v8/src/ast.cc

1151 lines
34 KiB
C++
Raw Normal View History

// Copyright 2012 the V8 project authors. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.
#include "src/ast.h"
#include <cmath> // For isfinite.
#include "src/builtins.h"
#include "src/code-stubs.h"
#include "src/contexts.h"
#include "src/conversions.h"
#include "src/hashmap.h"
#include "src/parser.h"
#include "src/property.h"
#include "src/property-details.h"
#include "src/scopes.h"
#include "src/string-stream.h"
#include "src/type-info.h"
namespace v8 {
namespace internal {
// ----------------------------------------------------------------------------
// All the Accept member functions for each syntax tree node type.
#define DECL_ACCEPT(type) \
void type::Accept(AstVisitor* v) { v->Visit##type(this); }
AST_NODE_LIST(DECL_ACCEPT)
#undef DECL_ACCEPT
// ----------------------------------------------------------------------------
// Implementation of other node functionality.
bool Expression::IsSmiLiteral() const {
return IsLiteral() && AsLiteral()->value()->IsSmi();
}
bool Expression::IsStringLiteral() const {
return IsLiteral() && AsLiteral()->value()->IsString();
}
bool Expression::IsNullLiteral() const {
return IsLiteral() && AsLiteral()->value()->IsNull();
}
bool Expression::IsUndefinedLiteral(Isolate* isolate) const {
const VariableProxy* var_proxy = AsVariableProxy();
if (var_proxy == NULL) return false;
Variable* var = var_proxy->var();
// The global identifier "undefined" is immutable. Everything
// else could be reassigned.
return var != NULL && var->location() == Variable::UNALLOCATED &&
var_proxy->raw_name()->IsOneByteEqualTo("undefined");
}
VariableProxy::VariableProxy(Zone* zone, Variable* var, int position)
: Expression(zone, position),
name_(var->raw_name()),
var_(NULL), // Will be set by the call to BindTo.
is_this_(var->is_this()),
is_trivial_(false),
is_lvalue_(false),
interface_(var->interface()) {
BindTo(var);
}
VariableProxy::VariableProxy(Zone* zone,
const AstRawString* name,
bool is_this,
Interface* interface,
int position)
: Expression(zone, position),
name_(name),
var_(NULL),
is_this_(is_this),
is_trivial_(false),
is_lvalue_(false),
interface_(interface) {
}
void VariableProxy::BindTo(Variable* var) {
ASSERT(var_ == NULL); // must be bound only once
ASSERT(var != NULL); // must bind
Get rid of static module allocation, do it in code. Modules now have their own local scope, represented by their own context. Module instance objects have an accessor for every export that forwards access to the respective slot from the module's context. (Exports that are modules themselves, however, are simple data properties.) All modules have a _hosting_ scope/context, which (currently) is the (innermost) enclosing global scope. To deal with recursion, nested modules are hosted by the same scope as global ones. For every (global or nested) module literal, the hosting context has an internal slot that points directly to the respective module context. This enables quick access to (statically resolved) module members by 2-dimensional access through the hosting context. For example, module A { let x; module B { let y; } } module C { let z; } allocates contexts as follows: [header| .A | .B | .C | A | C ] (global) | | | | | +-- [header| z ] (module) | | | +------- [header| y ] (module) | +------------ [header| x | B ] (module) Here, .A, .B, .C are the internal slots pointing to the hosted module contexts, whereas A, B, C hold the actual instance objects (note that every module context also points to the respective instance object through its extension slot in the header). To deal with arbitrary recursion and aliases between modules, they are created and initialized in several stages. Each stage applies to all modules in the hosting global scope, including nested ones. 1. Allocate: for each module _literal_, allocate the module contexts and respective instance object and wire them up. This happens in the PushModuleContext runtime function, as generated by AllocateModules (invoked by VisitDeclarations in the hosting scope). 2. Bind: for each module _declaration_ (i.e. literals as well as aliases), assign the respective instance object to respective local variables. This happens in VisitModuleDeclaration, and uses the instance objects created in the previous stage. For each module _literal_, this phase also constructs a module descriptor for the next stage. This happens in VisitModuleLiteral. 3. Populate: invoke the DeclareModules runtime function to populate each _instance_ object with accessors for it exports. This is generated by DeclareModules (invoked by VisitDeclarations in the hosting scope again), and uses the descriptors generated in the previous stage. 4. Initialize: execute the module bodies (and other code) in sequence. This happens by the separate statements generated for module bodies. To reenter the module scopes properly, the parser inserted ModuleStatements. R=mstarzinger@chromium.org,svenpanne@chromium.org BUG= Review URL: https://codereview.chromium.org/11093074 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13033 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-11-22 10:25:22 +00:00
ASSERT(!FLAG_harmony_modules || interface_->IsUnified(var->interface()));
ASSERT((is_this() && var->is_this()) || name_ == var->raw_name());
// Ideally CONST-ness should match. However, this is very hard to achieve
// because we don't know the exact semantics of conflicting (const and
// non-const) multiple variable declarations, const vars introduced via
// eval() etc. Const-ness and variable declarations are a complete mess
// in JS. Sigh...
