mirror of
https://github.com/RGBCube/serenity
synced 2025-07-26 04:57:44 +00:00
LibMarkdown: Rewrite Inline text parser to be more forgiving
The previous Text::parse was not able to give up on parsing a textual element, and just leave it as plain text. Because this is a very important part of markdown, I fully rewrote the parser to support this without having to backtrack. Also the parser now some other little features, such ast delimiter runs and flanking.
This commit is contained in:
Peter Elliott
Andreas Kling
parent
80e58dab9a
commit
ec9f892899
10 changed files with 462 additions and 397 deletions
|
|
@ -1,5 +1,6 @@
|
|||
/*
|
||||
* Copyright (c) 2019-2020, Sergey Bugaev <bugaevc@serenityos.org>
|
||||
* Copyright (c) 2021, Peter Elliott <pelliott@serenityos.org>
|
||||
*
|
||||
* SPDX-License-Identifier: BSD-2-Clause
|
||||
*/
|
||||
|
|
@ -8,268 +9,362 @@
|
|||
#include <AK/ScopeGuard.h>
|
||||
#include <AK/StringBuilder.h>
|
||||
#include <LibMarkdown/Text.h>
|
||||
#include <ctype.h>
|
||||
#include <string.h>
|
||||
|
||||
namespace Markdown {
|
||||
|
||||
static String unescape(const StringView& text)
|
||||
void Text::EmphasisNode::render_to_html(StringBuilder& builder) const
|
||||
{
|
||||
StringBuilder builder;
|
||||
for (size_t i = 0; i < text.length(); ++i) {
|
||||
if (text[i] == '\\' && i != text.length() - 1) {
|
||||
builder.append(text[i + 1]);
|
||||
i++;
|
||||
continue;
|
||||
}
|
||||
builder.append(text[i]);
|
||||
}
|
||||
return builder.build();
|
||||
builder.append((strong) ? "<strong>" : "<em>");
|
||||
child->render_to_html(builder);
|
||||
builder.append((strong) ? "</strong>" : "</em>");
|
||||
}
|
||||
|
||||
Text::Text(String&& text)
|
||||
void Text::EmphasisNode::render_for_terminal(StringBuilder&) const
|
||||
{
|
||||
m_spans.append({ move(text), Style {} });
|
||||
// FIXME.
|
||||
}
|
||||
|
||||
size_t Text::EmphasisNode::terminal_length() const
|
||||
{
|
||||
return child->terminal_length();
|
||||
}
|
||||
|
||||
void Text::CodeNode::render_to_html(StringBuilder& builder) const
|
||||
{
|
||||
builder.append("<code>");
|
||||
code->render_to_html(builder);
|
||||
builder.append("</code>");
|
||||
}
|
||||
|
||||
void Text::CodeNode::render_for_terminal(StringBuilder&) const
|
||||
{
|
||||
// FIXME.
|
||||
}
|
||||
|
||||
size_t Text::CodeNode::terminal_length() const
|
||||
{
|
||||
return code->terminal_length();
|
||||
}
|
||||
|
||||
void Text::TextNode::render_to_html(StringBuilder& builder) const
|
||||
{
|
||||
builder.append(escape_html_entities(text));
|
||||
}
|
||||
|
||||
void Text::TextNode::render_for_terminal(StringBuilder&) const
|
||||
{
|
||||
// FIXME.
|
||||
}
|
||||
|
||||
size_t Text::TextNode::terminal_length() const
|
||||
{
|
||||
return text.length();
|
||||
}
|
||||
|
||||
void Text::LinkNode::render_to_html(StringBuilder& builder) const
|
||||
{
|
||||
if (is_image) {
|
||||
builder.append("<img src=\"");
|
||||
href->render_to_html(builder);
|
||||
builder.append("\" alt=\"");
|
||||
text->render_to_html(builder);
|
||||
builder.append("\" >");
|
||||
} else {
|
||||
builder.append("<a href=\"");
|
||||
href->render_to_html(builder);
|
||||
builder.append("\">");
|
||||
text->render_to_html(builder);
|
||||
builder.append("</a>");
|
||||
}
|
||||
}
|
||||
|
||||
void Text::LinkNode::render_for_terminal(StringBuilder&) const
|
||||
{
|
||||
// FIXME.
