LibWeb/CSS: Make @page selector parsing accessible

Mostly minor changes - the code has moved, and has to support Token or
ComponentValue TokenStreams, but otherwise it's the same.
This commit is contained in:
Sam Atkins 2025-05-16 11:37:53 +01:00
commit d852ae17e8
Notes: github-actions[bot] 2025-05-16 15:45:11 +00:00
4 changed files with 87 additions and 64 deletions

View file

@ -106,6 +106,11 @@ Optional<CSS::SelectorList> parse_selector_for_nested_style_rule(CSS::Parser::Pa
return adapt_nested_relative_selector_list(*maybe_selectors); return adapt_nested_relative_selector_list(*maybe_selectors);
} }
Optional<CSS::PageSelectorList> parse_page_selector_list(CSS::Parser::ParsingParams const& params, StringView selector_text)
{
return CSS::Parser::Parser::create(params, selector_text).parse_as_page_selector_list();
}
Optional<CSS::Selector::PseudoElementSelector> parse_pseudo_element_selector(CSS::Parser::ParsingParams const& context, StringView selector_text) Optional<CSS::Selector::PseudoElementSelector> parse_pseudo_element_selector(CSS::Parser::ParsingParams const& context, StringView selector_text)
{ {
return CSS::Parser::Parser::create(context, selector_text).parse_as_pseudo_element_selector(); return CSS::Parser::Parser::create(context, selector_text).parse_as_pseudo_element_selector();

View file

@ -19,6 +19,7 @@
#include <LibWeb/CSS/Descriptor.h> #include <LibWeb/CSS/Descriptor.h>
#include <LibWeb/CSS/DescriptorID.h> #include <LibWeb/CSS/DescriptorID.h>
#include <LibWeb/CSS/MediaQuery.h> #include <LibWeb/CSS/MediaQuery.h>
#include <LibWeb/CSS/PageSelector.h>
#include <LibWeb/CSS/ParsedFontFace.h> #include <LibWeb/CSS/ParsedFontFace.h>
#include <LibWeb/CSS/Parser/ComponentValue.h> #include <LibWeb/CSS/Parser/ComponentValue.h>
#include <LibWeb/CSS/Parser/Dimension.h> #include <LibWeb/CSS/Parser/Dimension.h>
@ -117,6 +118,8 @@ public:
Optional<Selector::PseudoElementSelector> parse_as_pseudo_element_selector(); Optional<Selector::PseudoElementSelector> parse_as_pseudo_element_selector();
Optional<PageSelectorList> parse_as_page_selector_list();
Vector<NonnullRefPtr<MediaQuery>> parse_as_media_query_list(); Vector<NonnullRefPtr<MediaQuery>> parse_as_media_query_list();
RefPtr<MediaQuery> parse_as_media_query(); RefPtr<MediaQuery> parse_as_media_query();
@ -188,6 +191,9 @@ private:
template<typename T> template<typename T>
ParseErrorOr<SelectorList> parse_a_selector_list(TokenStream<T>&, SelectorType, SelectorParsingMode = SelectorParsingMode::Standard); ParseErrorOr<SelectorList> parse_a_selector_list(TokenStream<T>&, SelectorType, SelectorParsingMode = SelectorParsingMode::Standard);
template<typename T>
ParseErrorOr<PageSelectorList> parse_a_page_selector_list(TokenStream<T>&);
template<typename T> template<typename T>
Vector<NonnullRefPtr<MediaQuery>> parse_a_media_query_list(TokenStream<T>&); Vector<NonnullRefPtr<MediaQuery>> parse_a_media_query_list(TokenStream<T>&);
template<typename T> template<typename T>
@ -523,6 +529,7 @@ RefPtr<CSS::CSSStyleValue const> parse_css_value(CSS::Parser::ParsingParams cons
RefPtr<CSS::CSSStyleValue const> parse_css_descriptor(CSS::Parser::ParsingParams const&, CSS::AtRuleID, CSS::DescriptorID, StringView); RefPtr<CSS::CSSStyleValue const> parse_css_descriptor(CSS::Parser::ParsingParams const&, CSS::AtRuleID, CSS::DescriptorID, StringView);
Optional<CSS::SelectorList> parse_selector(CSS::Parser::ParsingParams const&, StringView); Optional<CSS::SelectorList> parse_selector(CSS::Parser::ParsingParams const&, StringView);
Optional<CSS::SelectorList> parse_selector_for_nested_style_rule(CSS::Parser::ParsingParams const&, StringView); Optional<CSS::SelectorList> parse_selector_for_nested_style_rule(CSS::Parser::ParsingParams const&, StringView);
Optional<CSS::PageSelectorList> parse_page_selector_list(CSS::Parser::ParsingParams const&, StringView);
Optional<CSS::Selector::PseudoElementSelector> parse_pseudo_element_selector(CSS::Parser::ParsingParams const&, StringView); Optional<CSS::Selector::PseudoElementSelector> parse_pseudo_element_selector(CSS::Parser::ParsingParams const&, StringView);
CSS::CSSRule* parse_css_rule(CSS::Parser::ParsingParams const&, StringView); CSS::CSSRule* parse_css_rule(CSS::Parser::ParsingParams const&, StringView);
RefPtr<CSS::MediaQuery> parse_media_query(CSS::Parser::ParsingParams const&, StringView); RefPtr<CSS::MediaQuery> parse_media_query(CSS::Parser::ParsingParams const&, StringView);

