Build-Test-Release Biweekly 2021-03-15

Build-Test-Release Biweekly 2021-03-15

Aaand we’re back!

MFEs on Lilac

On today’s meetup, in addition to our usual crowd (@nedbat, @sambapete, @regis, @blarghmatey, @Maksim_Sokolskiy, @Neo, @pdpinch, @BbrSofiane) we had special guests: @braden, @kmccormick, @djoy, @sarina, and @xitij2000 joined to help figure out what to do about Lilac’s support for MFEs.

Long story short, we’ve tentatively agreed to support the following MFEs on Lilac, both via the Native Installation and Tutor:

This means that until release, work will be coordinated (but not necessarily undertaken) by this group so that the following applies to all three MFEs:

  1. It passes a minimum standard of documentation: at the very least, the READMEs should have descriptive text that includes use cases, screenshots, and particularly, documentation on that MFE’s environment variables.
  2. Its basic feature-set (as documented above) works reliably with the rest of the Lilac codebase.
  3. It’s deployed by default using the Native Installation, and deployable as a plugin using Tutor.
  4. It is internationalized (i18n), localizable (l10n), and passes minimum accessibility (a11y) standards.
  5. It is reasonably themable

That last item - in particular, what “reasonable” means - is still up for grabs. We should know more once @xitij2000 completes his initial investigation on MFE theming.

Incoming ADR: the release timeline

We’ve been discussing the possibility of cutting Lilac from master on April 9th for some time, and now we finally have an ADR that can be used for concrete discussion.

Once again, there haven’t been strong objections to pushing this date back, but @pdpinch’s investigation into Identifying in-progress work that won’t land before Lilac is cut is still on-going. If you have any thoughts on the matter, please bring them up here or on that issue ASAP: April 9 is fast approaching!

For reference

New and unassigned issues

  • [#24] New, assigned to @blarghmatey (thanks!). We need to backport Mongo 4.0 to Koa, by koa.3.
  • [#27] New, unassigned (though @sarina has volunteered to help with the documentation bits - thanks!). Support for the Account MFE in Lilac.
  • [#28] New, unassigned. Support for the Profile MFE in Lilac.
  • [#29] New, unassigned. Support for the Gradebook MFE in Lilac.
  • [#5] Old, unassigned. Juniper-Koa upgrade instructions.

Open ADRs

Let’s try and wrap these up before April 9th, even if Lilac is not branched off of master on that date.


Thanks for reading! You can use this thread to discuss the above topics, or anything else that doesn’t warrant a new thread.


Hi @arbrandes - wrinkle in plans. I discovered Deprecation/Removal: Legacy Frontend Implementations (DEPR-17, DEPR-42, DEPR-109) this morning - this is a breaking change to Lilac. The legacy checkout page has not been supported in some time and no longer works after the PCI updates. We should make this one of the default MFEs supported; if we have to kick one out to remain at 3 MFEs, I guess I’d go with the Profile MFE (@kmccormick thoughts?)

1 Like

Just to clarify - all of those legacy frontends that Sarina linked have been deprecated, but it’s the continued functionining of two specific legacy frontends that are in question:

  • basket/checkout: this stopped working after Koa and is replaced by frontend-app-payment.
  • order history: this is replaced by frontend-app-ecommerce; the legacy version of this one may or may not still work; we have someone checking that.

I would kick out Profile and then Account (in that order), depending on whether one or both of the above MFEs need to work in Lilac.

(aside: do folks in the community actually use the ecommerce site?)


Thanks for the heads-up, @sarina, and the clarification, @kmccormick!

Regardless if a lot of people use it, I figure we can’t just drop ecommerce without at least 6 months of deprecation, so we’ll have to do what it takes to make it work. So as far as I can see, we’ll need to support:

  • frontend-app-payment ([#30] created)
  • frontend-app-ecommerce, depending on whether the legacy version works or not ([#31] created).

We need assignees for [#30], and since you said you have someone checking it, I’d like to assign [#31] to you, @kmccormick, if that’s ok.

This is certainly an important question. It’s something to ask on the next Contributor’s Meetup, but let’s do a quick, informal poll beforehand:

Do you or one of your customers use Open edX eCommerce?
  • Yes
  • No

0 voters

Tagging for awareness: @Felipe, @sambapete, @pdpinch, @regis, @Maksim_Sokolskiy, @BbrSofiane. Feel free to tag anybody else, particularly if you know they use eCommerce. :slight_smile:

1 Like

I’ve posted details about some of my initial findings, and our overall goals and approach here:

1 Like