[arch-commits] Commit in icu/repos (10 files)
Jan de Groot
jgc at archlinux.org
Sat Apr 22 11:59:20 UTC 2017
Date: Saturday, April 22, 2017 @ 11:59:19
Author: jgc
Revision: 293903
archrelease: copy trunk to testing-i686, testing-x86_64
Added:
icu/repos/testing-i686/
icu/repos/testing-i686/PKGBUILD
(from rev 293902, icu/trunk/PKGBUILD)
icu/repos/testing-i686/changeset-39671-CVE-2017-7867+7868.patch
(from rev 293902, icu/trunk/changeset-39671-CVE-2017-7867+7868.patch)
icu/repos/testing-i686/icu-58.1-iterator-reset.patch
(from rev 293902, icu/trunk/icu-58.1-iterator-reset.patch)
icu/repos/testing-i686/icu.8198.revert.icu5431.patch
(from rev 293902, icu/trunk/icu.8198.revert.icu5431.patch)
icu/repos/testing-x86_64/
icu/repos/testing-x86_64/PKGBUILD
(from rev 293902, icu/trunk/PKGBUILD)
icu/repos/testing-x86_64/changeset-39671-CVE-2017-7867+7868.patch
(from rev 293902, icu/trunk/changeset-39671-CVE-2017-7867+7868.patch)
icu/repos/testing-x86_64/icu-58.1-iterator-reset.patch
(from rev 293902, icu/trunk/icu-58.1-iterator-reset.patch)
icu/repos/testing-x86_64/icu.8198.revert.icu5431.patch
(from rev 293902, icu/trunk/icu.8198.revert.icu5431.patch)
---------------------------------------------------------+
testing-i686/PKGBUILD | 55 ++++
testing-i686/changeset-39671-CVE-2017-7867+7868.patch | 155 ++++++++++++++
testing-i686/icu-58.1-iterator-reset.patch | 124 +++++++++++
testing-i686/icu.8198.revert.icu5431.patch | 129 +++++++++++
testing-x86_64/PKGBUILD | 55 ++++
testing-x86_64/changeset-39671-CVE-2017-7867+7868.patch | 155 ++++++++++++++
testing-x86_64/icu-58.1-iterator-reset.patch | 124 +++++++++++
testing-x86_64/icu.8198.revert.icu5431.patch | 129 +++++++++++
8 files changed, 926 insertions(+)
Copied: icu/repos/testing-i686/PKGBUILD (from rev 293902, icu/trunk/PKGBUILD)
===================================================================
--- testing-i686/PKGBUILD (rev 0)
+++ testing-i686/PKGBUILD 2017-04-22 11:59:19 UTC (rev 293903)
@@ -0,0 +1,55 @@
+# $Id$
+# Maintainer: Andreas Radke <andyrtr at archlinux.org>
+# Contributor: Art Gramlich <art at gramlich-net.com>
+
+pkgname=icu
+pkgver=58.2
+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')
+#makedepends=('clang')
+# no https available
+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-58.1-iterator-reset.patch
+ changeset-39671-CVE-2017-7867+7868.patch)
+# upstream offers md5sum checks, only asc file for md5sum check
+md5sums=('fac212b32b7ec7ab007a12dff1f3aea1'
+ '8c09ae284967def053e9579d64d0f83c'
+ 'cea722333992ad6ede88ebd0bd21cfcd')
+sha256sums=('2b0a4410153a9b20de0e20c7d8b66049a72aef244b53683d0d7521371683da0c'
+ '8034928bcff89eca84d8d0f22fa9b6bdba5e0608a984f64fe7afe41b91bbea97'
+ 'cf0c6f946d7336bd46857c52308fd67171918f24fcf19015b5a47a8d2fb8ca0c')
+
+prepare() {
+ cd icu/source
+ # http://bugs.icu-project.org/trac/ticket/12827
+ patch -Np4 -i "${srcdir}"/icu-58.1-iterator-reset.patch
+
+ patch -Np4 -i "${srcdir}"/changeset-39671-CVE-2017-7867+7868.patch
+}
+
+build() {
+ cd icu/source
+ ./configure --prefix=/usr \
+ --sysconfdir=/etc \
+ --mandir=/usr/share/man \
+ --sbindir=/usr/bin
+ make
+}
+
+check() {
+ cd icu/source
+ make -k check # passes all
+}
+
+package() {
+ cd 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-39671-CVE-2017-7867+7868.patch (from rev 293902, icu/trunk/changeset-39671-CVE-2017-7867+7868.patch)
===================================================================
--- testing-i686/changeset-39671-CVE-2017-7867+7868.patch (rev 0)
+++ testing-i686/changeset-39671-CVE-2017-7867+7868.patch 2017-04-22 11:59:19 UTC (rev 293903)
@@ -0,0 +1,155 @@
+Index: /trunk/icu4c/source/common/utext.cpp
+===================================================================
+--- /trunk/icu4c/source/common/utext.cpp (revision 39670)
++++ /trunk/icu4c/source/common/utext.cpp (revision 39671)
+@@ -848,7 +848,13 @@
+
+ // Chunk size.
