Add word wrapping mode (#3597)

* feat: add word wrapping mode for --wrap flag

* Run `cargo fmt` and add CHANGELOG entry

* Add word wrap tests, update manpage and shell completions

- Add integration tests for word wrapping: basic word boundary breaking,
  fallback to character wrapping for long words, line numbers, and
  short lines that fit without wrapping
- Update manpage to document the new 'word' wrapping mode
- Update bash, fish, zsh, and PowerShell completions with 'word' option
- Avoid unnecessary clone of `line_buf` when word wrap is disabled

* make clippy and cargo fmt happy

---------

Co-authored-by: Keith Hall <keith-hall@users.noreply.github.com>

CHANGELOG.md

@@ -4,6 +4,7 @@
 
 ## Features
 
+- Add word wrapping mode via `--wrap=word`, see #3597 (@veeceey)
 - Implement `--unbuffered` mode for streaming input, allowing partial lines to display immediately (e.g. `tail -f | bat -u`). Closes #3555, see #3583 (@mainnebula)
 - Added an initial `flake.nix` for a ready made development environment; see #3578 (@vorburger)
 - Add `--quiet-empty` (`-E`) flag to suppress output when input is empty. Closes #1936, see #3563 (@NORMAL-EX)

assets/completions/_bat.ps1.in

@@ -9,7 +9,7 @@ Register-ArgumentCompleter -Native -CommandName '{{PROJECT_EXECUTABLE}}' -Script
     $ArrayCompletion = @('bash', 'fish', 'zsh', 'ps1')
     $ArrayWhen       = @('auto', 'never', 'always')
     $ArrayYesNo      = @('never', 'always')
-    $ArrayWrap       = @('always', 'never', 'character')
+    $ArrayWrap       = @('always', 'never', 'character', 'word')
     $ArrayBinary     = @('no-printing', 'as-text')
     $ArrayPrint      = @('unicode', 'caret')
 
@@ -135,7 +135,7 @@ Register-ArgumentCompleter -Native -CommandName '{{PROJECT_EXECUTABLE}}' -Script
             [CompletionResult]::new('--file-name'             , 'file-name'             , [CompletionResultType]::ParameterName, 'Specify the name to display for a file.')
             [CompletionResult]::new('--diff-context'          , 'diff-context'          , [CompletionResultType]::ParameterName, 'diff-context')
             [CompletionResult]::new('--tabs'                  , 'tabs'                  , [CompletionResultType]::ParameterName, 'Set the tab width to T spaces.')
-            [CompletionResult]::new('--wrap'                  , 'wrap'                  , [CompletionResultType]::ParameterName, 'Specify the text-wrapping mode (*auto*, character).')
+            [CompletionResult]::new('--wrap'                  , 'wrap'                  , [CompletionResultType]::ParameterName, 'Specify the text-wrapping mode (*auto*, never, character, word).')
             [CompletionResult]::new('--terminal-width'        , 'terminal-width'        , [CompletionResultType]::ParameterName, 'Explicitly set the width of the terminal instead of determining it automatically. If prefixed with ''+'' or ''-'', the value will be treated as an offset to the actual terminal width. See also: ''--wrap''.')
             [CompletionResult]::new('--color'                 , 'color'                 , [CompletionResultType]::ParameterName, 'When to use colors (*auto*, never, always).')
             [CompletionResult]::new('--italic-text'           , 'italic-text'           , [CompletionResultType]::ParameterName, 'Use italics in output (always, *never*)')

assets/completions/bat.bash.in

@@ -117,7 +117,7 @@ _bat() {
 		return 0
 		;;
 	--wrap)
-		COMPREPLY=($(compgen -W "auto never character" -- "$cur"))
+		COMPREPLY=($(compgen -W "auto never character word" -- "$cur"))
 		return 0
 		;;
 	--binary)

assets/completions/bat.fish.in

@@ -118,6 +118,7 @@ set -l wrap_opts '
     auto\tdefault
     never\t
     character\t
+    word\t
 '
 
