summaryrefslogtreecommitdiff
AgeCommit message (Collapse)AuthorFilesLines
2008-01-15Relicense HarfBuzzBehdad Esfahbod28-685/+506
Thanks to Werner Lemberg and David Turner of FreeType project for relicensing ftlayout, and Red Hat, Inc. and Trolltech ASA, HarfBuzz is now available under a much simpler (and more generous) license.
2008-01-04Fix a compiler warning with MSVC (C4138: '*/' found outside of comment)Simon Hausmann1-1/+1
2007-12-06Fix harfbuzz compilation on xlCThiago Macieira1-12/+12
Compiler error is: Bit-field lineBreakType must be of type signed int, unsigned int or int. Signed-off-by: Simon Hausmann <shausman@trolltech.com>
2007-11-18Remove unused macro.Behdad Esfahbod1-18/+0
2007-11-07Minor changes.Behdad Esfahbod6-108/+19
2007-11-07Port changes from PangoBehdad Esfahbod29-909/+995
2007-11-07improve error handlingBehdad Esfahbod11-203/+182
Call the now-noop macro ERR() at all error sites. The only exceptions are HB_Err_Ok (always) and HB_Err_Not_Covered (in most places). Also remove many redundant error types: - Replaced by HB_Err_Invalid_SubTable_Format and HB_Err_Invalid_SubTable: HB_Err_Invalid_GSUB_SubTable_Format HB_Err_Invalid_GSUB_SubTable HB_Err_Invalid_GPOS_SubTable_Format HB_Err_Invalid_GPOS_SubTable HB_Err_Invalid_GDEF_SubTable_Format HB_Err_Invalid_GDEF_SubTable It's always clear, from the function called, which one was intended. - Replaced by HB_Err_Not_Covered, but calling ERR() too: HB_Err_Too_Many_Nested_Contexts HB_Err_No_MM_Interpreter This makes the fonts work better. - Replaced by HB_Err_Not_Covered: HB_Err_Empty_Script HB_Err_Table_Missing And update the few places they are checked for. - Replaced by HB_Err_Invalid_Argument: HB_Err_Invalid_Face_Handle It's simply yet another invalid argument case. - Renamed to HB_Err_Read_Error: HB_Err_Invalid_Stream_Operation
2007-10-30Fix compilation for compilers with broken for scoping.Simon Hausmann2-5/+8
2007-10-27Fix compilation on SolarisMarius Bugge Monsen1-1/+1
Signed-off-by: Simon Hausmann <shausman@trolltech.com>
2007-10-23added API for finding word and sentence boundaries.Lars Knoll3-4/+180
Signed-off-by: Simon Hausmann <shausman@trolltech.com>
2007-10-23correctly set grapheme boundaries for hangul.Lars Knoll1-1/+1
Signed-off-by: Simon Hausmann <shausman@trolltech.com>
2007-10-23modification to harfbuzz to get proper support for grapheme boundaries in there.Lars Knoll2-7/+42
Signed-off-by: Simon Hausmann <shausman@trolltech.com>
2007-09-17Merge branch 'master' of git+ssh://git.freedesktop.org/git/harfbuzzSimon Hausmann1-1/+1
2007-09-17Enable the use of the calt property for Tibetan, as advised by Christopher ↵Simon Hausmann1-0/+1
Fynn <cfynn@gmx.net>
2007-09-02Merge branch 'stable'Simon Hausmann1-1/+1
2007-09-02Fix placement of a diacritic marks for an arabic ligature is not correctBehdad Esfahbod  1-1/+1
This one fixing issue that first ligature in a line was having wrong mark attachments.  I was debugging a bigger problem with mark attachments, that's ongoing. http://bugzilla.gnome.org/show_bug.cgi?id=302952 (fixed) http://bugzilla.gnome.org/show_bug.cgi?id=437633 (not fixed yet) 2007-08-29  Behdad Esfahbod  <behdad@gnome.org>         Bug 302952 – The placement of a diacritic marks for an arabic ligature         is not correct         * pango/opentype/harfbuzz-buffer.c (hb_buffer_allocate_ligid): Don't         use zero as allocated ligature id.  Zero means no ligature id.
2007-08-26Merge branch 'stable'Simon Hausmann2-6/+11
Conflicts: src/harfbuzz-gpos.c src/harfbuzz-gsub.c
2007-08-26[HarfBuzz] Another fixBehdad Esfahbod2-6/+11
The actual fix for this bug now:         http://bugzilla.gnome.org/show_bug.cgi?id=463430 The bug was that an intermediate index variable was of type FT_UShort and overflowing with more than 65536 input glyphs.  Yikes!  Patch attached.
