\r\n-compatible newline detection

metamath · Sep 30, 2023 · 854ab2e · 854ab2e
1 parent 363c92a
commit 854ab2e
Show file tree

Hide file tree

Showing 3 changed files with 17 additions and 8 deletions.
diff --git a/src/comment_parser.rs b/src/comment_parser.rs
@@ -304,7 +304,7 @@ impl<'a> CommentParser<'a> {
         while let Some(&c) = self.buf.get(self.pos) {
             match c {
                 b'[' | b'`' if self.buf.get(self.pos + 1) == Some(&c) => self.pos += 2,
-                b' ' | b'\n' | b'`' => break,
+                b' ' | b'\r' | b'\n' | b'`' => break,
                 b'[' if self.parse_bib().is_some() => break,
                 b'<' if self.buf[self.pos..].starts_with(b"<HTML>")
                     || self.buf[self.pos..].starts_with(b"</HTML>") =>
@@ -534,9 +534,9 @@ impl<'a> ParentheticalIter<'a> {
     pub fn new(buf: &'a [u8], span: Span) -> Self {
         lazy_static! {
             static ref PARENTHETICALS: Regex = Regex::new(concat!(
-                r"\((Contributed|Revised|Proof[ \n]+shortened)",
-                r"[ \n]+by[ \n]+([^,)]+),[ \n]+([0-9]{1,2}-[A-Z][a-z]{2}-[0-9]{4})\.\)|",
-                r"\((Proof[ \n]+modification|New[ \n]+usage)[ \n]+is[ \n]+discouraged\.\)",
+                r"\((Contributed|Revised|Proof[ \r\n]+shortened)",
+                r"[ \r\n]+by[ \r\n]+([^,)]+),[ \r\n]+([0-9]{1,2}-[A-Z][a-z]{2}-[0-9]{4})\.\)|",
+                r"\((Proof[ \r\n]+modification|New[ \r\n]+usage)[ \r\n]+is[ \r\n]+discouraged\.\)",
             ))
             .unwrap();
         }

diff --git a/src/parser.rs b/src/parser.rs
@@ -1102,13 +1102,13 @@ impl HeadingComment {
     pub(crate) fn parse(buf: &[u8], lvl: HeadingLevel, sp: Span) -> Option<Self> {
         lazy_static::lazy_static! {
             static ref MAJOR_PART: Regex =
-                Regex::new(r"^[ \n]+#{4,}\n *([^\n]*)\n#{4,}\n").unwrap();
+                Regex::new(r"^[ \r\n]+#{4,}\r?\n *([^\n]*)\r?\n#{4,}\r?\n").unwrap();
             static ref SECTION: Regex =
-                Regex::new(r"^[ \n]+(?:#\*){2,}#?\n *([^\n]*)\n(?:#\*){2,}#?\n").unwrap();
+                Regex::new(r"^[ \r\n]+(?:#\*){2,}#?\r?\n *([^\n]*)\r?\n(?:#\*){2,}#?\r?\n").unwrap();
             static ref SUBSECTION: Regex =
-                Regex::new(r"^[ \n]+(?:=-){2,}=?\n *([^\n]*)\n(?:=-){2,}=?\n").unwrap();
+                Regex::new(r"^[ \r\n]+(?:=-){2,}=?\r?\n *([^\n]*)\r?\n(?:=-){2,}=?\r?\n").unwrap();
             static ref SUBSUBSECTION: Regex =
-                Regex::new(r"^[ \n]+(?:-\.){2,}-?\n *([^\n]*)\n(?:-\.){2,}-?\n").unwrap();
+                Regex::new(r"^[ \r\n]+(?:-\.){2,}-?\r?\n *([^\n]*)\r?\n(?:-\.){2,}-?\r?\n").unwrap();
         }
         let regex = match lvl {
             HeadingLevel::MajorPart => &*MAJOR_PART,

diff --git a/src/verify_markup.rs b/src/verify_markup.rs
@@ -286,6 +286,15 @@ impl Database {
                         eol_check(&mut diags, &seg, line_start, i);
                         line_start = i + 1;
                     }
+                    b'\r' => {
+                        eol_check(&mut diags, &seg, line_start, i);
+                        line_start = if seg.buffer.get(i + 1) == Some(&b'\n') {
+                            iter.next();
+                            i + 2
+                        } else {
+                            i + 1
+                        }
+                    }
                     b'\t' => {
                         let count = seg.buffer[i..].iter().take_while(|&&c| c == b'\t').count();
                         for _ in 1..count {