Cherry-pick a regex compile patch to M49 branch
Set fLastSetLiteral to U_SENTINEL in the ctor and check if it's still
U_SENTINEL before adding the range.
https://codereview.chromium.org/1656003004/ is the master CL being
cherry-picked.
BUG=582470
TEST=See comments 13 and 15 in the bug
TBR=mmoroz
Review URL: https://codereview.chromium.org/1654213004 .
diff --git a/README.chromium b/README.chromium
index dae5e55..24b5118 100644
--- a/README.chromium
+++ b/README.chromium
@@ -327,3 +327,10 @@
- patches/utext.patch
- upstream bug (fixed in trunk in Jan, 2016. Will be in 57 release)
http://bugs.icu-project.org/trac/ticket/12130
+
+25. Fix a bug in regex compiler.
+ - patches/regexcmp.patch
+ - upstream bug
+ http://bugs.icu-project.org/trac/ticket/12138
+
+
diff --git a/patches/regexcmp.patch b/patches/regexcmp.patch
new file mode 100644
index 0000000..9aeabb4
--- /dev/null
+++ b/patches/regexcmp.patch
@@ -0,0 +1,43 @@
+diff --git a/source/i18n/regexcmp.cpp b/source/i18n/regexcmp.cpp
+index ea01d5a..29b789b 100644
+--- a/source/i18n/regexcmp.cpp
++++ b/source/i18n/regexcmp.cpp
+@@ -70,6 +70,7 @@ RegexCompile::RegexCompile(RegexPattern *rxp, UErrorCode &status) :
+
+ fMatchOpenParen = -1;
+ fMatchCloseParen = -1;
++ fLastSetLiteral = U_SENTINEL;
+
+ if (U_SUCCESS(status) && U_FAILURE(rxp->fDeferredStatus)) {
+ status = rxp->fDeferredStatus;
+@@ -1625,12 +1626,14 @@ UBool RegexCompile::doParseActions(int32_t action)
+ case doSetNamedRange:
+ // We have scanned literal-\N{CHAR NAME}. Add the range to the set.
+ // The left character is already in the set, and is saved in fLastSetLiteral.
++ // Nonetheless, check if |fLastSetLiteral| is indeed set because it's
++ // not set in some edge cases.
+ // The right side needs to be picked up, the scan is at the 'N'.
+ // Lower Limit > Upper limit being an error matches both Java
+ // and ICU UnicodeSet behavior.
+ {
+ UChar32 c = scanNamedChar();
+- if (U_SUCCESS(*fStatus) && fLastSetLiteral > c) {
++ if (U_SUCCESS(*fStatus) && (fLastSetLiteral == U_SENTINEL || fLastSetLiteral > c)) {
+ error(U_REGEX_INVALID_RANGE);
+ }
+ UnicodeSet *s = (UnicodeSet *)fSetStack.peek();
+@@ -1695,11 +1698,13 @@ UBool RegexCompile::doParseActions(int32_t action)
+ case doSetRange:
+ // We have scanned literal-literal. Add the range to the set.
+ // The left character is already in the set, and is saved in fLastSetLiteral.
++ // Nonetheless, check if |fLastSetLiteral| is indeed set because it's
++ // not set in some edge cases.
+ // The right side is the current character.
+ // Lower Limit > Upper limit being an error matches both Java
+ // and ICU UnicodeSet behavior.
+ {
+- if (fLastSetLiteral > fC.fChar) {
++ if (fLastSetLiteral == U_SENTINEL || fLastSetLiteral > fC.fChar) {
+ error(U_REGEX_INVALID_RANGE);
+ }
+ UnicodeSet *s = (UnicodeSet *)fSetStack.peek();
diff --git a/source/i18n/regexcmp.cpp b/source/i18n/regexcmp.cpp
index ea01d5a..29b789b 100644
--- a/source/i18n/regexcmp.cpp
+++ b/source/i18n/regexcmp.cpp
@@ -70,6 +70,7 @@
fMatchOpenParen = -1;
fMatchCloseParen = -1;
+ fLastSetLiteral = U_SENTINEL;
if (U_SUCCESS(status) && U_FAILURE(rxp->fDeferredStatus)) {
status = rxp->fDeferredStatus;
@@ -1625,12 +1626,14 @@
case doSetNamedRange:
// We have scanned literal-\N{CHAR NAME}. Add the range to the set.
// The left character is already in the set, and is saved in fLastSetLiteral.
+ // Nonetheless, check if |fLastSetLiteral| is indeed set because it's
+ // not set in some edge cases.
// The right side needs to be picked up, the scan is at the 'N'.
// Lower Limit > Upper limit being an error matches both Java
// and ICU UnicodeSet behavior.
{
UChar32 c = scanNamedChar();
- if (U_SUCCESS(*fStatus) && fLastSetLiteral > c) {
+ if (U_SUCCESS(*fStatus) && (fLastSetLiteral == U_SENTINEL || fLastSetLiteral > c)) {
error(U_REGEX_INVALID_RANGE);
}
UnicodeSet *s = (UnicodeSet *)fSetStack.peek();
@@ -1695,11 +1698,13 @@
case doSetRange:
// We have scanned literal-literal. Add the range to the set.
// The left character is already in the set, and is saved in fLastSetLiteral.
+ // Nonetheless, check if |fLastSetLiteral| is indeed set because it's
+ // not set in some edge cases.
// The right side is the current character.
// Lower Limit > Upper limit being an error matches both Java
// and ICU UnicodeSet behavior.
{
- if (fLastSetLiteral > fC.fChar) {
+ if (fLastSetLiteral == U_SENTINEL || fLastSetLiteral > fC.fChar) {
error(U_REGEX_INVALID_RANGE);
}
UnicodeSet *s = (UnicodeSet *)fSetStack.peek();