var_ = var;
var->set_is_used(true);
}
Assignment::Assignment(Zone* zone,
Token::Value op,
Expression* target,
Expression* value,
int pos)
: Expression(zone, pos),
op_(op),
target_(target),
value_(value),
binary_operation_(NULL),
assignment_id_(GetNextId(zone)),
is_uninitialized_(false),
store_mode_(STANDARD_STORE) { }
Token::Value Assignment::binary_op() const {
switch (op_) {
case Token::ASSIGN_BIT_OR: return Token::BIT_OR;
case Token::ASSIGN_BIT_XOR: return Token::BIT_XOR;
case Token::ASSIGN_BIT_AND: return Token::BIT_AND;
case Token::ASSIGN_SHL: return Token::SHL;
case Token::ASSIGN_SAR: return Token::SAR;
case Token::ASSIGN_SHR: return Token::SHR;
case Token::ASSIGN_ADD: return Token::ADD;
case Token::ASSIGN_SUB: return Token::SUB;
case Token::ASSIGN_MUL: return Token::MUL;
case Token::ASSIGN_DIV: return Token::DIV;
case Token::ASSIGN_MOD: return Token::MOD;
default: UNREACHABLE();
}
return Token::ILLEGAL;
}
bool FunctionLiteral::AllowsLazyCompilation() {
return scope()->AllowsLazyCompilation();
}
bool FunctionLiteral::AllowsLazyCompilationWithoutContext() {
return scope()->AllowsLazyCompilationWithoutContext();
}
int FunctionLiteral::start_position() const {
return scope()->start_position();
}
int FunctionLiteral::end_position() const {
return scope()->end_position();
}
StrictMode FunctionLiteral::strict_mode() const {
return scope()->strict_mode();
}
void FunctionLiteral::InitializeSharedInfo(
Handle<Code> unoptimized_code) {
for (RelocIterator it(*unoptimized_code); !it.done(); it.next()) {
RelocInfo* rinfo = it.rinfo();
if (rinfo->rmode() != RelocInfo::EMBEDDED_OBJECT) continue;
Object* obj = rinfo->target_object();
if (obj->IsSharedFunctionInfo()) {
SharedFunctionInfo* shared = SharedFunctionInfo::cast(obj);
if (shared->start_position() == start_position()) {
shared_info_ = Handle<SharedFunctionInfo>(shared);
break;
}
}
}
}
ObjectLiteralProperty::ObjectLiteralProperty(Zone* zone,
AstValueFactory* ast_value_factory,
Literal* key, Expression* value) {
emit_store_ = true;
key_ = key;
value_ = value;
if (key->raw_value()->EqualsString(ast_value_factory->proto_string())) {
kind_ = PROTOTYPE;
} else if (value_->AsMaterializedLiteral() != NULL) {
kind_ = MATERIALIZED_LITERAL;
} else if (value_->IsLiteral()) {
kind_ = CONSTANT;
} else {
kind_ = COMPUTED;
}
}
ObjectLiteralProperty::ObjectLiteralProperty(
Zone* zone, bool is_getter, FunctionLiteral* value) {
emit_store_ = true;
value_ = value;
kind_ = is_getter ? GETTER : SETTER;
}
bool ObjectLiteral::Property::IsCompileTimeValue() {
return kind_ == CONSTANT ||
(kind_ == MATERIALIZED_LITERAL &&
CompileTimeValue::IsCompileTimeValue(value_));
}
void ObjectLiteral::Property::set_emit_store(bool emit_store) {
emit_store_ = emit_store;
}
bool ObjectLiteral::Property::emit_store() {
return emit_store_;
}
void ObjectLiteral::CalculateEmitStore(Zone* zone) {
ZoneAllocationPolicy allocator(zone);
ZoneHashMap table(Literal::Match, ZoneHashMap::kDefaultHashMapCapacity,
allocator);
for (int i = properties()->length() - 1; i >= 0; i--) {
ObjectLiteral::Property* property = properties()->at(i);
Literal* literal = property->key();
if (literal->value()->IsNull()) continue;
uint32_t hash = literal->Hash();
// If the key of a computed property is in the table, do not emit
// a store for the property later.
if ((property->kind() == ObjectLiteral::Property::MATERIALIZED_LITERAL ||
property->kind() == ObjectLiteral::Property::COMPUTED) &&
table.Lookup(literal, hash, false, allocator) != NULL) {
property->set_emit_store(false);
} else {
// Add key to the table.
table.Lookup(literal, hash, true, allocator);
}
}
}
bool ObjectLiteral::IsBoilerplateProperty(ObjectLiteral::Property* property) {
return property != NULL &&
property->kind() != ObjectLiteral::Property::PROTOTYPE;
}
void ObjectLiteral::BuildConstantProperties(Isolate* isolate) {
if (!constant_properties_.is_null()) return;
// Allocate a fixed array to hold all the constant properties.
Handle<FixedArray> constant_properties = isolate->factory()->NewFixedArray(
boilerplate_properties_ * 2, TENURED);
int position = 0;
// Accumulate the value in local variables and store it at the end.
bool is_simple = true;
int depth_acc = 1;
uint32_t max_element_index = 0;
uint32_t elements = 0;
for (int i = 0; i < properties()->length(); i++) {
ObjectLiteral::Property* property = properties()->at(i);
if (!IsBoilerplateProperty(property)) {
is_simple = false;
continue;
}
MaterializedLiteral* m_literal = property->value()->AsMaterializedLiteral();
if (m_literal != NULL) {
m_literal->BuildConstants(isolate);
if (m_literal->depth() >= depth_acc) depth_acc = m_literal->depth() + 1;
}
// Add CONSTANT and COMPUTED properties to boilerplate. Use undefined
// value for COMPUTED properties, the real value is filled in at
// runtime. The enumeration order is maintained.
Handle<Object> key = property->key()->value();
Handle<Object> value = GetBoilerplateValue(property->value(), isolate);
// Ensure objects that may, at any point in time, contain fields with double
// representation are always treated as nested objects. This is true for
// computed fields (value is undefined), and smi and double literals
// (value->IsNumber()).
