Date, time, location

 Discussion topic(s)

🎥Recording

 Participants

Adolfo Brandes, Brian Smith, Diana Villalvazo Salas

🤖 Summary

Primary Topic: Dependency Freeze ADR proposal and release-cycle coordination

1. Dependency Freeze ADR Overview

Adolfo presents an ADR proposing a dependency freeze for core frontend libraries used across MFE development. The freeze would occur eight weeks before the master cutoff for each release, giving teams time to bring stragglers up to required versions.

Libraries included:

  • Node.js

  • React

  • React Router

  • React Query

  • Paragon

  • Frontend Base

  • Build Platform and Plugin Framework (for MFE authors)

Goals of the Freeze:

  • Provide predictability for teams developing or upgrading MFs.

  • Allow sufficient time to update dependencies without expecting full development cycles within this window.

  • Ensure consistency in MF environments across releases.

For new MFE development, authors should ask the Frontend Working Group early which versions and features will be targeted for the next release.

2. Version Targeting & Feature Levels

The group discusses whether the ADR should specify feature levels (major + minimum minor versions) rather than only major versions — especially for internal libraries like Paragon and Frontend Base, where nonbreaking features ship during a cycle.

Brian notes this prevents MFE authors or plugin developers from expecting features unavailable in some installations.

Adolfo will update the ADR wording accordingly.

3. Library-Specific Discussion

Node.js

  • Mostly upgraded for security reasons, not features.

  • Node has a predictable release cycle.

  • Past challenges (e.g., Node 16→20) involved packages requiring newer Node versions.

React

  • Updates may bring desirable new features (e.g., React 19, React 20).

  • React Compiler is now stable, potentially reducing the need for manual memoization.

  • Adoption requires ensuring code follows hooks rules (lint cleanup needed).

  • Group is positive about a React 19 upgrade soon.

React Router

  • Currently on React Router v6.

  • Future migration to TanStack Router v1 is appealing (better URL param handling).

  • Frontend Base currently relies on v6 syntax, so migration would be significant.

  • Group agrees migrating earlier may reduce long-term effort.

React Query

  • Breaking changes across versions require coordination.

  • Current version migration has required some work; future ones will too.

Paragon

  • New features or breaking changes require alignment between MFE authors and Paragon maintainers (Brian).

  • Pre-release versions will be used and pinned until stable releases occur.

  • Minimum minor version expectations will be set per release.

Frontend Base

  • Needs coordinated planning similar to Paragon.

  • Frontend Base's deduplication model requires all MFEs to share the same exact patch version of core libraries to avoid duplication issues at build time.

4. Possible Additional Dependencies

Webpack might be added to the dependency list for completeness, though it's been stable at v5 for a long time.

5. TanStack Ecosystem

  • Team already uses TanStack Table (via Paragon Data Table).

  • TanStack Router + TanStack Query could unify patterns.

  • Some concern that TanStack is becoming dominant, but overall confidence in their work.

6. Coordination Practices Going Forward

  • Dependency and feature-level decisions should be made collectively in the Frontend Working Group.

  • The group should become the central touchpoint for discussions with MFE authors about release targets.

 Action items

  • @Adolfo Brandes Update ADR wording to incorporate feature levels.

  • @Adolfo Brandes Review Diana’s PR for the Frontend Base navigation bar.

  • @Adolfo Brandes Plan and execute Paragon 23 upgrade in Frontend Base

 Decisions