v8/src/parsing/preparser.cc
verwaest f41e7ebd62 Don't use different function scopes when parsing with temp zones
Previously we'd have a scope in the main zone, and another in the temp zone. Then we carefully copied back data to the main zone. This CL changes it so that the scope is just fixed up to only contain data from the main zone. That avoids additional copies and additional allocations; while not increasing the care that needs to be taken. This will also make it easier to abort preparsing while parsing using a temp zone.

BUG=

Review-Url: https://codereview.chromium.org/2368313002
Cr-Commit-Position: refs/heads/master@{#39800}
2016-09-28 02:42:28 +00:00

327 lines
12 KiB
C++

// Copyright 2011 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 <cmath>
#include "src/allocation.h"
#include "src/base/logging.h"
#include "src/conversions-inl.h"
#include "src/conversions.h"
#include "src/globals.h"
#include "src/list.h"
#include "src/parsing/duplicate-finder.h"
#include "src/parsing/parser-base.h"
#include "src/parsing/preparse-data-format.h"
#include "src/parsing/preparse-data.h"
#include "src/parsing/preparser.h"
#include "src/unicode.h"
#include "src/utils.h"
namespace v8 {
namespace internal {
// ----------------------------------------------------------------------------
// The CHECK_OK macro is a convenient macro to enforce error
// handling for functions that may fail (by returning !*ok).
//
// CAUTION: This macro appends extra statements after a call,
// thus it must never be used where only a single statement
// is correct (e.g. an if statement branch w/o braces)!
#define CHECK_OK_VALUE(x) ok); \
if (!*ok) return x; \
((void)0
#define DUMMY ) // to make indentation work
#undef DUMMY
#define CHECK_OK CHECK_OK_VALUE(Statement::Default())
#define CHECK_OK_VOID CHECK_OK_VALUE(this->Void())
namespace {
PreParserIdentifier GetSymbolHelper(Scanner* scanner) {
switch (scanner->current_token()) {
case Token::ENUM:
return PreParserIdentifier::Enum();
case Token::AWAIT:
return PreParserIdentifier::Await();
case Token::FUTURE_STRICT_RESERVED_WORD:
return PreParserIdentifier::FutureStrictReserved();
case Token::LET:
return PreParserIdentifier::Let();
case Token::STATIC:
return PreParserIdentifier::Static();
case Token::YIELD:
return PreParserIdentifier::Yield();
case Token::ASYNC:
return PreParserIdentifier::Async();
default:
if (scanner->UnescapedLiteralMatches("eval", 4))
return PreParserIdentifier::Eval();
if (scanner->UnescapedLiteralMatches("arguments", 9))
return PreParserIdentifier::Arguments();
if (scanner->UnescapedLiteralMatches("undefined", 9))
return PreParserIdentifier::Undefined();
if (scanner->LiteralMatches("prototype", 9))
return PreParserIdentifier::Prototype();
if (scanner->LiteralMatches("constructor", 11))
return PreParserIdentifier::Constructor();
return PreParserIdentifier::Default();
}
}
} // unnamed namespace
PreParserIdentifier PreParser::GetSymbol() const {
PreParserIdentifier symbol = GetSymbolHelper(scanner());
if (track_unresolved_variables_) {
const AstRawString* result = scanner()->CurrentSymbol(ast_value_factory());
DCHECK_NOT_NULL(result);
symbol.string_ = result;
}
return symbol;
}
PreParser::PreParseResult PreParser::PreParseLazyFunction(
DeclarationScope* function_scope, bool parsing_module, ParserRecorder* log,
bool is_inner_function, bool may_abort, int* use_counts) {
DCHECK_EQ(FUNCTION_SCOPE, function_scope->scope_type());
parsing_module_ = parsing_module;
log_ = log;
use_counts_ = use_counts;
DCHECK(!track_unresolved_variables_);
track_unresolved_variables_ = is_inner_function;
// The caller passes the function_scope which is not yet inserted into the
// scope_state_. All scopes above the function_scope are ignored by the
// PreParser.
