| OLD | NEW |
| (Empty) | |
| 1 2015-11-28 Werner Lemberg <wl@gnu.org> |
| 2 |
| 3 * Version 2.6.2 released. |
| 4 ========================= |
| 5 |
| 6 |
| 7 Tag sources with `VER-2-6-2'. |
| 8 |
| 9 * docs/VERSION.DLL: Update documentation and bump version number to |
| 10 2.6.2. |
| 11 |
| 12 * README, Jamfile (RefDoc), builds/windows/vc2005/freetype.vcproj, |
| 13 builds/windows/vc2005/index.html, |
| 14 builds/windows/vc2008/freetype.vcproj, |
| 15 builds/windows/vc2008/index.html, |
| 16 builds/windows/vc2010/freetype.vcxproj, |
| 17 builds/windows/vc2010/index.html, |
| 18 builds/windows/visualc/freetype.dsp, |
| 19 builds/windows/visualc/freetype.vcproj, |
| 20 builds/windows/visualc/index.html, |
| 21 builds/windows/visualce/freetype.dsp, |
| 22 builds/windows/visualce/freetype.vcproj, |
| 23 builds/windows/visualce/index.html, |
| 24 builds/wince/vc2005-ce/freetype.vcproj, |
| 25 builds/wince/vc2005-ce/index.html, |
| 26 builds/wince/vc2008-ce/freetype.vcproj, |
| 27 builds/wince/vc2008-ce/index.html: s/2.6.1/2.6.2/, s/261/262/. |
| 28 |
| 29 * include/freetype/freetype.h (FREETYPE_PATCH): Set to 2. |
| 30 |
| 31 * builds/unix/configure.raw (version_info): Set to 18:2:12. |
| 32 * CMakeLists.txt (VERSION_PATCH): Set to 2. |
| 33 |
| 34 * docs/CHANGES: Updated. |
| 35 |
| 36 2015-11-28 Werner Lemberg <wl@gnu.org> |
| 37 |
| 38 Fix C++ compilation. |
| 39 |
| 40 * src/autofit/afloader.c: Include FT_INTERNAL_CALC_H. |
| 41 |
| 42 * src/truetype/ttgload.c (load_truetype_glyph): Pacify compiler. |
| 43 |
| 44 2015-11-28 Nikolaus Waxweiler <madigens@gmail.com> |
| 45 |
| 46 Change default LCD filter to be normalized and color-balanced. |
| 47 |
| 48 * src/base/ftlcdfil.c (FT_Library_SetLcdFilter): Update |
| 49 `default_filter'. |
| 50 |
| 51 2015-11-28 Werner Lemberg <wl@gnu.org> |
| 52 |
| 53 [docmaker] Allow references to section names. |
| 54 |
| 55 In the reference, we show the section's title enclosed in single |
| 56 quotes. |
| 57 |
| 58 * src/tools/docmaker/formatter.py (Formatter::__init__): Collect |
| 59 section names as identifiers. |
| 60 |
| 61 * src/tools/docmaker/tohtml.py (section_title_header): Split into... |
| 62 (section_title_header1, section_title_header2): ... these two |
| 63 strings. |
| 64 (HtmlFormatter::make_block_url, make_html_word, html_source_quote): |
| 65 Handle sections. |
| 66 (HtmlFormatter::section_enter): Updated to add `id' HTML attribute. |
| 67 |
| 68 2015-11-27 Tamas Kenez <tamas.kenez@adasworks.com> |
| 69 |
| 70 [cmake] Add script to test the config module. |
| 71 |
| 72 * builds/cmake/testbuild.sh: New file. |
| 73 |
| 74 2015-11-27 Tamas Kenez <tamas.kenez@adasworks.com> |
| 75 |
| 76 * CMakeLists.txt: Create `freetype-config.cmake' config module. |
| 77 |
| 78 2015-11-27 Tamas Kenez <tamas.kenez@adasworks.com> |
| 79 |
| 80 * CMakeLists.txt: Set CMAKE_DEBUG_POSTFIX to `d'. |
| 81 |
| 82 2015-11-27 Tamas Kenez <tamas.kenez@adasworks.com> |
| 83 |
| 84 [cmake] Add better control of library dependencies. |
| 85 |
| 86 * CMakeLists.txt: Add `WITH_*' variables to force/auto/omit |
| 87 ZLIB/BZip2/PNG/HarfBuzz. |
| 88 |
| 89 2015-11-27 Tamas Kenez <tamas.kenez@adasworks.com> |
| 90 |
| 91 [cmake] Make `FindHarfbuzz' observe the REQUIRED option. |
| 92 |
| 93 * builds/cmake/FindHarfBuzz.cmake: Implement it. |
| 94 |
| 95 2015-11-27 Werner Lemberg <wl@gnu.org> |
| 96 |
| 97 [cmake] Collect files specific to cmake in `builds/cmake'. |
| 98 |
| 99 * builds/FindHarfBuzz.cmake: Move to ... |
| 100 * builds/cmake/FindHarfBuzz.cmake: ... this place. |
| 101 |
| 102 * CMakeLists.txt (CMAKE_MODULE_PATH): Updated. |
| 103 |
| 104 2015-11-27 Alexander Bock <alexander.j.bock@nasa.gov> |
| 105 |
| 106 CMakeLists.txt: Honour new command line flag `FREETYPE_NO_DIST'. |
| 107 |
| 108 2015-11-26 Werner Lemberg <wl@gnu.org> |
| 109 |
| 110 [docmaker] Allow `foo[bar]' as identifier. |
| 111 |
| 112 We need this to handle equally named properties in different |
| 113 modules. |
| 114 |
| 115 * src/tools/docmaker/content.py (re_identifier), |
| 116 src/tools/docmaker/sources.py (re_crossref): Allow `foo[bar]'. |
| 117 |
| 118 * src/tools/docmaker/tohtml.py (HtmlFormatter::make_html_word, |
| 119 HtmlFormatter::index_exit, HtmlFormatter::section_enter, |
| 120 HtmlFormatter::block_enter): Handle `foo[bar]'. |
| 121 |
| 122 2015-11-25 Werner Lemberg <wl@gnu.org> |
| 123 |
| 124 * src/bdf/bdflib.c (bdf_load_font): Fix small memory leak (#46480). |
| 125 |
| 126 (_bdf_parse_glyphs): Always reset `p->glyph_name' after moving its |
| 127 contents. |
| 128 |
| 129 2015-11-21 Werner Lemberg <wl@gnu.org> |
| 130 |
| 131 * include/freetype/internal/ftcalc.h: Don't use `register' keyword. |
| 132 |
| 133 This fixes compiler warnings. |
| 134 |
| 135 Reported by Behdad. |
| 136 |
| 137 2015-11-20 Werner Lemberg <wl@gnu.org> |
| 138 |
| 139 Add `FT_LCD_FILTER_LEGACY1' enum value. |
| 140 |
| 141 This does the same as `FT_LCD_FILTER_LEGACY'. |
| 142 |
| 143 See |
| 144 |
| 145 https://bugs.freedesktop.org/show_bug.cgi?id=92981 |
| 146 |
| 147 for the reasoning. |
| 148 |
| 149 * include/freetype/ftlcdfil.h (FT_LcdFilter): New value |
| 150 `FT_LCD_FILTER_LEGACY1'. |
| 151 |
| 152 * src/base/ftlcdfil.c (FT_Library_SetLcdFilter): Use it. |
| 153 |
| 154 2015-11-15 Werner Lemberg <wl@gnu.org> |
| 155 |
| 156 * src/autofit/afhints.c (af_get_segment_index): Fix it. |
| 157 |
| 158 The old code was too simple, returning invalid values in most cases |
| 159 where a segment crosses the contour start. |
| 160 |
| 161 2015-11-15 Werner Lemberg <wl@gnu.org> |
| 162 |
| 163 * src/bdf/bdflib.c (bdf_load_font): Fix small memory leak (#46439). |
| 164 |
| 165 2015-11-11 Werner Lemberg <wl@gnu.org> |
| 166 |
| 167 [cff, autofit] Switch off stem darkening by default. |
| 168 |
| 169 * src/autofit/afmodule.c (af_autofitter_init), src/cff/cffobjs.c |
| 170 (cff_driver_init): Do it. |
| 171 |
| 172 2015-11-10 Jan Alexander Steffens (heftig) <jan.steffens@gmail.com> |
| 173 |
| 174 Allow native CFF hinter in FT_RENDER_MODE_LIGHT. |
| 175 |
| 176 Both the native CFF hinter and the auto-hinter now have a very |
| 177 similar rendering style. |
| 178 |
| 179 * include/freetype/freetype.h: Mention that FT_LOAD_TARGET_LIGHT no |
| 180 longer implies FT_LOAD_FORCE_AUTOHINT. |
| 181 |
| 182 * include/freetype/ftmodapi.h (FT_MODULE_DRIVER_HINTS_LIGHTLY): New |
| 183 macro. |
| 184 |
| 185 * include/freetype/internal/ftobjs.h (FT_DRIVER_HINTS_LIGHTLY): New |
| 186 macro. |
| 187 |
| 188 * src/cff/cffdrivr.c (cff_driver_class): Use it. |
| 189 |
| 190 * src/base/ftobjs.c (FT_Load_Glyph): Update auto-hinter selection |
| 191 logic. |
| 192 |
| 193 2015-11-09 Werner Lemberg <wl@gnu.org> |
| 194 |
| 195 * src/cid/cidload.c (cid_face_open): Fix GDBytes guard (#46408). |
| 196 |
| 197 2015-11-09 Werner Lemberg <wl@gnu.org> |
| 198 |
| 199 [truetype] Remove integer to pointer conversion compiler warning. |
| 200 |
| 201 Problem reported by Alexei. |
| 202 |
| 203 * src/truetype/ttgload.c (load_truetype_glyph): Use a solution found |
| 204 in the glib library to fix the issue. |
| 205 |
| 206 2015-11-08 Behdad Esfahbod <behdad@behdad.org> |
| 207 |
| 208 [sfnt] Accept version 3 of `EBLC' and `CBLC' tables also. |
| 209 |
| 210 * src/sfnt/ttsbit.c (tt_face_load_sbit): Implement it. |
| 211 |
| 212 2015-11-08 Philipp Knechtges <philipp-dev@knechtges.com> |
| 213 |
| 214 [autofit] Don't distort (latin) glyphs too much (#46195). |
| 215 |
| 216 * src/autofit/aflatin.h (AF_LatinBlueRec): Add `ascender' and |
| 217 `descender' fields. |
| 218 |
| 219 * src/autofit/aflatin.c (af_latin_metrics_init_blues): Collect |
| 220 ascender and descender data for blue zones. |
| 221 (af_latin_metrics_scale_dim): Reject vertical scaling values that |
| 222 change the result by more than two pixels. |
| 223 |
| 224 2015-11-05 Werner Lemberg <wl@gnu.org> |
| 225 |
| 226 [sfnt] Ignore embedded bitmaps with zero size (#46379). |
| 227 |
| 228 * src/sfnt/ttsbit.c (tt_sbit_decoder_load_bit_aligned): Implement |
| 229 it. |
| 230 |
| 231 2015-11-04 Werner Lemberg <wl@gnu.org> |
| 232 |
| 233 [truetype] Catch infinite recursion in subglyphs (#46372). |
| 234 |
| 235 * include/freetype/internal/tttypes.h (TT_LoaderRec): New field |
| 236 `composites'. |
| 237 |
| 238 * src/truetype/ttgload.c: Include FT_LIST_H. |
| 239 (load_truetype_glyph): Add composite subglyph index to a list; |
| 240 abort if index is already in list. |
| 241 (tt_loader_init): Updated. |
| 242 (tt_loader_done): New function. |
| 243 (TT_Load_Glyph): Call `tt_loader_done'. |
| 244 |
| 245 2015-11-04 Werner Lemberg <wl@gnu.org> |
| 246 |
| 247 [truetype] Better tracing of composite glyphs. |
| 248 |
| 249 * src/truetype/ttgload.c (TT_Load_Composite_Glyph, |
| 250 load_truetype_glyph): Implement it. |
| 251 |
| 252 2015-11-03 Werner Lemberg <wl@gnu.org> |
| 253 |
| 254 [sfnt] Protect against zero-size bitmaps (#46345). |
| 255 |
| 256 * src/sfnt/ttsbit.c (tt_sbit_decoder_load_bitmap): Check |
| 257 `glyph_size'. |
| 258 |
| 259 2015-11-02 Nikolaus Waxweiler <madigens@gmail.com> |
| 260 |
| 261 * src/autofit/afloader.c (af_loader_load_g): Implement emboldening. |
| 262 |
| 263 2015-11-02 Nikolaus Waxweiler <madigens@gmail.com> |
| 264 |
| 265 [autofit] Implement darkening computation function. |
| 266 |
| 267 This is a crude adaption of the original `cf2_computeDarkening' |
| 268 function. |
| 269 |
| 270 * src/autofit/afloader.c (af_intToFixed, af_fixedToInt, |
| 271 af_floatToFixed): New macros, taken from `cf2fixed.h'. |
| 272 (af_loader_compute_darkening): New function. |
| 273 * src/autofit/afloader.h: Updated. |
| 274 |
| 275 2015-11-02 Nikolaus Waxweiler <madigens@gmail.com> |
| 276 |
| 277 [autofit] Add functions to get standard widths for writing systems. |
| 278 |
| 279 We need the computed standard horizontal and vertical widths for the |
| 280 emboldening calculation. This method provides a convenient way to |
| 281 extract it from writing-system-specific metrics structures, which |
| 282 all script definitions must implement. |
| 283 |
| 284 * src/autofit/aftypes.h (AF_WritingSystem_GetStdWidthsFunc): New |
| 285 function type. |
| 286 (AF_WritingSystemClassRec): New member `style_metrics_getstdw'. |
| 287 (AF_DEFINE_WRITING_SYSTEM_CLASS): Updated. |
| 288 |
| 289 * src/autofit/afcjk.c (af_cjk_get_standard_width): New function. |
| 290 (af_cjk_writing_system_class): Updated. |
| 291 * src/autofit/afdummy.c (af_dummy_writing_system_class): Updated. |
| 292 * src/autofit/afindic.c (af_cjk_get_standard_width): New function. |
| 293 (af_indic_writing_system_class): Updated. |
| 294 * src/autofit/aflatin.c (af_latin_get_standard_width): New function. |
| 295 (af_indic_writing_system_class): Updated. |
| 296 * src/autofit/aflatin.c (af_latin_get_standard_width): New function. |
| 297 (af_indic_writing_system_class): Updated. |
| 298 |
| 299 2015-11-02 Nikolaus Waxweiler <madigens@gmail.com> |
| 300 |
| 301 [autofit] Extend `AF_FaceGlobalsRec' to hold emboldening data. |
| 302 |
| 303 * src/autofit/afglobal.h (AF_FaceGlobalsRec): Add fields. |
| 304 |
| 305 * src/autofit/afglobal.c (af_face_globals_new): Initialize new |
| 306 fields. |
| 307 (af_face_globals_free): Reset new fields. |
| 308 |
| 309 2015-11-02 Nikolaus Waxweiler <madigens@gmail.com> |
| 310 |
| 311 [autofit] Add stem-darkening properties. |
| 312 |
| 313 Actual code follows in a later commit. |
| 314 |
| 315 * include/freetype/ftautoh.h: Document `no-stem-darkening' and |
| 316 `darkening-parameters'. |
| 317 |
| 318 * src/autofit/afmodule.h: New fields `no_stem_darkening' and |
| 319 `darken_params'. |
| 320 |
| 321 * src/autofit/afmodule.c (af_property_set, af_property_get): |
| 322 Handle them. |
| 323 (af_autofitter_init): Initialize them. |
| 324 |
| 325 2015-11-02 Bungeman <bungeman@gmail.com> |
| 326 |
| 327 [ftfuzzer] Add support for multiple files (patch #8779). |
| 328 |
| 329 Currently, libFuzzer only supports mutation of a single file. We |
| 330 circumvent this problem by using an uncompressed tar archive as |
| 331 multiple-file input for the fuzzer. |
| 332 |
| 333 This patch enables tests of `FT_Attach_Stream' and AFM/PFM parsing; |
| 334 a constructed tarball should contain a font file as the first |
| 335 element, and files to be attached as further elements. |
| 336 |
| 337 * src/tools/ftfuzzer/ftfuzzer.cc: Include libarchive headers. |
| 338 (archive_read_entry_data, parse_data): New functions. |
| 339 (LLVMFuzzerTestOneInput): Updated. |
| 340 |
| 341 * src/tools/ftfuzzer/ftmutator.cc: New file, providing a custom |
| 342 mutator for libFuzzer that can mutate tarballs in a sensible way. |
| 343 |
| 344 2015-10-31 Werner Lemberg <wl@gnu.org> |
| 345 |
| 346 [sfnt] Fix cmap 14 validation (#46346). |
| 347 |
| 348 * src/sfnt/ttcmap.c (tt_cmap14_validate): Check limit before |
| 349 accessing `numRanges' and `numMappings'. |
| 350 Fix size check for non-default UVS table. |
| 351 |
| 352 2015-10-31 Werner Lemberg <wl@gnu.org> |
| 353 |
| 354 [sfnt] Handle infinite recursion in bitmap strikes (#46344). |
| 355 |
| 356 * src/sfnt/ttsbit.c (TT_SBitDecoder_LoadFunc, |
| 357 tt_sbit_decoder_load_bitmap, tt_sbit_decoder_load_byte_aligned, |
| 358 tt_sbit_decoder_load_bit_aligned, tt_sbit_decoder_load_png): Add |
| 359 argument for recursion depth. |
| 360 (tt_sbit_decoder_load_compound): Add argument for recursion depth. |
| 361 Increase recursion counter for recursive call. |
| 362 (tt_sbit_decoder_load_image): Add argument for recursion depth. |
| 363 Check recurse depth. |
| 364 (tt_face_load_sbit_image): Updated. |
| 365 |
| 366 2015-10-29 Werner Lemberg <wl@gnu.org> |
| 367 |
| 368 * src/autofit/afhints.c (af_glyph_hints_dump_points): Minor. |
| 369 |
| 370 2015-10-29 Werner Lemberg <wl@gnu.org> |
| 371 |
| 372 * CMakeLists.txt: Remove code to set MSVC's /FD compiler switch. |
| 373 |
| 374 Problem reported by David Capello <davidcapello@gmail.com>; see |
| 375 |
| 376 http://lists.nongnu.org/archive/html/freetype-devel/2015-10/msg00108.h
tml |
| 377 |
| 378 for details. |
| 379 |
| 380 2015-10-27 Werner Lemberg <wl@gnu.org> |
| 381 |
| 382 [pfr] Add some safety guards (#46302). |
| 383 |
| 384 * src/pfr/pfrload.h (PFR_CHECK): Rename to... |
| 385 (PFR_CHECK_SIZE): ... this. |
| 386 (PFR_SIZE): [!PFR_CONFIG_NO_CHECKS]: Define to PFR_CHECK_SIZE. |
| 387 |
| 388 * src/pfr/pfrload.c (pfr_log_font_count): Check `count'. |
| 389 (pfr_extra_item_load_kerning_pairs): Remove tracing message. |
| 390 (pfr_phy_font_load): Use PFR_CHECK_SIZE where appropriate. |
| 391 Allocate `chars' after doing a size checks. |
| 392 |
| 393 * src/pfr/pfrsbit.c (pfr_load_bitmap_bits): Move test for invalid |
| 394 bitmap format to... |
| 395 (pfr_slot_load_bitmap): ... this function. |
| 396 Check bitmap size. |
| 397 |
| 398 2015-10-26 Werner Lemberg <wl@gnu.org> |
| 399 |
| 400 [truetype] Fix sanitizing logic for `loca' (#46223). |
| 401 |
| 402 * src/truetype/ttpload.c (tt_face_load_loca): A thinko caused an |
| 403 incorrect adjustment of the number of glyphs, most often using far |
| 404 too large values. |
| 405 |
| 406 2015-10-25 Werner Lemberg <wl@gnu.org> |
| 407 |
| 408 [autofit] Improve tracing. |
| 409 |
| 410 * src/autofit/afhints.c (af_print_idx, af_get_segment_index, |
| 411 af_get_edge_index): New functions. |
| 412 |
| 413 (af_glyph_hints_dump_points): Remove unnecessary `|', `[', and `]'. |
| 414 Add segment and edge index for each point. |
| 415 Slightly change printing order of some elements. |
| 416 Don't print `-1' but `--' for missing elements. |
| 417 |
| 418 (af_glyph_hints_dump_segments, af_glyph_hints_dump_edges): Remove |
| 419 unnecessary `|', `[', and `]'. |
| 420 Don't print `-1' but `--' for missing elements. |
| 421 |
| 422 2015-10-24 Werner Lemberg <wl@gnu.org> |
| 423 |
| 424 [sfnt] Sanitize bitmap strike glyph height. |
| 425 |
| 426 Problem reported by Nikolay Sivov <bunglehead@gmail.com>. |
| 427 |
| 428 * src/sfnt/ttsbit.c (tt_face_load_strike_metrics): Avoid zero value |
| 429 for `metrics->height' by applying some heuristics. |
| 430 |
| 431 2015-10-22 Werner Lemberg <wl@gnu.org> |
| 432 |
| 433 [sfnt, type42] Fix clang compiler warnings. |
| 434 |
| 435 * src/sfnt/sfobjs.c (sfnt_init_face): Initialize `offset'. |
| 436 |
| 437 * src/type42/t42parse.c (t42_parse_sfnts): Use proper cast. |
| 438 |
| 439 2015-10-22 Dave Arnold <darnold@adobe.com> |
| 440 Werner Lemberg <wl@gnu.org> |
| 441 |
| 442 [cff] Avoid overflow/module arithmetic. |
| 443 |
| 444 This modifies the addition of subroutine number to subroutine bias |
| 445 from unsigned to signed, but does not change any results. |
| 446 |
| 447 * src/cff/cf2ft.c (cf2_initGlobalRegionBuffer, |
| 448 cf2_initLocalRegionBuffer): Change variable names from (unsigned) |
| 449 `idx' to (signed) `subrNum', since it is not an index until after |
| 450 the bias is added. |
| 451 * src/cff/cf2ft.h: Updated. |
| 452 |
| 453 * src/cff/cf2intrp.c (cf2_interpT2CharString) <cf2_cmdCALLSUBR>: |
| 454 Updated similarly. |
| 455 |
| 456 2015-10-22 Werner Lemberg <wl@gnu.org> |
| 457 |
| 458 [cid] Better check of `SubrCount' dictionary entry (#46272). |
| 459 |
| 460 * src/cid/cidload.c (cid_face_open): Add more sanity tests for |
| 461 `fd_bytes', `gd_bytes', `sd_bytes', and `num_subrs'. |
