awesome!!!

third-party/stb/tools/README.footer.md

@@ -0,0 +1,135 @@
+
+FAQ
+---
+
+#### What's the license?
+
+These libraries are in the public domain. You can do anything you
+want with them. You have no legal obligation
+to do anything else, although I appreciate attribution.
+
+They are also licensed under the MIT open source license, if you have lawyers
+who are unhappy with public domain. Every source file includes an explicit
+dual-license for you to choose from.
+
+#### How do I use these libraries?
+
+The idea behind single-header file libraries is that they're easy to distribute and deploy
+because all the code is contained in a single file. By default, the .h files in here act as
+their own header files, i.e. they declare the functions contained in the file but don't
+actually result in any code getting compiled.
+
+So in addition, you should select _exactly one_ C/C++ source file that actually instantiates
+the code, preferably a file you're not editing frequently. This file should define a
+specific macro (this is documented per-library) to actually enable the function definitions.
+For example, to use stb_image, you should have exactly one C/C++ file that doesn't
+include stb_image.h regularly, but instead does
+
+    #define STB_IMAGE_IMPLEMENTATION
+    #include "stb_image.h"
+
+The right macro to define is pointed out right at the top of each of these libraries.
+
+#### <a name="other_libs"></a> Are there other single-file public-domain/open source libraries with minimal dependencies out there?
+
+[Yes.](https://github.com/nothings/single_file_libs)
+
+#### If I wrap an stb library in a new library, does the new library have to be public domain/MIT?
+
+No, because it's public domain you can freely relicense it to whatever license your new
+library wants to be.
+
+#### What's the deal with SSE support in GCC-based compilers?
+
+stb_image will either use SSE2 (if you compile with -msse2) or
+will not use any SIMD at all, rather than trying to detect the
+processor at runtime and handle it correctly. As I understand it,
+the approved path in GCC for runtime-detection require
+you to use multiple source files, one for each CPU configuration.
+Because stb_image is a header-file library that compiles in only
+one source file, there's no approved way to build both an
+SSE-enabled and a non-SSE-enabled variation.
+
+While we've tried to work around it, we've had multiple issues over
+the years due to specific versions of gcc breaking what we're doing,
+so we've given up on it. See https://github.com/nothings/stb/issues/280
+and https://github.com/nothings/stb/issues/410 for examples.
+
+#### Some of these libraries seem redundant to existing open source libraries. Are they better somehow?
+
+Generally they're only better in that they're easier to integrate,
+easier to use, and easier to release (single file; good API; no
+attribution requirement). They may be less featureful, slower,
+and/or use more memory. If you're already using an equivalent
+library, there's probably no good reason to switch.
+
+#### Can I link directly to the table of stb libraries?
+
+You can use [this URL](https://github.com/nothings/stb#stb_libs) to link directly to that list.
+
+#### Why do you list "lines of code"? It's a terrible metric.
+
+Just to give you some idea of the internal complexity of the library,
+to help you manage your expectations, or to let you know what you're
+getting into. While not all the libraries are written in the same
+style, they're certainly similar styles, and so comparisons between
+the libraries are probably still meaningful.
+
+Note though that the lines do include both the implementation, the
+part that corresponds to a header file, and the documentation.
+
+#### Why single-file headers?
+
+Windows doesn't have standard directories where libraries
+live. That makes deploying libraries in Windows a lot more
+painful than open source developers on Unix-derivates generally
+realize. (It also makes library dependencies a lot worse in Windows.)
+
+There's also a common problem in Windows where a library was built
+against a different version of the runtime library, which causes
+link conflicts and confusion. Shipping the libs as headers means
+you normally just compile them straight into your project without
+making libraries, thus sidestepping that problem.
+
+Making them a single file makes it very easy to just
+drop them into a project that needs them. (Of course you can
+still put them in a proper shared library tree if you want.)
+
+Why not two files, one a header and one an implementation?
+The difference between 10 files and 9 files is not a big deal,
+but the difference between 2 files and 1 file is a big deal.
+You don't need to zip or tar the files up, you don't have to
+remember to attach *two* files, etc.
+
+#### Why "stb"? Is this something to do with Set-Top Boxes?
+
+No, they are just the initials for my name, Sean T. Barrett.
+This was not chosen out of egomania, but as a moderately sane
+way of namespacing the filenames and source function names.
+
+#### Will you add more image types to stb_image.h?
+
+No. As stb_image use has grown, it has become more important
+for us to focus on security of the codebase. Adding new image
+formats increases the amount of code we need to secure, so it
+is no longer worth adding new formats.
+
+#### Do you have any advice on how to create my own single-file library?
+
+Yes. https://github.com/nothings/stb/blob/master/docs/stb_howto.txt
+
+#### Why public domain?
+
+I prefer it over GPL, LGPL, BSD, zlib, etc. for many reasons.
+Some of them are listed here:
+https://github.com/nothings/stb/blob/master/docs/why_public_domain.md
+
+#### Why C?
+
+Primarily, because I use C, not C++. But it does also make it easier
+for other people to use them from other languages.
+
+#### Why not C99? stdint.h, declare-anywhere, etc.
+
+I still use MSVC 6 (1998) as my IDE because it has better human factors
+for me than later versions of MSVC.

third-party/stb/tools/README.header.md

@@ -0,0 +1,22 @@
+stb
+===
+
+single-file public domain (or MIT licensed) libraries for C/C++
+
+# This project discusses security-relevant bugs in public in Github Issues and Pull Requests, and it may take significant time for security fixes to be implemented or merged. If this poses an unreasonable risk to your project, do not use stb libraries.
+
+Noteworthy:
+
+* image loader: [stb_image.h](stb_image.h)
+* image writer: [stb_image_write.h](stb_image_write.h)
+* image resizer: [stb_image_resize2.h](stb_image_resize2.h)
+* font text rasterizer: [stb_truetype.h](stb_truetype.h)
+* typesafe containers: [stb_ds.h](stb_ds.h)
+
+Most libraries by stb, except: stb_dxt by Fabian "ryg" Giesen, stb_image_resize
+by Jorge L. "VinoBS" Rodriguez, and stb_sprintf by Jeff Roberts.
+
+<a name="stb_libs"></a>
+
+library    | lastest version | category | LoC | description
+--------------------- | ---- | -------- | --- | --------------------------------

third-party/stb/tools/README.list

@@ -0,0 +1,21 @@
+stb_vorbis.c                | audio            | decode ogg vorbis files from file/memory to float/16-bit signed output
+stb_hexwave.h               | audio            | audio waveform synthesizer
+stb_image.h                 | graphics         | image loading/decoding from file/memory: JPG, PNG, TGA, BMP, PSD, GIF, HDR, PIC
+stb_truetype.h              | graphics         | parse, decode, and rasterize characters from truetype fonts
+stb_image_write.h           | graphics         | image writing to disk: PNG, TGA, BMP
+stb_image_resize2.h         | graphics         | resize images larger/smaller with good quality
+stb_rect_pack.h             | graphics         | simple 2D rectangle packer with decent quality
+stb_perlin.h                | graphics         | perlin's revised simplex noise w/ different seeds
+stb_ds.h                    | utility          | typesafe dynamic array and hash tables for C, will compile in C++
+stb_sprintf.h               | utility          | fast sprintf, snprintf for C/C++
+stb_textedit.h              | user interface   | guts of a text editor for games etc implementing them from scratch
+stb_voxel_render.h          | 3D graphics      | Minecraft-esque voxel rendering "engine" with many more features
+stb_dxt.h                   | 3D graphics      | Fabian "ryg" Giesen's real-time DXT compressor
+stb_easy_font.h             | 3D graphics      | quick-and-dirty easy-to-deploy bitmap font for printing frame rate, etc
+stb_tilemap_editor.h        | game dev         | embeddable tilemap editor
+stb_herringbone_wang_tile.h | game dev         | herringbone Wang tile map generator
+stb_c_lexer.h               | parsing          | simplify writing parsers for C-like languages
+stb_divide.h                | math             | more useful 32-bit modulus e.g. "euclidean divide"
+stb_connected_components.h  | misc             | incrementally compute reachability on grids
+stb_leakcheck.h             | misc             | quick-and-dirty malloc/free leak-checking
+stb_include.h               | misc             | implement recursive #include support, particularly for GLSL

