feat: scan all href, src, or other valid url attrs (#25)

JustinBeckwith · Feb 13, 2019 · ddf1fa4 · ddf1fa4
1 parent 629b83c
commit ddf1fa4
Show file tree

Hide file tree

Showing 5 changed files with 46 additions and 7 deletions.
diff --git a/src/cli.ts b/src/cli.ts
@@ -3,7 +3,7 @@
 import * as meow from 'meow';
 import * as updateNotifier from 'update-notifier';
 import chalk from 'chalk';
-import {LinkChecker, LinkState, LinkResult, CheckOptions, check} from './index';
+import {LinkChecker, LinkState, LinkResult, CheckOptions} from './index';
 
 const pkg = require('../../package.json');
 updateNotifier({pkg}).notify();

diff --git a/src/links.ts b/src/links.ts
@@ -2,14 +2,39 @@ import * as cheerio from 'cheerio';
 import isAbsoluteUrl = require('is-absolute-url');
 import {URL} from 'url';
 
+const linksAttr = {
+  background: ['body'],
+  cite: ['blockquote', 'del', 'ins', 'q'],
+  data: ['object'],
+  href: ['a', 'area', 'embed', 'link'],
+  icon: ['command'],
+  longdesc: ['frame', 'iframe'],
+  manifest: ['html'],
+  poster: ['video'],
+  pluginspage: ['embed'],
+  pluginurl: ['embed'],
+  src: [
+    'audio', 'embed', 'frame', 'iframe', 'img', 'input', 'script', 'source',
+    'track', 'video'
+  ],
+} as {[index: string]: string[]};
+
 export function getLinks(source: string, baseUrl: string) {
   const $ = cheerio.load(source);
-  const links = $('a').toArray().map(e => e.attribs.href).filter(x => !!x);
-  const sanitized = links.map(link => {
-    const slink =
-        isAbsoluteUrl(link) ? new URL(link) : (new URL(link, baseUrl));
-    slink.hash = '';
-    return slink.href;
+  const links = new Array<string>();
+  Object.keys(linksAttr).forEach(attr => {
+    const elements = linksAttr[attr].map(tag => `${tag}[${attr}]`).join(',');
+    $(elements).each((i, element) => {
+      links.push(element.attribs[attr]);
+    });
   });
+  const sanitized = links.filter(link => !!link)
+                        .map(link => normalizeLink(link, baseUrl).href);
   return sanitized;
 }
+
+function normalizeLink(link: string, baseUrl: string): URL {
+  const slink = isAbsoluteUrl(link) ? new URL(link) : (new URL(link, baseUrl));
+  slink.hash = '';
+  return slink;
+}
diff --git a/test/fixtures/image/boo.jpg b/test/fixtures/image/boo.jpg
diff --git a/test/fixtures/image/index.html b/test/fixtures/image/index.html
@@ -0,0 +1,6 @@
+<html>
+<body>
+<img src="missing.png">
+<img src="boo.jpg">
+</body>
+</html>
diff --git a/test/test.ts b/test/test.ts
@@ -67,4 +67,12 @@ describe('linkinator', () => {
     assert.strictEqual(
         results.links.filter(x => x.state === LinkState.SKIPPED).length, 1);
   });
+
+  it('should detect broken image links', async () => {
+    const results = await check({path: 'test/fixtures/image'});
+    assert.strictEqual(
+        results.links.filter(x => x.state === LinkState.BROKEN).length, 1);
+    assert.strictEqual(
+        results.links.filter(x => x.state === LinkState.OK).length, 2);
+  });
 });