| 462 |
| 463 2015-10-21 Werner Lemberg <wl@gnu.org> |
| 464 |
| 465 [base] Pacify compiler (#46266). |
| 466 |
| 467 * src/base/ftoutln.c (FT_Outline_EmboldenXY): Initialize `in' and |
| 468 `anchor'. |
| 469 |
| 470 2015-10-21 Werner Lemberg <wl@gnu.org> |
| 471 |
| 472 [type42] Fix heap buffer overflow (#46269). |
| 473 |
| 474 * src/type42/t42parse.c (t42_parse_sfnts): Fix off-by-one error in |
| 475 bounds checking. |
| 476 |
| 477 2015-10-21 Dave Arnold <darnold@adobe.com> |
| 478 |
| 479 [cff] Fix limit in assert for max hints. |
| 480 |
| 481 * src/cff/cf2interp.c (cf2_hintmask_setAll): Allow mask equal to the |
| 482 limit (96 bits). |
| 483 |
| 484 2015-10-21 Dave Arnold <darnold@adobe.com> |
| 485 |
| 486 [cff] Remove an assert (#46107). |
| 487 |
| 488 * src/cff/cf2hints.c (cf2_hintmap_insertHint): Ignore paired edges |
| 489 in wrong order. |
| 490 |
| 491 2015-10-21 Werner Lemberg <wl@gnu.org> |
| 492 |
| 493 [sfnt] Avoid unnecessarily large allocation for WOFFs (#46257). |
| 494 |
| 495 * src/sfnt/sfobjs.c (woff_open_font): Use WOFF's `totalSfntSize' |
| 496 only after thorough checks. |
| 497 Add tracing messages. |
| 498 |
| 499 2015-10-21 Werner Lemberg <wl@gnu.org> |
| 500 |
| 501 [type42] Better check invalid `sfnts' array data (#46255). |
| 502 |
| 503 * src/type42/t42parse.c (t42_parse_sfnts): Table lengths must be |
| 504 checked individually against available data size. |
| 505 |
| 506 2015-10-20 Werner Lemberg <wl@gnu.org> |
| 507 |
| 508 [cid] Add a bunch of safety checks. |
| 509 |
| 510 * src/cid/cidload.c (parse_fd_array): Check `num_dicts' against |
| 511 stream size. |
| 512 (cid_read_subrs): Check largest offset against stream size. |
| 513 (cid_parse_dict): Move safety check to ... |
| 514 (cid_face_open): ... this function. |
| 515 Also test length of binary data and values of `SDBytes', |
| 516 `SubrMapOffset', `SubrCount', `CIDMapOffset', and `CIDCount'. |
| 517 |
| 518 2015-10-20 Werner Lemberg <wl@gnu.org> |
| 519 |
| 520 [cid] Avoid segfault with malformed input (#46250). |
| 521 |
| 522 * src/cid/cidload.c (cid_read_subrs): Return a proper error code for |
| 523 unsorted offsets. |
| 524 |
| 525 2015-10-20 StudioEtrange <nomorgan@gmail.com> |
| 526 |
| 527 * CMakeLists.txt: Enable shared library builds on MinGW (#46233). |
| 528 |
| 529 2015-10-20 Werner Lemberg <wl@gnu.org> |
| 530 |
| 531 * src/type1/t1afm.c (T1_Read_Metrics): Fix memory leak (#46229). |
| 532 |
| 533 2015-10-19 Bungeman <bungeman@gmail.com> |
| 534 |
| 535 [cid] Better handle invalid glyph stream offsets (#46221). |
| 536 |
| 537 * src/cid/cidgload.c (cid_load_glyph): Check minimum size of glyph |
| 538 length. |
| 539 |
| 540 2015-10-18 Werner Lemberg <wl@gnu.org> |
| 541 |
| 542 [psaux] Fix tracing of negative numbers. |
| 543 |
| 544 Due to incorrect casting negative numbers were shown as very large |
| 545 (positive) integers on 64bit systems. |
| 546 |
| 547 * src/psaux/t1decode.c (t1_decoder_parse_charstrings) <op_none>: |
| 548 Use division instead of shift. |
| 549 |
| 550 2015-10-18 Werner Lemberg <wl@gnu.org> |
| 551 |
| 552 [truetype] Improve TT_CONFIG_OPTION_MAX_RUNNABLE_OPCODES (#46223). |
| 553 |
| 554 * devel/ftoption.h, include/freetype/config/ftoption.h: Surround it |
| 555 with #ifndef ... #endif, as suggested in the tracker issue. |
| 556 |
| 557 2015-10-18 Werner Lemberg <wl@gnu.org> |
| 558 |
| 559 [truetype] Better protection against malformed `fpgm' (#46223). |
| 560 |
| 561 * src/truetype/ttobjs.c (tt_size_init_bytecode): Don't execute a |
| 562 malformed `fpgm' table more than once. |
| 563 |
| 564 2015-10-17 Werner Lemberg <wl@gnu.org> |
| 565 |
| 566 * src/cid/cidgload.c (cid_load_glyph): Fix memory leak. |
| 567 |
| 568 Reported by Kostya Serebryany <kcc@google.com>. |
| 569 |
| 570 2015-10-17 Werner Lemberg <wl@gnu.org> |
| 571 |
| 572 [bdf] Prevent memory leak (#46217). |
| 573 |
| 574 * src/bdf/bdflib.c (_bdf_parse_glyphs) <STARTCHAR>: Check |
| 575 _BDF_GLYPH_BITS. |
| 576 |
| 577 2015-10-17 Werner Lemberg <wl@gnu.org> |
| 578 |
| 579 [bdf] Use stream size to adjust number of glyphs. |
| 580 |
| 581 * src/bdf/bdflib.c (ACMSG17): New message macro. |
| 582 (_bdf_parse_t): Add member `size'. |
| 583 (bdf_load_font): Set `size'. |
| 584 (_bdf_parse_glyphs): Adjust `cnt' if necessary. |
| 585 |
| 586 2015-10-17 Werner Lemberg <wl@gnu.org> |
| 587 |
| 588 * src/cid/cidload.c (cid_parse_dict): Check `[FG]DBytes' size. |
| 589 |
| 590 2015-10-17 Werner Lemberg <wl@gnu.org> |
| 591 |
| 592 * src/cid/cidgload.c (cid_glyph_load): Check file offsets (#46222). |
| 593 |
| 594 2015-10-17 Werner Lemberg <wl@gnu.org> |
| 595 |
| 596 [psaux] Fix heap buffer overflow (#46221). |
| 597 |
| 598 * src/psaux/t1decode.c (t1_decoder_parse_charstring) <operator 12>: |
| 599 Fix limit check. |
| 600 |
| 601 2015-10-17 Werner Lemberg <wl@gnu.org> |
| 602 |
| 603 * src/cid/cidload.c (cid_parse_dict): Handle invalid input (#46220). |
| 604 |
| 605 2015-10-15 Kostya Serebryany <kcc@google.com> |
| 606 |
| 607 [ftfuzzer] Add README. |
| 608 |
| 609 * src/tools/ftfuzzer/README: New file. |
| 610 |
| 611 2015-10-15 Bungeman <bungeman@gmail.com> |
| 612 |
| 613 [bdf] Fix memory leak (#46213). |
| 614 |
| 615 * src/bdf/bdflib.c (bdf_load_font): Always go to label `Fail' in |
| 616 case of error. |
| 617 |
| 618 2015-10-15 Werner Lemberg <wl@gnu.org> |
| 619 |
| 620 [truetype] Add TT_CONFIG_OPTION_MAX_RUNNABLE_OPCODES (#46208). |
| 621 |
| 622 * devel/ftoption.h, include/freetype/config/ftoption.h |
| 623 (TT_CONFIG_OPTION_MAX_RUNNABLE_OPCODES): New configuration macro. |
| 624 |
| 625 * src/truetype/ttinterp.c (MAX_RUNNABLE_OPCODES): Removed. |
| 626 (TT_RunIns): Updated. |
| 627 |
| 628 2015-10-15 Werner Lemberg <wl@gnu.org> |
| 629 |
| 630 * src/truetype/ttinterp.c (TT_RunIns): Fix bytecode stack tracing. |
| 631 |
| 632 The used indices were off by 1. |
| 633 |
| 634 2015-10-15 Bungeman <bungeman@gmail.com> |
| 635 Werner Lemberg <wl@gnu.org> |
| 636 |
| 637 * src/tools/ftfuzzer/ftfuzzer.cc: Handle fixed sizes (#46211). |
| 638 |
| 639 2015-10-15 Werner Lemberg <wl@gnu.org> |
| 640 |
| 641 [base] Compute MD5 checksums only if explicitly requested. |
| 642 |
| 643 This improves profiling accuracy. |
| 644 |
| 645 * src/base/ftobjs.c (FT_Render_Glyph_Internal): Implement it. |
| 646 |
| 647 2015-10-14 Werner Lemberg <wl@gnu.org> |
| 648 |
| 649 [base] Use `FT_' namespace for MD5 functions (#42366). |
| 650 |
| 651 * src/base/ftobjs.c (MD5_*): Define as `FT_MD5_*'. |
| 652 Undefine HAVE_OPENSSL. |
| 653 |
| 654 2015-10-13 Werner Lemberg <wl@gnu.org> |
| 655 |
| 656 [type1] Correctly handle missing MM axis names (#46202). |
| 657 |
| 658 * src/type1/t1load.c (T1_Get_MM_Var): Implement it. |
| 659 |
| 660 2015-10-13 Werner Lemberg <wl@gnu.org> |
| 661 |
| 662 [pcf] Quickly exit if font index < 0. |
| 663 |
| 664 Similar to other font formats, this commit makes the parser no |
| 665 longer check the whole PCF file but only the header and the TOC if |
| 666 we just want to get the number of available faces (and a proper |
| 667 recognition of the font format). |
| 668 |
| 669 * src/pcf/pcfdrivr.c (PCF_Face_Init): Updated. |
| 670 Exit quickly if face_index < 0. |
| 671 |
| 672 * src/pcfread.c (pcf_load_font): Add `face_index' argument. |
| 673 Exit quickly if face_index < 0. |
| 674 |
| 675 * src/pcf/pcf.h: Updated. |
| 676 |
| 677 2015-10-13 Werner Lemberg <wl@gnu.org> |
| 678 |
| 679 [ftfuzzer] Handle TTCs and MM/GX variations. |
| 680 |
| 681 This patch also contains various other improvements. |
| 682 |
| 683 * src/tools/ftfuzzer/ftfuzzer.cc: Add preprocessor guard to reject |
| 684 pre-C++11 compilers. |
| 685 (FT_Global): New class. Use it to provide a global constructor and |
| 686 destructor for the `FT_Library' object. |
| 687 (setIntermediateAxis): New function to select an (arbitrary) |
| 688 instance. |
| 689 (LLVMFuzzerTestOneInput): Loop over all faces and named instances. |
| 690 Also call `FT_Set_Char_Size'. |
| 691 |
| 692 2015-10-13 Werner Lemberg <wl@gnu.org> |
| 693 |
| 694 [truetype] Refine some GX sanity tests. |
| 695 |
| 696 Use the `gvar' table size instead of the remaining bytes in the |
| 697 stream. |
| 698 |
| 699 * src/truetype/ttgxvar.h (GX_BlendRec): New field `gvar_size'. |
| 700 |
| 701 * src/truetype/ttgxvar.c (ft_var_load_gvar): Set `gvar_size'. |
| 702 (ft_var_readpackedpoints, ft_var_readpackeddeltas: New argument |
| 703 `size'. |
| 704 (tt_face_vary_cvt, TT_Vary_Apply_Glyph_Deltas): Updated. |
| 705 |
| 706 2015-10-13 Werner Lemberg <wl@gnu.org> |
| 707 |
| 708 [truetype] Another GX sanity test. |
| 709 |
| 710 * src/truetype/ttgxvar.c (TT_Vary_Apply_Glyph_Deltas): Check |
| 711 `tupleCount'. |
| 712 Add tracing message. |
| 713 |
| 714 2015-10-13 Werner Lemberg <wl@gnu.org> |
| 715 |
| 716 [truetype] Fix memory leak for broken GX fonts (#46188). |
| 717 |
| 718 * src/truetype/ttgxvar.c (TT_Vary_Apply_Glyph_Deltas): Fix scope of |
| 719 deallocation. |
| 720 |
| 721 2015-10-13 Werner Lemberg <wl@gnu.org> |
| 722 |
| 723 [truetype] Fix commit from 2015-10-10. |
| 724 |
| 725 * src/truetype/ttgxvar.c (ft_var_load_gvar): Add missing error |
| 726 handling body to condition. |
| 727 |
| 728 2015-10-12 Werner Lemberg <wl@gnu.org> |
| 729 |
| 730 [unix] Make MKDIR_P actually work. |
| 731 |
| 732 * builds/unix/configure.raw: Fix underquoting of `INSTALL' and |
| 733 `MKDIR_P'. |
| 734 |
| 735 Problem reported by Dan Liddell <lddll@yahoo.com>. |
| 736 |
| 737 2015-10-11 Werner Lemberg <wl@gnu.org> |
| 738 |
| 739 [sfnt] Improve extraction of number of named instances. |
| 740 |
| 741 * src/sfnt/sfobjs.c (sfnt_init_face) |
| 742 [TT_CONFIG_OPTION_GX_VAR_SUPPORT]: Check number of instances against |
| 743 `fvar' table size. |
| 744 |
| 745 2015-10-10 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 746 |
| 747 * src/base/ftoutln.c (FT_Outline_Get_Orientation): Fix overflow |
| 748 (#46149). |
| 749 |
| 750 2015-10-10 Werner Lemberg <wl@gnu.org> |
| 751 |
| 752 [sfnt] Fix infinite loops with broken cmaps (#46167). |
| 753 |
| 754 * src/sfnt/ttcmap.c (tt_cmap8_char_next, tt_cmap12_next): Take care |
| 755 of border conditions (i.e., if the loops exit naturally). |
| 756 |
| 757 2015-10-10 Werner Lemberg <wl@gnu.org> |
| 758 |
| 759 [truetype] More sanity tests for GX handling. |
| 760 |
| 761 These tests should mainly help avoid unnecessarily large memory |
| 762 allocations in case of malformed fonts. |
| 763 |
| 764 * src/truetype/ttgxvar.c (ft_var_readpackedpoints, |
| 765 ft_var_readpackeddeltas): Check number of points against stream |
| 766 size. |
| 767 (ft_var_load_avar): Check `pairCount' against table length. |
| 768 (ft_var_load_gvar): Check `globalCoordCount' and `glyphCount' |
| 769 against table length. |
| 770 (tt_face_vary_cvt): Check `tupleCount' and `offsetToData'. |
| 771 Fix trace. |
| 772 (TT_Vary_Apply_Glyph_Deltas): Fix trace. |
| 773 Free `sharedpoints' to avoid memory leak. |
| 774 |
| 775 2015-10-10 Werner Lemberg <wl@gnu.org> |
| 776 |
| 777 [truetype] Better protection against malformed GX data (#46166). |
| 778 |
| 779 * src/truetype/ttgxvar.c (TT_Vary_Apply_Glyph_Deltas): Correctly |
| 780 handle empty `localpoints' array. |
| 781 |
| 782 2015-10-10 Werner Lemberg <wl@gnu.org> |
| 783 |
| 784 * src/pcf/pcfread.c (pcf_read_TOC): Check stream size (#46162). |
| 785 |
| 786 2015-10-09 Werner Lemberg <wl@gnu.org> |
| 787 |
| 788 * src/gzip/ftgzip.c (FT_Stream_OpenGzip): Use real stream size. |
| 789 |
| 790 2015-10-08 Werner Lemberg <wl@gnu.org> |
| 791 |
| 792 [pcf] Protect against invalid number of TOC entries (#46159). |
| 793 |
| 794 * src/pcf/pcfread.c (pcf_read_TOC): Check number of TOC entries |
| 795 against size of data stream. |
| 796 |
| 797 2015-10-08 Werner Lemberg <wl@gnu.org> |
| 798 |
| 799 [type42] Protect against invalid number of glyphs (#46159). |
| 800 |
| 801 * src/type42/t42parse.c (t42_parse_charstrings): Check number of |
| 802 `CharStrings' dictionary entries against size of data stream. |
| 803 |
| 804 2015-10-08 Werner Lemberg <wl@gnu.org> |
| 805 |
| 806 [sfnt] Fix some signed overflows (#46149). |
| 807 |
| 808 * src/sfnt/ttsbit.c (tt_face_load_strike_metrics) |
| 809 <TT_SBIT_TABLE_TYPE_SBIX>: Use `FT_MulDiv'. |
| 810 |
| 811 2015-10-08 Werner Lemberg <wl@gnu.org> |
| 812 |
| 813 [type1] Protect against invalid number of subroutines (#46150). |
| 814 |
| 815 * src/type1/t1load.c (parse_subrs): Check number of |
| 816 `Subrs' dictionary entries against size of data stream. |
| 817 |
| 818 2015-10-07 Kostya Serebryany <kcc@google.com> |
| 819 |
| 820 [ftfuzzer] Add support for LLVM's LibFuzzer. |
| 821 |
| 822 * src/tools/ftfuzzer/ftfuzzer.cc, src/tools/runinput.cc: New files. |
| 823 |
| 824 2015-10-06 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 825 |
| 826 [smooth] Faster alternative line renderer. |
| 827 |
| 828 This implementation renders the entire line segment at once without |
| 829 subdividing it into scanlines. The main speed improvement comes from |
| 830 reducing the number of divisions to just two per line segment, which |
| 831 is a bare minimum to calculate cell coverage in a smooth rasterizer. |
| 832 Notably, the progression from cell to cell does not itself require any |
| 833 divisions at all. The speed improvement is more noticeable at larger |
| 834 sizes. |
| 835 |
| 836 * src/smooth/ftgrays.c (gray_render_line): New implementation. |
| 837 |
| 838 2015-10-06 Werner Lemberg <wl@gnu.org> |
| 839 |
| 840 [cff] Return correct PS names from pure CFF (#46130). |
| 841 |
| 842 * src/cff/cffdrivr.c (cff_get_ps_name): Use SFNT service only for |
| 843 SFNT. |
| 844 |
| 845 2015-10-04 Werner Lemberg <wl@gnu.org> |
| 846 |
| 847 [base] Replace left shifts with multiplication (#46118). |
| 848 |
| 849 * src/base/ftglyph.c (ft_bitmap_glyph_bbox, FT_Get_Glyph): Do it. |
| 850 |
| 851 2015-10-04 Werner Lemberg <wl@gnu.org> |
| 852 |
| 853 * Version 2.6.1 released. |
| 854 ========================= |
| 855 |
| 856 |
| 857 Tag sources with `VER-2-6-1'. |
| 858 |
| 859 * docs/VERSION.DLL: Update documentation and bump version number to |
| 860 2.6.1. |
| 861 |
| 862 * README, Jamfile (RefDoc), builds/windows/vc2005/freetype.vcproj, |
| 863 builds/windows/vc2005/index.html, |
| 864 builds/windows/vc2008/freetype.vcproj, |
| 865 builds/windows/vc2008/index.html, |
| 866 builds/windows/vc2010/freetype.vcxproj, |
| 867 builds/windows/vc2010/index.html, |
| 868 builds/windows/visualc/freetype.dsp, |
| 869 builds/windows/visualc/freetype.vcproj, |
| 870 builds/windows/visualc/index.html, |
| 871 builds/windows/visualce/freetype.dsp, |
| 872 builds/windows/visualce/freetype.vcproj, |
| 873 builds/windows/visualce/index.html, |
| 874 builds/wince/vc2005-ce/freetype.vcproj, |
| 875 builds/wince/vc2005-ce/index.html, |
| 876 builds/wince/vc2008-ce/freetype.vcproj, |
| 877 builds/wince/vc2008-ce/index.html: s/2.6/2.6.1/, s/26/261/. |
| 878 |
| 879 * include/freetype/freetype.h (FREETYPE_PATCH): Set to 1. |
| 880 |
| 881 * builds/unix/configure.raw (version_info): Set to 18:1:12. |
| 882 * CMakeLists.txt (VERSION_PATCH): Set to 1. |
| 883 |
| 884 * src/autofit/afmodule.c [AF_DEBUG_AUTOFIT]: Ensure C linking for |
| 885 dumping functions. |
| 886 |
| 887 2015-10-04 Werner Lemberg <wl@gnu.org> |
| 888 |
| 889 [bzip2, gzip] Avoid access of unitialized memory (#46109). |
| 890 |
| 891 * src/bzip2/ftbzip2.c (ft_bzip2_file_fill_input), src/gzip/ftgzip.c |
| 892 (ft_gzip_file_fill_input): In case of an error, adjust the limit to |
| 893 avoid copying uninitialized memory. |
| 894 |
| 895 2015-10-03 Werner Lemberg <wl@gnu.org> |
| 896 |
| 897 [bzip2, gzip] Avoid access of unitialized memory (#46109). |
| 898 |
| 899 * src/bzip2/ftbzip2.c (ft_bzip2_file_fill_output), src/gzip/ftgzip.c |
| 900 (ft_gzip_file_fill_output): In case of an error, adjust the limit to |
| 901 avoid copying uninitialized memory. |
| 902 |
| 903 2015-10-01 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 904 |
| 905 [smooth] Clean up worker. |
| 906 |
| 907 * src/smooth/ftgrays.c (gray_TWorker): Remove never used fields. |
| 908 |
| 909 2015-10-01 Werner Lemberg <wl@gnu.org> |
| 910 |
| 911 [sfnt] Make `tt_cmap4_char_map_linear' more robust (#46078). |
| 912 |
| 913 * src/sfnt/ttcmap.c (tt_cmap4_char_map_linear): Take care of |
| 914 border conditions (i.e., if the loop exits naturally). |
| 915 |
| 916 2015-10-01 Werner Lemberg <wl@gnu.org> |
| 917 |
| 918 * src/autofit/afranges.c (af_deva_nonbase_uniranges): Fix ranges. |
| 919 They should be a subset of `af_deva_uniranges'. |
| 920 |
| 921 2015-10-01 Werner Lemberg <wl@gnu.org> |
| 922 |
| 923 [sfnt] Make `tt_cmap4_char_map_linear' faster (#46078). |
| 924 |
| 925 * src/sfnt/ttcmap.c (tt_cmap4_char_map_linear): Use inner loop to |
| 926 reject too large glyph indices. |
| 927 |
| 928 2015-09-30 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 929 |
| 930 [smooth] Clean up worker. |
| 931 |
| 932 * src/smooth/ftgrays.c (gray_TWorker): Remove lightly used `last_ey'. |
| 933 (gray_start_cell, gray_render_line): Update. |
| 934 |
| 935 2015-09-30 Werner Lemberg <wl@gnu.org> |
| 936 |
| 937 [autofit] Replace `no-base' with `non-base'. |
| 938 |
| 939 * src/autofit/*: Do it. |
| 940 |
| 941 2015-09-30 Werner Lemberg <wl@gnu.org> |