|
||||
}
|
||||
|
||||
size_t Text::LinkNode::terminal_length() const
|
||||
{
|
||||
return text->terminal_length();
|
||||
}
|
||||
|
||||
void Text::MultiNode::render_to_html(StringBuilder& builder) const
|
||||
{
|
||||
for (auto& child : children) {
|
||||
child.render_to_html(builder);
|
||||
}
|
||||
}
|
||||
|
||||
void Text::MultiNode::render_for_terminal(StringBuilder&) const
|
||||
{
|
||||
// FIXME.
|
||||
}
|
||||
|
||||
size_t Text::MultiNode::terminal_length() const
|
||||
{
|
||||
size_t length = 0;
|
||||
for (auto& child : children) {
|
||||
length += child.terminal_length();
|
||||
}
|
||||
return length;
|
||||
}
|
||||
|
||||
size_t Text::terminal_length() const
|
||||
{
|
||||
return m_node->terminal_length();
|
||||
}
|
||||
|
||||
String Text::render_to_html() const
|
||||
{
|
||||
StringBuilder builder;
|
||||
|
||||
Vector<String> open_tags;
|
||||
Style current_style;
|
||||
|
||||
for (auto& span : m_spans) {
|
||||
struct TagAndFlag {
|
||||
String tag;
|
||||
bool Style::*flag;
|
||||
};
|
||||
TagAndFlag tags_and_flags[] = {
|
||||
{ "em", &Style::emph },
|
||||
{ "b", &Style::strong },
|
||||
{ "code", &Style::code }
|
||||
};
|
||||
auto it = open_tags.find_if([&](const String& open_tag) {
|
||||
if (open_tag == "a" && current_style.href != span.style.href)
|
||||
return true;
|
||||
if (open_tag == "img" && current_style.img != span.style.img)
|
||||
return true;
|
||||
for (auto& tag_and_flag : tags_and_flags) {
|
||||
if (open_tag == tag_and_flag.tag && !(span.style.*tag_and_flag.flag))
|
||||
return true;
|
||||
}
|
||||
return false;
|
||||
});
|
||||
|
||||
if (!it.is_end()) {
|
||||
// We found an open tag that should
|
||||
// not be open for the new span. Close
|
||||
// it and all the open tags that follow
|
||||
// it.
|
||||
for (ssize_t j = open_tags.size() - 1; j >= static_cast<ssize_t>(it.index()); --j) {
|
||||
auto& tag = open_tags[j];
|
||||
if (tag == "img") {
|
||||
builder.append("\" />");
|
||||
current_style.img = {};
|
||||
continue;
|
||||
}
|
||||
builder.appendff("</{}>", tag);
|
||||
if (tag == "a") {
|
||||
current_style.href = {};
|
||||
continue;
|
||||
}
|
||||
for (auto& tag_and_flag : tags_and_flags)
|
||||
if (tag == tag_and_flag.tag)
|
||||
current_style.*tag_and_flag.flag = false;
|
||||
}
|
||||
open_tags.shrink(it.index());
|
||||
}
|
||||
if (current_style.href.is_null() && !span.style.href.is_null()) {
|
||||
open_tags.append("a");
|
||||
builder.appendff("<a href=\"{}\">", span.style.href);
|
||||
}
|
||||
if (current_style.img.is_null() && !span.style.img.is_null()) {
|
||||
open_tags.append("img");
|
||||
builder.appendff("<img src=\"{}\" alt=\"", span.style.img);
|
||||
}
|
||||
for (auto& tag_and_flag : tags_and_flags) {
|
||||
if (current_style.*tag_and_flag.flag != span.style.*tag_and_flag.flag) {
|
||||
open_tags.append(tag_and_flag.tag);
|
||||
builder.appendff("<{}>", tag_and_flag.tag);
|
||||
}
|
||||
}
|
||||
|
||||
current_style = span.style;
|
||||
builder.append(escape_html_entities(span.text));
|
||||
}
|
||||
|
||||
for (ssize_t i = open_tags.size() - 1; i >= 0; --i) {
|
||||
auto& tag = open_tags[i];
|
||||
if (tag == "img") {
|
||||
builder.append("\" />");
|
||||
continue;
|
||||
}
|
||||
builder.appendff("</{}>", tag);
|
||||
}
|
||||
