False positive: js/incomplete-hostname-regexp treats LinkifyIt.match(text) as a regex call

檢查 js/incomplete-hostname-regexp 查詢及其程式庫模型，然後比較 PR 19854 中的 Sinon 先例。針對 test/link-recognition-qualification.test.js 重新執行連結的 CodeQL 警示；當不再回報 LinkifyIt.match(text) 字面值，同時一般的 String.match 正規表示式發現仍保持啟用時，即完成。

Description of the false positive

js/incomplete-hostname-regexp treats the argument to LinkifyIt.match(text) as a regular expression. The receiver is a LinkifyIt instance from linkify-it@6.1.0; its argument is document text to scan for links, not a regex pattern. Literal dots in these URLs are therefore correct.

Observed with CodeQL 2.26.4, JavaScript/TypeScript analysis, build-mode: none, and github/codeql-action@cdf488f595d80d6e07e03d4674febd5ab45fa938 (v4.37.9), using the default query suite without custom queries or exclusions.

Diagnostic on the text literal:

This string, which is used as a regular expression here, has an unescaped '.' before 'youtube.com/watc', so it might match more hosts than expected.

The linked use is scanner.match(text).

Code samples or links to source code

Small excerpt retaining the constructor, configuration, literal and call from the reported test:

import { LinkifyIt } from "linkify-it";

const scanner = new LinkifyIt({ fuzzyLink: false, fuzzyEmail: false })
  .add("ftp:", null)
  .add("mailto:", null)
  .add("//", null);
const text =
  "😀 *literal* (https://www.youtube.com/watch?v=tax4e4hBBZc), then https://store.steampowered.com/app/457140/.";
const matches = scanner.match(text);
console.log(matches.map((m) => m.raw));

With linkify-it@6.1.0, the API returns matches for the two literal URLs. The package's build/index.d.ts declares match(text: string): Match[] | null; build/index.mjs implements the method by scanning that document text with the library's link recognizers.

Exact reported source at the PR merge revision.

Source at the immutable PR head.

Validation: node --test test/link-recognition-qualification.test.js passes 1/1, including exact URL and offset assertions. The hosted CodeQL analysis reports the finding in the linked full test. The reduced excerpt above has not been separately analyzed with CodeQL; no claim is made that it is the smallest scanner reproducer.

Expected: recognize that this imported library's match method consumes document text, so these literals should not be classified as hostname regular expressions. Ordinary String.match regex findings should remain enabled.

Related precedent: https://github.com/github/codeql/pull/19854 explicitly models Sinon's match calls as non-RegExp. I found no existing linkify-it report in the upstream search.

URL to the alert on GitHub code scanning (optional)

https://github.com/MaksymShostak/steam-community-bbcode/security/code-scanning/2

Failing PR check.