/*
- * Copyright (C) 2007, 2008 Apple Inc. All rights reserved.
+ * Copyright (C) 2007, 2008, 2012-2015 Apple Inc. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
- * 3. Neither the name of Apple Computer, Inc. ("Apple") nor the names of
+ * 3. Neither the name of Apple Inc. ("Apple") nor the names of
* its contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
#ifndef SymbolTable_h
#define SymbolTable_h
+#include "ConcurrentJITLock.h"
+#include "ConstantMode.h"
+#include "InferredValue.h"
#include "JSObject.h"
-#include "UString.h"
-#include <wtf/AlwaysInline.h>
+#include "ScopedArgumentsTable.h"
+#include "TypeLocation.h"
+#include "VarOffset.h"
+#include "Watchpoint.h"
+#include <memory>
#include <wtf/HashTraits.h>
+#include <wtf/text/UniquedStringImpl.h>
namespace JSC {
- static ALWAYS_INLINE int missingSymbolMarker() { return std::numeric_limits<int>::max(); }
+class SymbolTable;
- // The bit twiddling in this class assumes that every register index is a
- // reasonably small positive or negative number, and therefore has its high
- // four bits all set or all unset.
+static ALWAYS_INLINE int missingSymbolMarker() { return std::numeric_limits<int>::max(); }
- struct SymbolTableEntry {
- SymbolTableEntry()
- : m_bits(0)
+// The bit twiddling in this class assumes that every register index is a
+// reasonably small positive or negative number, and therefore has its high
+// four bits all set or all unset.
+
+// In addition to implementing semantics-mandated variable attributes and
+// implementation-mandated variable indexing, this class also implements
+// watchpoints to be used for JIT optimizations. Because watchpoints are
+// meant to be relatively rare, this class optimizes heavily for the case
+// that they are not being used. To that end, this class uses the thin-fat
+// idiom: either it is thin, in which case it contains an in-place encoded
+// word that consists of attributes, the index, and a bit saying that it is
+// thin; or it is fat, in which case it contains a pointer to a malloc'd
+// data structure and a bit saying that it is fat. The malloc'd data
+// structure will be malloced a second time upon copy, to preserve the
+// property that in-place edits to SymbolTableEntry do not manifest in any
+// copies. However, the malloc'd FatEntry data structure contains a ref-
+// counted pointer to a shared WatchpointSet. Thus, in-place edits of the
+// WatchpointSet will manifest in all copies. Here's a picture:
+//
+// SymbolTableEntry --> FatEntry --> WatchpointSet
+//
+// If you make a copy of a SymbolTableEntry, you will have:
+//
+// original: SymbolTableEntry --> FatEntry --> WatchpointSet
+// copy: SymbolTableEntry --> FatEntry -----^
+
+struct SymbolTableEntry {
+private:
+ static VarOffset varOffsetFromBits(intptr_t bits)
+ {
+ VarKind kind;
+ intptr_t kindBits = bits & KindBitsMask;
+ if (kindBits <= UnwatchableScopeKindBits)
+ kind = VarKind::Scope;
+ else if (kindBits == StackKindBits)
+ kind = VarKind::Stack;
+ else
+ kind = VarKind::DirectArgument;
+ return VarOffset::assemble(kind, static_cast<int>(bits >> FlagBits));
+ }
+
+ static ScopeOffset scopeOffsetFromBits(intptr_t bits)
+ {
+ ASSERT((bits & KindBitsMask) <= UnwatchableScopeKindBits);
+ return ScopeOffset(static_cast<int>(bits >> FlagBits));
+ }
+
+public:
+
+ // Use the SymbolTableEntry::Fast class, either via implicit cast or by calling
+ // getFast(), when you (1) only care about isNull(), getIndex(), and isReadOnly(),
+ // and (2) you are in a hot path where you need to minimize the number of times
+ // that you branch on isFat() when getting the bits().