+-// Must be less than 85, because of byte mapping from UChar indexes to native indexes.
+-// Worst case is three native bytes to one UChar. (Supplemenaries are 4 native bytes
+-// to two UChars.)
++// Must be less than 42 (256/6), because of byte mapping from UChar indexes to native indexes.
++// Worst case there are six UTF-8 bytes per UChar.
++// obsolete 6 byte form fd + 5 trails maps to fffd
++// obsolete 5 byte form fc + 4 trails maps to fffd
++// non-shortest 4 byte forms maps to fffd
++// normal supplementaries map to a pair of utf-16, two utf8 bytes per utf-16 unit
++// mapToUChars array size must allow for the worst case, 6.
++// This could be brought down to 4, by treating fd and fc as pure illegal,
++// rather than obsolete lead bytes. But that is not compatible with the utf-8 access macros.
+ //
+ enum { UTF8_TEXT_CHUNK_SIZE=32 };
+@@ -890,5 +896,5 @@
+ // one for a supplementary starting in the last normal position,
+ // and one for an entry for the buffer limit position.
+- uint8_t mapToUChars[UTF8_TEXT_CHUNK_SIZE*3+6]; // Map native offset from bufNativeStart to
++ uint8_t mapToUChars[UTF8_TEXT_CHUNK_SIZE*6+6]; // Map native offset from bufNativeStart to
+ // correspoding offset in filled part of buf.
+ int32_t align;
+@@ -1033,4 +1039,5 @@
+ u8b = (UTF8Buf *)ut->p; // the current buffer
+ mapIndex = ix - u8b->toUCharsMapStart;
++ U_ASSERT(mapIndex < (int32_t)sizeof(UTF8Buf::mapToUChars));
+ ut->chunkOffset = u8b->mapToUChars[mapIndex] - u8b->bufStartIdx;
+ return TRUE;
+@@ -1299,4 +1306,8 @@
+ // If index is at the end, there is no character there to look at.
+ if (ix != ut->b) {
++ // Note: this function will only move the index back if it is on a trail byte
++ // and there is a preceding lead byte and the sequence from the lead
++ // through this trail could be part of a valid UTF-8 sequence
++ // Otherwise the index remains unchanged.
+ U8_SET_CP_START(s8, 0, ix);
+ }
+@@ -1312,5 +1323,8 @@
+ uint8_t *mapToNative = u8b->mapToNative;
+ uint8_t *mapToUChars = u8b->mapToUChars;
+- int32_t toUCharsMapStart = ix - (UTF8_TEXT_CHUNK_SIZE*3 + 1);
++ int32_t toUCharsMapStart = ix - sizeof(UTF8Buf::mapToUChars) + 1;
++ // Note that toUCharsMapStart can be negative. Happens when the remaining
++ // text from current position to the beginning is less than the buffer size.
++ // + 1 because mapToUChars must have a slot at the end for the bufNativeLimit entry.
+ int32_t destIx = UTF8_TEXT_CHUNK_SIZE+2; // Start in the overflow region
+ // at end of buffer to leave room
+@@ -1339,4 +1353,5 @@
+ // Special case ASCII range for speed.
