cclass_Unicode::StrChr needs to support non-BMP chars

...as seen with Clang's new -fsanitize=implicit-cast during
CppunitTest_starmath_qa_cppunit:

> i18npool/source/characterclassification/cclass_unicode_parser.cxx:565:46: runtime error: implicit conversion from type 'sal_uInt32' (aka 'unsigned int') of value 119886 (32-bit, unsigned) to type 'sal_Unicode' (aka 'char16_t') changed the value to 54350 (16-bit, unsigned)
>  #0 in i18npool::cclass_Unicode::getFlags(unsigned int) at i18npool/source/characterclassification/cclass_unicode_parser.cxx:565:46 (instdir/program/libi18npoollo.so +0x3ae807)
>  #1 in i18npool::cclass_Unicode::parseText(com::sun::star::i18n::ParseResult&, rtl::OUString const&, int, int) at i18npool/source/characterclassification/cclass_unicode_parser.cxx:712:29 (instdir/program/libi18npoollo.so +0x3b04c3)
>  #2 in i18npool::cclass_Unicode::parsePredefinedToken(int, rtl::OUString const&, int, com::sun::star::lang::Locale const&, int, rtl::OUString const&, int, rtl::OUString const&) at i18npool/source/characterclassification/cclass_unicode.cxx:275:5 (instdir/program/libi18npoollo.so +0x3a17ea)
>  #3 in non-virtual thunk to i18npool::cclass_Unicode::parsePredefinedToken(int, rtl::OUString const&, int, com::sun::star::lang::Locale const&, int, rtl::OUString const&, int, rtl::OUString const&) at i18npool/source/characterclassification/cclass_unicode.cxx (instdir/program/libi18npoollo.so +0x3a18dc)
>  #4 in i18npool::CharacterClassificationImpl::parsePredefinedToken(int, rtl::OUString const&, int, com::sun::star::lang::Locale const&, int, rtl::OUString const&, int, rtl::OUString const&) at i18npool/source/characterclassification/characterclassificationImpl.cxx:118:63 (instdir/program/libi18npoollo.so +0x3c48ba)
>  #5 in non-virtual thunk to i18npool::CharacterClassificationImpl::parsePredefinedToken(int, rtl::OUString const&, int, com::sun::star::lang::Locale const&, int, rtl::OUString const&, int, rtl::OUString const&) at i18npool/source/characterclassification/characterclassificationImpl.cxx (instdir/program/libi18npoollo.so +0x3c497c)
>  #6 in CharClass::parsePredefinedToken(int, rtl::OUString const&, int, int, rtl::OUString const&, int, rtl::OUString const&) const at unotools/source/i18n/charclass.cxx:443:25 (instdir/program/libutllo.so +0x904d17)
>  #7 in SmParser::NextToken() at starmath/source/parse.cxx:391:25 (workdir/LinkTarget/CppunitTest/libtest_starmath_qa_cppunit.so +0xa4a3e7)
>  #8 in SmParser::DoTerm(bool) at starmath/source/parse.cxx:1337:13 (workdir/LinkTarget/CppunitTest/libtest_starmath_qa_cppunit.so +0xa5951b)
>  #9 in SmParser::DoPower() at starmath/source/parse.cxx:1285:35 (workdir/LinkTarget/CppunitTest/libtest_starmath_qa_cppunit.so +0xa57d46)
>  #10 in SmParser::DoProduct() at starmath/source/parse.cxx:1105:19 (workdir/LinkTarget/CppunitTest/libtest_starmath_qa_cppunit.so +0xa5685a)
>  #11 in SmParser::DoSum() at starmath/source/parse.cxx:1087:19 (workdir/LinkTarget/CppunitTest/libtest_starmath_qa_cppunit.so +0xa55ebc)
>  #12 in SmParser::DoRelation() at starmath/source/parse.cxx:1069:19 (workdir/LinkTarget/CppunitTest/libtest_starmath_qa_cppunit.so +0xa558dc)
>  #13 in SmParser::DoExpression(bool) at starmath/source/parse.cxx:1043:29 (workdir/LinkTarget/CppunitTest/libtest_starmath_qa_cppunit.so +0xa54ff5)
>  #14 in SmParser::ParseExpression(rtl::OUString const&) at starmath/source/parse.cxx:2366:12 (workdir/LinkTarget/CppunitTest/libtest_starmath_qa_cppunit.so +0xa671dc)
>  #15 in (anonymous namespace)::Test::ParseAndCompare(char const*, char const*, char const*) at starmath/qa/cppunit/test_nodetotextvisitors.cxx:485:30 (workdir/LinkTarget/CppunitTest/libtest_starmath_qa_cppunit.so +0x5f7ea6)
>  #16 in (anonymous namespace)::Test::testMiscEquivalent() at starmath/qa/cppunit/test_nodetotextvisitors.cxx:637:5 (workdir/LinkTarget/CppunitTest/libtest_starmath_qa_cppunit.so +0x5f2dc8)

Change-Id: Iaf62efd60bd6132e005ab69ce385bbf5c2db5d19
Reviewed-on: https://gerrit.libreoffice.org/58979
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>

1 files changed, 6 insertions, 3 deletions

diff --git a/i18npool/source/characterclassification/cclass_unicode_parser.cxx b/i18npool/source/characterclassification/cclass_unicode_parser.cxx
index 8192d7ddab16..96882a8ef3c2 100644
--- a/i18npool/source/characterclassification/cclass_unicode_parser.cxx
+++ b/i18npool/source/characterclassification/cclass_unicode_parser.cxx
@@ -20,6 +20,7 @@
 
 #include <cclass_unicode.hxx>
 #include <unicode/uchar.h>
+#include <rtl/character.hxx>
 #include <rtl/math.hxx>
 #include <rtl/ustring.hxx>
 #include <com/sun/star/i18n/KParseTokens.hpp>
@@ -317,13 +318,15 @@ const sal_Int32 cclass_Unicode::pParseTokensType[ nDefCnt ] =
 
 
 // static
-const sal_Unicode* cclass_Unicode::StrChr( const sal_Unicode* pStr, sal_Unicode c )
+const sal_Unicode* cclass_Unicode::StrChr( const sal_Unicode* pStr, sal_uInt32 c )
 {
     if ( !pStr )
         return nullptr;
+    sal_Unicode cs[2];
+    auto const n = rtl::splitSurrogates(c, cs);
     while ( *pStr )
     {
-        if ( *pStr == c )
+        if ( *pStr == cs[0] && (n == 1 || pStr[1] == cs[1]) )
             return pStr;
         pStr++;
     }
@@ -659,7 +662,7 @@ ParserFlags cclass_Unicode::getFlagsExtended(sal_uInt32 const c)
 }
 
 
-ParserFlags cclass_Unicode::getStartCharsFlags( sal_Unicode c )
+ParserFlags cclass_Unicode::getStartCharsFlags( sal_uInt32 c )
 {
     if ( pStart )
     {