DCHECK_NULL(scope_state_);
FunctionState function_state(&function_state_, &scope_state_, function_scope);
DCHECK_EQ(Token::LBRACE, scanner()->current_token());
bool ok = true;
int start_position = peek_position();
LazyParsingResult result = ParseLazyFunctionLiteralBody(may_abort, &ok);
use_counts_ = nullptr;
track_unresolved_variables_ = false;
if (result == kLazyParsingAborted) {
return kPreParseAbort;
} else if (stack_overflow()) {
return kPreParseStackOverflow;
} else if (!ok) {
ReportUnexpectedToken(scanner()->current_token());
} else {
DCHECK_EQ(Token::RBRACE, scanner()->peek());
if (is_strict(function_scope->language_mode())) {
int end_pos = scanner()->location().end_pos;
CheckStrictOctalLiteral(start_position, end_pos, &ok);
CheckDecimalLiteralWithLeadingZero(start_position, end_pos);
}
}
return kPreParseSuccess;
}
// Preparsing checks a JavaScript program and emits preparse-data that helps
// a later parsing to be faster.
// See preparser-data.h for the data.
// The PreParser checks that the syntax follows the grammar for JavaScript,
// and collects some information about the program along the way.
// The grammar check is only performed in order to understand the program
// sufficiently to deduce some information about it, that can be used
// to speed up later parsing. Finding errors is not the goal of pre-parsing,
// rather it is to speed up properly written and correct programs.
// That means that contextual checks (like a label being declared where
// it is used) are generally omitted.
PreParser::Statement PreParser::ParseClassDeclaration(
ZoneList<const AstRawString*>* names, bool default_export, bool* ok) {
int pos = position();
bool is_strict_reserved = false;
Identifier name =
ParseIdentifierOrStrictReservedWord(&is_strict_reserved, CHECK_OK);
ExpressionClassifier no_classifier(this);
ParseClassLiteral(name, scanner()->location(), is_strict_reserved, pos,
CHECK_OK);
return Statement::Default();
}
PreParser::Statement PreParser::ParseFunctionDeclaration(bool* ok) {
Consume(Token::FUNCTION);
int pos = position();
ParseFunctionFlags flags = ParseFunctionFlags::kIsNormal;
if (Check(Token::MUL)) {
flags |= ParseFunctionFlags::kIsGenerator;
if (allow_harmony_restrictive_declarations()) {
ReportMessageAt(scanner()->location(),
MessageTemplate::kGeneratorInLegacyContext);
*ok = false;
return Statement::Default();
}
}
// PreParser is not able to parse "export default" yet (since PreParser is
// at the moment only used for functions, and it cannot occur
// there). TODO(marja): update this when it is.
return ParseHoistableDeclaration(pos, flags, nullptr, false, ok);
}
// Redefinition of CHECK_OK for parsing expressions.
#undef CHECK_OK
#define CHECK_OK CHECK_OK_VALUE(Expression::Default())
PreParser::Expression PreParser::ParseFunctionLiteral(
Identifier function_name, Scanner::Location function_name_location,
FunctionNameValidity function_name_validity, FunctionKind kind,
int function_token_pos, FunctionLiteral::FunctionType function_type,
LanguageMode language_mode, bool* ok) {
// Function ::
// '(' FormalParameterList? ')' '{' FunctionBody '}'
// Parse function body.
PreParserStatementList body;
bool outer_is_script_scope = scope()->is_script_scope();
DeclarationScope* function_scope = NewFunctionScope(kind);
function_scope->SetLanguageMode(language_mode);
FunctionState function_state(&function_state_, &scope_state_, function_scope);
DuplicateFinder duplicate_finder(scanner()->unicode_cache());
ExpressionClassifier formals_classifier(this, &duplicate_finder);
Expect(Token::LPAREN, CHECK_OK);
int start_position = scanner()->location().beg_pos;
function_scope->set_start_position(start_position);
PreParserFormalParameters formals(function_scope);
ParseFormalParameterList(&formals, CHECK_OK);
Expect(Token::RPAREN, CHECK_OK);
int formals_end_position = scanner()->location().end_pos;
CheckArityRestrictions(formals.arity, kind, formals.has_rest, start_position,
formals_end_position, CHECK_OK);
// See Parser::ParseFunctionLiteral for more information about lazy parsing
// and lazy compilation.
bool is_lazily_parsed = (outer_is_script_scope && allow_lazy() &&
!function_state_->this_function_is_parenthesized());
Expect(Token::LBRACE, CHECK_OK);
if (is_lazily_parsed) {
ParseLazyFunctionLiteralBody(false, CHECK_OK);
} else {
ParseStatementList(body, Token::RBRACE, CHECK_OK);
}
Expect(Token::RBRACE, CHECK_OK);
// Parsing the body may change the language mode in our scope.
language_mode = function_scope->language_mode();
// Validate name and parameter names. We can do this only after parsing the
// function, since the function can declare itself strict.
