *p = out+channel; + stbi_uc val = channel == 3 ? 255 : 0; + for (i = 0; i < pixelCount; i++, p += 4) + *p = val; + } + } else { + if (ri->bits_per_channel == 16) { // output bpc + stbi__uint16 *q = ((stbi__uint16 *) out) + channel; + for (i = 0; i < pixelCount; i++, q += 4) + *q = (stbi__uint16) stbi__get16be(s); + } else { + stbi_uc *p = out+channel; + if (bitdepth == 16) { // input bpc + for (i = 0; i < pixelCount; i++, p += 4) + *p = (stbi_uc) (stbi__get16be(s) >> 8); + } else { + for (i = 0; i < pixelCount; i++, p += 4) + *p = stbi__get8(s); + } + } + } + } + } + + // remove weird white matte from PSD + if (channelCount >= 4) { + if (ri->bits_per_channel == 16) { + for (i=0; i < w*h; ++i) { + stbi__uint16 *pixel = (stbi__uint16 *) out + 4*i; + if (pixel[3] != 0 && pixel[3] != 65535) { + float a = pixel[3] / 65535.0f; + float ra = 1.0f / a; + float inv_a = 65535.0f * (1 - ra); + pixel[0] = (stbi__uint16) (pixel[0]*ra + inv_a); + pixel[1] = (stbi__uint16) (pixel[1]*ra + inv_a); + pixel[2] = (stbi__uint16) (pixel[2]*ra + inv_a); + } + } + } else { + for (i=0; i < w*h; ++i) { + unsigned char *pixel = out + 4*i; + if (pixel[3] != 0 && pixel[3] != 255) { + float a = pixel[3] / 255.0f; + float ra = 1.0f / a; + float inv_a = 255.0f * (1 - ra); + pixel[0] = (unsigned char) (pixel[0]*ra + inv_a); + pixel[1] = (unsigned char) (pixel[1]*ra + inv_a); + pixel[2] = (unsigned char) (pixel[2]*ra + inv_a); + } + } + } + } + + // convert to desired output format + if (req_comp && req_comp != 4) { + if (ri->bits_per_channel == 16) + out = (stbi_uc *) stbi__convert_format16((stbi__uint16 *) out, 4, req_comp, w, h); + else + out = stbi__convert_format(out, 4, req_comp, w, h); + if (out == NULL) return out; // stbi__convert_format frees input on failure + } + + if (comp) *comp = 4; + *y = h; + *x = w; + + return out; +} +#endif + +// ************************************************************************************************* +// Softimage PIC loader +// by Tom Seddon +// +// See http://softimage.wiki.softimage.com/index.php/INFO:_PIC_file_format +// See http://ozviz.wasp.uwa.edu.au/~pbourke/dataformats/softimagepic/ + +#ifndef STBI_NO_PIC +static int stbi__pic_is4(stbi__context *s,const char *str) +{ + int i; + for (i=0; i<4; ++i) + if (stbi__get8(s) != (stbi_uc)str[i]) + return 0; + + return 1; +} + +static int stbi__pic_test_core(stbi__context *s) +{ + int i; + + if (!stbi__pic_is4(s,"\x53\x80\xF6\x34")) + return 0; + + for(i=0;i<84;++i) + stbi__get8(s); + + if (!stbi__pic_is4(s,"PICT")) + return 0; + + return 1; +} + +typedef struct +{ + stbi_uc size,type,channel; +} stbi__pic_packet; + +static stbi_uc *stbi__readval(stbi__context *s, int channel, stbi_uc *dest) +{ + int mask=0x80, i; + + for (i=0; i<4; ++i, mask>>=1) { + if (channel & mask) { + if (stbi__at_eof(s)) return stbi__errpuc("bad file","PIC file too short"); + dest[i]=stbi__get8(s); + } + } + + return dest; +} + +static void stbi__copyval(int channel,stbi_uc *dest,const stbi_uc *src) +{ + int mask=0x80,i; + + for (i=0;i<4; ++i, mask>>=1) + if (channel&mask) + dest[i]=src[i]; +} + +static stbi_uc *stbi__pic_load_core(stbi__context *s,int width,int height,int *comp, stbi_uc *result) +{ + int act_comp=0,num_packets=0,y,chained; + stbi__pic_packet packets[10]; + + // this will (should...) cater for even some bizarre stuff like having data + // for the same channel in multiple packets. + do { + stbi__pic_packet *packet; + + if (num_packets==sizeof(packets)/sizeof(packets[0])) + return stbi__errpuc("bad format","too many packets"); + + packet = &packets[num_packets++]; + + chained = stbi__get8(s); + packet->size = stbi__get8(s); + packet->type = stbi__get8(s); + packet->channel = stbi__get8(s); + + act_comp |= packet->channel; + + if (stbi__at_eof(s)) return stbi__errpuc("bad file","file too short (reading packets)"); + if (packet->size != 8) return stbi__errpuc("bad format","packet isn't 8bpp"); + } while (chained); + + *comp = (act_comp & 0x10 ? 4 : 3); // has alpha channel? + + for(y=0; ytype) { + default: + return stbi__errpuc("bad format","packet has bad compression type"); + + case 0: {//uncompressed + int x; + + for(x=0;xchannel,dest)) + return 0; + break; + } + + case 1://Pure RLE + { + int left=width, i; + + while (left>0) { + stbi_uc count,value[4]; + + count=stbi__get8(s); + if (stbi__at_eof(s)) return stbi__errpuc("bad file","file too short (pure read count)"); + + if (count > left) + count = (stbi_uc) left; + + if (!stbi__readval(s,packet->channel,value)) return 0; + + for(i=0; ichannel,dest,value); + left -= count; + } + } + break; + + case 2: {//Mixed RLE + int left=width; + while (left>0) { + int count = stbi__get8(s), i; + if (stbi__at_eof(s)) return stbi__errpuc("bad file","file too short (mixed read count)"); + + if (count >= 128) { // Repeated + stbi_uc value[4]; + + if (count==128) + count = stbi__get16be(s); + else + count -= 127; + if (count > left) + return stbi__errpuc("bad file","scanline overrun"); + + if (!stbi__readval(s,packet->channel,value)) + return 0; + + for(i=0;ichannel,dest,value); + } else { // Raw + ++count; + if (count>left) return stbi__errpuc("bad file","scanline overrun"); + + for(i=0;ichannel,dest)) + return 0; + } + left-=count; + } + break; + } + } + } + } + + return result; +} + +static void *stbi__pic_load(stbi__context *s,int *px,int *py,int *comp,int req_comp, stbi__result_info *ri) +{ + stbi_uc *result; + int i, x,y, internal_comp; + STBI_NOTUSED(ri); + + if (!comp) comp = &internal_comp; + + for (i=0; i<92; ++i) + stbi__get8(s); + + x = stbi__get16be(s); + y = stbi__get16be(s); + + if (y > STBI_MAX_DIMENSIONS) return stbi__errpuc("too large","Very large image (corrupt?)"); + if (x > STBI_MAX_DIMENSIONS) return stbi__errpuc("too large","Very large image (corrupt?)"); + + if (stbi__at_eof(s)) return stbi__errpuc("bad file","file too short (pic header)"); + if (!stbi__mad3sizes_valid(x, y, 4, 0)) return stbi__errpuc("too large", "PIC image too large to decode"); + + stbi__get32be(s); //skip `ratio' + stbi__get16be(s); //skip `fields' + stbi__get16be(s); //skip `pad' + + // intermediate buffer is RGBA + result = (stbi_uc *) stbi__malloc_mad3(x, y, 4, 0); + if (!result) return stbi__errpuc("outofmem", "Out of memory"); + memset(result, 0xff, x*y*4); + + if (!stbi__pic_load_core(s,x,y,comp, result)) { + STBI_FREE(result); + result=0; + } + *px = x; + *py = y; + if (req_comp == 0) req_comp = *comp; + result=stbi__convert_format(result,4,req_comp,x,y); + + return result; +} + +static int stbi__pic_test(stbi__context *s) +{ + int r = stbi__pic_test_core(s); + stbi__rewind(s); + return r; +} +#endif + +// ************************************************************************************************* +// GIF loader -- public domain by Jean-Marc Lienher -- simplified/shrunk by stb + +#ifndef STBI_NO_GIF +typedef struct +{ + stbi__int16 prefix; + stbi_uc first; + stbi_uc suffix; +} stbi__gif_lzw; + +typedef struct +{ + int w,h; + stbi_uc *out; // output buffer (always 4 components) + stbi_uc *background; // The current "background" as far as a gif is concerned + stbi_uc *history; + int flags, bgindex, ratio, transparent, eflags; + stbi_uc pal[256][4]; + stbi_uc lpal[256][4]; + stbi__gif_lzw codes[8192]; + stbi_uc *color_table; + int parse, step; + int lflags; + int start_x, start_y; + int max_x, max_y; + int cur_x, cur_y; + int line_size; + int delay; +} stbi__gif; + +static int stbi__gif_test_raw(stbi__context *s) +{ + int sz; + if (stbi__get8(s) != 'G' || stbi__get8(s) != 'I' || stbi__get8(s) != 'F' || stbi__get8(s) != '8') return 0; + sz = stbi__get8(s); + if (sz != '9' && sz != '7') return 0; + if (stbi__get8(s) != 'a') return 0; + return 1; +} + +static int stbi__gif_test(stbi__context *s) +{ + int r = stbi__gif_test_raw(s); + stbi__rewind(s); + return r; +} + +static void stbi__gif_parse_colortable(stbi__context *s, stbi_uc pal[256][4], int num_entries, int transp) +{ + int i; + for (i=0; i < num_entries; ++i) { + pal[i][2] = stbi__get8(s); + pal[i][1] = stbi__get8(s); + pal[i][0] = stbi__get8(s); + pal[i][3] = transp == i ? 