[arch-commits] Commit in icu/repos (10 files)
andyrtr at archlinux.org
andyrtr at archlinux.org
Fri Nov 16 17:10:46 UTC 2012
Date: Friday, November 16, 2012 @ 12:10:45
Author: andyrtr
Revision: 171107
db-move: moved icu from [staging] to [testing] (i686, x86_64)
Added:
icu/repos/testing-i686/
icu/repos/testing-i686/PKGBUILD
(from rev 171101, icu/repos/staging-i686/PKGBUILD)
icu/repos/testing-i686/changeset_32780.diff
(from rev 171101, icu/repos/staging-i686/changeset_32780.diff)
icu/repos/testing-i686/icu.8198.revert.icu5431.patch
(from rev 171101, icu/repos/staging-i686/icu.8198.revert.icu5431.patch)
icu/repos/testing-x86_64/
icu/repos/testing-x86_64/PKGBUILD
(from rev 171101, icu/repos/staging-x86_64/PKGBUILD)
icu/repos/testing-x86_64/changeset_32780.diff
(from rev 171101, icu/repos/staging-x86_64/changeset_32780.diff)
icu/repos/testing-x86_64/icu.8198.revert.icu5431.patch
(from rev 171101, icu/repos/staging-x86_64/icu.8198.revert.icu5431.patch)
Deleted:
icu/repos/staging-i686/
icu/repos/staging-x86_64/
----------------------------------------------+
testing-i686/PKGBUILD | 46 ++++++++
testing-i686/changeset_32780.diff | 67 ++++++++++++
testing-i686/icu.8198.revert.icu5431.patch | 129 +++++++++++++++++++++++++
testing-x86_64/PKGBUILD | 46 ++++++++
testing-x86_64/changeset_32780.diff | 67 ++++++++++++
testing-x86_64/icu.8198.revert.icu5431.patch | 129 +++++++++++++++++++++++++
6 files changed, 484 insertions(+)
Copied: icu/repos/testing-i686/PKGBUILD (from rev 171101, icu/repos/staging-i686/PKGBUILD)
===================================================================
--- testing-i686/PKGBUILD (rev 0)
+++ testing-i686/PKGBUILD 2012-11-16 17:10:45 UTC (rev 171107)
@@ -0,0 +1,46 @@
+# $Id$
+# Maintainer: Andreas Radke <andyrtr at archlinux.org>
+# Contributor: Art Gramlich <art at gramlich-net.com>
+
+pkgname=icu
+pkgver=50.1
+pkgrel=2
+pkgdesc="International Components for Unicode library"
+arch=(i686 x86_64)
+url="http://www.icu-project.org/"
+license=('custom:"icu"')
+depends=('gcc-libs>=4.7.1-5' 'sh')
+source=(#http://download.icu-project.org/files/${pkgname}4c/${pkgver}/${pkgname}4c-${pkgver/./_}-src.tgz
+ http://download.icu-project.org/files/${pkgname}4c/${pkgver}/${pkgname}4c-${pkgver//./_}-src.tgz
+ icu.8198.revert.icu5431.patch changeset_32780.diff)
+md5sums=('cf7bf9e56aa6c2057a8b6f464046483e'
+ 'ebd5470fc969c75e52baf4af94a9ee82'
+ '58f4b655e40dddc8e316600019b491b2')
+
+build() {
+ cd ${srcdir}/icu/source
+
+ # fix Malayalam encoding https://bugzilla.redhat.com/show_bug.cgi?id=654200
+ patch -Rp3 -i ${srcdir}/icu.8198.revert.icu5431.patch
+
+ # fix building clients without c++11 http://bugs.icu-project.org/trac/changeset/32780
+ patch -Np4 -i ${srcdir}/changeset_32780.diff
+
+ ./configure --prefix=/usr \
+ --sysconfdir=/etc \
+ --mandir=/usr/share/man
+ make
+}
+
+check() {
+ cd "$srcdir/icu/source"
+ make -k check # passes all
+}
+
+package() {
+ cd ${srcdir}/icu/source
+ make -j1 DESTDIR=${pkgdir} install
+
+ # Install license
+ install -Dm644 ${srcdir}/icu/license.html ${pkgdir}/usr/share/licenses/icu/license.html
+}
Copied: icu/repos/testing-i686/changeset_32780.diff (from rev 171101, icu/repos/staging-i686/changeset_32780.diff)
===================================================================
--- testing-i686/changeset_32780.diff (rev 0)
+++ testing-i686/changeset_32780.diff 2012-11-16 17:10:45 UTC (rev 171107)
@@ -0,0 +1,67 @@
+Index: /icu/trunk/source/test/intltest/strtest.cpp
+===================================================================
+--- /icu/trunk/source/test/intltest/strtest.cpp (revision 32779)
++++ /icu/trunk/source/test/intltest/strtest.cpp (revision 32780)
+@@ -1,5 +1,5 @@
+ /********************************************************************
+ * COPYRIGHT:
+- * Copyright (c) 1997-2011, International Business Machines Corporation and
++ * Copyright (c) 1997-2012, International Business Machines Corporation and
+ * others. All Rights Reserved.