| 942 |
| 943 [sfnt] Rewrite `tt_cmap4_char_map_linear' (#46078). |
| 944 |
| 945 * src/sfnt/ttcmap.c (tt_cmap4_char_map_linear): Add code to better |
| 946 skip invalid segments. |
| 947 If searching the next character, provide a more efficient logic to |
| 948 speed up the code. |
| 949 |
| 950 2015-09-30 Werner Lemberg <wl@gnu.org> |
| 951 |
| 952 [truetype] Adjust number of glyphs for malformed `loca' tables. |
| 953 |
| 954 * src/truetype/ttpload.c (tt_face_load_loca): Implement it. |
| 955 |
| 956 2015-09-29 Werner Lemberg <wl@gnu.org> |
| 957 |
| 958 [pshinter] Avoid harmless overflow (#45984). |
| 959 |
| 960 * src/pshinter/pshglob.c (psh_blues_set_zones): Fix it. |
| 961 |
| 962 2015-09-28 Werner Lemberg <wl@gnu.org> |
| 963 |
| 964 [autofit] Add support for Lao script. |
| 965 |
| 966 Thanks to Danh Hong <danhhong@gmail.com> for guidance with blue zone |
| 967 characters! |
| 968 |
| 969 * src/autofit/afblue.dat: Add blue zone data for Lao. |
| 970 |
| 971 * src/autofit/afblue.c, src/autofit/afblue.h: Regenerated. |
| 972 |
| 973 * src/autofit/afscript.h: Add Lao standard characters. |
| 974 |
| 975 * src/autofit/afranges.c: Add Lao data. |
| 976 |
| 977 * src/autofit/afstyles.h: Add Lao data. |
| 978 |
| 979 2015-09-27 suzuki toshiya <mpsuzuki@hiroshima-u.ac.jp> |
| 980 |
| 981 [base] Fix a leak by broken sfnt-PS or resource fork (#46028). |
| 982 |
| 983 open_face_from_buffer() frees passed buffer if valid font |
| 984 is not found. But if copying to the buffer is failed, |
| 985 the allocated buffer should be freed within the caller. |
| 986 |
| 987 * src/base/ftobjs.c (open_face_PS_from_sfnt_stream): Free |
| 988 the buffer `sfnt_ps' if an error caused before calling |
| 989 open_face_from_buffer(). |
| 990 (Mac_Read_sfnt_Resource): Free the buffer `sfnt_data' if |
| 991 an error caused before calling open_face_from_buffer(); |
| 992 |
| 993 2015-09-27 suzuki toshiya <mpsuzuki@hiroshima-u.ac.jp> |
| 994 |
| 995 [mac] Fix buffer size calculation for LWFN font. |
| 996 |
| 997 * src/base/ftmac.c (read_lwfn): Cast post_size to FT_ULong |
| 998 to prevent confused copy by too large chunk size. |
| 999 |
| 1000 2015-09-26 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 1001 |
| 1002 * src/smooth/ftgrays.c (PIXEL_MASK): Remove unused macro. |
| 1003 |
| 1004 2015-09-26 Werner Lemberg <wl@gnu.org> |
| 1005 |
| 1006 [autofit] Minor tracing improvement. |
| 1007 |
| 1008 * src/autofit/aflatin.c (af_latin_metrics_scale_dim): Don't emit |
| 1009 blue zones header line if there are no blue zones. |
| 1010 |
| 1011 2015-09-26 Werner Lemberg <wl@gnu.org> |
| 1012 |
| 1013 [bzip2, gzip, lzw] Harmonize function signatures with prototype. |
| 1014 |
| 1015 Suggested by Hin-Tak Leung. |
| 1016 |
| 1017 * src/bzip2/ftbzip2.c (ft_bzip2_stream_io), src/gzip/ftgzip.c |
| 1018 (ft_gzip_stream_io), src/lzw/ftlzw.c (ft_lzw_stream_io): Do it. |
| 1019 |
| 1020 2015-09-26 Hin-Tak Leung <htl10@users.sourceforge.net> |
| 1021 |
| 1022 Add new FT_LOAD_COMPUTE_METRICS load flag. |
| 1023 |
| 1024 * include/freetype/freetype.h (FT_LOAD_COMPUTE_METRICS): New macro. |
| 1025 * src/truetype/ttgload.c (compute_glyph_metrics): Usage. |
| 1026 |
| 1027 2015-09-26 Werner Lemberg <wl@gnu.org> |
| 1028 |
| 1029 * src/base/ftobjs.c (Mac_Read_sfnt_Resource): Add cast. |
| 1030 |
| 1031 2015-09-25 Werner Lemberg <wl@gnu.org> |
| 1032 |
| 1033 [type1] Protect against invalid number of glyphs (#46029). |
| 1034 |
| 1035 * src/type1/t1load.c (parse_charstrings): Check number of |
| 1036 `CharStrings' dictionary entries against size of data stream. |
| 1037 |
| 1038 2015-09-23 Werner Lemberg <wl@gnu.org> |
| 1039 |
| 1040 [sfnt] Better checks for invalid cmaps (2/2) (#46019). |
| 1041 |
| 1042 While the current code in `FT_Get_Next_Char' correctly rejects |
| 1043 out-of-bounds glyph indices, it can be extremely slow for malformed |
| 1044 cmaps that use 32bit values. This commit tries to improve that. |
| 1045 |
| 1046 * src/sfnt/ttcmap.c (tt_cmap8_char_next, tt_cmap12_next, |
| 1047 tt_cmap12_char_map_binary, tt_cmap13_next, |
| 1048 tt_cmap13_char_map_binary): Reject glyph indices larger than or |
| 1049 equal to the number of glyphs. |
| 1050 |
| 1051 2015-09-23 Werner Lemberg <wl@gnu.org> |
| 1052 |
| 1053 [base, sfnt] Better checks for invalid cmaps (1/2). |
| 1054 |
| 1055 * src/base/ftobjs.c (FT_Get_Char_Index): Don't return out-of-bounds |
| 1056 glyph indices. |
| 1057 (FT_Get_First_Char): Updated. |
| 1058 |
| 1059 * src/sfnt/ttcmap.c (tt_cmap6_char_next): Don't return character |
| 1060 codes greater than 0xFFFF. |
| 1061 |
| 1062 (tt_cmap8_char_index): Avoid integer overflow in computation of |
| 1063 glyph index. |
| 1064 (tt_cmap8_char_next): Avoid integer overflows in computation of |
| 1065 both next character code and glyph index. |
| 1066 |
| 1067 (tt_cmap10_char_index): Fix unsigned integer logic. |
| 1068 (tt_cmap10_char_next): Avoid integer overflow in computation of |
| 1069 next character code. |
| 1070 |
| 1071 (tt_cmap12_next): Avoid integer overflows in computation of both |
| 1072 next character code and glyph index. |
| 1073 (tt_cmap12_char_map_binary): Ditto. |
| 1074 (tt_cmap12_char_next): Simplify. |
| 1075 |
| 1076 (tt_cmap13_char_map_binary): Avoid integer overflow in computation |
| 1077 of next character code. |
| 1078 (tt_cmap13_char_next): Simplify. |
| 1079 |
| 1080 2015-09-21 suzuki toshiya <mpsuzuki@hiroshima-u.ac.jp> |
| 1081 |
| 1082 [base] Check too long POST and sfnt resource (#45919). |
| 1083 |
| 1084 * src/base/ftbase.h (FT_MAC_RFORK_MAX_LEN): Maximum length of the |
| 1085 resource fork for Mac OS. Resource forks larger than 16 MB can be |
| 1086 written but can't be handled correctly, at least in Carbon routine. |
| 1087 See https://support.microsoft.com/en-us/kb/130437. |
| 1088 |
| 1089 * src/base/ftobjs.c (Mac_Read_POST_Resource): No need to use `0x' |
| 1090 prefix for `%p' formatter. |
| 1091 |
| 1092 * src/base/ftbase.c (Mac_Read_POST_Resource): Check the fragment and |
| 1093 total size of the concatenated POST resource before buffer |
| 1094 allocation. |
| 1095 (Mac_Read_sfnt_Resource): Check the declared size of sfnt resource |
| 1096 before buffer allocation. |
| 1097 |
| 1098 * src/base/ftmac.c (read_lwfn, FT_New_Face_From_SFNT): Check the |
| 1099 total resource size before buffer allocation. |
| 1100 |
| 1101 2015-09-19 Werner Lemberg <wl@gnu.org> |
| 1102 |
| 1103 [sfnt] Improve handling of invalid SFNT table entries (#45987). |
| 1104 |
| 1105 This patch fixes weaknesses in function `tt_face_load_font_dir'. |
| 1106 |
| 1107 - It incorrectly assumed that valid tables are always at the |
| 1108 beginning. As a consequence, some valid tables after invalid |
| 1109 entries (which are ignored) were never seen. |
| 1110 |
| 1111 - Duplicate table entries (this is, having the same tag) were not |
| 1112 rejected. |
| 1113 |
| 1114 - The number of valid tables was sometimes too large, leading to |
| 1115 access of invalid tables. |
| 1116 |
| 1117 * src/sfnt/ttload.c (check_table_dir): Add argument to return number |
| 1118 of valid tables. |
| 1119 Add another tracing message. |
| 1120 (tt_face_load_font_dir): Only allocate table array for valid |
| 1121 entries as returned by `check_table_dir'. |
| 1122 Reject duplicate tables and adjust number of valid tables |
| 1123 accordingly. |
| 1124 |
| 1125 2015-09-19 Werner Lemberg <wl@gnu.org> |
| 1126 |
| 1127 [pcf] Improve `FT_ABS' fix from 2015-09-17 (#45999). |
| 1128 |
| 1129 * src/pcf/pcfread.c (pcf_load_font): Do first the cast to FT_Short, |
| 1130 then take the absolute value. |
| 1131 Also apply FT_ABS to `height'. |
| 1132 |
| 1133 2015-09-17 Werner Lemberg <wl@gnu.org> |
| 1134 |
| 1135 [type42] Fix memory leak (#45989). |
| 1136 |
| 1137 * src/type42/t42parse.c (t42_parse_charstrings): Allow only a single |
| 1138 `CharStrings' array. |
| 1139 |
| 1140 2015-09-17 Werner Lemberg <wl@gnu.org> |
| 1141 |
| 1142 [psaux] Fix memory leak (#45986). |
| 1143 |
| 1144 * src/psaux/psobjs.c (ps_parser_load_field) <T1_FIELD_TYPE_MM_BBOX>: |
| 1145 Free `temp' in case of error. |
| 1146 |
| 1147 2015-09-17 Werner Lemberg <wl@gnu.org> |
| 1148 |
| 1149 [psaux] Improve tracing message. |
| 1150 |
| 1151 * src/psaux/psobjs.c (ps_parser_load_field) <T1_FIELD_TYPE_MM_BBOX>: |
| 1152 Handle plural correctly. |
| 1153 |
| 1154 2015-09-17 Werner Lemberg <wl@gnu.org> |
| 1155 |
| 1156 [pcf] Fix integer overflows (#45985). |
| 1157 |
| 1158 * src/pcf/pcfread.c (pcf_load_font): Use FT_MulDiv. |
| 1159 |
| 1160 2015-09-17 Werner Lemberg <wl@gnu.org> |
| 1161 |
| 1162 [pcf] Use FT_ABS for some property values (#45893). |
| 1163 |
| 1164 * src/pcf/pcfread.c (pcf_load_font): Take absolute values for |
| 1165 AVERAGE_WIDTH, POINT_SIZE, PIXEL_SIZE, RESOLUTION_X, and |
| 1166 RESOLUTION_Y. In tracing mode, add warnings. |
| 1167 |
| 1168 2015-09-16 Werner Lemberg <wl@gnu.org> |
| 1169 |
| 1170 Minor fixes for some clang warnings. |
| 1171 |
| 1172 * src/base/ftoutln.c (FT_Outline_EmboldenXY): Cast, possible missing |
| 1173 initialization. |
| 1174 |
| 1175 * src/truetype/ttgload.c (TT_Process_Composite_Component): Cast. |
| 1176 |
| 1177 2015-09-15 Werner Lemberg <wl@gnu.org> |
| 1178 |
| 1179 [type1, type42] Fix memory leaks (#45966). |
| 1180 |
| 1181 * src/type1/t1load.c (parse_blend_axis_types): Handle multiple axis |
| 1182 names. |
| 1183 (parse_blend_design_map): Allow only a single design map. |
| 1184 (parse_encoding): Handle multiple encoding vectors. |
| 1185 |
| 1186 * src/type42/t42parse.c (t42_parse_encoding): Handle multiple |
| 1187 encoding vectors. |
| 1188 |
| 1189 2015-09-15 Werner Lemberg <wl@gnu.org> |
| 1190 |
| 1191 [truetype] Fix integer type (#45965). |
| 1192 |
| 1193 * src/truetype/ttobjs.c (tt_synth_sfnt_checksum): Implement it. |
| 1194 |
| 1195 2015-09-15 Werner Lemberg <wl@gnu.org> |
| 1196 |
| 1197 * src/pcf/pcfread.c (pcf_load_font): Fix integer overflow (#45964). |
| 1198 |
| 1199 2015-09-15 Werner Lemberg <wl@gnu.org> |
| 1200 |
| 1201 [type1, type42] Check encoding array size (#45961). |
| 1202 |
| 1203 * src/type1/t1load.c (parse_encoding), src/type42/t42parse.c |
| 1204 (t42_parse_encoding): Do it. |
| 1205 |
| 1206 2015-09-14 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 1207 |
| 1208 * src/base/ftcalc.c (FT_MulFix) [FT_LONG64]: Improve. |
| 1209 |
| 1210 2015-09-14 Werner Lemberg <wl@gnu.org> |
| 1211 |
| 1212 [type1] Fix another potential buffer overflow (#45955). |
| 1213 |
| 1214 * src/type1/t1parse (T1_Get_Private_Dict): Assure that check for |
| 1215 `eexec' doesn't exceed `limit'. |
| 1216 |
| 1217 2015-09-13 Werner Lemberg <wl@gnu.org> |
| 1218 |
| 1219 Replace `mkinstalldirs' with AC_PROG_MKDIR_P. |
| 1220 |
| 1221 * builds/unix/mkinstalldirs: Removed, no longer needed. |
| 1222 |
| 1223 * builds/unix/configure.raw: Call `AC_PROG_MKDIR_P'. |
| 1224 Update pwd call for `$INSTALL'. |
| 1225 |
| 1226 * builds/unix/unix-def.in (MKINSTALLDIRS): Use `@MKDIR_P@'. |
| 1227 |
| 1228 * autogen.sh: Updated. |
| 1229 |
| 1230 2015-09-13 Werner Lemberg <wl@gnu.org> |
| 1231 |
| 1232 [winfonts] Check alignment shift count for resource data (#45938). |
| 1233 |
| 1234 * src/winfonts/winfnt.c (fnt_face_get_dll_font): Implement it. |
| 1235 |
| 1236 2015-09-13 Werner Lemberg <wl@gnu.org> |
| 1237 |
| 1238 [type1] Fix potential buffer overflow (#45923). |
| 1239 |
| 1240 * src/type1/t1parse.c (T1_Get_Private_Dict): Assure `cur' doesn't |
| 1241 point to end of file buffer. |
| 1242 |
| 1243 2015-09-13 Werner Lemberg <wl@gnu.org> |
| 1244 |
| 1245 [gzip] Fix access of small compressed files (#45937). |
| 1246 |
| 1247 * src/gzip/ftgzip.c (ft_gzip_stream_close): Avoid memory leak. |
| 1248 |
| 1249 (ft_gzip_get_uncompressed_file): Correct byte order while reading |
| 1250 unsigned long value. Without this change, the whole optimization of |
| 1251 accessing small files in `FT_Stream_OpenGzip' is never executed! As |
| 1252 a consequence, access to PCF files in general (which are normally |
| 1253 small files) should be much improved now as originally intended. |
| 1254 |
| 1255 2015-09-11 Werner Lemberg <wl@gnu.org> |
| 1256 |
| 1257 [psaux] Fix potential buffer overflow (#45922). |
| 1258 |
| 1259 * src/psaux/psobjs.c (ps_parser_skip_PS_token): If a token is |
| 1260 enclosed in balanced expressions, ensure that the cursor position |
| 1261 doesn't get larger than the current limit. |
| 1262 |
| 1263 2015-09-11 Werner Lemberg <wl@gnu.org> |
| 1264 |
| 1265 [base] Avoid crash while tracing `load_mac_face'. |
| 1266 |
| 1267 Reported in Savannah bug #45919. |
| 1268 |
| 1269 * src/base/ftobjs.c (load_mac_face): Honour FT_OPEN_MEMORY while |
| 1270 tracing. |
| 1271 |
| 1272 2015-09-11 Werner Lemberg <wl@gnu.org> |
| 1273 |
| 1274 [type42] Fix endless loop (#45920). |
| 1275 |
| 1276 * src/type42/t42parse.c (t42_parse_encoding): Synchronize with |
| 1277 type1's `parse_encoding'. |
| 1278 |
| 1279 2015-09-10 Werner Lemberg <wl@gnu.org> |
| 1280 |
| 1281 [docmaker] Allow `-' in bold and italic markup. |
| 1282 |
| 1283 * src/tools/docmaker/sources.py (re_italic, re_bold): Adjust |
| 1284 accordingly. |
| 1285 |
| 1286 2015-09-09 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 1287 |
| 1288 * src/base/ftcalc.c (FT_RoundFix): Improve. |
| 1289 |
| 1290 2015-09-09 Wojciech Mamrak <wmamrak@gmail.com> |
| 1291 |
| 1292 * src/base/ftcalc.c (FT_CeilFix, FT_FloorFix): Normalize. |
| 1293 |
| 1294 This commit makes the functions behave as expected, this is, |
| 1295 rounding towards plus or minus infinity. |
| 1296 |
| 1297 2015-09-07 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 1298 |
| 1299 * src/smooth/ftgrays.c (gray_render_line): Simplify clipping. |
| 1300 |
| 1301 2015-09-04 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 1302 |
| 1303 [raster,smooth] Microoptimizations. |
| 1304 |
| 1305 * src/raster/ftraster.c (Insert_Y_Turn, Finalize_Profile_Table, |
| 1306 Beziier_Up, ): Use do-while loops. |
| 1307 |
| 1308 * src/smooth/ftgrays.c (gray_render_scanline, gray_render_line, |
| 1309 gray_convert_glyph): Ditto. |
| 1310 |
| 1311 2015-09-04 Werner Lemberg <wl@gnu.org> |
| 1312 |
| 1313 [autofit] Redesign code ranges (2/2). |
| 1314 |
| 1315 This commit adds two fallback scripts (`latb', `latp') and |
| 1316 implements support for the no-base character ranges introduced in |
| 1317 the previous commit. |
| 1318 |
| 1319 * src/autofit/aftypes.h (AF_ScriptClassRec): Add |
| 1320 `script_uni_nobase_ranges' field. |
| 1321 (AF_DEFINE_SCRIPT_CLASS): Updated. |
| 1322 |
| 1323 * src/autofit/afscript.h, src/autofit/afstyles.h: Add `latb' and |
| 1324 `latp' fallback scripts. |
| 1325 |
| 1326 * src/autofit/afblue.dat: Add blue zones for Latin subscript and |
| 1327 superscript fallback scripts. |
| 1328 |
| 1329 * src/autofit/afblue.c, src/autofit/afblue.h: Regenerated. |
| 1330 |
| 1331 * src/autofit/afglobal.h (AF_NOBASE): New style flag for no-base |
| 1332 characters. |
| 1333 (AF_STYLE_MASK): Updated. |
| 1334 |
| 1335 * src/autofit/afglobal.c (SCRIPT): Updated. |
| 1336 (af_face_globals_compute_style_coverage): Handle new style flag. |
| 1337 |
| 1338 * src/autofit/aflatin.c (af_latin_hints_apply): Handle new style |
| 1339 flag. |
| 1340 |
| 1341 * src/autofit/afranges.h (SCRIPT): Use it to export no-base ranges. |
| 1342 |
| 1343 2015-09-04 Werner Lemberg <wl@gnu.org> |
| 1344 |
| 1345 [autofit] Redesign code ranges (1/2). |
| 1346 |
| 1347 This patch introduces auxiliary code ranges that identify no-base |
| 1348 characters; they refer to glyphs of a script that should be hinted |
| 1349 without alignments to blue zones (mostly diacritics). |
| 1350 |
| 1351 It also splits off ranges for fallback scripts that handle subscript |
| 1352 and superscript characters not covered by OpenType features. For |
| 1353 example, this greatly helps improve the hinting of various phonetic |
| 1354 alphabets, which contain a large amount characters that look like |
| 1355 superscript glyphs. |
| 1356 |
| 1357 Finally, code ranges are updated to Unicode 8.0, and enclosed |
| 1358 characters are removed in general since they normally look better if |
| 1359 they stay unhinted. |
| 1360 |
| 1361 * src/autofit/afranges.c (af_latn_uniranges): Updated to Unicode |
| 1362 8.0. |
| 1363 Split off superscript-like and subscript-like glyphs into... |
| 1364 |
| 1365 (af_latb_uniranges, af_latp_uniranges): ... these two new arrays. |
| 1366 |
| 1367 (af_xxxx_nobase_uniranges): New arrays that hold no-base characters |
| 1368 of the corresponding character ranges. |
| 1369 |
| 1370 2015-09-03 Werner Lemberg <wl@gnu.org> |
| 1371 |
| 1372 [autofit] Pass glyph index to hinting function. |
| 1373 |
| 1374 No functionality change yet. |
| 1375 |
| 1376 * src/autofit/aftypes.h (AF_WritingSystem_ApplyHintsFunc): Pass |
| 1377 glyph index. |
| 1378 |
| 1379 * src/autofit/afcjk.c, src/autofit/afcjk.h (af_cjk_hints_apply), |
| 1380 src/autofit/afdummy.c (af_dummy_hints_apply), src/autofit/afindic.c |
| 1381 (af_indic_hints_apply), src/autofit/aflatin.c |
| 1382 (af_latin_hints_apply), src/autofit/aflatin2.c |
| 1383 (af_latin2_hints_apply), src/autofit/afloader.c (af_loader_load_g): |
| 1384 Updated. |
| 1385 |
| 1386 2015-08-30 Werner Lemberg <wl@gnu.org> |