third-party/stb/tools/build_matrix.c

@@ -0,0 +1,137 @@
+#define STB_DEFINE
+#include "stb.h"
+
+// true if no error
+int run_command(char *batch_file, char *command)
+{
+   char buffer[4096];
+   if (batch_file[0]) {
+      sprintf(buffer, "%s && %s", batch_file, command);
+      return system(buffer) == 0;
+   } else {
+      return system(command) == 0;
+   }
+}
+
+typedef struct
+{
+   char *compiler_name;
+   char *batchfile;
+   char *objdir;
+   char *compiler;
+   char *args;
+   char *link;
+} compiler_info;
+
+compiler_info *compilers;
+char *shared_args;
+char *shared_link;
+
+typedef struct
+{
+   char *filelist;
+} project_info;
+
+project_info *projects;
+
+enum { NONE, IN_COMPILERS, IN_ARGS, IN_PROJECTS, IN_LINK };
+
+int main(int argc, char **argv)
+{
+   int state = NONE;
+   int i,j,n;
+   char **line;
+   if (argc != 2) stb_fatal("Usage: stb_build_matrix {build-file}\n");
+   line = stb_stringfile(argv[1], &n);
+   if (line == 0) stb_fatal("Couldn't open file '%s'\n", argv[1]);
+
+   for (i=0; i < n; ++i) {
+      char *p = stb_trimwhite(line[i]);
+      if (p[0] == 0) continue;
+      else if (p[0] == '#') continue;
+      else if (0 == stricmp(p, "[compilers]")) { state = IN_COMPILERS;                     }
+      else if (0 == stricmp(p, "[args]"     )) { state = IN_ARGS     ; shared_args = NULL; }
+      else if (0 == stricmp(p, "[projects]" )) { state = IN_PROJECTS ;                     }
+      else if (0 == stricmp(p, "[link]"     )) { state = IN_LINK     ; shared_link = NULL; }
+      else {
+         switch (state) {
+            case NONE: stb_fatal("Invalid text outside section at line %d.", i+1);
+            case IN_COMPILERS: {
+               char buffer[4096];
+               int count;
+               compiler_info ci;
+               char **tokens = stb_tokens_stripwhite(p, ",", &count), *batch;
+               if (count > 3) stb_fatal("Expecting name and batch file name at line %d.", i+1);
+               batch = (count==1 ? tokens[0] : tokens[1]);
+               if (strlen(batch)) 
+                  sprintf(buffer, "c:\\%s.bat", batch);
+               else
+                  strcpy(buffer, "");
+               ci.compiler_name = strdup(tokens[0]);
+               ci.batchfile     = strdup(buffer);
+               ci.compiler      = count==3 ? strdup(tokens[2]) : "cl";
+               if (0==strnicmp(batch, "vcvars_", 7))
+                  ci.objdir     = strdup(stb_sprintf("vs_%s_%d", batch+7, i));
+               else
+                  ci.objdir     = strdup(stb_sprintf("%s_%d", batch, i));
+               ci.args = shared_args;
+               ci.link = shared_link;
+               stb_arr_push(compilers, ci);
+               break;
+            }
+            case IN_ARGS: {
+               stb_arr_push(shared_args, ' ');
+               for (j=0; p[j] != 0; ++j)
+                  stb_arr_push(shared_args, p[j]);
+               break;
+            }
+            case IN_LINK: {
+               stb_arr_push(shared_link, ' ');
+               for (j=0; p[j] != 0; ++j)
+                  stb_arr_push(shared_link, p[j]);
+               break;
+            }
+            case IN_PROJECTS: {
+               project_info pi;
+               pi.filelist = strdup(p);
+               stb_arr_push(projects, pi);
+               break;
+            }
+         }
+      }
+   }
+
+   _mkdir("obj");
+   for (j=0; j < stb_arr_len(compilers); ++j) {
+      char command[4096];
+      for (i=0; i < stb_arr_len(projects); ++i) {
+         int r;
+         _mkdir(stb_sprintf("obj/%s", compilers[j].objdir));
+         if (stb_suffix(compilers[j].compiler, "cl"))
+            sprintf(command, "%s %.*s %s /link %.*s",
+                                     compilers[j].compiler,
+                                     stb_arr_len(compilers[j].args), compilers[j].args,
+                                     projects[i].filelist,
+                                     stb_arr_len(compilers[j].link), compilers[j].link);
+         else
+            sprintf(command, "%s %.*s %s %.*s",
+                                     compilers[j].compiler,
+                                     stb_arr_len(compilers[j].args), compilers[j].args,
+                                     projects[i].filelist,
+                                     stb_arr_len(compilers[j].link), compilers[j].link);
+         r = run_command(compilers[j].batchfile, command);
+         stbprint("{%c== Compiler %s == Building %s}\n", r ? '$' : '!', compilers[j].compiler_name, projects[i].filelist);
+         stb_copyfile("a.exe", stb_sprintf("obj/%s/a.exe", compilers[j].objdir));
+         //printf("Copy: %s to %s\n", "a.exe", stb_sprintf("obj/%s/a.exe", compilers[j].objdir));
+         stb_copyfile("temp.exe", stb_sprintf("obj/%s/temp.exe", compilers[j].objdir));
+         system("if EXIST a.exe del /q a.exe");
+         system("if EXIST temp.exe del /q temp.exe");
+         system("if EXIST *.obj del /q *.obj");
+         system("if EXIST *.o del /q *.o");
+         if (!r)
+            return 1;
+      }
+   }
+
+   return 0;
+}

