dmaax/ladybird
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

LibWeb: Explicitly disallow skipping 0 characters in HTML tokenizer

Browse source 
Previously, this caused the source position of an ambiguous ampersand
to be incorrect.
This commit is contained in:
Tim Ledbetter 2025-11-06 07:23:01 +00:00 committed by Andreas Kling
parent 9037f4e832
commit 478cf4f42c
2 changed files with 11 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

3
Libraries/LibWeb/HTML/Parser/HTMLTokenizer.cpp
View file

@ -220,6 +220,7 @@ Optional<u32> HTMLTokenizer::next_code_point(StopAtInsertionPoint stop_at_insert
void HTMLTokenizer::skip(size_t count)
{
VERIFY(count > 0);
if (!m_source_positions.is_empty())
m_source_positions.append(m_source_positions.last());
for (size_t i = 0; i < count; ++i) {
@ -1725,7 +1726,7 @@ _StartOfFunction:
auto num_consumed = m_temporary_buffer.size() - starting_consumed_count;
if (num_consumed == 0) {
DONT_CONSUME_NEXT_INPUT_CHARACTER;
} else {
} else if (num_consumed > 1) {
skip(num_consumed - 1);
}
}

9
Tests/LibWeb/TestHTMLTokenizer.cpp
View file

@ -266,3 +266,12 @@ TEST_CASE(regression)
u32 hash = hash_tokens(tokens);
EXPECT_EQ(hash, 3657343287u);
}
TEST_CASE(ambiguous_ampersand_offset)
{
auto tokens = run_tokenizer("&a"sv);
auto& token = tokens.first();
EXPECT_EQ(token.type(), Token::Type::Character);
EXPECT_EQ(token.start_position().line, 0u);
EXPECT_EQ(token.start_position().column, 1u);
}