| 1387 |
| 1388 [autofit] Code clean-up. |
| 1389 |
| 1390 * src/autofit/afglobal.h (AF_STYLE_MASK): New macro. |
| 1391 (AF_STYLE_UNASSIGNED): Use AF_STYLE_MASK for definition. |
| 1392 |
| 1393 * src/autofit/afglobal.c (af_face_globals_compute_style_coverage): |
| 1394 Updated. |
| 1395 |
| 1396 2015-08-30 Werner Lemberg <wl@gnu.org> |
| 1397 |
| 1398 [autofit] Make glyph style array use 16bit values. |
| 1399 |
| 1400 * include/freetype/ftautoh.h (FT_Prop_GlyphToScriptMap): Use |
| 1401 `FT_UShort' for `map' field. |
| 1402 |
| 1403 * src/autofit/afglobal.c (af_face_globals_compute_style_coverage, |
| 1404 af_face_globals_new), src/autofit/hbshim.c, src/autofit/hbshim.h |
| 1405 (af_get_coverage): Use FT_UShort for `glyph_styles' array. |
| 1406 |
| 1407 * src/autofit/afglobal.h (AF_STYLE_UNASSIGNED, AF_DIGIT): Extend to |
| 1408 16 bits. |
| 1409 (AF_FaceGlobalsRec): Use `FT_UShort' for `glyph_styles' field. |
| 1410 |
| 1411 2015-08-26 Werner Lemberg <wl@gnu.org> |
| 1412 |
| 1413 * builds/unix/configure.raw: Need harfbuzz >= 0.9.21 (#45828). |
| 1414 |
| 1415 2015-08-25 Werner Lemberg <wl@gnu.org> |
| 1416 |
| 1417 [base] Improve kerning tracing and documentation. |
| 1418 |
| 1419 * src/base/ftobjs.c (FT_Get_Kerning): Emit tracing message if |
| 1420 scaled-down kerning values differ. |
| 1421 |
| 1422 2015-08-18 Werner Lemberg <wl@gnu.org> |
| 1423 |
| 1424 [raster] Remove last remnants of `raster5' driver. |
| 1425 |
| 1426 * src/raster/ftrend1.h (ft_raster5_renderer_class): Removed. |
| 1427 |
| 1428 * src/raster/rastpic.c, src/raster/rastpic.h |
| 1429 (ft_raster5_renderer_class_pic_init, |
| 1430 ft_raster5_renderer_class_pic_free): Removed. |
| 1431 |
| 1432 2015-08-17 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 1433 |
| 1434 [base] Improve emboldener (#45596). |
| 1435 |
| 1436 * src/base/ftoutln.c (FT_Outline_EmboldenXY): Correct displacement |
| 1437 of zero-lenght segments. |
| 1438 |
| 1439 2015-08-16 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 1440 |
| 1441 [base] Reoptimize arithmetic. |
| 1442 |
| 1443 * src/base/ftcalc.c (FT_MulDiv, FT_MulFix) [!FT_LONG64]: Remove |
| 1444 special cases that slow down the general use. |
| 1445 |
| 1446 2015-08-15 pazer <ibemad@gmail.com> |
| 1447 |
| 1448 Fix C++ compilation (#45762). |
| 1449 |
| 1450 * src/base/ftstroke.c (ft_outline_glyph_class): Use |
| 1451 FT_CALLBACK_TABLE. |
| 1452 |
| 1453 2015-08-14 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 1454 |
| 1455 [truetype] Clean up. |
| 1456 |
| 1457 * src/truetype/ttgload.c (TT_Process_Composite_Component): Use |
| 1458 `FT_Outline_Transform' and `FT_Outline_Translate'. |
| 1459 (translate_array): Dropped. |
| 1460 |
| 1461 2015-08-14 Andreas Enge <andreas.enge@inria.fr> |
| 1462 |
| 1463 * builds/unix/detect.mk (CONFIG_SHELL): Don't handle it (#44261). |
| 1464 |
| 1465 2015-08-13 Werner Lemberg <wl@gnu.org> |
| 1466 |
| 1467 [truetype] Introduce named instance access to GX fonts. |
| 1468 |
| 1469 For functions querying a face, bits 16-30 of the face index can hold |
| 1470 the named instance index if we have a GX font. The indices start |
| 1471 with value 1; value 0 indicates font access without GX variation |
| 1472 data. |
| 1473 |
| 1474 * include/freetype/freetype.h (FT_FaceRec): Update documentation. |
| 1475 * include/freetype/internal/sfnt.h: Ditto. |
| 1476 |
| 1477 * src/sfnt/sfobjs.c (sfnt_init_face) |
| 1478 [TT_CONFIG_OPTION_GX_VAR_SUPPORT]: Get number of named instances and |
| 1479 do argument checks. |
| 1480 (sfnt_load_face): Updated. |
| 1481 |
| 1482 * src/truetype/ttobjs.c (tt_face_init) |
| 1483 [TT_CONFIG_OPTION_GX_VAR_SUPPORT]: Use named instance, overwriting |
| 1484 the style name. |
| 1485 |
| 1486 * src/base/ftobjs.c (open_face_from_buffer, |
| 1487 open_face_PS_from_sfnt_stream): Updated. |
| 1488 * src/bdf/bdfdrivr.c (BDF_Face_Init): Updated. |
| 1489 * src/cff/cffload.c (cff_font_load): Updated. |
| 1490 |
| 1491 * src/cff/cffobjs.c (cff_face_init): Make function exit early for |
| 1492 pure CFF fonts if `font_index < 0'. |
| 1493 Updated. |
| 1494 |
| 1495 * src/cid/cidobjs.c (cid_face_init): Updated. |
| 1496 * src/pcf/pcfdrivr.c (PCF_Face_Init): Updated. |
| 1497 * src/pfr/pfrobjs.c (pfr_face_init): Updated. |
| 1498 * src/type1/t1objs.c (T1_Face_Init): Updated. |
| 1499 * src/type42/t42objs.c (T42_Face_Init): Updated. |
| 1500 * src/winfonts/winfnt.c (fnt_face_get_dll_font, FNT_Face_Init): |
| 1501 Updated. |
| 1502 |
| 1503 * docs/CHANGES: Updated. |
| 1504 |
| 1505 2015-08-12 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 1506 |
| 1507 [type1,cff,cid] Streamline font matrix application. |
| 1508 |
| 1509 * src/type1/t1gload.c (T1_Load_Glyph): Directly modify advances only |
| 1510 if font matrix is not trivial. |
| 1511 * src/cff/cffgload.c (cff_slot_load): Ditto. |
| 1512 * sff/cid/cidgload.c (cid_slot_load_glyph): Ditto for advances and the |
| 1513 entire outline. |
| 1514 |
| 1515 2015-08-11 Werner Lemberg <wl@gnu.org> |
| 1516 |
| 1517 [builds/unix] Minor. |
| 1518 |
| 1519 * builds/unix/configure.raw: |
| 1520 s/lib{priv,staticconf}/libs{priv,staticconf}/ for orthogonality with |
| 1521 similarly named uppercase variables. |
| 1522 |
| 1523 2015-08-10 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 1524 |
| 1525 [type1,cid,type42] Minor improvements. |
| 1526 |
| 1527 * src/type1/t1load.c (t1_parse_font_matrix): Scale units per EM only |
| 1528 when necessary. Refresh comments. |
| 1529 * src/cid/cidload.c (cid_parse_font_matrix): Ditto. |
| 1530 * src/type42/t42parse.c (t42_parse_font_matrix): Refresh comments. |
| 1531 |
| 1532 2015-08-08 Werner Lemberg <wl@gnu.org> |
| 1533 |
| 1534 [type42] Fix glyph access. |
| 1535 |
| 1536 This is a severe bug: We've missed one level of indirection, as |
| 1537 described in the Type 42 specification. As a result, ftview |
| 1538 sometimes showed incorrect glyphs for given glyph names, and even |
| 1539 displayed `error 0x0006' (invalid argument!) in case the number of |
| 1540 glyph indices differed between the Type 42 font and the embedded |
| 1541 TTF. |
| 1542 |
| 1543 Apparently, noone ever noticed it; this shows how much Type 42 fonts |
| 1544 are in use... |
| 1545 |
| 1546 * src/type42/t42objs.c (T42_GlyphSlot_Load): Map Type 42 glyph index |
| 1547 to embedded TTF's glyph index. |
| 1548 |
| 1549 2015-08-08 Werner Lemberg <wl@gnu.org> |
| 1550 |
| 1551 [type42] Minor clean-up. |
| 1552 |
| 1553 * src/type42/t42parse.c (t42_parse_font_matrix): Remove unused |
| 1554 variable. |
| 1555 |
| 1556 2015-08-06 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 1557 |
| 1558 [type42] Parse FontMatrix according to specifications. |
| 1559 |
| 1560 * src/type42/t42parse.c (t42_parse_font_matrix): Type 42 FontMatrix |
| 1561 does not need scaling by 1000. Units_per_EM are taken from the |
| 1562 embedded TrueType. |
| 1563 |
| 1564 2015-08-06 Werner Lemberg <wl@gnu.org> |
| 1565 |
| 1566 [autofit] Improve Arabic hinting. |
| 1567 |
| 1568 Problem reported by Titus Nemeth <tn@tntypography.eu> (by using |
| 1569 ttfautohint). |
| 1570 |
| 1571 * src/autofit/afblue.dat: Add neutral blue zone for the tatweel |
| 1572 character. |
| 1573 |
| 1574 * src/autofit/afblue.c, src/autofit/afblue.h: Regenerated. |
| 1575 |
| 1576 2015-08-05 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 1577 |
| 1578 [truetype] Clean up types. |
| 1579 |
| 1580 * src/truetype/ttobjs.c (TT_Size): Move declaration from here. |
| 1581 * include/freetype/internal/tttypes.h (TT_Size): ... to here. |
| 1582 (TT_LoaderRec): Switch to appropriate types for `face' and `size'. |
| 1583 * src/truetype/ttgload.c: Remove corresponding type casts. |
| 1584 * src/truetype/ttsubpix.c: Ditto. |
| 1585 |
| 1586 2015-08-05 Werner Lemberg <wl@gnu.org> |
| 1587 |
| 1588 [autofit] Improve recognition of flat vs. rounded segments. |
| 1589 |
| 1590 Lower the flatness threshold from upem/8 to upem/14, making the |
| 1591 auto-hinter accept shorter elements. |
| 1592 |
| 1593 Synchronize flat/round stem selection algorithm with blue zone code. |
| 1594 |
| 1595 * src/autofit/aflatin.c (FLAT_THRESHOLD): New macro. |
| 1596 (af_latin_metrics_init_blues): Use it. |
| 1597 (af_latin_hints_compute_segments): Collect information on maximum |
| 1598 and minimum coordinates of `on' points; use this to add a constraint |
| 1599 for the flat/round decision similar to |
| 1600 `af_latin_metrics_init_blues'. |
| 1601 |
| 1602 2015-08-04 Werner Lemberg <wl@gnu.org> |
| 1603 |
| 1604 Another left-shift bug (#45681). |
| 1605 |
| 1606 * src/base/ftobjs.c (IsMacBinary): Only accept positive values for |
| 1607 `dlen'. |
| 1608 |
| 1609 2015-08-03 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 1610 |
| 1611 [base] Fix `ft_corner_orientation'. |
| 1612 |
| 1613 Remove casting from `FT_Long' to `FT_Int' that might change the sign |
| 1614 of the return value and make it faster too. |
| 1615 |
| 1616 * src/base/ftcalc.c (ft_corner_orientation): On 32-bit systems, stay |
| 1617 with 32-bit arithmetic when safe. Use plain math on 64-bit systems. |
| 1618 * src/pshinter/pshalgo.c: Remove old unused code. |
| 1619 |
| 1620 2015-08-03 Werner Lemberg <wl@gnu.org> |
| 1621 |
| 1622 * src/truetype/ttgload.c (load_truetype_glyph) |
| 1623 [TT_CONFIG_OPTION_GX_VAR_SUPPORT]: Fix crash for composite glyphs |
| 1624 having a depth greater than 1. |
| 1625 |
| 1626 2015-08-03 Werner Lemberg <wl@gnu.org> |
| 1627 |
| 1628 Fix typo in clang bug from 2015-07-31 (#45678). |
| 1629 |
| 1630 * src/base/ftrfork.c (FT_Raccess_Get_HeaderInfo): Fix inequality. |
| 1631 |
| 1632 2015-08-02 Werner Lemberg <wl@gnu.org> |
| 1633 |
| 1634 * CMakeLists.txt: Improve shared library support. |
| 1635 |
| 1636 Based on a patch from John Cary <cary@txcorp.com>. |
| 1637 |
| 1638 2015-08-02 Werner Lemberg <wl@gnu.org> |
| 1639 |
| 1640 * builds/unix/freetype-config.in (enable_shared): Remove. Unused. |
| 1641 |
| 1642 2015-08-02 Werner Lemberg <wl@gnu.org> |
| 1643 |
| 1644 Fix more invalid left-shifts. |
| 1645 |
| 1646 * src/pfr/pfrgload.c (pfr_glyph_load_compound): Use multiplication, |
| 1647 not left-shift. |
| 1648 |
| 1649 * src/truetype/ttgxvar.c (ft_var_load_avar, ft_var_load_gvar, |
| 1650 tt_face_vary_cvt, TT_Vary_Apply_Glyph_Deltas): Use multiplication, |
| 1651 not left-shift. |
| 1652 |
| 1653 2015-07-31 Werner Lemberg <wl@gnu.org> |
| 1654 |
| 1655 Fix some bugs found by clang's `-fsanitize=undefined' (#45661). |
| 1656 |
| 1657 * src/base/ftrfork.c (FT_Raccess_Get_HeaderInfo): Only accept |
| 1658 positive values from header. |
| 1659 Check overflow. |
| 1660 |
| 1661 * src/base/ftoutln.c (SCALED): Correctly handle left-shift of |
| 1662 negative values. |
| 1663 |
| 1664 * src/bdf/bdf.h (_bdf_glyph_modified, _bdf_set_glyph_modified, |
| 1665 _bdf_clear_glyph_modified): Use unsigned long constant. |
| 1666 |
| 1667 * src/bdf/bdfdrivr.c (BDF_Size_Select, BDF_Glyph_Load): Don't |
| 1668 left-shift values that can be negative. |
| 1669 |
| 1670 * src/pcf/pcfdrivr.c (PCF_Size_Select, PCF_Glyph_Load): Don't |
| 1671 left-shift values that can be negative. |
| 1672 |
| 1673 * src/raster/ftraster.c (SCALED): Correctly handle left-shift of |
| 1674 negative values. |
| 1675 |
| 1676 * src/sfnt/ttsbit.c (tt_face_load_strike_metrics): Don't left-shift |
| 1677 values that can be negative. |
| 1678 |
| 1679 * src/truetype/ttgload.c (TT_Load_Composite_Glyph, |
| 1680 compute_glyph_metrics, load_sbit_image): Don't left-shift values |
| 1681 that can be negative. |
| 1682 |
| 1683 2015-07-31 Werner Lemberg <wl@gnu.org> |
| 1684 |
| 1685 Define FT_LONG_MAX. |
| 1686 |
| 1687 * include/freetype/config/ftstdlib.h (FT_LONG_MAX): New macro. |
| 1688 * src/cff/cf2arrst.c (cf2_arrstack_setNumElements): Use it. |
| 1689 |
| 1690 2015-07-28 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 1691 |
| 1692 * src/base/ftcalc.c (FT_Vector_NormLen): Clarify. |
| 1693 |
| 1694 2015-07-27 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 1695 |
| 1696 * src/base/ftcalc.c (FT_Vector_NormLen): Explicate type conversions. |
| 1697 |
| 1698 2015-07-26 Matthias Clasen <matthias.clasen@gmail.com> |
| 1699 |
| 1700 [cff] Don't use `hmtx' table for LSB (#45520). |
| 1701 |
| 1702 * src/cff/cffgload.c (cff_slot_load): Use `htmx' table for advance |
| 1703 width only. Bug introduced 2015-04-10. |
| 1704 |
| 1705 2015-07-09 Werner Lemberg <wl@gnu.org> |
| 1706 |
| 1707 Better support of user-supplied C++ namespaces. |
| 1708 |
| 1709 See |
| 1710 |
| 1711 http://lists.nongnu.org/archive/html/freetype-devel/2015-07/msg00008.h
tml |
| 1712 |
| 1713 for a rationale. |
| 1714 |
| 1715 * src/autofit/afpic.h, src/base/basepic.h, src/cff/cffpic.h, |
| 1716 src/pshinter/pshpic.h, src/psnames/pspic.h, src/raster/rastpic.h, |
| 1717 src/sfnt/sfntpic.h, src/smooth/ftspic.h, src/truetype/ttpic.h |
| 1718 (FT_BEGIN_HEADER, FT_END_HEADER): Move macro calls to not enclose |
| 1719 header files that contain FT_{BEGIN,END}_HEADER macros by |
| 1720 themselves. |
| 1721 |
| 1722 * src/autofit/aftypes.h [FT_DEBUG_AUTOFIT]: Include |
| 1723 FT_CONFIG_STANDARD_LIBRARY_H earlier. |
| 1724 |
| 1725 * src/truetype/ttpic.h: Include FT_INTERNL_PIC_H. |
| 1726 |
| 1727 2015-07-07 Werner Lemberg <wl@gnu.org> |
| 1728 |
| 1729 [sfnt] Make `tt_face_get_name' member of the SFNT interface. |
| 1730 |
| 1731 * include/freetype/internal/sfnt.h (TT_Get_Name_Func): New |
| 1732 prototype. |
| 1733 (SFNT_Interface, FT_DEFINE_SFNT_INTERFACE): New member `get_name'. |
| 1734 |
| 1735 * src/sfnt/sfdriver.c (sfnt_interface): Updated. |
| 1736 |
| 1737 * src/sfnt/sfobjs.c (tt_face_get_name): Tag it with `LOCAL_DEF'. |
| 1738 * src/sfnt/sfobjs.h: Add prototype for it. |
| 1739 |
| 1740 2015-06-30 Werner Lemberg <wl@gnu.org> |
| 1741 |
| 1742 Fix some clang compiler warnings. |
| 1743 |
| 1744 * src/base/ftoutln.c (FT_Outline_EmboldenXY), src/cff/cf2intrp.c |
| 1745 (cf2_interpT2CharString), src/truetype/ttgload.c |
| 1746 (load_truetype_glyph), src/truetype/ttgxvar.c (tt_handle_deltas), |
| 1747 src/truetype/ttinterp.c (Ins_INSTCTRL): Fix signedness issues. |
| 1748 |
| 1749 2015-06-29 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 1750 |
| 1751 [truetype] Speed up bytecode interpreter. |
| 1752 |
| 1753 * src/truetype/ttinterp.c (Normalize): Use `FT_Vector_NormLen'. |
| 1754 |
| 1755 2015-06-29 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 1756 |
| 1757 [base] Speed up emboldening. |
| 1758 |
| 1759 * src/base/ftoutln.c (FT_Outline_EmboldenXY): Use |
| 1760 `FT_Vector_NormLen'. |
| 1761 |
| 1762 2015-06-29 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 1763 |
| 1764 [base] Implement fast vector normalization. |
| 1765 |
| 1766 The function uses Newton's iterations instead of dividing vector |
| 1767 components by its length, which needs a square root. This is, |
| 1768 literally, a bit less accurate but a lot faster. |
| 1769 |
| 1770 * src/base/ftcalc.c (FT_Vector_NormLen): New function. |
| 1771 |
| 1772 2015-06-28 Werner Lemberg <wl@gnu.org> |
| 1773 |
| 1774 * CMakeLists.txt: Always create `ftconfig.h'. |
| 1775 |
| 1776 For non-UNIX builds, the file stays unmodified. However, it's |
| 1777 better to have the main configuration files at the same place |
| 1778 regardless of the OS. |
| 1779 |
| 1780 2015-06-28 Werner Lemberg <wl@gnu.org> |
| 1781 |
| 1782 * CMakeLists.txt: Improve MSVC support (#43737). |
| 1783 |
| 1784 2015-06-28 Werner Lemberg <wl@gnu.org> |
| 1785 |
| 1786 [cmake] Check for libraries and create `ftoption.h'. |
| 1787 |
| 1788 * builds/FindHarfBuzz.cmake: New file, taken from |
| 1789 |
| 1790 https://trac.webkit.org/browser/trunk/Source/cmake/FindHarfBuzz.cmake |
| 1791 |
| 1792 * CMakeLists.Txt: Add path to local cmake modules. |
| 1793 Find dependencies for zlib, bzip2, libpng, and harfbuzz. |
| 1794 Create `ftoption.h' file. |
| 1795 Set up include and linker stuff for libraries. |
| 1796 |
| 1797 2015-06-28 Werner Lemberg <wl@gnu.org> |
| 1798 |
| 1799 * CMakeLists.txt: Fix creation of `ftconfig.h'. |
| 1800 Check for UNIX header files using `check_include_file'. |
| 1801 Set up correct header include directories. |
| 1802 |
| 1803 2015-06-28 Werner Lemberg <wl@gnu.org> |
| 1804 |
| 1805 * CMakeLists.txt: Disallow in-source builds. |
| 1806 |
| 1807 2015-06-27 Werner Lemberg <wl@gnu.org> |
| 1808 |
| 1809 * src/tools/docmaker/utils.py (check_output): Add missing `\n'. |
| 1810 |
| 1811 2015-06-26 Werner Lemberg <wl@gnu.org> |
| 1812 |
| 1813 * CMakeLists.txt: Select platform-dependent `ftdebug.c'. |
| 1814 |
| 1815 2015-06-25 Werner Lemberg <wl@gnu.org> |
| 1816 |
| 1817 * CMakeLists.txt: Use cmake functions for generating `ftconfig.h'. |
| 1818 Additionally, do this for UNIX only. |
| 1819 |
| 1820 2015-06-25 Werner Lemberg <wl@gnu.org> |
| 1821 |
| 1822 * CMakeLists.txt (BASE_SRCS): Use `ftbase.c' and `psnames.c'. |
| 1823 |
| 1824 2015-06-25 Werner Lemberg <wl@gnu.org> |
| 1825 |
| 1826 Another adjustment to header locations. |
| 1827 |
| 1828 This change is a result of a discussion thread on freetype-devel |
| 1829 |
| 1830 http://lists.nongnu.org/archive/html/freetype-devel/2015-06/msg00041.h
tml |
| 1831 |
| 1832 Re-introduce the `freetype2' subdirectory for all FreeType header |