+ ********************************************************************/
+@@ -95,5 +95,5 @@
+ StringTest::Test_U_STRING() {
+ U_STRING_INIT(ustringVar, "aZ0 -", 5);
+- if( sizeof(ustringVar)/sizeof(*ustringVar)!=6 ||
++ if( u_strlen(ustringVar)!=5 ||
+ ustringVar[0]!=0x61 ||
+ ustringVar[1]!=0x5a ||
+Index: /icu/trunk/source/common/unicode/ustring.h
+===================================================================
+--- /icu/trunk/source/common/unicode/ustring.h (revision 32779)
++++ /icu/trunk/source/common/unicode/ustring.h (revision 32780)
+@@ -940,5 +940,5 @@
+ */
+ #if defined(U_DECLARE_UTF16)
+-# define U_STRING_DECL(var, cs, length) static const UChar var[(length)+1]=U_DECLARE_UTF16(cs)
++# define U_STRING_DECL(var, cs, length) static const UChar *var=(const UChar *)U_DECLARE_UTF16(cs)
+ /**@stable ICU 2.0 */
+ # define U_STRING_INIT(var, cs, length)
+Index: /icu/trunk/source/common/unicode/platform.h
+===================================================================
+--- /icu/trunk/source/common/unicode/platform.h (revision 32779)
++++ /icu/trunk/source/common/unicode/platform.h (revision 32780)
+@@ -665,7 +665,7 @@
+ * gcc 4.4 defines the __CHAR16_TYPE__ macro to a usable type but
+ * does not support u"abc" string literals.
+- * C++11 requires support for UTF-16 literals
++ * C++11 and C11 require support for UTF-16 literals
+ */
+-# if (defined(__cplusplus) && __cplusplus >= 201103L)
++# if (defined(__cplusplus) && __cplusplus >= 201103L) || (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 201112L)
+ # define U_HAVE_CHAR16_T 1
+ # else
+Index: /icu/trunk/source/common/unicode/umachine.h
+===================================================================
+--- /icu/trunk/source/common/unicode/umachine.h (revision 32779)
++++ /icu/trunk/source/common/unicode/umachine.h (revision 32780)
+@@ -257,5 +257,5 @@
+ /**
+ * \var UChar
+- * Define UChar to be char16_t, if available,
++ * Define UChar to be UCHAR_TYPE, if that is #defined (for example, to char16_t),
+ * or wchar_t if that is 16 bits wide; always assumed to be unsigned.
+ * If neither is available, then define UChar to be uint16_t.
+@@ -267,8 +267,8 @@
+ * @stable ICU 4.4
+ */
+-
+-/* Define UChar to be compatible with char16_t or wchar_t if possible. */
+-#if U_HAVE_CHAR16_T
+- typedef char16_t UChar;
++#if defined(UCHAR_TYPE)
++ typedef UCHAR_TYPE UChar;
++/* Not #elif U_HAVE_CHAR16_T -- because that is type-incompatible with pre-C++11 callers
++ typedef char16_t UChar; */
+ #elif U_SIZEOF_WCHAR_T==2
+ typedef wchar_t UChar;
Copied: icu/repos/testing-i686/icu.8198.revert.icu5431.patch (from rev 171101, icu/repos/staging-i686/icu.8198.revert.icu5431.patch)
===================================================================
--- testing-i686/icu.8198.revert.icu5431.patch (rev 0)
+++ testing-i686/icu.8198.revert.icu5431.patch 2012-11-16 17:10:45 UTC (rev 171107)
@@ -0,0 +1,129 @@
+Index: icu/trunk/source/layout/IndicReordering.cpp
+===================================================================
+--- icu/trunk/source/layout/IndicReordering.cpp (revision 25772)
++++ icu/trunk/source/layout/IndicReordering.cpp (revision 26090)
+@@ -126,4 +126,8 @@
+ FeatureMask fSMFeatures;
+
++ LEUnicode fPreBaseConsonant;
++ LEUnicode fPreBaseVirama;
++ le_int32 fPBCIndex;
++ FeatureMask fPBCFeatures;
+
+ void saveMatra(LEUnicode matra, le_int32 matraIndex, IndicClassTable::CharClass matraClass)
+@@ -172,5 +176,6 @@
+ fMatraFeatures(0), fMPreOutIndex(-1), fMPreFixups(mpreFixups),
+ fVMabove(0), fVMpost(0), fVMIndex(0), fVMFeatures(0),
+- fSMabove(0), fSMbelow(0), fSMIndex(0), fSMFeatures(0)
++ fSMabove(0), fSMbelow(0), fSMIndex(0), fSMFeatures(0),
++ fPreBaseConsonant(0), fPreBaseVirama(0), fPBCIndex(0), fPBCFeatures(0)
+ {
+ // nothing else to do...
