Commit Graph

397 Commits

Author SHA1 Message Date
Pierre Schweitzer 06f65740b2 usp10: Fix memory leak. 2011-12-09 21:15:25 +01:00
Aric Stewart 7b95ac8238 usp10: Some Indic scripts, Sinhala, want GDEF to set glyph properties. 2011-12-08 18:04:30 +01:00
Aric Stewart 746679e186 usp10: Handle dangling joiners when processing Indic GlyphProps. 2011-12-07 17:57:42 +01:00
Pierre Schweitzer 783683be2d usp10: Fix memory leaks. 2011-12-07 17:49:58 +01:00
Aric Stewart c481cdd25e usp10: Unicode does not compose the Gurmukhi letter SHA. 2011-12-07 17:49:57 +01:00
Aric Stewart acf1343f34 usp10: Handle Halant, ZWJ and ZWNJ in ShapeCharGlyphProp_BaseIndic. 2011-12-06 17:04:21 +01:00
Aric Stewart 93a8499b95 usp10: Sinhala, while behaving like a base Indic, does not set GlyphProps based on indic syllable. 2011-12-06 17:04:15 +01:00
Aric Stewart a9058605d6 usp10: Use UpdateClusters in DecomposeVowels. 2011-12-06 17:04:09 +01:00
Aric Stewart 4dffb0aa43 usp10: The top logclust value may be different than the number of chars. 2011-12-06 17:04:04 +01:00
Aric Stewart 894f2dc9ba usp10: Use GDEF in baseIndic GlyphProp processing. 2011-12-06 17:03:59 +01:00
Aric Stewart 9cddd18cb9 usp10: Initialize the GSUB table before doing any contextual shaping. 2011-12-06 17:03:51 +01:00
Aric Stewart 6673694847 usp10: Correct issues with LTR logclust indexing after Multiple substitution. 2011-12-05 17:39:10 +01:00
Aric Stewart 225de02a29 usp10: Handle Thai SARA AM differently than GDEF. 2011-12-02 17:25:52 +01:00
Aric Stewart 183fdedb90 usp10: More properly handle the glyph properties on glyphs added by multiple substitutions. 2011-12-02 15:59:59 +01:00
Aric Stewart ceca716999 usp10: Phags-Pa wants the calt feature enabled. 2011-12-02 15:59:54 +01:00
Aric Stewart ff9ae10ba5 usp10: Handle the Gurmukhi addak. 2011-12-01 18:16:20 +01:00
Aric Stewart 302409ebe4 usp10: Always apply features in logical order. 2011-12-01 18:16:11 +01:00
Aric Stewart c4c9f8b041 usp10: Apply lookups in order over the whole range of glyphs instead of individually to each glyph. 2011-12-01 18:16:03 +01:00
Aric Stewart 711dd9b13a usp10: Standard features are only ccmp and locl. 2011-12-01 18:15:06 +01:00
Aric Stewart 5ed8160ec7 usp10: Turn off dlig feature for hebrew. It is not used by default. 2011-12-01 18:14:59 +01:00
Francois Gouget 51a256e65f Assorted spelling fixes and wording tweaks. 2011-11-28 16:40:50 +01:00
Aric Stewart 25d43d7a09 usp10: Cache the script with the feature so that in caches with multiple scripts we do not fetch the wrong feature from the cache. 2011-11-21 11:02:53 +01:00
Aric Stewart 9097226a3c usp10: Itemize numbers outside of the basic ascii set in a seperate script. 2011-11-21 11:02:39 +01:00
Aric Stewart 2fe60ff696 usp10: Windows breaks punctuation into 2 different script types. 2011-11-21 11:02:34 +01:00
Aric Stewart 714408684d usp10: Handle Combining Diacritical Marks. 2011-11-21 11:02:28 +01:00
Aric Stewart 2abb8193b9 usp10: Add a facility to mark itemization test elements as broken. 2011-11-21 11:02:22 +01:00
Aric Stewart fce22eadc1 usp10: Some unicode code points will force us into bidi mode. 2011-11-21 11:02:10 +01:00
Aric Stewart bba75dd0a7 usp10: Improve handling for ZWJ and ZWNJ in itemization. 2011-11-21 11:01:58 +01:00
Aric Stewart ed15053837 usp10: In Arabic only fallback to presentation form B for isolated glyphs. 2011-11-21 11:01:50 +01:00
Aric Stewart 105704534d usp10: When converting common indic items use the main indic script not the Numeric or currency script, even if that was the most recent. 2011-11-17 18:10:19 +01:00
Aric Stewart 41db79a855 usp10: If leading with a strong character use its bidi level for our intial LayoutRTL. 2011-11-17 18:10:11 +01:00
Aric Stewart 2c8fb3aba1 usp10: Keep the levels array in more cases. 2011-11-17 18:09:56 +01:00
Aric Stewart 92f706cd74 usp10: Windows 7 defines a Tibetan numeric script but never uses it. 2011-11-17 18:09:52 +01:00
Aric Stewart 7868be6e70 usp10: Properly set LayoutRTL. 2011-11-15 19:37:28 +01:00
Aric Stewart e5d33f08c7 usp10: Devanagari danda and double danda are used for all Indic scripts. 2011-11-15 19:37:15 +01:00
Aric Stewart 6c3398d957 usp10: Pay attention to script changes for bidi run itemization. 2011-11-15 16:09:10 +01:00
Aric Stewart 145102c7fd usp10: Merge neutral scripts. 2011-11-15 16:09:10 +01:00
Aric Stewart 3767cebc88 usp10: Move scripts to an array initialized once to minimize get_char_script calls. 2011-11-15 16:09:09 +01:00
Aric Stewart 9c9513f94a usp10: Break with a better understanding on the handling of bidi strengths. 2011-11-15 16:09:09 +01:00
Aric Stewart e6ee421e88 usp10: Add some additional Latin script ranges. 2011-11-14 15:25:24 +01:00
Aric Stewart 65710af292 usp10: Better understand and handle numbers level in RTL runs. 2011-11-09 11:34:49 +01:00
Aric Stewart 2781ac1e6d usp10: Better handle the absorption of whitespace in bidi runs. 2011-11-09 11:34:45 +01:00
Aric Stewart c5db60200f usp10/tests: Add Kannada shaping test. 2011-11-08 18:19:18 +01:00
Aric Stewart cd3d305b60 usp10: Add a few chars that Windows itemizes as punctuation. 2011-11-08 18:19:18 +01:00
Aric Stewart 38aa7696c5 usp10: Avoid an end of buffer overrun when seeking non space characters at an end of a BiDi string. 2011-11-08 18:19:17 +01:00
Aric Stewart 89a53d6f81 usp10: Sinhala does glyph props like Indic languages. 2011-10-28 12:46:19 +02:00
Aric Stewart ec23fe6543 usp10: All glyphs in a single indic syllable are part of the same cluster. 2011-10-28 12:46:14 +02:00
Aric Stewart 729de083f3 usp10: Correct itemizing multiple spaces at the beginning of a run. 2011-10-24 11:14:06 +02:00
Aric Stewart f0c85ff83d usp10: Correct some reading buffer overruns when parsing indic syllables.
Found with help from Dan Kegel and Valgrind.
2011-10-20 11:02:47 +02:00
Aric Stewart 8836da818e usp10/tests: Initialize our SCRIPT_VISATTR array.
Found by Valgrind and Dan Kegel.
2011-10-20 11:02:37 +02:00