TEST(HTMLTreeBuilderSimulatorTest, SelfClosingMathFollowedByScript) {
HTMLParserOptions options;
HTMLTreeBuilderSimulator simulator(options);
std::unique_ptr<HTMLTokenizer> tokenizer = HTMLTokenizer::create(options);
SegmentedString input("<math/><script></script>");
HTMLToken token;
EXPECT_TRUE(tokenizer->nextToken(input, token));
EXPECT_EQ(HTMLTreeBuilderSimulator::OtherToken,
simulator.simulate(CompactHTMLToken(&token, TextPosition()),
tokenizer.get()));
token.clear();
EXPECT_TRUE(tokenizer->nextToken(input, token));
EXPECT_EQ(HTMLTreeBuilderSimulator::ScriptStart,
simulator.simulate(CompactHTMLToken(&token, TextPosition()),
tokenizer.get()));
EXPECT_EQ(HTMLTokenizer::ScriptDataState, tokenizer->getState());
token.clear();
EXPECT_TRUE(tokenizer->nextToken(input, token));
EXPECT_EQ(HTMLTreeBuilderSimulator::ScriptEnd,
simulator.simulate(CompactHTMLToken(&token, TextPosition()),
tokenizer.get()));
}
void HTMLDocumentParser::constructTreeFromHTMLToken(HTMLToken& rawToken)
{
RefPtr<AtomicHTMLToken> token = AtomicHTMLToken::create(rawToken);
// We clear the rawToken in case constructTreeFromAtomicToken
// synchronously re-enters the parser. We don't clear the token immedately
// for Character tokens because the AtomicHTMLToken avoids copying the
// characters by keeping a pointer to the underlying buffer in the
// HTMLToken. Fortunately, Character tokens can't cause us to re-enter
// the parser.
//
// FIXME: Stop clearing the rawToken once we start running the parser off
// the main thread or once we stop allowing synchronous JavaScript
// execution from parseAttribute.
if (rawToken.type() != HTMLTokenTypes::Character)
rawToken.clear();
m_treeBuilder->constructTree(token.get());
// AtomicHTMLToken keeps a pointer to the HTMLToken's buffer instead
// of copying the characters for performance.
// Clear the external characters pointer before the raw token is cleared
// to make sure that we won't have a dangling pointer.
token->clearExternalCharacters();
if (!rawToken.isUninitialized()) {
ASSERT(rawToken.type() == HTMLTokenTypes::Character);
rawToken.clear();
}
}
void HTMLDocumentParser::constructTreeFromHTMLToken(HTMLToken& rawToken)
{
AtomicHTMLToken token(rawToken);
// We clear the rawToken in case constructTreeFromAtomicToken
// synchronously re-enters the parser. We don't clear the token immedately
// for Character tokens because the AtomicHTMLToken avoids copying the
// characters by keeping a pointer to the underlying buffer in the
// HTMLToken. Fortunately, Character tokens can't cause us to re-enter
// the parser.
//
// FIXME: Stop clearing the rawToken once we start running the parser off
// the main thread or once we stop allowing synchronous JavaScript
// execution from parseAttribute.
if (rawToken.type() != HTMLToken::Character)
rawToken.clear();
m_treeBuilder->constructTree(&token);
if (!rawToken.isUninitialized()) {
ASSERT(rawToken.type() == HTMLToken::Character);
rawToken.clear();
}
}