+@@ -191,4 +196,6 @@
+ fVMabove = fVMpost = 0;
+ fSMabove = fSMbelow = 0;
++
++ fPreBaseConsonant = fPreBaseVirama = 0;
+ }
+
+@@ -386,4 +393,12 @@
+ }
+
++ void notePreBaseConsonant(le_uint32 index,LEUnicode PBConsonant, LEUnicode PBVirama, FeatureMask features)
++ {
++ fPBCIndex = index;
++ fPreBaseConsonant = PBConsonant;
++ fPreBaseVirama = PBVirama;
++ fPBCFeatures = features;
++ }
++
+ void noteBaseConsonant()
+ {
+@@ -465,4 +480,20 @@
+ }
+
++ 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()
+ {
+@@ -723,4 +754,5 @@
+ }
+
++
+ IndicClassTable::CharClass charClass = CC_RESERVED;
+ IndicClassTable::CharClass nextClass = CC_RESERVED;
+@@ -730,7 +762,9 @@
+ 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])) {
+@@ -746,12 +780,20 @@
+ 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)) {
++ !(hasBelowBaseForm || hasPostBaseForm || hasPreBaseForm)) {
+ 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;
+@@ -786,10 +828,12 @@
+
+ // 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 'blwf' on first consonant.
+- FeatureMask features = (i == baseLimit? tagArray2 : tagArray1);
++ // Don't put 'pstf' or 'blwf' on anything before the base consonant.
++ FeatureMask features = tagArray1 & ~( pstfFeatureMask | blwfFeatureMask );
+
+ charClass = classTable->getCharClass(ch);
+@@ -842,5 +886,5 @@
+
+ // write below-base consonants
+- if (baseConsonant != lastConsonant) {
++ if (baseConsonant != lastConsonant && !seenPreBaseForm) {
+ for (i = bcSpan + 1; i < postBase; i += 1) {
+ output.writeChar(chars[i], i, tagArray1);
+@@ -872,5 +916,5 @@
+ // write post-base consonants
+ // FIXME: does this put the right tags on post-base consonants?
+- if (baseConsonant != lastConsonant) {
++ if (baseConsonant != lastConsonant && !seenPreBaseForm) {
+ if (postBase <= lastConsonant) {
+ for (i = postBase; i <= lastConsonant; i += 1) {
Copied: icu/repos/testing-x86_64/PKGBUILD (from rev 171101, icu/repos/staging-x86_64/PKGBUILD)
===================================================================
--- testing-x86_64/PKGBUILD (rev 0)
+++ testing-x86_64/PKGBUILD 2012-11-16 17:10:45 UTC (rev 171107)
@@ -0,0 +1,46 @@
+# $Id$
+# Maintainer: Andreas Radke <andyrtr at archlinux.org>
+# Contributor: Art Gramlich <art at gramlich-net.com>
+
+pkgname=icu
+pkgver=50.1
+pkgrel=2
+pkgdesc="International Components for Unicode library"
+arch=(i686 x86_64)
+url="http://www.icu-project.org/"
+license=('custom:"icu"')
+depends=('gcc-libs>=4.7.1-5' 'sh')
+source=(#http://download.icu-project.org/files/${pkgname}4c/${pkgver}/${pkgname}4c-${pkgver/./_}-src.tgz
+ http://download.icu-project.org/files/${pkgname}4c/${pkgver}/${pkgname}4c-${pkgver//./_}-src.tgz
+ icu.8198.revert.icu5431.patch changeset_32780.diff)
+md5sums=('cf7bf9e56aa6c2057a8b6f464046483e'
+ 'ebd5470fc969c75e52baf4af94a9ee82'
+ '58f4b655e40dddc8e316600019b491b2')
+
+build() {
+ cd ${srcdir}/icu/source
+
+ # fix Malayalam encoding https://bugzilla.redhat.com/show_bug.cgi?id=654200
+ patch -Rp3 -i ${srcdir}/icu.8198.revert.icu5431.patch
+
+ # fix building clients without c++11 http://bugs.icu-project.org/trac/changeset/32780
+ patch -Np4 -i ${srcdir}/changeset_32780.diff
+
+ ./configure --prefix=/usr \
+ --sysconfdir=/etc \
+ --mandir=/usr/share/man
+ make
+}
+
+check() {