+ buf[destIx] = (UChar)c;
++ U_ASSERT(toUCharsMapStart <= srcIx);
+ mapToUChars[srcIx - toUCharsMapStart] = (uint8_t)destIx;
+ mapToNative[destIx] = (uint8_t)(srcIx - toUCharsMapStart);
+@@ -1368,4 +1383,5 @@
+ mapToUChars[sIx-- - toUCharsMapStart] = (uint8_t)destIx;
+ } while (sIx >= srcIx);
++ U_ASSERT(toUCharsMapStart <= (srcIx+1));
+
+ // Set native indexing limit to be the current position.
+@@ -1542,4 +1558,5 @@
+ U_ASSERT(index<=ut->chunkNativeLimit);
+ int32_t mapIndex = index - u8b->toUCharsMapStart;
++ U_ASSERT(mapIndex < (int32_t)sizeof(UTF8Buf::mapToUChars));
+ int32_t offset = u8b->mapToUChars[mapIndex] - u8b->bufStartIdx;
+ U_ASSERT(offset>=0 && offset<=ut->chunkLength);
+Index: /trunk/icu4c/source/test/intltest/utxttest.cpp
+===================================================================
+--- /trunk/icu4c/source/test/intltest/utxttest.cpp (revision 39670)
++++ /trunk/icu4c/source/test/intltest/utxttest.cpp (revision 39671)
+@@ -68,4 +68,6 @@
+ case 7: name = "Ticket12130";
+ if (exec) Ticket12130(); break;
++ case 8: name = "Ticket12888";
++ if (exec) Ticket12888(); break;
+ default: name = ""; break;
+ }
+@@ -1584,2 +1586,62 @@
+ utext_close(&ut);
+ }
++
++// Ticket 12888: bad handling of illegal utf-8 containing many instances of the archaic, now illegal,
++// six byte utf-8 forms. Original implementation had an assumption that
++// there would be at most three utf-8 bytes per UTF-16 code unit.
++// The five and six byte sequences map to a single replacement character.
++
++void UTextTest::Ticket12888() {
++ const char *badString =
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80";
++
++ UErrorCode status = U_ZERO_ERROR;
++ LocalUTextPointer ut(utext_openUTF8(NULL, badString, -1, &status));
++ TEST_SUCCESS(status);
++ for (;;) {
++ UChar32 c = utext_next32(ut.getAlias());
++ if (c == U_SENTINEL) {
++ break;
++ }
++ }
++ int32_t endIdx = utext_getNativeIndex(ut.getAlias());
++ if (endIdx != (int32_t)strlen(badString)) {
++ errln("%s:%d expected=%d, actual=%d", __FILE__, __LINE__, strlen(badString), endIdx);
++ return;
++ }
++
++ for (int32_t prevIndex = endIdx; prevIndex>0;) {
++ UChar32 c = utext_previous32(ut.getAlias());
++ int32_t currentIndex = utext_getNativeIndex(ut.getAlias());
++ if (c != 0xfffd) {
++ errln("%s:%d (expected, actual, index) = (%d, %d, %d)\n",
++ __FILE__, __LINE__, 0xfffd, c, currentIndex);
++ break;
++ }
++ if (currentIndex != prevIndex - 6) {
++ errln("%s:%d: wrong index. Expected, actual = %d, %d",
++ __FILE__, __LINE__, prevIndex - 6, currentIndex);
++ break;
++ }
++ prevIndex = currentIndex;
++ }
++}
+Index: /trunk/icu4c/source/test/intltest/utxttest.h
+===================================================================
+--- /trunk/icu4c/source/test/intltest/utxttest.h (revision 39670)
++++ /trunk/icu4c/source/test/intltest/utxttest.h (revision 39671)
+@@ -39,4 +39,5 @@
+ void Ticket10983();
+ void Ticket12130();
++ void Ticket12888();
+
+ private:
Copied: icu/repos/testing-i686/icu-58.1-iterator-reset.patch (from rev 293902, icu/trunk/icu-58.1-iterator-reset.patch)
===================================================================
--- testing-i686/icu-58.1-iterator-reset.patch (rev 0)
+++ testing-i686/icu-58.1-iterator-reset.patch 2017-04-22 11:59:19 UTC (rev 293903)
@@ -0,0 +1,124 @@
