fix(serializers): preserve angle-bracket URLs as text

[rfgamaral]

Overview

Treats angle-bracket HTTP(S) URLs as plain text when loading stored Markdown into rich-text editors. This prevents unsupported Markdown syntax from becoming a link whether or not the link extension is enabled.

Reference

Related to Doist/Issues#20812

PR Checklist

• Pull request title follows the Conventional Commits Specification
• Added/updated unit test cases and/or end-to-end test cases

Test plan

1. Open the deployed Storybook preview
2. Open the Default story under Typist Editor → Rich-text
3. Set the content control to <https://duckduckgo.com>
   • Observe that the editor displays <https://duckduckgo.com> as plain text rather than a link
   • Observe that Markdown Output contains <https://duckduckgo.com> unchanged

Demo

Before	After

[netlify]

✅ Deploy Preview for doist-typist ready!

Name	Link
🔨 Latest commit	116677e
🔍 Latest deploy log	https://app.netlify.com/projects/doist-typist/deploys/6a7b6332fcf2a300087aa6cd
😎 Deploy Preview	https://deploy-preview-1466--doist-typist.netlify.app
📱 Preview on mobile	Toggle QR Code... Use your smartphone camera to open QR code link.
🤖 Make changes	Run an agent on this branch

---

To edit notification comments on pull requests, go to your Netlify project configuration.

[doistbot]

The PR treats angle-bracket HTTP(S) URLs as plain text when loading stored Markdown into rich-text editors, preventing them from being parsed as links.

Few things worth tightening:

• Consider reusing the existing remarkDisableConstructs plugin instead of the new AST-walking plugin. Adding 'autolink' to its disabledConstructs list would disable angle-bracket URL parsing at the micromark level—no AST traversal, source slicing, or regex needed. Since the behavior applies unconditionally, the push should go outside the schema conditionals in that plugin.

I also left one optional follow-up note in the details below.

Optional follow-up note (1)

• src/serializers/html/plugins/remark-ignore-angle-bracket-autolinks.ts:12: visit(tree as Parent, ...) invokes the callback for every node in the AST, then early-returns for non-link nodes after accessing position properties. Every other visitor in this directory uses the typed form (e.g. visit(tree, 'element', ...)) to skip non-matching nodes. Use visit(tree, 'link', (node, index, parent) => { ... }) to avoid the per-node callback overhead on the serialization hot path, which runs on every document load and scales with document size.

_{Share Feedback • Review Logs}

[doist-release-bot]

🎉 This PR is included in version 17.0.1 🎉

The release is available on:

• npm package (@latest dist-tag)
• GitHub release

Your semantic-release bot 📦🚀