// TODO(verwaest): Remove once we can store them inline.
if (FLAG_track_double_fields &&
(value->IsNumber() || value->IsUninitialized())) {
may_store_doubles_ = true;
}
is_simple = is_simple && !value->IsUninitialized();
// Keep track of the number of elements in the object literal and
// the largest element index. If the largest element index is
// much larger than the number of elements, creating an object
// literal with fast elements will be a waste of space.
uint32_t element_index = 0;
if (key->IsString()
&& Handle<String>::cast(key)->AsArrayIndex(&element_index)
&& element_index > max_element_index) {
max_element_index = element_index;
elements++;
} else if (key->IsSmi()) {
int key_value = Smi::cast(*key)->value();
if (key_value > 0
&& static_cast<uint32_t>(key_value) > max_element_index) {
max_element_index = key_value;
}
elements++;
}
// Add name, value pair to the fixed array.
constant_properties->set(position++, *key);
constant_properties->set(position++, *value);
}
constant_properties_ = constant_properties;
fast_elements_ =
(max_element_index <= 32) || ((2 * elements) >= max_element_index);
set_is_simple(is_simple);
set_depth(depth_acc);
}
void ArrayLiteral::BuildConstantElements(Isolate* isolate) {
if (!constant_elements_.is_null()) return;
// Allocate a fixed array to hold all the object literals.
Handle<JSArray> array =
isolate->factory()->NewJSArray(0, FAST_HOLEY_SMI_ELEMENTS);
JSArray::Expand(array, values()->length());
// Fill in the literals.
bool is_simple = true;
int depth_acc = 1;
bool is_holey = false;
for (int i = 0, n = values()->length(); i < n; i++) {
Expression* element = values()->at(i);
MaterializedLiteral* m_literal = element->AsMaterializedLiteral();
if (m_literal != NULL) {
m_literal->BuildConstants(isolate);
if (m_literal->depth() + 1 > depth_acc) {
depth_acc = m_literal->depth() + 1;
}
}
Handle<Object> boilerplate_value = GetBoilerplateValue(element, isolate);
if (boilerplate_value->IsTheHole()) {
is_holey = true;
} else if (boilerplate_value->IsUninitialized()) {
is_simple = false;
JSObject::SetOwnElement(
array, i, handle(Smi::FromInt(0), isolate), SLOPPY).Assert();
} else {
JSObject::SetOwnElement(array, i, boilerplate_value, SLOPPY).Assert();
}
}
Handle<FixedArrayBase> element_values(array->elements());
// Simple and shallow arrays can be lazily copied, we transform the
// elements array to a copy-on-write array.
if (is_simple && depth_acc == 1 && values()->length() > 0 &&
array->HasFastSmiOrObjectElements()) {
element_values->set_map(isolate->heap()->fixed_cow_array_map());
}
// Remember both the literal's constant values as well as the ElementsKind
// in a 2-element FixedArray.
Handle<FixedArray> literals = isolate->factory()->NewFixedArray(2, TENURED);
ElementsKind kind = array->GetElementsKind();
kind = is_holey ? GetHoleyElementsKind(kind) : GetPackedElementsKind(kind);
literals->set(0, Smi::FromInt(kind));
literals->set(1, *element_values);
constant_elements_ = literals;
set_is_simple(is_simple);
set_depth(depth_acc);
}
Handle<Object> MaterializedLiteral::GetBoilerplateValue(Expression* expression,
Isolate* isolate) {
if (expression->IsLiteral()) {
return expression->AsLiteral()->value();
}
if (CompileTimeValue::IsCompileTimeValue(expression)) {
return CompileTimeValue::GetValue(isolate, expression);
}
return isolate->factory()->uninitialized_value();
}
void MaterializedLiteral::BuildConstants(Isolate* isolate) {
if (IsArrayLiteral()) {
return AsArrayLiteral()->BuildConstantElements(isolate);
}
if (IsObjectLiteral()) {
return AsObjectLiteral()->BuildConstantProperties(isolate);
}
ASSERT(IsRegExpLiteral());
ASSERT(depth() >= 1); // Depth should be initialized.
}
void TargetCollector::AddTarget(Label* target, Zone* zone) {
// Add the label to the collector, but discard duplicates.
int length = targets_.length();
for (int i = 0; i < length; i++) {
if (targets_[i] == target) return;
}
targets_.Add(target, zone);
}
void UnaryOperation::RecordToBooleanTypeFeedback(TypeFeedbackOracle* oracle) {
// TODO(olivf) If this Operation is used in a test context, then the
// expression has a ToBoolean stub and we want to collect the type
// information. However the GraphBuilder expects it to be on the instruction
// corresponding to the TestContext, therefore we have to store it here and
// not on the operand.
set_to_boolean_types(oracle->ToBooleanTypes(expression()->test_id()));
}
void BinaryOperation::RecordToBooleanTypeFeedback(TypeFeedbackOracle* oracle) {
// TODO(olivf) If this Operation is used in a test context, then the right
// hand side has a ToBoolean stub and we want to collect the type information.
