Merge pull request #9 from estrattonbailey/fix-recursion-failure

Fix infinite loop in `HtmlParser`
holtwick · Oct 11, 2023 · 3669c26 · 3669c26
2 parents e00555f + 9a3b2d7
commit 3669c26
Show file tree

Hide file tree

Showing 2 changed files with 27 additions and 4 deletions.
diff --git a/src/htmlparser.spec.ts b/src/htmlparser.spec.ts
@@ -8,8 +8,8 @@ describe("htmlparser", () => {
     <title>Test</title>
     </head>
     <body>
-    <pre><code class="lang-jsx">               
-      <span class="hljs-tag">&lt;/<span class="hljs-name">a</span>&gt;</span><span class="hljs-tag">&lt;/<span class="hljs-name">li</span>&gt;</span>)}    
+    <pre><code class="lang-jsx">
+      <span class="hljs-tag">&lt;/<span class="hljs-name">a</span>&gt;</span><span class="hljs-tag">&lt;/<span class="hljs-name">li</span>&gt;</span>)}
     </code></pre>
     </body>
     </html>`
@@ -107,4 +107,24 @@ $('body')
       </script>"
     `)
   })
+
+  it("should not recurse on bad fragment", () => {
+    const tests = [
+      '<',
+      '<<',
+      '<<<',
+      '<<div',
+      '<div<div',
+      '>',
+      '>>',
+      '>>>',
+      '>>div',
+      '>div>div',
+    ]
+
+    for (const t of tests) {
+      const { textContent: text } = parseHTML(t) as VHTMLDocument
+      expect(text).toEqual(t)
+    }
+  })
 })
diff --git a/src/htmlparser.ts b/src/htmlparser.ts
@@ -81,19 +81,22 @@ export class HtmlParser {
 
       if (treatAsChars) {
         index = html.indexOf('<')
+        let offset = index
 
         if (index === 0) {
           // First char is a < so find the next one
           index = html.substring(1).indexOf('<')
+          // We're at substring(1) so add 1 to the index
+          offset = offset + 1
         }
 
         if (index === -1) {
           characters = html
           html = ''
         }
         else {
-          characters = html.substring(0, index)
-          html = html.substring(index)
+          characters = html.substring(0, offset)
+          html = html.substring(offset)
         }
 
         if (!this.options.ignoreWhitespaceText || !/^\s*$/.test(characters))