| 1833 files after installation, and rename the `freetype2' subdirectory in |
| 1834 the git repository to `freetype'. |
| 1835 |
| 1836 * include/freetype2: Renamed to... |
| 1837 * include/freetype: This. |
| 1838 |
| 1839 * CMakeLists.txt (PUBLIC_HEADERS, PUBLIC_CONFIG_HEADERS, |
| 1840 PRIVATE_HEADERS): Updated. |
| 1841 Update creation of `ftconfig.h'. |
| 1842 Install generated `ftconfig.h'. |
| 1843 |
| 1844 * Jamfile (HDRMACRO, RefDoc), autogen.sh: Updated. |
| 1845 |
| 1846 * builds/amiga/include/config/ftconfig.h, builds/freetype.mk |
| 1847 (PUBLIC_DIR), builds/symbian/bld.inf, builds/toplevel.mk (work), |
| 1848 builds/unix/freetype2.in: Updated. |
| 1849 |
| 1850 * builds/unix/freetype-config.in: Updated. |
| 1851 * builds/unix/configure.raw: Don't check for `rmdir'. |
| 1852 * builds/unix/unix-def.in (DELDIR): Use `rm -rf', which is portable |
| 1853 according to the autoconf info manual. |
| 1854 * builds/unix/install.mk (install, uninstall, |
| 1855 distclean_project_unix): Update and simplify. |
| 1856 |
| 1857 * builds/wince/*, builds/windows/*: Updated. |
| 1858 |
| 1859 * devel/ft2build.h, include/ft2build.h: Updated. |
| 1860 |
| 1861 * include/freetype2/config/ftheader.h, |
| 1862 include/freetype2/internal/ftserv.h, |
| 1863 include/freetype2/internal/internal.h: Update all header file |
| 1864 macros. |
| 1865 |
| 1866 * src/tools/chktrcmp.py (TRACE_DEF_FILES): Updated. |
| 1867 |
| 1868 * docs/*: Updated. |
| 1869 |
| 1870 2015-06-24 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 1871 |
| 1872 * src/bdf/bdflib.c (_bdf_parse_start): Disallow 0 bpp. |
| 1873 |
| 1874 2015-06-24 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 1875 |
| 1876 * src/bdf/bdflib.c (_bdf_parse_start): Simplify bpp parsing. |
| 1877 |
| 1878 2015-06-23 Werner Lemberg <wl@gnu.org> |
| 1879 |
| 1880 s/TYPEOF/FT_TYPEOF/ (#45376). |
| 1881 |
| 1882 * builds/unix/ftconfig.in, builds/vms/ftconfig.in, |
| 1883 include/freetype2/config/ftconfig.h, |
| 1884 include/freetype2/internal/ftobjs.h, src/autofit/afwarp.h: Do it. |
| 1885 |
| 1886 2015-06-22 Werner Lemberg <wl@gnu.org> |
| 1887 |
| 1888 Fix Savannah bug #45097. |
| 1889 |
| 1890 We no longer `pollute' the namespace of possible header file names; |
| 1891 instead we move `ft2build.h' up by one level so that it gets |
| 1892 installed in the default include directory (e.g., |
| 1893 /usr/local/include). After this commit, only `ft2build.h' stays in |
| 1894 the compiler's include path. |
| 1895 |
| 1896 No visible changes for the user who follows the standard FreeType |
| 1897 header inclusion rules. |
| 1898 |
| 1899 * include/*: Move to ... |
| 1900 * include/freetype2/*: This directory, except `ft2build.h'. |
| 1901 |
| 1902 * CMakeLists.txt (PUBLIC_HEADERS, PUBLIC_CONFIG_HEADERS, |
| 1903 PRIVATE_HEADERS), Jamfile (HDRMACRO, RefDoc), autogen.sh: Updated. |
| 1904 |
| 1905 * builds/amiga/include/config/ftconfig.h, builds/freetype.mk |
| 1906 (PUBLIC_DIR), builds/symbian/bld.inf, builds/toplevel.mk (work), |
| 1907 builds/unix/install.mk (install, uninstall), |
| 1908 builds/unix/freetype2.in: Updated. |
| 1909 |
| 1910 * builds/unix/freetype-config.in: Updated. |
| 1911 Emit -I directory only if it is not `/usr/include'. |
| 1912 |
| 1913 * builds/wince/*, builds/windows/*: Updated. |
| 1914 |
| 1915 * devel/ft2build.h, include/ft2build.h: Updated. |
| 1916 |
| 1917 * include/freetype2/config/ftheader.h, |
| 1918 include/freetype2/internal/ftserv.h, |
| 1919 include/freetype2/internal/internal.h: Update all header file |
| 1920 macros. |
| 1921 |
| 1922 * src/tools/chktrcmp.py (TRACE_DEF_FILES): Updated. |
| 1923 |
| 1924 2015-06-21 Werner Lemberg <wl@gnu.org> |
| 1925 |
| 1926 Make Jam support work again. |
| 1927 |
| 1928 This is just very basic stuff and just a little bit tested on |
| 1929 GNU/Linux only. I won't delve into this since I'm not a Jam user. |
| 1930 |
| 1931 * Jamfile: Call `HDRMACRO' for `ftserv.h' also. |
| 1932 (DEFINES): Replace with... |
| 1933 (CCFLAGS): ... this. |
| 1934 |
| 1935 * src/Jamfile: Don't call `HDRMACRO' for `internal.h'; this is |
| 1936 already handled in the top-level Jamfile. |
| 1937 |
| 1938 * src/autofit/Jamfile (DEFINES): Replace with... |
| 1939 (CCFLAGS): ... this. |
| 1940 (_sources): Add missing files. |
| 1941 |
| 1942 * src/cache/Jamfile: Don't call `HDRMACRO' for `ftcache.h'; it no |
| 1943 longer contains macro header definitions. |
| 1944 |
| 1945 * src/base/Jamfile, src/cff/Jamfile, src/sfnt/Jamfile, |
| 1946 src/truetype/Jamfile (_sources): Add missing files. |
| 1947 |
| 1948 2015-06-16 Werner Lemberg <wl@gnu.org> |
| 1949 |
| 1950 Fix Savannah bug #45326. |
| 1951 |
| 1952 * src/sfnt/sfntpic.h (SFNT_SERVICES_GET): Remove duplicate |
| 1953 definitions. |
| 1954 |
| 1955 2015-06-07 Werner Lemberg <wl@gnu.org> |
| 1956 |
| 1957 * Version 2.6 released. |
| 1958 ======================= |
| 1959 |
| 1960 |
| 1961 Tag sources with `VER-2-6'. |
| 1962 |
| 1963 * docs/VERSION.DLL: Update documentation and bump version number to |
| 1964 2.6. |
| 1965 |
| 1966 * README, Jamfile (RefDoc), builds/windows/vc2005/freetype.vcproj, |
| 1967 builds/windows/vc2005/index.html, |
| 1968 builds/windows/vc2008/freetype.vcproj, |
| 1969 builds/windows/vc2008/index.html, |
| 1970 builds/windows/vc2010/freetype.vcxproj, |
| 1971 builds/windows/vc2010/index.html, |
| 1972 builds/windows/visualc/freetype.dsp, |
| 1973 builds/windows/visualc/freetype.vcproj, |
| 1974 builds/windows/visualc/index.html, |
| 1975 builds/windows/visualce/freetype.dsp, |
| 1976 builds/windows/visualce/freetype.vcproj, |
| 1977 builds/windows/visualce/index.html, |
| 1978 builds/wince/vc2005-ce/freetype.vcproj, |
| 1979 builds/wince/vc2005-ce/index.html, |
| 1980 builds/wince/vc2008-ce/freetype.vcproj, |
| 1981 builds/wince/vc2008-ce/index.html: s/2.5.5/2.6/, s/255/26/. |
| 1982 |
| 1983 * include/freetype/freetype.h (FREETYPE_MINOR): Set to 6. |
| 1984 (FREETYPE_PATCH): Set to 0. |
| 1985 |
| 1986 * builds/unix/configure.raw (version_info): Set to 18:0:12. |
| 1987 * CMakeLists.txt (VERSION_MINOR): Set to 6. |
| 1988 (VERSION_PATCH): Set to 0. |
| 1989 |
| 1990 * src/autofit/afmodule.c [!FT_MAKE_OPTION_SINGLE_OBJECT]: Add |
| 1991 declarations for dumping functions. |
| 1992 |
| 1993 * src/truetype/ttinterp.c (TT_New_Context): Pacify compiler. |
| 1994 |
| 1995 * builds/toplevel.mk: Use `freetype.mk's code to compute the version |
| 1996 string. |
| 1997 Don't include a zero patch level in version string. |
| 1998 * builds/freetype.mk: Remove code for computing the version string. |
| 1999 |
| 2000 2015-06-06 Ashish Azad <ashish.azad@samsung.com> |
| 2001 |
| 2002 Fix Savannah bug #45260. |
| 2003 |
| 2004 * src/pfr/pfrdrivr.c (pfr_get_kerning): Fix typo. |
| 2005 |
| 2006 2015-06-03 Werner Lemberg <wl@gnu.org> |
| 2007 |
| 2008 [truetype] Fix memory leak. |
| 2009 |
| 2010 Problem reported by Grissiom <chaos.proton@gmail.com>; in |
| 2011 |
| 2012 http://lists.nongnu.org/archive/html/freetype/2015-05/msg00013.html |
| 2013 |
| 2014 there is an example code to trigger the bug. |
| 2015 |
| 2016 * src/truetype/ttobjs.c (tt_size_init_bytecode): Free old `size' |
| 2017 data before allocating again. Bug most probably introduced four |
| 2018 years ago in version 2.4.3. |
| 2019 |
| 2020 2015-06-02 Werner Lemberg <wl@gnu.org> |
| 2021 |
| 2022 [raster] Add more tracing. |
| 2023 |
| 2024 * src/raster/ftraster.c (FT_TRACE7) [_STANDALONE_]: Define. |
| 2025 (Vertical_Sweep_Span, Vertical_Sweep_Drop, Horizontal_Sweep_Span, |
| 2026 Horizontal_Sweep_Drop, Render_Glyph): Add tracing calls. |
| 2027 |
| 2028 2015-06-01 Werner Lemberg <wl@gnu.org> |
| 2029 |
| 2030 [truetype] While tracing opcodes, show code position and stack. |
| 2031 |
| 2032 * src/truetype/ttinterp.c: Change all existing TRACE7 calls to |
| 2033 TRACE6. |
| 2034 (opcode_name): Add string lengths. |
| 2035 (TT_RunIns): Implement display of code position and stack. |
| 2036 |
| 2037 2015-05-31 Werner Lemberg <wl@gnu.org> |
| 2038 |
| 2039 [truetype] In GX, make private point numbers work correctly. |
| 2040 |
| 2041 This is completely missing in Apple's documentation: If a `gvar' |
| 2042 tuple uses private point numbers (this is, deltas are specified for |
| 2043 some points only), the uncovered points must be interpolated for |
| 2044 this tuple similar to the IUP bytecode instruction. Examples that |
| 2045 need this functionality are glyphs `Oslash' and `Q' in Skia.ttf. |
| 2046 |
| 2047 * src/truetype/ttgxvar.c (tt_delta_shift, tt_delta_interpolate, |
| 2048 tt_handle_deltas): New functions. |
| 2049 (TT_Vary_Get_Glyph_Deltas): Renamed to... |
| 2050 (TT_Vary_Apply_Glyph_Deltas): ... this; it directly processes the |
| 2051 points and does no longer return an array of deltas. |
| 2052 Add tracing information. |
| 2053 Call `tt_handle_deltas' to interpolate missing deltas. |
| 2054 Also fix a minor memory leak in case of error. |
| 2055 |
| 2056 * src/truetype/ttgxvar.h: Updated. |
| 2057 |
| 2058 * src/truetype/ttgload.c (TT_Process_Simple_Glyph, |
| 2059 load_truetype_glyph): Updated. |
| 2060 |
| 2061 2015-05-31 Werner Lemberg <wl@gnu.org> |
| 2062 |
| 2063 [truetype] In GX, make intermediate tuplets work at extrema. |
| 2064 |
| 2065 * src/truetype/ttgxvar.c (ft_var_apply_tuple): Fix range condition. |
| 2066 |
| 2067 2015-05-31 Werner Lemberg <wl@gnu.org> |
| 2068 |
| 2069 [truetype] Add tracing information to GX code. |
| 2070 |
| 2071 * src/truetype/ttgxvar.c (ft_var_load_avar, ft_var_load_gvar, |
| 2072 ft_var_apply_tuple, TT_Get_MM_Var, TT_Set_MM_Blend, |
| 2073 TT_Set_Var_Design, tt_face_vary_cvt): Do it. |
| 2074 |
| 2075 2015-05-28 Werner Lemberg <wl@gnu.org> |
| 2076 |
| 2077 * src/tools/apinames.c (names_dump): Fix invalid reference. |
| 2078 |
| 2079 Problem reported by Guzman Mosqueda, Jose R |
| 2080 <jose.r.guzman.mosqueda@intel.com>. |
| 2081 |
| 2082 2015-05-24 Werner Lemberg <wl@gnu.org> |
| 2083 |
| 2084 [truetype] Fix commit from 2015-05-22. |
| 2085 |
| 2086 * src/truetype/ttgload.c, src/truetype/ttinterp.c: Guard new code |
| 2087 with `TT_CONFIG_OPTION_SUBPIXEL_HINTING'. |
| 2088 |
| 2089 Problem reported by Nikolaus Waxweiler <madigens@gmail.com>. |
| 2090 |
| 2091 2015-05-23 Werner Lemberg <wl@gnu.org> |
| 2092 |
| 2093 [truetype] Fix return values of GETINFO bytecode instruction. |
| 2094 |
| 2095 * src/truetype/ttinterp.h (TT_ExecContextRec): New fields |
| 2096 `vertical_lcd' and `gray_cleartype'. |
| 2097 |
| 2098 * src/truetype/ttgload.c (tt_loader_init): Initialize new fields. |
| 2099 Change `symmetrical smoothing' to TRUE, since FreeType produces |
| 2100 exactly this. |
| 2101 |
| 2102 * src/truetype/ttinterp.c (Ins_GETINFO): Fix selector/return bit |
| 2103 values for symmetrical smoothing, namely 11/18. |
| 2104 Handle bits for vertical LCD subpixels (8/15) and Gray ClearType |
| 2105 (12/19). |
| 2106 |
| 2107 2015-05-23 Werner Lemberg <wl@gnu.org> |
| 2108 |
| 2109 [truetype] Minor. |
| 2110 |
| 2111 * src/truetype/ttinterp.h (TT_ExecContext): |
| 2112 s/subpixel/subpixel_hinting. |
| 2113 |
| 2114 * src/truetype/ttgload.c, src/truetype/ttgload.h: Updated. |
| 2115 |
| 2116 2015-05-22 Werner Lemberg <wl@gnu.org> |
| 2117 |
| 2118 [truetype] Support selector index 3 of the INSTCTRL instruction. |
| 2119 |
| 2120 This flag activates `native ClearType hinting', disabling backwards |
| 2121 compatibility mode as described in Greg Hitchcocks whitepaper. In |
| 2122 other words, it enables unrestricted functionality of all TrueType |
| 2123 instructions in ClearType. |
| 2124 |
| 2125 * src/truetype/ttgload.c (tt_get_metrics): Call `sph_set_tweaks' |
| 2126 unconditionally. |
| 2127 (tt_loader_init): Unset `ignore_x_mode' flag if bit 2 of |
| 2128 `GS.instruct_control' is active. |
| 2129 |
| 2130 * src/truetype/ttinterp.c (Ins_INSTCTRL): Handle selector index 3. |
| 2131 (Ins_GETINFO): Updated. |
| 2132 |
| 2133 * docs/CHANGES: Document it. |
| 2134 |
| 2135 2015-05-20 Werner Lemberg <wl@gnu.org> |
| 2136 |
| 2137 [truetype] Minor. |
| 2138 |
| 2139 * src/truetype/ttinterp.h (SetSuperRound): Fix type of `GridPeriod' |
| 2140 argument. |
| 2141 |
| 2142 2015-05-17 Werner Lemberg <wl@gnu.org> |
| 2143 |
| 2144 [truetype] Fix loading of composite glyphs. |
| 2145 |
| 2146 * src/truetype/ttgload.c (TT_Load_Composite_Glyph): If the |
| 2147 ARGS_ARE_XY_VALUES flag is not set, handle argument values as |
| 2148 unsigned. I trust `ttx' (which has exactly such code) that it does |
| 2149 the right thing here... |
| 2150 |
| 2151 The reason that noone has ever noticed this bug is probably the fact |
| 2152 that point-aligned subglyphs are rare, as are subglyphs with a |
| 2153 number of points in the range [128;255], which is quite large (or |
| 2154 even in the range [32768;65535], which is extremely unlikely). |
| 2155 |
| 2156 2015-05-12 Chris Liddell <chris.liddell@artifex.com> |
| 2157 |
| 2158 [cff] Make the `*curveto' operators more tolerant. |
| 2159 |
| 2160 * src/cff/cf2intrp.c (cf2_interpT2CharString): The opcodes |
| 2161 `vvcurveto', `hhcurveto', `vhcurveto', and `hvcurveto' all iterate, |
| 2162 pulling values off the stack until the stack is exhausted. |
| 2163 Implicitly the stack must be a multiple (or for subtly different |
| 2164 behaviour) a multiple plus a specific number of extra values deep. |
| 2165 If that's not the case, enforce it (as the old code did). |
| 2166 |
| 2167 2015-05-12 Chris Liddell <chris.liddell@artifex.com> |
| 2168 |
| 2169 [cff] fix incremental interface with new cff code. |
| 2170 |
| 2171 * src/cff/cf2ft.c (cf2_getSeacComponent): When using the incremental |
| 2172 interface to retrieve glyph data for a SEAC, it be left to the |
| 2173 incremental interface callback to apply the encoding to raw |
| 2174 character index (as it was in the previous code). |
| 2175 |
| 2176 2015-04-29 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 2177 |
| 2178 [autofit] Speed up IUP. |
| 2179 |
| 2180 * src/autofit/afhints.c (af_iup_interp): Separate trivial snapping to |
| 2181 the same position from true interpolation, use `scale' to reduce |
| 2182 divisions. |
| 2183 |
| 2184 2015-04-28 Werner Lemberg <wl@gnu.org> |
| 2185 |
| 2186 [cff] Use `name' table for PS name if we have a SFNT-CFF. |
| 2187 |
| 2188 This follows the OpenType 1.7 specification. See |
| 2189 |
| 2190 http://tug.org/pipermail/tex-live/2015-April/036634.html |
| 2191 |
| 2192 for a discussion. |
| 2193 |
| 2194 * src/cff/cffdrivr.c (cff_get_ps_name): Use the `sfnt' service if we |
| 2195 have an SFNT. |
| 2196 |
| 2197 2015-04-27 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 2198 |
| 2199 [truetype] Speed up IUP. |
| 2200 |
| 2201 * src/truetype/ttinterp.c (_iup_worker_interpolate): Separate trivial |
| 2202 snapping to the same position from true interpolation. |
| 2203 |
| 2204 2015-04-21 Werner Lemberg <wl@gnu.org> |
| 2205 |
| 2206 [autofit] By default, enable warping code but switch off warping. |
| 2207 |
| 2208 Suggested by Behdad. |
| 2209 |
| 2210 * include/config/ftoption.h: Define AF_CONFIG_OPTION_USE_WARPER. |
| 2211 |
| 2212 * src/autofit/afmodule.c (af_autofitter_init): Initialize `warping' |
| 2213 with `false'. |
| 2214 |
| 2215 2015-04-21 Werner Lemberg <wl@gnu.org> |
| 2216 |
| 2217 * docs/CHANGES: Updated. |
| 2218 |
| 2219 2015-04-21 Werner Lemberg <wl@gnu.org> |
| 2220 |
| 2221 [autofit] Introduce `warping' property. |
| 2222 |
| 2223 This code replaces the debugging hook from the previous commit with |
| 2224 a better, more generic solution. |
| 2225 |
| 2226 * include/ftautoh.h: Document it. |
| 2227 |
| 2228 * src/autofit/afmodule.h (AF_ModuleRec) |
| 2229 [AF_CONFIG_OPTION_USE_WARPER]: Add `warping' field. |
| 2230 |
| 2231 * src/autofit/afmodule.c (_af_debug_disable_warper): Remove. |
| 2232 (af_property_set, af_property_get, af_autofitter_init) |
| 2233 [AF_CONFIG_OPTION_USE_WARPER]: Handle `warping' option. |
| 2234 |
| 2235 * src/autofit/afhints.h (AF_HINTS_DO_WARP): Remove use of the no |
| 2236 longer existing `_af_debug_disable_warper'. |
| 2237 |
| 2238 * src/autofit/afcjk.c (af_cjk_hints_init), src/autofit/aflatin.c |
| 2239 (af_latin_hints_init), src/autofit/aflatin2.c (af_latin2_hints_init) |
| 2240 [AF_CONFIG_OPTION_USE_WARPER]: Add `AF_SCALER_FLAG_NO_WARPER' to the |
| 2241 scaler flags if warping is off. |
| 2242 |
| 2243 * src/autofit/aftypes.h: Updated. |
| 2244 |
| 2245 2015-04-16 Werner Lemberg <wl@gnu.org> |
| 2246 |
| 2247 [autofit] Add debugging hook to disable warper. |
| 2248 |
| 2249 * src/autofit/afmodule.c (_af_debug_disable_warper) |
| 2250 [FT_DEBUG_AUTOFIT]: New global variable. |
| 2251 |
| 2252 * src/autofit/aftypes.h: Updated. |
| 2253 (AF_SCALER_FLAG_NO_WARPER): New macro (not actively used yet). |
| 2254 |
| 2255 * src/autofit/afhints.h (AF_HINTS_DO_WARP): New macro. |
| 2256 |
| 2257 * src/autofi/aflatin.c (af_latin_hints_apply) |
| 2258 [AF_CONFIG_OPTION_USE_WARPER]: Use `AF_HINTS_DO_WARP' to control use |
| 2259 of warper. |
| 2260 |
| 2261 * src/autofit/afcjk.c (af_cjk_hints_init, af_cjk_hints_apply) |
| 2262 [AF_CONFIG_OPTION_USE_WARPER]: Synchronize with `aflatin.c'. |
| 2263 |
| 2264 * src/autofit/aflatin2.c (af_latin2_hints_apply) |
| 2265 [AF_CONFIG_OPTION_USE_WARPER]: Synchronize with `aflatin.c'. |
