Diff
Modified: trunk/Source/_javascript_Core/ChangeLog (112563 => 112564)
--- trunk/Source/_javascript_Core/ChangeLog 2012-03-29 19:48:21 UTC (rev 112563)
+++ trunk/Source/_javascript_Core/ChangeLog 2012-03-29 20:16:03 UTC (rev 112564)
@@ -1,3 +1,31 @@
+2012-03-29 Gavin Barraclough <[email protected]>
+
+ Template the Yarr::Interpreter on the character type
+ https://bugs.webkit.org/show_bug.cgi?id=82637
+
+ Reviewed by Sam Weinig.
+
+ We should be able to call to the interpreter after having already checked the character type,
+ without having to re-package the character pointer back up into a string!
+
+ * runtime/RegExp.cpp:
+ (JSC::RegExp::match):
+ (JSC::RegExp::matchCompareWithInterpreter):
+ - Don't pass length.
+ * yarr/Yarr.h:
+ - moved function declarations to YarrInterpreter.h.
+ * yarr/YarrInterpreter.cpp:
+ (Yarr):
+ (Interpreter):
+ (JSC::Yarr::Interpreter::InputStream::InputStream):
+ (InputStream):
+ (JSC::Yarr::Interpreter::Interpreter):
+ (JSC::Yarr::interpret):
+ - templated Interpreter class on CharType.
+ * yarr/YarrInterpreter.h:
+ (Yarr):
+ - added function declarations.
+
2012-03-29 David Kilzer <[email protected]>
Don't use a flattened framework path when building on OS X
Modified: trunk/Source/_javascript_Core/runtime/RegExp.cpp (112563 => 112564)
--- trunk/Source/_javascript_Core/runtime/RegExp.cpp 2012-03-29 19:48:21 UTC (rev 112563)
+++ trunk/Source/_javascript_Core/runtime/RegExp.cpp 2012-03-29 20:16:03 UTC (rev 112564)
@@ -343,7 +343,7 @@
#endif
} else
#endif
- result = Yarr::interpret(m_regExpBytecode.get(), s, startOffset, s.length(), reinterpret_cast<unsigned*>(offsetVector));
+ result = Yarr::interpret(m_regExpBytecode.get(), s, startOffset, reinterpret_cast<unsigned*>(offsetVector));
// FIXME: The YARR engine should handle unsigned or size_t length matches.
// The YARR Interpreter is "unsigned" clean, while the YARR JIT hasn't been addressed.
@@ -467,7 +467,7 @@
Vector<int, 32> nonReturnedOvector;
nonReturnedOvector.resize(offsetVectorSize);
offsetVector = nonReturnedOvector.data();
- int r = Yarr::interpret(m_regExpBytecode.get(), s, startOffset, s.length(), reinterpret_cast<unsigned*>(offsetVector));
+ int r = Yarr::interpret(m_regExpBytecode.get(), s, startOffset, reinterpret_cast<unsigned*>(offsetVector));
#if REGEXP_FUNC_TEST_DATA_GEN
RegExpFunctionalTestCollector::get()->outputOneTest(this, s, startOffset, offsetVector, result);
#endif
@@ -509,7 +509,7 @@
for (unsigned j = 0, i = 0; i < m_numSubpatterns + 1; j += 2, i++)
interpreterOffsetVector[j] = -1;
- interpreterResult = Yarr::interpret(m_regExpBytecode.get(), s, startOffset, s.length(), interpreterOffsetVector);
+ interpreterResult = Yarr::interpret(m_regExpBytecode.get(), s, startOffset, interpreterOffsetVector);
if (jitResult != interpreterResult)
differences++;
Modified: trunk/Source/_javascript_Core/yarr/Yarr.h (112563 => 112564)
--- trunk/Source/_javascript_Core/yarr/Yarr.h 2012-03-29 19:48:21 UTC (rev 112563)
+++ trunk/Source/_javascript_Core/yarr/Yarr.h 2012-03-29 20:16:03 UTC (rev 112564)
@@ -63,9 +63,6 @@
