AuroraMiddleware/scuffed-code
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

ICU-2908 fix crash on regexp patterns with quantifiers on a surrogate pair e.g. \ud800\udc00*

Browse Source 
X-SVN-Rev: 12065
This commit is contained in:
Andy Heninger 2003-05-23 01:32:25 +00:00
parent e731ba6034
commit 9c02edfe2d
2 changed files with 32 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

25
icu4c/source/i18n/regexcmp.cpp
View File

@ -1005,7 +1005,7 @@ UBool RegexCompile::doParseActions(EParseAction action)
}
break;
case doPossesiveInterval:
case doPossessiveInterval:
// Finished scanning a Possessive {lower,upper}+ interval. Generate the code for it.
{
// Remember the loc for the top of the block being looped over.
@ -1215,7 +1215,7 @@ UBool RegexCompile::doParseActions(EParseAction action)
case doPossesivePlus:
case doPossessivePlus:
// Possessive ++ quantifier.
// Compiles to
// 1. STO_SP
@ -1250,7 +1250,7 @@ UBool RegexCompile::doParseActions(EParseAction action)
}
break;
case doPossesiveStar:
case doPossessiveStar:
// Possessive *+ quantifier.
// Compiles to
// 1. STO_SP loc
@ -1286,7 +1286,7 @@ UBool RegexCompile::doParseActions(EParseAction action)
}
break;
case doPossesiveOpt:
case doPossessiveOpt:
// Possessive ?+ quantifier.
// Compiles to
// 1. STO_SP loc
@ -1445,11 +1445,26 @@ void RegexCompile::literalChar(UChar32 c) {
// We are adding onto an existing string
fRXPat->fLiteralText.append(c);
// If the most recently emitted op is a URX_ONECHAR, change it to a string op.
op = fRXPat->fCompiledPat->lastElementi();
opType = URX_TYPE(op);
U_ASSERT(opType == URX_ONECHAR || opType == URX_ONECHAR_I || opType == URX_STRING_LEN);
// If the most recently emitted op is a URX_ONECHAR,
if (opType == URX_ONECHAR || opType == URX_ONECHAR_I) {
if (U16_IS_TRAIL(c) && U16_IS_LEAD(URX_VAL(op))) {
// The most recently emitted op is a ONECHAR that was the first half
// of a surrogate pair. Update the ONECHAR's operand to be the
// supplementary code point resulting from both halves of the pair.
c = U16_GET_SUPPLEMENTARY(URX_VAL(op), c);
op = URX_BUILD(opType, c);
patternLoc = fRXPat->fCompiledPat->size() - 1;
fRXPat->fCompiledPat->setElementAt(op, patternLoc);
return;
}
// The most recently emitted op is a ONECHAR.
// We've now received another adjacent char. Change the ONECHAR op
// to a string op.
if (fModeFlags & UREGEX_CASE_INSENSITIVE) {
op = URX_BUILD(URX_STRING_I, fStringOpStart);
} else {

11
icu4c/source/test/testdata/regextst.txt vendored
View File

@ -338,6 +338,17 @@
"ab\x{101234}c" "<0>ab\U00101234c</0>"
"abα" "<0>abα</0>"
#
# \u Surrogate Pairs
#
"\ud800\udc00" "<0>\U00010000</0>"
"\ud800\udc00*" "<0>\U00010000\U00010000\U00010000</0>\U00010001"
"\ud800\ud800\udc00" "<0>\ud800\U00010000</0>\U00010000\U00010000\U00010001"
"(\ud800)(\udc00)" "\u00010000"
#
# Random debugging, Temporary
#