Only strip overstrike when a syntax highlighting theme is used

Author: akirk

CHANGELOG.md

@@ -4,7 +4,7 @@
 
 ## Bugfixes
 
-- Strip overstriking to better support man pages, see #3517 (@akirk)
+- Strip overstriking before applying syntax highlighting to better support man pages, see #3517 (@akirk)
 - `--help` now correctly honors `--pager=builtin`. See #3516 (@keith-hall)
 
 ## Other

src/printer.rs

@@ -167,20 +167,12 @@ impl Printer for SimplePrinter<'_> {
                     }
                 }
             } else {
+                // SimplePrinter doesn't have syntax info, so we don't strip overstrike.
+                // This preserves raw data for plain output mode.
                 match handle {
-                    OutputHandle::IoWrite(handle) => {
-                        // Only strip overstrike for valid UTF-8, otherwise write raw bytes
-                        if let Ok(line) = std::str::from_utf8(line_buffer) {
-                            let line = strip_overstrike(line);
-                            handle.write_all(line.as_bytes())?;
-                        } else {
-                            handle.write_all(line_buffer)?;
-                        }
-                    }
+                    OutputHandle::IoWrite(handle) => handle.write_all(line_buffer)?,
                     OutputHandle::FmtWrite(handle) => {
-                        let line = String::from_utf8_lossy(line_buffer);
-                        let line = strip_overstrike(&line);
-                        write!(handle, "{line}")?;
+                        write!(handle, "{}", String::from_utf8_lossy(line_buffer))?;
                     }
                 }
             };
@@ -216,6 +208,7 @@ pub(crate) struct InteractivePrinter<'a> {
     background_color_highlight: Option<Color>,
     consecutive_empty_lines: usize,
     strip_ansi: bool,
+    strip_overstrike: bool,
 }
 
 impl<'a> InteractivePrinter<'a> {
@@ -314,6 +307,10 @@ impl<'a> InteractivePrinter<'a> {
             _ => false,
         };
 
+        // Strip overstrike only when we have syntax highlighting (not plain text).
+        // Overstrike is used by man pages for bold/underline formatting.
+        let strip_overstrike = !is_plain_text;
+
         Ok(InteractivePrinter {
             panel_width,
             colors,
@@ -327,6 +324,7 @@ impl<'a> InteractivePrinter<'a> {
             background_color_highlight,
             consecutive_empty_lines: 0,
             strip_ansi,
+            strip_overstrike,
         })
     }
 
@@ -640,7 +638,7 @@ impl Printer for InteractivePrinter<'_> {
             };
 
             // Strip overstrike sequences (used by man pages for bold/underline).
-            if line.contains('\x08') {
+            if self.strip_overstrike && line.contains('\x08') {
                 line = Cow::Owned(strip_overstrike(&line).into_owned());
             }
 

tests/examples/git-commit.man

@@ -0,0 +1,30 @@
+GIT-COMMIT(1)                     Git Manual                     GIT-COMMIT(1)
+
+NNAAMMEE
+       git-commit - Record changes to the repository
+
+SSYYNNOOPPSSIISS
+       ggiitt ccoommmmiitt [--aa | ----iinntteerraaccttiivvee | ----ppaattcchh] [--ss] [--vv] [--uu[_<_m_o_d_e_>]] [----aammeenndd]
+               [----ddrryy--rruunn] [(--cc | --CC | ----ssqquuaasshh) _<_c_o_m_m_i_t_> | ----ffiixxuupp [(aammeenndd|rreewwoorrdd)::]_<_c_o_m_m_i_t_>]
+               [--FF _<_f_i_l_e_> | --mm _<_m_s_g_>] [----rreesseett--aauutthhoorr] [----aallllooww--eemmppttyy]
+               [----aallllooww--eemmppttyy--mmeessssaaggee] [----nnoo--vveerriiffyy] [--ee] [----aauutthhoorr==_<_a_u_t_h_o_r_>]
+               [----ddaattee==_<_d_a_t_e_>] [----cclleeaannuupp==_<_m_o_d_e_>] [----[nnoo--]ssttaattuuss]
+               [--ii | --oo] [----ppaatthhssppeecc--ffrroomm--ffiillee==_<_f_i_l_e_> [----ppaatthhssppeecc--ffiillee--nnuull]]
+               [(----ttrraaiilleerr _<_t_o_k_e_n_>[(==|::)_<_v_a_l_u_e_>])...] [--SS[_<_k_e_y_i_d_>]]
+               [----] [_<_p_a_t_h_s_p_e_c_>...]
+
+DDEESSCCRRIIPPTTIIOONN
+       Create a new commit containing the current contents of the index and
+       the given log message describing the changes. The new commit is a
+       direct child of HEAD, usually the tip of the current branch, and the
+       branch is updated to point to it (unless no branch is associated with
+       the working tree, in which case HHEEAADD is "detached" as described in ggiitt--
+       cchheecckkoouutt(1)).
+
+       The content to be committed can be specified in several ways:
+
+        1. by using ggiitt--aadddd(1) to incrementally "add" changes to the index
+           before using the ccoommmmiitt command (Note: even modified files must be
+           "added");
+
+        2. by using ggiitt--rrmm(1) to remove files from the working tree and the

tests/examples/overstrike.txt

@@ -0,0 +1 @@
+BBold tteexxtt and _u_n_d_e_r_l_i_n_e

tests/integration_tests.rs

@@ -2541,6 +2541,46 @@ fn binary_as_text() {
         .stderr("");
 }
 
+#[test]
+fn no_strip_overstrike_for_plain_text() {
+    // Overstrike is preserved for plain text files (no syntax highlighting)
+    bat()
+        .arg("--color=never")
+        .arg("--decorations=never")
+        .arg("overstrike.txt")
+        .assert()
+        .success()
+        .stdout("B\x08Bold t\x08te\x08ex\x08xt\x08t and _\x08u_\x08n_\x08d_\x08e_\x08r_\x08l_\x08i_\x08n_\x08e\n")
+        .stderr("");
+}
+
+#[test]
+fn strip_overstrike_with_syntax_highlighting() {
+    // Overstrike is stripped when syntax highlighting is applied (e.g., for help)
+    bat()
+        .arg("--force-colorization")
+        .arg("--language=help")
+        .arg("overstrike.txt")
+        .assert()
+        .success()
+        .stdout(predicate::str::contains("Bold text and underline"))
+        .stderr("");
+}
+
+#[test]
+fn strip_overstrike_for_manpage_syntax() {
+    // Overstrike is stripped for .man files (Manpage syntax)
+    bat()
+        .arg("--force-colorization")
+        .arg("git-commit.man")
+        .assert()
+        .success()
+        .stdout(predicate::str::contains("NAME"))
+        .stdout(predicate::str::contains("git-commit - Record changes"))
+        .stdout(predicate::str::is_match(r"\x1b\[38;[0-9;]+m--interactive\x1b\[").unwrap())
+        .stderr("");
+}
+
 #[test]
 fn no_paging_arg() {
     bat()