From 1f2a7604150f8c92c3572528e9379c80344a5620 Mon Sep 17 00:00:00 2001
From: Astro <astro@spaceboyz.net>
Date: Fri, 25 Mar 2022 01:53:21 +0100
Subject: [PATCH] minidom: normalize newlines

---
 minidom/src/token.rs | 25 +++++++++++++++++++++++--
 1 file changed, 23 insertions(+), 2 deletions(-)

diff --git a/minidom/src/token.rs b/minidom/src/token.rs
index 2f79e6a..5c86a37 100644
--- a/minidom/src/token.rs
+++ b/minidom/src/token.rs
@@ -1,5 +1,6 @@
 //! Parsed XML token
 
+use std::borrow::Cow;
 use nom::{
     branch::alt,
     bytes::streaming::{tag, take_while1},
@@ -212,8 +213,8 @@ impl Token {
                     let (s, _) = not(peek(char(until)))(s)?;
                     let (s, text) = take_while1(|b| b != until as u8 && b != b'&')(s)?;
                     let text = Self::str_from_utf8(text)?;
-                    // TODO: CoW
-                    Ok((s, text.to_owned()))
+                    let text = Self::normalize_newlines(text);
+                    Ok((s, text.into_owned()))
                 })
             )
         )(s)?;
@@ -226,6 +227,18 @@ impl Token {
         std::str::from_utf8(s)
             .map_err(|_| nom::Err::Failure(nom::error::Error::new(s, nom::error::ErrorKind::Fail)))
     }
+
+    /// https://www.w3.org/TR/2008/REC-xml-20081126/#sec-line-ends
+    fn normalize_newlines(s: &str) -> Cow<str> {
+        let mut s = Cow::from(s);
+        if s.find("\r\n").is_some() {
+            s = Cow::from(s.replace("\r\n", "\n"));
+        }
+        if s.find("\r").is_some() {
+            s = Cow::from(s.replace("\r", "\n"));
+        }
+        s
+    }
 }
 
 #[cfg(test)]
@@ -247,6 +260,14 @@ mod tests {
         );
     }
 
+    #[test]
+    fn test_newlines() {
+        assert_eq!(
+            Ok((&b"</x"[..], Token::Text("a\nb\nc\nd".to_string()))),
+            Token::parse(b"a\nb\rc\r\nd</x")
+        );
+    }
+
     #[test]
     fn test_text_entities() {
         assert_eq!(