+ class Fast {
+ public:
+ Fast()
+ : m_bits(SlimFlag)
{
}
-
- SymbolTableEntry(int index)
+
+ ALWAYS_INLINE Fast(const SymbolTableEntry& entry)
+ : m_bits(entry.bits())
{
- ASSERT(isValidIndex(index));
- pack(index, false, false);
+ }
+
+ bool isNull() const
+ {
+ return !(m_bits & ~SlimFlag);
}
- SymbolTableEntry(int index, unsigned attributes)
+ VarOffset varOffset() const
{
- ASSERT(isValidIndex(index));
- pack(index, attributes & ReadOnly, attributes & DontEnum);
+ return varOffsetFromBits(m_bits);
}
- bool isNull() const
+ // Asserts if the offset is anything but a scope offset. This structures the assertions
+ // in a way that may result in better code, even in release, than doing
+ // varOffset().scopeOffset().
+ ScopeOffset scopeOffset() const
{
- return !m_bits;
+ return scopeOffsetFromBits(m_bits);
}
-
- int getIndex() const
+
+ bool isReadOnly() const
{
- return m_bits >> FlagBits;
+ return m_bits & ReadOnlyFlag;
}
-
+
+ bool isDontEnum() const
+ {
+ return m_bits & DontEnumFlag;
+ }
+
unsigned getAttributes() const
{
unsigned attributes = 0;
- if (m_bits & ReadOnlyFlag)
+ if (isReadOnly())
attributes |= ReadOnly;
- if (m_bits & DontEnumFlag)
+ if (isDontEnum())
attributes |= DontEnum;
return attributes;
}
- void setAttributes(unsigned attributes)
+ bool isFat() const
{
- pack(getIndex(), attributes & ReadOnly, attributes & DontEnum);
+ return !(m_bits & SlimFlag);
}
+
+ private:
+ friend struct SymbolTableEntry;
+ intptr_t m_bits;
+ };
- bool isReadOnly() const
- {
- return m_bits & ReadOnlyFlag;
- }
+ SymbolTableEntry()
+ : m_bits(SlimFlag)
+ {
+ }
- private:
- static const unsigned ReadOnlyFlag = 0x1;
- static const unsigned DontEnumFlag = 0x2;
- static const unsigned NotNullFlag = 0x4;
- static const unsigned FlagBits = 3;
+ SymbolTableEntry(VarOffset offset)
+ : m_bits(SlimFlag)
+ {
+ ASSERT(isValidVarOffset(offset));
+ pack(offset, true, false, false);
+ }
- void pack(int index, bool readOnly, bool dontEnum)
+ SymbolTableEntry(VarOffset offset, unsigned attributes)
+ : m_bits(SlimFlag)
+ {
+ ASSERT(isValidVarOffset(offset));
+ pack(offset, true, attributes & ReadOnly, attributes & DontEnum);
+ }
+
+ ~SymbolTableEntry()
+ {
+ freeFatEntry();
+ }
+
+ SymbolTableEntry(const SymbolTableEntry& other)
+ : m_bits(SlimFlag)
+ {
+ *this = other;
+ }
+
+ SymbolTableEntry& operator=(const SymbolTableEntry& other)
+ {
+ if (UNLIKELY(other.isFat()))
+ return copySlow(other);
+ freeFatEntry();
+ m_bits = other.m_bits;
+ return *this;
+ }
+
+ bool isNull() const
+ {
+ return !(bits() & ~SlimFlag);
+ }
+
+ VarOffset varOffset() const
+ {
+ return varOffsetFromBits(bits());
+ }
+
+ bool isWatchable() const
+ {
+ return (m_bits & KindBitsMask) == ScopeKindBits;
+ }
+
+ // Asserts if the offset is anything but a scope offset. This structures the assertions
+ // in a way that may result in better code, even in release, than doing
+ // varOffset().scopeOffset().