Char16
};
-JS_EXPORT_PRIVATE PassOwnPtr<BytecodePattern> byteCompile(YarrPattern&, BumpPointerAllocator*);
-JS_EXPORT_PRIVATE unsigned interpret(BytecodePattern*, const UString& input, unsigned start, unsigned length, unsigned* output);
-
} } // namespace JSC::Yarr
#endif // Yarr_h
Modified: trunk/Source/_javascript_Core/yarr/YarrInterpreter.cpp (112563 => 112564)
--- trunk/Source/_javascript_Core/yarr/YarrInterpreter.cpp 2012-03-29 19:48:21 UTC (rev 112563)
+++ trunk/Source/_javascript_Core/yarr/YarrInterpreter.cpp 2012-03-29 20:16:03 UTC (rev 112564)
@@ -42,6 +42,7 @@
namespace JSC { namespace Yarr {
+template<typename CharType>
class Interpreter {
public:
struct ParenthesesDisjunctionContext;
@@ -170,55 +171,9 @@
allocatorPool = allocatorPool->dealloc(context);
}
- // This class is a placeholder for future character iterator, current
- // proposed name StringConstCharacterIterator.
- class CharAccess {
- public:
- CharAccess(const UString& s)
- {
- if (s.is8Bit()) {
- m_charSize = Char8;
- m_ptr.ptr8 = s.characters8();
- } else {
- m_charSize = Char16;
- m_ptr.ptr16 = s.characters16();
- }
- }
-
- CharAccess(const LChar* ptr)
- : m_charSize(Char8)
- {
- m_ptr.ptr8 = ptr;
- }
-
- CharAccess(const UChar* ptr)
- : m_charSize(Char16)
- {
- m_ptr.ptr16 = ptr;
- }
-
- ~CharAccess()
- {
- }
-
- inline UChar operator[](unsigned index)
- {
- if (m_charSize == Char8)
- return m_ptr.ptr8[index];
- return m_ptr.ptr16[index];
- }
-
- private:
- union {
- const LChar* ptr8;
- const UChar* ptr16;
- } m_ptr;
- YarrCharSize m_charSize;
- };
-
class InputStream {
public:
- InputStream(const UString& input, unsigned start, unsigned length)
+ InputStream(const CharType* input, unsigned start, unsigned length)
: input(input)
, pos(start)
, length(length)
@@ -332,7 +287,7 @@
}
private:
- CharAccess input;
+ const CharType* input;
unsigned pos;
unsigned length;
};
@@ -1489,7 +1444,7 @@
return output[0];
}
- Interpreter(BytecodePattern* pattern, unsigned* output, const UString input, unsigned start, unsigned length)
+ Interpreter(BytecodePattern* pattern, unsigned* output, const CharType* input, unsigned length, unsigned start)
: pattern(pattern)
, output(output)
, input(input, start, length)
@@ -1979,18 +1934,31 @@
return ByteCompiler(pattern).compile(allocator);
}
-unsigned interpret(BytecodePattern* bytecode, const UString& input, unsigned start, unsigned length, unsigned* output)
+unsigned interpret(BytecodePattern* bytecode, const UString& input, unsigned start, unsigned* output)
{
- return Interpreter(bytecode, output, input, start, length).interpret();
+ if (input.is8Bit())
+ return Interpreter<LChar>(bytecode, output, input.characters8(), input.length(), start).interpret();
+ return Interpreter<UChar>(bytecode, output, input.characters16(), input.length(), start).interpret();
}
-COMPILE_ASSERT(sizeof(Interpreter::BackTrackInfoPatternCharacter) == (YarrStackSpaceForBackTrackInfoPatternCharacter * sizeof(uintptr_t)), CheckYarrStackSpaceForBackTrackInfoPatternCharacter);
