Applied multiple character set support diff by Rob Goodwin

@Multiple character set support in gdttf.c (Rob Goodwin)

diff --git a/ext/gd/gdttf.c b/ext/gd/gdttf.c
index 13587ffd88b2e92401e6673901536267a2cd5e76..3db6b806e5bb5b968bd7f60559d448ad9b415029 100644 (file)
--- a/ext/gd/gdttf.c
+++ b/ext/gd/gdttf.c
@@ -80,7 +80,12 @@ typedef struct {
        TT_Face                         face;
        TT_Face_Properties  properties;
        TT_Instance                     instance;
-       TT_CharMap                      char_map;
+       TT_CharMap                      char_map_Unicode;
+       TT_CharMap                      char_map_Big5;
+       TT_CharMap                      char_map_Roman;
+       int                                     have_char_map_Unicode;
+       int                                     have_char_map_Big5;
+       int                                     have_char_map_Roman;
        TT_Matrix                       matrix;
        TT_Instance_Metrics     imetrics;
        gdCache_head_t          *glyphCache;
@@ -153,6 +158,8 @@ static void bitmapRelease( void *element );
 /* local prototype */
 char *gdttfchar(gdImage *im, int fg, font_t *font, int x, int y, TT_F26Dot6 x1,  TT_F26Dot6 y1, TT_F26Dot6 *advance, TT_BBox **bbox, char **next);
 
+
+
 /********************************************************************
  * gdTcl_UtfToUniChar is borrowed from ...
  */
@@ -328,7 +335,7 @@ fontFetch ( char **error, void *key )
        TT_Error                err;
        font_t                  *a;
        fontkey_t               *b=(fontkey_t *)key;
-       int                             i, n;
+       int                             i, n, map_found;
        short                   platform, encoding;
 
        a = (font_t *)malloc(sizeof(font_t));
@@ -372,6 +379,10 @@ fontFetch ( char **error, void *key )
        if (TT_Set_Instance_Resolutions(a->instance, RESOLUTION, RESOLUTION)) {
                *error = "Could not set device resolutions";
                return NULL;
+map_found = 0;
+a->have_char_map_Unicode = 0;
+a->have_char_map_Big5 = 0;
+a->have_char_map_Roman = 0;
        }
 
        if (TT_Set_Instance_CharSize(a->instance, (TT_F26Dot6)(a->ptsize*64))) {
@@ -385,15 +396,28 @@ fontFetch ( char **error, void *key )
        n = TT_Get_CharMap_Count(a->face);
 
        for (i = 0; i < n; i++) {
-               TT_Get_CharMap_ID(a->face, i, &platform, &encoding);
-               if ((platform == 3 && encoding == 1)  ||
-                   (platform == 2 && encoding == 1)  ||
-                   (platform == 0)) {
-                       TT_Get_CharMap(a->face, i, &a->char_map);
-                       i = n+1;
-               }
+        TT_Get_CharMap_ID(a->face, i, &platform, &encoding);
+               if ((platform == 3 && encoding == 1)           /* Windows Unicode */
+                       || (platform == 0 && encoding == 0)) {        /* ?? Unicode */
+                       TT_Get_CharMap(a->face, i, &a->char_map_Unicode);
+                       a->have_char_map_Unicode = 1;
+                       map_found++;
+               } else if (platform == 3 && encoding == 4) {   /* Windows Big5 */
+                       TT_Get_CharMap(a->face, i, &a->char_map_Big5);
+                       a->have_char_map_Big5 = 1;
+                       map_found++;
+               } else if (platform == 1 && encoding == 0) {   /* Apple Roman */
+                       TT_Get_CharMap(a->face, i, &a->char_map_Roman);
+                       a->have_char_map_Roman = 1;
+                       map_found++;
+        }
        }
 
+       if (! map_found) {
+               *error = "Unable to find a CharMap that I can handle";
+        return NULL;
+    }
+
        if (i == n) {
                *error = "Sorry, but this font doesn't contain any Unicode mapping table";
                return NULL;
@@ -462,7 +486,13 @@ glyphFetch ( char **error, void *key )
        if (a->hinting && b->font->angle == 0.0) {
                flags |= TTLOAD_HINT_GLYPH;
        }
-       glyph_code = TT_Char_Index(b->font->char_map, a->character);
+       if (b->font->have_char_map_Unicode) {
+               glyph_code = TT_Char_Index(b->font->char_map_Unicode, a->character);
+       } else if (a->character < 161 && b->font->have_char_map_Roman) {
+               glyph_code = TT_Char_Index(b->font->char_map_Roman, a->character);
+       } else if ( b->font->have_char_map_Big5) {
+               glyph_code = TT_Char_Index(b->font->char_map_Big5, a->character);
+       }
        if ((err=TT_Load_Glyph(b->font->instance, a->glyph, glyph_code, flags))) {
                *error = "TT_Load_Glyph problem";
                return NULL;
@@ -657,8 +687,23 @@ gdttfchar(gdImage *im, int fg, font_t *font,
        }
        /**************/
 
-       len = gdTcl_UtfToUniChar(*next, &ch);
-       *next += len;
+       if (font->have_char_map_Unicode) { /* use UTF-8 mapping from ASCII */
+        len = gdTcl_UtfToUniChar(*next, &ch);
+        *next += len;
+       } else {
+       /*
+        * Big 5 mapping:
+        * use "JIS-8 half-width katakana" coding from 8-bit characters.  Ref:
+        * ftp://ftp.ora.com/pub/examples/nutshell/ujip/doc/japan.inf-032092.sjs
+        */
+               ch = (**next) & 255;         /* don't extend sign */
+               (*next)++;
+               if (ch >= 161                /* first code of JIS-8 pair */
+                       && **next) {                /* don't advance past '\0' */
+                       ch = (ch * 256) + **next;
+                       (*next)++;
+               }
+       }
 
        glyphkey.character = ch;
        glyphkey.hinting = 1;