| 1 | #include "mupdf/fitz.h" |
| 2 | #include "mupdf/pdf.h" |
| 3 | |
| 4 | #include "../fitz/fitz-imp.h" |
| 5 | |
| 6 | #include <ft2build.h> |
| 7 | #include FT_FREETYPE_H |
| 8 | #ifdef FT_FONT_FORMATS_H |
| 9 | #include FT_FONT_FORMATS_H |
| 10 | #else |
| 11 | #include FT_XFREE86_H |
| 12 | #endif |
| 13 | #include FT_TRUETYPE_TABLES_H |
| 14 | |
| 15 | #ifndef FT_SFNT_HEAD |
| 16 | #define FT_SFNT_HEAD ft_sfnt_head |
| 17 | #endif |
| 18 | |
| 19 | static int ft_font_file_kind(FT_Face face) |
| 20 | { |
| 21 | #ifdef FT_FONT_FORMATS_H |
| 22 | const char *kind = FT_Get_Font_Format(face); |
| 23 | #else |
| 24 | const char *kind = FT_Get_X11_Font_Format(face); |
| 25 | #endif |
| 26 | if (!strcmp(kind, "TrueType" )) return 2; |
| 27 | if (!strcmp(kind, "Type 1" )) return 1; |
| 28 | if (!strcmp(kind, "CFF" )) return 3; |
| 29 | if (!strcmp(kind, "CID Type 1" )) return 1; |
| 30 | return 0; |
| 31 | } |
| 32 | |
| 33 | static int is_ttc(fz_font *font) |
| 34 | { |
| 35 | return !memcmp(font->buffer->data, "ttcf" , 4); |
| 36 | } |
| 37 | |
| 38 | static int is_truetype(FT_Face face) |
| 39 | { |
| 40 | return ft_font_file_kind(face) == 2; |
| 41 | } |
| 42 | |
| 43 | static int is_postscript(FT_Face face) |
| 44 | { |
| 45 | int kind = ft_font_file_kind(face); |
| 46 | return (kind == 1 || kind == 3); |
| 47 | } |
| 48 | |
| 49 | static int is_builtin_font(fz_context *ctx, fz_font *font) |
| 50 | { |
| 51 | int size; |
| 52 | unsigned char *data; |
| 53 | if (!font->buffer) |
| 54 | return 0; |
| 55 | fz_buffer_storage(ctx, font->buffer, &data); |
| 56 | return fz_lookup_base14_font(ctx, pdf_clean_font_name(font->name), &size) == data; |
| 57 | } |
| 58 | |
| 59 | static pdf_obj* |
| 60 | pdf_add_font_file(fz_context *ctx, pdf_document *doc, fz_font *font) |
| 61 | { |
| 62 | fz_buffer *buf = font->buffer; |
| 63 | pdf_obj *obj = NULL; |
| 64 | pdf_obj *ref = NULL; |
| 65 | |
| 66 | fz_var(obj); |
| 67 | fz_var(ref); |
| 68 | |
| 69 | /* Check for substitute fonts */ |
| 70 | if (font->flags.ft_substitute) |
| 71 | return NULL; |
| 72 | |
| 73 | fz_try(ctx) |
| 74 | { |
| 75 | size_t len = fz_buffer_storage(ctx, buf, NULL); |
| 76 | obj = pdf_new_dict(ctx, doc, 3); |
| 77 | pdf_dict_put_int(ctx, obj, PDF_NAME(Length1), (int)len); |
| 78 | switch (ft_font_file_kind(font->ft_face)) |
| 79 | { |
| 80 | case 1: |
| 81 | /* TODO: these may not be the correct values, but I doubt it matters */ |
| 82 | pdf_dict_put_int(ctx, obj, PDF_NAME(Length2), len); |
| 83 | pdf_dict_put_int(ctx, obj, PDF_NAME(Length3), 0); |
| 84 | break; |
| 85 | case 2: |
| 86 | break; |
| 87 | case 3: |
| 88 | if (FT_Get_Sfnt_Table(font->ft_face, FT_SFNT_HEAD)) |
| 89 | pdf_dict_put(ctx, obj, PDF_NAME(Subtype), PDF_NAME(OpenType)); |
| 90 | else |
| 91 | pdf_dict_put(ctx, obj, PDF_NAME(Subtype), PDF_NAME(Type1C)); |
| 92 | break; |
| 93 | } |
| 94 | ref = pdf_add_object(ctx, doc, obj); |
| 95 | pdf_update_stream(ctx, doc, ref, buf, 0); |
| 96 | } |
| 97 | fz_always(ctx) |
| 98 | { |
| 99 | pdf_drop_obj(ctx, obj); |
| 100 | } |
| 101 | fz_catch(ctx) |
| 102 | { |
| 103 | pdf_drop_obj(ctx, ref); |
| 104 | fz_rethrow(ctx); |
| 105 | } |
| 106 | return ref; |
| 107 | } |
| 108 | |
| 109 | static void |
| 110 | pdf_add_font_descriptor(fz_context *ctx, pdf_document *doc, pdf_obj *fobj, fz_font *font) |
| 111 | { |
| 112 | FT_Face face = font->ft_face; |
