mirrors/ladybird
0
0
Fork 0
mirror of https://github.com/LadybirdBrowser/ladybird.git synced 2025-05-02 09:18:52 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 9

LibWeb: Handle iso-8859-1 web content a little bit better

Browse source 
We now look at the HTTP response headers for a Content-Type header and
try to parse it if present to find the text encoding.

If the text encoding is iso-8859-1, we turn all non-ASCII characters
into question marks. This makes Swedish Google load on my machine! :^)
This commit is contained in:
Andreas Kling 2020-05-03 22:22:05 +02:00
parent eb6e35a1be
commit f3676ebef5
Notes: sideshowbarker 2024-07-19 07:00:27 +09:00
3 changed files with 44 additions and 6 deletions
Download patch file Download diff file Expand all files Collapse all files

27
Libraries/LibWeb/Parser/HTMLParser.cpp
View file

@ -383,16 +383,37 @@ static bool parse_html_document(const StringView& html, Document& document, Pare
return true;
}
RefPtr<DocumentFragment> parse_html_fragment(Document& document, const StringView& html)
String to_utf8(const StringView& input, const String& encoding)
{
String output;
if (encoding == "utf-8") {
output = input;
} else if (encoding == "iso-8859-1") {
StringBuilder builder(input.length());
for (size_t i = 0; i < input.length(); ++i) {
u8 ch = input[i];
builder.append(ch >= 0x80 ? '?' : ch);
}
output = builder.to_string();
} else {
dbg() << "Unknown encoding " << encoding;
ASSERT_NOT_REACHED();
}
return output;
}
RefPtr<DocumentFragment> parse_html_fragment(Document& document, const StringView& raw_html, const String& encoding)
{
auto fragment = adopt(*new DocumentFragment(document));
if (!parse_html_document(html, document, *fragment))
if (!parse_html_document(to_utf8(raw_html, encoding), document, *fragment))
return nullptr;
return fragment;
}
RefPtr<Document> parse_html_document(const StringView& html, const URL& url)
RefPtr<Document> parse_html_document(const StringView& raw_html, const URL& url, const String& encoding)
{
String html = to_utf8(raw_html, encoding);
auto document = adopt(*new Document(url));
document->set_source(html);