fix: respect recursive rules (#26)

JustinBeckwith · web-flow · commit 877a32bd1948 · 2019-02-13T12:06:41.000-08:00
diff --git a/package.json b/package.json
@@ -26,6 +26,7 @@
     "gaxios": "^1.6.0",
     "is-absolute-url": "^2.1.0",
     "meow": "^5.0.0",
+    "server-destroy": "^1.0.1",
     "update-notifier": "^2.5.0"
   },
   "devDependencies": {
@@ -37,6 +38,7 @@
     "@types/mocha": "^5.2.5",
     "@types/nock": "^9.3.1",
     "@types/node": "^11.9.3",
+    "@types/server-destroy": "^1.0.0",
     "@types/sinon": "^7.0.5",
     "@types/update-notifier": "^2.5.0",
     "codecov": "^3.1.0",
diff --git a/src/index.ts b/src/index.ts
@@ -1,6 +1,7 @@
 import {EventEmitter} from 'events';
 import * as gaxios from 'gaxios';
 import * as http from 'http';
+import enableDestroy = require('server-destroy');
 
 import {getLinks} from './links';
 
@@ -55,6 +56,7 @@ export class LinkChecker extends EventEmitter {
     if (!options.path.startsWith('http')) {
       const port = options.port || 5000 + Math.round(Math.random() * 1000);
       server = await this.startWebServer(options.path, port);
+      enableDestroy(server);
       options.path = `http://localhost:${port}`;
     }
     const results = await this.crawl(
@@ -64,7 +66,7 @@ export class LinkChecker extends EventEmitter {
       passed: results.filter(x => x.state === LinkState.BROKEN).length === 0
     };
     if (server) {
-      server.close();
+      server.destroy();
     }
     return result;
   }
@@ -141,7 +143,8 @@ export class LinkChecker extends EventEmitter {
       const urls = getLinks(data, opts.checkOptions.path);
       for (const url of urls) {
         // only crawl links that start with the same host
-        const crawl = url.startsWith(opts.checkOptions.path);
+        const crawl = opts.checkOptions.recurse! &&
+            url.startsWith(opts.checkOptions.path);
         await this.crawl({
           url,
           crawl,
diff --git a/test/fixtures/recurse/fake.html b/test/fixtures/recurse/fake.html
@@ -0,0 +1,6 @@
+<html>
+<body>
+<a href='/nothere.html'></a>
+<a href='http://www.google.com'></a>
+</body>
+</html>
diff --git a/test/fixtures/recurse/first.html b/test/fixtures/recurse/first.html
@@ -0,0 +1,6 @@
+<html>
+<body>
+<a href='/'></a>
+<a href='/second.html'></a>
+</body>
+</html>
diff --git a/test/fixtures/recurse/index.html b/test/fixtures/recurse/index.html
@@ -0,0 +1,5 @@
+<html>
+<body>
+<a href="first.html">go right over there</a>
+</body>
+</html>
diff --git a/test/fixtures/recurse/second.html b/test/fixtures/recurse/second.html
@@ -0,0 +1,5 @@
+<html>
+<body>
+<a href='http://fake.local'>Link to another castle</a>
+</body>
+</html>
diff --git a/test/test.ts b/test/test.ts
@@ -75,4 +75,20 @@ describe('linkinator', () => {
     assert.strictEqual(
         results.links.filter(x => x.state === LinkState.OK).length, 2);
   });
+
+  it('should perform a recursive scan', async () => {
+    // This test is making sure that we do a recursive scan of links,
+    // but also that we don't follow links to another site
+    const scope = nock('http://fake.local')
+                      .get('/')
+                      .replyWithFile(200, 'test/fixtures/recurse/fake.html');
+    const results = await check({path: 'test/fixtures/recurse', recurse: true});
+    assert.strictEqual(results.links.length, 5);
+    scope.done();
+  });
+
+  it('should not recurse by default', async () => {
+    const results = await check({path: 'test/fixtures/recurse'});
+    assert.strictEqual(results.links.length, 2);
+  });
 });

-Original file line number
+Diff line change
@@ @@ -0,0 +1,6 @@ @@
 +<html>
 +<body>
 +<a href='/nothere.html'></a>
 +<a href='http://www.google.com'></a>
 +</body>
 +</html>