| 113 | pdf_obj *fdobj = NULL; |
| 114 | pdf_obj *fileref; |
| 115 | fz_rect bbox; |
| 116 | |
| 117 | fdobj = pdf_new_dict(ctx, doc, 10); |
| 118 | fz_try(ctx) |
| 119 | { |
| 120 | pdf_dict_put(ctx, fdobj, PDF_NAME(Type), PDF_NAME(FontDescriptor)); |
| 121 | pdf_dict_put_name(ctx, fdobj, PDF_NAME(FontName), font->name); |
| 122 | |
| 123 | bbox.x0 = font->bbox.x0 * 1000; |
| 124 | bbox.y0 = font->bbox.y0 * 1000; |
| 125 | bbox.x1 = font->bbox.x1 * 1000; |
| 126 | bbox.y1 = font->bbox.y1 * 1000; |
| 127 | pdf_dict_put_rect(ctx, fdobj, PDF_NAME(FontBBox), bbox); |
| 128 | |
| 129 | pdf_dict_put_int(ctx, fdobj, PDF_NAME(ItalicAngle), 0); |
| 130 | pdf_dict_put_int(ctx, fdobj, PDF_NAME(Ascent), face->ascender * 1000.0f / face->units_per_EM); |
| 131 | pdf_dict_put_int(ctx, fdobj, PDF_NAME(Descent), face->descender * 1000.0f / face->units_per_EM); |
| 132 | pdf_dict_put_int(ctx, fdobj, PDF_NAME(StemV), 80); |
| 133 | pdf_dict_put_int(ctx, fdobj, PDF_NAME(Flags), PDF_FD_NONSYMBOLIC); |
| 134 | |
| 135 | fileref = pdf_add_font_file(ctx, doc, font); |
| 136 | if (fileref) |
| 137 | { |
| 138 | switch (ft_font_file_kind(face)) |
| 139 | { |
| 140 | default: |
| 141 | case 1: pdf_dict_put_drop(ctx, fdobj, PDF_NAME(FontFile), fileref); break; |
| 142 | case 2: pdf_dict_put_drop(ctx, fdobj, PDF_NAME(FontFile2), fileref); break; |
| 143 | case 3: pdf_dict_put_drop(ctx, fdobj, PDF_NAME(FontFile3), fileref); break; |
| 144 | } |
| 145 | } |
| 146 | |
| 147 | pdf_dict_put_drop(ctx, fobj, PDF_NAME(FontDescriptor), pdf_add_object(ctx, doc, fdobj)); |
| 148 | } |
| 149 | fz_always(ctx) |
| 150 | pdf_drop_obj(ctx, fdobj); |
| 151 | fz_catch(ctx) |
| 152 | fz_rethrow(ctx); |
| 153 | } |
| 154 | |
| 155 | static void |
| 156 | pdf_add_simple_font_widths(fz_context *ctx, pdf_document *doc, pdf_obj *fobj, fz_font *font, const char * const encoding[]) |
| 157 | { |
| 158 | int width_table[256]; |
| 159 | pdf_obj *widths; |
| 160 | int i, first, last; |
| 161 | |
| 162 | first = 0; |
| 163 | last = 0; |
| 164 | |
| 165 | for (i = 0; i < 256; ++i) |
| 166 | { |
| 167 | int glyph = 0; |
| 168 | if (encoding[i]) |
| 169 | { |
| 170 | glyph = fz_encode_character_by_glyph_name(ctx, font, encoding[i]); |
| 171 | } |
| 172 | if (glyph > 0) |
| 173 | { |
| 174 | if (!first) |
| 175 | first = i; |
| 176 | last = i; |
| 177 | width_table[i] = fz_advance_glyph(ctx, font, glyph, 0) * 1000; |
| 178 | } |
| 179 | else |
| 180 | width_table[i] = 0; |
| 181 | } |
| 182 | |
| 183 | widths = pdf_new_array(ctx, doc, last - first + 1); |
| 184 | pdf_dict_put_drop(ctx, fobj, PDF_NAME(Widths), widths); |
| 185 | for (i = first; i <= last; ++i) |
| 186 | pdf_array_push_int(ctx, widths, width_table[i]); |
| 187 | pdf_dict_put_int(ctx, fobj, PDF_NAME(FirstChar), first); |
| 188 | pdf_dict_put_int(ctx, fobj, PDF_NAME(LastChar), last); |
| 189 | } |
| 190 | |
| 191 | static void |
| 192 | pdf_add_cid_system_info(fz_context *ctx, pdf_document *doc, pdf_obj *fobj, const char *reg, const char *ord, int supp) |
| 193 | { |
| 194 | pdf_obj *csi = pdf_dict_put_dict(ctx, fobj, PDF_NAME(CIDSystemInfo), 3); |
| 195 | pdf_dict_put_string(ctx, csi, PDF_NAME(Registry), reg, strlen(reg)); |
| 196 | pdf_dict_put_string(ctx, csi, PDF_NAME(Ordering), ord, strlen(ord)); |
| 197 | pdf_dict_put_int(ctx, csi, PDF_NAME(Supplement), supp); |
| 198 | } |
| 199 | |