third-party/stb/tools/easy_font_maker.c

@@ -0,0 +1,211 @@
+// This program was used to encode the data for stb_simple_font.h
+
+#define STB_DEFINE
+#include "stb.h"
+#define STB_IMAGE_IMPLEMENTATION
+#include "stb_image.h"
+
+int w,h;
+uint8 *data;
+
+int last_x[2], last_y[2];
+int num_seg[2], non_empty;
+#if 0
+typedef struct
+{
+   unsigned short first_segment;
+   unsigned char advance;
+} chardata;
+
+typedef struct
+{
+   unsigned char x:4;
+   unsigned char y:4;
+   unsigned char len:3;
+   unsigned char dir:1;
+} segment;
+
+segment *segments;
+
+void add_seg(int x, int y, int len, int horizontal)
+{
+   segment s;
+   s.x = x;
+   s.y = y;
+   s.len = len;
+   s.dir = horizontal;
+   assert(s.x == x);
+   assert(s.y == y);
+   assert(s.len == len);
+   stb_arr_push(segments, s);
+}
+#else
+typedef struct
+{
+   unsigned char first_segment:8;
+   unsigned char first_v_segment:8;
+   unsigned char advance:5;
+   unsigned char voff:1;
+} chardata;
+
+#define X_LIMIT 1
+#define LEN_LIMIT 7
+
+typedef struct
+{
+   unsigned char dx:1;
+   unsigned char y:4;
+   unsigned char len:3;
+} segment;
+
+segment *segments;
+segment *vsegments;
+
+void add_seg(int x, int y, int len, int horizontal)
+{
+   segment s;
+
+   while (x - last_x[horizontal] > X_LIMIT) {
+      add_seg(last_x[horizontal] + X_LIMIT, 0, 0, horizontal);
+   }
+   while (len > LEN_LIMIT) {
+      add_seg(x, y, LEN_LIMIT, horizontal);
+      len -= LEN_LIMIT;
+      x += LEN_LIMIT*horizontal;
+      y += LEN_LIMIT*!horizontal;
+   }
+
+   s.dx = x - last_x[horizontal];
+   s.y = y;
+   s.len = len;
+   non_empty += len != 0;
+   //assert(s.x == x);
+   assert(s.y == y);
+   assert(s.len == len);
+   ++num_seg[horizontal];
+   if (horizontal)
+      stb_arr_push(segments, s);
+   else
+      stb_arr_push(vsegments, s);
+   last_x[horizontal] = x;
+}
+
+void print_segments(segment *s)
+{
+   int i, hpos;
+   printf("   ");
+   hpos = 4;
+   for (i=0; i < stb_arr_len(s); ++i) {
+      // repack for portability
+      unsigned char seg = s[i].len + s[i].dx*8 + s[i].y*16;
+      hpos += printf("%d,", seg);
+      if (hpos > 72 && i+1 < stb_arr_len(s)) {
+         hpos = 4;
+         printf("\n    ");
+      }
+   }
+   printf("\n");
+}
+
+#endif
+
+chardata charinfo[128];
+
+int parse_char(int x, chardata *c, int offset)
+{
+   int start_x = x, end_x, top_y = 0, y;
+
+   c->first_segment = stb_arr_len(segments);
+   c->first_v_segment = stb_arr_len(vsegments) - offset;
+   assert(c->first_segment == stb_arr_len(segments));
+   assert(c->first_v_segment + offset == stb_arr_len(vsegments));
+
+   // find advance distance
+   end_x = x+1;
+   while (data[end_x*3] == 255)
+      ++end_x;
+   c->advance = end_x - start_x + 1;
+
+   last_x[0] = last_x[1] = 0;
+   last_y[0] = last_y[1] = 0;
+
+   for (y=2; y < h; ++y) {
+      for (x=start_x; x < end_x; ++x) {
+         if (data[y*3*w+x*3+1] < 255) {
+            top_y = y;
+            break;
+         }
+      }
+      if (top_y)
+         break;
+   }
+   c->voff = top_y > 2;
+   if (top_y > 2) 
+      top_y = 3;
+
+   for (x=start_x; x < end_x; ++x) {
+      int y;
+      for (y=2; y < h; ++y) {
+         if (data[y*3*w+x*3+1] < 255) {
+            if (data[y*3*w+x*3+0] == 255) { // red
+               int len=0;
+               while (y+len < h && data[(y+len)*3*w+x*3+0] == 255 && data[(y+len)*3*w+x*3+1] == 0) {
+                  data[(y+len)*3*w+x*3+0] = 0;
+                  ++len;
+               }
+               add_seg(x-start_x,y-top_y,len,0);
+            }
+            if (data[y*3*w+x*3+2] == 255) { // blue
+               int len=0;
+               while (x+len < end_x && data[y*3*w+(x+len)*3+2] == 255 && data[y*3*w+(x+len)*3+1] == 0) {
+                  data[y*3*w+(x+len)*3+2] = 0;
+                  ++len;
+               }
+               add_seg(x-start_x,y-top_y,len,1);
+            }
+         }
+      }
+   }
+   return end_x;
+}
+
+
+int main(int argc, char **argv)
+{
+   int c, x=0;
+   data = stbi_load("easy_font_raw.png", &w, &h, 0, 3);
+   for (c=32; c < 127; ++c) {
+      x = parse_char(x, &charinfo[c], 0);
+      printf("%3d -- %3d %3d\n", c, charinfo[c].first_segment, charinfo[c].first_v_segment);
+   }
+   printf("===\n");
+   printf("%d %d %d\n", num_seg[0], num_seg[1], non_empty);
+   printf("%d\n", sizeof(segments[0]) * stb_arr_len(segments));
+   printf("%d\n", sizeof(segments[0]) * stb_arr_len(segments) + sizeof(segments[0]) * stb_arr_len(vsegments) + sizeof(charinfo[32])*95);
+
+   printf("struct {\n"
+          "    unsigned char advance;\n"
+          "    unsigned char h_seg;\n"
+          "    unsigned char v_seg;\n"
+          "} stb_easy_font_charinfo[96] = {\n");
+   charinfo[c].first_segment = stb_arr_len(segments);
+   charinfo[c].first_v_segment = stb_arr_len(vsegments);
+   for (c=32; c < 128; ++c) {
+      if ((c & 3) == 0) printf("    ");
+      printf("{ %2d,%3d,%3d },",
+         charinfo[c].advance + 16*charinfo[c].voff,
+         charinfo[c].first_segment,
+         charinfo[c].first_v_segment);
+      if ((c & 3) == 3) printf("\n"); else printf("  ");
+   }
+   printf("};\n\n");
+
+   printf("unsigned char stb_easy_font_hseg[%d] = {\n", stb_arr_len(segments));
+      print_segments(segments);
+   printf("};\n\n");
+
+   printf("unsigned char stb_easy_font_vseg[%d] = {\n", stb_arr_len(vsegments));
+      print_segments(vsegments);
+   printf("};\n");
+   return 0;
+}

