mirrors/ladybird
0
0
Fork 0
mirror of https://github.com/LadybirdBrowser/ladybird.git synced 2025-05-04 10:18:51 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 10

LibWeb: Remove most uses of the old HTML parser

Browse source 
The only remaining client of the old parser is the fragment parser used
by the Element.innerHTML setter. We'll need to implement a bit more
stuff in the new parser before we can switch that over.
This commit is contained in:
Andreas Kling 2020-06-21 22:29:05 +02:00
parent c9d55e3b80
commit 07d976716f
Notes: sideshowbarker 2024-07-19 05:27:42 +09:00
5 changed files with 16 additions and 8 deletions
Download patch file Download diff file Expand all files Collapse all files

6
Libraries/LibWeb/Loader/FrameLoader.cpp
View file

@ -54,7 +54,7 @@ static RefPtr<Document> create_markdown_document(const ByteBuffer& data, const U
if (!markdown_document) if (!markdown_document)
return nullptr; return nullptr;
return parse_html_document(markdown_document->render_to_html(), url); return parse_html_document(markdown_document->render_to_html(), url, "utf-8");
} }
static RefPtr<Document> create_text_document(const ByteBuffer& data, const URL& url) static RefPtr<Document> create_text_document(const ByteBuffer& data, const URL& url)
@ -116,7 +116,7 @@ static RefPtr<Document> create_gemini_document(const ByteBuffer& data, const URL
{ {
auto markdown_document = Gemini::Document::parse({ (const char*)data.data(), data.size() }, url); auto markdown_document = Gemini::Document::parse({ (const char*)data.data(), data.size() }, url);
return parse_html_document(markdown_document->render_to_html(), url); return parse_html_document(markdown_document->render_to_html(), url, "utf-8");
} }
RefPtr<Document> FrameLoader::create_document_from_mime_type(const ByteBuffer& data, const URL& url, const String& mime_type, const String& encoding) RefPtr<Document> FrameLoader::create_document_from_mime_type(const ByteBuffer& data, const URL& url, const String& mime_type, const String& encoding)
@ -190,7 +190,7 @@ void FrameLoader::load_error_page(const URL& failed_url, const String& error)
String::copy(data).characters(), String::copy(data).characters(),
escape_html_entities(failed_url.to_string()).characters(), escape_html_entities(failed_url.to_string()).characters(),
escape_html_entities(error).characters()); escape_html_entities(error).characters());
auto document = parse_html_document(html, failed_url); auto document = parse_html_document(html, failed_url, "utf-8");
ASSERT(document); ASSERT(document);
frame().set_document(document); frame().set_document(document);
frame().page().client().page_did_change_title(document->title()); frame().page().client().page_did_change_title(document->title());

7
Libraries/LibWeb/Parser/HTMLDocumentParser.cpp
View file

@ -46,6 +46,13 @@
namespace Web { namespace Web {
RefPtr<Document> parse_html_document(const StringView& data, const URL& url, const String& encoding)
{
HTMLDocumentParser parser(data, encoding);
parser.run(url);
return parser.document();
}
HTMLDocumentParser::HTMLDocumentParser(const StringView& input, const String& encoding) HTMLDocumentParser::HTMLDocumentParser(const StringView& input, const String& encoding)
: m_tokenizer(input, encoding) : m_tokenizer(input, encoding)
{ {

2
Libraries/LibWeb/Parser/HTMLDocumentParser.h
View file

@ -59,6 +59,8 @@
namespace Web { namespace Web {
RefPtr<Document> parse_html_document(const StringView&, const URL&, const String& encoding);
class HTMLDocumentParser { class HTMLDocumentParser {
public: public:
HTMLDocumentParser(const StringView& input, const String& encoding); HTMLDocumentParser(const StringView& input, const String& encoding);

8
Libraries/LibWeb/Parser/HTMLParser.cpp
View file

@ -82,7 +82,7 @@ static Vector<char> codepoint_to_bytes(const u32 codepoint)
return bytes; return bytes;
} }
static bool parse_html_document(const StringView& html, Document& document, ParentNode& root) static bool deprecated_parse_html_document(const StringView& html, Document& document, ParentNode& root)
{ {
NonnullRefPtrVector<ParentNode> node_stack; NonnullRefPtrVector<ParentNode> node_stack;
node_stack.append(root); node_stack.append(root);
@ -466,19 +466,19 @@ String to_utf8(const StringView& input, const String& encoding)
RefPtr<DocumentFragment> parse_html_fragment(Document& document, const StringView& raw_html, const String& encoding) RefPtr<DocumentFragment> parse_html_fragment(Document& document, const StringView& raw_html, const String& encoding)
{ {
auto fragment = adopt(*new DocumentFragment(document)); auto fragment = adopt(*new DocumentFragment(document));
if (!parse_html_document(to_utf8(raw_html, encoding), document, *fragment)) if (!deprecated_parse_html_document(to_utf8(raw_html, encoding), document, *fragment))
return nullptr; return nullptr;
return fragment; return fragment;
} }
RefPtr<Document> parse_html_document(const StringView& raw_html, const URL& url, const String& encoding) RefPtr<Document> deprecated_parse_html_document(const StringView& raw_html, const URL& url, const String& encoding)
{ {
String html = to_utf8(raw_html, encoding); String html = to_utf8(raw_html, encoding);
auto document = adopt(*new Document(url)); auto document = adopt(*new Document(url));
document->set_source(html); document->set_source(html);
if (!parse_html_document(html, *document, *document)) if (!deprecated_parse_html_document(html, *document, *document))
return nullptr; return nullptr;
document->fixup(); document->fixup();

1
Libraries/LibWeb/Parser/HTMLParser.h
View file

@ -33,7 +33,6 @@ namespace Web {
class DocumentFragment; class DocumentFragment;
RefPtr<Document> parse_html_document(const StringView&, const URL& = URL(), const String& encoding = "utf-8");
RefPtr<DocumentFragment> parse_html_fragment(Document&, const StringView&, const String& encoding = "utf-8"); RefPtr<DocumentFragment> parse_html_fragment(Document&, const StringView&, const String& encoding = "utf-8");
} }