| 200 | /* Different states of starting, same width as last, or consecutive glyph */ |
| 201 | enum { FW_START, FW_SAME, FW_RUN }; |
| 202 | |
| 203 | /* ToDo: Ignore the default sized characters */ |
| 204 | static void |
| 205 | pdf_add_cid_font_widths(fz_context *ctx, pdf_document *doc, pdf_obj *fobj, fz_font *font) |
| 206 | { |
| 207 | FT_Face face = font->ft_face; |
| 208 | pdf_obj *run_obj = NULL; |
| 209 | pdf_obj *fw; |
| 210 | int curr_code; |
| 211 | int prev_code; |
| 212 | int curr_size; |
| 213 | int prev_size; |
| 214 | int first_code; |
| 215 | int new_first_code; |
| 216 | int state = FW_START; |
| 217 | int new_state = FW_START; |
| 218 | int publish = 0; |
| 219 | |
| 220 | fz_var(run_obj); |
| 221 | |
| 222 | fw = pdf_add_new_array(ctx, doc, 10); |
| 223 | fz_try(ctx) |
| 224 | { |
| 225 | prev_code = 0; |
| 226 | prev_size = fz_advance_glyph(ctx, font, 0, 0) * 1000; |
| 227 | first_code = prev_code; |
| 228 | |
| 229 | for (;;) |
| 230 | { |
| 231 | curr_code = prev_code + 1; |
| 232 | if (curr_code >= face->num_glyphs) |
| 233 | break; |
| 234 | curr_size = fz_advance_glyph(ctx, font, curr_code, 0) * 1000; |
| 235 | |
| 236 | switch (state) |
| 237 | { |
| 238 | case FW_SAME: |
| 239 | if (curr_size != prev_size) |
| 240 | { |
| 241 | /* End of same widths for consecutive ids. Current will |
| 242 | * be pushed as prev. below during next iteration */ |
| 243 | publish = 1; |
| 244 | if (curr_code < face->num_glyphs) |
| 245 | run_obj = pdf_new_array(ctx, doc, 10); |
| 246 | new_state = FW_RUN; |
| 247 | /* And the new first code is our current code */ |
| 248 | new_first_code = curr_code; |
| 249 | } |
| 250 | break; |
| 251 | case FW_RUN: |
| 252 | if (curr_size == prev_size) |
| 253 | { |
| 254 | /* Same width, so start a new same entry starting with |
| 255 | * the previous code. i.e. the prev size is not put |
| 256 | * in the run */ |
| 257 | publish = 1; |
| 258 | new_state = FW_SAME; |
| 259 | new_first_code = prev_code; |
| 260 | } |
| 261 | else |
| 262 | { |
| 263 | /* Add prev size to run_obj */ |
| 264 | pdf_array_push_int(ctx, run_obj, prev_size); |
| 265 | } |
| 266 | break; |
| 267 | case FW_START: |
| 268 | /* Starting fresh. Determine our state */ |
| 269 | if (curr_size == prev_size) |
| 270 | { |
| 271 | state = FW_SAME; |
| 272 | } |
| 273 | else |
| 274 | { |
| 275 | run_obj = pdf_new_array(ctx, doc, 10); |
| 276 | pdf_array_push_int(ctx, run_obj, prev_size); |
| 277 | state = FW_RUN; |
| 278 | } |
| 279 | new_first_code = prev_code; |
| 280 | break; |
| 281 | } |
| 282 | |
| 283 | if (publish || curr_code == face->num_glyphs) |
| 284 | { |
| 285 | switch (state) |
| 286 | { |
| 287 | case FW_SAME: |
| 288 | /* Add three entries. First cid, last cid and width */ |
| 289 | pdf_array_push_int(ctx, fw, first_code); |
| 290 | pdf_array_push_int(ctx, fw, prev_code); |
| 291 | pdf_array_push_int(ctx, fw, prev_size); |
| 292 | break; |
| 293 | case FW_RUN: |
| 294 | if (pdf_array_len(ctx, run_obj) > 0) |
| 295 | { |
| 296 | pdf_array_push_int(ctx, fw, first_code); |
| 297 | pdf_array_push(ctx, fw, run_obj); |
| 298 | } |
| 299 | pdf_drop_obj(ctx, run_obj); |
| 300 | run_obj = NULL; |
| 301 | break; |
| 302 | case FW_START: |
| 303 | /* Lone wolf. Not part of a consecutive run */ |
| 304 | pdf_array_push_int(ctx, fw, prev_code); |
| 305 | pdf_array_push_int(ctx, fw, prev_code); |
| 306 | pdf_array_push_int(ctx, fw, prev_size); |