+ ScopeOffset scopeOffset() const
+ {
+ return scopeOffsetFromBits(bits());
+ }
+
+ ALWAYS_INLINE Fast getFast() const
+ {
+ return Fast(*this);
+ }
+
+ ALWAYS_INLINE Fast getFast(bool& wasFat) const
+ {
+ Fast result;
+ wasFat = isFat();
+ if (wasFat)
+ result.m_bits = fatEntry()->m_bits | SlimFlag;
+ else
+ result.m_bits = m_bits;
+ return result;
+ }
+
+ unsigned getAttributes() const
+ {
+ return getFast().getAttributes();
+ }
+
+ void setAttributes(unsigned attributes)
+ {
+ pack(varOffset(), isWatchable(), attributes & ReadOnly, attributes & DontEnum);
+ }
+
+ bool isReadOnly() const
+ {
+ return bits() & ReadOnlyFlag;
+ }
+
+ ConstantMode constantMode() const
+ {
+ return modeForIsConstant(isReadOnly());
+ }
+
+ bool isDontEnum() const
+ {
+ return bits() & DontEnumFlag;
+ }
+
+ void disableWatching()
+ {
+ if (WatchpointSet* set = watchpointSet())
+ set->invalidate("Disabling watching in symbol table");
+ if (varOffset().isScope())
+ pack(varOffset(), false, isReadOnly(), isDontEnum());
+ }
+
+ void prepareToWatch();
+
+ void addWatchpoint(Watchpoint*);
+
+ // This watchpoint set is initialized clear, and goes through the following state transitions:
+ //
+ // First write to this var, in any scope that has this symbol table: Clear->IsWatched.
+ //
+ // Second write to this var, in any scope that has this symbol table: IsWatched->IsInvalidated.
+ //
+ // We ensure that we touch the set (i.e. trigger its state transition) after we do the write. This
+ // means that if you're in the compiler thread, and you:
+ //
+ // 1) Observe that the set IsWatched and commit to adding your watchpoint.
+ // 2) Load a value from any scope that has this watchpoint set.
+ //
+ // Then you can be sure that that value is either going to be the correct value for that var forever,
+ // or the watchpoint set will invalidate and you'll get fired.
+ //
+ // It's possible to write a program that first creates multiple scopes with the same var, and then
+ // initializes that var in just one of them. This means that a compilation could constant-fold to one
+ // of the scopes that still has an undefined value for this variable. That's fine, because at that
+ // point any write to any of the instances of that variable would fire the watchpoint.
+ WatchpointSet* watchpointSet()
+ {
+ if (!isFat())
+ return 0;
+ return fatEntry()->m_watchpoints.get();
+ }
+
+private:
+ static const intptr_t SlimFlag = 0x1;
+ static const intptr_t ReadOnlyFlag = 0x2;
+ static const intptr_t DontEnumFlag = 0x4;
+ static const intptr_t NotNullFlag = 0x8;
+ static const intptr_t KindBitsMask = 0x30;
+ static const intptr_t ScopeKindBits = 0x00;
+ static const intptr_t UnwatchableScopeKindBits = 0x10;
+ static const intptr_t StackKindBits = 0x20;
+ static const intptr_t DirectArgumentKindBits = 0x30;
+ static const intptr_t FlagBits = 6;
+
+ class FatEntry {
+ WTF_MAKE_FAST_ALLOCATED;
+ public:
+ FatEntry(intptr_t bits)
+ : m_bits(bits & ~SlimFlag)
{
- m_bits = (index << FlagBits) | NotNullFlag;
- if (readOnly)
- m_bits |= ReadOnlyFlag;
- if (dontEnum)
- m_bits |= DontEnumFlag;
}
- bool isValidIndex(int index)
- {
- return ((index << FlagBits) >> FlagBits) == index;
+ intptr_t m_bits; // always has FatFlag set and exactly matches what the bits would have been if this wasn't fat.