2007-08-21Merge branch 'master' of git+ssh://git.freedesktop.org/git/harfbuzzSimon Hausmann1-4/+4
2007-08-21Merge branch 'stable'Simon Hausmann1-1/+1
2007-08-21While trying to see if the patch in this bug:Behdad Esfahbod1-1/+1
  http://bugzilla.gnome.org/show_bug.cgi?id=463430 is correct (it's not), I found a small bug in harfbuzz-gsub.
2007-08-01Use unsigned int instead of uint to compile on OS X.Simon Hausmann1-4/+4
2007-08-01Get rid of the unused faceData field in HB_Font.Simon Hausmann2-6/+4
2007-07-31Don't crash if HB_FreeFace is called with a null pointer.Simon Hausmann1-0/+2
2007-07-31Moved all HB_Err macros into the HB_Error enum.Simon Hausmann5-21/+13
2007-07-05Ignore some more generated filesSimon Hausmann1-0/+2
2007-07-05Behdad says: Don't use --enable-maintainer mode. Turns out it's indeed broken.Simon Hausmann2-3/+2
2007-04-07Fix justified arabic text (automatic kashida insertion). We forgot to copy theSimon Hausmann1-0/+2
justification class from the output of getArabicProperties into the glyph attributes array.
2007-04-04Apply the same fix as in the Qt 4.3 branch: The return values from the OpenTypeSimon Hausmann2-3/+9
layer for the advances have a fraction. Unless design metrics are requested these need to be rounded in order to give consistent positioning (for example kerning). For example "TeTeTeTeTe" otherwise has different/inconsistent kerning for "Te" that is especially for small sizes very visible. Discussed with Lars.
2007-04-03[harfbuzz-open] Fix GDEF type-2 class lookup array overflowBehdad Esfahbod1-1/+1
Reviewed by Simon Hausmann
2007-04-03[harfbuzz-open] Fix GDEF type-2 class lookup array overflowBehdad Esfahbod1-1/+1
Reviewed by Simon Hausmann
2007-04-03Header cleanup, don't include harfbuzz.h from harfbuzz-shaper.h, asSimon Hausmann1-1/+5
required by Andreas
2007-04-03Merge branch 'stable'Simon Hausmann1-0/+6
Conflicts: src/harfbuzz-gpos.c
2007-04-03Fixed gpos pair adjustments sometimes not being applied if the first glyphSimon Hausmann1-0/+6
was covered but not the second. In that case we did not try looking up the second glyph in the next iteration. This mean for example that while "Te" was nicely kerned "QTe" wasn't because after we failed to look up "T" for the "QT" pair we skipped looking up "T" for "TE" in the next iteration. Reviewed by Denis Jacquerye
2007-04-02Fix uninitialized variableSimon Hausmann1-0/+2
2007-04-02removed old cvsignore filesSimon Hausmann1-7/+0
2007-04-02removed generated object file :)Simon Hausmann1-0/+0
2007-04-02close a mem leak.lars1-0/+2
2007-03-29Fix stupid logic error in HB_GetCharAttributesSimon Hausmann1-2/+2
2007-03-28Adapt to the latest API changesSimon Hausmann1-2/+2
2007-03-28API cleanup, rename getAscent to getFontMetricSimon Hausmann2-2/+6
2007-03-28Some naming cleanupsSimon Hausmann10-47/+50
2007-03-28Use short instead of int for the bitfields in HB_GlyphAttributes. That makes ↵Simon Hausmann1-6/+6
it work with MSVC
2007-03-28Apply Lars' patch from perforce change 253707 in Qt/4.3:Simon Hausmann1-17/+17
Fixes: some indic shaping issues Task: reported by several tasks Details: Add indic character classes for characters that have been added to the unicode standard since the last time I went through this table.
2007-03-28Use standard forward declarations for HB_FaceSimon Hausmann2-2/+3
2007-03-28Added a comment about HB_GlyphAttributes and SCRIPT_VISATTRSimon Hausmann1-0/+3
2007-03-28Fixed missing uint32Simon Hausmann2-3/+3
2007-03-28Don't use stdint.h, Msvc doesn't provide itSimon Hausmann13-97/+103
2007-03-28Don't try to re-use the initially provided glyphs a second timeSimon Hausmann1-1/+5
2007-03-28Make it possible for the calling application to provide the glyph indices inSimon Hausmann4-5/+18
advance
generated by cgit v1.2.3 (git 2.39.1) at 2024-11-25 03:50:44 +0000