| 307 | break; |
| 308 | } |
| 309 | |
| 310 | if (curr_code < face->num_glyphs) |
| 311 | { |
| 312 | state = new_state; |
| 313 | first_code = new_first_code; |
| 314 | publish = 0; |
| 315 | } |
| 316 | } |
| 317 | |
| 318 | prev_size = curr_size; |
| 319 | prev_code = curr_code; |
| 320 | } |
| 321 | |
| 322 | if (font->width_table != NULL) |
| 323 | pdf_dict_put_int(ctx, fobj, PDF_NAME(DW), font->width_default); |
| 324 | if (pdf_array_len(ctx, fw) > 0) |
| 325 | pdf_dict_put(ctx, fobj, PDF_NAME(W), fw); |
| 326 | } |
| 327 | fz_always(ctx) |
| 328 | pdf_drop_obj(ctx, fw); |
| 329 | fz_catch(ctx) |
| 330 | fz_rethrow(ctx); |
| 331 | } |
| 332 | |
| 333 | /* Descendant font construction used for CID font creation from ttf or Adobe type1 */ |
| 334 | static pdf_obj* |
| 335 | pdf_add_descendant_cid_font(fz_context *ctx, pdf_document *doc, fz_font *font) |
| 336 | { |
| 337 | FT_Face face = font->ft_face; |
| 338 | pdf_obj *fobj, *fref; |
| 339 | const char *ps_name; |
| 340 | |
| 341 | fobj = pdf_new_dict(ctx, doc, 3); |
| 342 | fz_try(ctx) |
| 343 | { |
| 344 | pdf_dict_put(ctx, fobj, PDF_NAME(Type), PDF_NAME(Font)); |
| 345 | if (is_truetype(face)) |
| 346 | pdf_dict_put(ctx, fobj, PDF_NAME(Subtype), PDF_NAME(CIDFontType2)); |
| 347 | else |
| 348 | pdf_dict_put(ctx, fobj, PDF_NAME(Subtype), PDF_NAME(CIDFontType0)); |
| 349 | |
| 350 | pdf_add_cid_system_info(ctx, doc, fobj, "Adobe" , "Identity" , 0); |
| 351 | |
| 352 | ps_name = FT_Get_Postscript_Name(face); |
| 353 | if (ps_name) |
| 354 | pdf_dict_put_name(ctx, fobj, PDF_NAME(BaseFont), ps_name); |
| 355 | else |
| 356 | pdf_dict_put_name(ctx, fobj, PDF_NAME(BaseFont), font->name); |
| 357 | |
| 358 | pdf_add_font_descriptor(ctx, doc, fobj, font); |
| 359 | |
| 360 | /* We may have a cid font already with width info in source font and no cmap in the ft face */ |
| 361 | pdf_add_cid_font_widths(ctx, doc, fobj, font); |
| 362 | |
| 363 | fref = pdf_add_object(ctx, doc, fobj); |
| 364 | } |
| 365 | fz_always(ctx) |
| 366 | pdf_drop_obj(ctx, fobj); |
| 367 | fz_catch(ctx) |
| 368 | fz_rethrow(ctx); |
| 369 | return fref; |
| 370 | } |
| 371 | |
| 372 | static int next_range(int *table, int size, int k) |
| 373 | { |
| 374 | int n; |
| 375 | for (n = 1; k + n < size; ++n) |
| 376 | { |
| 377 | if ((k & 0xFF00) != ((k+n) & 0xFF00)) /* high byte changes */ |
| 378 | break; |
| 379 | if (table[k] + n != table[k+n]) |
| 380 | break; |
| 381 | } |
| 382 | return n; |
| 383 | } |
| 384 | |
| 385 | /* Create the ToUnicode CMap. */ |
| 386 | static void |
| 387 | pdf_add_to_unicode(fz_context *ctx, pdf_document *doc, pdf_obj *fobj, fz_font *font) |
| 388 | { |
| 389 | FT_Face face = font->ft_face; |
| 390 | fz_buffer *buf; |
| 391 | |
| 392 | int *table; |
| 393 | int num_seq = 0; |
| 394 | int num_chr = 0; |
| 395 | int n, k; |
| 396 | |
| 397 | /* Populate reverse cmap table */ |
| 398 | { |
| 399 | FT_ULong ucs; |
| 400 | FT_UInt gid; |
| 401 | |
| 402 | table = fz_calloc(ctx, face->num_glyphs, sizeof *table); |
| 403 | fz_lock(ctx, FZ_LOCK_FREETYPE); |
| 404 | ucs = FT_Get_First_Char(face, &gid); |
| 405 | while (gid > 0) |
| 406 | { |
| 407 | if (gid < (FT_ULong)face->num_glyphs && face->num_glyphs > 0) |
| 408 | table[gid] = ucs; |
| 409 | ucs = FT_Get_Next_Char(face, ucs, &gid); |
| 410 | } |
| 411 | fz_unlock(ctx, FZ_LOCK_FREETYPE); |
| 412 | } |
| 413 | |
