aboutsummaryrefslogtreecommitdiffstats
path: root/src
diff options
context:
space:
mode:
authorLibravatar Titus Wormer <tituswormer@gmail.com>2022-10-13 12:50:52 +0200
committerLibravatar Titus Wormer <tituswormer@gmail.com>2022-10-13 12:50:52 +0200
commitbdbd90b7c25d62bc7ebc20246e9f8dffea1228b3 (patch)
tree8c9c5bc53021e7a7c1b154f22aca4f170bf9e28a /src
parentcaf2c774ac8a4713e8be753e5cddb09b499dbe1c (diff)
downloadmarkdown-rs-bdbd90b7c25d62bc7ebc20246e9f8dffea1228b3.tar.gz
markdown-rs-bdbd90b7c25d62bc7ebc20246e9f8dffea1228b3.tar.bz2
markdown-rs-bdbd90b7c25d62bc7ebc20246e9f8dffea1228b3.zip
Remove unneeded code from `gfm_autolink_literal`
Diffstat (limited to '')
-rw-r--r--src/construct/gfm_autolink_literal.rs100
1 files changed, 9 insertions, 91 deletions
diff --git a/src/construct/gfm_autolink_literal.rs b/src/construct/gfm_autolink_literal.rs
index c4ef637..0704da7 100644
--- a/src/construct/gfm_autolink_literal.rs
+++ b/src/construct/gfm_autolink_literal.rs
@@ -650,7 +650,15 @@ pub fn resolve(tokenizer: &mut Tokenizer) {
byte_index + 1,
kind == Name::GfmAutolinkLiteralXmpp,
) {
- range = (start, peek_bytes_truncate(bytes, start, end), kind);
+ // Note: normally we’d truncate trailing
+ // punctuation from the link.
+ // However, email autolink literals cannot
+ // contain any of those markers, except for
+ // `.`, but that can only occur if it isn’t
+ // trailing.
+ // So we can ignore truncating while
+ // postprocessing!
+ range = (start, end, kind);
}
}
@@ -838,93 +846,3 @@ fn peek_bytes_email_domain(bytes: &[u8], start: usize, xmpp: bool) -> Option<usi
None
}
}
-
-/// Move back past punctuation.
-///
-/// Moving back is only used when post processing text: so for the email address
-/// algorithm.
-///
-/// This is much more complex that needed, because GH allows a lot of
-/// punctuation in the protocol and www algorithms.
-/// However, those aren’t implemented like the email algo.
-///
-/// ```markdown
-/// > | a contact@example.org”) b
-/// ^-- from
-/// ^-- to
-/// ```
-fn peek_bytes_truncate(bytes: &[u8], start: usize, mut end: usize) -> usize {
- let mut index = start;
-
- // Source: <https://github.com/github/cmark-gfm/blob/ef1cfcb/extensions/autolink.c#L42>
- while index < end {
- if bytes[index] == b'<' {
- end = index;
- break;
- }
- index += 1;
- }
-
- let mut split = end;
-
- // Move before trailing punctuation.
- while split > start {
- match bytes[split - 1] {
- b'!' | b'"' | b'&' | b'\'' | b')' | b',' | b'.' | b':' | b'<' | b'>' | b'?' | b']'
- | b'}' => {}
- // Source: <https://github.com/github/cmark-gfm/blob/ef1cfcb/extensions/autolink.c#L61>.
- // Note: we can’t move across actual references, because those have been parsed already.
- b';' => {
- let mut new_split = split - 1;
- // Move back past alphabeticals.
- while new_split > start && matches!(bytes[new_split - 1], b'A'..=b'Z' | b'a'..=b'z')
- {
- new_split -= 1;
- }
-
- // Nonempty character reference:
- if new_split > start && bytes[new_split - 1] == b'&' && new_split < split - 1 {
- split = new_split - 1;
- continue;
- }
-
- // Otherwise it’s just a `;`.
- }
- _ => break,
- }
- split -= 1;
- }
-
- // If there was trailing punctuation, try to balance parens.
- if split != end {
- let mut open = 0;
- let mut close = 0;
- let mut paren_index = start;
-
- // Count parens in `url` (not in trail).
- while paren_index < split {
- match bytes[paren_index] {
- b'(' => open += 1,
- b')' => close += 1,
- _ => {}
- }
-
- paren_index += 1;
- }
-
- let mut trail_index = split;
-
- // If there are more opening than closing parens, try to balance them
- // from the trail.
- while open > close && trail_index < end {
- if bytes[trail_index] == b')' {
- split = trail_index;
- close += 1;
- }
-
- trail_index += 1;
- }
- }
-
- split
-}