third-party/stb/tools/make_readme.c

@@ -0,0 +1,65 @@
+#define STB_DEFINE
+#include "../stb.h"
+
+int main(int argc, char  **argv)
+{
+   int i;
+   int hlen, flen, listlen, total_lines = 0;
+   char *header = stb_file("README.header.md", &hlen);      // stb_file - read file into malloc()ed buffer
+   char *footer = stb_file("README.footer.md", &flen);      // stb_file - read file into malloc()ed buffer
+   char **list  = stb_stringfile("README.list", &listlen);  // stb_stringfile - read file lines into malloced array of strings
+
+   FILE *f = fopen("../README.md", "wb");
+
+   fprintf(f, "<!---   THIS FILE IS AUTOMATICALLY GENERATED, DO NOT CHANGE IT BY HAND   --->\r\n\r\n");
+   fwrite(header, 1, hlen, f);
+
+   for (i=0; i < listlen; ++i) {
+      int num,j;
+      char **tokens = stb_tokens_stripwhite(list[i], "|", &num);  // stb_tokens -- tokenize string into malloced array of strings
+      int num_lines;
+      char **lines = stb_stringfile(stb_sprintf("../%s", tokens[0]), &num_lines);
+      char *s1, *s2,*s3;
+      if (lines == NULL) stb_fatal("Couldn't open '%s'", tokens[0]);
+      s1 = strchr(lines[0], '-');
+      if (!s1) stb_fatal("Couldn't find '-' before version number in %s", tokens[0]); // stb_fatal -- print error message & exit
+      s2 = strchr(s1+2, '-');
+      if (!s2) stb_fatal("Couldn't find '-' after version number in %s", tokens[0]);  // stb_fatal -- print error message & exit
+      *s2 = 0;
+      s1 += 1;
+      s1 = stb_trimwhite(s1);                  // stb_trimwhite -- advance pointer to after whitespace & delete trailing whitespace
+      if (*s1 == 'v') ++s1;
+      s3 = tokens[0];
+      stb_trimwhite(s3);
+      fprintf(f, "**[");
+      if (strlen(s3) < 21) {
+         fprintf(f, "%s", tokens[0]);
+      } else {
+         char buffer[256];
+         strncpy(buffer, s3, 18);
+         buffer[18] = 0;   
+         strcat(buffer, "...");
+         fprintf(f, "%s", buffer);
+      }
+      fprintf(f, "](%s)**", tokens[0]);
+      fprintf(f, " | %s", s1);
+      s1 = stb_trimwhite(tokens[1]);           // stb_trimwhite -- advance pointer to after whitespace & delete trailing whitespace
+      s2 = stb_dupreplace(s1, " ", "&nbsp;");  // stb_dupreplace -- search & replace string and malloc result
+      fprintf(f, " | %s", s2);
+      free(s2);
+      fprintf(f, " | %d", num_lines);
+      total_lines += num_lines;
+      for (j=2; j < num; ++j)
+         fprintf(f, " | %s", tokens[j]);
+      fprintf(f, "\r\n");
+   }
+
+   fprintf(f, "\r\n");
+   fprintf(f, "Total libraries: %d\r\n", listlen);
+   fprintf(f, "Total lines of C code: %d\r\n\r\n", total_lines);
+
+   fwrite(footer, 1, flen, f);
+   fclose(f);
+
+   return 0;
+}

third-party/stb/tools/make_readme.dsp

@@ -0,0 +1,97 @@
+# Microsoft Developer Studio Project File - Name="make_readme" - Package Owner=<4>
+# Microsoft Developer Studio Generated Build File, Format Version 6.00
+# ** DO NOT EDIT **
+
+# TARGTYPE "Win32 (x86) Console Application" 0x0103
+
+CFG=make_readme - Win32 Debug
+!MESSAGE This is not a valid makefile. To build this project using NMAKE,
+!MESSAGE use the Export Makefile command and run
+!MESSAGE 
+!MESSAGE NMAKE /f "make_readme.mak".
+!MESSAGE 
+!MESSAGE You can specify a configuration when running NMAKE
+!MESSAGE by defining the macro CFG on the command line. For example:
+!MESSAGE 
+!MESSAGE NMAKE /f "make_readme.mak" CFG="make_readme - Win32 Debug"
+!MESSAGE 
+!MESSAGE Possible choices for configuration are:
+!MESSAGE 
+!MESSAGE "make_readme - Win32 Release" (based on "Win32 (x86) Console Application")
+!MESSAGE "make_readme - Win32 Debug" (based on "Win32 (x86) Console Application")
+!MESSAGE 
+
+# Begin Project
+# PROP AllowPerConfigDependencies 0
+# PROP Scc_ProjName ""
+# PROP Scc_LocalPath ""
+CPP=cl.exe
+RSC=rc.exe
+
+!IF  "$(CFG)" == "make_readme - Win32 Release"
+
+# PROP BASE Use_MFC 0
+# PROP BASE Use_Debug_Libraries 0
+# PROP BASE Output_Dir "Release"
+# PROP BASE Intermediate_Dir "Release"
+# PROP BASE Target_Dir ""
+# PROP Use_MFC 0
+# PROP Use_Debug_Libraries 0
+# PROP Output_Dir "Release"
+# PROP Intermediate_Dir "Release"
+# PROP Target_Dir ""
+# ADD BASE CPP /nologo /W3 /GX /O2 /D "WIN32" /D "NDEBUG" /D "_CONSOLE" /D "_MBCS" /YX /FD /c
+# ADD CPP /nologo /W3 /GX /O2 /D "WIN32" /D "NDEBUG" /D "_CONSOLE" /D "_MBCS" /YX /FD /c
+# ADD BASE RSC /l 0x409 /d "NDEBUG"
+# ADD RSC /l 0x409 /d "NDEBUG"
+BSC32=bscmake.exe
+# ADD BASE BSC32 /nologo
+# ADD BSC32 /nologo
+LINK32=link.exe
+# ADD BASE LINK32 kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib /nologo /subsystem:console /machine:I386
+# ADD LINK32 kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib /nologo /subsystem:console /machine:I386
+
+!ELSEIF  "$(CFG)" == "make_readme - Win32 Debug"
+
+# PROP BASE Use_MFC 0
+# PROP BASE Use_Debug_Libraries 1
+# PROP BASE Output_Dir "Debug"
+# PROP BASE Intermediate_Dir "Debug"
+# PROP BASE Target_Dir ""
+# PROP Use_MFC 0
+# PROP Use_Debug_Libraries 1
+# PROP Output_Dir "Debug"
+# PROP Intermediate_Dir "Debug\make_readme"
+# PROP Target_Dir ""
+# ADD BASE CPP /nologo /W3 /Gm /GX /ZI /Od /D "WIN32" /D "_DEBUG" /D "_CONSOLE" /D "_MBCS" /YX /FD /GZ /c
+# ADD CPP /nologo /W3 /Gm /GX /ZI /Od /D "WIN32" /D "_DEBUG" /D "_CONSOLE" /D "_MBCS" /FD /GZ /c
+# SUBTRACT CPP /YX
+# ADD BASE RSC /l 0x409 /d "_DEBUG"
+# ADD RSC /l 0x409 /d "_DEBUG"
+BSC32=bscmake.exe
+# ADD BASE BSC32 /nologo
+# ADD BSC32 /nologo
+LINK32=link.exe
+# ADD BASE LINK32 kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib /nologo /subsystem:console /debug /machine:I386 /pdbtype:sept
+# ADD LINK32 kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib /nologo /subsystem:console /debug /machine:I386 /pdbtype:sept
+
+!ENDIF 
+
+# Begin Target
+
+# Name "make_readme - Win32 Release"
+# Name "make_readme - Win32 Debug"
+# Begin Source File
+
+SOURCE=.\make_readme.c
+# End Source File
+# Begin Source File
+
+SOURCE=.\README.header.md
+# End Source File
+# Begin Source File
+
+SOURCE=.\README.list
+# End Source File
+# End Target
+# End Project

third-party/stb/tools/mr.bat

@@ -0,0 +1 @@
+debug\make_readme

third-party/stb/tools/trailing_whitespace.c

@@ -0,0 +1,32 @@
+#define STB_DEFINE
+#include "stb.h"
+
+int main(int argc, char **argv)
+{
+   int i;
+   for (i=1; i < argc; ++i) {
+      int len;
+      FILE *f;
+      char *s = stb_file(argv[i], &len);
+      char *end, *src, *dest;
+      if (s == NULL) {
+         printf("Couldn't read file '%s'.\n", argv[i]);
+         continue; 
+      }
+      end = s + len;
+      src = dest = s;
+      while (src < end) {
+         char *start=0;
+         while (src < end && *src != '\n' && *src != '\r')
+            *dest++ = *src++;
+         while (dest-1 > s && (dest[-1] == ' ' || dest[-1] == '\t'))
+            --dest;
+         while (src < end && (*src == '\n' || *src == '\r'))
+            *dest++ = *src++;
+      }
+      f = fopen(argv[i], "wb");
+      fwrite(s, 1, dest-s, f);
+      fclose(f);
+   }
+    return 0;
+}