| 414 | for (k = 0; k < face->num_glyphs; k += n) |
| 415 | { |
| 416 | n = next_range(table, face->num_glyphs, k); |
| 417 | if (n > 1) |
| 418 | ++num_seq; |
| 419 | else if (table[k] > 0) |
| 420 | ++num_chr; |
| 421 | } |
| 422 | |
| 423 | /* No mappings available... */ |
| 424 | if (num_seq + num_chr == 0) |
| 425 | { |
| 426 | fz_warn(ctx, "cannot create ToUnicode mapping for %s" , font->name); |
| 427 | fz_free(ctx, table); |
| 428 | return; |
| 429 | } |
| 430 | |
| 431 | buf = fz_new_buffer(ctx, 0); |
| 432 | fz_try(ctx) |
| 433 | { |
| 434 | /* Header boiler plate */ |
| 435 | fz_append_string(ctx, buf, "/CIDInit /ProcSet findresource begin\n" ); |
| 436 | fz_append_string(ctx, buf, "12 dict begin\n" ); |
| 437 | fz_append_string(ctx, buf, "begincmap\n" ); |
| 438 | fz_append_string(ctx, buf, "/CIDSystemInfo <</Registry(Adobe)/Ordering(UCS)/Supplement 0>> def\n" ); |
| 439 | fz_append_string(ctx, buf, "/CMapName /Adobe-Identity-UCS def\n" ); |
| 440 | fz_append_string(ctx, buf, "/CMapType 2 def\n" ); |
| 441 | fz_append_string(ctx, buf, "1 begincodespacerange\n" ); |
| 442 | fz_append_string(ctx, buf, "<0000> <FFFF>\n" ); |
| 443 | fz_append_string(ctx, buf, "endcodespacerange\n" ); |
| 444 | |
| 445 | /* Note to have a valid CMap, the number of entries in table set can |
| 446 | * not exceed 100, so we have to break into multiple tables. Also, note |
| 447 | * that to reduce the file size we should be looking for sequential |
| 448 | * ranges. Per Adobe technical note #5411, we can't have a range |
| 449 | * cross a boundary where the high order byte changes */ |
| 450 | |
| 451 | /* First the ranges */ |
| 452 | if (num_seq > 0) |
| 453 | { |
| 454 | int count = 0; |
| 455 | if (num_seq > 100) |
| 456 | { |
| 457 | fz_append_string(ctx, buf, "100 beginbfrange\n" ); |
| 458 | num_seq -= 100; |
| 459 | } |
| 460 | else |
| 461 | fz_append_printf(ctx, buf, "%d beginbfrange\n" , num_seq); |
| 462 | for (k = 0; k < face->num_glyphs; k += n) |
| 463 | { |
| 464 | n = next_range(table, face->num_glyphs, k); |
| 465 | if (n > 1) |
| 466 | { |
| 467 | if (count == 100) |
| 468 | { |
| 469 | fz_append_string(ctx, buf, "endbfrange\n" ); |
| 470 | if (num_seq > 100) |
| 471 | { |
| 472 | fz_append_string(ctx, buf, "100 beginbfrange\n" ); |
| 473 | num_seq -= 100; |
| 474 | } |
| 475 | else |
| 476 | fz_append_printf(ctx, buf, "%d beginbfrange\n" , num_seq); |
| 477 | count = 0; |
| 478 | } |
| 479 | fz_append_printf(ctx, buf, "<%04x> <%04x> <%04x>\n" , k, k+n-1, table[k]); |
| 480 | ++count; |
| 481 | } |
| 482 | } |
| 483 | fz_append_string(ctx, buf, "endbfrange\n" ); |
| 484 | } |
| 485 | |
| 486 | /* Then the singles */ |
| 487 | if (num_chr > 0) |
| 488 | { |
| 489 | int count = 0; |
| 490 | if (num_chr > 100) |
| 491 | { |
| 492 | fz_append_string(ctx, buf, "100 beginbfchar\n" ); |
| 493 | num_chr -= 100; |
| 494 | } |
| 495 | else |
| 496 | fz_append_printf(ctx, buf, "%d beginbfchar\n" , num_chr); |
| 497 | for (k = 0; k < face->num_glyphs; k += n) |
| 498 | { |
| 499 | n = next_range(table, face->num_glyphs, k); |
| 500 | if (n == 1 && table[k] > 0) |
| 501 | { |
| 502 | if (count == 100) |
| 503 | { |
| 504 | fz_append_string(ctx, buf, "endbfchar\n" ); |
| 505 | if (num_chr > 100) |
| 506 | { |
| 507 | fz_append_string(ctx, buf, "100 beginbfchar\n" ); |
| 508 | num_chr -= 100; |
| 509 | } |
| 510 | else |