+ cd "$srcdir/icu/source"
+ make -k check # passes all
+}
+
+package() {
+ cd ${srcdir}/icu/source
+ make -j1 DESTDIR=${pkgdir} install
+
+ # Install license
+ install -Dm644 ${srcdir}/icu/license.html ${pkgdir}/usr/share/licenses/icu/license.html
+}
Copied: icu/repos/testing-x86_64/changeset_32780.diff (from rev 171101, icu/repos/staging-x86_64/changeset_32780.diff)
===================================================================
--- testing-x86_64/changeset_32780.diff (rev 0)
+++ testing-x86_64/changeset_32780.diff 2012-11-16 17:10:45 UTC (rev 171107)
@@ -0,0 +1,67 @@
+Index: /icu/trunk/source/test/intltest/strtest.cpp
+===================================================================
+--- /icu/trunk/source/test/intltest/strtest.cpp (revision 32779)
++++ /icu/trunk/source/test/intltest/strtest.cpp (revision 32780)
+@@ -1,5 +1,5 @@
+ /********************************************************************
+ * COPYRIGHT:
+- * Copyright (c) 1997-2011, International Business Machines Corporation and
++ * Copyright (c) 1997-2012, International Business Machines Corporation and
+ * others. All Rights Reserved.
+ ********************************************************************/
+@@ -95,5 +95,5 @@
+ StringTest::Test_U_STRING() {
+ U_STRING_INIT(ustringVar, "aZ0 -", 5);
+- if( sizeof(ustringVar)/sizeof(*ustringVar)!=6 ||
++ if( u_strlen(ustringVar)!=5 ||
+ ustringVar[0]!=0x61 ||
+ ustringVar[1]!=0x5a ||
+Index: /icu/trunk/source/common/unicode/ustring.h
+===================================================================
+--- /icu/trunk/source/common/unicode/ustring.h (revision 32779)
++++ /icu/trunk/source/common/unicode/ustring.h (revision 32780)
+@@ -940,5 +940,5 @@
+ */
+ #if defined(U_DECLARE_UTF16)
+-# define U_STRING_DECL(var, cs, length) static const UChar var[(length)+1]=U_DECLARE_UTF16(cs)
++# define U_STRING_DECL(var, cs, length) static const UChar *var=(const UChar *)U_DECLARE_UTF16(cs)
+ /**@stable ICU 2.0 */
+ # define U_STRING_INIT(var, cs, length)
+Index: /icu/trunk/source/common/unicode/platform.h
+===================================================================
+--- /icu/trunk/source/common/unicode/platform.h (revision 32779)
++++ /icu/trunk/source/common/unicode/platform.h (revision 32780)
+@@ -665,7 +665,7 @@
+ * gcc 4.4 defines the __CHAR16_TYPE__ macro to a usable type but
+ * does not support u"abc" string literals.
+- * C++11 requires support for UTF-16 literals
++ * C++11 and C11 require support for UTF-16 literals
+ */
+-# if (defined(__cplusplus) && __cplusplus >= 201103L)
++# if (defined(__cplusplus) && __cplusplus >= 201103L) || (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 201112L)
+ # define U_HAVE_CHAR16_T 1
+ # else
+Index: /icu/trunk/source/common/unicode/umachine.h
+===================================================================
+--- /icu/trunk/source/common/unicode/umachine.h (revision 32779)
++++ /icu/trunk/source/common/unicode/umachine.h (revision 32780)
+@@ -257,5 +257,5 @@
+ /**
+ * \var UChar
+- * Define UChar to be char16_t, if available,
++ * Define UChar to be UCHAR_TYPE, if that is #defined (for example, to char16_t),
+ * or wchar_t if that is 16 bits wide; always assumed to be unsigned.
+ * If neither is available, then define UChar to be uint16_t.