+Index: /icu/trunk/source/common/ulist.c
+===================================================================
+--- /icu/trunk/source/common/ulist.c (revision 39483)
++++ /icu/trunk/source/common/ulist.c (revision 39484)
+@@ -30,5 +30,4 @@
+
+ int32_t size;
+- int32_t currentIndex;
+ };
+
+@@ -52,5 +51,4 @@
+ newList->tail = NULL;
+ newList->size = 0;
+- newList->currentIndex = -1;
+
+ return newList;
+@@ -81,6 +79,7 @@
+ p->next->previous = p->previous;
+ }
+- list->curr = NULL;
+- list->currentIndex = 0;
++ if (p == list->curr) {
++ list->curr = p->next;
++ }
+ --list->size;
+ if (p->forceDelete) {
+@@ -151,5 +150,4 @@
+ list->head->previous = newItem;
+ list->head = newItem;
+- list->currentIndex++;
+ }
+
+@@ -194,5 +192,4 @@
+ curr = list->curr;
+ list->curr = curr->next;
+- list->currentIndex++;
+
+ return curr->data;
+@@ -210,5 +207,4 @@
+ if (list != NULL) {
+ list->curr = list->head;
+- list->currentIndex = 0;
+ }
+ }
+@@ -273,3 +269,2 @@
+ return (UList *)(en->context);
+ }
+-
+Index: /icu/trunk/source/i18n/ucol_res.cpp
+===================================================================
+--- /icu/trunk/source/i18n/ucol_res.cpp (revision 39483)
++++ /icu/trunk/source/i18n/ucol_res.cpp (revision 39484)
+@@ -681,4 +681,5 @@
+ }
+ memcpy(en, &defaultKeywordValues, sizeof(UEnumeration));
++ ulist_resetList(sink.values); // Initialize the iterator.
+ en->context = sink.values;
+ sink.values = NULL; // Avoid deletion in the sink destructor.
+Index: /icu/trunk/source/test/intltest/apicoll.cpp
+===================================================================
+--- /icu/trunk/source/test/intltest/apicoll.cpp (revision 39483)
++++ /icu/trunk/source/test/intltest/apicoll.cpp (revision 39484)
+@@ -82,14 +82,7 @@
+ col = Collator::createInstance(Locale::getEnglish(), success);
+ if (U_FAILURE(success)){
+- errcheckln(success, "Default Collator creation failed. - %s", u_errorName(success));
+- return;
+- }
+-
+- StringEnumeration* kwEnum = col->getKeywordValuesForLocale("", Locale::getEnglish(),true,success);
+- if (U_FAILURE(success)){
+- errcheckln(success, "Get Keyword Values for Locale failed. - %s", u_errorName(success));
+- return;
+- }
+- delete kwEnum;
++ errcheckln(success, "English Collator creation failed. - %s", u_errorName(success));
++ return;
++ }
+
+ col->getVersion(versionArray);
+@@ -230,4 +223,27 @@
+ delete aFrCol;
+ delete junk;
++}
++
++void CollationAPITest::TestKeywordValues() {
++ IcuTestErrorCode errorCode(*this, "TestKeywordValues");
++ LocalPointer<Collator> col(Collator::createInstance(Locale::getEnglish(), errorCode));
++ if (errorCode.logIfFailureAndReset("English Collator creation failed")) {
++ return;
++ }
++
++ LocalPointer<StringEnumeration> kwEnum(
++ col->getKeywordValuesForLocale("collation", Locale::getEnglish(), TRUE, errorCode));
++ if (errorCode.logIfFailureAndReset("Get Keyword Values for English Collator failed")) {
++ return;
++ }
++ assertTrue("expect at least one collation tailoring for English", kwEnum->count(errorCode) > 0);
++ const char *kw;
++ UBool hasStandard = FALSE;
++ while ((kw = kwEnum->next(NULL, errorCode)) != NULL) {
++ if (strcmp(kw, "standard") == 0) {
++ hasStandard = TRUE;
++ }
++ }
++ assertTrue("expect at least the 'standard' collation tailoring for English", hasStandard);
+ }
+