+
+ RefPtr<WatchpointSet> m_watchpoints;
+ };
+
+ SymbolTableEntry& copySlow(const SymbolTableEntry&);
+ JS_EXPORT_PRIVATE void notifyWriteSlow(VM&, JSValue, const FireDetail&);
+
+ bool isFat() const
+ {
+ return !(m_bits & SlimFlag);
+ }
+
+ const FatEntry* fatEntry() const
+ {
+ ASSERT(isFat());
+ return bitwise_cast<const FatEntry*>(m_bits);
+ }
+
+ FatEntry* fatEntry()
+ {
+ ASSERT(isFat());
+ return bitwise_cast<FatEntry*>(m_bits);
+ }
+
+ FatEntry* inflate()
+ {
+ if (LIKELY(isFat()))
+ return fatEntry();
+ return inflateSlow();
+ }
+
+ FatEntry* inflateSlow();
+
+ ALWAYS_INLINE intptr_t bits() const
+ {
+ if (isFat())
+ return fatEntry()->m_bits;
+ return m_bits;
+ }
+
+ ALWAYS_INLINE intptr_t& bits()
+ {
+ if (isFat())
+ return fatEntry()->m_bits;
+ return m_bits;
+ }
+
+ void freeFatEntry()
+ {
+ if (LIKELY(!isFat()))
+ return;
+ freeFatEntrySlow();
+ }
+
+ JS_EXPORT_PRIVATE void freeFatEntrySlow();
+
+ void pack(VarOffset offset, bool isWatchable, bool readOnly, bool dontEnum)
+ {
+ ASSERT(!isFat());
+ intptr_t& bitsRef = bits();
+ bitsRef =
+ (static_cast<intptr_t>(offset.rawOffset()) << FlagBits) | NotNullFlag | SlimFlag;
+ if (readOnly)
+ bitsRef |= ReadOnlyFlag;
+ if (dontEnum)
+ bitsRef |= DontEnumFlag;
+ switch (offset.kind()) {
+ case VarKind::Scope:
+ if (isWatchable)
+ bitsRef |= ScopeKindBits;
+ else
+ bitsRef |= UnwatchableScopeKindBits;
+ break;
+ case VarKind::Stack:
+ bitsRef |= StackKindBits;
+ break;
+ case VarKind::DirectArgument:
+ bitsRef |= DirectArgumentKindBits;
+ break;
+ default:
+ RELEASE_ASSERT_NOT_REACHED();
+ break;
}
+ }
+
+ static bool isValidVarOffset(VarOffset offset)
+ {
+ return ((static_cast<intptr_t>(offset.rawOffset()) << FlagBits) >> FlagBits) == static_cast<intptr_t>(offset.rawOffset());
+ }
- int m_bits;
- };
+ intptr_t m_bits;
+};
- struct SymbolTableIndexHashTraits : HashTraits<SymbolTableEntry> {
- static const bool emptyValueIsZero = true;
- static const bool needsDestruction = false;
- };
+struct SymbolTableIndexHashTraits : HashTraits<SymbolTableEntry> {
+ static const bool needsDestruction = true;
+};
- typedef HashMap<RefPtr<StringImpl>, SymbolTableEntry, IdentifierRepHash, HashTraits<RefPtr<StringImpl> >, SymbolTableIndexHashTraits> SymbolTable;
+class SymbolTable final : public JSCell {
+public:
+ typedef JSCell Base;
+ static const unsigned StructureFlags = Base::StructureFlags | StructureIsImmortal;
- class SharedSymbolTable : public SymbolTable, public RefCounted<SharedSymbolTable> {
- WTF_MAKE_FAST_ALLOCATED;
- public:
- static PassRefPtr<SharedSymbolTable> create() { return adoptRef(new SharedSymbolTable); }
- private:
- SharedSymbolTable() { turnOffVerifier(); }
+ typedef HashMap<RefPtr<UniquedStringImpl>, SymbolTableEntry, IdentifierRepHash, HashTraits<RefPtr<UniquedStringImpl>>, SymbolTableIndexHashTraits> Map;
+ typedef HashMap<RefPtr<UniquedStringImpl>, GlobalVariableID, IdentifierRepHash> UniqueIDMap;
+ typedef HashMap<RefPtr<UniquedStringImpl>, RefPtr<TypeSet>, IdentifierRepHash> UniqueTypeSetMap;
+ typedef HashMap<VarOffset, RefPtr<UniquedStringImpl>> OffsetToVariableMap;
+ typedef Vector<SymbolTableEntry*> LocalToEntryVec;
+
+ static SymbolTable* create(VM& vm)
+ {
+ SymbolTable* symbolTable = new (NotNull, allocateCell<SymbolTable>(vm.heap)) SymbolTable(vm);
+ symbolTable->finishCreation(vm);
+ return symbolTable;
+ }
+
+ static SymbolTable* createNameScopeTable(VM& vm, const Identifier& ident, unsigned attributes)
+ {
+ SymbolTable* result = create(vm);
+ result->add(ident.impl(), SymbolTableEntry(VarOffset(ScopeOffset(0)), attributes));
+ return result;
+ }
+
+ static const bool needsDestruction = true;
+ static void destroy(JSCell*);
+
+ static Structure* createStructure(VM& vm, JSGlobalObject* globalObject, JSValue prototype)
+ {
+ return Structure::create(vm, globalObject, prototype, TypeInfo(CellType, StructureFlags), info());
+ }
+
+ // You must hold the lock until after you're done with the iterator.
