fix(frontend): strip closing hashes from web fetch titles (#6103)

* fix(frontend): strip closing hashes from web fetch titles

Signed-off-by: 97three <2212371308@qq.com>

* fix(frontend): scan web fetch heading suffixes in linear time

Signed-off-by: 97three <2212371308@qq.com>

---------

Signed-off-by: 97three <2212371308@qq.com>

frontend/AGENTS.md

@@ -87,8 +87,13 @@ More specific `AGENTS.md` files under `src/` contain the frontend sections split
 
 `core/utils/markdown.ts` reads web-fetch titles from the first nonblank line.
 Match zero to three literal spaces before `# ` without trimming indentation;
-mixed space/tab code blocks must fall back to the URL. Keep this local to title
-extraction rather than changing the shared streamdown fence parser.
+mixed space/tab code blocks must fall back to the URL. Strip optional closing
+hashes with a backwards scan so long fetched titles cannot cause quadratic regex
+backtracking. Closing hashes require a preceding space/tab (or an empty title)
+and only ASCII spaces/tabs afterward, apart from the terminal CR in CRLF input.
+Non-ASCII trailing whitespace must not turn literal hashes into closing syntax.
+Keep this local to title extraction rather than changing the shared streamdown
+fence parser.
 
 Custom Agent `display_name` is an optional Unicode UI label, edited in
 `AgentSettingsDialog`. Use it with a fallback to `name` for gallery/chat text;

frontend/src/core/utils/markdown.ts

@@ -9,5 +9,24 @@ export function extractTitleFromMarkdown(markdown: string) {
   if (!headingPrefix) {
     return undefined;
   }
-  return firstLine.slice(headingPrefix[0].length).trim() || undefined;
+  const title = firstLine.slice(headingPrefix[0].length);
+  // Scan backwards once: a whitespace-search regex can backtrack quadratically
+  // on long fetched titles. Only ASCII spaces/tabs may follow closing hashes.
+  let end = title.endsWith("\r") ? title.length - 1 : title.length;
+  while (end > 0 && (title[end - 1] === " " || title[end - 1] === "\t")) {
+    end--;
+  }
+  let hashStart = end;
+  while (hashStart > 0 && title[hashStart - 1] === "#") {
+    hashStart--;
+  }
+  if (
+    hashStart < end &&
+    (hashStart === 0 ||
+      title[hashStart - 1] === " " ||
+      title[hashStart - 1] === "\t")
+  ) {
+    return title.slice(0, hashStart).trim() || undefined;
+  }
+  return title.trim() || undefined;
 }

frontend/tests/unit/core/utils/markdown.test.ts

@@ -6,6 +6,48 @@ test("reads the title from a leading ATX heading", () => {
   expect(extractTitleFromMarkdown("# Real Title\n\nbody")).toBe("Real Title");
 });
 
+test.each([
+  ["# Release Notes ###", "Release Notes"],
+  ["# Release Notes ###   ", "Release Notes"],
+  ["# Release Notes\t###\t", "Release Notes"],
+  ["# Release Notes ###\r\nbody", "Release Notes"],
+  ["# ###", undefined],
+])("removes a closing ATX heading sequence from %j", (markdown, expected) => {
+  expect(extractTitleFromMarkdown(markdown)).toBe(expected);
+});
+
+test.each([
+  ["# C#", "C#"],
+  ["# Release ### notes", "Release ### notes"],
+  ["# Release ###x", "Release ###x"],
+  ["# Release \\###", "Release \\###"],
+])("preserves literal hashes in %j", (markdown, expected) => {
+  expect(extractTitleFromMarkdown(markdown)).toBe(expected);
+});
+
+test.each(["\u00a0", "\u2003"])(
+  "preserves literal hashes followed by non-ASCII whitespace %j",
+  (whitespace) => {
+    expect(extractTitleFromMarkdown(`# Release ###${whitespace}`)).toBe(
+      "Release ###",
+    );
+    expect(extractTitleFromMarkdown(`# #${whitespace}\r\nbody`)).toBe("#");
+  },
+);
+
+test("handles long whitespace and hash runs without changing literal content", () => {
+  const spaces = " ".repeat(100_000);
+  const hashes = "#".repeat(100_000);
+  expect(extractTitleFromMarkdown(`# Title${spaces}#x`)).toBe(
+    `Title${spaces}#x`,
+  );
+  expect(extractTitleFromMarkdown(`# Title${spaces}${hashes}\t\r\nbody`)).toBe(
+    "Title",
+  );
+  expect(extractTitleFromMarkdown(`# ${hashes}x`)).toBe(`${hashes}x`);
+  expect(extractTitleFromMarkdown(`# ${hashes}\t`)).toBeUndefined();
+});
+
 test("skips blank lines before the first heading", () => {
   expect(extractTitleFromMarkdown("\n# Real Title\n\nbody")).toBe("Real Title");
   expect(extractTitleFromMarkdown("  \n\n# Real Title")).toBe("Real Title");