+@@ -2467,4 +2483,5 @@
+ TESTCASE_AUTO_BEGIN;
+ TESTCASE_AUTO(TestProperty);
++ TESTCASE_AUTO(TestKeywordValues);
+ TESTCASE_AUTO(TestOperators);
+ TESTCASE_AUTO(TestDuplicate);
+Index: /icu/trunk/source/test/intltest/apicoll.h
+===================================================================
+--- /icu/trunk/source/test/intltest/apicoll.h (revision 39483)
++++ /icu/trunk/source/test/intltest/apicoll.h (revision 39484)
+@@ -36,4 +36,5 @@
+ */
+ void TestProperty(/* char* par */);
++ void TestKeywordValues();
+
+ /**
Copied: icu/repos/testing-i686/icu.8198.revert.icu5431.patch (from rev 293902, icu/trunk/icu.8198.revert.icu5431.patch)
===================================================================
--- testing-i686/icu.8198.revert.icu5431.patch (rev 0)
+++ testing-i686/icu.8198.revert.icu5431.patch 2017-04-22 11:59:19 UTC (rev 293903)
@@ -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 293902, icu/trunk/PKGBUILD)
===================================================================
--- testing-x86_64/PKGBUILD (rev 0)
+++ testing-x86_64/PKGBUILD 2017-04-22 11:59:19 UTC (rev 293903)
@@ -0,0 +1,55 @@
+# $Id$
+# Maintainer: Andreas Radke <andyrtr at archlinux.org>
+# Contributor: Art Gramlich <art at gramlich-net.com>
+
+pkgname=icu
+pkgver=58.2
+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')
+#makedepends=('clang')
+# no https available
+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-58.1-iterator-reset.patch
+ changeset-39671-CVE-2017-7867+7868.patch)
+# upstream offers md5sum checks, only asc file for md5sum check
+md5sums=('fac212b32b7ec7ab007a12dff1f3aea1'
+ '8c09ae284967def053e9579d64d0f83c'
+ 'cea722333992ad6ede88ebd0bd21cfcd')
+sha256sums=('2b0a4410153a9b20de0e20c7d8b66049a72aef244b53683d0d7521371683da0c'
+ '8034928bcff89eca84d8d0f22fa9b6bdba5e0608a984f64fe7afe41b91bbea97'
+ 'cf0c6f946d7336bd46857c52308fd67171918f24fcf19015b5a47a8d2fb8ca0c')
+
+prepare() {
+ cd icu/source
+ # http://bugs.icu-project.org/trac/ticket/12827
+ patch -Np4 -i "${srcdir}"/icu-58.1-iterator-reset.patch
+
+ patch -Np4 -i "${srcdir}"/changeset-39671-CVE-2017-7867+7868.patch
+}
+
+build() {
+ cd icu/source
+ ./configure --prefix=/usr \
+ --sysconfdir=/etc \
+ --mandir=/usr/share/man \
+ --sbindir=/usr/bin
+ make
+}
+
+check() {
+ cd icu/source
+ make -k check # passes all
+}
+
+package() {
+ cd 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-39671-CVE-2017-7867+7868.patch (from rev 293902, icu/trunk/changeset-39671-CVE-2017-7867+7868.patch)
===================================================================
--- testing-x86_64/changeset-39671-CVE-2017-7867+7868.patch (rev 0)
+++ testing-x86_64/changeset-39671-CVE-2017-7867+7868.patch 2017-04-22 11:59:19 UTC (rev 293903)
@@ -0,0 +1,155 @@
+Index: /trunk/icu4c/source/common/utext.cpp
+===================================================================
+--- /trunk/icu4c/source/common/utext.cpp (revision 39670)
++++ /trunk/icu4c/source/common/utext.cpp (revision 39671)
+@@ -848,7 +848,13 @@
+
+ // Chunk size.
+-// Must be less than 85, because of byte mapping from UChar indexes to native indexes.
+-// Worst case is three native bytes to one UChar. (Supplemenaries are 4 native bytes
+-// to two UChars.)
++// Must be less than 42 (256/6), because of byte mapping from UChar indexes to native indexes.
