Avoid unnecessary proxy URL lookups (#3377)

Co-authored-by: Nicolas Dorseuil <[email protected]>

Author: conico974

packages/gitbook/src/lib/data/urls.test.ts

@@ -362,6 +362,35 @@ describe('getURLLookupAlternatives', () => {
             ],
         });
     });
+
+    it('should skip proxy root URLs', () => {
+        expect(
+            getURLLookupAlternatives(
+                new URL('https://proxy.gitbook.site/sites/site_foo/hello/world')
+            )
+        ).toEqual({
+            revision: undefined,
+            changeRequest: undefined,
+            basePath: undefined,
+            urls: [
+                {
+                    url: 'https://proxy.gitbook.site/sites/site_foo',
+                    extraPath: 'hello/world',
+                    primary: false,
+                },
+                {
+                    url: 'https://proxy.gitbook.site/sites/site_foo/hello',
+                    extraPath: 'world',
+                    primary: false,
+                },
+                {
+                    url: 'https://proxy.gitbook.site/sites/site_foo/hello/world',
+                    extraPath: '',
+                    primary: true,
+                },
+            ],
+        });
+    });
 });
 
 describe('normalizeURL', () => {

packages/gitbook/src/lib/data/urls.ts

@@ -1,3 +1,5 @@
+import { isProxyRootRequest } from '../proxy';
+
 /**
  * For a given GitBook URL, return a list of alternative URLs that could be matched against to lookup the content.
  * The approach is optimized to aim at reusing cached lookup results as much as possible.
@@ -39,6 +41,11 @@ export function getURLLookupAlternatives(input: URL) {
             return;
         }
 
+        // We don't want to push the url if it is the root proxy URL (either https://proxy.gitbook.site or https://proxy.gitbook.site/sites)
+        if (isProxyRootRequest(adding)) {
+            return;
+        }
+
         alternatives.push({
             url: adding.toString(),
             extraPath,

packages/gitbook/src/lib/proxy.test.ts

@@ -1,5 +1,5 @@
 import { describe, expect, it } from 'bun:test';
-import { getProxyRequestIdentifier, isProxyRequest } from './proxy';
+import { getProxyRequestIdentifier, isProxyRequest, isProxyRootRequest } from './proxy';
 
 describe('isProxyRequest', () => {
     it('should return true for proxy requests', () => {
@@ -19,3 +19,32 @@ describe('getProxyRequestIdentifier', () => {
         expect(getProxyRequestIdentifier(proxyRequestURL)).toBe('sites/site_foo');
     });
 });
+
+describe('isProxyRootRequest', () => {
+    it('should return true for root proxy requests', () => {
+        const rootProxyRequestURL = new URL('https://proxy.gitbook.site/');
+        expect(isProxyRootRequest(rootProxyRequestURL)).toBe(true);
+    });
+
+    it('should return true for sites root proxy requests', () => {
+        const sitesRootProxyRequestURL = new URL('https://proxy.gitbook.site/sites');
+        expect(isProxyRootRequest(sitesRootProxyRequestURL)).toBe(true);
+    });
+
+    it('should return true for sites root proxy requests with trailing slash', () => {
+        const sitesRootProxyRequestURL = new URL('https://proxy.gitbook.site/sites/');
+        expect(isProxyRootRequest(sitesRootProxyRequestURL)).toBe(true);
+    });
+
+    it('should return false for non proxy requests', () => {
+        const nonRootProxyRequestURL = new URL('https://example.com/docs/foo/hello/world');
+        expect(isProxyRootRequest(nonRootProxyRequestURL)).toBe(false);
+    });
+
+    it('should return false for non-root proxy requests', () => {
+        const nonRootProxyRequestURL = new URL(
+            'https://proxy.gitbook.site/sites/site_foo/hello/world'
+        );
+        expect(isProxyRootRequest(nonRootProxyRequestURL)).toBe(false);
+    });
+});

packages/gitbook/src/lib/proxy.ts

@@ -13,3 +13,17 @@ export function getProxyRequestIdentifier(requestURL: URL): string {
     const pathname = requestURL.pathname.slice(1).split('/');
     return pathname.slice(0, 2).join('/');
 }
+
+/**
+ * Check if the request is for the root of the proxy site (i.e. https://proxy.gitbook.site/ or https://proxy.gitbook.site/sites).
+ * @param requestURL The URL of the request to check if it is a proxy root request.
+ */
+export function isProxyRootRequest(requestURL: URL): boolean {
+    // Check if the request is for the root of the proxy site
+    return (
+        isProxyRequest(requestURL) &&
+        (requestURL.pathname === '/' ||
+            requestURL.pathname === '/sites' ||
+            requestURL.pathname === '/sites/')
+    );
+}