third-party/stb/tools/unicode.c

@@ -0,0 +1,749 @@
+#define STB_DEFINE
+#include "../stb.h"
+
+// create unicode mappings
+//
+// Two kinds of mappings:
+//     map to a number
+//     map to a bit
+//
+// For mapping to a number, we use the following strategy:
+//
+//   User supplies:
+//     1. a table of numbers (for now we use uint16, so full Unicode table is 4MB)
+//     2. a "don't care" value
+//     3. define a 'fallback' value (typically 0)
+//     4. define a fast-path range (typically 0..255 or 0..1023) [@TODO: automate detecting this]
+//
+//   Code:
+//     1. Determine range of *end* of unicode codepoints (U+10FFFF and down) which
+//        all have the same value (or don't care). If large enough, emit this as a
+//        special case in the code.
+//     2. Repeat above, limited to at most U+FFFF.
+//     3. Cluster the data into intervals of 8,16,32,64,128,256 numeric values.
+//        3a. If all the values in an interval are fallback/dont-care, no further processing
+//        3b. Find the "trimmed range" outside which all the values are the fallback or don't care
+//        3c. Find the "special trimmed range" outside which all the values are some constant or don't care
+//     4. Pack the clusters into continuous memory, and find previous instances of
+//        the cluster. Repeat for trimmed & special-trimmed. In the first case, find
+//        previous instances of the cluster (allow don't-care to match in either
+//        direction), both aligned and mis-aligned; in the latter, starting where
+//        things start or mis-aligned. Build an index table specifying the
+//        location of each cluster (and its length). Allow an extra indirection here;
+//        the full-sized index can index a smaller table which has the actual offset
+//        (and lengths).
+//     5. Associate with each packed continuous memory above the amount of memory
+//        required to store the data w/ smallest datatype (of uint8, uint16, uint32).
+//        Discard the continuous memory. Recurse on each index table, but avoid the
+//        smaller packing.
+//
+// For mapping to a bit, we pack the results for 8 characters into a byte, and then apply
+// the above strategy. Note that there may be more optimal approaches with e.g. packing
+// 8 different bits into a single structure, though, which we should explore eventually.
+
+
+// currently we limit *indices* to being 2^16, and we pack them as
+//      index + end_trim*2^16 + start_trim*2^24; specials have to go in a separate table
+typedef uint32 uval;
+#define UVAL_DONT_CARE_DEFAULT 0xffffffff
+
+typedef struct
+{
+   uval *input;
+   uint32 dont_care;
+   uint32 fallback;
+   int  fastpath;
+   int  length;
+   int  depth;
+   int  has_sign;
+   int  splittable;
+   int  replace_fallback_with_codepoint;
+   size_t input_size;
+   size_t inherited_storage;
+} table;
+
+typedef struct
+{
+   int   split_log2;
+   table result; // index into not-returned table
+   int   storage;
+} output;
+
+typedef struct
+{
+   table t;
+   char **output_name;
+} info;
+
+typedef struct
+{
+   size_t path;
+   size_t size;
+} result;
+
+typedef struct
+{
+   uint8 trim_end;
+   uint8 trim_start;
+   uint8 special;
+   uint8 aligned;
+   uint8 indirect;
+
+   uint16 overhead; // add some forced overhead for each mode to avoid getting complex encoding when it doesn't save much
+
+} mode_info;
+
+mode_info modes[] =
+{
+   {   0,0,0,0,0,    32, },
+   {   0,0,0,0,1,   100, },
+   {   0,0,0,1,0,    32, },
+   {   0,0,0,1,1,   100, },
+   {   0,0,1,0,1,   100, },
+   {   0,0,1,1,0,    32, },
+   {   0,0,1,1,1,   200, },
+   {   1,0,0,0,0,   100, },
+   {   1,0,0,0,1,   120, },
+   {   1,1,0,0,0,   100, },
+   {   1,1,0,0,1,   130, },
+   {   1,0,1,0,0,   130, },
+   {   1,0,1,0,1,   180, },
+   {   1,1,1,0,0,   180, },
+   {   1,1,1,0,1,   200, },
+};
+
+#define MODECOUNT  (sizeof(modes)/sizeof(modes[0]))
+#define CLUSTERSIZECOUNT   6    // 8,16, 32,64,  128,256
+
+size_t size_for_max_number(uint32 number)
+{
+   if (number == 0) return 0;
+   if (number < 256) return 1;
+   if (number < 256*256) return 2;
+   if (number < 256*256*256) return 3;
+   return 4;
+}
+
+size_t size_for_max_number_aligned(uint32 number)
+{
+   size_t n = size_for_max_number(number);
+   return n == 3 ? 4 : n;
+}
+
+uval get_data(uval *data, int offset, uval *end)
+{
+   if (data + offset >= end)
+      return 0;
+   else
+      return data[offset];
+}
+
+int safe_len(uval *data, int len, uval *end)
+{
+   if (len > end - data)
+      return end - data;
+   return len;
+}
+
+uval tempdata[256];
+int dirty=0;
+
+size_t find_packed(uval **packed, uval *data, int len, int aligned, int fastpath, uval *end, int offset, int replace)
+{
+   int packlen = stb_arr_len(*packed);
+   int i,p;
+
+   if (data+len > end || replace) {
+      int safelen = safe_len(data, len, end);
+      memset(tempdata, 0, dirty*sizeof(tempdata[0]));
+      memcpy(tempdata, data, safelen * sizeof(data[0]));
+      data = tempdata;
+      dirty = len;
+   }
+   if (replace) {
+      int i;
+      int safelen = safe_len(data, len, end);
+      for (i=0; i < safelen; ++i)
+         if (data[i] == 0)
+            data[i] = offset+i;
+   }
+
+   if (len <= 0)
+      return 0;
+   if (!fastpath) {
+      if (aligned) {
+         for (i=0; i < packlen; i += len)
+            if ((*packed)[i] == data[0] && 0==memcmp(&(*packed)[i], data, len * sizeof(uval)))
+               return i / len;
+      } else {
+         for (i=0; i < packlen-len+1; i +=  1 )
+            if ((*packed)[i] == data[0] && 0==memcmp(&(*packed)[i], data, len * sizeof(uval)))
+               return i;
+      }
+   }
+   p = stb_arr_len(*packed);
+   for (i=0; i < len; ++i)
+      stb_arr_push(*packed, data[i]);
+   return p;
+}
+
+void output_table(char *name1, char *name2, uval *data, int length, int sign, char **names)
+{
+   char temp[20];
+   uval maxv = 0;
+   int bytes, numlen, at_newline;
+   int linelen = 79; // @TODO: make table more readable by choosing a length that's a multiple?
+   int i,pos, do_split=0;
+   for (i=0; i < length; ++i)
+      if (sign)
+         maxv = stb_max(maxv, (uval)abs((int)data[i]));
+      else
+         maxv = stb_max(maxv, data[i]);
+   bytes = size_for_max_number_aligned(maxv);
+   sprintf(temp, "%d", maxv);
+   numlen=strlen(temp);
+   if (sign)
+      ++numlen;
+   
+   if (bytes == 0)
+      return;
+
+   printf("uint%d %s%s[%d] = {\n", bytes*8, name1, name2, length);
+   at_newline = 1;
+   for (i=0; i < length; ++i) {
+      if (pos + numlen + 2 > linelen) {
+         printf("\n");
+         at_newline = 1;
+         pos = 0;
+      }
+      if (at_newline) {
+         printf("  ");
+         pos = 2;
+         at_newline = 0;
+      } else {
+         printf(" ");
+         ++pos;
+      }
+      printf("%*d,", numlen, data[i]);
+      pos += numlen+1;
+   }
+   if (!at_newline) printf("\n");
+   printf("};\n");
+}
+
+void output_table_with_trims(char *name1, char *name2, uval *data, int length)
+{
+   uval maxt=0, maxp=0;
+   int i,d,s,e, count;
+   // split the table into two pieces
+   uval *trims = NULL;
+
+   if (length == 0)
+      return;
+
+   for (i=0; i < stb_arr_len(data); ++i) {
+      stb_arr_push(trims, data[i] >> 16);
+      data[i] &= 0xffff;
+      maxt = stb_max(maxt, trims[i]);
+      maxp = stb_max(maxp, data[i]);
+   }
+
+   d=s=e=1;
+   if (maxt >= 256) {
+      // need to output start & end values
+      if (maxp >= 256) {
+         // can pack into a single table