// However the GraphBuilder expects it to be on the instruction corresponding
// to the TestContext, therefore we have to store it here and not on the
// right hand operand.
set_to_boolean_types(oracle->ToBooleanTypes(right()->test_id()));
}
bool BinaryOperation::ResultOverwriteAllowed() const {
switch (op_) {
case Token::COMMA:
case Token::OR:
case Token::AND:
return false;
case Token::BIT_OR:
case Token::BIT_XOR:
case Token::BIT_AND:
case Token::SHL:
case Token::SAR:
case Token::SHR:
case Token::ADD:
case Token::SUB:
case Token::MUL:
case Token::DIV:
case Token::MOD:
return true;
default:
UNREACHABLE();
}
return false;
}
2011-09-19 14:50:33 +00:00
static bool IsTypeof(Expression* expr) {
UnaryOperation* maybe_unary = expr->AsUnaryOperation();
return maybe_unary != NULL && maybe_unary->op() == Token::TYPEOF;
}
2011-09-19 14:50:33 +00:00
// Check for the pattern: typeof <expression> equals <string literal>.
static bool MatchLiteralCompareTypeof(Expression* left,
Token::Value op,
Expression* right,
Expression** expr,
Handle<String>* check) {
if (IsTypeof(left) && right->IsStringLiteral() && Token::IsEqualityOp(op)) {
*expr = left->AsUnaryOperation()->expression();
*check = Handle<String>::cast(right->AsLiteral()->value());
return true;
}
return false;
}
2011-09-19 14:50:33 +00:00
bool CompareOperation::IsLiteralCompareTypeof(Expression** expr,
Handle<String>* check) {
return MatchLiteralCompareTypeof(left_, op_, right_, expr, check) ||
MatchLiteralCompareTypeof(right_, op_, left_, expr, check);
}
2011-09-19 14:50:33 +00:00
static bool IsVoidOfLiteral(Expression* expr) {
UnaryOperation* maybe_unary = expr->AsUnaryOperation();
return maybe_unary != NULL &&
maybe_unary->op() == Token::VOID &&
maybe_unary->expression()->IsLiteral();
2011-09-19 14:50:33 +00:00
}
// Check for the pattern: void <literal> equals <expression> or
// undefined equals <expression>
2011-09-19 14:50:33 +00:00
static bool MatchLiteralCompareUndefined(Expression* left,
Token::Value op,
Expression* right,
Expression** expr,
Isolate* isolate) {
2011-09-19 14:50:33 +00:00
if (IsVoidOfLiteral(left) && Token::IsEqualityOp(op)) {
*expr = right;
return true;
}
if (left->IsUndefinedLiteral(isolate) && Token::IsEqualityOp(op)) {
2011-09-19 14:50:33 +00:00
*expr = right;
return true;
}
return false;
}
bool CompareOperation::IsLiteralCompareUndefined(
Expression** expr, Isolate* isolate) {
return MatchLiteralCompareUndefined(left_, op_, right_, expr, isolate) ||
MatchLiteralCompareUndefined(right_, op_, left_, expr, isolate);
2011-09-19 14:50:33 +00:00
}
2011-09-19 14:50:33 +00:00
// Check for the pattern: null equals <expression>
static bool MatchLiteralCompareNull(Expression* left,
Token::Value op,
Expression* right,
Expression** expr) {
if (left->IsNullLiteral() && Token::IsEqualityOp(op)) {
*expr = right;
return true;
}
return false;
}
2011-09-19 14:50:33 +00:00
bool CompareOperation::IsLiteralCompareNull(Expression** expr) {
return MatchLiteralCompareNull(left_, op_, right_, expr) ||
MatchLiteralCompareNull(right_, op_, left_, expr);
}
// ----------------------------------------------------------------------------
// Inlining support
bool Declaration::IsInlineable() const {
return proxy()->var()->IsStackAllocated();
}
bool FunctionDeclaration::IsInlineable() const {
return false;
}
// ----------------------------------------------------------------------------
// Recording of type feedback
// TODO(rossberg): all RecordTypeFeedback functions should disappear
// once we use the common type field in the AST consistently.
void Expression::RecordToBooleanTypeFeedback(TypeFeedbackOracle* oracle) {
to_boolean_types_ = oracle->ToBooleanTypes(test_id());
}
bool Call::IsUsingCallFeedbackSlot(Isolate* isolate) const {
CallType call_type = GetCallType(isolate);
return (call_type != POSSIBLY_EVAL_CALL);
}
Call::CallType Call::GetCallType(Isolate* isolate) const {
VariableProxy* proxy = expression()->AsVariableProxy();
if (proxy != NULL) {
if (proxy->var()->is_possibly_eval(isolate)) {
return POSSIBLY_EVAL_CALL;
} else if (proxy->var()->IsUnallocated()) {
return GLOBAL_CALL;
} else if (proxy->var()->IsLookupSlot()) {
return LOOKUP_SLOT_CALL;
}
}
Property* property = expression()->AsProperty();
return property != NULL ? PROPERTY_CALL : OTHER_CALL;
}
bool Call::ComputeGlobalTarget(Handle<GlobalObject> global,
LookupResult* lookup) {
target_ = Handle<JSFunction>::null();
cell_ = Handle<Cell>::null();
ASSERT(lookup->IsFound() &&
lookup->type() == NORMAL &&
lookup->holder() == *global);
cell_ = Handle<Cell>(global->GetPropertyCell(lookup));
if (cell_->value()->IsJSFunction()) {
Handle<JSFunction> candidate(JSFunction::cast(cell_->value()));
// If the function is in new space we assume it's more likely to
// change and thus prefer the general IC code.
