fix(backend): decode percent-encoded characters in generic git URL repo names

CHANGELOG.md

@@ -20,6 +20,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Upgraded `engine.io` to `^6.6.11`, `source-map-js` to `^1.2.2`, `smol-toml` to `^1.9.0`, `fast-copy` to `^3.1.0`, and `dompurify` to `^3.4.16`. [#1714](https://github.com/sourcebot-dev/sourcebot/pull/1714)
 - Upgraded `golang.org/x/crypto` to `v0.56.0` and OpenTelemetry-Go exporters and SDK to `v1.45.0` in Zoekt. [#1716](https://github.com/sourcebot-dev/sourcebot/pull/1716)
 - Upgraded `@modelcontextprotocol/sdk` to `^1.32.1`. [#1719](https://github.com/sourcebot-dev/sourcebot/pull/1719)
+- Decoded percent-encoded characters in repo names derived from direct generic git URLs. [#1666](https://github.com/sourcebot-dev/sourcebot/pull/1666)
 - Fixed `.gitattributes` `linguist-language` overrides being ignored for files in subdirectories. [#1692](https://github.com/sourcebot-dev/sourcebot/pull/1692)
 
 ## [5.1.15] - 2026-09-29

packages/backend/src/repoCompileUtils.test.ts

@@ -334,4 +334,24 @@ describe('compileGenericGitHostConfig_url', () => {
         const metadata = result[0].metadata as { gitConfig?: Record<string, string> };
         expect(metadata.gitConfig!['zoekt.name']).toBe('github.com/test/repo');
     });
+
+    test('should decode percent-encoded characters in the repo name', async () => {
+        mockedIsUrlAValidGitRepo.mockResolvedValue(true);
+
+        const config = {
+            type: 'git' as const,
+            url: 'https://github.com/test/Project%20Name%20With%20Spaces.git',
+        };
+
+        const result = await compileGenericGitHostConfig_url(config, 1);
+
+        expect(result).toHaveLength(1);
+        // The repo name should have decoded spaces, not %20
+        expect(result[0].name).toBe('github.com/test/Project Name With Spaces');
+        expect(result[0].displayName).toBe('github.com/test/Project Name With Spaces');
+
+        const metadata = result[0].metadata as { gitConfig?: Record<string, string> };
+        expect(metadata.gitConfig!['zoekt.name']).toBe('github.com/test/Project Name With Spaces');
+        expect(metadata.gitConfig!['zoekt.display-name']).toBe('github.com/test/Project Name With Spaces');
+    });
 });

packages/backend/src/repoCompileUtils.ts

@@ -725,7 +725,28 @@ export const compileGenericGitHostConfig_url = async (
 
     // @note: matches the naming here:
     // https://github.com/sourcebot-dev/zoekt/blob/main/gitindex/index.go#L293
-    const repoName = path.join(remoteUrl.host, remoteUrl.pathname.replace(/\.git$/, ''));
+    // Decode URL-encoded characters (e.g., %20 -> space) to ensure consistent repo names
+    let decodedPathname: string;
+    try {
+        decodedPathname = decodeURIComponent(remoteUrl.pathname);
+    } catch (e) {
+        if (e instanceof URIError) {
+            const warning = `Skipping ${remoteUrl.toString()} - malformed URL encoding.`;
+            logger.warn(warning);
+            reportRepositoryDiscoveryIssue({
+                code: "INVALID_REPOSITORY_SOURCE",
+                effect: "TARGET_SKIPPED",
+                subject: {
+                    kind: "url",
+                    value: remoteUrl.toString(),
+                },
+                message: warning,
+            });
+            return [];
+        }
+        throw e;
+    }
+    const repoName = path.join(remoteUrl.host, decodedPathname.replace(/\.git$/, ''));
 
     const repo: RepoData = {
         external_codeHostType: 'genericGitHost',