| 511 | fz_append_printf(ctx, buf, "%d beginbfchar\n" , num_chr); |
| 512 | count = 0; |
| 513 | } |
| 514 | fz_append_printf(ctx, buf, "<%04x> <%04x>\n" , k, table[k]); |
| 515 | ++count; |
| 516 | } |
| 517 | } |
| 518 | fz_append_string(ctx, buf, "endbfchar\n" ); |
| 519 | } |
| 520 | |
| 521 | /* Trailer boiler plate */ |
| 522 | fz_append_string(ctx, buf, "endcmap\n" ); |
| 523 | fz_append_string(ctx, buf, "CMapName currentdict /CMap defineresource pop\n" ); |
| 524 | fz_append_string(ctx, buf, "end\nend\n" ); |
| 525 | |
| 526 | pdf_dict_put_drop(ctx, fobj, PDF_NAME(ToUnicode), pdf_add_stream(ctx, doc, buf, NULL, 0)); |
| 527 | } |
| 528 | fz_always(ctx) |
| 529 | { |
| 530 | fz_free(ctx, table); |
| 531 | fz_drop_buffer(ctx, buf); |
| 532 | } |
| 533 | fz_catch(ctx) |
| 534 | fz_rethrow(ctx); |
| 535 | } |
| 536 | |
| 537 | /* Creates CID font with Identity-H CMap and a ToUnicode CMap that is created by |
| 538 | * using the TTF cmap table "backwards" to go from the GID to a Unicode value. |
| 539 | * We can possibly get width information that may have been embedded in |
| 540 | * the PDF /W array (or W2 if vertical text) */ |
| 541 | pdf_obj * |
| 542 | pdf_add_cid_font(fz_context *ctx, pdf_document *doc, fz_font *font) |
| 543 | { |
| 544 | pdf_obj *fobj = NULL; |
| 545 | pdf_obj *fref = NULL; |
| 546 | pdf_obj *dfonts = NULL; |
| 547 | unsigned char digest[16]; |
| 548 | |
| 549 | fref = pdf_find_font_resource(ctx, doc, PDF_CID_FONT_RESOURCE, 0, font, digest); |
| 550 | if (fref) |
| 551 | return fref; |
| 552 | |
| 553 | fobj = pdf_add_new_dict(ctx, doc, 10); |
| 554 | fz_try(ctx) |
| 555 | { |
| 556 | pdf_dict_put(ctx, fobj, PDF_NAME(Type), PDF_NAME(Font)); |
| 557 | pdf_dict_put(ctx, fobj, PDF_NAME(Subtype), PDF_NAME(Type0)); |
| 558 | pdf_dict_put_name(ctx, fobj, PDF_NAME(BaseFont), font->name); |
| 559 | pdf_dict_put(ctx, fobj, PDF_NAME(Encoding), PDF_NAME(Identity_H)); |
| 560 | pdf_add_to_unicode(ctx, doc, fobj, font); |
| 561 | |
| 562 | dfonts = pdf_dict_put_array(ctx, fobj, PDF_NAME(DescendantFonts), 1); |
| 563 | pdf_array_push_drop(ctx, dfonts, pdf_add_descendant_cid_font(ctx, doc, font)); |
| 564 | |
| 565 | fref = pdf_insert_font_resource(ctx, doc, digest, fobj); |
| 566 | } |
| 567 | fz_always(ctx) |
| 568 | pdf_drop_obj(ctx, fobj); |
| 569 | fz_catch(ctx) |
| 570 | fz_rethrow(ctx); |
| 571 | return fref; |
| 572 | } |
| 573 | |
| 574 | /* Create simple (8-bit encoding) fonts */ |
| 575 | |
| 576 | static void |
| 577 | pdf_add_simple_font_encoding_imp(fz_context *ctx, pdf_document *doc, pdf_obj *font, const char *glyph_names[]) |
| 578 | { |
| 579 | pdf_obj *enc, *diff; |
| 580 | int i, last; |
| 581 | |
| 582 | enc = pdf_dict_put_dict(ctx, font, PDF_NAME(Encoding), 2); |
| 583 | pdf_dict_put(ctx, enc, PDF_NAME(BaseEncoding), PDF_NAME(WinAnsiEncoding)); |
| 584 | diff = pdf_dict_put_array(ctx, enc, PDF_NAME(Differences), 129); |
| 585 | last = 0; |
| 586 | for (i = 128; i < 256; ++i) |
| 587 | { |
| 588 | const char *glyph = glyph_names[i]; |
| 589 | if (glyph) |
| 590 | { |
| 591 | if (last != i-1) |
| 592 | pdf_array_push_int(ctx, diff, i); |
| 593 | last = i; |
| 594 | pdf_array_push_name(ctx, diff, glyph); |
| 595 | } |
| 596 | } |
| 597 | } |
| 598 | |
| 599 | static void |
| 600 | pdf_add_simple_font_encoding(fz_context *ctx, pdf_document *doc, pdf_obj *fobj, int encoding) |
| 601 | { |
| 602 | switch (encoding) |