if (!lookup->isolate()->heap()->InNewSpace(*candidate)) {
target_ = candidate;
return true;
}
}
return false;
}
void CallNew::RecordTypeFeedback(TypeFeedbackOracle* oracle) {
int allocation_site_feedback_slot = FLAG_pretenuring_call_new
? AllocationSiteFeedbackSlot()
: CallNewFeedbackSlot();
allocation_site_ =
oracle->GetCallNewAllocationSite(allocation_site_feedback_slot);
is_monomorphic_ = oracle->CallNewIsMonomorphic(CallNewFeedbackSlot());
if (is_monomorphic_) {
target_ = oracle->GetCallNewTarget(CallNewFeedbackSlot());
if (!allocation_site_.is_null()) {
elements_kind_ = allocation_site_->GetElementsKind();
}
}
}
void ObjectLiteral::Property::RecordTypeFeedback(TypeFeedbackOracle* oracle) {
TypeFeedbackId id = key()->LiteralFeedbackId();
SmallMapList maps;
oracle->CollectReceiverTypes(id, &maps);
receiver_type_ = maps.length() == 1 ? maps.at(0)
: Handle<Map>::null();
}
// ----------------------------------------------------------------------------
// Implementation of AstVisitor
void AstVisitor::VisitDeclarations(ZoneList<Declaration*>* declarations) {
for (int i = 0; i < declarations->length(); i++) {
Visit(declarations->at(i));
}
}
void AstVisitor::VisitStatements(ZoneList<Statement*>* statements) {
for (int i = 0; i < statements->length(); i++) {
Statement* stmt = statements->at(i);
Visit(stmt);
if (stmt->IsJump()) break;
}
}
void AstVisitor::VisitExpressions(ZoneList<Expression*>* expressions) {
for (int i = 0; i < expressions->length(); i++) {
// The variable statement visiting code may pass NULL expressions
// to this code. Maybe this should be handled by introducing an
// undefined expression or literal? Revisit this code if this
// changes
Expression* expression = expressions->at(i);
if (expression != NULL) Visit(expression);
}
}
// ----------------------------------------------------------------------------
// Regular expressions
#define MAKE_ACCEPT(Name) \
void* RegExp##Name::Accept(RegExpVisitor* visitor, void* data) { \
return visitor->Visit##Name(this, data); \
}
FOR_EACH_REG_EXP_TREE_TYPE(MAKE_ACCEPT)
#undef MAKE_ACCEPT
#define MAKE_TYPE_CASE(Name) \
RegExp##Name* RegExpTree::As##Name() { \
return NULL; \
} \
bool RegExpTree::Is##Name() { return false; }
FOR_EACH_REG_EXP_TREE_TYPE(MAKE_TYPE_CASE)
#undef MAKE_TYPE_CASE
#define MAKE_TYPE_CASE(Name) \
RegExp##Name* RegExp##Name::As##Name() { \
return this; \
} \
bool RegExp##Name::Is##Name() { return true; }
FOR_EACH_REG_EXP_TREE_TYPE(MAKE_TYPE_CASE)
#undef MAKE_TYPE_CASE
static Interval ListCaptureRegisters(ZoneList<RegExpTree*>* children) {
Interval result = Interval::Empty();
for (int i = 0; i < children->length(); i++)
result = result.Union(children->at(i)->CaptureRegisters());
return result;
}
Interval RegExpAlternative::CaptureRegisters() {
return ListCaptureRegisters(nodes());
}
Interval RegExpDisjunction::CaptureRegisters() {
return ListCaptureRegisters(alternatives());
}
Interval RegExpLookahead::CaptureRegisters() {
return body()->CaptureRegisters();
}
Interval RegExpCapture::CaptureRegisters() {
Interval self(StartRegister(index()), EndRegister(index()));
return self.Union(body()->CaptureRegisters());
}
Interval RegExpQuantifier::CaptureRegisters() {
return body()->CaptureRegisters();
}
bool RegExpAssertion::IsAnchoredAtStart() {
return assertion_type() == RegExpAssertion::START_OF_INPUT;
}
bool RegExpAssertion::IsAnchoredAtEnd() {
return assertion_type() == RegExpAssertion::END_OF_INPUT;
}
bool RegExpAlternative::IsAnchoredAtStart() {
ZoneList<RegExpTree*>* nodes = this->nodes();
for (int i = 0; i < nodes->length(); i++) {
RegExpTree* node = nodes->at(i);
if (node->IsAnchoredAtStart()) { return true; }
if (node->max_match() > 0) { return false; }
}
return false;
}
bool RegExpAlternative::IsAnchoredAtEnd() {
ZoneList<RegExpTree*>* nodes = this->nodes();
for (int i = nodes->length() - 1; i >= 0; i--) {
RegExpTree* node = nodes->at(i);
if (node->IsAnchoredAtEnd()) { return true; }
if (node->max_match() > 0) { return false; }
}
return false;
}
bool RegExpDisjunction::IsAnchoredAtStart() {
ZoneList<RegExpTree*>* alternatives = this->alternatives();
for (int i = 0; i < alternatives->length(); i++) {
if (!alternatives->at(i)->IsAnchoredAtStart())
return false;
}
return true;
}
bool RegExpDisjunction::IsAnchoredAtEnd() {
ZoneList<RegExpTree*>* alternatives = this->alternatives();
for (int i = 0; i < alternatives->length(); i++) {
if (!alternatives->at(i)->IsAnchoredAtEnd())
return false;
}
return true;
}
bool RegExpLookahead::IsAnchoredAtStart() {
return is_positive() && body()->IsAnchoredAtStart();
}
bool RegExpCapture::IsAnchoredAtStart() {
return body()->IsAnchoredAtStart();
}
bool RegExpCapture::IsAnchoredAtEnd() {
return body()->IsAnchoredAtEnd();
}
// Convert regular expression trees to a simple sexp representation.
// This representation should be different from the input grammar
// in as many cases as possible, to make it more difficult for incorrect
// parses to look as correct ones which is likely if the input and
// output formats are alike.
class RegExpUnparser V8_FINAL : public RegExpVisitor {
public:
explicit RegExpUnparser(Zone* zone);
void VisitCharacterRange(CharacterRange that);
SmartArrayPointer<const char> ToString() { return stream_.ToCString(); }
#define MAKE_CASE(Name) virtual void* Visit##Name(RegExp##Name*, \
void* data) V8_OVERRIDE;
FOR_EACH_REG_EXP_TREE_TYPE(MAKE_CASE)
#undef MAKE_CASE
private:
StringStream* stream() { return &stream_; }
HeapStringAllocator alloc_;
StringStream stream_;
Zone* zone_;
};
RegExpUnparser::RegExpUnparser(Zone* zone) : stream_(&alloc_), zone_(zone) {
}
void* RegExpUnparser::VisitDisjunction(RegExpDisjunction* that, void* data) {
stream()->Add("(|");
for (int i = 0; i < that->alternatives()->length(); i++) {
stream()->Add(" ");
that->alternatives()->at(i)->Accept(this, data);
}
stream()->Add(")");
return NULL;
}
void* RegExpUnparser::VisitAlternative(RegExpAlternative* that, void* data) {
stream()->Add("(:");
for (int i = 0; i < that->nodes()->length(); i++) {
stream()->Add(" ");
that->nodes()->at(i)->Accept(this, data);
}
stream()->Add(")");
return NULL;
}
void RegExpUnparser::VisitCharacterRange(CharacterRange that) {
stream()->Add("%k", that.from());
if (!that.IsSingleton()) {
stream()->Add("-%k", that.to());
}
}
void* RegExpUnparser::VisitCharacterClass(RegExpCharacterClass* that,
void* data) {
if (that->is_negated())
stream()->Add("^");
stream()->Add("[");
for (int i = 0; i < that->ranges(zone_)->length(); i++) {
if (i > 0) stream()->Add(" ");
VisitCharacterRange(that->ranges(zone_)->at(i));
}
stream()->Add("]");
return NULL;
}
void* RegExpUnparser::VisitAssertion(RegExpAssertion* that, void* data) {
switch (that->assertion_type()) {
case RegExpAssertion::START_OF_INPUT:
stream()->Add("@^i");
break;
case RegExpAssertion::END_OF_INPUT:
stream()->Add("@$i");
break;
case RegExpAssertion::START_OF_LINE:
stream()->Add("@^l");
break;
case RegExpAssertion::END_OF_LINE:
stream()->Add("@$l");
break;
case RegExpAssertion::BOUNDARY:
stream()->Add("@b");
break;
case RegExpAssertion::NON_BOUNDARY:
stream()->Add("@B");
break;
}
return NULL;
}
void* RegExpUnparser::VisitAtom(RegExpAtom* that, void* data) {
stream()->Add("'");
Vector<const uc16> chardata = that->data();
for (int i = 0; i < chardata.length(); i++) {
stream()->Add("%k", chardata[i]);
}
stream()->Add("'");
return NULL;
}
void* RegExpUnparser::VisitText(RegExpText* that, void* data) {
if (that->elements()->length() == 1) {
that->elements()->at(0).tree()->Accept(this, data);
} else {
stream()->Add("(!");
for (int i = 0; i < that->elements()->length(); i++) {
stream()->Add(" ");
that->elements()->at(i).tree()->Accept(this, data);
}
stream()->Add(")");
}
return NULL;
}
void* RegExpUnparser::VisitQuantifier(RegExpQuantifier* that, void* data) {
stream()->Add("(# %i ", that->min());
if (that->max() == RegExpTree::kInfinity) {
stream()->Add("- ");
} else {
stream()->Add("%i ", that->max());
}
stream()->Add(that->is_greedy() ? "g " : that->is_possessive() ? "p " : "n ");
that->body()->Accept(this, data);
stream()->Add(")");
return NULL;
}
void* RegExpUnparser::VisitCapture(RegExpCapture* that, void* data) {
stream()->Add("(^ ");
that->body()->Accept(this, data);
stream()->Add(")");
return NULL;
}
void* RegExpUnparser::VisitLookahead(RegExpLookahead* that, void* data) {
stream()->Add("(-> ");
stream()->Add(that->is_positive() ? "+ " : "- ");
that->body()->Accept(this, data);
stream()->Add(")");
return NULL;
}
void* RegExpUnparser::VisitBackReference(RegExpBackReference* that,
void* data) {
stream()->Add("(<- %i)", that->index());
return NULL;
}
void* RegExpUnparser::VisitEmpty(RegExpEmpty* that, void* data) {
stream()->Put('%');
return NULL;
}
SmartArrayPointer<const char> RegExpTree::ToString(Zone* zone) {
RegExpUnparser unparser(zone);
Accept(&unparser, NULL);
return unparser.ToString();
}
RegExpDisjunction::RegExpDisjunction(ZoneList<RegExpTree*>* alternatives)
: alternatives_(alternatives) {
ASSERT(alternatives->length() > 1);
RegExpTree* first_alternative = alternatives->at(0);
min_match_ = first_alternative->min_match();
max_match_ = first_alternative->max_match();
for (int i = 1; i < alternatives->length(); i++) {
RegExpTree* alternative = alternatives->at(i);
min_match_ = Min(min_match_, alternative->min_match());
max_match_ = Max(max_match_, alternative->max_match());
}
}
static int IncreaseBy(int previous, int increase) {
if (RegExpTree::kInfinity - previous < increase) {
return RegExpTree::kInfinity;
} else {
return previous + increase;
}
}
RegExpAlternative::RegExpAlternative(ZoneList<RegExpTree*>* nodes)
: nodes_(nodes) {
ASSERT(nodes->length() > 1);
min_match_ = 0;
max_match_ = 0;
for (int i = 0; i < nodes->length(); i++) {
RegExpTree* node = nodes->at(i);
int node_min_match = node->min_match();
min_match_ = IncreaseBy(min_match_, node_min_match);
int node_max_match = node->max_match();
max_match_ = IncreaseBy(max_match_, node_max_match);
}
}
CaseClause::CaseClause(Zone* zone,
Expression* label,
ZoneList<Statement*>* statements,
int pos)
: Expression(zone, pos),
label_(label),
statements_(statements),
compare_type_(Type::None(zone)),
compare_id_(AstNode::GetNextId(zone)),
entry_id_(AstNode::GetNextId(zone)) {
}
#define REGULAR_NODE(NodeType) \
void AstConstructionVisitor::Visit##NodeType(NodeType* node) { \
increase_node_count(); \
}
#define REGULAR_NODE_WITH_FEEDBACK_SLOTS(NodeType) \
void AstConstructionVisitor::Visit##NodeType(NodeType* node) { \
increase_node_count(); \
add_slot_node(node); \
}
#define DONT_OPTIMIZE_NODE(NodeType) \
void AstConstructionVisitor::Visit##NodeType(NodeType* node) { \
increase_node_count(); \
set_dont_optimize_reason(k##NodeType); \
add_flag(kDontInline); \
add_flag(kDontSelfOptimize); \
}
#define DONT_SELFOPTIMIZE_NODE(NodeType) \
void AstConstructionVisitor::Visit##NodeType(NodeType* node) { \
increase_node_count(); \
add_flag(kDontSelfOptimize); \
}
#define DONT_SELFOPTIMIZE_NODE_WITH_FEEDBACK_SLOTS(NodeType) \
void AstConstructionVisitor::Visit##NodeType(NodeType* node) { \
increase_node_count(); \
add_slot_node(node); \
add_flag(kDontSelfOptimize); \
}
#define DONT_CACHE_NODE(NodeType) \
void AstConstructionVisitor::Visit##NodeType(NodeType* node) { \
increase_node_count(); \
set_dont_optimize_reason(k##NodeType); \
add_flag(kDontInline); \
add_flag(kDontSelfOptimize); \
add_flag(kDontCache); \
}
REGULAR_NODE(VariableDeclaration)
REGULAR_NODE(FunctionDeclaration)
REGULAR_NODE(Block)
REGULAR_NODE(ExpressionStatement)
REGULAR_NODE(EmptyStatement)
REGULAR_NODE(IfStatement)
REGULAR_NODE(ContinueStatement)
REGULAR_NODE(BreakStatement)
REGULAR_NODE(ReturnStatement)
REGULAR_NODE(SwitchStatement)
REGULAR_NODE(CaseClause)
REGULAR_NODE(Conditional)
REGULAR_NODE(Literal)
REGULAR_NODE(ArrayLiteral)
REGULAR_NODE(ObjectLiteral)
REGULAR_NODE(RegExpLiteral)
REGULAR_NODE(FunctionLiteral)
REGULAR_NODE(Assignment)
REGULAR_NODE(Throw)
REGULAR_NODE(Property)
REGULAR_NODE(UnaryOperation)
REGULAR_NODE(CountOperation)
REGULAR_NODE(BinaryOperation)
REGULAR_NODE(CompareOperation)
REGULAR_NODE(ThisFunction)
REGULAR_NODE_WITH_FEEDBACK_SLOTS(Call)
REGULAR_NODE_WITH_FEEDBACK_SLOTS(CallNew)
// In theory, for VariableProxy we'd have to add:
// if (node->var()->IsLookupSlot()) add_flag(kDontInline);
// But node->var() is usually not bound yet at VariableProxy creation time, and
// LOOKUP variables only result from constructs that cannot be inlined anyway.
REGULAR_NODE(VariableProxy)
Get rid of static module allocation, do it in code. Modules now have their own local scope, represented by their own context. Module instance objects have an accessor for every export that forwards access to the respective slot from the module's context. (Exports that are modules themselves, however, are simple data properties.) All modules have a _hosting_ scope/context, which (currently) is the (innermost) enclosing global scope. To deal with recursion, nested modules are hosted by the same scope as global ones. For every (global or nested) module literal, the hosting context has an internal slot that points directly to the respective module context. This enables quick access to (statically resolved) module members by 2-dimensional access through the hosting context. For example, module A { let x; module B { let y; } } module C { let z; } allocates contexts as follows: [header| .A | .B | .C | A | C ] (global) | | | | | +-- [header| z ] (module) | | | +------- [header| y ] (module) | +------------ [header| x | B ] (module) Here, .A, .B, .C are the internal slots pointing to the hosted module contexts, whereas A, B, C hold the actual instance objects (note that every module context also points to the respective instance object through its extension slot in the header). To deal with arbitrary recursion and aliases between modules, they are created and initialized in several stages. Each stage applies to all modules in the hosting global scope, including nested ones. 1. Allocate: for each module _literal_, allocate the module contexts and respective instance object and wire them up. This happens in the PushModuleContext runtime function, as generated by AllocateModules (invoked by VisitDeclarations in the hosting scope). 2. Bind: for each module _declaration_ (i.e. literals as well as aliases), assign the respective instance object to respective local variables. This happens in VisitModuleDeclaration, and uses the instance objects created in the previous stage. For each module _literal_, this phase also constructs a module descriptor for the next stage. This happens in VisitModuleLiteral. 3. Populate: invoke the DeclareModules runtime function to populate each _instance_ object with accessors for it exports. This is generated by DeclareModules (invoked by VisitDeclarations in the hosting scope again), and uses the descriptors generated in the previous stage. 4. Initialize: execute the module bodies (and other code) in sequence. This happens by the separate statements generated for module bodies. To reenter the module scopes properly, the parser inserted ModuleStatements. R=mstarzinger@chromium.org,svenpanne@chromium.org BUG= Review URL: https://codereview.chromium.org/11093074 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13033 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-11-22 10:25:22 +00:00
// We currently do not optimize any modules.