++// Worst case there are six UTF-8 bytes per UChar.
++// obsolete 6 byte form fd + 5 trails maps to fffd
++// obsolete 5 byte form fc + 4 trails maps to fffd
++// non-shortest 4 byte forms maps to fffd
++// normal supplementaries map to a pair of utf-16, two utf8 bytes per utf-16 unit
++// mapToUChars array size must allow for the worst case, 6.
++// This could be brought down to 4, by treating fd and fc as pure illegal,
++// rather than obsolete lead bytes. But that is not compatible with the utf-8 access macros.
+ //
+ enum { UTF8_TEXT_CHUNK_SIZE=32 };
+@@ -890,5 +896,5 @@
+ // one for a supplementary starting in the last normal position,
+ // and one for an entry for the buffer limit position.
+- uint8_t mapToUChars[UTF8_TEXT_CHUNK_SIZE*3+6]; // Map native offset from bufNativeStart to
++ uint8_t mapToUChars[UTF8_TEXT_CHUNK_SIZE*6+6]; // Map native offset from bufNativeStart to
+ // correspoding offset in filled part of buf.
+ int32_t align;
+@@ -1033,4 +1039,5 @@
+ u8b = (UTF8Buf *)ut->p; // the current buffer
+ mapIndex = ix - u8b->toUCharsMapStart;
++ U_ASSERT(mapIndex < (int32_t)sizeof(UTF8Buf::mapToUChars));
+ ut->chunkOffset = u8b->mapToUChars[mapIndex] - u8b->bufStartIdx;
+ return TRUE;
+@@ -1299,4 +1306,8 @@
+ // If index is at the end, there is no character there to look at.
+ if (ix != ut->b) {
++ // Note: this function will only move the index back if it is on a trail byte
++ // and there is a preceding lead byte and the sequence from the lead
++ // through this trail could be part of a valid UTF-8 sequence
++ // Otherwise the index remains unchanged.
+ U8_SET_CP_START(s8, 0, ix);
+ }
+@@ -1312,5 +1323,8 @@
+ uint8_t *mapToNative = u8b->mapToNative;
+ uint8_t *mapToUChars = u8b->mapToUChars;
+- int32_t toUCharsMapStart = ix - (UTF8_TEXT_CHUNK_SIZE*3 + 1);
++ int32_t toUCharsMapStart = ix - sizeof(UTF8Buf::mapToUChars) + 1;
++ // Note that toUCharsMapStart can be negative. Happens when the remaining
++ // text from current position to the beginning is less than the buffer size.
++ // + 1 because mapToUChars must have a slot at the end for the bufNativeLimit entry.
+ int32_t destIx = UTF8_TEXT_CHUNK_SIZE+2; // Start in the overflow region
+ // at end of buffer to leave room
+@@ -1339,4 +1353,5 @@
+ // Special case ASCII range for speed.
+ buf[destIx] = (UChar)c;
++ U_ASSERT(toUCharsMapStart <= srcIx);
+ mapToUChars[srcIx - toUCharsMapStart] = (uint8_t)destIx;
+ mapToNative[destIx] = (uint8_t)(srcIx - toUCharsMapStart);
+@@ -1368,4 +1383,5 @@
+ mapToUChars[sIx-- - toUCharsMapStart] = (uint8_t)destIx;
+ } while (sIx >= srcIx);
++ U_ASSERT(toUCharsMapStart <= (srcIx+1));
+
+ // Set native indexing limit to be the current position.
+@@ -1542,4 +1558,5 @@
+ U_ASSERT(index<=ut->chunkNativeLimit);
+ int32_t mapIndex = index - u8b->toUCharsMapStart;
++ U_ASSERT(mapIndex < (int32_t)sizeof(UTF8Buf::mapToUChars));
+ int32_t offset = u8b->mapToUChars[mapIndex] - u8b->bufStartIdx;
+ U_ASSERT(offset>=0 && offset<=ut->chunkLength);
+Index: /trunk/icu4c/source/test/intltest/utxttest.cpp
+===================================================================
+--- /trunk/icu4c/source/test/intltest/utxttest.cpp (revision 39670)
++++ /trunk/icu4c/source/test/intltest/utxttest.cpp (revision 39671)
+@@ -68,4 +68,6 @@
+ case 7: name = "Ticket12130";
+ if (exec) Ticket12130(); break;
++ case 8: name = "Ticket12888";
++ if (exec) Ticket12888(); break;
+ default: name = ""; break;
+ }
+@@ -1584,2 +1586,62 @@
+ utext_close(&ut);
+ }
++
++// Ticket 12888: bad handling of illegal utf-8 containing many instances of the archaic, now illegal,
++// six byte utf-8 forms. Original implementation had an assumption that
++// there would be at most three utf-8 bytes per UTF-16 code unit.
