diff options
author | Peter Breitenlohner <peb@mppmu.mpg.de> | 2012-06-28 08:09:18 +0000 |
---|---|---|
committer | Peter Breitenlohner <peb@mppmu.mpg.de> | 2012-06-28 08:09:18 +0000 |
commit | 1a70c47822d9d7b990ea89c6fdb09ea203b074be (patch) | |
tree | 84c47052c65e16d1e6a3a1fe9a6596910be71465 /Build/source/libs/icu/icu-49.1-PATCHES/patch-82-Indic-bugfix | |
parent | 72073e34d626c482f4458495b66c3a0c56f496dc (diff) |
icu: bug fix for Indic scripts
git-svn-id: svn://tug.org/texlive/trunk@26904 c570f23f-e606-0410-a88d-b1316a301751
Diffstat (limited to 'Build/source/libs/icu/icu-49.1-PATCHES/patch-82-Indic-bugfix')
-rw-r--r-- | Build/source/libs/icu/icu-49.1-PATCHES/patch-82-Indic-bugfix | 230 |
1 files changed, 230 insertions, 0 deletions
diff --git a/Build/source/libs/icu/icu-49.1-PATCHES/patch-82-Indic-bugfix b/Build/source/libs/icu/icu-49.1-PATCHES/patch-82-Indic-bugfix new file mode 100644 index 00000000000..0a8a6ee871f --- /dev/null +++ b/Build/source/libs/icu/icu-49.1-PATCHES/patch-82-Indic-bugfix @@ -0,0 +1,230 @@ +diff -ur icu-49.1.orig/source/layout/IndicClassTables.cpp icu-49.1/source/layout/IndicClassTables.cpp +--- icu-49.1.orig/source/layout/IndicClassTables.cpp 2012-03-21 18:38:10.000000000 +0100 ++++ icu-49.1/source/layout/IndicClassTables.cpp 2012-06-18 14:23:38.000000000 +0200 +@@ -66,7 +66,6 @@ + // special forms... (Bengali RA?) + #define _bb (_ct | CF_BELOW_BASE) + #define _pb (_ct | CF_POST_BASE) +-#define _fb (_ct | CF_PRE_BASE) + #define _vt (_bb | CF_VATTU) + #define _rv (_vt | CF_REPH) + #define _rp (_pb | CF_REPH) +@@ -198,7 +197,7 @@ + _xx, _xx, _mp, _mp, _xx, _iv, _iv, _iv, _iv, _iv, _iv, _iv, _iv, _xx, _iv, _iv, // 0D00 - 0D0F + _iv, _xx, _iv, _iv, _iv, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, // 0D10 - 0D1F + _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _xx, _ct, _ct, _ct, _ct, _ct, _pb, // 0D20 - 0D2F +- _fb, _fb, _bb, _ct, _ct, _pb, _ct, _ct, _ct, _ct, _xx, _xx, _xx, _xx, _r2, _dr, // 0D30 - 0D3F ++ _pb, _cn, _bb, _ct, _ct, _pb, _ct, _ct, _ct, _ct, _xx, _xx, _xx, _xx, _r2, _dr, // 0D30 - 0D3F + _dr, _dr, _dr, _dr, _xx, _xx, _l1, _l1, _dl, _xx, _s1, _s2, _s3, _vr, _xx, _xx, // 0D40 - 0D4F + _xx, _xx, _xx, _xx, _xx, _xx, _xx, _m2, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, // 0D50 - 0D5F + _iv, _iv, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx // 0D60 - 0D6F +diff -ur icu-49.1.orig/source/layout/IndicReordering.cpp icu-49.1/source/layout/IndicReordering.cpp +--- icu-49.1.orig/source/layout/IndicReordering.cpp 2012-03-21 18:38:10.000000000 +0100 ++++ icu-49.1/source/layout/IndicReordering.cpp 2012-06-18 14:38:25.000000000 +0200 +@@ -129,10 +129,6 @@ + le_int32 fSMIndex; + FeatureMask fSMFeatures; + +- LEUnicode fPreBaseConsonant; +- LEUnicode fPreBaseVirama; +- le_int32 fPBCIndex; +- FeatureMask fPBCFeatures; + + void saveMatra(LEUnicode matra, le_int32 matraIndex, IndicClassTable::CharClass matraClass) + { +@@ -179,8 +175,7 @@ + fMpost(0), fMpostIndex(0), fLengthMark(0), fLengthMarkIndex(0), fAlLakuna(0), fAlLakunaIndex(0), + fMatraFeatures(0), fMPreOutIndex(-1), fMPreFixups(mpreFixups), + fVMabove(0), fVMpost(0), fVMIndex(0), fVMFeatures(0), +- fSMabove(0), fSMbelow(0), fSMIndex(0), fSMFeatures(0), +- fPreBaseConsonant(0), fPreBaseVirama(0), fPBCIndex(0), fPBCFeatures(0) ++ fSMabove(0), fSMbelow(0), fSMIndex(0), fSMFeatures(0) + { + // nothing else to do... + } +@@ -199,8 +194,6 @@ + + fVMabove = fVMpost = 0; + fSMabove = fSMbelow = 0; +- +- fPreBaseConsonant = fPreBaseVirama = 0; + } + + void writeChar(LEUnicode ch, le_uint32 charIndex, FeatureMask charFeatures) +@@ -396,14 +389,6 @@ + } + } + +- void notePreBaseConsonant(le_uint32 index,LEUnicode PBConsonant, LEUnicode PBVirama, FeatureMask features) +- { +- fPBCIndex = index; +- fPreBaseConsonant = PBConsonant; +- fPreBaseVirama = PBVirama; +- fPBCFeatures = features; +- } +- + void noteBaseConsonant() + { + if (fMPreFixups != NULL && fMPreOutIndex >= 0) { +@@ -483,22 +468,6 @@ + } + } + +- void writePreBaseConsonant() +- { +- // The TDIL spec says that consonant + virama + RRA should produce a rakar in Malayalam. However, +- // it seems that almost none of the fonts for Malayalam are set up to handle this. +- // So, we're going to force the issue here by using the rakar as defined with RA in most fonts. +- +- if (fPreBaseConsonant == 0x0d31) { // RRA +- fPreBaseConsonant = 0x0d30; // RA +- } +- +- if (fPreBaseConsonant != 0) { +- writeChar(fPreBaseConsonant, fPBCIndex, fPBCFeatures); +- writeChar(fPreBaseVirama,fPBCIndex-1,fPBCFeatures); +- } +- } +- + le_int32 getOutputIndex() + { + return fOutIndex; +@@ -765,7 +734,6 @@ + lastConsonant -= 1; + } + +- + IndicClassTable::CharClass charClass = CC_RESERVED; + IndicClassTable::CharClass nextClass = CC_RESERVED; + le_int32 baseConsonant = lastConsonant; +@@ -773,11 +741,9 @@ + le_int32 postBaseLimit = classTable->scriptFlags & SF_POST_BASE_LIMIT_MASK; + le_bool seenVattu = FALSE; + le_bool seenBelowBaseForm = FALSE; +- le_bool seenPreBaseForm = FALSE; + le_bool hasNukta = FALSE; + le_bool hasBelowBaseForm = FALSE; + le_bool hasPostBaseForm = FALSE; +- le_bool hasPreBaseForm = FALSE; + + if (postBase < markStart && classTable->isNukta(chars[postBase])) { + charClass = CC_NUKTA; +@@ -791,22 +757,14 @@ + + hasBelowBaseForm = IndicClassTable::hasBelowBaseForm(charClass) && !hasNukta; + hasPostBaseForm = IndicClassTable::hasPostBaseForm(charClass) && !hasNukta; +- hasPreBaseForm = IndicClassTable::hasPreBaseForm(charClass) && !hasNukta; + + if (IndicClassTable::isConsonant(charClass)) { + if (postBaseLimit == 0 || seenVattu || + (baseConsonant > baseLimit && !classTable->isVirama(chars[baseConsonant - 1])) || +- !(hasBelowBaseForm || hasPostBaseForm || hasPreBaseForm)) { ++ !(hasBelowBaseForm || hasPostBaseForm)) { + break; + } + +- // Note any pre-base consonants +- if ( baseConsonant == lastConsonant && lastConsonant > 0 && +- hasPreBaseForm && classTable->isVirama(chars[baseConsonant - 1])) { +- output.notePreBaseConsonant(lastConsonant,chars[lastConsonant],chars[lastConsonant-1],tagArray2); +- seenPreBaseForm = TRUE; +- +- } + // consonants with nuktas are never vattus + seenVattu = IndicClassTable::isVattu(charClass) && !hasNukta; + +@@ -839,14 +797,12 @@ + } + + // write any pre-base consonants +- output.writePreBaseConsonant(); +- + le_bool supressVattu = TRUE; + + for (i = baseLimit; i < baseConsonant; i += 1) { + LEUnicode ch = chars[i]; +- // Don't put 'pstf' or 'blwf' on anything before the base consonant. +- FeatureMask features = tagArray1 & ~( pstfFeatureMask | blwfFeatureMask ); ++ // Don't put 'blwf' on first consonant. ++ FeatureMask features = (i == baseLimit? tagArray2 : tagArray1); + + charClass = classTable->getCharClass(ch); + nextClass = classTable->getCharClass(chars[i + 1]); +@@ -897,7 +853,7 @@ + } + + // write below-base consonants +- if (baseConsonant != lastConsonant && !seenPreBaseForm) { ++ if (baseConsonant != lastConsonant) { + for (i = bcSpan + 1; i < postBase; i += 1) { + output.writeChar(chars[i], i, tagArray1); + } +@@ -927,7 +883,7 @@ + + // write post-base consonants + // FIXME: does this put the right tags on post-base consonants? +- if (baseConsonant != lastConsonant && !seenPreBaseForm) { ++ if (baseConsonant != lastConsonant) { + if (postBase <= lastConsonant) { + for (i = postBase; i <= lastConsonant; i += 1) { + output.writeChar(chars[i], i, tagArray3); +@@ -1195,7 +1151,7 @@ + } + + +-void IndicReordering::getDynamicProperties( DynamicProperties *, const IndicClassTable *classTable ) { ++void IndicReordering::getDynamicProperties( DynamicProperties */*dProps*/, const IndicClassTable *classTable ) { + + + LEUnicode currentChar; +diff -ur icu-49.1.orig/source/layout/IndicReordering.h icu-49.1/source/layout/IndicReordering.h +--- icu-49.1.orig/source/layout/IndicReordering.h 2012-03-21 18:38:08.000000000 +0100 ++++ icu-49.1/source/layout/IndicReordering.h 2012-06-18 14:27:48.000000000 +0200 +@@ -50,7 +50,6 @@ + #define CF_BELOW_BASE 0x10000000U + #define CF_POST_BASE 0x08000000U + #define CF_LENGTH_MARK 0x04000000U +-#define CF_PRE_BASE 0x02000000U + + #define CF_POS_BEFORE 0x00300000U + #define CF_POS_BELOW 0x00200000U +@@ -121,7 +120,6 @@ + inline le_bool hasPostBaseForm(LEUnicode ch) const; + inline le_bool hasBelowBaseForm(LEUnicode ch) const; + inline le_bool hasAboveBaseForm(LEUnicode ch) const; +- inline le_bool hasPreBaseForm(LEUnicode ch) const; + + inline static le_bool isVowelModifier(CharClass charClass); + inline static le_bool isStressMark(CharClass charClass); +@@ -138,7 +136,6 @@ + inline static le_bool hasPostBaseForm(CharClass charClass); + inline static le_bool hasBelowBaseForm(CharClass charClass); + inline static le_bool hasAboveBaseForm(CharClass charClass); +- inline static le_bool hasPreBaseForm(CharClass charClass); + + static const IndicClassTable *getScriptClassTable(le_int32 scriptCode); + }; +@@ -260,11 +257,6 @@ + return (charClass & CF_POST_BASE) != 0; + } + +-inline le_bool IndicClassTable::hasPreBaseForm(CharClass charClass) +-{ +- return (charClass & CF_PRE_BASE) != 0; +-} +- + inline le_bool IndicClassTable::hasBelowBaseForm(CharClass charClass) + { + return (charClass & CF_BELOW_BASE) != 0; +@@ -345,11 +337,6 @@ + return hasBelowBaseForm(getCharClass(ch)); + } + +-inline le_bool IndicClassTable::hasPreBaseForm(LEUnicode ch) const +-{ +- return hasPreBaseForm(getCharClass(ch)); +-} +- + inline le_bool IndicClassTable::hasAboveBaseForm(LEUnicode ch) const + { + return hasAboveBaseForm(getCharClass(ch)); |