| 2266 |
| 2267 2015-04-10 Werner Lemberg <wl@gnu.org> |
| 2268 |
| 2269 [cff] Update advance width handling to OpenType 1.7. |
| 2270 |
| 2271 Problem reported by Behdad. |
| 2272 |
| 2273 * src/cff/cffdrivr.c (cff_get_advances): Handle SFNT case |
| 2274 separately. |
| 2275 |
| 2276 * src/cff/cffgload.c (cff_slot_load): Use advance width and side |
| 2277 bearing values from `hmtx' table if present. |
| 2278 |
| 2279 2015-04-03 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 2280 |
| 2281 * src/autofit/afhints.c (af_glyph_hints_reload): Use do-while loop. |
| 2282 |
| 2283 2015-04-02 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 2284 |
| 2285 * src/autofit/aflatin.c (af_latin_hint_edges): Reduce logic. |
| 2286 |
| 2287 2015-04-01 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 2288 |
| 2289 [autofit] Finish the thought. |
| 2290 |
| 2291 * src/autofit/afhints.c (af_direction_compute): make sure the long arm |
| 2292 is never negative so that its `FT_ABS' is not necessary. |
| 2293 |
| 2294 2015-04-01 Werner Lemberg <wl@gnu.org> |
| 2295 |
| 2296 [autofit] Call dumper functions for tracing. |
| 2297 |
| 2298 * src/autofit/afcjk.c (af_cjk_hints_apply): Remove dead code. |
| 2299 * src/autofit/afhints.c (af_glyph_hints_dump_points): Minor |
| 2300 improvement. |
| 2301 * src/autofit/afmodule.c (af_autofitter_load_glyph): Implement it. |
| 2302 |
| 2303 2015-04-01 Werner Lemberg <wl@gnu.org> |
| 2304 |
| 2305 [autofit] Make debugging stuff work again. |
| 2306 |
| 2307 The interface to ftgrid was broken in the series of commits starting |
| 2308 with |
| 2309 |
| 2310 [autofit] Allocate AF_Loader on the stack instead of AF_Module. |
| 2311 |
| 2312 from 2015-01-14. |
| 2313 |
| 2314 * src/autofit/afmodule.c (_af_debug_hints_rec) [FT_DEBUG_AUTOFIT]: |
| 2315 Use a global AF_GlyphHintsRec object for debugging. |
| 2316 (af_autofitter_done, af_autofitter_load_glyph): Updated. |
| 2317 |
| 2318 * src/autofit/afloader.c (af_loader_init, af_loader_done): Updated. |
| 2319 |
| 2320 2015-04-01 Werner Lemberg <wl@gnu.org> |
| 2321 |
| 2322 * src/autofit/afhints.c (af_glyph_hints_done): Fix minor thinko. |
| 2323 |
| 2324 2015-03-29 Werner Lemberg <wl@gnu.org> |
| 2325 |
| 2326 [cff] Fix Savannah bug #44629. |
| 2327 |
| 2328 * src/cff/cf2font.h (CF2_MAX_SUBR), src/cff/cffgload.h |
| 2329 (CFF_MAX_SUBRS_CALLS): Set to 16. |
| 2330 |
| 2331 2015-03-29 Werner Lemberg <wl@gnu.org> |
| 2332 |
| 2333 [type1, truetype] Make the MM API more flexible w.r.t. `num_coords'. |
| 2334 |
| 2335 This commit allows `num_coords' to be larger or smaller than the |
| 2336 number of available axes while selecting a design instance, either |
| 2337 ignoring excess data or using defaults if data is missing. |
| 2338 |
| 2339 * src/truetype/ttgxvar.c (TT_Set_MM_Blend, TT_Set_Var_Design): |
| 2340 Implement it. |
| 2341 |
| 2342 * src/type1/t1load.c (T1_Set_MM_Blend, T1_Set_MM_Design, |
| 2343 T1_Set_Var_Design): Ditto. |
| 2344 |
| 2345 2015-03-29 Werner Lemberg <wl@gnu.org> |
| 2346 |
| 2347 [type1] Minor. |
| 2348 |
| 2349 * src/type1/t1load.c (T1_Set_MM_Blend, T1_Set_MM_Design): Use |
| 2350 FT_THROW. |
| 2351 (T1_Set_Var_Design): Use T1_MAX_MM_AXIS and FT_THROW. |
| 2352 |
| 2353 2015-03-27 Werner Lemberg <wl@gnu.org> |
| 2354 |
| 2355 [cff] Trace charstring nesting levels. |
| 2356 |
| 2357 * src/cff/cf2intrp.c (cf2_interpT2CharString) <cf2_cmdCALLGSUBR, |
| 2358 cf2_cmdCALLSUBR, cf2_cmdRETURN>: Implement it. |
| 2359 |
| 2360 * src/cff/cffgload.c (cff_decoder_parse_charstrings) |
| 2361 <cff_op_callsubr, cff_op_callgsubr, cff_op_return>: Ditto. |
| 2362 |
| 2363 2015-03-21 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 2364 |
| 2365 [base] Optimize `FT_Angle_Diff'. |
| 2366 |
| 2367 Under normal circumstances we are usually close to the desired range |
| 2368 of angle values, so that the remainder is not really necessary. |
| 2369 |
| 2370 * src/base/fttrigon.c (FT_Angle_Diff): Use loops instead of remainder. |
| 2371 |
| 2372 * src/autofit/aftypes.h (AF_ANGLE_DIFF): Ditto in the unused macro. |
| 2373 |
| 2374 2015-03-21 Werner Lemberg <wl@gnu.org> |
| 2375 |
| 2376 [truetype] Improve `gvar' handling. |
| 2377 |
| 2378 * src/truetype/ttgxvar.c (ft_var_readpackedpoints): Correctly handle |
| 2379 single-element runs. Cf. glyph `Q' in Skia.ttf with weights larger |
| 2380 than the default. |
| 2381 |
| 2382 2015-03-20 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 2383 |
| 2384 * src/base/fttrigon.c (FT_Vector_Rotate): Minor refactoring. |
| 2385 |
| 2386 2015-03-17 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 2387 |
| 2388 Fix Savannah bug #44412 (part 2). |
| 2389 |
| 2390 * src/base/fttrigon.c (FT_Sin, FT_Cos, FT_Tan): Call `FT_Vector_Unit'. |
| 2391 |
| 2392 2015-03-11 Werner Lemberg <wl@gnu.org> |
| 2393 |
| 2394 [autofit] Add support for Arabic script. |
| 2395 |
| 2396 Thanks to Titus Nemeth <tn@tntypography.eu> for guidance! |
| 2397 |
| 2398 * src/autofit/afblue.dat: Add blue zone data for Arabic. |
| 2399 |
| 2400 * src/autofit/afblue.c, src/autofit/afblue.h: Regenerated. |
| 2401 |
| 2402 * src/autofit/afscript.h: Add Arabic standard characters. |
| 2403 |
| 2404 * src/autofit/afranges.c: Add Arabic data. |
| 2405 |
| 2406 * src/autofit/afstyles.h: Add Arabic data. |
| 2407 |
| 2408 * docs/CHANGES: Document it. |
| 2409 |
| 2410 2015-03-11 Werner Lemberg <wl@gnu.org> |
| 2411 |
| 2412 Rename `svxf86nm.h' to `svfntfmt.h'; update related symbols. |
| 2413 |
| 2414 * include/internal/ftserv.h (FT_SERVICE_XFREE86_NAME_H): Renamed |
| 2415 to... |
| 2416 (FT_SERVICE_FONT_FORMAT_H): This. |
| 2417 |
| 2418 * include/internal/services/svfntfmt.h (FT_XF86_FORMAT_*): Renamed |
| 2419 to ... |
| 2420 (FT_FONT_FORMAT_*): This. |
| 2421 |
| 2422 src/base/ftfntfmt.c, src/bdf/bdfdrivr.c, src/cff/cffdrivr.c, |
| 2423 src/cid/cidriver.c, src/pcf/pcfdrivr.c, src/pfr/pfrdrivr.c, |
| 2424 src/truetype/ttdriver.c, src/type1/t1driver.c, |
| 2425 src/type42/t42drivr.c, src/winfonts/winfnt.c: Updated. |
| 2426 |
| 2427 2015-03-11 Werner Lemberg <wl@gnu.org> |
| 2428 |
| 2429 [base] Rename `FT_XFREE86_H' to `FT_FONT_FORMATS_H'. |
| 2430 |
| 2431 * include/config/ftheader.h: Implement it. |
| 2432 * src/base/ftfntfmt.c, docs/CHANGES: Updated. |
| 2433 |
| 2434 2015-03-11 Werner Lemberg <wl@gnu.org> |
| 2435 |
| 2436 [base] Rename `FT_Get_X11_Font_Format' to `FT_Get_Font_Format'. |
| 2437 |
| 2438 * include/ftfntfmt.h, src/base/ftfntfmt.c: Implement it. |
| 2439 |
| 2440 * docs/CHANGES: Updated. |
| 2441 |
| 2442 2015-03-11 Werner Lemberg <wl@gnu.org> |
| 2443 |
| 2444 Fix automatic copyright updating. |
| 2445 |
| 2446 * src/tools/update-copyright: Make scanning of `no-copyright' |
| 2447 actually work. |
| 2448 |
| 2449 * src/tools/no-copyright: Don't include README in general. |
| 2450 |
| 2451 2015-03-11 Werner Lemberg <wl@gnu.org> |
| 2452 |
| 2453 Rename `ftxf86.[ch]' to `ftfntfmt.[ch]'. |
| 2454 |
| 2455 CMakeLists.txt, builds/amiga/makefile, builds/amiga/makefile.os4, |
| 2456 builds/amiga/smakefile, builds/mac/FreeType.m68k_cfm.make.txt, |
| 2457 builds/mac/FreeType.m68k_far.make.txt, |
| 2458 builds/mac/FreeType.ppc_carbon.make.txt, |
| 2459 builds/mac/FreeType.ppc_classic.make.txt, builds/symbian/bld.inf, |
| 2460 builds/symbian/freetype.mmp, builds/wince/vc2005-ce/freetype.vcproj, |
| 2461 builds/wince/vc2008-ce/freetype.vcproj, |
| 2462 builds/windows/vc2005/freetype.vcproj, |
| 2463 builds/windows/vc2008/freetype.vcproj, |
| 2464 builds/windows/vc2010/freetype.vcxproj, |
| 2465 builds/windows/vc2010/freetype.vcxproj.filters, |
| 2466 builds/windows/visualc/freetype.dsp, |
| 2467 builds/windows/visualc/freetype.vcproj, |
| 2468 builds/windows/visualce/freetype.dsp, |
| 2469 builds/windows/visualce/freetype.vcproj, docs/INSTALL.ANY, |
| 2470 include/config/ftheader.h, include/ftfntfmt.h, modules.cfg, |
| 2471 src/base/ftfntfmt.c, vms_make.com: Updated. |
| 2472 |
| 2473 2015-03-10 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 2474 |
| 2475 Fix Savannah bug #44412 (part 1). |
| 2476 |
| 2477 * src/base/ftstroke.c (ft_stroker_inside): Handle near U-turns. |
| 2478 |
| 2479 2015-03-10 Werner Lemberg <wl@gnu.org> |
| 2480 |
| 2481 [base] Rename `FT_Bitmap_New' to `FT_Bitmap_Init'. |
| 2482 |
| 2483 * include/ftbitmap.h, src/base/ftbitmap.c: Implement it. |
| 2484 Update all callers. |
| 2485 |
| 2486 * docs/CHANGES: Updated. |
| 2487 |
| 2488 2015-03-06 Werner Lemberg <wl@gnu.org> |
| 2489 |
| 2490 * src/sfnt/ttload.c (tt_face_load_font_dir): Fix compiler warning. |
| 2491 |
| 2492 Found by Alexei. |
| 2493 |
| 2494 2015-03-05 Alexei Podtelezhnikov <apodtele@gmail.com> |
| 2495 |
| 2496 * src/base/ftstroke.c: Simplify. |
| 2497 |
| 2498 2015-03-04 Werner Lemberg <wl@gnu.org> |
| 2499 |
| 2500 [truetype] Some fixes and code refactoring in `ttgxvar.c'. |
| 2501 |
| 2502 * src/truetype/ttgxvar.c (ft_var_readpackedpoints): Fix return value |
| 2503 of `point_cnt' if two bytes are read. |
| 2504 Use a more vertical coding style. |
| 2505 (ft_var_readpackeddeltas): Use FT_UInt for `delta_cnt' parameter. |
| 2506 Use a more vertical coding style. |
| 2507 |
| 2508 2015-03-03 Werner Lemberg <wl@gnu.org> |
| 2509 |
| 2510 [autofit] Fix Savannah bug #44241. |
| 2511 |
| 2512 * src/autofit/aflatin.c (af_latin_metrics_init_blues): Reject glyphs |
| 2513 with less than 3 points. |
| 2514 |
| 2515 2015-03-02 Werner Lemberg <wl@gnu.org> |
| 2516 |
| 2517 Simplify `TYPEOF' macro. |
| 2518 |
| 2519 No need for two arguments. |
| 2520 |
| 2521 * include/config/ftconfig.h, builds/unix/ftconfig.in, |
| 2522 builds/vms/ftconfig.h (TYPEOF): Updated. |
| 2523 |
| 2524 * include/internal/ftobjs.h (FT_PAD_FLOOR, FT_PIX_FLOOR), |
| 2525 src/autofit/afwarp.h (AF_WARPER_FLOOR): Updated. |
| 2526 |
| 2527 2015-03-01 Werner Lemberg <wl@gnu.org> |
| 2528 |
| 2529 Various compiler warning fixes for `make multi'. |
| 2530 |
| 2531 * src/autofit/afcjk.c (af_cjk_hints_compute_blue_edges), |
| 2532 src/autofit/aflatin.c (af_latin_hint_compute_blue_edges, |
| 2533 af_latin_hint_edges), src/autofit/aflatin2.c |
| 2534 (af_latin2_hints_compute_blue_edges, af_latin2_hint_edges): Declare |
| 2535 as `static'. |
| 2536 |
| 2537 * src/cache/ftccmap.c (FTC_CMAP_QUERY_HASH, FTC_CMAP_NODE_HASH): |
| 2538 Removed. Unused. |
| 2539 * src/cache/ftcimage.c: Include FT_INTERNAL_OBJECTS_H. |
| 2540 * src/cache/ftcmanag.c (FTC_LRU_GET_MANAGER): Removed. Unused. |
| 2541 |
| 2542 * src/cff/cf2intrp.c: Include `cf2intrp.h'. |
| 2543 * src/cff/cffdrivr.c (PAIR_TAG): Removed. Unused. |
| 2544 |
| 2545 * src/gzip/ftgzip.c (NO_DUMMY_DECL): Removed. Unused. |
| 2546 |
| 2547 * src/psaux/afmparse.c (afm_parser_read_int): Declare as `static'. |
| 2548 |
| 2549 * src/pshinter/pshalgo.c (STRONGER, PSH_ZONE_MIN, PSH_ZONE_MAX): |
| 2550 Removed. Unused. |
| 2551 |
| 2552 * src/raster/ftraster.c (Render_Glyph): Declare as `static'. |
| 2553 |
| 2554 * src/sfnt/ttpost.c (load_format_20): Fix signedness warning. |
| 2555 |
| 2556 * src/truetype/ttdriver.c (PAIR_TAG): Removed. Unused. |
| 2557 * src/truetype/ttsubpix.c (is_member_of_family_class, |
| 2558 is_member_of_style_class): Declare as `static'. |
| 2559 |
| 2560 * src/type1/t1gload.c (T1_Parse_Glyph_And_Get_Char_String): Declare |
| 2561 as `static'. |
| 2562 * src/type1/t1load.c (mm_axis_unmap, mm_weights_unmap): Declare as |
| 2563 `static'. |
| 2564 (T1_FIELD_COUNT): Removed. Unused. |
| 2565 * src/type1/t1parse.h (T1_Done_Table): Removed. Unused. |
| 2566 |
| 2567 * src/type42/t42parse.c (T1_Done_Table): Removed. Unused. |
| 2568 |
| 2569 2015-02-25 Werner Lemberg <wl@gnu.org> |
| 2570 |
| 2571 [psaux] Signedness fixes. |
| 2572 |
| 2573 * include/internal/psaux.h, src/psaux/afmparse.c, |
| 2574 src/psaux/afmparse.h, src/psaux/psconv.c, src/psaux/psobjs.c, |
| 2575 src/psaux/t1cmap.c, src/psaux/t1decode.c: Apply. |
| 2576 |
| 2577 2015-02-25 Werner Lemberg <wl@gnu.org> |
| 2578 |
| 2579 [otvalid] Signedness fixes. |
| 2580 |
| 2581 * src/otvalid/otvcommn.c, src/otvalid/otvgdef.c, |
| 2582 src/otvalid/otvgpos.c, src/otvalid/otvgsub.c, src/otvalid/otvmath.c: |
| 2583 Apply. |
| 2584 |
| 2585 2015-02-25 Werner Lemberg <wl@gnu.org> |
| 2586 |
| 2587 * src/bzip2/ftbzip2.c (ft_bzip2_alloc): Signedness fix. |
| 2588 |
| 2589 2015-02-25 Werner Lemberg <wl@gnu.org> |
| 2590 |
| 2591 [lzw] Signedness fixes. |
| 2592 |
| 2593 * src/lzw/ftzopen.c, src/lzw/ftzopen.h: Apply. |
| 2594 |
| 2595 2015-02-25 Werner Lemberg <wl@gnu.org> |
| 2596 |
| 2597 [gxvalid] Signedness fixes. |
| 2598 |
| 2599 * src/gxvalid/gxvbsln.c, src/gxvalid/gxvcommn.c, |
| 2600 src/gxvalid/gxvcommn.h, src/gxvalid/gxvjust.c, |
| 2601 src/gxvalid/gxvkern.c, src/gxvalid/gxvlcar.c, src/gxvalid/gxvmort.c, |
| 2602 src/gxvalid/gxvmort1.c, src/gxvalid/gxvmort2.c, |
| 2603 src/gxvalid/gxvmorx.c, src/gxvalid/gxvmorx1.c, |
| 2604 src/gxvalid/gxvmorx2.c, src/gxvalid/gxvopbd.c, |
| 2605 src/gxvalid/gxvprop.c, src/gxvalid/gxvtrak.c: Apply. |
| 2606 |
| 2607 2015-02-25 Werner Lemberg <wl@gnu.org> |
| 2608 |
| 2609 [cache] Signedness fixes. |
| 2610 |
| 2611 * src/cache/ftcbasic.c, src/cache/ftccmap.c, src/cache/ftcimage.c, |
| 2612 src/cache/ftcmanag.c, src/cache/ftcsbits.c: Apply. |
| 2613 |
| 2614 2015-02-25 Werner Lemberg <wl@gnu.org> |
| 2615 |
| 2616 Change dimension fields in `FTC_ImageTypeRec' to unsigned type. |
| 2617 |
| 2618 This doesn't break ABI. |
| 2619 |
| 2620 * include/ftcache.h (FTC_ImageTypeRec): Use unsigned types for |
| 2621 `width' and `height'. |
| 2622 |
| 2623 * docs/CHANGES: Document it. |
| 2624 |
| 2625 2015-02-25 Werner Lemberg <wl@gnu.org> |
| 2626 |
| 2627 [cache] Don't use `labs'. |
| 2628 |
| 2629 This is the only place in FreeType where this function was used. |
| 2630 |
| 2631 * include/config/ftstdlib.h (ft_labs): Remove. |
| 2632 |
| 2633 * src/cache/ftcimage.c (ftc_inode_weight): Replace `ft_labs' with |
| 2634 `FT_ABS'. |
| 2635 |
| 2636 2015-02-23 Werner Lemberg <wl@gnu.org> |
| 2637 |
| 2638 [cache] Replace `FT_PtrDist' with `FT_Offset'. |
| 2639 |
| 2640 * src/cache/ftccache.h (FTC_NodeRec): `FT_Offset' (a.k.a. `size_t') |
| 2641 is a better choice for `hash' to hold a pointer than `FT_PtrDist' |
| 2642 (a.k.a. `ptrdiff_t'), especially since the latter is signed, |
| 2643 causing zillions of signedness warnings. [Note that `hash' was of |
| 2644 type `FT_UInt32' before the change to `FT_PtrDist'.] |
| 2645 Update all users. |
| 2646 |
| 2647 * src/cache/ftcbasic.c, src/cache/ftccache.c, src/cache/ftccmap.c, |
| 2648 src/cache/ftcglyph.c, src/cache/ftcglyph.h: Updated. |
| 2649 |
| 2650 2015-02-23 Werner Lemberg <wl@gnu.org> |
| 2651 |
| 2652 [smooth, raster] Re-enable standalone compilation. |
| 2653 |
| 2654 * src/raster/ftraster.c (FT_RENDER_POOL_SIZE, FT_MAX) |
| 2655 [_STANDALONE_]: Define macros. |
| 2656 |
| 2657 * src/smooth/ftgrays.c (FT_RENDER_POOL_SIZE, FT_MAX, FT_ABS, |
| 2658 FT_HYPOT) [_STANDALONE_]: Define macros. |
| 2659 |
| 2660 2015-02-22 Werner Lemberg <wl@gnu.org> |
| 2661 |
| 2662 [smooth] Signedness fixes. |
| 2663 |
| 2664 * src/smooth/ftgrays.c, src/smooth/ftsmooth.c: Apply. |
| 2665 |
| 2666 2015-02-22 Werner Lemberg <wl@gnu.org> |
| 2667 |
| 2668 * src/raster/ftraster.c: Use the file's typedefs everywhere. |
| 2669 |
| 2670 2015-02-22 Werner Lemberg <wl@gnu.org> |
| 2671 |
| 2672 * src/sfnt/ttpost.c (load_format_20): Fix error tracing message. |
| 2673 |
| 2674 Bug introduced 6 commits earlier. |
| 2675 |
| 2676 2015-02-22 Werner Lemberg <wl@gnu.org> |
| 2677 |
| 2678 [pshinter] Fix thinko. |
| 2679 |
| 2680 * src/pshinter/pshalgo.c (psh_glyph_find_strong_points): Correctly |
| 2681 check `count'. |
| 2682 Bug introduced two commits earlier. |
| 2683 |
| 2684 2015-02-22 Werner Lemberg <wl@gnu.org> |
| 2685 |
| 2686 [raster] Signedness fixes. |
| 2687 |
| 2688 * src/raster/ftraster.c, src/raster/ftrend1.c: Apply. |
| 2689 |
| 2690 2015-02-22 Werner Lemberg <wl@gnu.org> |
| 2691 |
| 2692 [pshinter] Signedness fixes. |
| 2693 |
| 2694 * src/pshinter/pshalgo.c, src/pshinter/pshglob.c, |
| 2695 src/pshinter/pshrec.c: Apply. |
| 2696 |
| 2697 2015-02-22 Werner Lemberg <wl@gnu.org> |
| 2698 |
| 2699 [pshinter] Use macros for (unsigned) flags, not enumerations. |
| 2700 |
| 2701 * src/pshinter/pshalgo.h (PSH_Hint_Flags): Replace with macros. |
| 2702 Updated. |
| 2703 * src/pshinter/pshrec.h (PS_Hint_Flags): Replace with macros. |
| 2704 |
| 2705 2015-02-22 Werner Lemberg <wl@gnu.org> |
| 2706 |
| 2707 * src/pshinter/pshrec.c: Simplify. |
| 2708 (ps_hints_open, ps_hints_stem): Remove switch statement. |
| 2709 |
| 2710 2015-02-22 Werner Lemberg <wl@gnu.org> |
| 2711 |
| 2712 [sfnt] Signedness fixes. |
| 2713 |
| 2714 * src/sfnt/pngshim.c, src/sfnt/sfobjs.c, src/sfnt/ttcmap.c, |
| 2715 src/sfnt/ttkern.c, src/sfnt/ttload.c, src/sfnt/ttpost.c, |
| 2716 src/sfnt/ttsbit.c: Apply. |
| 2717 * src/sfnt/sfdriver.c: Apply. |
| 2718 (sfnt_get_ps_name): Simplify. |
| 2719 |
| 2720 2015-02-22 Werner Lemberg <wl@gnu.org> |
| 2721 |
| 2722 [bdf] Signedness fixes. |
| 2723 |
| 2724 * src/bdf/bdf.h, src/bdf/bdfdrivr.c, src/bdf/bdfdrivr.h, |
| 2725 src/bdf/bdflib.c: Apply. |
| 2726 |
| 2727 2015-02-22 Werner Lemberg <wl@gnu.org> |
| 2728 |
| 2729 * src/bdf/bdflib.c (_bdf_atous): New function. |
| 2730 (_bdf_parse_glyphs, _bdf_parse_start): Use it. |