| 603 | { |
| 604 | default: |
| 605 | case PDF_SIMPLE_ENCODING_LATIN: |
| 606 | pdf_dict_put(ctx, fobj, PDF_NAME(Encoding), PDF_NAME(WinAnsiEncoding)); |
| 607 | break; |
| 608 | case PDF_SIMPLE_ENCODING_GREEK: |
| 609 | pdf_add_simple_font_encoding_imp(ctx, doc, fobj, fz_glyph_name_from_iso8859_7); |
| 610 | break; |
| 611 | case PDF_SIMPLE_ENCODING_CYRILLIC: |
| 612 | pdf_add_simple_font_encoding_imp(ctx, doc, fobj, fz_glyph_name_from_koi8u); |
| 613 | break; |
| 614 | } |
| 615 | } |
| 616 | |
| 617 | pdf_obj * |
| 618 | pdf_add_simple_font(fz_context *ctx, pdf_document *doc, fz_font *font, int encoding) |
| 619 | { |
| 620 | FT_Face face = font->ft_face; |
| 621 | pdf_obj *fobj = NULL; |
| 622 | pdf_obj *fref = NULL; |
| 623 | const char **enc; |
| 624 | unsigned char digest[16]; |
| 625 | |
| 626 | fref = pdf_find_font_resource(ctx, doc, PDF_SIMPLE_FONT_RESOURCE, encoding, font, digest); |
| 627 | if (fref) |
| 628 | return fref; |
| 629 | |
| 630 | switch (encoding) |
| 631 | { |
| 632 | default: |
| 633 | case PDF_SIMPLE_ENCODING_LATIN: enc = fz_glyph_name_from_windows_1252; break; |
| 634 | case PDF_SIMPLE_ENCODING_GREEK: enc = fz_glyph_name_from_iso8859_7; break; |
| 635 | case PDF_SIMPLE_ENCODING_CYRILLIC: enc = fz_glyph_name_from_koi8u; break; |
| 636 | } |
| 637 | |
| 638 | fobj = pdf_add_new_dict(ctx, doc, 10); |
| 639 | fz_try(ctx) |
| 640 | { |
| 641 | pdf_dict_put(ctx, fobj, PDF_NAME(Type), PDF_NAME(Font)); |
| 642 | if (is_truetype(face)) |
| 643 | pdf_dict_put(ctx, fobj, PDF_NAME(Subtype), PDF_NAME(TrueType)); |
| 644 | else |
| 645 | pdf_dict_put(ctx, fobj, PDF_NAME(Subtype), PDF_NAME(Type1)); |
| 646 | |
| 647 | if (!is_builtin_font(ctx, font)) |
| 648 | { |
| 649 | const char *ps_name = FT_Get_Postscript_Name(face); |
| 650 | if (!ps_name) |
| 651 | ps_name = font->name; |
| 652 | pdf_dict_put_name(ctx, fobj, PDF_NAME(BaseFont), ps_name); |
| 653 | pdf_add_simple_font_encoding(ctx, doc, fobj, encoding); |
| 654 | pdf_add_simple_font_widths(ctx, doc, fobj, font, enc); |
| 655 | pdf_add_font_descriptor(ctx, doc, fobj, font); |
| 656 | } |
| 657 | else |
| 658 | { |
| 659 | pdf_dict_put_name(ctx, fobj, PDF_NAME(BaseFont), pdf_clean_font_name(font->name)); |
| 660 | pdf_add_simple_font_encoding(ctx, doc, fobj, encoding); |
| 661 | if (encoding != PDF_SIMPLE_ENCODING_LATIN) |
| 662 | pdf_add_simple_font_widths(ctx, doc, fobj, font, enc); |
| 663 | } |
| 664 | |
| 665 | fref = pdf_insert_font_resource(ctx, doc, digest, fobj); |
| 666 | } |
| 667 | fz_always(ctx) |
| 668 | { |
| 669 | pdf_drop_obj(ctx, fobj); |
| 670 | } |
| 671 | fz_catch(ctx) |
| 672 | fz_rethrow(ctx); |
| 673 | return fref; |
| 674 | } |
| 675 | |
| 676 | int |
| 677 | pdf_font_writing_supported(fz_font *font) |
| 678 | { |
| 679 | if (font->ft_face == NULL || font->buffer == NULL || font->buffer->len < 4) |
| 680 | return 0; |
| 681 | if (is_ttc(font)) |
| 682 | return 0; |
| 683 | if (is_truetype(font->ft_face)) |
| 684 | return 1; |
| 685 | if (is_postscript(font->ft_face)) |
| 686 | return 1; |
| 687 | return 0; |
| 688 | } |
| 689 | |
| 690 | /* Add a non-embedded UTF16-encoded CID-font for the CJK scripts: CNS1, GB1, Japan1, or Korea1 */ |
| 691 | pdf_obj * |
| 692 | pdf_add_cjk_font(fz_context *ctx, pdf_document *doc, fz_font *fzfont, int script, int wmode, int serif) |
| 693 | { |
| 694 | pdf_obj *fref, *font, *subfont, *fontdesc; |