|
||||
return builder.build();
|
||||
m_node->render_to_html(builder);
|
||||
return builder.build().trim(" \n\t");
|
||||
}
|
||||
|
||||
String Text::render_for_terminal() const
|
||||
{
|
||||
StringBuilder builder;
|
||||
|
||||
for (auto& span : m_spans) {
|
||||
bool needs_styling = span.style.strong || span.style.emph || span.style.code;
|
||||
if (needs_styling) {
|
||||
builder.append("\033[");
|
||||
bool first = true;
|
||||
if (span.style.strong || span.style.code) {
|
||||
builder.append('1');
|
||||
first = false;
|
||||
}
|
||||
if (span.style.emph) {
|
||||
if (!first)
|
||||
builder.append(';');
|
||||
builder.append('4');
|
||||
}
|
||||
builder.append('m');
|
||||
}
|
||||
|
||||
if (!span.style.href.is_null()) {
|
||||
if (strstr(span.style.href.characters(), "://") != nullptr) {
|
||||
builder.append("\033]8;;");
|
||||
builder.append(span.style.href);
|
||||
builder.append("\033\\");
|
||||
}
|
||||
}
|
||||
|
||||
builder.append(span.text.characters());
|
||||
|
||||
if (needs_styling)
|
||||
builder.append("\033[0m");
|
||||
|
||||
if (!span.style.href.is_null()) {
|
||||
// When rendering for the terminal, ignore any
|
||||
// non-absolute links, because the user has no
|
||||
// chance to follow them anyway.
|
||||
if (strstr(span.style.href.characters(), "://") != nullptr) {
|
||||
builder.appendff(" <{}>", span.style.href);
|
||||
builder.append("\033]8;;\033\\");
|
||||
}
|
||||
}
|
||||
if (!span.style.img.is_null()) {
|
||||
if (strstr(span.style.img.characters(), "://") != nullptr) {
|
||||
builder.appendff(" <{}>", span.style.img);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
return builder.build();
|
||||
m_node->render_for_terminal(builder);
|
||||
return builder.build().trim(" \n\t");
|
||||
}
|
||||
|
||||
Optional<Text> Text::parse(const StringView& str)
|
||||
Text Text::parse(StringView const& str)
|
||||
{
|
||||
Style current_style;
|
||||
size_t current_span_start = 0;
|
||||
int first_span_in_the_current_link = -1;
|
||||
bool current_link_is_actually_img = false;
|
||||
Vector<Span> spans;
|
||||
Text text;
|
||||
auto const tokens = tokenize(str);
|
||||
auto iterator = tokens.begin();
|
||||
text.m_node = parse_sequence(iterator, false);
|
||||
return text;
|
||||
}
|
||||
|
||||
auto append_span_if_needed = [&](size_t offset) {
|
||||
VERIFY(current_span_start <= offset);
|
||||
if (current_span_start != offset) {
|
||||
Span span {
|
||||
unescape(str.substring_view(current_span_start, offset - current_span_start)),
|
||||
current_style
|
||||
};
|
||||
spans.append(move(span));
|
||||
current_span_start = offset;
|
||||
}
|
||||
Vector<Text::Token> Text::tokenize(StringView const& str)
|
||||
{
|
||||
Vector<Token> tokens;
|
||||
StringBuilder current_token;
|
||||
|
||||
auto flush_token = [&](bool left_flanking, bool right_flanking, bool is_run) {
|
||||
if (current_token.is_empty())
|
||||
return;
|
||||
|
||||
tokens.append({
|
||||
current_token.build(),
|
||||
left_flanking,
|
||||
right_flanking,
|
||||
is_run,
|
||||
});
|
||||
current_token.clear();
|
||||
};
|
||||
|
||||
for (size_t offset = 0; offset < str.length(); offset++) {
|
||||
for (size_t offset = 0; offset < str.length(); ++offset) {
|
||||
auto has = [&](StringView const& seq) {
|
||||
