RGBCube/serenity
1
Fork 0
mirror of https://github.com/RGBCube/serenity synced 2025-05-31 11:28:12 +00:00
Code Issues Activity Actions

LibWeb: Fix parsing of character references in attribute values

Browse source
This commit is contained in:
ovf 2021-07-24 23:15:47 +01:00 • committed by Andreas Kling
parent 0505a815f7
commit 13c7d55320
2 changed files with 23 additions and 13 deletions
Download patch file Download diff file Expand all files Collapse all files

27
Tests/LibWeb/TestHTMLTokenizer.cpp
View file

@ -57,7 +57,7 @@ using Token = Web::HTML::HTMLToken;
#define EXPECT_TAG_TOKEN_ATTRIBUTE(name, value) \ #define EXPECT_TAG_TOKEN_ATTRIBUTE(name, value) \
VERIFY(last_token); \ VERIFY(last_token); \
EXPECT_EQ(last_token->attribute(#name), #value); EXPECT_EQ(last_token->attribute(#name), value);
#define EXPECT_TAG_TOKEN_ATTRIBUTE_COUNT(count) \ #define EXPECT_TAG_TOKEN_ATTRIBUTE_COUNT(count) \
VERIFY(last_token); \ VERIFY(last_token); \
@ -124,7 +124,7 @@ TEST_CASE(unquoted_attributes)
BEGIN_ENUMERATION(tokens); BEGIN_ENUMERATION(tokens);
EXPECT_START_TAG_TOKEN(p); EXPECT_START_TAG_TOKEN(p);
EXPECT_TAG_TOKEN_ATTRIBUTE_COUNT(1); EXPECT_TAG_TOKEN_ATTRIBUTE_COUNT(1);
EXPECT_TAG_TOKEN_ATTRIBUTE(foo, bar); EXPECT_TAG_TOKEN_ATTRIBUTE(foo, "bar");
EXPECT_END_OF_FILE_TOKEN(); EXPECT_END_OF_FILE_TOKEN();
END_ENUMERATION(); END_ENUMERATION();
} }
@ -135,7 +135,7 @@ TEST_CASE(single_quoted_attributes)
BEGIN_ENUMERATION(tokens); BEGIN_ENUMERATION(tokens);
EXPECT_START_TAG_TOKEN(p); EXPECT_START_TAG_TOKEN(p);
EXPECT_TAG_TOKEN_ATTRIBUTE_COUNT(1); EXPECT_TAG_TOKEN_ATTRIBUTE_COUNT(1);
EXPECT_TAG_TOKEN_ATTRIBUTE(foo, bar); EXPECT_TAG_TOKEN_ATTRIBUTE(foo, "bar");
EXPECT_END_OF_FILE_TOKEN(); EXPECT_END_OF_FILE_TOKEN();
END_ENUMERATION(); END_ENUMERATION();
} }
@ -146,7 +146,7 @@ TEST_CASE(double_quoted_attributes)
BEGIN_ENUMERATION(tokens); BEGIN_ENUMERATION(tokens);
EXPECT_START_TAG_TOKEN(p); EXPECT_START_TAG_TOKEN(p);
EXPECT_TAG_TOKEN_ATTRIBUTE_COUNT(1); EXPECT_TAG_TOKEN_ATTRIBUTE_COUNT(1);
EXPECT_TAG_TOKEN_ATTRIBUTE(foo, bar); EXPECT_TAG_TOKEN_ATTRIBUTE(foo, "bar");
EXPECT_END_OF_FILE_TOKEN(); EXPECT_END_OF_FILE_TOKEN();
END_ENUMERATION(); END_ENUMERATION();
} }
@ -157,9 +157,22 @@ TEST_CASE(multiple_attributes)
BEGIN_ENUMERATION(tokens); BEGIN_ENUMERATION(tokens);
EXPECT_START_TAG_TOKEN(p); EXPECT_START_TAG_TOKEN(p);
EXPECT_TAG_TOKEN_ATTRIBUTE_COUNT(3); EXPECT_TAG_TOKEN_ATTRIBUTE_COUNT(3);
EXPECT_TAG_TOKEN_ATTRIBUTE(foo, bar); EXPECT_TAG_TOKEN_ATTRIBUTE(foo, "bar");
EXPECT_TAG_TOKEN_ATTRIBUTE(baz, foobar); EXPECT_TAG_TOKEN_ATTRIBUTE(baz, "foobar");
EXPECT_TAG_TOKEN_ATTRIBUTE(foo2, bar2); EXPECT_TAG_TOKEN_ATTRIBUTE(foo2, "bar2");
EXPECT_END_OF_FILE_TOKEN();
END_ENUMERATION();
}
TEST_CASE(character_reference_in_attribute)
{
auto tokens = run_tokenizer("<p foo=a&amp;b bar='a&amp;b' baz=\"a&amp;b\">");
BEGIN_ENUMERATION(tokens);
EXPECT_START_TAG_TOKEN(p);
EXPECT_TAG_TOKEN_ATTRIBUTE_COUNT(3);
EXPECT_TAG_TOKEN_ATTRIBUTE(foo, "a&b");
EXPECT_TAG_TOKEN_ATTRIBUTE(bar, "a&b");
EXPECT_TAG_TOKEN_ATTRIBUTE(baz, "a&b");
EXPECT_END_OF_FILE_TOKEN(); EXPECT_END_OF_FILE_TOKEN();
END_ENUMERATION(); END_ENUMERATION();
} }

9
Userland/Libraries/LibWeb/HTML/Parser/HTMLTokenizer.cpp
View file

@ -1179,9 +1179,8 @@ _StartOfFunction:
} }
ON('&') ON('&')
{ {
m_current_token.last_attribute().value = consume_current_builder();
m_return_state = State::AttributeValueDoubleQuoted; m_return_state = State::AttributeValueDoubleQuoted;
SWITCH_TO(CharacterReference); SWITCH_TO_WITH_UNCLEAN_BUILDER(CharacterReference);
} }
ON(0) ON(0)
{ {
@ -1211,9 +1210,8 @@ _StartOfFunction:
} }
ON('&') ON('&')
{ {
m_current_token.last_attribute().value = consume_current_builder();
m_return_state = State::AttributeValueSingleQuoted; m_return_state = State::AttributeValueSingleQuoted;
SWITCH_TO(CharacterReference); SWITCH_TO_WITH_UNCLEAN_BUILDER(CharacterReference);
} }
ON(0) ON(0)
{ {
@ -1244,9 +1242,8 @@ _StartOfFunction:
} }
ON('&') ON('&')
{ {
m_current_token.last_attribute().value = consume_current_builder();
m_return_state = State::AttributeValueUnquoted; m_return_state = State::AttributeValueUnquoted;
SWITCH_TO(CharacterReference); SWITCH_TO_WITH_UNCLEAN_BUILDER(CharacterReference);
} }
ON('>') ON('>')
{ {