+@@ -267,8 +267,8 @@
+ * @stable ICU 4.4
+ */
+-
+-/* Define UChar to be compatible with char16_t or wchar_t if possible. */
+-#if U_HAVE_CHAR16_T
+- typedef char16_t UChar;
++#if defined(UCHAR_TYPE)
++ typedef UCHAR_TYPE UChar;
++/* Not #elif U_HAVE_CHAR16_T -- because that is type-incompatible with pre-C++11 callers
++ typedef char16_t UChar; */
+ #elif U_SIZEOF_WCHAR_T==2
+ typedef wchar_t UChar;
Copied: icu/repos/testing-x86_64/icu.8198.revert.icu5431.patch (from rev 171101, icu/repos/staging-x86_64/icu.8198.revert.icu5431.patch)
===================================================================
--- testing-x86_64/icu.8198.revert.icu5431.patch (rev 0)
+++ testing-x86_64/icu.8198.revert.icu5431.patch 2012-11-16 17:10:45 UTC (rev 171107)
@@ -0,0 +1,129 @@
+Index: icu/trunk/source/layout/IndicReordering.cpp
+===================================================================
+--- icu/trunk/source/layout/IndicReordering.cpp (revision 25772)
++++ icu/trunk/source/layout/IndicReordering.cpp (revision 26090)
+@@ -126,4 +126,8 @@
+ FeatureMask fSMFeatures;
+
++ LEUnicode fPreBaseConsonant;
++ LEUnicode fPreBaseVirama;
++ le_int32 fPBCIndex;
++ FeatureMask fPBCFeatures;
+
+ void saveMatra(LEUnicode matra, le_int32 matraIndex, IndicClassTable::CharClass matraClass)
+@@ -172,5 +176,6 @@
+ fMatraFeatures(0), fMPreOutIndex(-1), fMPreFixups(mpreFixups),
+ fVMabove(0), fVMpost(0), fVMIndex(0), fVMFeatures(0),
+- fSMabove(0), fSMbelow(0), fSMIndex(0), fSMFeatures(0)
++ fSMabove(0), fSMbelow(0), fSMIndex(0), fSMFeatures(0),
++ fPreBaseConsonant(0), fPreBaseVirama(0), fPBCIndex(0), fPBCFeatures(0)
+ {
+ // nothing else to do...
+@@ -191,4 +196,6 @@
+ fVMabove = fVMpost = 0;
+ fSMabove = fSMbelow = 0;
++
++ fPreBaseConsonant = fPreBaseVirama = 0;
+ }
+
+@@ -386,4 +393,12 @@
+ }
+
++ void notePreBaseConsonant(le_uint32 index,LEUnicode PBConsonant, LEUnicode PBVirama, FeatureMask features)
++ {
++ fPBCIndex = index;
++ fPreBaseConsonant = PBConsonant;
++ fPreBaseVirama = PBVirama;
++ fPBCFeatures = features;
++ }
++
+ void noteBaseConsonant()
+ {
+@@ -465,4 +480,20 @@
+ }
+
++ 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()
+ {
+@@ -723,4 +754,5 @@
+ }
+
++
+ IndicClassTable::CharClass charClass = CC_RESERVED;
+ IndicClassTable::CharClass nextClass = CC_RESERVED;
+@@ -730,7 +762,9 @@
+ 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])) {
+@@ -746,12 +780,20 @@
+ 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)) {
++ !(hasBelowBaseForm || hasPostBaseForm || hasPreBaseForm)) {
+ 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;
+@@ -786,10 +828,12 @@
+
+ // 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 'blwf' on first consonant.
+- FeatureMask features = (i == baseLimit? tagArray2 : tagArray1);
++ // Don't put 'pstf' or 'blwf' on anything before the base consonant.
++ FeatureMask features = tagArray1 & ~( pstfFeatureMask | blwfFeatureMask );
+
+ charClass = classTable->getCharClass(ch);
+@@ -842,5 +886,5 @@
+
+ // write below-base consonants
+- if (baseConsonant != lastConsonant) {
++ if (baseConsonant != lastConsonant && !seenPreBaseForm) {
+ for (i = bcSpan + 1; i < postBase; i += 1) {
+ output.writeChar(chars[i], i, tagArray1);
+@@ -872,5 +916,5 @@
+ // write post-base consonants
+ // FIXME: does this put the right tags on post-base consonants?
+- if (baseConsonant != lastConsonant) {
++ if (baseConsonant != lastConsonant && !seenPreBaseForm) {
+ if (postBase <= lastConsonant) {
+ for (i = postBase; i <= lastConsonant; i += 1) {
More information about the arch-commits
mailing list