0 : 255; + } +} + +static int stbi__gif_header(stbi__context *s, stbi__gif *g, int *comp, int is_info) +{ + stbi_uc version; + if (stbi__get8(s) != 'G' || stbi__get8(s) != 'I' || stbi__get8(s) != 'F' || stbi__get8(s) != '8') + return stbi__err("not GIF", "Corrupt GIF"); + + version = stbi__get8(s); + if (version != '7' && version != '9') return stbi__err("not GIF", "Corrupt GIF"); + if (stbi__get8(s) != 'a') return stbi__err("not GIF", "Corrupt GIF"); + + stbi__g_failure_reason = ""; + g->w = stbi__get16le(s); + g->h = stbi__get16le(s); + g->flags = stbi__get8(s); + g->bgindex = stbi__get8(s); + g->ratio = stbi__get8(s); + g->transparent = -1; + + if (g->w > STBI_MAX_DIMENSIONS) return stbi__err("too large","Very large image (corrupt?)"); + if (g->h > STBI_MAX_DIMENSIONS) return stbi__err("too large","Very large image (corrupt?)"); + + if (comp != 0) *comp = 4; // can't actually tell whether it's 3 or 4 until we parse the comments + + if (is_info) return 1; + + if (g->flags & 0x80) + stbi__gif_parse_colortable(s,g->pal, 2 << (g->flags & 7), -1); + + return 1; +} + +static int stbi__gif_info_raw(stbi__context *s, int *x, int *y, int *comp) +{ + stbi__gif* g = (stbi__gif*) stbi__malloc(sizeof(stbi__gif)); + if (!g) return stbi__err("outofmem", "Out of memory"); + if (!stbi__gif_header(s, g, comp, 1)) { + STBI_FREE(g); + stbi__rewind( s ); + return 0; + } + if (x) *x = g->w; + if (y) *y = g->h; + STBI_FREE(g); + return 1; +} + +static void stbi__out_gif_code(stbi__gif *g, stbi__uint16 code) +{ + stbi_uc *p, *c; + int idx; + + // recurse to decode the prefixes, since the linked-list is backwards, + // and working backwards through an interleaved image would be nasty + if (g->codes[code].prefix >= 0) + stbi__out_gif_code(g, g->codes[code].prefix); + + if (g->cur_y >= g->max_y) return; + + idx = g->cur_x + g->cur_y; + p = &g->out[idx]; + g->history[idx / 4] = 1; + + c = &g->color_table[g->codes[code].suffix * 4]; + if (c[3] > 128) { // don't render transparent pixels; + p[0] = c[2]; + p[1] = c[1]; + p[2] = c[0]; + p[3] = c[3]; + } + g->cur_x += 4; + + if (g->cur_x >= g->max_x) { + g->cur_x = g->start_x; + g->cur_y += g->step; + + while (g->cur_y >= g->max_y && g->parse > 0) { + g->step = (1 << g->parse) * g->line_size; + g->cur_y = g->start_y + (g->step >> 1); + --g->parse; + } + } +} + +static stbi_uc *stbi__process_gif_raster(stbi__context *s, stbi__gif *g) +{ + stbi_uc lzw_cs; + stbi__int32 len, init_code; + stbi__uint32 first; + stbi__int32 codesize, codemask, avail, oldcode, bits, valid_bits, clear; + stbi__gif_lzw *p; + + lzw_cs = stbi__get8(s); + if (lzw_cs > 12) return NULL; + clear = 1 << lzw_cs; + first = 1; + codesize = lzw_cs + 1; + codemask = (1 << codesize) - 1; + bits = 0; + valid_bits = 0; + for (init_code = 0; init_code < clear; init_code++) { + g->codes[init_code].prefix = -1; + g->codes[init_code].first = (stbi_uc) init_code; + g->codes[init_code].suffix = (stbi_uc) init_code; + } + + // support no starting clear code + avail = clear+2; + oldcode = -1; + + len = 0; + for(;;) { + if (valid_bits < codesize) { + if (len == 0) { + len = stbi__get8(s); // start new block + if (len == 0) + return g->out; + } + --len; + bits |= (stbi__int32) stbi__get8(s) << valid_bits; + valid_bits += 8; + } else { + stbi__int32 code = bits & codemask; + bits >>= codesize; + valid_bits -= codesize; + // @OPTIMIZE: is there some way we can accelerate the non-clear path? + if (code == clear) { // clear code + codesize = lzw_cs + 1; + codemask = (1 << codesize) - 1; + avail = clear + 2; + oldcode = -1; + first = 0; + } else if (code == clear + 1) { // end of stream code + stbi__skip(s, len); + while ((len = stbi__get8(s)) > 0) + stbi__skip(s,len); + return g->out; + } else if (code <= avail) { + if (first) { + return stbi__errpuc("no clear code", "Corrupt GIF"); + } + + if (oldcode >= 0) { + p = &g->codes[avail++]; + if (avail > 8192) { + return stbi__errpuc("too many codes", "Corrupt GIF"); + } + + p->prefix = (stbi__int16) oldcode; + p->first = g->codes[oldcode].first; + p->suffix = (code == avail) ? p->first : g->codes[code].first; + } else if (code == avail) + return stbi__errpuc("illegal code in raster", "Corrupt GIF"); + + stbi__out_gif_code(g, (stbi__uint16) code); + + if ((avail & codemask) == 0 && avail <= 0x0FFF) { + codesize++; + codemask = (1 << codesize) - 1; + } + + oldcode = code; + } else { + return stbi__errpuc("illegal code in raster", "Corrupt GIF"); + } + } + } +} + +// this function is designed to support animated gifs, although stb_image doesn't support it +// two back is the image from two frames ago, used for a very specific disposal format +static stbi_uc *stbi__gif_load_next(stbi__context *s, stbi__gif *g, int *comp, int req_comp, stbi_uc *two_back) +{ + int dispose; + int first_frame; + int pi; + int pcount; + STBI_NOTUSED(req_comp); + + // on first frame, any non-written pixels get the background colour (non-transparent) + first_frame = 0; + if (g->out == 0) { + if (!stbi__gif_header(s, g, comp,0)) return 0; // stbi__g_failure_reason set by stbi__gif_header + if (!stbi__mad3sizes_valid(4, g->w, g->h, 0)) + return stbi__errpuc("too large", "GIF image is too large"); + pcount = g->w * g->h; + g->out = (stbi_uc *) stbi__malloc(4 * pcount); + g->background = (stbi_uc *) stbi__malloc(4 * pcount); + g->history = (stbi_uc *) stbi__malloc(pcount); + if (!g->out || !g->background || !g->history) + return stbi__errpuc("outofmem", "Out of memory"); + + // image is treated as "transparent" at the start - ie, nothing overwrites the current background; + // background colour is only used for pixels that are not rendered first frame, after that "background" + // color refers to the color that was there the previous frame. + memset(g->out, 0x00, 4 * pcount); + memset(g->background, 0x00, 4 * pcount); // state of the background (starts transparent) + memset(g->history, 0x00, pcount); // pixels that were affected previous frame + first_frame = 1; + } else { + // second frame - how do we dispose of the previous one? + dispose = (g->eflags & 0x1C) >> 2; + pcount = g->w * g->h; + + if ((dispose == 3) && (two_back == 0)) { + dispose = 2; // if I don't have an image to revert back to, default to the old background + } + + if (dispose == 3) { // use previous graphic + for (pi = 0; pi < pcount; ++pi) { + if (g->history[pi]) { + memcpy( &g->out[pi * 4], &two_back[pi * 4], 4 ); + } + } + } else if (dispose == 2) { + // restore what was changed last frame to background before that frame; + for (pi = 0; pi < pcount; ++pi) { + if (g->history[pi]) { + memcpy( &g->out[pi * 4], &g->background[pi * 4], 4 ); + } + } + } else { + // This