RGBCube/serenity
1
Fork 0
mirror of https://github.com/RGBCube/serenity synced 2025-05-31 12:48:10 +00:00
Code Issues Activity Actions

LibHTML: Parse HTML escape sequences

Browse source
This commit is contained in:
Sergey Bugaev 2019-09-25 12:42:56 +03:00 • committed by Andreas Kling
parent 6491493e26
commit db8a1a6aba
1 changed files with 29 additions and 5 deletions
Download patch file Download diff file Expand all files Collapse all files

34
Libraries/LibHTML/Parser/HTMLParser.cpp
View file

@ -1,4 +1,5 @@
#include <AK/NonnullRefPtrVector.h> #include <AK/NonnullRefPtrVector.h>
#include <AK/StringBuilder.h>
#include <LibHTML/DOM/Element.h> #include <LibHTML/DOM/Element.h>
#include <LibHTML/DOM/Text.h> #include <LibHTML/DOM/Text.h>
#include <LibHTML/Parser/HTMLParser.h> #include <LibHTML/Parser/HTMLParser.h>
@ -54,7 +55,7 @@ NonnullRefPtr<Document> parse_html(const String& html)
auto state = State::Free; auto state = State::Free;
Vector<char, 256> text_buffer; StringBuilder text_buffer;
Vector<char, 32> tag_name_buffer; Vector<char, 32> tag_name_buffer;
@ -74,9 +75,8 @@ NonnullRefPtr<Document> parse_html(const String& html)
attribute_name_buffer.clear(); attribute_name_buffer.clear();
if (new_state == State::BeforeAttributeValue) if (new_state == State::BeforeAttributeValue)
attribute_value_buffer.clear(); attribute_value_buffer.clear();
if (state == State::Free && !text_buffer.is_empty()) { if (state == State::Free && !text_buffer.string_view().is_empty()) {
auto text_node = adopt(*new Text(String::copy(text_buffer))); auto text_node = adopt(*new Text(text_buffer.to_string()));
text_buffer.clear();
node_stack.last().append_child(text_node); node_stack.last().append_child(text_node);
} }
state = new_state; state = new_state;
@ -120,7 +120,31 @@ NonnullRefPtr<Document> parse_html(const String& html)
move_to_state(State::BeforeTagName); move_to_state(State::BeforeTagName);
break; break;
} }
text_buffer.append(ch); if (ch != '&') {
text_buffer.append(ch);
} else {
struct Escape {
const char* code;
const char* value;
};
static Escape escapes[] = {
{ "&lt;", "<" },
{ "&gt;", ">" },
{ "&amp;", "&" }
};
auto rest_of_html = html.substring_view(i, html.length() - i);
bool found = false;
for (auto& escape : escapes) {
if (rest_of_html.starts_with(escape.code)) {
text_buffer.append(escape.value);
found = true;
i += strlen(escape.code) - 1;
break;
}
}
if (!found)
dbg() << "Unhandled escape sequence";
}
break; break;
case State::BeforeTagName: case State::BeforeTagName:
if (ch == '/') { if (ch == '/') {