Allow users to override the tab width

Fixes #274

CHANGELOG.md

@@ -11,6 +11,10 @@ names are recognised as JSON, e.g. `.jsonl`.
 
 Fixed crash in inline mode.
 
+Added an option `--tab-width` that controls how many spaces are used
+to display tabs. The default value is 8, consistent with older
+versions.
+
 ### Diffing
 
 Difftastic now diffs files in parallel when diffing whole directories,

sample_files/compare.expected

@@ -145,6 +145,9 @@ eeab25a68552f051a6392b5e713fbd29  -
 sample_files/syntax_error_before.js sample_files/syntax_error_after.js
 fe636ad27b1aa75e0b153dfe248023bb  -
 
+sample_files/tab_before.c sample_files/tab_after.c
+36ba3231eeba6f0b67a6be9db454de19  -
+
 sample_files/text_before.txt sample_files/text_after.txt
 dfc3495b8d5931029b479f0c878a3219  -
 

sample_files/tab_after.c

@@ -0,0 +1,6 @@
+#include <stdio.h>
+
+int main() {
+	printf("Goodbye World");
+	return 0;
+}

sample_files/tab_before.c

@@ -0,0 +1,6 @@
+#include <stdio.h>
+
+int main() {
+	printf("Hello World");
+	return 0;
+}

src/main.rs

@@ -142,6 +142,7 @@ fn main() {
             language_override,
             lhs_path,
             rhs_path,
+            tab_width,
             ..
         } => {
             let use_color = should_use_color(color_output);
@@ -164,6 +165,7 @@ fn main() {
                 diff_directories(
                     lhs_path,
                     rhs_path,
+                    tab_width,
                     node_limit,
                     byte_limit,
                     language_override,
@@ -183,6 +185,7 @@ fn main() {
                     &display_path,
                     lhs_path,
                     rhs_path,
+                    tab_width,
                     missing_as_empty,
                     node_limit,
                     byte_limit,
@@ -206,6 +209,7 @@ fn diff_file(
     display_path: &str,
     lhs_path: &Path,
     rhs_path: &Path,
+    tab_width: usize,
     missing_as_empty: bool,
     node_limit: u32,
     byte_limit: usize,
@@ -216,6 +220,7 @@ fn diff_file(
         display_path,
         &lhs_bytes,
         &rhs_bytes,
+        tab_width,
         node_limit,
         byte_limit,
         language_override,
@@ -226,6 +231,7 @@ fn diff_file_content(
     display_path: &str,
     lhs_bytes: &[u8],
     rhs_bytes: &[u8],
+    tab_width: usize,
     node_limit: u32,
     byte_limit: usize,
     language_override: Option<guess_language::Language>,
@@ -242,12 +248,13 @@ fn diff_file_content(
     }
 
     // TODO: don't replace tab characters inside string literals.
+    let tab_as_spaces = " ".repeat(tab_width);
     let mut lhs_src = String::from_utf8_lossy(lhs_bytes)
         .to_string()
-        .replace('\t', "        ");
+        .replace('\t', &tab_as_spaces);
     let mut rhs_src = String::from_utf8_lossy(rhs_bytes)
         .to_string()
-        .replace('\t', "        ");
+        .replace('\t', &tab_as_spaces);
 
     // Ignore the trailing newline, if present.
     // TODO: highlight if this has changes (#144).
@@ -371,6 +378,7 @@ fn diff_file_content(
 fn diff_directories<'a>(
     lhs_dir: &'a Path,
     rhs_dir: &'a Path,
+    tab_width: usize,
     node_limit: u32,
     byte_limit: usize,
     language_override: Option<guess_language::Language>,
@@ -390,6 +398,7 @@ fn diff_directories<'a>(
             &rel_path.to_string_lossy(),
             &lhs_path,
             &rhs_path,
+            tab_width,
             true,
             node_limit,
             byte_limit,
@@ -521,7 +530,7 @@ fn max_num_nodes(roots_vec: &[(Vec<&syntax::Syntax>, Vec<&syntax::Syntax>)]) ->
 #[cfg(test)]
 mod tests {
     use super::*;
-    use crate::options::{DEFAULT_BYTE_LIMIT, DEFAULT_NODE_LIMIT};
+    use crate::options::{DEFAULT_BYTE_LIMIT, DEFAULT_NODE_LIMIT, DEFAULT_TAB_WIDTH};
 
     #[test]
     fn test_diff_identical_content() {
@@ -530,6 +539,7 @@ mod tests {
             "foo.el",
             s.as_bytes(),
             s.as_bytes(),
+            DEFAULT_TAB_WIDTH,
             DEFAULT_NODE_LIMIT,
             DEFAULT_BYTE_LIMIT,
             None,

src/options.rs

@@ -8,6 +8,7 @@ use crate::{guess_language, style::BackgroundColor};
 
 pub const DEFAULT_NODE_LIMIT: u32 = 30_000;
 pub const DEFAULT_BYTE_LIMIT: usize = 1_000_000;
+pub const DEFAULT_TAB_WIDTH: usize = 8;
 
 const USAGE: &str = concat!(env!("CARGO_BIN_NAME"), " [OPTIONS] OLD-PATH NEW-PATH");
 
@@ -64,6 +65,17 @@ fn app() -> clap::Command<'static> {
                 .validator(|s| s.parse::<usize>())
                 .required(false),
         )
+        .arg(
+            Arg::new("tab-width")
+                .long("tab-width")
+                .takes_value(true)
+                .value_name("NUM_SPACES")
+                .long_help("Treat a tab as this many spaces.")
+                .env("DFT_TAB_WIDTH")
+                .default_value(formatcp!("{}", DEFAULT_TAB_WIDTH))
+                .validator(|s| s.parse::<usize>())
+                .required(false),
+        )
         .arg(
             Arg::new("display").long("display")
                 .possible_values(["side-by-side", "side-by-side-show-both", "inline", ])
@@ -142,6 +154,7 @@ pub enum Mode {
         byte_limit: usize,
         print_unchanged: bool,
         missing_as_empty: bool,
+        tab_width: usize,
         display_mode: DisplayMode,
         background_color: BackgroundColor,
         color_output: ColorOutput,
@@ -292,6 +305,12 @@ pub fn parse_args() -> Mode {
         .parse::<usize>()
         .expect("Value already validated by clap");
 
+    let tab_width = matches
+        .value_of("tab-width")
+        .expect("Always present as we've given clap a default")
+        .parse::<usize>()
+        .expect("Value already validated by clap");
+
     let print_unchanged = !matches.is_present("skip-unchanged");
     let missing_as_empty = matches.is_present("missing-as-empty");
 
@@ -300,6 +319,7 @@ pub fn parse_args() -> Mode {
         byte_limit,
         print_unchanged,
         missing_as_empty,
+        tab_width,
         display_mode,
         background_color,
         color_output,