fix: start a GFM www autolink only at a word boundary - #4166

Marked version: 18.1.1 (master at 4b7b5b2)

Markdown flavor: GitHub Flavored Markdown

Description

Related to #4118, which handles the <www. case. This applies the general rule, so it covers that case too.

Expectation

The extended www autolink "can only come at the beginning of a line, after whitespace, or any of the delimiting characters *, _, ~, and (". So none of these are links:

xwww.example.com

a.www.example.com

<www.example.com>

while visit www.example.com, *www.example.com*, (www.example.com) and _www.example.com_ are. micromark (with the GFM extension) and GitHub render it this way.

Result

<p>x<a href="http://www.example.com">www.example.com</a></p>

inline.gfm.text stops before every www., so the url tokenizer gets to start one anywhere in a word.

Change

The text rule only stops before a www. that follows the start, whitespace, *, _, ~ or ( ((?<![^\s*_~(])www\.). Like the existing precode- rule, this uses lookbehind only when supportsLookbehind, and keeps the current behavior otherwise.

• npm run test:specs / test:unit / lint pass; new spec test/specs/new/autolink_www_word_start.
• npm run bench (7 interleaved runs, median): master 2707 ms, with this change 2698 ms.
• test:redos reports inline.gfm.text safe, and long inputs stay linear.

Found while comparing marked with commonmark.js and micromark on fuzzed input; written with help from an AI assistant (Claude) and checked against marked's test suite as described above.

Contributor

• Test(s) exist to ensure functionality and minimize regression
• no tests required for this PR.
• If submitting new feature, it has been documented in the appropriate places.

Committer

In most cases, this should be a different person than the contributor.

• CI is green (no forced merge required).
• Squash and Merge PR following conventional commit guidelines.