| 2731 |
| 2732 2015-02-22 Werner Lemberg <wl@gnu.org> |
| 2733 |
| 2734 [pcf] Signedness fixes. |
| 2735 |
| 2736 * src/pcf/pcf.h, src/pcf/pcfdrivr.c: Apply. |
| 2737 * src/pcf/pcfread.c: Apply. |
| 2738 (pcf_get_encodings): Ignore invalid negative encoding offsets. |
| 2739 |
| 2740 2015-02-21 Werner Lemberg <wl@gnu.org> |
| 2741 |
| 2742 * src/winfonts/winfnt.c: Signedness fixes. |
| 2743 |
| 2744 2015-02-21 Werner Lemberg <wl@gnu.org> |
| 2745 |
| 2746 [type42] Signedness fixes. |
| 2747 |
| 2748 * src/type42/t42parse.c, src/type42/t42parse.h, |
| 2749 src/type42/t42types.h: Apply. |
| 2750 |
| 2751 2015-02-21 Werner Lemberg <wl@gnu.org> |
| 2752 |
| 2753 [pfr] Signedness fixes. |
| 2754 |
| 2755 * src/pfr/pfrdrivr.c, src/pfr/pfrgload.c, src/pfr/pfrload.c, |
| 2756 src/pfr/pfrload.h, src/pfr/pfrobjs.c, src/pfr/pfrsbit.c, |
| 2757 src/pfr/pfrtypes.h: Apply. |
| 2758 |
| 2759 2015-02-21 Werner Lemberg <wl@gnu.org> |
| 2760 |
| 2761 [cff] Minor signedness fixes related to last commit. |
| 2762 |
| 2763 * src/cff/cf2ft.c, src/cff/cf2intrp.c, src/cff/cffgload.c: Apply. |
| 2764 |
| 2765 2015-02-20 Werner Lemberg <wl@gnu.org> |
| 2766 |
| 2767 [cff] Thinkos in bias handling. |
| 2768 |
| 2769 Only the final result is always positive. |
| 2770 |
| 2771 Bug introduced three commits earlier. |
| 2772 |
| 2773 * src/cff/cffgload.c, src/cff/cffgload.h: Apply. |
| 2774 |
| 2775 2015-02-20 Werner Lemberg <wl@gnu.org> |
| 2776 |
| 2777 [cid] Fix signedness issues and emit some better error codes. |
| 2778 |
| 2779 * src/cid/cidgload.c, src/cid/cidload.h, src/cid/cidobjs.c, |
| 2780 src/cid/cidparse.h: Apply. |
| 2781 * src/cid/cidload.c: Apply. |
| 2782 (parse_fd_array): Reject negative values for number of dictionaries. |
| 2783 * src/cid/cidparse.c: Apply. |
| 2784 (cid_parser_new): Reject negative values for hex data length. |
| 2785 |
| 2786 2015-02-20 Werner Lemberg <wl@gnu.org> |
| 2787 |
| 2788 [cff] Signedness fixes for new engine. |
| 2789 |
| 2790 * src/cff/cf2arrst.c, src/cff/cf2fixed.h, src/cff/cf2ft.c, |
| 2791 src/cff/cf2ft.h, src/cff/cf2hints.c, src/cff/cf2intrp.c: Apply. |
| 2792 |
| 2793 2015-02-20 Werner Lemberg <wl@gnu.org> |
| 2794 |
| 2795 [cff] Signedness fixes for basic infrastructure and old engine. |
| 2796 |
| 2797 * include/internal/pshints.h, src/cff/cffdrivr.c, |
| 2798 src/cff/cffgload.c, src/cff/cffgload.h, src/cff/cffload.c, |
| 2799 src/cff/cffobjs.c, src/cff/cffparse.c, src/pshinter/pshrec.c: Apply. |
| 2800 |
| 2801 2015-02-19 Werner Lemberg <wl@gnu.org> |
| 2802 |
| 2803 * src/truetype/ttgxvar.c (TT_Get_MM_Var): Ignore `countSizePairs'. |
| 2804 |
| 2805 This is hard-coded to value 2 in `fvar' version 1.0 (and no newer |
| 2806 version exists), but some fonts set it incorrectly. |
| 2807 |
| 2808 Problem reported by Adam Twardoch <adam@fontlab.com>. |
| 2809 |
| 2810 2015-02-19 Werner Lemberg <wl@gnu.org> |
| 2811 |
| 2812 [cff] Emit better error code for invalid private dict size. |
| 2813 |
| 2814 * src/cff/cffparse.c (cff_parse_private_dict): Reject negative |
| 2815 values for size and offset. |
| 2816 |
| 2817 2015-02-19 Werner Lemberg <wl@gnu.org> |
| 2818 |
| 2819 [autofit] Fix signedness issues. |
| 2820 |
| 2821 * src/autofit/afangles.c, src/autofit/afcjk.c, |
| 2822 src/autofit/afglobal.c, src/autofit/afhints.c, |
| 2823 src/autofit/aflatin.c, src/autofit/aflatin2.c, src/autofit/afwarp.c, |
| 2824 src/autofit/hbshim.c: Apply. |
| 2825 |
| 2826 2015-02-19 Werner Lemberg <wl@gnu.org> |
| 2827 |
| 2828 [autofit] Use macros for (unsigned) flags, not enumerations. |
| 2829 |
| 2830 This harmonizes with other code in FreeType (and reduces the number |
| 2831 of necessary casts to avoid compiler warnings). |
| 2832 |
| 2833 * src/autofit/afblue.hin: Make flag macros unsigned. |
| 2834 * src/autofit/afblue.h: Regenerated. |
| 2835 |
| 2836 * src/autofit/afcjk.h: Replace flag enumeration with macros. |
| 2837 * src/autofit/afcjk.c: Updated. |
| 2838 |
| 2839 * src/autofit/afhints.h (AF_Flags, AF_Edge_Flags): Replace with |
| 2840 macros. |
| 2841 * src/autofit/afhints.c: Updated. |
| 2842 |
| 2843 * src/autofit/aflatin.h: Replace flag enumerations with macros. |
| 2844 * src/autofit/aflatin.c, src/autofit/aflatin2.c: Updated. |
| 2845 |
| 2846 * src/autofit/aftypes.h (AF_ScalerFlags): Replace with macros. |
| 2847 |
| 2848 2015-02-18 Werner Lemberg <wl@gnu.org> |
| 2849 |
| 2850 [type1] Fix signedness issues. |
| 2851 |
| 2852 * include/internal/psaux.h, include/internal/t1types.h, |
| 2853 src/psaux/psobjs.c, src/psaux/psobjs.h, src/psaux/t1decode.c, |
| 2854 src/type1/t1gload.c, src/type1/t1load.c, src/type1/t1parse.c: Apply. |
| 2855 |
| 2856 2015-02-18 Werner Lemberg <wl@gnu.org> |
| 2857 |
| 2858 [psaux, type1] Fix minor AFM issues. |
| 2859 |
| 2860 * include/internal/t1types.h (AFM_KernPairRec): Make indices |
| 2861 unsigned. |
| 2862 Update users. |
| 2863 (AFM_FontInfoRec): Make element counters unsigned. |
| 2864 Update users. |
| 2865 * src/psaux/afmparse.h (AFM_ValueRec): Add union member for unsigned |
| 2866 int. |
| 2867 |
| 2868 * src/psaux/afmparse.c (afm_parse_track_kern, afm_parse_kern_pairs): |
| 2869 Reject negative values for number of kerning elements. |
| 2870 |
| 2871 * src/type1/t1afm.c, src/tools/test_afm.c: Updated. |
| 2872 |
| 2873 2015-02-18 Werner Lemberg <wl@gnu.org> |
| 2874 |
| 2875 Don't use `FT_PtrDist' for lengths. |
| 2876 |
| 2877 Use FT_UInt instead. |
| 2878 |
| 2879 * include/internal/psaux.h (PS_Table_FuncsRec, PS_TableRec, |
| 2880 T1_DecoderRec): Do it. |
| 2881 |
| 2882 * include/internal/t1types.h (T1_FontRec): Ditto. |
| 2883 |
| 2884 * src/cid/cidload.c (cid_parse_dict): Updated. |
| 2885 * src/pfr/pfrload.c (pfr_extra_item_load_font_id): Ditto. |
| 2886 * src/psaux/psobjs.c (ps_table_add), src/psaux/psobjs.h: Ditto. |
| 2887 * src/type1/t1load.c (parse_blend_axis_types, parse_encoding, |
| 2888 parse_chharstrings, parse_dict): Ditto. |
| 2889 * src/type42/t42parse.c (t42_parse_encoding, t42_parse_charstrings, |
| 2890 t42_parse_dict): Ditto. |
| 2891 |
| 2892 2015-02-18 Werner Lemberg <wl@gnu.org> |
| 2893 |
| 2894 * src/type1/t1driver.c (t1_ps_get_font_value): Clean up. |
| 2895 This handles negative values better, avoiding many casts. |
| 2896 |
| 2897 2015-02-17 Werner Lemberg <wl@gnu.org> |
| 2898 |
| 2899 [base] Fix Savannah bug #44284. |
| 2900 |
| 2901 * src/base/ftcalc.c (FT_MulFix): Typos. |
| 2902 |
| 2903 2015-02-17 Werner Lemberg <wl@gnu.org> |
| 2904 |
| 2905 [truetype] Finish compiler warning fixes for signedness issues. |
| 2906 |
| 2907 * src/truetype/ttgxvar.c, src/truetype/ttsubpix.c, |
| 2908 src/truetype/ttsubpix.h: Apply. |
| 2909 |
| 2910 2015-02-17 Werner Lemberg <wl@gnu.org> |
| 2911 |
| 2912 * src/truetype/ttsubpix.c: Adding missing `static' keywords. |
| 2913 |
| 2914 2015-02-17 Werner Lemberg <wl@gnu.org> |
| 2915 |
| 2916 [truetype] More signedness fixes. |
| 2917 |
| 2918 * include/internal/tttypes.h, src/truetype/ttinterp.h, |
| 2919 src/truetype/ttobjs.h, src/truetype/ttinterp.c, |
| 2920 src/truetype/ttobjs.c: Apply. |
| 2921 |
| 2922 2015-02-17 Werner Lemberg <wl@gnu.org> |
| 2923 |
| 2924 [truetype] Various signedness fixes. |
| 2925 |
| 2926 * include/internal/ftgloadr.h, src/truetype/ttpload.c: Apply. |
| 2927 |
| 2928 * src/truetype/ttgload.c: Apply. |
| 2929 (TT_Get_VMetrics): Protect against invalid ascenders and descenders |
| 2930 while constructing advance height. |
| 2931 |
| 2932 2015-02-16 Werner Lemberg <wl@gnu.org> |
| 2933 |
| 2934 [base] Finish compiler warning fixes for signedness issues. |
| 2935 |
| 2936 * src/base/ftglyph.c, src/base/ftlcdfil.c, src/base/ftstroke.c: |
| 2937 Apply. |
| 2938 |
| 2939 2015-02-16 Werner Lemberg <wl@gnu.org> |
| 2940 |
| 2941 * include/tttables.h (TT_OS2): `fsType' must be FT_UShort. |
| 2942 |
| 2943 2015-02-16 Werner Lemberg <wl@gnu.org> |
| 2944 |
| 2945 More minor signedness warning fixes. |
| 2946 |
| 2947 * src/base/ftbbox.c, src/base/ftbitmap.c, src/base/fttrigon.c, |
| 2948 src/base/ftutil.c: Apply. |
| 2949 |
| 2950 2015-02-16 Werner Lemberg <wl@gnu.org> |
| 2951 |
| 2952 Next round of minor compiler warning fixes. |
| 2953 |
| 2954 * include/internal/ftrfork.h (FT_RFork_Ref): Change `offset' member |
| 2955 type to `FT_Long'. |
| 2956 (CONST_FT_RFORK_RULE_ARRAY_BEGIN): Add `static' keyword. |
| 2957 |
| 2958 * include/internal/ftstream.h (FT_Stream_Pos): Return `FT_ULong'. |
| 2959 |
| 2960 * src/base/ftoutln.c, src/base/ftrfork.c, src/base/ftstream.c: |
| 2961 Signedess fixes. |
| 2962 |
| 2963 2015-02-16 Werner Lemberg <wl@gnu.org> |
| 2964 |
| 2965 Various minor signedness fixes. |
| 2966 |
| 2967 * include/ftadvanc.h, include/internal/ftobjs.h, |
| 2968 src/base/ftgloadr.c, src/base/ftobjs.c: Apply. |
| 2969 |
| 2970 2015-02-16 Werner Lemberg <wl@gnu.org> |
| 2971 |
| 2972 New `TYPEOF' macro. |
| 2973 |
| 2974 This helps suppress signedness warnings, avoiding issues with |
| 2975 implicit conversion changes. |
| 2976 |
| 2977 * include/config/ftconfig.h, builds/unix/ftconfig.in, |
| 2978 builds/vms/ftconfig.h (TYPEOF): Define. |
| 2979 |
| 2980 * include/internal/ftobjs.h (FT_PAD_FLOOR, FT_PIX_FLOOR), |
| 2981 src/autofit/afwarp.h (AF_WARPER_FLOOR): Use it. |
| 2982 |
| 2983 2015-02-16 Werner Lemberg <wl@gnu.org> |
| 2984 |
| 2985 * src/base/ftsystem.c: Use casts in standard C function wrappers. |
| 2986 (ft_alloc, ft_realloc, ft_ansi_stream_io, FT_Stream_Open): Do it. |
| 2987 |
| 2988 2015-02-16 Werner Lemberg <wl@gnu.org> |
| 2989 |
| 2990 Fix Savannah bug #44261. |
| 2991 |
| 2992 * builds/unix/detect.mk (setup) [unix]: Set `CONFIG_SHELL' in the |
| 2993 environment also while calling the configure script. |
| 2994 |
| 2995 2015-02-16 Werner Lemberg <wl@gnu.org> |
| 2996 |
| 2997 * include/internal/ftmemory.h: Add some `FT_Offset' casts. |
| 2998 (FT_MEM_SET, FT_MEM_COPY, FT_MEM_MOVE, FT_ARRAY_ZERO, FT_ARRAY_COPY, |
| 2999 FT_MEM_MOVE): Do it. |
| 3000 |
| 3001 2015-02-15 Werner Lemberg <wl@gnu.org> |
| 3002 |
| 3003 [base] Clean up signedness issues in `ftdbgmem.c'. |
| 3004 |
| 3005 Also fix other minor issues. |
| 3006 |
| 3007 * src/base/ftdbgmem.c (FT_MemTableRec): Replace all FT_ULong types |
| 3008 with FT_Long for consistency. |
| 3009 (ft_mem_primes): Change type to `FT_Int'. |
| 3010 (ft_mem_closest_prime, ft_mem_table_set): Updated. |
| 3011 |
| 3012 (ft_mem_debug_panic, ft_mem_debug_alloc, ft_mem_debug_free, |
| 3013 ft_mem_debug_realloc): Use `static' keyword and fix signedness |
| 3014 warnings where necessary. |
| 3015 |
| 3016 (ft_mem_table_resize, ft_mem_table_new, ft_mem_table_destroy, |
| 3017 ft_mem_table_get_nodep, ft_mem_debug_init, FT_DumpMemory): Fix types |
| 3018 and add or remove casts to avoid signedness warnings. |
| 3019 |
| 3020 2015-02-15 Werner Lemberg <wl@gnu.org> |
| 3021 |
| 3022 [base] Clean up signedness in arithmetic functions. |
| 3023 |
| 3024 This makes the code more readable and reduces compiler warnings. |
| 3025 |
| 3026 * src/base/ftcalc.c (FT_MulDiv, FT_MulDiv_No_Round, FT_MulFix, |
| 3027 FT_DivFix): Convert input parameters to unsigned, do the |
| 3028 computation, then convert the result back to signed. |
| 3029 (ft_corner_orientation): Fix casts. |
| 3030 |
| 3031 2015-02-07 Werner Lemberg <wl@gnu.org> |
| 3032 |
| 3033 [sfnt] Fix Savannah bug #44184. |
| 3034 |
| 3035 * src/sfnt/ttload.c (check_table_dir, tt_face_load_font_dir): No |
| 3036 longer reject `htmx' and `vmtx' tables with invalid length but |
| 3037 sanitize them. |
| 3038 |
| 3039 2015-02-06 Jon Anderson <jon@websupergoo.com> |
| 3040 |
| 3041 [truetype] Fix regression in the incremental glyph loader. |
| 3042 |
| 3043 * src/truetype/ttgload.c (load_truetype_glyph): For incremental |
| 3044 fonts, the glyph index may be greater than the number of glyphs |
| 3045 indicated, so guard the check with a preprocessor conditional. |
| 3046 |
| 3047 2015-02-06 Werner Lemberg <wl@gnu.org> |
| 3048 |
| 3049 [autofit] Fix potential memory leak. |
| 3050 |
| 3051 While this doesn't show up with FreeType, exactly the same code |
| 3052 leaks with ttfautohint's modified auto-hinter code (which gets used |
| 3053 in a slightly different way). |
| 3054 |
| 3055 It certainly doesn't harm since it is similar to already existing |
| 3056 checks in the code for embedded arrays. |
| 3057 |
| 3058 * src/autofit/afhints.c (af_glyph_hints_reload): Set `max_contours' |
| 3059 and `max_points' for all cases. |
| 3060 |
| 3061 2015-01-31 Werner Lemberg <wl@gnu.org> |
| 3062 |
| 3063 [autofit] Add support for Thai script. |
| 3064 |
| 3065 Thanks to Ben Mitchell <ben@rosettatype.com> for guidance with blue |
| 3066 zone characters! |
| 3067 |
| 3068 * src/autofit/afblue.dat: Add blue zone data for Thai. |
| 3069 |
| 3070 * src/autofit/afblue.c, src/autofit/afblue.h: Regenerated. |
| 3071 |
| 3072 * src/autofit/afscript.h: Add Thai standard characters. |
| 3073 |
| 3074 * src/autofit/afranges.c: Add Thai data. |
| 3075 |
| 3076 * src/autofit/afstyles.h: Add Thai data. |
| 3077 |
| 3078 2015-01-23 Behdad Esfahbod <behdad@behdad.org> |
| 3079 |
| 3080 [raster] Handle `FT_RASTER_FLAG_AA' correctly. |
| 3081 |
| 3082 This fixes a breakage caused by the commit `[raster] Remove |
| 3083 5-level gray AA mode from monochrome rasterizer.'. |
| 3084 |
| 3085 Problem reported by Markus Trippelsdorf <markus@trippelsdorf.de>. |
| 3086 |
| 3087 * src/raster/ftraster.c (ft_black_render): Handle |
| 3088 `FT_RASTER_FLAG_AA'. |
| 3089 |
| 3090 * src/raster/ftrend1.c (ft_raster1_render): Remove gray AA mode |
| 3091 remnants. |
| 3092 |
| 3093 2015-01-18 Werner Lemberg <wl@gnu.org> |
| 3094 |
| 3095 * src/base/ftobjs.c (FT_New_Library): Fix compiler warning. |
| 3096 |
| 3097 2015-01-18 Chris Liddell <chris.liddell@artifex.com> |
| 3098 |
| 3099 [raster] Fix Savannah bug #44022. |
| 3100 |
| 3101 Add fallback for glyphs with degenerate bounding boxes. |
| 3102 |
| 3103 If a glyph has only one very narrow feature, the bbox can end up |
| 3104 with either the width or height of the bbox being 0, in which case |
| 3105 no raster memory is allocated and no attempt is made to render the |
| 3106 glyph. This is less than ideal when the drop-out compensation in |
| 3107 the rendering code would actually result in the glyph being |
| 3108 rendered. |
| 3109 |
| 3110 This problem can be observed with the `I' glyph (gid 47) in the |
| 3111 Autodesk RomanS TrueType font. |
| 3112 |
| 3113 * src/raster/ftrend1.c (ft_raster1_render): Add a fallback if either |
| 3114 dimension is zero to explicitly round up/down (instead of simply |
| 3115 round). |
| 3116 |
| 3117 2015-01-17 Werner Lemberg <wl@gnu.org> |
| 3118 |
| 3119 Add some tools to handle yearly copyright notice updates. |
| 3120 |
| 3121 We are now following the GNU guidelines: A new release automatically |
| 3122 means that the copyright year of all affected files gets updated; it |
| 3123 is no longer used to track years of modification changes. |
| 3124 |
| 3125 * src/tools/update-copyright-year: New Perl script. |
| 3126 * src/tools/update-copyright: New shell script that calls |
| 3127 `update-copyright-year' on all files. |
| 3128 * src/tools/no-copyright: Exceptions that should not be handled by |
| 3129 `update-copyright' |
| 3130 |
| 3131 2015-01-14 Werner Lemberg <wl@gnu.org> |
| 3132 |
| 3133 * docs/CHANGES: Updated, using a description from Behdad. |
| 3134 |
| 3135 2015-01-14 Behdad Esfahbod <behdad@behdad.org> |
| 3136 |
| 3137 * src/autofit/afmodule.c (af_autofitter_done): Fix compiler warning. |
| 3138 |
| 3139 2015-01-14 Behdad Esfahbod <behdad@behdad.org> |
| 3140 |
| 3141 [autofit] Add embedded array of segments and edges. |
| 3142 |
| 3143 Avoids multiple mallocs per typical glyphs. |
| 3144 |
| 3145 With this and recent changes to avoid mallocs, the thread-safe |
| 3146 stack-based loader is now as fast as the previous model that had one |
| 3147 cached singleton. |
| 3148 |
| 3149 * src/autofit/afhints.h (AF_SEGMENTS_EMBEDDED, AF_EDGES_EMBEDDED): |
| 3150 New macros. |
| 3151 (AF_AxisHintsRec): Add two arrays for segments and edges. |
| 3152 |
| 3153 * src/autofit/afhints.c (af_axis_hints_new_segment): Only allocate |
| 3154 data if number of segments exceeds given threshold value. |
| 3155 (af_axis_hints_new_edge): Only allocate data if number of edges |
| 3156 exceeds given threshold value. |
| 3157 (af_glyph_hints_done): Updated. |
| 3158 |
| 3159 2015-01-14 Behdad Esfahbod <behdad@behdad.org> |
| 3160 |
| 3161 [autofit] Add embedded arrays for points and contours. |
| 3162 |
| 3163 This avoids at least two malloc calls for typical glyphs. |
| 3164 |
| 3165 * src/autofit/afhints.h (AF_POINTS_EMBEDDED, AF_CONTOURS_EMBEDDED): |
| 3166 New macros. |
| 3167 (AF_GlyphHintsRec): Add two arrays for contours and points. |
| 3168 |
| 3169 * src/autofit/afhints.c (af_glyph_hints_init, af_glyph_hints_done): |
| 3170 Updated. |
| 3171 (af_glyph_hints_reload): Only allocate data if number of contours or |
| 3172 points exceeds given threshold values. |
| 3173 |
| 3174 2015-01-14 Behdad Esfahbod <behdad@behdad.org> |
| 3175 |
| 3176 [autofit] Allocate hints object on the stack. |
| 3177 |
| 3178 This avoids one malloc per load. |
| 3179 |
| 3180 * src/autofit/afloader.h (AF_LoaderRec): Change type of `hints' to |