++// The five and six byte sequences map to a single replacement character.
++
++void UTextTest::Ticket12888() {
++ const char *badString =
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80"
++ "\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80\xfd\x80\x80\x80\x80\x80";
++
++ UErrorCode status = U_ZERO_ERROR;
++ LocalUTextPointer ut(utext_openUTF8(NULL, badString, -1, &status));
++ TEST_SUCCESS(status);
++ for (;;) {
++ UChar32 c = utext_next32(ut.getAlias());
++ if (c == U_SENTINEL) {
++ break;
++ }
++ }
++ int32_t endIdx = utext_getNativeIndex(ut.getAlias());
++ if (endIdx != (int32_t)strlen(badString)) {
++ errln("%s:%d expected=%d, actual=%d", __FILE__, __LINE__, strlen(badString), endIdx);
++ return;
++ }
++
++ for (int32_t prevIndex = endIdx; prevIndex>0;) {
++ UChar32 c = utext_previous32(ut.getAlias());
++ int32_t currentIndex = utext_getNativeIndex(ut.getAlias());
++ if (c != 0xfffd) {
++ errln("%s:%d (expected, actual, index) = (%d, %d, %d)\n",
++ __FILE__, __LINE__, 0xfffd, c, currentIndex);
++ break;
++ }
++ if (currentIndex != prevIndex - 6) {
++ errln("%s:%d: wrong index. Expected, actual = %d, %d",
++ __FILE__, __LINE__, prevIndex - 6, currentIndex);
++ break;
++ }
++ prevIndex = currentIndex;
++ }
++}
+Index: /trunk/icu4c/source/test/intltest/utxttest.h
+===================================================================
+--- /trunk/icu4c/source/test/intltest/utxttest.h (revision 39670)
++++ /trunk/icu4c/source/test/intltest/utxttest.h (revision 39671)
+@@ -39,4 +39,5 @@
+ void Ticket10983();
+ void Ticket12130();
++ void Ticket12888();
+
+ private:
Copied: icu/repos/testing-x86_64/icu-58.1-iterator-reset.patch (from rev 293902, icu/trunk/icu-58.1-iterator-reset.patch)
===================================================================
--- testing-x86_64/icu-58.1-iterator-reset.patch (rev 0)
+++ testing-x86_64/icu-58.1-iterator-reset.patch 2017-04-22 11:59:19 UTC (rev 293903)
@@ -0,0 +1,124 @@
+Index: /icu/trunk/source/common/ulist.c
+===================================================================
+--- /icu/trunk/source/common/ulist.c (revision 39483)
++++ /icu/trunk/source/common/ulist.c (revision 39484)
+@@ -30,5 +30,4 @@
+
+ int32_t size;
+- int32_t currentIndex;
+ };
+
+@@ -52,5 +51,4 @@
+ newList->tail = NULL;
+ newList->size = 0;
+- newList->currentIndex = -1;
+
+ return newList;
+@@ -81,6 +79,7 @@
+ p->next->previous = p->previous;
+ }
+- list->curr = NULL;
+- list->currentIndex = 0;
++ if (p == list->curr) {
++ list->curr = p->next;
++ }
+ --list->size;
+ if (p->forceDelete) {
+@@ -151,5 +150,4 @@
+ list->head->previous = newItem;
+ list->head = newItem;
+- list->currentIndex++;
+ }
+
+@@ -194,5 +192,4 @@
+ curr = list->curr;
+ list->curr = curr->next;
+- list->currentIndex++;
+
+ return curr->data;
+@@ -210,5 +207,4 @@
+ if (list != NULL) {
+ list->curr = list->head;
+- list->currentIndex = 0;
+ }
+ }
+@@ -273,3 +269,2 @@
+ return (UList *)(en->context);
+ }
+-
+Index: /icu/trunk/source/i18n/ucol_res.cpp
+===================================================================
+--- /icu/trunk/source/i18n/ucol_res.cpp (revision 39483)
++++ /icu/trunk/source/i18n/ucol_res.cpp (revision 39484)
+@@ -681,4 +681,5 @@
+ }
+ memcpy(en, &defaultKeywordValues, sizeof(UEnumeration));
++ ulist_resetList(sink.values); // Initialize the iterator.