| 695 | pdf_obj *dfonts; |
| 696 | fz_rect bbox = { -200, -200, 1200, 1200 }; |
| 697 | unsigned char digest[16]; |
| 698 | int flags; |
| 699 | |
| 700 | const char *basefont, *encoding, *ordering; |
| 701 | int supplement; |
| 702 | |
| 703 | switch (script) |
| 704 | { |
| 705 | default: |
| 706 | script = FZ_ADOBE_CNS; |
| 707 | /* fall through */ |
| 708 | case FZ_ADOBE_CNS: /* traditional chinese */ |
| 709 | basefont = serif ? "Ming" : "Fangti" ; |
| 710 | encoding = wmode ? "UniCNS-UTF16-V" : "UniCNS-UTF16-H" ; |
| 711 | ordering = "CNS1" ; |
| 712 | supplement = 7; |
| 713 | break; |
| 714 | case FZ_ADOBE_GB: /* simplified chinese */ |
| 715 | basefont = serif ? "Song" : "Heiti" ; |
| 716 | encoding = wmode ? "UniGB-UTF16-V" : "UniGB-UTF16-H" ; |
| 717 | ordering = "GB1" ; |
| 718 | supplement = 5; |
| 719 | break; |
| 720 | case FZ_ADOBE_JAPAN: |
| 721 | basefont = serif ? "Mincho" : "Gothic" ; |
| 722 | encoding = wmode ? "UniJIS-UTF16-V" : "UniJIS-UTF16-H" ; |
| 723 | ordering = "Japan1" ; |
| 724 | supplement = 6; |
| 725 | break; |
| 726 | case FZ_ADOBE_KOREA: |
| 727 | basefont = serif ? "Batang" : "Dotum" ; |
| 728 | encoding = wmode ? "UniKS-UTF16-V" : "UniKS-UTF16-H" ; |
| 729 | ordering = "Korea1" ; |
| 730 | supplement = 2; |
| 731 | break; |
| 732 | } |
| 733 | |
| 734 | flags = PDF_FD_SYMBOLIC; |
| 735 | if (serif) |
| 736 | flags |= PDF_FD_SERIF; |
| 737 | |
| 738 | fref = pdf_find_font_resource(ctx, doc, PDF_CJK_FONT_RESOURCE, script, fzfont, digest); |
| 739 | if (fref) |
| 740 | return fref; |
| 741 | |
| 742 | font = pdf_add_new_dict(ctx, doc, 5); |
| 743 | fz_try(ctx) |
| 744 | { |
| 745 | pdf_dict_put(ctx, font, PDF_NAME(Type), PDF_NAME(Font)); |
| 746 | pdf_dict_put(ctx, font, PDF_NAME(Subtype), PDF_NAME(Type0)); |
| 747 | pdf_dict_put_name(ctx, font, PDF_NAME(BaseFont), basefont); |
| 748 | pdf_dict_put_name(ctx, font, PDF_NAME(Encoding), encoding); |
| 749 | dfonts = pdf_dict_put_array(ctx, font, PDF_NAME(DescendantFonts), 1); |
| 750 | pdf_array_push_drop(ctx, dfonts, subfont = pdf_add_new_dict(ctx, doc, 5)); |
| 751 | { |
| 752 | pdf_dict_put(ctx, subfont, PDF_NAME(Type), PDF_NAME(Font)); |
| 753 | pdf_dict_put(ctx, subfont, PDF_NAME(Subtype), PDF_NAME(CIDFontType0)); |
| 754 | pdf_dict_put_name(ctx, subfont, PDF_NAME(BaseFont), basefont); |
| 755 | pdf_add_cid_system_info(ctx, doc, subfont, "Adobe" , ordering, supplement); |
| 756 | fontdesc = pdf_add_new_dict(ctx, doc, 8); |
| 757 | pdf_dict_put_drop(ctx, subfont, PDF_NAME(FontDescriptor), fontdesc); |
| 758 | { |
| 759 | pdf_dict_put(ctx, fontdesc, PDF_NAME(Type), PDF_NAME(FontDescriptor)); |
| 760 | pdf_dict_put_text_string(ctx, fontdesc, PDF_NAME(FontName), basefont); |
| 761 | pdf_dict_put_rect(ctx, fontdesc, PDF_NAME(FontBBox), bbox); |
| 762 | pdf_dict_put_int(ctx, fontdesc, PDF_NAME(Flags), flags); |
| 763 | pdf_dict_put_int(ctx, fontdesc, PDF_NAME(ItalicAngle), 0); |
| 764 | pdf_dict_put_int(ctx, fontdesc, PDF_NAME(Ascent), 1000); |
| 765 | pdf_dict_put_int(ctx, fontdesc, PDF_NAME(Descent), -200); |
| 766 | pdf_dict_put_int(ctx, fontdesc, PDF_NAME(StemV), 80); |
| 767 | } |
| 768 | } |
| 769 | |
| 770 | fref = pdf_insert_font_resource(ctx, doc, digest, font); |
| 771 | } |
| 772 | fz_always(ctx) |
| 773 | pdf_drop_obj(ctx, font); |
| 774 | fz_catch(ctx) |
| 775 | fz_rethrow(ctx); |
| 776 | |
| 777 | return fref; |
| 778 | } |
| 779 | |