-COMPILE_ASSERT(sizeof(Interpreter::BackTrackInfoCharacterClass) == (YarrStackSpaceForBackTrackInfoCharacterClass * sizeof(uintptr_t)), CheckYarrStackSpaceForBackTrackInfoCharacterClass);
-COMPILE_ASSERT(sizeof(Interpreter::BackTrackInfoBackReference) == (YarrStackSpaceForBackTrackInfoBackReference * sizeof(uintptr_t)), CheckYarrStackSpaceForBackTrackInfoBackReference);
-COMPILE_ASSERT(sizeof(Interpreter::BackTrackInfoAlternative) == (YarrStackSpaceForBackTrackInfoAlternative * sizeof(uintptr_t)), CheckYarrStackSpaceForBackTrackInfoAlternative);
-COMPILE_ASSERT(sizeof(Interpreter::BackTrackInfoParentheticalAssertion) == (YarrStackSpaceForBackTrackInfoParentheticalAssertion * sizeof(uintptr_t)), CheckYarrStackSpaceForBackTrackInfoParentheticalAssertion);
-COMPILE_ASSERT(sizeof(Interpreter::BackTrackInfoParenthesesOnce) == (YarrStackSpaceForBackTrackInfoParenthesesOnce * sizeof(uintptr_t)), CheckYarrStackSpaceForBackTrackInfoParenthesesOnce);
-COMPILE_ASSERT(sizeof(Interpreter::BackTrackInfoParentheses) == (YarrStackSpaceForBackTrackInfoParentheses * sizeof(uintptr_t)), CheckYarrStackSpaceForBackTrackInfoParentheses);
+unsigned interpret(BytecodePattern* bytecode, const LChar* input, unsigned length, unsigned start, unsigned* output)
+{
+ return Interpreter<LChar>(bytecode, output, input, length, start).interpret();
+}
+unsigned interpret(BytecodePattern* bytecode, const UChar* input, unsigned length, unsigned start, unsigned* output)
+{
+ return Interpreter<UChar>(bytecode, output, input, length, start).interpret();
+}
+// These should be the same for both UChar & LChar.
+COMPILE_ASSERT(sizeof(Interpreter<UChar>::BackTrackInfoPatternCharacter) == (YarrStackSpaceForBackTrackInfoPatternCharacter * sizeof(uintptr_t)), CheckYarrStackSpaceForBackTrackInfoPatternCharacter);
+COMPILE_ASSERT(sizeof(Interpreter<UChar>::BackTrackInfoCharacterClass) == (YarrStackSpaceForBackTrackInfoCharacterClass * sizeof(uintptr_t)), CheckYarrStackSpaceForBackTrackInfoCharacterClass);
+COMPILE_ASSERT(sizeof(Interpreter<UChar>::BackTrackInfoBackReference) == (YarrStackSpaceForBackTrackInfoBackReference * sizeof(uintptr_t)), CheckYarrStackSpaceForBackTrackInfoBackReference);
+COMPILE_ASSERT(sizeof(Interpreter<UChar>::BackTrackInfoAlternative) == (YarrStackSpaceForBackTrackInfoAlternative * sizeof(uintptr_t)), CheckYarrStackSpaceForBackTrackInfoAlternative);
+COMPILE_ASSERT(sizeof(Interpreter<UChar>::BackTrackInfoParentheticalAssertion) == (YarrStackSpaceForBackTrackInfoParentheticalAssertion * sizeof(uintptr_t)), CheckYarrStackSpaceForBackTrackInfoParentheticalAssertion);
+COMPILE_ASSERT(sizeof(Interpreter<UChar>::BackTrackInfoParenthesesOnce) == (YarrStackSpaceForBackTrackInfoParenthesesOnce * sizeof(uintptr_t)), CheckYarrStackSpaceForBackTrackInfoParenthesesOnce);
+COMPILE_ASSERT(sizeof(Interpreter<UChar>::BackTrackInfoParentheses) == (YarrStackSpaceForBackTrackInfoParentheses * sizeof(uintptr_t)), CheckYarrStackSpaceForBackTrackInfoParentheses);
+
+
} }
Modified: trunk/Source/_javascript_Core/yarr/YarrInterpreter.h (112563 => 112564)