View file

@ -639,74 +639,13 @@ GC::Ptr<CSSFontFaceRule> Parser::convert_to_font_face_rule(AtRule const& rule)
return CSSFontFaceRule::create(realm(), CSSFontFaceDescriptors::create(realm(), descriptors.release_descriptors())); return CSSFontFaceRule::create(realm(), CSSFontFaceDescriptors::create(realm(), descriptors.release_descriptors()));
} }
static Optional<PageSelectorList> parse_page_selector_list(Vector<ComponentValue> const& component_values)
{
// https://drafts.csswg.org/css-page-3/#syntax-page-selector
// <page-selector-list> = <page-selector>#
// <page-selector> = [ <ident-token>? <pseudo-page>* ]!
// <pseudo-page> = : [ left | right | first | blank ]
PageSelectorList selector_list;
TokenStream tokens { component_values };
tokens.discard_whitespace();
while (tokens.has_next_token()) {
// First optional ident
Optional<FlyString> maybe_ident;
if (tokens.next_token().is(Token::Type::Ident))
maybe_ident = tokens.consume_a_token().token().ident();
// Then an optional series of pseudo-classes
Vector<PagePseudoClass> pseudo_classes;
while (tokens.next_token().is(Token::Type::Colon)) {
tokens.discard_a_token(); // :
if (!tokens.next_token().is(Token::Type::Ident)) {
dbgln_if(CSS_PARSER_DEBUG, "Invalid @page selector: pseudo-class is not an ident: `{}`", tokens.next_token().to_debug_string());
return {};
}
auto pseudo_class_name = tokens.consume_a_token().token().ident();
if (auto pseudo_class = page_pseudo_class_from_string(pseudo_class_name); pseudo_class.has_value()) {
pseudo_classes.append(*pseudo_class);
} else {
dbgln_if(CSS_PARSER_DEBUG, "Invalid @page selector: unrecognized pseudo-class `:{}`", pseudo_class_name);
return {};
}
}
if (!maybe_ident.has_value() && pseudo_classes.is_empty()) {
// Nothing parsed
dbgln_if(CSS_PARSER_DEBUG, "Invalid @page selector: is empty");
return {};
}
selector_list.empend(move(maybe_ident), move(pseudo_classes));
tokens.discard_whitespace();
if (tokens.next_token().is(Token::Type::Comma)) {
tokens.discard_a_token(); // ,
tokens.discard_whitespace();
if (!tokens.has_next_token()) {
dbgln_if(CSS_PARSER_DEBUG, "Invalid @page selector: trailing comma");
return {};
}
} else if (tokens.has_next_token()) {
dbgln_if(CSS_PARSER_DEBUG, "Invalid @page selector: trailing token `{}`", tokens.next_token().to_debug_string());
return {};
}
}
return selector_list;
}
GC::Ptr<CSSPageRule> Parser::convert_to_page_rule(AtRule const& page_rule) GC::Ptr<CSSPageRule> Parser::convert_to_page_rule(AtRule const& page_rule)
{ {
// https://drafts.csswg.org/css-page-3/#syntax-page-selector // https://drafts.csswg.org/css-page-3/#syntax-page-selector
// @page = @page <page-selector-list>? { <declaration-rule-list> } // @page = @page <page-selector-list>? { <declaration-rule-list> }
auto page_selectors = parse_page_selector_list(page_rule.prelude); TokenStream tokens { page_rule.prelude };
if (!page_selectors.has_value()) auto page_selectors = parse_a_page_selector_list(tokens);
if (page_selectors.is_error())
return nullptr; return nullptr;
GC::RootVector<GC::Ref<CSSRule>> child_rules { realm().heap() }; GC::RootVector<GC::Ref<CSSRule>> child_rules { realm().heap() };