is a non-disposal case eithe way, so just + // leave the pixels as is, and they will become the new background + // 1: do not dispose + // 0: not specified. + } + + // background is what out is after the undoing of the previou frame; + memcpy( g->background, g->out, 4 * g->w * g->h ); + } + + // clear my history; + memset( g->history, 0x00, g->w * g->h ); // pixels that were affected previous frame + + for (;;) { + int tag = stbi__get8(s); + switch (tag) { + case 0x2C: /* Image Descriptor */ + { + stbi__int32 x, y, w, h; + stbi_uc *o; + + x = stbi__get16le(s); + y = stbi__get16le(s); + w = stbi__get16le(s); + h = stbi__get16le(s); + if (((x + w) > (g->w)) || ((y + h) > (g->h))) + return stbi__errpuc("bad Image Descriptor", "Corrupt GIF"); + + g->line_size = g->w * 4; + g->start_x = x * 4; + g->start_y = y * g->line_size; + g->max_x = g->start_x + w * 4; + g->max_y = g->start_y + h * g->line_size; + g->cur_x = g->start_x; + g->cur_y = g->start_y; + + // if the width of the specified rectangle is 0, that means + // we may not see *any* pixels or the image is malformed; + // to make sure this is caught, move the current y down to + // max_y (which is what out_gif_code checks). + if (w == 0) + g->cur_y = g->max_y; + + g->lflags = stbi__get8(s); + + if (g->lflags & 0x40) { + g->step = 8 * g->line_size; // first interlaced spacing + g->parse = 3; + } else { + g->step = g->line_size; + g->parse = 0; + } + + if (g->lflags & 0x80) { + stbi__gif_parse_colortable(s,g->lpal, 2 << (g->lflags & 7), g->eflags & 0x01 ? g->transparent : -1); + g->color_table = (stbi_uc *) g->lpal; + } else if (g->flags & 0x80) { + g->color_table = (stbi_uc *) g->pal; + } else + return stbi__errpuc("missing color table", "Corrupt GIF"); + + o = stbi__process_gif_raster(s, g); + if (!o) return NULL; + + // if this was the first frame, + pcount = g->w * g->h; + if (first_frame && (g->bgindex > 0)) { + // if first frame, any pixel not drawn to gets the background color + for (pi = 0; pi < pcount; ++pi) { + if (g->history[pi] == 0) { + g->pal[g->bgindex][3] = 255; // just in case it was made transparent, undo that; It will be reset next frame if need be; + memcpy( &g->out[pi * 4], &g->pal[g->bgindex], 4 ); + } + } + } + + return o; + } + + case 0x21: // Comment Extension. + { + int len; + int ext = stbi__get8(s); + if (ext == 0xF9) { // Graphic Control Extension. + len = stbi__get8(s); + if (len == 4) { + g->eflags = stbi__get8(s); + g->delay = 10 * stbi__get16le(s); // delay - 1/100th of a second, saving as 1/1000ths. + + // unset old transparent + if (g->transparent >= 0) { + g->pal[g->transparent][3] = 255; + } + if (g->eflags & 0x01) { + g->transparent = stbi__get8(s); + if (g->transparent >= 0) { + g->pal[g->transparent][3] = 0; + } + } else { + // don't need transparent + stbi__skip(s, 1); + g->transparent = -1; + } + } else { + stbi__skip(s, len); + break; + } + } + while ((len = stbi__get8(s)) != 0) { + stbi__skip(s, len); + } + break; + } + + case 0x3B: // gif stream termination code + return (stbi_uc *) s; // using '1' causes warning on some compilers + + default: + return stbi__errpuc("unknown code", "Corrupt GIF"); + } + } +} + +static void *stbi__load_gif_main_outofmem(stbi__gif *g, stbi_uc *out, int **delays) +{ + STBI_FREE(g->out); + STBI_FREE(g->history); + STBI_FREE(g->background); + + if (out) STBI_FREE(out); + if (delays && *delays) STBI_FREE(*delays); + return stbi__errpuc("outofmem", "Out of memory"); +} + +static void *stbi__load_gif_main(stbi__context *s, int **delays, int *x, int *y, int *z, int *comp, int req_comp) +{ + if (stbi__gif_test(s)) { + int layers = 0; + stbi_uc *u = 0; + stbi_uc *out = 0; + stbi_uc *two_back = 0; + stbi__gif g; + int stride; + int out_size = 0; + int delays_size = 0; + + STBI_NOTUSED(out_size); + STBI_NOTUSED(delays_size); + + memset(&g, 0, sizeof(g)); + if (delays) { + *delays = 0; + } + + do { + u = stbi__gif_load_next(s, &g, comp, req_comp, two_back); + if (u == (stbi_uc *) s) u = 0; // end of animated gif marker + + if (u) { + *x = g.w; + *y = g.h; + ++layers; + stride = g.w * g.h * 4; + + if (out) { + void *tmp = (stbi_uc*) STBI_REALLOC_SIZED( out, out_size, layers * stride ); + if (!tmp) + return stbi__load_gif_main_outofmem(&g, out, delays); + else { + out = (stbi_uc*) tmp; + out_size = layers * stride; + } + + if (delays) { + int *new_delays = (int*) STBI_REALLOC_SIZED( *delays, delays_size, sizeof(int) * layers ); + if (!new_delays) + return stbi__load_gif_main_outofmem(&g, out, delays); + *delays = new_delays; + delays_size = layers * sizeof(int); + } + } else { + out = (stbi_uc*)stbi__malloc( layers * stride ); + if (!out) + return stbi__load_gif_main_outofmem(&g, out, delays); + out_size = layers * stride; + if (delays) { + *delays = (int*) stbi__malloc( layers * sizeof(int) ); + if (!*delays) + return stbi__load_gif_main_outofmem(&g, out, delays); + delays_size = layers * sizeof(int); + } + } + memcpy( out + ((layers - 1) * stride), u, stride ); + if (layers >= 2) { + two_back = out - 2 * stride; + } + + if (delays) { + (*delays)[layers - 1U] = g.delay; + } + } + } while (u != 0); + + // free temp buffer; + STBI_FREE(g.out); + STBI_FREE(g.history); + STBI_FREE(g.background); + + // do the final conversion after loading everything; + if (req_comp && req_comp != 4) + out = stbi__convert_format(out, 4, req_comp, layers * g.w, g.h); + + *z = layers; + return out; + } else { + return stbi__errpuc("not GIF", "Image was not as a gif type."); + } +} + +static void *stbi__gif_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri) +{ + stbi_uc *u = 0; + stbi__gif g; + memset(&g, 0, sizeof(g)); + STBI_NOTUSED(ri); + + u = stbi__gif_load_next(s, &g, comp, req_comp, 0); + if (u == (stbi_uc *) s) u = 0; // end of animated gif marker + if (u) { + *x = g.w; + *y = g.h; + + // moved conversion to after successful load so that the same + // can be done for multiple frames. + if (req_comp && req_comp != 4) + u = stbi__convert_format(u, 4, req_comp, g.w, g.h); + } else if (g.out) { + // if there was an error and we allocated an image buffer, free it! + STBI_FREE(g.out); + } + + // free buffers needed for multiple frame loading; + STBI_FREE(g.history); + STBI_FREE(g.background); + + return u; +} + +static int stbi__gif_info(stbi__context *s, int *x, int *y, int *comp) +{ + return stbi__gif_info_raw(s,x,y,comp); +} +#endif + +// ************************************************************************************************* +// Radiance RGBE HDR loader +// originally by Nicolas Schulz +#ifndef STBI_NO_HDR +static int stbi__hdr_test_core(stbi__context *s, const char *signature) +{ + int i; + for (i=0; signature[i]; ++i) + if (stbi__get8(s) != signature[i]) + return 0; + stbi__rewind(s); + return 1; +} + +static int stbi__hdr_test(stbi__context* s) +{ + int r = stbi__hdr_test_core(s, "#?RADIANCE\n"); + stbi__rewind(s); + if(!r) { + r = stbi__hdr_test_core(s, "#?RGBE\n"); + stbi__rewind(s); + } + return r; +} + +#define STBI__HDR_BUFLEN 1024 +static char *stbi__hdr_gettoken(stbi__context *z, char *buffer) +{ + int len=0; + char c = '\0'; + + c = (char) stbi__get8(z); + + while (!stbi__at_eof(z) && c != '\n') { + buffer[len++] = c; + if (len == STBI__HDR_BUFLEN-1) { + // flush to end of line + while (!stbi__at_eof(z) && stbi__get8(z) != '\n') + ; + break; + } + c = (char) stbi__get8(z); + } + + buffer[len] = 