if (offset + seq.length() > str.length())
|
||||
return false;
|
||||
|
||||
return str.substring_view(offset, seq.length()) == seq;
|
||||
};
|
||||
|
||||
auto expect = [&](StringView const& seq) {
|
||||
VERIFY(has(seq));
|
||||
flush_token(false, false, false);
|
||||
current_token.append(seq);
|
||||
flush_token(false, false, false);
|
||||
offset += seq.length() - 1;
|
||||
};
|
||||
|
||||
char ch = str[offset];
|
||||
|
||||
bool is_escape = ch == '\\';
|
||||
if (is_escape && offset != str.length() - 1) {
|
||||
offset++;
|
||||
continue;
|
||||
if (ch == '\\' && offset + 1 < str.length()) {
|
||||
current_token.append(str[offset + 1]);
|
||||
++offset;
|
||||
} else if (ch == '*' || ch == '_' || ch == '`') {
|
||||
flush_token(false, false, false);
|
||||
|
||||
char delim = ch;
|
||||
size_t run_offset;
|
||||
for (run_offset = offset; run_offset < str.length() && str[run_offset] == delim; ++run_offset) {
|
||||
current_token.append(str[run_offset]);
|
||||
}
|
||||
|
||||
bool left_flanking = run_offset < str.length() && !isspace(str[run_offset]);
|
||||
bool right_flanking = offset > 0 && !isspace(str[offset - 1]);
|
||||
flush_token(left_flanking, right_flanking, true);
|
||||
offset = run_offset - 1;
|
||||
|
||||
} else if (ch == '\n') {
|
||||
flush_token(false, false, false);
|
||||
current_token.append(ch);
|
||||
flush_token(false, false, false);
|
||||
} else if (has("[")) {
|
||||
expect("[");
|
||||
} else if (has("![")) {
|
||||
expect("![");
|
||||
} else if (has("](")) {
|
||||
expect("](");
|
||||
} else if (has(")")) {
|
||||
expect(")");
|
||||
} else {
|
||||
current_token.append(ch);
|
||||
}
|
||||
}
|
||||
flush_token(false, false, false);
|
||||
return tokens;
|
||||
}
|
||||
|
||||
NonnullOwnPtr<Text::MultiNode> Text::parse_sequence(Vector<Token>::ConstIterator& tokens, bool in_link)
|
||||
{
|
||||
auto node = make<MultiNode>();
|
||||
|
||||
for (; !tokens.is_end(); ++tokens) {
|
||||
if (tokens->is_run) {
|
||||
switch (tokens->run_char()) {
|
||||
case '*':
|
||||
case '_':
|
||||
node->children.append(parse_emph(tokens, in_link));
|
||||
break;
|
||||
case '`':
|
||||
node->children.append(parse_code(tokens));
|
||||
break;
|
||||
}
|
||||
} else if (!in_link && (*tokens == "[" || *tokens == "![")) {
|
||||
node->children.append(parse_link(tokens));
|
||||
} else if (in_link && *tokens == "](") {
|
||||
return node;
|
||||
} else {
|
||||
node->children.append(make<TextNode>(tokens->data));
|
||||
}
|
||||
|
||||
bool is_special_character = false;
|
||||
is_special_character |= ch == '`';
|
||||
if (!current_style.code)
|
||||
is_special_character |= ch == '*' || ch == '_' || ch == '[' || ch == ']' || (ch == '!' && offset + 1 < str.length() && str[offset + 1] == '[');
|
||||
if (!is_special_character)
|
||||
continue;
|
||||
if (in_link && !tokens.is_end() && *tokens == "](")
|
||||
return node;
|
||||
|
||||
append_span_if_needed(offset);
|
||||
if (tokens.is_end())
|
||||
break;
|
||||
}
|
||||
return node;
|
||||
}
|
||||
|
||||
switch (ch) {
|
||||
case '`':
|
||||
current_style.code = !current_style.code;
|
||||
break;
|
||||
case '*':
|
||||
case '_':
|
||||
if (offset + 1 < str.length() && str[offset + 1] == ch) {
|
||||
offset++;
|
||||
current_style.strong = !current_style.strong;
|
||||