--- trunk/Source/_javascript_Core/yarr/YarrInterpreter.h 2012-03-29 19:48:21 UTC (rev 112563)
+++ trunk/Source/_javascript_Core/yarr/YarrInterpreter.h 2012-03-29 20:16:03 UTC (rev 112564)
@@ -375,6 +375,11 @@
Vector<CharacterClass*> m_userCharacterClasses;
};
+JS_EXPORT_PRIVATE PassOwnPtr<BytecodePattern> byteCompile(YarrPattern&, BumpPointerAllocator*);
+JS_EXPORT_PRIVATE unsigned interpret(BytecodePattern*, const UString& input, unsigned start, unsigned* output);
+unsigned interpret(BytecodePattern*, const LChar* input, unsigned length, unsigned start, unsigned* output);
+unsigned interpret(BytecodePattern*, const UChar* input, unsigned length, unsigned start, unsigned* output);
+
} } // namespace JSC::Yarr
#endif // YarrInterpreter_h
Modified: trunk/Source/WebCore/ChangeLog (112563 => 112564)
--- trunk/Source/WebCore/ChangeLog 2012-03-29 19:48:21 UTC (rev 112563)
+++ trunk/Source/WebCore/ChangeLog 2012-03-29 20:16:03 UTC (rev 112564)
@@ -1,3 +1,19 @@
+2012-03-29 Gavin Barraclough <[email protected]>
+
+ Template the Yarr::Interpreter on the character type
+ https://bugs.webkit.org/show_bug.cgi?id=82637
+
+ Reviewed by Sam Weinig.
+
+ We should be able to call to the interpreter after having already checked the character type,
+ without having to re-package the character pointer back up into a string!
+
+ * inspector/ContentSearchUtils.cpp:
+ (WebCore::ContentSearchUtils::findMagicComment):
+ * platform/text/RegularExpression.cpp:
+ (WebCore::RegularExpression::match):
+ - Don't pass length.
+
2012-03-29 Sheriff Bot <[email protected]>
Unreviewed, rolling out r112553.
Modified: trunk/Source/WebCore/inspector/ContentSearchUtils.cpp (112563 => 112564)
--- trunk/Source/WebCore/inspector/ContentSearchUtils.cpp 2012-03-29 19:48:21 UTC (rev 112563)
+++ trunk/Source/WebCore/inspector/ContentSearchUtils.cpp 2012-03-29 20:16:03 UTC (rev 112564)
@@ -152,7 +152,7 @@
ASSERT(pattern.m_numSubpatterns == 1);
Vector<int, 4> matches;
matches.resize(4);
- unsigned result = JSC::Yarr::interpret(bytecodePattern.get(), JSC::UString(content.impl()), 0, content.length(), reinterpret_cast<unsigned*>(matches.data()));
+ unsigned result = JSC::Yarr::interpret(bytecodePattern.get(), JSC::UString(content.impl()), 0, reinterpret_cast<unsigned*>(matches.data()));
if (result == JSC::Yarr::offsetNoMatch)
return String();
ASSERT(matches[2] > 0 && matches[3] > 0);
Modified: trunk/Source/WebCore/platform/text/RegularExpression.cpp (112563 => 112564)
--- trunk/Source/WebCore/platform/text/RegularExpression.cpp 2012-03-29 19:48:21 UTC (rev 112563)
+++ trunk/Source/WebCore/platform/text/RegularExpression.cpp 2012-03-29 20:16:03 UTC (rev 112564)
@@ -112,7 +112,7 @@
unsigned result;
if (str.length() <= INT_MAX)
- result = JSC::Yarr::interpret(d->m_regExpByteCode.get(), JSC::UString(str.impl()), startFrom, str.length(), offsetVector);
+ result = JSC::Yarr::interpret(d->m_regExpByteCode.get(), JSC::UString(str.impl()), startFrom, offsetVector);
else {
// This code can't handle unsigned offsets. Limit our processing to strings with offsets that
// can be represented as ints.
