fix(web): match .gitattributes patterns in subdirectories

CHANGELOG.md

@@ -17,6 +17,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ### Fixed
 - Made the default home page configurable with `DEFAULT_HOME_VIEW_PAGE`, defaulting to Code Search and supporting Ask. [#1677](https://github.com/sourcebot-dev/sourcebot/pull/1677)
 - Require authentication for the streaming and blocking Ask APIs in Public SaaS deployments. [#1679](https://github.com/sourcebot-dev/sourcebot/pull/1679)
+- Fixed `.gitattributes` `linguist-language` overrides being ignored for files in subdirectories. [#1692](https://github.com/sourcebot-dev/sourcebot/pull/1692)
 
 ## [5.1.14] - 2026-09-17
 

packages/web/src/lib/gitattributes.test.ts

@@ -0,0 +1,54 @@
+import { describe, expect, test } from 'vitest';
+import { parseGitAttributes, resolveLanguageFromGitAttributes } from './gitattributes';
+
+describe('resolveLanguageFromGitAttributes', () => {
+    test('a pattern without a slash matches files in any directory', () => {
+        const attrs = parseGitAttributes('*.h linguist-language=C\n');
+
+        expect(resolveLanguageFromGitAttributes('foo.h', attrs)).toBe('C');
+        expect(resolveLanguageFromGitAttributes('src/include/foo.h', attrs)).toBe('C');
+    });
+
+    test('a pattern with a leading slash is anchored to the repository root', () => {
+        const attrs = parseGitAttributes('/config.in linguist-language=Makefile\n');
+
+        expect(resolveLanguageFromGitAttributes('config.in', attrs)).toBe('Makefile');
+        expect(resolveLanguageFromGitAttributes('sub/config.in', attrs)).toBeUndefined();
+    });
+
+    test('a pattern with an inner slash is matched relative to the repository root', () => {
+        const attrs = parseGitAttributes('docs/*.txt linguist-language=Markdown\n');
+
+        expect(resolveLanguageFromGitAttributes('docs/intro.txt', attrs)).toBe('Markdown');
+        expect(resolveLanguageFromGitAttributes('./docs/intro.txt', attrs)).toBe('Markdown');
+        expect(resolveLanguageFromGitAttributes('./docs/./intro.txt', attrs)).toBe('Markdown');
+        expect(resolveLanguageFromGitAttributes('other/docs/intro.txt', attrs)).toBeUndefined();
+    });
+
+    test('the last matching rule wins', () => {
+        const attrs = parseGitAttributes('*.inc linguist-language=PHP\nlegacy/*.inc linguist-language=Pascal\n');
+
+        expect(resolveLanguageFromGitAttributes('src/a.inc', attrs)).toBe('PHP');
+        expect(resolveLanguageFromGitAttributes('legacy/a.inc', attrs)).toBe('Pascal');
+    });
+
+    test('dotfiles match slashless patterns', () => {
+        const attrs = parseGitAttributes('*.conf linguist-language=TOML\n');
+
+        expect(resolveLanguageFromGitAttributes('src/.config.conf', attrs)).toBe('TOML');
+    });
+
+    test.each(['-linguist-language', '!linguist-language'])('%s clears the previous language override', (reset) => {
+        const attrs = parseGitAttributes(`*.inc linguist-language=PHP\nlegacy/*.inc ${reset}\n`);
+
+        expect(resolveLanguageFromGitAttributes('legacy/a.inc', attrs)).toBeUndefined();
+    });
+
+    test('a later language value replaces a reset', () => {
+        const attrs = parseGitAttributes(
+            '*.inc linguist-language=PHP\nlegacy/*.inc -linguist-language\nlegacy/special.inc linguist-language=Pascal\n',
+        );
+
+        expect(resolveLanguageFromGitAttributes('legacy/special.inc', attrs)).toBe('Pascal');
+    });
+});

packages/web/src/lib/gitattributes.ts

@@ -58,14 +58,31 @@ export function parseGitAttributes(content: string): GitAttributes {
     return { rules };
 }
 
+// matchesGitAttributesPattern applies gitattributes pattern rules: a pattern
+// without a slash matches the file name at any depth, and any other pattern
+// is matched against the full path from the repository root.
+// @see https://git-scm.com/docs/gitattributes#_description
+function matchesGitAttributesPattern(filePath: string, pattern: string): boolean {
+    const normalizedFilePath = filePath
+        .split('/')
+        .filter((pathSegment) => pathSegment !== '.')
+        .join('/');
+    if (!pattern.includes('/')) {
+        const fileName = normalizedFilePath.slice(normalizedFilePath.lastIndexOf('/') + 1);
+        return micromatch.isMatch(fileName, pattern, { dot: true });
+    }
+    return micromatch.isMatch(normalizedFilePath, pattern.replace(/^\//, ''), { dot: true });
+}
+
 // resolveLanguageFromGitAttributes returns the linguist-language override for
 // the given file path based on the parsed .gitattributes rules, or undefined
 // if no rule matches. Last matching rule wins, consistent with gitattributes semantics.
 export function resolveLanguageFromGitAttributes(filePath: string, gitAttributes: GitAttributes): string | undefined {
     let language: string | undefined;
     for (const rule of gitAttributes.rules) {
-        if (micromatch.isMatch(filePath, rule.pattern) && rule.attrs['linguist-language']) {
-            language = rule.attrs['linguist-language'];
+        const languageAttribute = rule.attrs['linguist-language'];
+        if (matchesGitAttributesPattern(filePath, rule.pattern) && languageAttribute !== undefined) {
+            language = languageAttribute === 'false' || languageAttribute === 'unspecified' ? undefined : languageAttribute;
         }
     }
     return language;