perf: skip unclosed suffixes when stripping review markdown tags (#20329)

mobile/src/components/pr-sidebar/markdown-blocks.ts

@@ -44,7 +44,14 @@ const SUMMARY = /<summary\b[^>]*>([\s\S]*?)<\/summary>/i
 // show literally. Conservative: only matches `<tag ...>` / `</tag>` shapes, so a bare
 // "a < b" in prose is left alone.
 export function stripHtmlTags(text: string): string {
-  return text.replace(/<\/?[a-zA-Z][a-zA-Z0-9-]*(?:\s[^>]*)?\/?>/g, '')
+  const end = text.lastIndexOf('>') + 1
+  if (end === 0) {
+    return text
+  }
+  // No tag can close in this suffix; keep it literal without retrying every opener.
+  return (
+    text.slice(0, end).replace(/<\/?[a-zA-Z][a-zA-Z0-9-]*(?:\s[^>]*)?\/?>/g, '') + text.slice(end)
+  )
 }
 
 export function parseMarkdownBlocks(content: string): MarkdownBlock[] {

mobile/src/components/pr-sidebar/markdown-tag-stripping-performance.test.ts

@@ -0,0 +1,47 @@
+import { runInNewContext } from 'node:vm'
+import { describe, expect, it } from 'vitest'
+import { parseInline, stripHtmlTags } from './markdown-blocks'
+
+const ORIGINAL_TAGS = /<\/?[a-zA-Z][a-zA-Z0-9-]*(?:\s[^>]*)?\/?>/g
+
+describe('review Markdown unclosed tags', () => {
+  it.each(['', '<b>before</b> '])('preserves an unclosed suffix after %s', (prefix) => {
+    const suffix = '<a '.repeat(20_000)
+    const text = prefix + suffix
+    const result = runInNewContext('parse(text)', { parse: parseInline, text }, { timeout: 250 })
+    expect(result).toEqual([{ kind: 'text', text: prefix.replace(ORIGINAL_TAGS, '') + suffix }])
+  })
+
+  it('preserves the original stripping grammar over generated markup', () => {
+    const parts = [
+      '<a ',
+      '<b>',
+      '</b>',
+      '<a href="x">',
+      '<custom-tag x>',
+      '<',
+      '>',
+      'a',
+      ' ',
+      '/',
+      '\n',
+      '"',
+      '<1>',
+      '<a/>',
+      '<a x<',
+      '<a'
+    ]
+    let seed = 17
+    const random = () => {
+      seed = (Math.imul(seed, 1664525) + 1013904223) >>> 0
+      return seed
+    }
+    for (let index = 0; index < 5000; index++) {
+      const text = Array.from(
+        { length: 1 + (random() % 40) },
+        () => parts[random() % parts.length]
+      ).join('')
+      expect(stripHtmlTags(text), text).toBe(text.replace(ORIGINAL_TAGS, ''))
+    }
+  })
+})