Fix handling of trailing bare CR in str::lines

KamilaBorowska · ijackson · KamilaBorowska · commit e5377776044e · 2022-08-09T07:27:37.000Z
Previously "bare\r" was split into ["bare"] even though the documentation said that only LF and CRLF count as newlines. This fix is a behavioural change, even though it brings the behaviour into line with the documentation, and into line with that of `std::io::BufRead::lines()`. This is an alternative to #91051, which proposes to document rather than fix the behaviour. Fixes #94435. Co-authored-by: Ian Jackson <ijackson@chiark.greenend.org.uk>
diff --git a/library/alloc/tests/str.rs b/library/alloc/tests/str.rs
@@ -1499,13 +1499,25 @@ fn test_split_whitespace() {
 
 #[test]
 fn test_lines() {
-    let data = "\nMäry häd ä little lämb\n\r\nLittle lämb\n";
-    let lines: Vec<&str> = data.lines().collect();
-    assert_eq!(lines, ["", "Märy häd ä little lämb", "", "Little lämb"]);
-
-    let data = "\r\nMäry häd ä little lämb\n\nLittle lämb"; // no trailing \n
-    let lines: Vec<&str> = data.lines().collect();
-    assert_eq!(lines, ["", "Märy häd ä little lämb", "", "Little lämb"]);
+    fn t(data: &str, expected: &[&str]) {
+        let lines: Vec<&str> = data.lines().collect();
+        assert_eq!(lines, expected);
+    }
+    t("", &[]);
+    t("\n", &[""]);
+    t("\n2nd", &["", "2nd"]);
+    t("\r\n", &[""]);
+    t("bare\r", &["bare\r"]);
+    t("bare\rcr", &["bare\rcr"]);
+    t("Text\n\r", &["Text", "\r"]);
+    t(
+        "\nMäry häd ä little lämb\n\r\nLittle lämb\n",
+        &["", "Märy häd ä little lämb", "", "Little lämb"],
+    );
+    t(
+        "\r\nMäry häd ä little lämb\n\nLittle lämb",
+        &["", "Märy häd ä little lämb", "", "Little lämb"],
+    );
 }
 
 #[test]
diff --git a/library/core/src/lib.rs b/library/core/src/lib.rs
@@ -184,6 +184,7 @@
 #![feature(intra_doc_pointers)]
 #![feature(intrinsics)]
 #![feature(lang_items)]
+#![feature(let_else)]
 #![feature(link_llvm_intrinsics)]
 #![feature(macro_metavar_expr)]
 #![feature(min_specialization)]
diff --git a/library/core/src/str/iter.rs b/library/core/src/str/iter.rs
@@ -1091,7 +1091,7 @@ generate_pattern_iterators! {
 #[stable(feature = "rust1", since = "1.0.0")]
 #[must_use = "iterators are lazy and do nothing unless consumed"]
 #[derive(Clone, Debug)]
-pub struct Lines<'a>(pub(super) Map<SplitTerminator<'a, char>, LinesMap>);
+pub struct Lines<'a>(pub(super) Map<SplitInclusive<'a, char>, LinesMap>);
 
 #[stable(feature = "rust1", since = "1.0.0")]
 impl<'a> Iterator for Lines<'a> {
diff --git a/library/core/src/str/mod.rs b/library/core/src/str/mod.rs
@@ -995,7 +995,7 @@ impl str {
     #[stable(feature = "rust1", since = "1.0.0")]
     #[inline]
     pub fn lines(&self) -> Lines<'_> {
-        Lines(self.split_terminator('\n').map(LinesMap))
+        Lines(self.split_inclusive('\n').map(LinesMap))
     }
 
     /// An iterator over the lines of a string.
@@ -2589,9 +2589,9 @@ impl_fn_for_zst! {
     /// A nameable, cloneable fn type
     #[derive(Clone)]
     struct LinesMap impl<'a> Fn = |line: &'a str| -> &'a str {
-        let l = line.len();
-        if l > 0 && line.as_bytes()[l - 1] == b'\r' { &line[0 .. l - 1] }
-        else { line }
+        let Some(line) = line.strip_suffix('\n') else { return line };
+        let Some(line) = line.strip_suffix('\r') else { return line };
+        line
     };
 
     #[derive(Clone)]