 # While --tabs theoretically takes any number, most people should be OK with these.

assets/completions/bat.zsh.in

@@ -34,7 +34,7 @@ _{{PROJECT_EXECUTABLE}}_main() {
         '(-d --diff)'--diff'[only show lines that have been added/removed/modified]'
         --diff-context='[specify lines of context around added/removed/modified lines when using `--diff`]:lines'
         --tabs='[set the tab width]:tab width [4]'
-        --wrap='[specify the text-wrapping mode]:mode [auto]:(auto never character)'
+        --wrap='[specify the text-wrapping mode]:mode [auto]:(auto never character word)'
         '!(--wrap)'{-S,--chop-long-lines}
         --terminal-width='[explicitly set the width of the terminal instead of determining it automatically]:width'
         '(-n --number --diff --diff-context)'{-n,--number}'[show line numbers]'

assets/manual/bat.1.in

@@ -102,8 +102,10 @@ Set the tab width to T spaces. Use a width of 0 to pass tabs through directly
 .HP
 \fB\-\-wrap\fR <mode>
 .IP
-Specify the text\-wrapping mode (*auto*, never, character). The '\-\-terminal\-width' option
+Specify the text\-wrapping mode (*auto*, never, character, word). The '\-\-terminal\-width' option
-can be used in addition to control the output width.
+can be used in addition to control the output width. In \fBword\fR mode, lines are broken at
+whitespace boundaries. If a single word exceeds the terminal width, it falls back to
+character wrapping.
 .HP
 \fB\-S\fR, \fB\-\-chop\-long\-lines\fR
 .IP

doc/long-help.txt

@@ -61,8 +61,8 @@ Options:
           Set the tab width to T spaces. Use a width of 0 to pass tabs through directly
 
       --wrap <mode>
-          Specify the text-wrapping mode (*auto*, never, character). The '--terminal-width' option
+          Specify the text-wrapping mode (*auto*, never, character, word). The '--terminal-width'
-          can be used in addition to control the output width.
+          option can be used in addition to control the output width.
 
   -S, --chop-long-lines
           Truncate all lines longer than screen width. Alias for '--wrap=never'.

doc/short-help.txt

@@ -26,7 +26,7 @@ Options:
       --tabs <T>
           Set the tab width to T spaces.
       --wrap <mode>
-          Specify the text-wrapping mode (*auto*, never, character).
+          Specify the text-wrapping mode (*auto*, never, character, word).
   -S, --chop-long-lines
           Truncate all lines longer than screen width. Alias for '--wrap=never'.
   -n, --number

src/bin/bat/app.rs

@@ -405,6 +405,7 @@ impl App {
                 } else {
                     match self.matches.get_one::<String>("wrap").map(|s| s.as_str()) {
                         Some("character") => WrappingMode::Character,
+                        Some("word") => WrappingMode::Word,
                         Some("never") => WrappingMode::NoWrapping(true),
                         Some("auto") | None => {
                             if self.interactive_output || maybe_term_width.is_some() {

src/bin/bat/clap_app.rs

@@ -211,11 +211,11 @@ pub fn build_app(interactive_output: bool) -> Command {
                 .long("wrap")
                 .overrides_with("wrap")
                 .value_name("mode")
-                .value_parser(["auto", "never", "character"])
+                .value_parser(["auto", "never", "character", "word"])
                 .default_value("auto")
                 .hide_default_value(true)
-                .help("Specify the text-wrapping mode (*auto*, never, character).")
+                .help("Specify the text-wrapping mode (*auto*, never, character, word).")
-                .long_help("Specify the text-wrapping mode (*auto*, never, character). \
+                .long_help("Specify the text-wrapping mode (*auto*, never, character, word). \
                            The '--terminal-width' option can be used in addition to \
                            control the output width."),
         )

src/printer.rs

@@ -781,11 +781,21 @@ impl Printer for InteractivePrinter<'_> {
                             // Displayed width of line_buf
                             let mut current_width = 0;
 
+                            let word_wrap = matches!(self.config.wrapping_mode, WrappingMode::Word);
+
+                            // For word wrapping, track last whitespace position.
+                            let mut last_ws_idx: Option<usize> = None;
+
                             for c in text.chars() {
                                 // calculate the displayed width for next character
                                 let cw = c.width().unwrap_or(0);
                                 current_width += cw;
 
+                                // Track whitespace positions for word wrapping.
+                                if word_wrap && c.is_whitespace() {
+                                    last_ws_idx = Some(line_buf.len());
+                                }
+
                                 // if next character cannot be printed on this line,
                                 // flush the buffer.
                                 if current_width > max_width {
@@ -807,13 +817,37 @@ impl Printer for InteractivePrinter<'_> {
                                         }
                                     }
 