+ Map::iterator find(const ConcurrentJITLocker&, UniquedStringImpl* key)
+ {
+ return m_map.find(key);
+ }
+
+ Map::iterator find(const GCSafeConcurrentJITLocker&, UniquedStringImpl* key)
+ {
+ return m_map.find(key);
+ }
+
+ SymbolTableEntry get(const ConcurrentJITLocker&, UniquedStringImpl* key)
+ {
+ return m_map.get(key);
+ }
+
+ SymbolTableEntry get(UniquedStringImpl* key)
+ {
+ ConcurrentJITLocker locker(m_lock);
+ return get(locker, key);
+ }
+
+ SymbolTableEntry inlineGet(const ConcurrentJITLocker&, UniquedStringImpl* key)
+ {
+ return m_map.inlineGet(key);
+ }
+
+ SymbolTableEntry inlineGet(UniquedStringImpl* key)
+ {
+ ConcurrentJITLocker locker(m_lock);
+ return inlineGet(locker, key);
+ }
+
+ Map::iterator begin(const ConcurrentJITLocker&)
+ {
+ return m_map.begin();
+ }
+
+ Map::iterator end(const ConcurrentJITLocker&)
+ {
+ return m_map.end();
+ }
+
+ Map::iterator end(const GCSafeConcurrentJITLocker&)
+ {
+ return m_map.end();
+ }
+
+ size_t size(const ConcurrentJITLocker&) const
+ {
+ return m_map.size();
+ }
+
+ size_t size() const
+ {
+ ConcurrentJITLocker locker(m_lock);
+ return size(locker);
+ }
+
+ ScopeOffset maxScopeOffset() const
+ {
+ return m_maxScopeOffset;
+ }
+
+ void didUseScopeOffset(ScopeOffset offset)
+ {
+ if (!m_maxScopeOffset || m_maxScopeOffset < offset)
+ m_maxScopeOffset = offset;
+ }
+
+ void didUseVarOffset(VarOffset offset)
+ {
+ if (offset.isScope())
+ didUseScopeOffset(offset.scopeOffset());
+ }
+
+ unsigned scopeSize() const
+ {
+ ScopeOffset maxScopeOffset = this->maxScopeOffset();
+
+ // Do some calculation that relies on invalid scope offset plus one being zero.
+ unsigned fastResult = maxScopeOffset.offsetUnchecked() + 1;
+
+ // Assert that this works.
+ ASSERT(fastResult == (!maxScopeOffset ? 0 : maxScopeOffset.offset() + 1));
+
+ return fastResult;
+ }
+
+ ScopeOffset nextScopeOffset() const
+ {
+ return ScopeOffset(scopeSize());
+ }
+
+ ScopeOffset takeNextScopeOffset(const ConcurrentJITLocker&)
+ {
+ ScopeOffset result = nextScopeOffset();
+ m_maxScopeOffset = result;
+ return result;
+ }
+
+ ScopeOffset takeNextScopeOffset()
+ {
+ ConcurrentJITLocker locker(m_lock);
+ return takeNextScopeOffset(locker);
+ }
+
+ void add(const ConcurrentJITLocker&, UniquedStringImpl* key, const SymbolTableEntry& entry)
+ {
+ RELEASE_ASSERT(!m_localToEntry);
+ didUseVarOffset(entry.varOffset());
+ Map::AddResult result = m_map.add(key, entry);
+ ASSERT_UNUSED(result, result.isNewEntry);
+ }
+
+ void add(UniquedStringImpl* key, const SymbolTableEntry& entry)
+ {
+ ConcurrentJITLocker locker(m_lock);
+ add(locker, key, entry);
+ }
+
+ void set(const ConcurrentJITLocker&, UniquedStringImpl* key, const SymbolTableEntry& entry)
+ {
+ RELEASE_ASSERT(!m_localToEntry);
+ didUseVarOffset(entry.varOffset());
+ m_map.set(key, entry);
+ }
+
+ void set(UniquedStringImpl* key, const SymbolTableEntry& entry)
+ {
+ ConcurrentJITLocker locker(m_lock);
+ set(locker, key, entry);
+ }
+
+ bool contains(const ConcurrentJITLocker&, UniquedStringImpl* key)
+ {
+ return m_map.contains(key);
+ }
+
+ bool contains(UniquedStringImpl* key)
+ {
+ ConcurrentJITLocker locker(m_lock);
+ return contains(locker, key);
+ }
+
+ // The principle behind ScopedArgumentsTable modifications is that we will create one and
+ // leave it unlocked - thereby allowing in-place changes - until someone asks for a pointer to
+ // the table. Then, we will lock it. Then both our future changes and their future changes
+ // will first have to make a copy. This discipline means that usually when we create a
+ // ScopedArguments object, we don't have to make a copy of the ScopedArgumentsTable - instead
+ // we just take a reference to one that we already have.