} else {
|
||||
current_style.emph = !current_style.emph;
|
||||
}
|
||||
break;
|
||||
case '!':
|
||||
current_link_is_actually_img = true;
|
||||
break;
|
||||
case '[':
|
||||
if constexpr (MARKDOWN_DEBUG) {
|
||||
if (first_span_in_the_current_link != -1)
|
||||
dbgln("Dropping the outer link");
|
||||
}
|
||||
first_span_in_the_current_link = spans.size();
|
||||
break;
|
||||
case ']': {
|
||||
if (first_span_in_the_current_link == -1) {
|
||||
dbgln_if(MARKDOWN_DEBUG, "Unmatched ]");
|
||||
continue;
|
||||
}
|
||||
ScopeGuard guard = [&] {
|
||||
first_span_in_the_current_link = -1;
|
||||
current_link_is_actually_img = false;
|
||||
};
|
||||
if (offset + 2 >= str.length() || str[offset + 1] != '(')
|
||||
continue;
|
||||
offset += 2;
|
||||
size_t start_of_href = offset;
|
||||
bool Text::can_open(Token const& opening)
|
||||
{
|
||||
return (opening.run_char() == '*' && opening.left_flanking) || (opening.run_char() == '_' && opening.left_flanking && !opening.right_flanking);
|
||||
}
|
||||
|
||||
do
|
||||
offset++;
|
||||
while (offset < str.length() && str[offset] != ')');
|
||||
if (offset == str.length())
|
||||
offset--;
|
||||
bool Text::can_close_for(Token const& opening, Text::Token const& closing)
|
||||
{
|
||||
if (opening.run_char() != closing.run_char())
|
||||
return false;
|
||||
|
||||
const StringView href = str.substring_view(start_of_href, offset - start_of_href);
|
||||
for (size_t i = first_span_in_the_current_link; i < spans.size(); i++) {
|
||||
if (current_link_is_actually_img)
|
||||
spans[i].style.img = href;
|
||||
else
|
||||
spans[i].style.href = href;
|
||||
if (opening.run_length() != closing.run_length())
|
||||
return false;
|
||||
|
||||
return (opening.run_char() == '*' && closing.right_flanking) || (opening.run_char() == '_' && !closing.left_flanking && closing.right_flanking);
|
||||
}
|
||||
|
||||
NonnullOwnPtr<Text::Node> Text::parse_emph(Vector<Token>::ConstIterator& tokens, bool in_link)
|
||||
{
|
||||
auto opening = *tokens;
|
||||
|
||||
// Check that the opening delimiter run is properly flanking.
|
||||
if (!can_open(opening))
|
||||
return make<TextNode>(opening.data);
|
||||
|
||||
auto child = make<MultiNode>();
|
||||
for (++tokens; !tokens.is_end(); ++tokens) {
|
||||
if (tokens->is_run) {
|
||||
if (can_close_for(opening, *tokens)) {
|
||||
return make<EmphasisNode>(opening.run_length() >= 2, move(child));
|
||||
}
|
||||
break;
|
||||
}
|
||||
default:
|
||||
VERIFY_NOT_REACHED();
|
||||
|
||||
switch (tokens->run_char()) {
|
||||
case '*':
|
||||
case '_':
|
||||
child->children.append(parse_emph(tokens, in_link));
|
||||
break;
|
||||
case '`':
|
||||
child->children.append(parse_code(tokens));
|
||||
break;
|
||||
}
|
||||
} else if (*tokens == "[" || *tokens == "![") {
|
||||
child->children.append(parse_link(tokens));
|
||||
} else if (in_link && *tokens == "](") {
|
||||
child->children.prepend(make<TextNode>(opening.data));
|
||||
return child;
|
||||
} else {
|
||||
child->children.append(make<TextNode>(tokens->data));
|
||||
}
|
||||
|
||||
// We've processed the character as a special, so the next offset will
|
||||
// start after it. Note that explicit continue statements skip over this
|
||||
// line, effectively treating the character as not special.