0; + return buffer; +} + +static void stbi__hdr_convert(float *output, stbi_uc *input, int req_comp) +{ + if ( input[3] != 0 ) { + float f1; + // Exponent + f1 = (float) ldexp(1.0f, input[3] - (int)(128 + 8)); + if (req_comp <= 2) + output[0] = (input[0] + input[1] + input[2]) * f1 / 3; + else { + output[0] = input[0] * f1; + output[1] = input[1] * f1; + output[2] = input[2] * f1; + } + if (req_comp == 2) output[1] = 1; + if (req_comp == 4) output[3] = 1; + } else { + switch (req_comp) { + case 4: output[3] = 1; /* fallthrough */ + case 3: output[0] = output[1] = output[2] = 0; + break; + case 2: output[1] = 1; /* fallthrough */ + case 1: output[0] = 0; + break; + } + } +} + +static float *stbi__hdr_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri) +{ + char buffer[STBI__HDR_BUFLEN]; + char *token; + int valid = 0; + int width, height; + stbi_uc *scanline; + float *hdr_data; + int len; + unsigned char count, value; + int i, j, k, c1,c2, z; + const char *headerToken; + STBI_NOTUSED(ri); + + // Check identifier + headerToken = stbi__hdr_gettoken(s,buffer); + if (strcmp(headerToken, "#?RADIANCE") != 0 && strcmp(headerToken, "#?RGBE") != 0) + return stbi__errpf("not HDR", "Corrupt HDR image"); + + // Parse header + for(;;) { + token = stbi__hdr_gettoken(s,buffer); + if (token[0] == 0) break; + if (strcmp(token, "FORMAT=32-bit_rle_rgbe") == 0) valid = 1; + } + + if (!valid) return stbi__errpf("unsupported format", "Unsupported HDR format"); + + // Parse width and height + // can't use sscanf() if we're not using stdio! + token = stbi__hdr_gettoken(s,buffer); + if (strncmp(token, "-Y ", 3)) return stbi__errpf("unsupported data layout", "Unsupported HDR format"); + token += 3; + height = (int) strtol(token, &token, 10); + while (*token == ' ') ++token; + if (strncmp(token, "+X ", 3)) return stbi__errpf("unsupported data layout", "Unsupported HDR format"); + token += 3; + width = (int) strtol(token, NULL, 10); + + if (height > STBI_MAX_DIMENSIONS) return stbi__errpf("too large","Very large image (corrupt?)"); + if (width > STBI_MAX_DIMENSIONS) return stbi__errpf("too large","Very large image (corrupt?)"); + + *x = width; + *y = height; + + if (comp) *comp = 3; + if (req_comp == 0) req_comp = 3; + + if (!stbi__mad4sizes_valid(width, height, req_comp, sizeof(float), 0)) + return stbi__errpf("too large", "HDR image is too large"); + + // Read data + hdr_data = (float *) stbi__malloc_mad4(width, height, req_comp, sizeof(float), 0); + if (!hdr_data) + return stbi__errpf("outofmem", "Out of memory"); + + // Load image data + // image data is stored as some number of sca + if ( width < 8 || width >= 32768) { + // Read flat data + for (j=0; j < height; ++j) { + for (i=0; i < width; ++i) { + stbi_uc rgbe[4]; + main_decode_loop: + stbi__getn(s, rgbe, 4); + stbi__hdr_convert(hdr_data + j * width * req_comp + i * req_comp, rgbe, req_comp); + } + } + } else { + // Read RLE-encoded data + scanline = NULL; + + for (j = 0; j < height; ++j) { + c1 = stbi__get8(s); + c2 = stbi__get8(s); + len = stbi__get8(s); + if (c1 != 2 || c2 != 2 || (len & 0x80)) { + // not run-length encoded, so we have to actually use THIS data as a decoded + // pixel (note this can't be a valid pixel--one of RGB must be >= 128) + stbi_uc rgbe[4]; + rgbe[0] = (stbi_uc) c1; + rgbe[1] = (stbi_uc) c2; + rgbe[2] = (stbi_uc) len; + rgbe[3] = (stbi_uc) stbi__get8(s); + stbi__hdr_convert(hdr_data, rgbe, req_comp); + i = 1; + j = 0; + STBI_FREE(scanline); + goto main_decode_loop; // yes, this makes no sense + } + len <<= 8; + len |= stbi__get8(s); + if (len != width) { STBI_FREE(hdr_data); STBI_FREE(scanline); return stbi__errpf("invalid decoded scanline length", "corrupt HDR"); } + if (scanline == NULL) { + scanline = (stbi_uc *) stbi__malloc_mad2(width, 4, 0); + if (!scanline) { + STBI_FREE(hdr_data); + return stbi__errpf("outofmem", "Out of memory"); + } + } + + for (k = 0; k < 4; ++k) { + int nleft; + i = 0; + while ((nleft = width - i) > 0) { + count = stbi__get8(s); + if (count > 128) { + // Run + value = stbi__get8(s); + count -= 128; + if ((count == 0) || (count > nleft)) { STBI_FREE(hdr_data); STBI_FREE(scanline); return stbi__errpf("corrupt", "bad RLE data in HDR"); } + for (z = 0; z < count; ++z) + scanline[i++ * 4 + k] = value; + } else { + // Dump + if ((count == 0) || (count > nleft)) { STBI_FREE(hdr_data); STBI_FREE(scanline); return stbi__errpf("corrupt", "bad RLE data in HDR"); } + for (z = 0; z < count; ++z) + scanline[i++ * 4 + k] = stbi__get8(s); + } + } + } + for (i=0; i < width; ++i) + stbi__hdr_convert(hdr_data+(j*width + i)*req_comp, scanline + i*4, req_comp); + } + if (scanline) + STBI_FREE(scanline); + } + + return hdr_data; +} + +static int stbi__hdr_info(stbi__context *s, int *x, int *y, int *comp) +{ + char buffer[STBI__HDR_BUFLEN]; + char *token; + int valid = 0; + int dummy; + + if (!x) x = &dummy; + if (!y) y = &dummy; + if (!comp) comp = &dummy; + + if (stbi__hdr_test(s) == 0) { + stbi__rewind( s ); + return 0; + } + + for(;;) { + token = stbi__hdr_gettoken(s,buffer); + if (token[0] == 0) break; + if (strcmp(token, "FORMAT=32-bit_rle_rgbe") == 0) valid = 1; + } + + if (!valid) { + stbi__rewind( s ); + return 0; + } + token = stbi__hdr_gettoken(s,buffer); + if (strncmp(token, "-Y ", 3)) { + stbi__rewind( s ); + return 0; + } + token += 3; + *y = (int) strtol(token, &token, 10); + while (*token == ' ') ++token; + if (strncmp(token, "+X ", 3)) { + stbi__rewind( s ); + return 0; + } + token += 3; + *x = (int) strtol(token, NULL, 10); + *comp = 3; + return 1; +} +#endif // STBI_NO_HDR + +#ifndef STBI_NO_BMP +static int stbi__bmp_info(stbi__context *s, int *x, int *y, int *comp) +{ + void *p; + stbi__bmp_data info; + + info.all_a = 255; + p = stbi__bmp_parse_header(s, &info); + if (p == NULL) { + stbi__rewind( s ); + return 0; + } + if (x) *x = s->img_x; + if (y) *y = s->img_y; + if (comp) { + if (info.bpp == 24 && info.ma == 0xff000000) + *comp = 3; + else + *comp = info.ma ? 4 : 3; + } + return 1; +} +#endif + +#ifndef STBI_NO_PSD +static int stbi__psd_info(stbi__context *s, int *x, int *y, int *comp) +{ + int channelCount, dummy, depth; + if (!x) x = &dummy; + if (!y) y = &dummy; + if (!comp) comp = &dummy; + if (stbi__get32be(s) != 0x38425053) { + stbi__rewind( s ); + return 0; + } + if (stbi__get16be(s) != 1) { + stbi__rewind( s ); + return 0; + } + stbi__skip(s, 6); + channelCount = stbi__get16be(s); + if (channelCount < 0 || channelCount > 16) { + stbi__rewind( s ); + return 0; + } + *y = stbi__get32be(s); + *x = stbi__get32be(s); + depth = stbi__get16be(s); + if (depth != 8 && depth != 16) { + stbi__rewind( s ); + return 0; + } + if (stbi__get16be(s) != 3) { + stbi__rewind( s ); + return 0; + } + *comp = 4; + return 1; +} + +static int stbi__psd_is16(stbi__context *s) +{ + int channelCount, depth; + if (stbi__get32be(s) != 0x38425053) { + stbi__rewind( s ); + return 0; + } + if (stbi__get16be(s) != 1) { + stbi__rewind( s ); + return 0; + } + stbi__skip(s, 6); + channelCount = stbi__get16be(s); + if (channelCount < 0 || channelCount > 16) { + stbi__rewind( s ); + return 0; + } + STBI_NOTUSED(stbi__get32be(s)); + STBI_NOTUSED(stbi__get32be(s)); + depth = stbi__get16be(s); + if (depth != 16) { + stbi__rewind( s ); + return 0; + } + return 1; +} +#endif + +#ifndef STBI_NO_PIC +static int stbi__pic_info(stbi__context *s, int *x, int *y, int *comp) +{ + int act_comp=0,num_packets=0,chained,dummy; + stbi__pic_packet packets[10]; + + if (!x) x = &dummy; + if (!y) y = &dummy; + if (!comp) comp = &dummy; + + if (!stbi__pic_is4(s,"\x53\x80\xF6\x34")) { + stbi__rewind(s); + return 0; + } + + stbi__skip(s, 88); + + *x = stbi__get16be(s); + *y = stbi__get16be(s); + if (stbi__at_eof(s)) { + stbi__rewind( s); + return 0; + } + if ( (*x) != 0 && (1 << 28) / (*x) < (*y)) { + stbi__rewind( s ); + return 0; + } + + stbi__skip(s, 8); + + do { + stbi__pic_packet *packet; + + if (num_packets==sizeof(packets)/sizeof(packets[0])) + return 0; + + packet = &packets[num_packets++]; + chained = stbi__get8(s); + packet->size = stbi__get8(s); + packet->type = stbi__get8(s); + packet->channel = stbi__get8(s); + act_comp |= packet->channel; + + if (stbi__at_eof(s)) { + stbi__rewind( s ); + return 0; + } + if (packet->size != 8) { + stbi__rewind( s ); + return 0; + } + } while (chained); + + *comp = (act_comp & 0x10 ? 