+                                    // Determine the break point and remainder
+                                    // for word wrapping.
+                                    let (emit_end, rest_start) = if word_wrap {
+                                        if let Some(ws_idx) = last_ws_idx {
+                                            // Skip the whitespace character itself
+                                            // and carry the rest to the next line.
+                                            let rs = ws_idx
+                                                + line_buf[ws_idx..]
+                                                    .chars()
+                                                    .next()
+                                                    .map(|ch| ch.len_utf8())
+                                                    .unwrap_or(0);
+                                            (ws_idx, Some(rs))
+                                        } else {
+                                            (line_buf.len(), None)
+                                        }
+                                    } else {
+                                        (line_buf.len(), None)
+                                    };
+
                                     // It wraps.
                                     write!(
                                         handle,
                                         "{}{}\n{}",
                                         as_terminal_escaped(
                                             style,
-                                            &format!("{}{line_buf}", self.ansi_style),
+                                            &format!(
+                                                "{}{}",
+                                                self.ansi_style,
+                                                &line_buf[..emit_end]
+                                            ),
                                             self.config.true_color,
                                             self.config.colored_output,
                                             self.config.use_italic_text,
@@ -826,8 +860,21 @@ impl Printer for InteractivePrinter<'_> {
                                     cursor = 0;
                                     max_width = cursor_max;
 
-                                    line_buf.clear();
+                                    if let Some(rs) = rest_start {
-                                    current_width = cw;
+                                        // Word wrap: carry remainder to next line.
+                                        let remainder = line_buf[rs..].to_string();
+                                        let rem_width: usize = remainder
+                                            .chars()
+                                            .map(|ch| ch.width().unwrap_or(0))
+                                            .sum();
+                                        line_buf.clear();
+                                        line_buf.push_str(&remainder);
+                                        current_width = rem_width + cw;
+                                    } else {
+                                        line_buf.clear();
+                                        current_width = cw;
+                                    }
+                                    last_ws_idx = None;
                                 }
 
                                 line_buf.push(c);

src/wrapping.rs

@@ -1,6 +1,7 @@
 #[derive(Debug, Clone, Copy, PartialEq, Eq)]
 pub enum WrappingMode {
     Character,
+    Word,
     // The bool specifies whether wrapping has been explicitly disabled by the user via --wrap=never
     NoWrapping(bool),
 }

tests/examples/word-wrap.txt

@@ -0,0 +1,3 @@
+The quick brown fox jumps over the lazy dog and then runs away
+superlongwordthatdefinitelyexceedstheterminalwidthandshouldfallbacktocharacterwrapping
+short words here

tests/integration_tests.rs

@@ -3791,3 +3791,62 @@ fn unbuffered_mode_plain_output() {
         .success()
         .stdout("hello world\n");
 }
+
+#[test]
+fn word_wrap_breaks_at_word_boundaries() {
+    bat()
+        .arg("word-wrap.txt")
+        .arg("--wrap=word")
+        .arg("--terminal-width=40")
+        .arg("--style=plain")
+        .arg("--decorations=always")
+        .arg("--color=never")
+        .assert()
+        .success()
+        .stdout(
+            "\
+The quick brown fox jumps over the lazy
+dog and then runs away
+superlongwordthatdefinitelyexceedstheter
+minalwidthandshouldfallbacktocharacterwr
+apping
+short words here
+",
+        );
+}
+
+#[test]
+fn word_wrap_with_line_numbers() {
+    bat()
+        .arg("word-wrap.txt")
+        .arg("--wrap=word")
+        .arg("--terminal-width=40")
+        .arg("--style=numbers")
+        .arg("--decorations=always")
+        .arg("--color=never")
+        .assert()
+        .success()
+        .stdout(
+            "   1 The quick brown fox jumps over the
+     lazy dog and then runs away
+   2 superlongwordthatdefinitelyexceedst
+     heterminalwidthandshouldfallbacktoc
+     haracterwrapping
+   3 short words here
+",
+        );
+}
+
+#[test]
+fn word_wrap_short_line_no_wrap() {
+    bat()
+        .arg("--wrap=word")
+        .arg("--terminal-width=80")
+        .arg("--style=plain")
+        .arg("--decorations=always")
+        .arg("--color=never")
+        .arg("single-line.txt")
+        .assert()
+        .success()
+        .stdout("Single Line\n");
+}