Add narrow no-break space support

book/src/configuration.md

@@ -253,8 +253,8 @@ Options for rendering whitespace with visible characters. Use `:set whitespace.r
 
 | Key | Description | Default |
 |-----|-------------|---------|
-| `render` | Whether to render whitespace. May either be `"all"` or `"none"`, or a table with sub-keys `space`, `nbsp`, `tab`, and `newline` | `"none"` |
-| `characters` | Literal characters to use when rendering whitespace. Sub-keys may be any of `tab`, `space`, `nbsp`, `newline` or `tabpad` | See example below |
+| `render` | Whether to render whitespace. May either be `"all"` or `"none"`, or a table with sub-keys `space`, `nbsp`, `nnbsp`, `tab`, and `newline` | `"none"` |
+| `characters` | Literal characters to use when rendering whitespace. Sub-keys may be any of `tab`, `space`, `nbsp`, `nnbsp`, `newline` or `tabpad` | See example below |
 
 Example
 
@@ -265,11 +265,14 @@ render = "all"
 [editor.whitespace.render]
 space = "all"
 tab = "all"
+nbsp = "none"
+nnbsp = "none"
 newline = "none"
 
 [editor.whitespace.characters]
 space = "·"
 nbsp = "⍽"
+nnbsp = "␣"
 tab = "→"
 newline = "⏎"
 tabpad = "·" # Tabs will look like "→···" (depending on tab width)

helix-core/src/chars.rs

@@ -45,8 +45,8 @@ pub fn char_is_whitespace(ch: char) -> bool {
         '\u{0009}' | // Character Tabulation
         '\u{0020}' | // Space
         '\u{00A0}' | // No-break Space
-        '\u{180E}' | // Mongolian Vowel Separator
         '\u{202F}' | // Narrow No-break Space
+        '\u{180E}' | // Mongolian Vowel Separator
         '\u{205F}' | // Medium Mathematical Space
         '\u{3000}' | // Ideographic Space
         '\u{FEFF}'   // Zero Width No-break Space

helix-term/src/ui/document.rs

@@ -341,6 +341,7 @@ pub struct TextRenderer<'a> {
     pub indent_guide_style: Style,
     pub newline: String,
     pub nbsp: String,
+    pub nnbsp: String,
     pub space: String,
     pub tab: String,
     pub virtual_tab: String,
@@ -395,6 +396,11 @@ impl<'a> TextRenderer<'a> {
         } else {
             " ".to_owned()
         };
+        let nnbsp = if ws_render.nnbsp() == WhitespaceRenderValue::All {
+            ws_chars.nnbsp.into()
+        } else {
+            " ".to_owned()
+        };
 
         let text_style = theme.get("ui.text");
 
@@ -405,6 +411,7 @@ impl<'a> TextRenderer<'a> {
             indent_guide_char: editor_config.indent_guides.character.into(),
             newline,
             nbsp,
+            nnbsp,
             space,
             tab,
             virtual_tab,
@@ -448,6 +455,7 @@ impl<'a> TextRenderer<'a> {
         let width = grapheme.width();
         let space = if is_virtual { " " } else { &self.space };
         let nbsp = if is_virtual { " " } else { &self.nbsp };
+        let nnbsp = if is_virtual { " " } else { &self.nnbsp };
         let tab = if is_virtual {
             &self.virtual_tab
         } else {
@@ -461,6 +469,7 @@ impl<'a> TextRenderer<'a> {
             // TODO special rendering for other whitespaces?
             Grapheme::Other { ref g } if g == " " => space,
             Grapheme::Other { ref g } if g == "\u{00A0}" => nbsp,
+            Grapheme::Other { ref g } if g == "\u{202F}" => nnbsp,
             Grapheme::Other { ref g } => g,
             Grapheme::Newline => &self.newline,
         };

helix-tui/src/widgets/reflow.rs

@@ -4,6 +4,7 @@ use helix_core::unicode::width::UnicodeWidthStr;
 use unicode_segmentation::UnicodeSegmentation;
 
 const NBSP: &str = "\u{00a0}";
+const NNBSP: &str = "\u{202f}";
 
 /// A state machine to pack styled symbols into lines.
 /// Cannot implement it as Iterator since it yields slices of the internal buffer (need streaming
@@ -58,7 +59,8 @@ impl<'a, 'b> LineComposer<'a> for WordWrapper<'a, 'b> {
         let mut symbols_exhausted = true;
         for StyledGrapheme { symbol, style } in &mut self.symbols {
             symbols_exhausted = false;
-            let symbol_whitespace = symbol.chars().all(&char::is_whitespace) && symbol != NBSP;
+            let symbol_whitespace =
+                symbol.chars().all(&char::is_whitespace) && symbol != NBSP && symbol != NNBSP;
 
             // Ignore characters wider that the total max width.
             if symbol.width() as u16 > self.max_line_width
@@ -496,6 +498,20 @@ mod test {
         assert_eq!(word_wrapper_space, vec!["AAAAAAAAAAAAAAA AAAA", "AAA",]);
     }
 
+    #[test]
+    fn line_composer_word_wrapper_nnbsp() {
+        let width = 20;
+        let text = "AAAAAAAAAAAAAAA AAAA\u{202f}AAA";
+        let (word_wrapper, _) = run_composer(Composer::WordWrapper { trim: true }, text, width);
+        assert_eq!(word_wrapper, vec!["AAAAAAAAAAAAAAA", "AAAA\u{202f}AAA",]);
+
+        // Ensure that if the character was a regular space, it would be wrapped differently.
+        let text_space = text.replace('\u{202f}', " ");
+        let (word_wrapper_space, _) =
+            run_composer(Composer::WordWrapper { trim: true }, &text_space, width);
+        assert_eq!(word_wrapper_space, vec!["AAAAAAAAAAAAAAA AAAA", "AAA",]);
+    }
+
     #[test]
     fn line_composer_word_wrapper_preserve_indentation() {
         let width = 20;

helix-view/src/editor.rs

@@ -676,6 +676,7 @@ pub enum WhitespaceRender {
         default: Option<WhitespaceRenderValue>,
         space: Option<WhitespaceRenderValue>,
         nbsp: Option<WhitespaceRenderValue>,
+        nnbsp: Option<WhitespaceRenderValue>,
         tab: Option<WhitespaceRenderValue>,
         newline: Option<WhitespaceRenderValue>,
     },
@@ -707,6 +708,14 @@ impl WhitespaceRender {
             }
         }
     }
+    pub fn nnbsp(&self) -> WhitespaceRenderValue {
+        match *self {
+            Self::Basic(val) => val,
+            Self::Specific { default, nnbsp, .. } => {
+                nnbsp.or(default).unwrap_or(WhitespaceRenderValue::None)
+            }
+        }
+    }
     pub fn tab(&self) -> WhitespaceRenderValue {
         match *self {
             Self::Basic(val) => val,
@@ -730,6 +739,7 @@ impl WhitespaceRender {
 pub struct WhitespaceCharacters {
     pub space: char,
     pub nbsp: char,
+    pub nnbsp: char,
     pub tab: char,
     pub tabpad: char,
     pub newline: char,
@@ -740,6 +750,7 @@ impl Default for WhitespaceCharacters {
         Self {
             space: '·',   // U+00B7
             nbsp: '⍽',    // U+237D
+            nnbsp: '␣',   // U+2423
             tab: '→',     // U+2192
             newline: '⏎', // U+23CE
             tabpad: ' ',