+         printf("struct { uint16 val; uint8 start, end; } %s%s[%d] = {\n", name1, name2, length);
+      } else {
+         output_table(name1, name2, data, length, 0, 0);
+         d=0;
+         printf("struct { uint8 start, end; } %s%s_trim[%d] = {\n", name1, name2, length);
+      }
+   } else if (maxt > 0) {
+      if (maxp >= 256) {
+         output_table(name1, name2, data, length, 0, 0);
+         output_table(name1, stb_sprintf("%s_end", name2), trims, length, 0, 0);
+         return;
+      } else {
+         printf("struct { uint8 val, end; } %s%s[%d] = {\n", name1, name2, length);
+         s=0;
+      }
+   } else {
+      output_table(name1, name2, data, length, 0, 0);
+      return;
+   }
+   // d or s can be zero (but not both), e is always present and last
+   count = d + s + e;
+   assert(count >= 2 && count <= 3);
+
+   {
+      char temp[60];
+      uval maxv = 0;
+      int numlen, at_newline, len;
+      int linelen = 79; // @TODO: make table more readable by choosing a length that's a multiple?
+      int i,pos, do_split=0;
+      numlen = 0;
+      for (i=0; i < length; ++i) {
+         if (count == 2)
+            sprintf(temp, "{%d,%d}", d ? data[i] : (trims[i]>>8), trims[i]&255);
+         else
+            sprintf(temp, "{%d,%d,%d}", data[i], trims[i]>>8, trims[i]&255);
+         len = strlen(temp);
+         numlen = stb_max(len, numlen);
+      }
+   
+      at_newline = 1;
+      for (i=0; i < length; ++i) {
+         if (pos + numlen + 2 > linelen) {
+            printf("\n");
+            at_newline = 1;
+            pos = 0;
+         }
+         if (at_newline) {
+            printf("  ");
+            pos = 2;
+            at_newline = 0;
+         } else {
+            printf(" ");
+            ++pos;
+         }
+         if (count == 2)
+            sprintf(temp, "{%d,%d}", d ? data[i] : (trims[i]>>8), trims[i]&255);
+         else
+            sprintf(temp, "{%d,%d,%d}", data[i], trims[i]>>8, trims[i]&255);
+         printf("%*s,", numlen, temp);
+         pos += numlen+1;
+      }
+      if (!at_newline) printf("\n");
+      printf("};\n");
+   }
+}
+
+int weight=1;
+
+table pack_for_mode(table *t, int mode, char *table_name)
+{
+   size_t extra_size;
+   int i;
+   uval maxv;
+   mode_info mi = modes[mode % MODECOUNT];
+   int size = 8 << (mode / MODECOUNT);
+   table newtab;
+   uval *packed = NULL;
+   uval *index = NULL;
+   uval *indirect = NULL;
+   uval *specials = NULL;
+   newtab.dont_care = UVAL_DONT_CARE_DEFAULT;
+   if (table_name)
+      printf("// clusters of %d\n", size);
+   for (i=0; i < t->length; i += size) {
+      uval newval;
+      int fastpath = (i < t->fastpath);
+      if (mi.special) {
+         int end_trim = size-1;
+         int start_trim = 0;
+         uval special;
+         // @TODO: pick special from start or end instead of only end depending on which is longer
+         for(;;) {
+            special = t->input[i + end_trim];
+            if (special != t->dont_care || end_trim == 0)
+               break;
+            --end_trim;
+         }
+         // at this point, special==inp[end_trim], and end_trim >= 0
+         if (special == t->dont_care && !fastpath) {
+            // entire block is don't care, so OUTPUT don't care
+            stb_arr_push(index, newtab.dont_care);
+            continue;
+         } else {
+            uval pos, trim;
+            if (mi.trim_end && !fastpath) {
+               while (end_trim >= 0) {
+                  if (t->input[i + end_trim] == special || t->input[i + end_trim] == t->dont_care)
+                     --end_trim;
+                  else
+                     break;
+               }
+            }
+
+            if (mi.trim_start && !fastpath) {
+               while (start_trim < end_trim) {
+                  if (t->input[i + start_trim] == special || t->input[i + start_trim] == t->dont_care)
+                     ++start_trim;
+                  else
+                     break;
+               }
+            }
+
+            // end_trim points to the last character we have to output
+
+            // find the first match, or add it
+            pos = find_packed(&packed, &t->input[i+start_trim], end_trim-start_trim+1, mi.aligned, fastpath, &t->input[t->length], i+start_trim, t->replace_fallback_with_codepoint);
+
+            // encode as a uval
+            if (!mi.trim_end) {
+               if (end_trim == 0)
+                  pos = special;
+               else
+                  pos = pos | 0x80000000;
+            } else {
+               assert(end_trim < size && end_trim >= -1);
+               if (!fastpath) assert(end_trim < size-1); // special always matches last one
+               assert(end_trim < size && end_trim+1 >= 0);
+               if (!fastpath) assert(end_trim+1 < size);
+
+               if (mi.trim_start)
+                  trim = start_trim*256 + (end_trim+1);
+               else
+                  trim = end_trim+1;
+
+               assert(pos < 65536); // @TODO: if this triggers, just bail on this search path
+               pos = pos + (trim << 16);
+            }
+
+            newval = pos;
+
+            stb_arr_push(specials, special);
+         }
+      } else if (mi.trim_end) {
+         int end_trim = size-1;
+         int start_trim = 0;
+         uval pos, trim;
+
+         while (end_trim >= 0 && !fastpath)
+            if (t->input[i + end_trim] == t->fallback || t->input[i + end_trim] == t->dont_care)
+               --end_trim;
+            else
+               break;
+
+         if (mi.trim_start && !fastpath) {
+            while (start_trim < end_trim) {
+               if (t->input[i + start_trim] == t->fallback || t->input[i + start_trim] == t->dont_care)
+                  ++start_trim;
+               else
+                  break;
+            }
+         }
+
+         // end_trim points to the last character we have to output, and can be -1
+         ++end_trim; // make exclusive at end
+
+         if (end_trim == 0 && size == 256)
+            start_trim = end_trim = 1;  // we can't make encode a length from 0..256 in 8 bits, so restrict end_trim to 1..256
+
+         // find the first match, or add it
+         pos = find_packed(&packed, &t->input[i+start_trim], end_trim - start_trim, mi.aligned, fastpath, &t->input[t->length], i+start_trim, t->replace_fallback_with_codepoint);
+
+         assert(end_trim <= size && end_trim >= 0);
+         if (size == 256)
+            assert(end_trim-1 < 256 && end_trim-1 >= 0);
+         else
+            assert(end_trim < 256 && end_trim >= 0);
+         if (size == 256)
+            --end_trim;
+
+         if (mi.trim_start)
+            trim = start_trim*256 + end_trim;
+         else
+            trim = end_trim;
+
+         assert(pos < 65536); // @TODO: if this triggers, just bail on this search path
+         pos = pos + (trim << 16);
+
+         newval = pos;
+      } else {
+         newval = find_packed(&packed, &t->input[i], size, mi.aligned, fastpath, &t->input[t->length], i, t->replace_fallback_with_codepoint);
+      }
+
+      if (mi.indirect) {
+         int j;
+         for (j=0; j < stb_arr_len(indirect); ++j)
+            if (indirect[j] == newval)
+               break;
+         if (j == stb_arr_len(indirect))
+            stb_arr_push(indirect, newval);
+         stb_arr_push(index, j);
+      } else {
+         stb_arr_push(index, newval);
+      }
+   }
+
+   // total up the new size for everything but the index table
