Remove CharKind::Newline

Nathan Sobo and Keith Simmons created

This is just a character, and so it seems clearer to refer to it specifically when we want to know if a character is a newline. There was only one case where we relied on Newline being different from Whitespace, and we special-cased that instance. Changing this actually makes us match the behavior of VS Code when double-clicking runs of multiple newlines.

/cc @as-cii

Co-Authored-By: Keith Simmons <keith@the-simmons.net>

Change summary

crates/editor/src/movement.rs     | 20 ++++++++++----------
crates/editor/src/multi_buffer.rs |  4 ++--
crates/language/src/buffer.rs     |  5 +----
3 files changed, 13 insertions(+), 16 deletions(-)

Detailed changes

crates/editor/src/movement.rs 🔗

@@ -151,17 +151,14 @@ pub fn previous_word_start(map: &DisplaySnapshot, mut point: DisplayPoint) -> Di
 
     let mut boundary = DisplayPoint::new(point.row(), 0);
     let mut column = 0;
-    let mut prev_char_kind = CharKind::Newline;
+    let mut prev_char_kind = CharKind::Whitespace;
     for c in map.chars_at(DisplayPoint::new(point.row(), 0)) {
         if column >= point.column() {
             break;
         }
 
         let char_kind = char_kind(c);
-        if char_kind != prev_char_kind
-            && char_kind != CharKind::Whitespace
-            && char_kind != CharKind::Newline
-        {
+        if char_kind != prev_char_kind && char_kind != CharKind::Whitespace && c != '\n' {
             *boundary.column_mut() = column;
         }
 
@@ -179,10 +176,7 @@ pub fn next_word_end(map: &DisplaySnapshot, mut point: DisplayPoint) -> DisplayP
             if c == '\n' {
                 break;
             }
-            if prev_char_kind != char_kind
-                && prev_char_kind != CharKind::Whitespace
-                && prev_char_kind != CharKind::Newline
-            {
+            if prev_char_kind != char_kind && prev_char_kind != CharKind::Whitespace {
                 break;
             }
         }
@@ -441,7 +435,7 @@ mod tests {
             .select_font(family_id, &Default::default())
             .unwrap();
         let font_size = 14.0;
-        let buffer = MultiBuffer::build_simple("lorem ipsum   dolor\n    sit", cx);
+        let buffer = MultiBuffer::build_simple("lorem ipsum   dolor\n    sit\n\n\n\n", cx);
         let display_map = cx
             .add_model(|cx| DisplayMap::new(buffer, tab_size, font_id, font_size, None, 1, 1, cx));
         let snapshot = display_map.update(cx, |map, cx| map.snapshot(cx));
@@ -502,5 +496,11 @@ mod tests {
             surrounding_word(&snapshot, DisplayPoint::new(1, 7)),
             DisplayPoint::new(1, 4)..DisplayPoint::new(1, 7),
         );
+
+        // Don't consider runs of multiple newlines to be a "word"
+        assert_eq!(
+            surrounding_word(&snapshot, DisplayPoint::new(3, 0)),
+            DisplayPoint::new(3, 0)..DisplayPoint::new(3, 0),
+        );
     }
 }

crates/editor/src/multi_buffer.rs 🔗

@@ -1407,7 +1407,7 @@ impl MultiBufferSnapshot {
         );
 
         for ch in prev_chars {
-            if Some(char_kind(ch)) == word_kind {
+            if Some(char_kind(ch)) == word_kind && ch != '\n' {
                 start -= ch.len_utf8();
             } else {
                 break;
@@ -1415,7 +1415,7 @@ impl MultiBufferSnapshot {
         }
 
         for ch in next_chars {
-            if Some(char_kind(ch)) == word_kind {
+            if Some(char_kind(ch)) == word_kind && ch != '\n' {
                 end += ch.len_utf8();
             } else {
                 break;

crates/language/src/buffer.rs 🔗

@@ -271,7 +271,6 @@ pub(crate) struct DiagnosticEndpoint {
 
 #[derive(Copy, Clone, Eq, PartialEq, PartialOrd, Ord, Debug)]
 pub enum CharKind {
-    Newline,
     Punctuation,
     Whitespace,
     Word,
@@ -2259,9 +2258,7 @@ pub fn contiguous_ranges(
 }
 
 pub fn char_kind(c: char) -> CharKind {
-    if c == '\n' {
-        CharKind::Newline
-    } else if c.is_whitespace() {
+    if c.is_whitespace() {
         CharKind::Whitespace
     } else if c.is_alphanumeric() || c == '_' {
         CharKind::Word