View file

@ -1147,4 +1147,76 @@ Optional<Selector::SimpleSelector::ANPlusBPattern> Parser::parse_a_n_plus_b_patt
return syntax_error(); return syntax_error();
} }
Optional<PageSelectorList> Parser::parse_as_page_selector_list()
{
auto selector_list = parse_a_page_selector_list(m_token_stream);
if (!selector_list.is_error())
return selector_list.release_value();
return {};
}
template<typename T>
Parser::ParseErrorOr<PageSelectorList> Parser::parse_a_page_selector_list(TokenStream<T>& tokens)
{
// https://drafts.csswg.org/css-page-3/#syntax-page-selector
// <page-selector-list> = <page-selector>#
// <page-selector> = [ <ident-token>? <pseudo-page>* ]!
// <pseudo-page> = : [ left | right | first | blank ]
PageSelectorList selector_list;
tokens.discard_whitespace();
while (tokens.has_next_token()) {
// First optional ident
Optional<FlyString> maybe_ident;
if (tokens.next_token().is(Token::Type::Ident))
maybe_ident = static_cast<Token>(tokens.consume_a_token()).ident();
// Then an optional series of pseudo-classes
Vector<PagePseudoClass> pseudo_classes;
while (tokens.next_token().is(Token::Type::Colon)) {
tokens.discard_a_token(); // :
if (!tokens.next_token().is(Token::Type::Ident)) {
dbgln_if(CSS_PARSER_DEBUG, "Invalid @page selector: pseudo-class is not an ident: `{}`", tokens.next_token().to_debug_string());
return ParseError::SyntaxError;
}
auto pseudo_class_name = static_cast<Token>(tokens.consume_a_token()).ident();
if (auto pseudo_class = page_pseudo_class_from_string(pseudo_class_name); pseudo_class.has_value()) {
pseudo_classes.append(*pseudo_class);
} else {
dbgln_if(CSS_PARSER_DEBUG, "Invalid @page selector: unrecognized pseudo-class `:{}`", pseudo_class_name);
return ParseError::SyntaxError;
}
}
if (!maybe_ident.has_value() && pseudo_classes.is_empty()) {
// Nothing parsed
dbgln_if(CSS_PARSER_DEBUG, "Invalid @page selector: is empty");
return ParseError::SyntaxError;
}
selector_list.empend(move(maybe_ident), move(pseudo_classes));
tokens.discard_whitespace();
if (tokens.next_token().is(Token::Type::Comma)) {
tokens.discard_a_token(); // ,
tokens.discard_whitespace();
if (!tokens.has_next_token()) {
dbgln_if(CSS_PARSER_DEBUG, "Invalid @page selector: trailing comma");
return ParseError::SyntaxError;
}
} else if (tokens.has_next_token()) {
dbgln_if(CSS_PARSER_DEBUG, "Invalid @page selector: trailing token `{}`", tokens.next_token().to_debug_string());
return ParseError::SyntaxError;
}
}
return selector_list;
}
template Parser::ParseErrorOr<PageSelectorList> Parser::parse_a_page_selector_list(TokenStream<ComponentValue>&);
template Parser::ParseErrorOr<PageSelectorList> Parser::parse_a_page_selector_list(TokenStream<Token>&);
} }