4 : 3); + + return 1; +} +#endif + +// ************************************************************************************************* +// Portable Gray Map and Portable Pixel Map loader +// by Ken Miller +// +// PGM: http://netpbm.sourceforge.net/doc/pgm.html +// PPM: http://netpbm.sourceforge.net/doc/ppm.html +// +// Known limitations: +// Does not support comments in the header section +// Does not support ASCII image data (formats P2 and P3) + +#ifndef STBI_NO_PNM + +static int stbi__pnm_test(stbi__context *s) +{ + char p, t; + p = (char) stbi__get8(s); + t = (char) stbi__get8(s); + if (p != 'P' || (t != '5' && t != '6')) { + stbi__rewind( s ); + return 0; + } + return 1; +} + +static void *stbi__pnm_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri) +{ + stbi_uc *out; + STBI_NOTUSED(ri); + + ri->bits_per_channel = stbi__pnm_info(s, (int *)&s->img_x, (int *)&s->img_y, (int *)&s->img_n); + if (ri->bits_per_channel == 0) + return 0; + + if (s->img_y > STBI_MAX_DIMENSIONS) return stbi__errpuc("too large","Very large image (corrupt?)"); + if (s->img_x > STBI_MAX_DIMENSIONS) return stbi__errpuc("too large","Very large image (corrupt?)"); + + *x = s->img_x; + *y = s->img_y; + if (comp) *comp = s->img_n; + + if (!stbi__mad4sizes_valid(s->img_n, s->img_x, s->img_y, ri->bits_per_channel / 8, 0)) + return stbi__errpuc("too large", "PNM too large"); + + out = (stbi_uc *) stbi__malloc_mad4(s->img_n, s->img_x, s->img_y, ri->bits_per_channel / 8, 0); + if (!out) return stbi__errpuc("outofmem", "Out of memory"); + if (!stbi__getn(s, out, s->img_n * s->img_x * s->img_y * (ri->bits_per_channel / 8))) { + STBI_FREE(out); + return stbi__errpuc("bad PNM", "PNM file truncated"); + } + + if (req_comp && req_comp != s->img_n) { + if (ri->bits_per_channel == 16) { + out = (stbi_uc *) stbi__convert_format16((stbi__uint16 *) out, s->img_n, req_comp, s->img_x, s->img_y); + } else { + out = stbi__convert_format(out, s->img_n, req_comp, s->img_x, s->img_y); + } + if (out == NULL) return out; // stbi__convert_format frees input on failure + } + return out; +} + +static int stbi__pnm_isspace(char c) +{ + return c == ' ' || c == '\t' || c == '\n' || c == '\v' || c == '\f' || c == '\r'; +} + +static void stbi__pnm_skip_whitespace(stbi__context *s, char *c) +{ + for (;;) { + while (!stbi__at_eof(s) && stbi__pnm_isspace(*c)) + *c = (char) stbi__get8(s); + + if (stbi__at_eof(s) || *c != '#') + break; + + while (!stbi__at_eof(s) && *c != '\n' && *c != '\r' ) + *c = (char) stbi__get8(s); + } +} + +static int stbi__pnm_isdigit(char c) +{ + return c >= '0' && c <= '9'; +} + +static int stbi__pnm_getinteger(stbi__context *s, char *c) +{ + int value = 0; + + while (!stbi__at_eof(s) && stbi__pnm_isdigit(*c)) { + value = value*10 + (*c - '0'); + *c = (char) stbi__get8(s); + if((value > 214748364) || (value == 214748364 && *c > '7')) + return stbi__err("integer parse overflow", "Parsing an integer in the PPM header overflowed a 32-bit int"); + } + + return value; +} + +static int stbi__pnm_info(stbi__context *s, int *x, int *y, int *comp) +{ + int maxv, dummy; + char c, p, t; + + if (!x) x = &dummy; + if (!y) y = &dummy; + if (!comp) comp = &dummy; + + stbi__rewind(s); + + // Get identifier + p = (char) stbi__get8(s); + t = (char) stbi__get8(s); + if (p != 'P' || (t != '5' && t != '6')) { + stbi__rewind(s); + return 0; + } + + *comp = (t == '6') ? 3 : 1; // '5' is 1-component .pgm; '6' is 3-component .ppm + + c = (char) stbi__get8(s); + stbi__pnm_skip_whitespace(s, &c); + + *x = stbi__pnm_getinteger(s, &c); // read width + if(*x == 0) + return stbi__err("invalid width", "PPM image header had zero or overflowing width"); + stbi__pnm_skip_whitespace(s, &c); + + *y = stbi__pnm_getinteger(s, &c); // read height + if (*y == 0) + return stbi__err("invalid width", "PPM image header had zero or overflowing width"); + stbi__pnm_skip_whitespace(s, &c); + + maxv = stbi__pnm_getinteger(s, &c); // read max value + if (maxv > 65535) + return stbi__err("max value > 65535", "PPM image supports only 8-bit and 16-bit images"); + else if (maxv > 255) + return 16; + else + return 8; +} + +static int stbi__pnm_is16(stbi__context *s) +{ + if (stbi__pnm_info(s, NULL, NULL, NULL) == 16) + return 1; + return 0; +} +#endif + +static int stbi__info_main(stbi__context *s, int *x, int *y, int *comp) +{ + #ifndef STBI_NO_JPEG + if (stbi__jpeg_info(s, x, y, comp)) return 1; + #endif + + #ifndef STBI_NO_PNG + if (stbi__png_info(s, x, y, comp)) return 1; + #endif + + #ifndef STBI_NO_GIF + if (stbi__gif_info(s, x, y, comp)) return 1; + #endif + + #ifndef STBI_NO_BMP + if (stbi__bmp_info(s, x, y, comp)) return 1; + #endif + + #ifndef STBI_NO_PSD + if (stbi__psd_info(s, x, y, comp)) return 1; + #endif + + #ifndef STBI_NO_PIC + if (stbi__pic_info(s, x, y, comp)) return 1; + #endif + + #ifndef STBI_NO_PNM + if (stbi__pnm_info(s, x, y, comp)) return 1; + #endif + + #ifndef STBI_NO_HDR + if (stbi__hdr_info(s, x, y, comp)) return 1; + #endif + + // test tga last because it's a crappy test! + #ifndef STBI_NO_TGA + if (stbi__tga_info(s, x, y, comp)) + return 1; + #endif + return stbi__err("unknown image type", "Image not of any known type, or corrupt"); +} + +static int stbi__is_16_main(stbi__context *s) +{ + #ifndef STBI_NO_PNG + if (stbi__png_is16(s)) return 1; + #endif + + #ifndef STBI_NO_PSD + if (stbi__psd_is16(s)) return 1; + #endif + + #ifndef STBI_NO_PNM + if (stbi__pnm_is16(s)) return 1; + #endif + return 0; +} + +#ifndef STBI_NO_STDIO +STBIDEF int stbi_info(char const *filename, int *x, int *y, int *comp) +{ + FILE *f = stbi__fopen(filename, "rb"); + int result; + if (!f) return stbi__err("can't fopen", "Unable to open file"); + result = stbi_info_from_file(f, x, y, comp); + fclose(f); + return result; +} + +STBIDEF int stbi_info_from_file(FILE *f, int *x, int *y, int *comp) +{ + int r; + stbi__context s; + long pos = ftell(f); + stbi__start_file(&s, f); + r = stbi__info_main(&s,x,y,comp); + fseek(f,pos,SEEK_SET); + return r; +} + +STBIDEF int stbi_is_16_bit(char const *filename) +{ + FILE *f = stbi__fopen(filename, "rb"); + int result; + if (!f) return stbi__err("can't fopen", "Unable to open file"); + result = stbi_is_16_bit_from_file(f); + fclose(f); + return result; +} + +STBIDEF int stbi_is_16_bit_from_file(FILE *f) +{ + int r; + stbi__context s; + long pos = ftell(f); + stbi__start_file(&s, f); + r = stbi__is_16_main(&s); + fseek(f,pos,SEEK_SET); + return r; +} +#endif // !STBI_NO_STDIO + +STBIDEF int stbi_info_from_memory(stbi_uc const *buffer, int len, int *x, int *y, int *comp) +{ + stbi__context s; + stbi__start_mem(&s,buffer,len); + return stbi__info_main(&s,x,y,comp); +} + +STBIDEF int stbi_info_from_callbacks(stbi_io_callbacks const *c, void *user, int *x, int *y, int *comp) +{ + stbi__context s; + stbi__start_callbacks(&s, (stbi_io_callbacks *) c, user); + return stbi__info_main(&s,x,y,comp); +} + +STBIDEF int stbi_is_16_bit_from_memory(stbi_uc const *buffer, int len) +{ + stbi__context s; + stbi__start_mem(&s,buffer,len); + return stbi__is_16_main(&s); +} + +STBIDEF int stbi_is_16_bit_from_callbacks(stbi_io_callbacks const *c, void *user) +{ + stbi__context s; + stbi__start_callbacks(&s, (stbi_io_callbacks *) c, user); + return stbi__is_16_main(&s); +} + +#endif // STB_IMAGE_IMPLEMENTATION + +/* + revision history: + 2.20 (2019-02-07) support utf8 filenames in Windows; fix warnings and platform ifdefs + 2.19 (2018-02-11) fix warning + 2.18 (2018-01-30) fix warnings + 2.17 (2018-01-29) change sbti__shiftsigned to avoid clang -O2 bug + 1-bit BMP + *_is_16_bit api + avoid warnings + 2.16 (2017-07-23) all functions have 16-bit variants; + STBI_NO_STDIO works again; + compilation fixes; + fix rounding in unpremultiply; + optimize vertical flip; + disable raw_len validation; + documentation fixes + 2.15 (2017-03-18) fix png-1,2,4 bug; now all Imagenet JPGs decode; + warning fixes; disable run-time SSE detection on gcc; + uniform handling of optional "return" values; + thread-safe initialization of zlib tables + 2.14 (2017-03-03) remove deprecated STBI_JPEG_OLD; fixes for Imagenet JPGs + 2.13 (2016-11-29) add 16-bit API, only supported for PNG right now + 2.12 (2016-04-02) fix typo in 2.11 PSD fix that caused crashes + 2.11 (2016-04-02) allocate large structures on the stack + remove white matting for transparent PSD + fix reported channel count for PNG & BMP + re-enable SSE2 in non-gcc 64-bit + support RGB-formatted JPEG + read 16-bit PNGs (only as 8-bit) + 2.10 (2016-01-22) avoid warning introduced in 2.09 by STBI_REALLOC_SIZED + 2.09 (2016-01-16) allow comments in PNM files + 16-bit-per-pixel TGA (not bit-per-component) + info() for TGA could break due to .hdr handling + info() for BMP to shares code instead of sloppy parse + can use STBI_REALLOC_SIZED if allocator doesn't support realloc + code cleanup + 2.08 (2015-09-13) fix to 2.07 cleanup, reading RGB PSD as RGBA + 2.07 (2015-09-13) fix compiler warnings + partial animated GIF support + limited 16-bpc PSD support + #ifdef unused functions + bug with < 92 byte PIC,PNM,HDR,TGA + 2.06 (2015-04-19) fix bug where PSD returns wrong '*comp' value + 2.05 (2015-04-19) fix bug in progressive JPEG handling, fix warning + 2.04 (2015-04-15) try to re-enable SIMD on MinGW 64-bit + 2.03 (2015-04-12) extra corruption checking (mmozeiko) + stbi_set_flip_vertically_on_load (nguillemot) + fix NEON support; fix mingw support + 2.02 (2015-01-19) fix incorrect assert, fix warning + 2.01 (2015-01-17) fix various warnings; suppress SIMD on gcc 32-bit without -msse2 + 2.00b (2014-12-25) fix STBI_MALLOC in progressive JPEG + 2.00 (2014-12-25) optimize JPG, including x86 SSE2 & NEON SIMD (ryg) + progressive JPEG (stb) + PGM/PPM support (Ken Miller) + STBI_MALLOC,STBI_REALLOC,STBI_FREE + GIF bugfix -- seemingly never worked + STBI_NO_*, STBI_ONLY_* + 1.48 (2014-12-14) fix incorrectly-named assert() + 1.47 (2014-12-14) 1/2/4-bit PNG support, both direct and paletted (Omar Cornut & stb) + optimize PNG (ryg) + fix bug in interlaced PNG with user-specified channel count (stb) + 1.46 (2014-08-26) + fix broken tRNS chunk (colorkey-style transparency) in non-paletted PNG + 1.45 (2014-08-16) + fix MSVC-ARM internal compiler error by wrapping malloc + 1.44 (2014-08-07) + various warning fixes from Ronny Chevalier + 1.43 (2014-07-15) + fix MSVC-only compiler problem in code changed in 1.42 + 1.42 (2014-07-09) + don't define _CRT_SECURE_NO_WARNINGS (affects user code) + fixes to stbi__cleanup_jpeg path + added STBI_ASSERT to avoid requiring assert.h + 1.41 (2014-06-25) + fix search&replace from 1.36 that messed up comments/error messages + 1.40 (2014-06-22) + fix gcc struct-initialization warning + 1.39 (2014-06-15) + fix to TGA optimization when req_comp != number of components in TGA; + fix to GIF loading because BMP wasn't rewinding (whoops, no GIFs in my test suite) + add support for BMP version 5 (more ignored fields) + 1.38 (2014-06-06) + suppress MSVC warnings on integer casts truncating values + fix accidental rename of 'skip' field of I/O + 1.37 (2014-06-04) + remove duplicate typedef + 1.36 (2014-06-03) + convert to header file single-file library + if de-iphone isn't set, load iphone images color-swapped instead of returning NULL + 1.35 (2014-05-27) + various warnings + fix broken STBI_SIMD path + fix bug where stbi_load_from_file no longer left file pointer in correct place + fix broken non-easy path for 32-bit BMP (possibly never used) + TGA optimization by Arseny Kapoulkine + 1.34 (unknown) + use STBI_NOTUSED in stbi__resample_row_generic(), fix one more leak in tga failure case + 1.33 (2011-07-14) + make stbi_is_hdr work in STBI_NO_HDR (as specified), minor compiler-friendly improvements + 1.32 (2011-07-13) + support for "info" function for all supported filetypes (SpartanJ) + 1.31 (2011-06-20) + a few more leak fixes, bug in PNG handling (SpartanJ) + 1.30 (2011-06-11) + added ability to load files via callbacks to accomidate custom input streams (Ben Wenger) + removed deprecated format-specific test/load functions + removed support for installable file formats (stbi_loader) -- would have been broken for IO callbacks anyway + error cases in bmp and tga give messages and don't leak (Raymond Barbiero, grisha) + fix inefficiency in decoding 32-bit BMP (David Woo) + 1.29 (2010-08-16) + various warning fixes from Aurelien Pocheville + 1.28 (2010-08-01) + fix bug in GIF palette transparency (SpartanJ) + 1.27 (2010-08-01) + cast-to-stbi_uc to fix warnings + 1.26 (2010-07-24) + fix bug in file buffering for PNG reported by SpartanJ + 1.25 (2010-07-17) + refix trans_data warning (Won Chun) + 1.24 (2010-07-12) + perf improvements reading from files on platforms with lock-heavy fgetc() + minor perf improvements for jpeg + deprecated type-specific functions so we'll get feedback if they're needed + attempt to fix trans_data warning (Won Chun) + 1.23 fixed bug in iPhone support + 1.22 (2010-07-10) + removed image *writing* support + stbi_info support from Jetro Lauha + GIF support from Jean-Marc Lienher + iPhone PNG-extensions from James Brown + warning-fixes from Nicolas Schulz and Janez Zemva (i.stbi__err. Janez (U+017D)emva) + 1.21 fix use of 'stbi_uc' in header (reported by jon blow) + 1.20 added support for Softimage PIC, by Tom Seddon + 1.19 bug in interlaced PNG corruption check (found by ryg) + 1.18 (2008-08-02) + fix a threading bug (local mutable static) + 1.17 support interlaced PNG + 1.16 major bugfix - stbi__convert_format converted one too many pixels + 1.15 initialize some fields for thread safety + 1.14 fix threadsafe conversion bug + header-file-only version (#define STBI_HEADER_FILE_ONLY before including) + 1.13 threadsafe + 1.12 const qualifiers in the API + 1.11 Support installable IDCT, colorspace