From 17e6952ab891853b954c53ae7d46f7a494e1f991 Mon Sep 17 00:00:00 2001
From: krikera <prafulrai522@gmail.com>
Date: Sat, 31 May 2025 02:58:47 +0530
Subject: [PATCH] Fix UTF-8 BOM file type detection for first-line syntax
 patterns - Fixes #3314

---
 CHANGELOG.md  |  1 +
 src/assets.rs | 41 ++++++++++++++++++++++++++++++++++++++++-
 2 files changed, 41 insertions(+), 1 deletion(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index ecf1c76b..73d19363 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -7,6 +7,7 @@
 
 ## Bugfixes
 
+- Fix UTF-8 BOM not being stripped for syntax detection, see #3314 (@krikera)
 - Fix `BAT_THEME_DARK` and `BAT_THEME_LIGHT` being ignored, see issue #3171 and PR #3168 (@bash)
 - Prevent `--list-themes` from outputting default theme info to stdout when it is piped, see #3189 (@einfachIrgendwer0815)
 - Rename some submodules to fix Dependabot submodule updates, see issue #3198 and PR #3201 (@victor-gp)
diff --git a/src/assets.rs b/src/assets.rs
index e6c50219..a5608e3b 100644
--- a/src/assets.rs
+++ b/src/assets.rs
@@ -298,7 +298,11 @@ impl HighlightingAssets {
         let syntax_set = self.get_syntax_set()?;
         Ok(String::from_utf8(reader.first_line.clone())
             .ok()
-            .and_then(|l| syntax_set.find_syntax_by_first_line(&l))
+            .and_then(|l| {
+                // Strip UTF-8 BOM if present
+                let line = l.strip_prefix('\u{feff}').unwrap_or(&l);
+                syntax_set.find_syntax_by_first_line(line)
+            })
             .map(|syntax| SyntaxReferenceInSet { syntax, syntax_set }))
     }
 }
@@ -533,6 +537,41 @@ mod tests {
         );
     }
 
+    #[test]
+    fn syntax_detection_first_line_with_utf8_bom() {
+        let test = SyntaxDetectionTest::new();
+
+        // Test that XML files are detected correctly even with UTF-8 BOM
+        // The BOM should be stripped before first-line syntax detection
+        let xml_with_bom = "\u{feff}<?xml version=\"1.0\" encoding=\"utf-8\"?>";
+        assert_eq!(
+            test.syntax_for_file_with_content("unknown_file", xml_with_bom),
+            "XML"
+        );
+
+        // Test the specific .csproj case mentioned in the issue
+        // Even if .csproj has extension mapping, this tests first-line fallback
+        let csproj_content_with_bom = "\u{feff}<?xml version=\"1.0\" encoding=\"utf-8\"?>\n<Project ToolsVersion=\"15.0\" xmlns=\"http://schemas.microsoft.com/developer/msbuild/2003\">";
+        assert_eq!(
+            test.syntax_for_file_with_content("test.csproj", csproj_content_with_bom),
+            "XML"
+        );
+
+        // Test that shell scripts are detected correctly even with UTF-8 BOM
+        let script_with_bom = "\u{feff}#!/bin/bash";
+        assert_eq!(
+            test.syntax_for_file_with_content("unknown_script", script_with_bom),
+            "Bourne Again Shell (bash)"
+        );
+
+        // Test that PHP files are detected correctly even with UTF-8 BOM
+        let php_with_bom = "\u{feff}<?php";
+        assert_eq!(
+            test.syntax_for_file_with_content("unknown_php", php_with_bom),
+            "PHP"
+        );
+    }
+
     #[test]
     fn syntax_detection_with_custom_mapping() {
         let mut test = SyntaxDetectionTest::new();