| 3181 `AF_GlyphHints'. |
| 3182 Update prototype. |
| 3183 |
| 3184 * src/autofit/afloader.c (af_loader_init): Use `AF_GlyphHints' |
| 3185 parameter instead of `FT_Memory'. |
| 3186 (af_loader_done): Directly reset `load_hints'. |
| 3187 (af_loader_load_g): Updated. |
| 3188 |
| 3189 * src/autofit/afmodule.c (af_autofitter_load_glyph): Use local |
| 3190 `hints' object. |
| 3191 |
| 3192 2015-01-14 Behdad Esfahbod <behdad@behdad.org> |
| 3193 |
| 3194 [autofit] Reuse slot glyph loader. |
| 3195 |
| 3196 No need to create a new glyph loader; we can reuse the one from |
| 3197 `slot->internal->loader'. It's hard to tell why it was written that |
| 3198 way originally, but new code looks sound and correct to me, and |
| 3199 avoids lots of allocations. |
| 3200 |
| 3201 * src/autofit/afloader.c (af_loader_init): Change return type to |
| 3202 `void'. |
| 3203 Don't call `FT_GlyphLoader_New'. |
| 3204 (af_loader_reset): Don't call `FT_GlyphLoader_Rewind'. |
| 3205 (af_loader_load_g): Update code to use `internal->loader', which |
| 3206 doesn't need copying of data. |
| 3207 |
| 3208 * src/autofit/afloader.h (AF_LoaderRec): Remove `gloader' member. |
| 3209 Update prototype. |
| 3210 |
| 3211 * src/autofit/afmodule.c (af_autofitter_load_glyph): Updated. |
| 3212 |
| 3213 2015-01-14 Behdad Esfahbod <behdad@behdad.org> |
| 3214 |
| 3215 [autofit] Remove (unused) support for composite glyphs. |
| 3216 |
| 3217 We never have to deal with composite glyphs in the autohinter, as |
| 3218 those will be loaded into FORMAT_OUTLINE by the recursed |
| 3219 `FT_Load_Glyph' function. |
| 3220 |
| 3221 In the rare cases that FT_LOAD_NO_RECURSE is set, it will imply |
| 3222 FT_LOAD_NO_SCALE as per `FT_Load_Glyph', which then implies |
| 3223 FT_LOAD_NO_HINTING: |
| 3224 |
| 3225 /* resolve load flags dependencies */ |
| 3226 |
| 3227 if ( load_flags & FT_LOAD_NO_RECURSE ) |
| 3228 load_flags |= FT_LOAD_NO_SCALE | |
| 3229 FT_LOAD_IGNORE_TRANSFORM; |
| 3230 |
| 3231 if ( load_flags & FT_LOAD_NO_SCALE ) |
| 3232 { |
| 3233 load_flags |= FT_LOAD_NO_HINTING | |
| 3234 FT_LOAD_NO_BITMAP; |
| 3235 |
| 3236 load_flags &= ~FT_LOAD_RENDER; |
| 3237 } |
| 3238 |
| 3239 and as such the auto-hinter is never called. Thus, the recursion in |
| 3240 `af_loader_load_g' never actually happens. So remove the depth |
| 3241 counter as well. |
| 3242 |
| 3243 * src/autofit/afloader.c (af_loader_load_g): Remove `depth' |
| 3244 parameter. |
| 3245 <FT_GLYPH_FORMAT_COMPOSITE>: Remove associated code. |
| 3246 (af_loader_load_glyph): Updated. |
| 3247 |
| 3248 2015-01-14 Behdad Esfahbod <behdad@behdad.org> |
| 3249 |
| 3250 [raster] Fix uninitialized memory access. |
| 3251 |
| 3252 Apparently `ras.cProfile' might be uninitialized. This will be the |
| 3253 case if `ras.top == ras.cProfile->offset', as can be seen in |
| 3254 `End_Profile'. The overshoot code introduced in a change `Fix B/W |
| 3255 rasterization of subglyphs with different drop-out modes.' (from |
| 3256 2009-06-18) violated this, accessing `ras.cProfile->flags' |
| 3257 unconditionally just before calling `End_Profile' (which then |
| 3258 detected that `cProfile' is uninitialized and didn't touch it). |
| 3259 |
| 3260 This was harmless, and was not detected by valgrind before because |
| 3261 the objects were allocated on the `raster_pool', which was always |
| 3262 initialized. With recent change to allocate raster buffers on the |
| 3263 stack, valgrind now reported this invalid access. |
| 3264 |
| 3265 * src/raster/ftraster.c (Convert_Glyph): Don't access an |
| 3266 uninitialized `cProfile'. |
| 3267 |
| 3268 2015-01-14 Behdad Esfahbod <behdad@behdad.org> |
| 3269 |
| 3270 [smooth] Fix uninitialized memory access. |
| 3271 |
| 3272 Looks like `ras.span_y' could always be used without initialization. |
| 3273 This was never detected by valgrind before because the library-wide |
| 3274 `raster_pool' was used for the worker object and `raster_pool' was |
| 3275 originally zero'ed. But subsequent reuses of it were using `span_y' |
| 3276 uninitialized. With the recent change to not use `render_pool' and |
| 3277 allocate worker and buffer on the stack, valgrind now detects this |
| 3278 uninitialized access. |
| 3279 |
| 3280 * src/smooth/ftgrays.c (gray_raster_render): Initialize |
| 3281 `ras.span_y'. |
| 3282 |
| 3283 2015-01-14 Behdad Esfahbod <behdad@behdad.org> |
| 3284 |
| 3285 [base] Don't initialize unused `driver->glyph_loader'. |
| 3286 |
| 3287 * src/base/ftobjs.c (Destroy_Driver): Don't call |
| 3288 `FT_GlyphLoader_Done'. |
| 3289 (FT_Add_Module): Don't call `FT_GlyphLoader_New'. |
| 3290 |
| 3291 2015-01-14 Behdad Esfahbod <behdad@behdad.org> |
| 3292 |
| 3293 [base] Don't allocate `library->raster_pool' anymore. |
| 3294 |
| 3295 It's unused after the following commits: |
| 3296 |
| 3297 [raster] Allocate render pool for mono rasterizer on the stack. |
| 3298 [raster] Remove 5-level gray AA mode from monochrome rasterizer. |
| 3299 |
| 3300 The value of FT_RENDER_POOL_SIZE still serves the purpose it used to |
| 3301 serve, which is, to adjust the pool size. But the pool is now |
| 3302 allocated on the stack on demand. |
| 3303 |
| 3304 * src/base/ftobjs.c (FT_New_Library, FT_Done_Library): Implement. |
| 3305 |
| 3306 2015-01-14 Behdad Esfahbod <behdad@behdad.org> |
| 3307 |
| 3308 [base] Do not reorder library->renderers upon use. |
| 3309 |
| 3310 Instead of keeping `library->renderers' in a MRU order, just leave |
| 3311 it as-is. The MRU machinery wasn't thread-safe. |
| 3312 |
| 3313 With this patch, rasterizing glyphs from different faces from |
| 3314 different threads doesn't fail choosing rasterizer |
| 3315 (FT_Err_Cannot_Render_Glyph). |
| 3316 |
| 3317 Easiest to see that crash was to add a `printf' (or otherwise let |
| 3318 thread yield in FT_Throw with debugging enabled). |
| 3319 |
| 3320 * src/base/ftobjs.c (FT_Render_Glyph_Internal), src/base/ftoutln.c |
| 3321 (FT_Outline_Render): Don't call `FT_Set_Renderer'. |
| 3322 |
| 3323 2015-01-14 Behdad Esfahbod <behdad@behdad.org> |
| 3324 |
| 3325 [raster] Allocate render pool for mono rasterizer on the stack. |
| 3326 |
| 3327 Instead of using the `render_pool' member of `FT_Library' that is |
| 3328 provided down to the rasterizer, completely ignore that and allocate |
| 3329 needed objects on the stack instead. |
| 3330 |
| 3331 With this patch, rasterizing glyphs from different faces from |
| 3332 different threads doesn't crash in the monochrome rasterizer. |
| 3333 |
| 3334 * src/raster/ftraster.c (black_TRaster): Remove `buffer', |
| 3335 `buffer_size', and `worker' members. |
| 3336 |
| 3337 (ft_black_render): Create `buffer' locally. |
| 3338 (ft_black_reset): Updated. |
| 3339 |
| 3340 2015-01-14 Behdad Esfahbod <behdad@behdad.org> |
| 3341 |
| 3342 [raster] Remove 5-level gray AA mode from monochrome rasterizer. |
| 3343 |
| 3344 It was off by default and couldn't be turned on at runtime. And the |
| 3345 smooth rasterizer superceded it over ten years ago. No point in |
| 3346 keeping. Comments suggested that it was there for compatibility |
| 3347 with FreeType 1. |
| 3348 |
| 3349 550 lines down. |
| 3350 |
| 3351 * src/raster/ftraster.c (FT_RASTER_OPTION_ANTI_ALIASING, |
| 3352 RASTER_GRAY_LINES): Remove macros and all associated code. |
| 3353 |
| 3354 (black_TWorker): Remove `gray_min_x' and `gray_max_x'. |
| 3355 (black_TRaster): Remove `grays' and `gray_width'. |
| 3356 |
| 3357 (Vertical_Sweep_Init, Vertical_Sweep_Span, Vertical_Sweep_Drop, |
| 3358 ft_black_render): Updated. |
| 3359 |
| 3360 * src/raster/ftrend1.c (ft_raster1_render): Simplify code. |
| 3361 (ft_raster5_renderer_class): Removed. |
| 3362 |
| 3363 2015-01-14 Behdad Esfahbod <behdad@behdad.org> |
| 3364 |
| 3365 [smooth] Allocate render pool for smooth rasterizer on the stack. |
| 3366 |
| 3367 Instead of using the `render_pool' member of `FT_Library' that is |
| 3368 provided down to the rasterizer, completely ignore that and allocate |
| 3369 needed objects on the stack instead. |
| 3370 |
| 3371 With this patch, rasterizing glyphs from different faces from |
| 3372 different threads doesn't crash in the smooth rasterizer. |
| 3373 |
| 3374 Bugs: |
| 3375 |
| 3376 https://bugzilla.redhat.com/show_bug.cgi?id=678397 |
| 3377 https://bugzilla.redhat.com/show_bug.cgi?id=1004315 |
| 3378 https://bugzilla.redhat.com/show_bug.cgi?id=1165471 |
| 3379 https://bugs.freedesktop.org/show_bug.cgi?id=69034 |
| 3380 |
| 3381 * src/smooth/ftgrays.c (gray_TRaster): Remove `buffer', |
| 3382 `buffer_size', `band_size', and `worker' members. |
| 3383 |
| 3384 (gray_raster_render): Create `buffer', `buffer_size', and |
| 3385 `band_size' locally. |
| 3386 (gray_raster_reset): Updated. |
| 3387 |
| 3388 2015-01-14 Behdad Esfahbod <behdad@behdad.org> |
| 3389 |
| 3390 [truetype] Allocate TT_ExecContext in TT_Size instead of TT_Driver. |
| 3391 |
| 3392 Previously the code had stipulation for using a per-TT_Size exec |
| 3393 context if `size->debug' was true. But there was no way that |
| 3394 `size->debug' could *ever* be true. As such, the code was always |
| 3395 using the singleton `TT_ExecContext' that was stored in `TT_Driver'. |
| 3396 This was, clearly, not threadsafe. |
| 3397 |
| 3398 With this patch, loading glyphs from different faces from different |
| 3399 threads doesn't crash in the bytecode loader code. |
| 3400 |
| 3401 * src/truetype/ttobjs.h (TT_SizeRec): Remove `debug' member. |
| 3402 (TT_DriverRec): Remove `context' member. |
| 3403 |
| 3404 * src/truetype/ttobjs.c (tt_size_run_fpgm, tt_size_run_prep): Remove |
| 3405 `TT_ExecContext' code related to a global `TT_Driver' object. |
| 3406 |
| 3407 (tt_driver_done): Don't remove `TT_ExecContext' object here but ... |
| 3408 (tt_size_done_bytecode): ... here. |
| 3409 |
| 3410 (tt_driver_init): Don't create `TT_ExecContext' object here but ... |
| 3411 (tt_size_init_bytecode): ... here, only on demand. |
| 3412 |
| 3413 * src/truetype/ttinterp.c (TT_Run_Context): Remove defunct debug |
| 3414 code. |
| 3415 (TT_New_Context): Remove `TT_ExecContext' code related to a global |
| 3416 `TT_Driver' object. |
| 3417 |
| 3418 * src/truetype/ttinterp.h: Updated. |
| 3419 |
| 3420 * src/truetype/ttgload.c (TT_Hint_Glyph, tt_loader_init): Updated. |
| 3421 |
| 3422 2015-01-14 Behdad Esfahbod <behdad@behdad.org> |
| 3423 |
| 3424 [autofit] Allocate AF_Loader on the stack instead of AF_Module. |
| 3425 |
| 3426 Stop sharing a global `AF_Loader'. Allocate one on the stack during |
| 3427 glyph load. |
| 3428 |
| 3429 Right now this results in about 25% slowdown, to be fixed in a |
| 3430 following commit. |
| 3431 |
| 3432 With this patch loading glyphs from different faces from different |
| 3433 threads doesn't immediately crash in the autohinting loader code. |
| 3434 |
| 3435 Bugs: |
| 3436 |
| 3437 https://bugzilla.redhat.com/show_bug.cgi?id=1164941 |
| 3438 |
| 3439 * src/autofit/afloader.c (af_loader_init): Pass |
| 3440 `AF_Loader' and `FT_Memory' instead of `AF_Module' as arguments. |
| 3441 (af_loader_reset, af_loader_load_glyph): Also pass `loader' as |
| 3442 argument. |
| 3443 (af_loader_done): Use `AF_Loader' instead of `AF_Module' as |
| 3444 argument. |
| 3445 |
| 3446 * src/autofit/afmodule.c (af_autofitter_init): Don't call |
| 3447 `af_loader_init'. |
| 3448 (af_autofitter_done): Don't call `af_loader_done'. |
| 3449 (af_autofitter_load_glyph): Use a local `AF_Loader' object. |
| 3450 |
| 3451 * src/autofit/afloader.h: Include `afmodule.h'. |
| 3452 Update prototypes. |
| 3453 Move typedef for `AF_Module' to... |
| 3454 |
| 3455 * src/autofit/afmodule.h: ... this place. |
| 3456 No longer include `afloader.h'. |
| 3457 |
| 3458 2015-01-14 Behdad Esfahbod <behdad@behdad.org> |
| 3459 |
| 3460 * src/type42/t42objs.h (T42_DriverRec): Remove unused member. |
| 3461 |
| 3462 2015-01-12 Werner Lemberg <wl@gnu.org> |
| 3463 |
| 3464 Fix Savannah bug #43976. |
| 3465 |
| 3466 Assure that FreeType's internal include directories are found before |
| 3467 `CPPFLAGS' (which might be set by the user in the environment), and |
| 3468 `CPPFLAGS' before `CFLAGS'. |
| 3469 |
| 3470 * builds/freetype.mk (FT_CFLAGS): Don't add `INCLUDE_FLAGS'. |
| 3471 (FT_COMPILE): Make this a special variable for compiling only the |
| 3472 files handled in `freetype.mk'. |
| 3473 (.c.$O): Removed, unused. |
| 3474 |
| 3475 * src/*/rules.mk (*_COMPILE): Fix order of include directories. |
| 3476 |
| 3477 2015-01-11 Werner Lemberg <wl@gnu.org> |
| 3478 |
| 3479 [truetype] Prettyfing. |
| 3480 |
| 3481 * src/truetype/ttinterp.c (project, dualproj, fast_project, |
| 3482 fast_dualproj): Rename to... |
| 3483 (PROJECT, DUALPROJ, FAST_PROJECT, FAST_DUALPROJ): ... this. |
| 3484 |
| 3485 2015-01-11 Werner Lemberg <wl@gnu.org> |
| 3486 |
| 3487 * src/truetype/ttinterp.c (Ins_JROT, Ins_JROF): Simplify. |
| 3488 |
| 3489 Based on a patch from Behdad. |
| 3490 |
| 3491 2015-01-11 Werner Lemberg <wl@gnu.org> |
| 3492 |
| 3493 * src/truetype/ttinterp.c (Ins_SxVTL): Simplify function call. |
| 3494 |
| 3495 2015-01-11 Werner Lemberg <wl@gnu.org> |
| 3496 |
| 3497 * src/truetype/ttinterp.c (Normalize): Remove unused argument. |
| 3498 |
| 3499 2015-01-11 Werner Lemberg <wl@gnu.org> |
| 3500 |
| 3501 [truetype] More macro expansions. |
| 3502 |
| 3503 * src/truetype/ttinterp.c (FT_UNUSED_EXEC): Remove macro by |
| 3504 expansion. |
| 3505 |
| 3506 2015-01-11 Werner Lemberg <wl@gnu.org> |
| 3507 |
| 3508 [truetype] More macro expansions. |
| 3509 |
| 3510 * src/truetype/ttinterp.c (INS_ARG): Remove macro by expansion, |
| 3511 adjusting funtion calls where necessary. |
| 3512 (FT_UNUSED_ARG): Removed, no longer needed. |
| 3513 |
| 3514 2015-01-10 Werner Lemberg <wl@gnu.org> |
| 3515 |
| 3516 [truetype] More macro expansions. |
| 3517 |
| 3518 Based on a patch from Behdad. |
| 3519 |
| 3520 * src/truetype/ttinterp.c (DO_*): Expand macros into corresponding |
| 3521 `Ins_*' functions. |
| 3522 (TT_RunIns): Replace `DO_*' macros with `Ins_*' function calls. |
| 3523 (ARRAY_BOUND_ERROR): Remove second definition, which is no longer |
| 3524 needed. |
| 3525 (Ins_SVTCA, Ins_SPVTCA, Ins_SFVTCA): Replaced with... |
| 3526 (Ins_SxyTCA): New function. |
| 3527 |
| 3528 2015-01-10 Werner Lemberg <wl@gnu.org> |
| 3529 |
| 3530 [truetype] Remove TT_CONFIG_OPTION_INTERPRETER_SWITCH. |
| 3531 |
| 3532 Behdad suggested this code simplification, and nobody objected... |
| 3533 |
| 3534 * include/config/ftoption.h, devel/ftoption.h |
| 3535 (TT_CONFIG_OPTION_INTERPRETER_SWITCH): Remove. |
| 3536 |
| 3537 * src/truetype/ttinterp.c [TT_CONFIG_OPTION_INTERPRETER_SWITCH]: |
| 3538 Remove related code. |
| 3539 (ARRAY_BOUND_ERROR): Use do-while loop. |
| 3540 |
| 3541 2015-01-10 Werner Lemberg <wl@gnu.org> |
| 3542 |
| 3543 [truetype] More macro expansions. |
| 3544 |
| 3545 * src/truetype/ttinterp.c, src/truetype/ttinterp.h (EXEC_ARG_, |
| 3546 EXEC_ARG): Remove by replacing with expansion. |
| 3547 |
| 3548 2015-01-10 Werner Lemberg <wl@gnu.org> |
| 3549 |
| 3550 [truetype] More macro expansions. |
| 3551 |
| 3552 Based on a patch from Behdad. |
| 3553 |
| 3554 * src/truetype/ttinterp.c (SKIP_Code, GET_ShortIns, NORMalize, |
| 3555 SET_SuperRound, ROUND_None, INS_Goto_CodeRange, CUR_Func_move, |
| 3556 CUR_Func_move_orig, CUR_Func_round, CUR_Func_cur_ppem, |
| 3557 CUR_Func_read_cvt, CUR_Func_write_cvt, CUR_Func_move_cvt, |
| 3558 CURRENT_Ratio, INS_SxVTL, COMPUTE_Funcs, COMPUTE_Round, |
| 3559 COMPUTE_Point_Displacement, MOVE_Zp2_Point): Remove by replacing |
| 3560 with expansion. |
| 3561 |
| 3562 (Cur_Func_project, CUR_Func_dualproj, CUR_fast_project, |
| 3563 CUR_fast_dualproj): Replace with macros `project', `dualproj', |
| 3564 `fast_project', `fast_dualproj'. |
| 3565 |
| 3566 2015-01-10 Werner Lemberg <wl@gnu.org> |
| 3567 |
| 3568 [truetype] More macro expansions. |
| 3569 |
| 3570 * src/truetype/ttinterp.c (EXEC_OP_, EXEC_OP): Remove by replacing |
| 3571 with expansion. |
| 3572 |
| 3573 2015-01-10 Werner Lemberg <wl@gnu.org> |
| 3574 |
| 3575 [truetype] Remove code for static TrueType interpreter. |
| 3576 |
| 3577 This is a follow-up patch. |
| 3578 |
| 3579 * src/truetype/ttinterp.c, src/truetype/ttinterp.h |
| 3580 [TT_CONFIG_OPTION_STATIC_INTERPRETER, |
| 3581 TT_CONFIG_OPTION_STATIC_RASTER]: Remove macros and related code. |
| 3582 |
| 3583 2015-01-10 Werner Lemberg <wl@gnu.org> |
| 3584 |
| 3585 * src/truetype/ttinterp.c (CUR): Remove by replacing with expansion. |
| 3586 |
| 3587 This starts a series of patches that simplifies the code of the |
| 3588 bytecode interpreter. |
| 3589 |
| 3590 |
| 3591 ---------------------------------------------------------------------------- |
| 3592 |
| 3593 Copyright 2015 by |
| 3594 David Turner, Robert Wilhelm, and Werner Lemberg. |
| 3595 |
| 3596 This file is part of the FreeType project, and may only be used, modified, |
| 3597 and distributed under the terms of the FreeType project license, |
| 3598 LICENSE.TXT. By continuing to use, modify, or distribute this file you |
| 3599 indicate that you have read the license and understand and accept it |
| 3600 fully. |
| 3601 |
| 3602 |
| 3603 Local Variables: |
| 3604 version-control: never |
| 3605 coding: utf-8 |
| 3606 End: |
| OLD | NEW |