+   extra_size = mi.overhead * weight; // not the actual overhead cost; a penalty to avoid excessive complexity
+   extra_size += 150; // per indirection
+   if (table_name)
+      extra_size = 0;
+   
+   if (t->has_sign) {
+      // 'packed' contains two values, which should be packed positive & negative for size
+      uval maxv2;
+      for (i=0; i < stb_arr_len(packed); ++i)
+         if (packed[i] & 0x80000000)
+            maxv2 = stb_max(maxv2, packed[i]);
+         else
+            maxv  = stb_max(maxv, packed[i]);
+      maxv = stb_max(maxv, maxv2) << 1;
+   } else {
+      maxv = 0;
+      for (i=0; i < stb_arr_len(packed); ++i)
+         if (packed[i] > maxv && packed[i] != t->dont_care)
+            maxv = packed[i];
+   }
+   extra_size += stb_arr_len(packed) * (t->splittable ? size_for_max_number(maxv) : size_for_max_number_aligned(maxv));
+   if (table_name) {
+      if (t->splittable)
+         output_table_with_trims(table_name, "", packed, stb_arr_len(packed));
+      else
+         output_table(table_name, "", packed, stb_arr_len(packed), t->has_sign, NULL);
+   }
+
+   maxv = 0;
+   for (i=0; i < stb_arr_len(specials); ++i)
+      if (specials[i] > maxv)
+         maxv = specials[i];
+   extra_size += stb_arr_len(specials) * size_for_max_number_aligned(maxv);
+   if (table_name)
+      output_table(table_name, "_default", specials, stb_arr_len(specials), 0, NULL);
+
+   maxv = 0;
+   for (i=0; i < stb_arr_len(indirect); ++i)
+      if (indirect[i] > maxv)
+         maxv = indirect[i];
+   extra_size += stb_arr_len(indirect) * size_for_max_number(maxv);
+
+   if (table_name && stb_arr_len(indirect)) {
+      if (mi.trim_end)
+         output_table_with_trims(table_name, "_index", indirect, stb_arr_len(indirect));
+      else {
+         assert(0); // this case should only trigger in very extreme circumstances
+         output_table(table_name, "_index", indirect, stb_arr_len(indirect), 0, NULL);
+      }
+      mi.trim_end = mi.special = 0;
+   }
+
+   if (table_name)
+      printf("// above tables should be %d bytes\n", extra_size);
+
+   maxv = 0;
+   for (i=0; i < stb_arr_len(index); ++i)
+      if (index[i] > maxv && index[i] != t->dont_care)
+         maxv = index[i];
+   newtab.splittable = mi.trim_end;
+   newtab.input_size = newtab.splittable ? size_for_max_number(maxv) : size_for_max_number_aligned(maxv);
+   newtab.input = index;
+   newtab.length = stb_arr_len(index);
+   newtab.inherited_storage = t->inherited_storage + extra_size;
+   newtab.fastpath = 0;
+   newtab.depth = t->depth+1;
+   stb_arr_free(indirect);
+   stb_arr_free(packed);
+   stb_arr_free(specials);
+
+   return newtab;
+}
+
+result pack_table(table *t, size_t path, int min_storage)
+{
+   int i;
+   result best;
+   best.size = t->inherited_storage + t->input_size * t->length;
+   best.path = path;
+
+   if ((int) t->inherited_storage > min_storage) {
+      best.size = stb_max(best.size, t->inherited_storage);
+      return best;
+   }
+
+   if (t->length <= 256 || t->depth >= 4) {
+      //printf("%08x: %7d\n", best.path, best.size);
+      return best;
+   }
+
+   path <<= 7;
+   for (i=0; i < MODECOUNT * CLUSTERSIZECOUNT; ++i) {
+      table newtab;
+      result r;
+      newtab = pack_for_mode(t, i, 0);
+      r = pack_table(&newtab, path+i+1, min_storage);
+      if (r.size < best.size)
+         best = r;
+      stb_arr_free(newtab.input);
+      //printf("Size: %6d + %6d\n", newtab.inherited_storage, newtab.input_size * newtab.length);
+   }
+   return best;
+}
+
+int pack_table_by_modes(table *t, int *modes)
+{
+   table s = *t;
+   while (*modes > -1) {
+      table newtab;
+      newtab = pack_for_mode(&s, *modes, 0);
+      if (s.input != t->input)
+         stb_arr_free(s.input);
+      s = newtab;
+      ++modes;
+   }
+   return s.inherited_storage + s.input_size * s.length;
+}
+
+int strip_table(table *t, int exceptions)
+{
+   uval terminal_value;
+   int p = t->length-1;
+   while (t->input[p] == t->dont_care)
+      --p;
+   terminal_value = t->input[p];
+
+   while (p >= 0x10000) {
+      if (t->input[p] != terminal_value && t->input[p] != t->dont_care) {
+         if (exceptions)
+            --exceptions;
+         else
+            break;
+      }
+      --p;
+   }
+   return p+1; // p is a character we must output
+}
+
+void optimize_table(table *t, char *table_name)
+{
+   int modelist[3] = { 85, -1 };
+   int modes[8];
+   int num_modes = 0;
+   int decent_size;
+   result r;
+   size_t path;
+   table s;
+
+   // strip tail end of table
+   int orig_length = t->length;
+   int threshhold = 0xffff;
+   int p = strip_table(t, 2);
+   int len_saved = t->length - p;
+   if (len_saved >= threshhold) {
+      t->length = p;
+      while (p > 0x10000) {
+         p = strip_table(t, 0);
+         len_saved = t->length - p;
+         if (len_saved < 0x10000)
+            break;
+         len_saved = orig_length - p;
+         if (len_saved < threshhold)
+            break;
+         threshhold *= 2;
+      }
+   }
+
+   t->depth = 1;
+
+
+   // find size of table if we use path 86
+   decent_size = pack_table_by_modes(t, modelist);
+
+
+   #if 1
+   // find best packing of remainder of table by exploring tree of packings
+   r = pack_table(t, 0, decent_size);
+   // use the computed 'path' to evaluate and output tree
+   path = r.path;
+   #else
+   path = 86;//90;//132097;
+   #endif
+
+   while (path) {
+      modes[num_modes++] = (path & 127) - 1;
+      path >>= 7;
+   }
+
+   printf("// modes: %d\n", r.path);
+   s = *t;
+   while (num_modes > 0) {
+      char name[256];
+      sprintf(name, "%s_%d", table_name, num_modes+1);
+      --num_modes;
+      s = pack_for_mode(&s, modes[num_modes], name);
+   }
+   // output the final table as-is
+   if (s.splittable)
+      output_table_with_trims(table_name, "_1", s.input, s.length);
+   else
+      output_table(table_name, "_1", s.input, s.length, 0, NULL);
+}
+
+uval unicode_table[0x110000];
+
+typedef struct
+{
+   uval lo,hi;
+} char_range;
+
+char_range get_range(char *str)
+{
+   char_range cr;
+   char *p;
+   cr.lo = strtol(str, &p, 16);
+   p = stb_skipwhite(p);
+   if (*p == '.')
+      cr.hi = strtol(p+2, NULL, 16);
+   else
+      cr.hi = cr.lo;
+   return cr;
+}
+
+char *skip_semi(char *s, int count)
+{
+   while (count) {
+      s = strchr(s, ';');
+      assert(s != NULL);
+      ++s;
+      --count;
+   }
+   return s;
+}
+
+int main(int argc, char **argv)
+{
+   table t;
+   uval maxv=0;
+   int i,n=0;
+   char **s = stb_stringfile("../../data/UnicodeData.txt", &n);
+   assert(s);
+   for (i=0; i < n; ++i) {
+      if (s[i][0] == '#' || s[i][0] == '\n' || s[i][0] == 0)
+         ;
+      else {
+         char_range cr = get_range(s[i]);
+         char *t = skip_semi(s[i], 13);
+         uval j, v;
+         if (*t == ';' || *t == '\n' || *t == 0)
+            v = 0;
+         else {
+            v = strtol(t, NULL, 16);
+            if (v < 65536) {
+               maxv = stb_max(v, maxv);
+               for (j=cr.lo; j <= cr.hi; ++j) {
+                  unicode_table[j] = v;
+                  //printf("%06x => %06x\n", j, v);
+               }
+            }
+         }
+      }
+   }
+
+   t.depth = 0;
+   t.dont_care = UVAL_DONT_CARE_DEFAULT;
+   t.fallback = 0;
+   t.fastpath = 256;
+   t.inherited_storage = 0;
+   t.has_sign = 0;
+   t.splittable = 0;
+   t.input = unicode_table;
+   t.input_size = size_for_max_number(maxv);
+   t.length = 0x110000;
+   t.replace_fallback_with_codepoint = 1;
+
+   optimize_table(&t, "stbu_upppercase");
+   return 0;
+}