|
||||
current_span_start = offset + 1;
|
||||
if (in_link && !tokens.is_end() && *tokens == "](") {
|
||||
child->children.prepend(make<TextNode>(opening.data));
|
||||
return child;
|
||||
}
|
||||
|
||||
if (tokens.is_end())
|
||||
break;
|
||||
}
|
||||
child->children.prepend(make<TextNode>(opening.data));
|
||||
return child;
|
||||
}
|
||||
|
||||
NonnullOwnPtr<Text::Node> Text::parse_code(Vector<Token>::ConstIterator& tokens)
|
||||
{
|
||||
auto opening = *tokens;
|
||||
|
||||
auto is_closing = [&](Token const& token) {
|
||||
return token.is_run && token.run_char() == '`' && token.run_length() == opening.run_length();
|
||||
};
|
||||
|
||||
bool is_all_whitespace = true;
|
||||
auto code = make<MultiNode>();
|
||||
for (auto iterator = tokens + 1; !iterator.is_end(); ++iterator) {
|
||||
if (is_closing(*iterator)) {
|
||||
tokens = iterator;
|
||||
|
||||
// Strip first and last space, when appropriate.
|
||||
if (!is_all_whitespace) {
|
||||
auto& first = dynamic_cast<TextNode&>(code->children.first());
|
||||
auto& last = dynamic_cast<TextNode&>(code->children.last());
|
||||
if (first.text.starts_with(" ") && last.text.ends_with(" ")) {
|
||||
first.text = first.text.substring(1);
|
||||
last.text = last.text.substring(0, last.text.length() - 1);
|
||||
}
|
||||
}
|
||||
|
||||
return make<CodeNode>(move(code));
|
||||
}
|
||||
|
||||
is_all_whitespace = is_all_whitespace && iterator->data.is_whitespace();
|
||||
code->children.append(make<TextNode>((*iterator == "\n") ? " " : iterator->data));
|
||||
}
|
||||
|
||||
append_span_if_needed(str.length());
|
||||
|
||||
return Text(move(spans));
|
||||
return make<TextNode>(opening.data);
|
||||
}
|
||||
|
||||
NonnullOwnPtr<Text::Node> Text::parse_link(Vector<Token>::ConstIterator& tokens)
|
||||
{
|
||||
auto opening = *tokens++;
|
||||
bool is_image = opening == "![";
|
||||
|
||||
auto link_text = parse_sequence(tokens, true);
|
||||
|
||||
if (tokens.is_end() || *tokens != "](") {
|
||||
link_text->children.prepend(make<TextNode>(opening.data));
|
||||
return link_text;
|
||||
}
|
||||
auto seperator = *tokens;
|
||||
VERIFY(seperator == "](");
|
||||
|
||||
auto address = make<MultiNode>();
|
||||
for (auto iterator = tokens + 1; !iterator.is_end(); ++iterator) {
|
||||
if (*iterator == ")") {
|
||||
tokens = iterator;
|
||||
return make<LinkNode>(is_image, move(link_text), move(address));
|
||||
}
|
||||
|
||||
address->children.append(make<TextNode>(iterator->data));
|
||||
}
|
||||
|
||||
link_text->children.prepend(make<TextNode>(opening.data));
|
||||
link_text->children.append(make<TextNode>(seperator.data));
|
||||
return link_text;
|
||||
}
|
||||
}
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue