aboutsummaryrefslogtreecommitdiffstats
path: root/src/construct/gfm_footnote_definition.rs
diff options
context:
space:
mode:
authorLibravatar Titus Wormer <tituswormer@gmail.com>2022-09-12 13:50:48 +0200
committerLibravatar Titus Wormer <tituswormer@gmail.com>2022-09-12 13:54:43 +0200
commit2011d2ea4d6ec9d1bd7409c22e4258aceaaa3afe (patch)
treee462579538174cb983dcc2be4b0418b8c5077050 /src/construct/gfm_footnote_definition.rs
parent640c103c0a2b92f7f2a49cfc0721577f40aa90e0 (diff)
downloadmarkdown-rs-2011d2ea4d6ec9d1bd7409c22e4258aceaaa3afe.tar.gz
markdown-rs-2011d2ea4d6ec9d1bd7409c22e4258aceaaa3afe.tar.bz2
markdown-rs-2011d2ea4d6ec9d1bd7409c22e4258aceaaa3afe.zip
Fix whitespace in GFM footnote definition identifiers
* Fix a crash on line endings in footnote definitions * Fix to match spaces and tabs in identifiers to `cmark-gfm` * Fix order of one attribute
Diffstat (limited to 'src/construct/gfm_footnote_definition.rs')
-rw-r--r--src/construct/gfm_footnote_definition.rs114
1 files changed, 96 insertions, 18 deletions
diff --git a/src/construct/gfm_footnote_definition.rs b/src/construct/gfm_footnote_definition.rs
index cbe816f..d3f72d7 100644
--- a/src/construct/gfm_footnote_definition.rs
+++ b/src/construct/gfm_footnote_definition.rs
@@ -127,6 +127,8 @@
//! β€” while `CommonMark` prevents links in links, GitHub does not prevent footnotes (which turn into links) in links
//! * [Footnote-like brackets around image, break that image](https://github.com/github/cmark-gfm/issues/275)\
//! β€” images can’t be used in what looks like a footnote call
+//! * [GFM footnotes: line ending in footnote definition label causes text to disappear](https://github.com/github/cmark-gfm/issues/282)\
+//! β€” line endings in footnote definitions cause text to disappear
//!
//! ## Tokens
//!
@@ -164,11 +166,11 @@
//! [html_sup]: https://html.spec.whatwg.org/multipage/text-level-semantics.html#the-sub-and-sup-elements
use crate::construct::partial_space_or_tab::space_or_tab_min_max;
-use crate::event::Name;
+use crate::event::{Content, Link, Name};
use crate::state::{Name as StateName, State};
use crate::tokenizer::Tokenizer;
use crate::util::{
- constant::TAB_SIZE,
+ constant::{LINK_REFERENCE_SIZE_MAX, TAB_SIZE},
normalize_identifier::normalize_identifier,
skip,
slice::{Position, Slice},
@@ -220,22 +222,104 @@ pub fn start(tokenizer: &mut Tokenizer) -> State {
pub fn label_before(tokenizer: &mut Tokenizer) -> State {
match tokenizer.current {
Some(b'[') => {
- tokenizer.tokenize_state.token_1 = Name::GfmFootnoteDefinitionLabel;
- tokenizer.tokenize_state.token_2 = Name::GfmFootnoteDefinitionLabelMarker;
- tokenizer.tokenize_state.token_3 = Name::GfmFootnoteDefinitionLabelString;
- tokenizer.tokenize_state.token_4 = Name::GfmFootnoteDefinitionMarker;
- tokenizer.tokenize_state.marker = b'^';
tokenizer.enter(Name::GfmFootnoteDefinitionPrefix);
- tokenizer.attempt(
- State::Next(StateName::GfmFootnoteDefinitionLabelAfter),
- State::Nok,
- );
- State::Retry(StateName::LabelStart)
+ tokenizer.enter(Name::GfmFootnoteDefinitionLabel);
+ tokenizer.enter(Name::GfmFootnoteDefinitionLabelMarker);
+ tokenizer.consume();
+ tokenizer.exit(Name::GfmFootnoteDefinitionLabelMarker);
+ State::Next(StateName::GfmFootnoteDefinitionLabelAtMarker)
}
_ => State::Nok,
}
}
+/// In label, at caret.
+///
+/// ```markdown
+/// > | [^a]: b
+/// ^
+/// ```
+pub fn label_at_marker(tokenizer: &mut Tokenizer) -> State {
+ if tokenizer.current == Some(b'^') {
+ tokenizer.enter(Name::GfmFootnoteDefinitionMarker);
+ tokenizer.consume();
+ tokenizer.exit(Name::GfmFootnoteDefinitionMarker);
+ tokenizer.enter(Name::GfmFootnoteDefinitionLabelString);
+ tokenizer.enter_link(
+ Name::Data,
+ Link {
+ previous: None,
+ next: None,
+ content: Content::String,
+ },
+ );
+ State::Next(StateName::GfmFootnoteDefinitionLabelInside)
+ } else {
+ State::Nok
+ }
+}
+
+/// In label.
+///
+/// > πŸ‘‰ **Note**: `cmark-gfm` prevents whitespace from occurring in footnote
+/// > definition labels.
+///
+/// ```markdown
+/// > | [^a]: b
+/// ^
+/// ```
+pub fn label_inside(tokenizer: &mut Tokenizer) -> State {
+ // Too long.
+ if tokenizer.tokenize_state.size > LINK_REFERENCE_SIZE_MAX
+ // Space or tab is not supported by GFM for some reason (`\n` and
+ // `[` make sense).
+ || matches!(tokenizer.current, None | Some(b'\t' | b'\n' | b' ' | b'['))
+ // Closing brace with nothing.
+ || (matches!(tokenizer.current, Some(b']')) && tokenizer.tokenize_state.size == 0)
+ {
+ tokenizer.tokenize_state.size = 0;
+ State::Nok
+ } else if matches!(tokenizer.current, Some(b']')) {
+ tokenizer.tokenize_state.size = 0;
+ tokenizer.exit(Name::Data);
+ tokenizer.exit(Name::GfmFootnoteDefinitionLabelString);
+ tokenizer.enter(Name::GfmFootnoteDefinitionLabelMarker);
+ tokenizer.consume();
+ tokenizer.exit(Name::GfmFootnoteDefinitionLabelMarker);
+ tokenizer.exit(Name::GfmFootnoteDefinitionLabel);
+ State::Next(StateName::GfmFootnoteDefinitionLabelAfter)
+ } else {
+ let next = if matches!(tokenizer.current.unwrap(), b'\\') {
+ StateName::GfmFootnoteDefinitionLabelEscape
+ } else {
+ StateName::GfmFootnoteDefinitionLabelInside
+ };
+ tokenizer.consume();
+ tokenizer.tokenize_state.size += 1;
+ State::Next(next)
+ }
+}
+
+/// After `\`, at a special character.
+///
+/// > πŸ‘‰ **Note**: `cmark-gfm` currently does not support escaped brackets:
+/// > <https://github.com/github/cmark-gfm/issues/240>
+///
+/// ```markdown
+/// > | [^a\*b]: c
+/// ^
+/// ```
+pub fn label_escape(tokenizer: &mut Tokenizer) -> State {
+ match tokenizer.current {
+ Some(b'[' | b'\\' | b']') => {
+ tokenizer.tokenize_state.size += 1;
+ tokenizer.consume();
+ State::Next(StateName::GfmFootnoteDefinitionLabelInside)
+ }
+ _ => State::Retry(StateName::GfmFootnoteDefinitionLabelInside),
+ }
+}
+
/// After definition label.
///
/// ```markdown
@@ -243,12 +327,6 @@ pub fn label_before(tokenizer: &mut Tokenizer) -> State {
/// ^
/// ```
pub fn label_after(tokenizer: &mut Tokenizer) -> State {
- tokenizer.tokenize_state.token_1 = Name::Data;
- tokenizer.tokenize_state.token_2 = Name::Data;
- tokenizer.tokenize_state.token_3 = Name::Data;
- tokenizer.tokenize_state.token_4 = Name::Data;
- tokenizer.tokenize_state.marker = 0;
-
match tokenizer.current {
Some(b':') => {
let end = skip::to_back(