conversion routines + 1.10 Fixes for 64-bit (don't use "unsigned long") + optimized upsampling by Fabian "ryg" Giesen + 1.09 Fix format-conversion for PSD code (bad global variables!) + 1.08 Thatcher Ulrich's PSD code integrated by Nicolas Schulz + 1.07 attempt to fix C++ warning/errors again + 1.06 attempt to fix C++ warning/errors again + 1.05 fix TGA loading to return correct *comp and use good luminance calc + 1.04 default float alpha is 1, not 255; use 'void *' for stbi_image_free + 1.03 bugfixes to STBI_NO_STDIO, STBI_NO_HDR + 1.02 support for (subset of) HDR files, float interface for preferred access to them + 1.01 fix bug: possible bug in handling right-side up bmps... not sure + fix bug: the stbi__bmp_load() and stbi__tga_load() functions didn't work at all + 1.00 interface to zlib that skips zlib header + 0.99 correct handling of alpha in palette + 0.98 TGA loader by lonesock; dynamically add loaders (untested) + 0.97 jpeg errors on too large a file; also catch another malloc failure + 0.96 fix detection of invalid v value - particleman@mollyrocket forum + 0.95 during header scan, seek to markers in case of padding + 0.94 STBI_NO_STDIO to disable stdio usage; rename all #defines the same + 0.93 handle jpegtran output; verbose errors + 0.92 read 4,8,16,24,32-bit BMP files of several formats + 0.91 output 24-bit Windows 3.0 BMP files + 0.90 fix a few more warnings; bump version number to approach 1.0 + 0.61 bugfixes due to Marc LeBlanc, Christopher Lloyd + 0.60 fix compiling as c++ + 0.59 fix warnings: merge Dave Moore's -Wall fixes + 0.58 fix bug: zlib uncompressed mode len/nlen was wrong endian + 0.57 fix bug: jpg last huffman symbol before marker was >9 bits but less than 16 available + 0.56 fix bug: zlib uncompressed mode len vs. nlen + 0.55 fix bug: restart_interval not initialized to 0 + 0.54 allow NULL for 'int *comp' + 0.53 fix bug in png 3->4; speedup png decoding + 0.52 png handles req_comp=3,4 directly; minor cleanup; jpeg comments + 0.51 obey req_comp requests, 1-component jpegs return as 1-component, + on 'test' only check type, not whether we support this variant + 0.50 (2006-11-19) + first released version +*/ + + +/* +------------------------------------------------------------------------------ +This software is available under 2 licenses -- choose whichever you prefer. +------------------------------------------------------------------------------ +ALTERNATIVE A - MIT License +Copyright (c) 2017 Sean Barrett +Permission is hereby granted, free of charge, to any person obtaining a copy of +this software and associated documentation files (the "Software"), to deal in +the Software without restriction, including without limitation the rights to +use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies +of the Software, and to permit persons to whom the Software is furnished to do +so, subject to the following conditions: +The above copyright notice and this permission notice shall be included in all +copies or substantial portions of the Software. +THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE +AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER +LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE +SOFTWARE. +------------------------------------------------------------------------------ +ALTERNATIVE B - Public Domain (www.unlicense.org) +This is free and unencumbered software released into the public domain. +Anyone is free to copy, modify, publish, use, compile, sell, or distribute this +software, either in source code form or as a compiled binary, for any purpose, +commercial or non-commercial, and by any means. +In jurisdictions that recognize copyright laws, the author or authors of this +software dedicate any and all copyright interest in the software to the public +domain. We make this dedication for the benefit of the public at large and to +the detriment of our heirs and successors. We intend this dedication to be an +overt act of relinquishment in perpetuity of all present and future rights to +this software under copyright law. +THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE +AUTHORS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN +ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION +WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. +------------------------------------------------------------------------------ +*/ diff --git a/src/codegen/codegen.d b/src/codegen/codegen.d index d6b8632..95e26aa 100644 --- a/src/codegen/codegen.d +++ b/src/codegen/codegen.d @@ -4,6 +4,9 @@ import std.stdio; import std.file; import std.string; import std.path; +import std.uni; +import std.conv; +import std.array; void main() @@ -55,21 +58,50 @@ CreateAssetLookups() file_hashes[dir] = hashes; } - auto f = File("src/generated/assets.d", "w"); + auto f = File("src/generated/assets_codegen.d", "w"); + scope(exit) f.close(); + + string asset_str = "import aliases;\n\n"; + + asset_str ~= "enum AssetType : u32\n{\n\tNone,\n"; + foreach(i, dir; dirs) + { + auto graphemes = dir.byGrapheme.array; + string upper = [graphemes[0]].byCodePoint.text.toUpper; + graphemes[0] = upper.decodeGrapheme; + + string asset_type = graphemes.byCodePoint.text.stripRight("s"); + asset_str ~= format("\t%s,\n", asset_type); + } + asset_str ~= "}\n\nalias AT = AssetType;\n\n"; foreach(dir; dirs) { string asset_type = stripRight(dir, "s").toUpper(); - string file_array = format("static immutable string[] %s_FILES = [\n", asset_type); + asset_str ~= format("static immutable string[] %s_FILES = [\n", asset_type); foreach(file; file_names[dir]) { - file_array ~= format("\t\"%s\",\n", file); + asset_str ~= format("\t\"%s\",\n", file); } - file_array ~= "];\n\n"; + asset_str ~= "];\n\n"; + + + asset_str ~= format("static immutable string[] %s_NAMES = [\n", asset_type); + foreach(name; base_names[dir]) + { + asset_str ~= format("\t\"%s\",\n", name); + } + asset_str ~= "];\n\n"; + + asset_str ~= format("static immutable u64[] %s_HASHES = [\n", asset_type); + foreach(hash; file_hashes[dir]) + { + asset_str ~= format("\t%d,\n", hash); + } + asset_str ~= "];\n\n"; - f.write(file_array); } - f.close(); + f.write(asset_str); } diff --git a/src/generated/assets_codegen.d b/src/generated/assets_codegen.d new file mode 100644 index 0000000..e3ebbc2 --- /dev/null +++ b/src/generated/assets_codegen.d @@ -0,0 +1,84 @@ +import aliases; + +enum AssetType : u32 +{ + None, + Model, + Shader, + Texture, +} + +alias AT = AssetType; + +static immutable string[] MODEL_FILES = [ + "models/test_char.m3d", + "models/yoda.m3d", +]; + +static immutable string[] MODEL_NAMES = [ + "models/test_char", + "models/yoda", +]; + +static immutable u64[] MODEL_HASHES = [ + 13826959199295087925, + 4559395153940738542, +]; + +static immutable string[] SHADER_FILES = [ + "shaders/gui.frag.spv", + "shaders/pbr.frag.spv", + "shaders/gui.vert.spv", + "shaders/quad.frag.spv", + "shaders/quad.vert.spv", + "shaders/pbr.vert.spv", +]; + +static immutable string[] SHADER_NAMES = [ + "shaders/gui.frag", + "shaders/pbr.frag", + "shaders/gui.vert", + "shaders/quad.frag", + "shaders/quad.vert", + "shaders/pbr.vert", +]; + +static immutable u64[] SHADER_HASHES = [ + 15780387719315455808, + 2230071466542309169, + 14797956403837654625, + 8430018914716708078, + 14432191255225961360, + 8518761701216801634, +]; + +static immutable string[] TEXTURE_FILES = [ + "textures/ham_smoke.png", + "textures/cheesoid.png", + "textures/hog.jpg", + "textures/patamon.png", + "textures/pattermon.png", + "textures/hamster.png", + "textures/purplemon.png", +]; + +static immutable string[] TEXTURE_NAMES = [ + "textures/ham_smoke", + "textures/cheesoid", + "textures/hog", + "textures/patamon", + "textures/pattermon", + "textures/hamster", + "textures/purplemon", +]; + +static immutable u64[] TEXTURE_HASHES = [ + 15457434128510259736, + 12443444479937967236, + 5538438794723924882, + 16650761267170532297, + 11718504567089932798, + 7627422980398294448, + 14316598952102237724, +]; + diff --git a/src/generated/codegen.d b/src/generated/codegen.d deleted file mode 100644 index a0c8a66..0000000 --- a/src/generated/codegen.d +++ /dev/null @@ -1,24 +0,0 @@ -static immutable string[] MODEL_FILES = [ - "models/test_char.m3d", - "models/yoda.m3d", -]; - -static immutable string[] SHADER_FILES = [ - "shaders/gui.frag.spv", - "shaders/pbr.frag.spv", - "shaders/gui.vert.spv", - "shaders/quad.frag.spv", - "shaders/quad.vert.spv", - "shaders/pbr.vert.spv", -]; - -static immutable string[] TEXTURE_FILES = [ - "textures/ham_smoke.png", - "textures/cheesoid.png", - "textures/hog.jpg", - "textures/patamon.png", - "textures/pattermon.png", - "textures/hamster.png", - "textures/purplemon.png", -]; - diff --git a/src/packer/packer.d b/src/packer/packer.d index deadc27..bf6d6c9 100644 --- a/src/packer/packer.d +++ b/src/packer/packer.d @@ -1,9 +1,12 @@ import aliases; +import includes; +import std.stdio; import std.file; import util; import assets; - -const u32 FILE_VERSION = 1; +import assets_codegen; +import std.traits; +import std.algorithm.comparison; void main() { @@ -12,9 +15,190 @@ void main() chdir("build"); } - static foreach(file; AssetFiles) + + PackFile(); + TestFile(); +} + +void +PackFile() +{ + File ap = File("./assets.sgp", "wb"); + scope(exit) { - Log(file); + ap.flush(); + ap.close(); + chdir("../build"); + } + + chdir("../assets"); + + FileHeader h = InitHeader(); + ap.rawWrite([h]); + + u64 offset = FileHeader.sizeof + (AssetInfo.sizeof * h.asset_count); + AssetInfo[] asset_info; + foreach(type; [ EnumMembers!AssetType ]) + { + auto files = GetAssetFiles(type); + auto names = GetAssetNames(type); + auto hashes = GetAssetHashes(type); + + foreach(i, file; files) + { + auto f = File(file, "rb"); + scope(exit) f.close(); + + u64 length = cast(u64)f.size(); + + AssetInfo info = { + hash: hashes[i], + offset: offset, + length: length, + type: type, + }; + + auto data = f.rawRead(new u8[length]); + assert(length == data.length, "rawRead failure: data length returned doesn't match"); + + ap.seek(offset); + ap.rawWrite(data); + + offset += length; + + if (type == AT.Texture) + { + info.texture = GetTexMeta(data); + } + + if (type == AT.Model) + { + info.model = GetModelMeta(data); + } + + asset_info ~= info; + } + } + + ap.seek(FileHeader.sizeof); + ap.rawWrite(asset_info); +} + +TexMeta +GetTexMeta(u8[] data) +{ + int ch = 4; + int width, height, has_ch; + auto img = stbi_load_from_memory(data.ptr, cast(int)data.length, &width, &height, &has_ch, ch); + + assert(img != null, "stbi_load_from_image failure: image is NULL"); + assert(width > 0 && height > 0 && has_ch > 0, "stbi_load_from_image failure: dimensions are invalid"); + + stbi_image_free(img); + + return TexMeta(w: width, h: height, ch: has_ch); +} + +ModelMeta +GetModelMeta(u8[] data) +{ + m3d_t* model = m3d_load(data.ptr, null, null, null); + assert(model != null, "m3d_load failure: model is null"); + + u64 index_count = cast(u64)(model.numface * 3); + + m3d_free(model); + + return ModelMeta(index_count: index_count); +} + +void +TestFile() +{ + File ap = File("assets.sgp", "rb"); + scope(exit) + { + ap.flush(); + ap.close(); + chdir("../build"); + } + + FileHeader file_header = ap.rawRead(new FileHeader[1])[0]; + FileHeader test_header = InitHeader(); + assert(file_header == test_header, "TestFile failure: Header is incorrect"); + + AssetInfo[] file_info = ap.rawRead(new AssetInfo[file_header.asset_count]); + assert(file_info.length == file_header.asset_count, "TestFile failure: Incorrect AssetInfo length returned"); + + chdir("../assets"); + + u64 asset_index = 0; + foreach(type; [ EnumMembers!AssetType ]) + { + auto files = GetAssetFiles(type); + auto names = GetAssetNames(type); + auto hashes = GetAssetHashes(type); + + foreach(i, file; files) + { + scope(exit) asset_index += 1; + + AssetInfo* info = file_info.ptr + asset_index; + File asset = File(file, "rb"); + + u8[] data = asset.rawRead(new u8[asset.size()]); + assert(data.length == info.length, "TestFile failure: File length read is incorrect"); + + ap.seek(info.offset); + u8[] pack_data = ap.rawRead(new u8[info.length]); + assert(equal!((a, b) => a == b)(data[], pack_data[]), "TestFile failure: Asset data does not match file data"); + + if (type == AT.Texture) + { + assert(info.texture == GetTexMeta(pack_data), "TestFile failure: Texture meta is incorrect"); + } + + if (type == AT.Model) + { + assert(info.model == GetModelMeta(pack_data), "TestFile failure: Model meta is incorrect"); + } + } + } +} + +immutable(string[]) +GetAssetFiles(AssetType type) +{ + switch (type) + { + case AT.Model: return MODEL_FILES; + case AT.Shader: return SHADER_FILES; + case AT.Texture: return TEXTURE_FILES; + default: return []; + } +} + +immutable(string[]) +GetAssetNames(AssetType type) +{ + switch (type) + { + case AT.Model: return MODEL_NAMES; + case AT.Shader: return SHADER_NAMES; + case AT.Texture: return TEXTURE_NAMES; + default: return []; + } +} + +immutable(u64[]) +GetAssetHashes(AssetType type) +{ + switch (type) + { + case AT.Model: return MODEL_HASHES; + case AT.Shader: return SHADER_HASHES; + case AT.Texture: return TEXTURE_HASHES; + default: return []; } } @@ -25,12 +209,14 @@ MagicValue(string str) return cast(u32)(cast(u32)(str[0] << 24) | cast(u32)(str[1] << 16) | cast(u32)(str[2] << 8) | cast(u32)(str[3] << 0)); } -FileHeader +static FileHeader InitHeader() { FileHeader header = { magic: MagicValue("steg"), file_version: FILE_VERSION, + asset_count: cast(u64)(MODEL_FILES.length + SHADER_FILES.length + TEXTURE_FILES.length), + asset_info_offset: FileHeader.sizeof, }; return header; diff --git a/src/shared/assets.d b/src/shared/assets.d index a45a509..ffc1985 100644 --- a/src/shared/assets.d +++ b/src/shared/assets.d @@ -1,12 +1,15 @@ import aliases; import std.file; +import assets_codegen; + +const u32 FILE_VERSION = 1; struct FileHeader { u32 magic; u32 file_version; - u64 asset_counts; - u64 asset_offset; + u64 asset_count; + u64 asset_info_offset; } struct ModelMeta @@ -21,16 +24,6 @@ struct TexMeta u32 ch; } -enum AssetType : u32 -{ - None, - Shader, - Texture, - Model, -} - -alias AT = AssetType; - struct AssetInfo { union @@ -40,6 +33,6 @@ struct AssetInfo }; u64 hash; u64 offset; - u64 len; + u64 length; AssetType type; } diff --git a/src/shared/includes.c b/src/shared/includes.c index d4455c5..45efd25 100644 --- a/src/shared/includes.c +++ b/src/shared/includes.c @@ -22,4 +22,12 @@ #include "../../external/vma/vk_mem_alloc.h" -#include "../../external/xxhash/xxhash.h" +#include "../../external/xxhash/xxhash.c" + +#define STB_IMAGE_IMPLEMENTATION + +#include "../../external/stb/stb_image.h" + +#define M3D_IMPLEMENTATION + +#include "../../external/m3d/m3d.h"