+
+ uint32_t argumentsLength() const
+ {
+ if (!m_arguments)
+ return 0;
+ return m_arguments->length();
+ }
+
+ void setArgumentsLength(VM& vm, uint32_t length)
+ {
+ if (UNLIKELY(!m_arguments))
+ m_arguments.set(vm, this, ScopedArgumentsTable::create(vm));
+ m_arguments.set(vm, this, m_arguments->setLength(vm, length));
+ }
+
+ ScopeOffset argumentOffset(uint32_t i) const
+ {
+ ASSERT_WITH_SECURITY_IMPLICATION(m_arguments);
+ return m_arguments->get(i);
+ }
+
+ void setArgumentOffset(VM& vm, uint32_t i, ScopeOffset offset)
+ {
+ ASSERT_WITH_SECURITY_IMPLICATION(m_arguments);
+ m_arguments.set(vm, this, m_arguments->set(vm, i, offset));
+ }
+
+ ScopedArgumentsTable* arguments() const
+ {
+ if (!m_arguments)
+ return nullptr;
+ m_arguments->lock();
+ return m_arguments.get();
+ }
+
+ const LocalToEntryVec& localToEntry(const ConcurrentJITLocker&);
+ SymbolTableEntry* entryFor(const ConcurrentJITLocker&, ScopeOffset);
+
+ GlobalVariableID uniqueIDForVariable(const ConcurrentJITLocker&, UniquedStringImpl* key, VM&);
+ GlobalVariableID uniqueIDForOffset(const ConcurrentJITLocker&, VarOffset, VM&);
+ RefPtr<TypeSet> globalTypeSetForOffset(const ConcurrentJITLocker&, VarOffset, VM&);
+ RefPtr<TypeSet> globalTypeSetForVariable(const ConcurrentJITLocker&, UniquedStringImpl* key, VM&);
+
+ bool usesNonStrictEval() { return m_usesNonStrictEval; }
+ void setUsesNonStrictEval(bool usesNonStrictEval) { m_usesNonStrictEval = usesNonStrictEval; }
+
+ SymbolTable* cloneScopePart(VM&);
+
+ void prepareForTypeProfiling(const ConcurrentJITLocker&);
+
+ InferredValue* singletonScope() { return m_singletonScope.get(); }
+
+ static void visitChildren(JSCell*, SlotVisitor&);
+
+ DECLARE_EXPORT_INFO;
+
+private:
+ JS_EXPORT_PRIVATE SymbolTable(VM&);
+ ~SymbolTable();
+
+ JS_EXPORT_PRIVATE void finishCreation(VM&);
+
+ Map m_map;
+ ScopeOffset m_maxScopeOffset;
+
+ struct TypeProfilingRareData {
+ UniqueIDMap m_uniqueIDMap;
+ OffsetToVariableMap m_offsetToVariableMap;
+ UniqueTypeSetMap m_uniqueTypeSetMap;
};
+ std::unique_ptr<TypeProfilingRareData> m_typeProfilingRareData;
+
+ bool m_usesNonStrictEval;
+ WriteBarrier<ScopedArgumentsTable> m_arguments;
+ WriteBarrier<InferredValue> m_singletonScope;
+
+ std::unique_ptr<LocalToEntryVec> m_localToEntry;
+
+public:
+ mutable ConcurrentJITLock m_lock;
+};
+
} // namespace JSC
#endif // SymbolTable_h
projects / apple / javascriptcore.git / blobdiff