DONT_OPTIMIZE_NODE(ModuleDeclaration)
DONT_OPTIMIZE_NODE(ImportDeclaration)
DONT_OPTIMIZE_NODE(ExportDeclaration)
DONT_OPTIMIZE_NODE(ModuleVariable)
DONT_OPTIMIZE_NODE(ModulePath)
DONT_OPTIMIZE_NODE(ModuleUrl)
Get rid of static module allocation, do it in code. Modules now have their own local scope, represented by their own context. Module instance objects have an accessor for every export that forwards access to the respective slot from the module's context. (Exports that are modules themselves, however, are simple data properties.) All modules have a _hosting_ scope/context, which (currently) is the (innermost) enclosing global scope. To deal with recursion, nested modules are hosted by the same scope as global ones. For every (global or nested) module literal, the hosting context has an internal slot that points directly to the respective module context. This enables quick access to (statically resolved) module members by 2-dimensional access through the hosting context. For example, module A { let x; module B { let y; } } module C { let z; } allocates contexts as follows: [header| .A | .B | .C | A | C ] (global) | | | | | +-- [header| z ] (module) | | | +------- [header| y ] (module) | +------------ [header| x | B ] (module) Here, .A, .B, .C are the internal slots pointing to the hosted module contexts, whereas A, B, C hold the actual instance objects (note that every module context also points to the respective instance object through its extension slot in the header). To deal with arbitrary recursion and aliases between modules, they are created and initialized in several stages. Each stage applies to all modules in the hosting global scope, including nested ones. 1. Allocate: for each module _literal_, allocate the module contexts and respective instance object and wire them up. This happens in the PushModuleContext runtime function, as generated by AllocateModules (invoked by VisitDeclarations in the hosting scope). 2. Bind: for each module _declaration_ (i.e. literals as well as aliases), assign the respective instance object to respective local variables. This happens in VisitModuleDeclaration, and uses the instance objects created in the previous stage. For each module _literal_, this phase also constructs a module descriptor for the next stage. This happens in VisitModuleLiteral. 3. Populate: invoke the DeclareModules runtime function to populate each _instance_ object with accessors for it exports. This is generated by DeclareModules (invoked by VisitDeclarations in the hosting scope again), and uses the descriptors generated in the previous stage. 4. Initialize: execute the module bodies (and other code) in sequence. This happens by the separate statements generated for module bodies. To reenter the module scopes properly, the parser inserted ModuleStatements. R=mstarzinger@chromium.org,svenpanne@chromium.org BUG= Review URL: https://codereview.chromium.org/11093074 git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13033 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-11-22 10:25:22 +00:00
DONT_OPTIMIZE_NODE(ModuleStatement)
DONT_OPTIMIZE_NODE(Yield)
DONT_OPTIMIZE_NODE(WithStatement)
DONT_OPTIMIZE_NODE(TryCatchStatement)
DONT_OPTIMIZE_NODE(TryFinallyStatement)
DONT_OPTIMIZE_NODE(DebuggerStatement)
DONT_OPTIMIZE_NODE(NativeFunctionLiteral)
DONT_SELFOPTIMIZE_NODE(DoWhileStatement)
DONT_SELFOPTIMIZE_NODE(WhileStatement)
DONT_SELFOPTIMIZE_NODE(ForStatement)
DONT_SELFOPTIMIZE_NODE_WITH_FEEDBACK_SLOTS(ForInStatement)
DONT_SELFOPTIMIZE_NODE(ForOfStatement)
DONT_CACHE_NODE(ModuleLiteral)
void AstConstructionVisitor::VisitCallRuntime(CallRuntime* node) {
increase_node_count();
if (node->is_jsruntime()) {
// Don't try to inline JS runtime calls because we don't (currently) even
// optimize them.
add_flag(kDontInline);
} else if (node->function()->intrinsic_type == Runtime::INLINE &&
node->raw_name()->IsOneByteEqualTo("_Arguments")) {
// Don't inline the %_Arguments because it's implementation will not work.
// There is no stack frame to get them from.
add_flag(kDontInline);
}
}
#undef REGULAR_NODE
#undef DONT_OPTIMIZE_NODE
#undef DONT_SELFOPTIMIZE_NODE
#undef DONT_CACHE_NODE
Handle<String> Literal::ToString() {
if (value_->IsString()) return value_->AsString()->string();
ASSERT(value_->IsNumber());
char arr[100];
Vector<char> buffer(arr, ARRAY_SIZE(arr));
const char* str;
if (value()->IsSmi()) {
// Optimization only, the heap number case would subsume this.
SNPrintF(buffer, "%d", Smi::cast(*value())->value());
str = arr;
} else {
str = DoubleToCString(value()->Number(), buffer);
}
return isolate_->factory()->NewStringFromAsciiChecked(str);
}
} } // namespace v8::internal