Page Comparison

Maintain (Infrastructure)

• upgrade yarn from v1 to v4.1, or migrate to npm >= v10

• POC or analysis to determine estimate or technical approach

1. Encourage teams to implement -

   Jira Legacy
   server System Jira serverId 01505d01-b853-3c2e-90f1-ee9b165564fc key STRIPES-904

   1. Next step: (Action item John creates Jira issues) for

      1. PR stripes-core

         1. Noah has sample PR1393

      2. PR stripes-component

      3. PR ui-development

   2. Then create issues for other ui-repos after John and Zak validate

2. Spike - define technical approach

   1. Outline options (Action item: Zak) - Estimated time (2 days)

      1. Review options with John and send a note to stripes-update/stripes-architecture for awareness and additional participation

      2. (Get all feedback timeframe: 2 weeks)

         1. KG can create stories for Spitfire/Corsair to provide feedback.

         2. DevOps review > need TC support?

• Start in Ramsons > Release Sunflower

  • Get Devs to review in Ramsons and decide to use it Ramsons or wait for Sunflower

  • Get TC approval for Sunflower.

$40

• v4.1 seems to Just Work for repos, for platforms, for workspaces. YAY!

• it’s our cruftiest cruft. fix the worst first.

• blocks other dependency updates (favicons, maybe others)

• It’s easy and we could use the quick win.

• Can be rolled out incrementally across repos/teams

• Must be done in each repository but does not need to be coordinated.

$20

• yarn v4 is less promising than initially anticipated due to poor support on Windows.

• npm v10 is less promising than initially anticipated due to buggy handling of registry data in the latest version.

• crap. it feels like “npm v10 with work-arounds” may be the best option, but those work-arounds are real papercuts: low grade, but constantly annoying. 😞

Maintain

• upgrade react-router from v5 to v6 - I suspect the impact will be very small or very large, as in, some repos won't be affected at all but others will have to do some heavy lifting. This work will have to be done all at once.

• react-router

  • POC or analysis to determine estimate

  • Impacts all ui-repositories. All must repos must upgrade. BREAKING CHANGE

  • sidecar environment?

  • no-need for a rancher environment / use a ui-build

  • Communicate!!!!

X

$0

• Next time, I’m putting most of my cash in this basket since and sprinkling the rest among the other “update old things”.

• Must be done in each repository AND coordinated across them, so even if the work is easy the effort is large.

$5

Innovation

• transpile all modules on publish (update stripes-webpack)

• Blocks Marva from using stripes-components

• Industry best practice

$20

• gotta follow ecosystem SOP

• enables MARVA (or any third party) to consume us

• must be done in each repo but can be done separately

$20

Improved dependency handling support

1. commit yarn.lock in all modules

2. enable dependabot in all modules

To consider this feature > Verify with Noah’s work and/or with DevOps

$30

• This won’t erase any debt but should prevent/slow new debt. That’d be nice.

• Must be done separately in each repository but does not need to be coordinated.

$10

Improved developer adoption

Accessibility

• Improve with overlay components' behavior on narrow screens

• Review and address Deque accessibility results related to time-picker/date-picker/ (will address VPAT)

Next steps: John to revise goals/objectives cell.

Accessibility

• Post-css revisions / will review color contrast Deque accessibility issues to determine resolution

Break tower of knowledge/support

Bolster UX consistency

Improved test coverage

DayJS - Switch ui date-time library from using the EOL’d moment library.

$8

Maintain

Popper -> convert to Floating UI

Dependency update

Possible delegate candidate

X

$3

Maintain

Downshift -> Update to most recent version (lots of breaking versions since)

Dependency update

Possible delegate candidate

X

$3

Maintain

react-highlight-words

Dependency update

Possible delegate candidate

Migration

redux-form and react-form and their related libraries have been abandoned. We already contend with some peer-dep warnings for related libraries that haven't provided a React v18 release (which means they've been out of date since at least March, 2022). The main contenders for replacement are formik and react-hook-form. We don't need a "thou shalt switch" decree, but we should do a spike to see how stripes-components interacts, pick our favorite, and write a guide. I would vote for RHF first. This work can be done incrementally.

Select a new form library selection

$10

• don’t change the old stuff, but provide a clear path for new stuff

• stops us from accumulating debt

• can be done in isolation in stripes-components

$10

Maintain

react-query: we're on v3; current is v5. Dunno the impact. This work may be doable incrementally but all at once is preferable.

• POC or analysis to determine estimate or technical approach

$10