third-party/stb/tools/unicode/unicode.dsp

@@ -0,0 +1,88 @@
+# Microsoft Developer Studio Project File - Name="unicode" - Package Owner=<4>
+# Microsoft Developer Studio Generated Build File, Format Version 6.00
+# ** DO NOT EDIT **
+
+# TARGTYPE "Win32 (x86) Console Application" 0x0103
+
+CFG=unicode - Win32 Debug
+!MESSAGE This is not a valid makefile. To build this project using NMAKE,
+!MESSAGE use the Export Makefile command and run
+!MESSAGE 
+!MESSAGE NMAKE /f "unicode.mak".
+!MESSAGE 
+!MESSAGE You can specify a configuration when running NMAKE
+!MESSAGE by defining the macro CFG on the command line. For example:
+!MESSAGE 
+!MESSAGE NMAKE /f "unicode.mak" CFG="unicode - Win32 Debug"
+!MESSAGE 
+!MESSAGE Possible choices for configuration are:
+!MESSAGE 
+!MESSAGE "unicode - Win32 Release" (based on "Win32 (x86) Console Application")
+!MESSAGE "unicode - Win32 Debug" (based on "Win32 (x86) Console Application")
+!MESSAGE 
+
+# Begin Project
+# PROP AllowPerConfigDependencies 0
+# PROP Scc_ProjName ""
+# PROP Scc_LocalPath ""
+CPP=cl.exe
+RSC=rc.exe
+
+!IF  "$(CFG)" == "unicode - Win32 Release"
+
+# PROP BASE Use_MFC 0
+# PROP BASE Use_Debug_Libraries 0
+# PROP BASE Output_Dir "Release"
+# PROP BASE Intermediate_Dir "Release"
+# PROP BASE Target_Dir ""
+# PROP Use_MFC 0
+# PROP Use_Debug_Libraries 0
+# PROP Output_Dir "Release"
+# PROP Intermediate_Dir "Release"
+# PROP Target_Dir ""
+# ADD BASE CPP /nologo /W3 /GX /O2 /D "WIN32" /D "NDEBUG" /D "_CONSOLE" /D "_MBCS" /YX /FD /c
+# ADD CPP /nologo /W3 /GX /O2 /D "WIN32" /D "NDEBUG" /D "_CONSOLE" /D "_MBCS" /YX /FD /c
+# ADD BASE RSC /l 0x409 /d "NDEBUG"
+# ADD RSC /l 0x409 /d "NDEBUG"
+BSC32=bscmake.exe
+# ADD BASE BSC32 /nologo
+# ADD BSC32 /nologo
+LINK32=link.exe
+# ADD BASE LINK32 kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib  kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib /nologo /subsystem:console /machine:I386
+# ADD LINK32 kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib  kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib /nologo /subsystem:console /machine:I386
+
+!ELSEIF  "$(CFG)" == "unicode - Win32 Debug"
+
+# PROP BASE Use_MFC 0
+# PROP BASE Use_Debug_Libraries 1
+# PROP BASE Output_Dir "Debug"
+# PROP BASE Intermediate_Dir "Debug"
+# PROP BASE Target_Dir ""
+# PROP Use_MFC 0
+# PROP Use_Debug_Libraries 1
+# PROP Output_Dir "Debug"
+# PROP Intermediate_Dir "Debug"
+# PROP Target_Dir ""
+# ADD BASE CPP /nologo /W3 /Gm /GX /ZI /Od /D "WIN32" /D "_DEBUG" /D "_CONSOLE" /D "_MBCS" /YX /FD /GZ  /c
+# ADD CPP /nologo /W3 /Gm /GX /ZI /Od /D "WIN32" /D "_DEBUG" /D "_CONSOLE" /D "_MBCS" /YX /FD /GZ  /c
+# ADD BASE RSC /l 0x409 /d "_DEBUG"
+# ADD RSC /l 0x409 /d "_DEBUG"
+BSC32=bscmake.exe
+# ADD BASE BSC32 /nologo
+# ADD BSC32 /nologo
+LINK32=link.exe
+# ADD BASE LINK32 kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib  kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib /nologo /subsystem:console /debug /machine:I386 /pdbtype:sept
+# ADD LINK32 kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib  kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib /nologo /subsystem:console /debug /machine:I386 /pdbtype:sept
+
+!ENDIF 
+
+# Begin Target
+
+# Name "unicode - Win32 Release"
+# Name "unicode - Win32 Debug"
+# Begin Source File
+
+SOURCE=..\unicode.c
+# End Source File
+# End Target
+# End Project