CheckFunctionName(language_mode, function_name, function_name_validity,
function_name_location, CHECK_OK);
const bool allow_duplicate_parameters =
is_sloppy(language_mode) && formals.is_simple && !IsConciseMethod(kind);
ValidateFormalParameters(language_mode, allow_duplicate_parameters, CHECK_OK);
if (is_strict(language_mode)) {
int end_position = scanner()->location().end_pos;
CheckStrictOctalLiteral(start_position, end_position, CHECK_OK);
CheckDecimalLiteralWithLeadingZero(start_position, end_position);
}
return Expression::Default();
}
PreParser::LazyParsingResult PreParser::ParseLazyFunctionLiteralBody(
bool may_abort, bool* ok) {
int body_start = position();
PreParserStatementList body;
LazyParsingResult result = ParseStatementList(
body, Token::RBRACE, may_abort, CHECK_OK_VALUE(kLazyParsingComplete));
if (result == kLazyParsingAborted) return result;
// Position right after terminal '}'.
DCHECK_EQ(Token::RBRACE, scanner()->peek());
int body_end = scanner()->peek_location().end_pos;
DeclarationScope* scope = this->scope()->AsDeclarationScope();
DCHECK(scope->is_function_scope());
log_->LogFunction(body_start, body_end,
function_state_->materialized_literal_count(),
function_state_->expected_property_count(), language_mode(),
scope->uses_super_property(), scope->calls_eval());
return kLazyParsingComplete;
}
PreParserExpression PreParser::ParseClassLiteral(
PreParserIdentifier name, Scanner::Location class_name_location,
bool name_is_strict_reserved, int pos, bool* ok) {
// All parts of a ClassDeclaration and ClassExpression are strict code.
if (name_is_strict_reserved) {
ReportMessageAt(class_name_location,
MessageTemplate::kUnexpectedStrictReserved);
*ok = false;
return EmptyExpression();
}
if (IsEvalOrArguments(name)) {
ReportMessageAt(class_name_location, MessageTemplate::kStrictEvalArguments);
*ok = false;
return EmptyExpression();
}
LanguageMode class_language_mode = language_mode();
BlockState block_state(zone(), &scope_state_);
scope()->SetLanguageMode(
static_cast<LanguageMode>(class_language_mode | STRICT));
// TODO(marja): Make PreParser use scope names too.
// this->scope()->SetScopeName(name);
bool has_extends = Check(Token::EXTENDS);
if (has_extends) {
ExpressionClassifier extends_classifier(this);
ParseLeftHandSideExpression(CHECK_OK);
CheckNoTailCallExpressions(CHECK_OK);
ValidateExpression(CHECK_OK);
impl()->AccumulateFormalParameterContainmentErrors();
}
ClassLiteralChecker checker(this);
bool has_seen_constructor = false;
Expect(Token::LBRACE, CHECK_OK);
while (peek() != Token::RBRACE) {
if (Check(Token::SEMICOLON)) continue;
bool is_computed_name = false; // Classes do not care about computed
// property names here.
ExpressionClassifier property_classifier(this);
ParseClassPropertyDefinition(&checker, has_extends, &is_computed_name,
&has_seen_constructor, CHECK_OK);
ValidateExpression(CHECK_OK);
impl()->AccumulateFormalParameterContainmentErrors();
}
Expect(Token::RBRACE, CHECK_OK);
return Expression::Default();
}
PreParserExpression PreParser::ExpressionFromIdentifier(
PreParserIdentifier name, int start_position, int end_position,
InferName infer) {
if (track_unresolved_variables_) {
AstNodeFactory factory(ast_value_factory());
// Setting the Zone is necessary because zone_ might be the temp Zone, and
// AstValueFactory doesn't know about it.
factory.set_zone(zone());
DCHECK_NOT_NULL(name.string_);
scope()->NewUnresolved(&factory, name.string_, start_position, end_position,
NORMAL_VARIABLE);
}
return PreParserExpression::FromIdentifier(name);
}
#undef CHECK_OK
#undef CHECK_OK_CUSTOM
} // namespace internal
} // namespace v8