+ en->context = sink.values;
+ sink.values = NULL; // Avoid deletion in the sink destructor.
+Index: /icu/trunk/source/test/intltest/apicoll.cpp
+===================================================================
+--- /icu/trunk/source/test/intltest/apicoll.cpp (revision 39483)
++++ /icu/trunk/source/test/intltest/apicoll.cpp (revision 39484)
+@@ -82,14 +82,7 @@
+ col = Collator::createInstance(Locale::getEnglish(), success);
+ if (U_FAILURE(success)){
+- errcheckln(success, "Default Collator creation failed. - %s", u_errorName(success));
+- return;
+- }
+-
+- StringEnumeration* kwEnum = col->getKeywordValuesForLocale("", Locale::getEnglish(),true,success);
+- if (U_FAILURE(success)){
+- errcheckln(success, "Get Keyword Values for Locale failed. - %s", u_errorName(success));
+- return;
+- }
+- delete kwEnum;
++ errcheckln(success, "English Collator creation failed. - %s", u_errorName(success));
++ return;
++ }
+
+ col->getVersion(versionArray);
+@@ -230,4 +223,27 @@
+ delete aFrCol;
+ delete junk;
++}
++
++void CollationAPITest::TestKeywordValues() {
++ IcuTestErrorCode errorCode(*this, "TestKeywordValues");
++ LocalPointer<Collator> col(Collator::createInstance(Locale::getEnglish(), errorCode));
++ if (errorCode.logIfFailureAndReset("English Collator creation failed")) {
++ return;
++ }
++
++ LocalPointer<StringEnumeration> kwEnum(
++ col->getKeywordValuesForLocale("collation", Locale::getEnglish(), TRUE, errorCode));
++ if (errorCode.logIfFailureAndReset("Get Keyword Values for English Collator failed")) {
++ return;
++ }
++ assertTrue("expect at least one collation tailoring for English", kwEnum->count(errorCode) > 0);
++ const char *kw;
++ UBool hasStandard = FALSE;
++ while ((kw = kwEnum->next(NULL, errorCode)) != NULL) {
++ if (strcmp(kw, "standard") == 0) {
++ hasStandard = TRUE;
++ }
++ }
++ assertTrue("expect at least the 'standard' collation tailoring for English", hasStandard);
+ }
+
+@@ -2467,4 +2483,5 @@
+ TESTCASE_AUTO_BEGIN;
+ TESTCASE_AUTO(TestProperty);
++ TESTCASE_AUTO(TestKeywordValues);
+ TESTCASE_AUTO(TestOperators);
+ TESTCASE_AUTO(TestDuplicate);
+Index: /icu/trunk/source/test/intltest/apicoll.h
+===================================================================
+--- /icu/trunk/source/test/intltest/apicoll.h (revision 39483)
++++ /icu/trunk/source/test/intltest/apicoll.h (revision 39484)
+@@ -36,4 +36,5 @@
+ */
+ void TestProperty(/* char* par */);
++ void TestKeywordValues();
+
+ /**
Copied: icu/repos/testing-x86_64/icu.8198.revert.icu5431.patch (from rev 293902, icu/trunk/icu.8198.revert.icu5431.patch)
===================================================================
--- testing-x86_64/icu.8198.revert.icu5431.patch (rev 0)
+++ testing-x86_64/icu.8198.revert.icu5431.patch 2017-04-22 11:59:19 UTC (rev 293903)
@@ -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