Maintenance Working Group Meeting - 2024/12/12 08:58 EST - Transcript

Attendees

Adolfo Brandes, Feanil Patel, Feanil Patel's Presentation, Michelle Philbrick, Robert Raposa, Sarina Canelake

Transcript

Feanil Patel: Hello. Sorry I'm late. It was nominating containers. I got distracted. I can see getting busy.

Michelle Philbrick: As you can see, it's very active here.

Feanil Patel: I think just loading up the sheet now  Share my screen.

Feanil Patel: is kind of the approach for the next couple of bits of maintenance. let's see if there's anything and then I have an update on the maintainers at large. Check the next thing out. Jeremy and Kyle aren't here.

Feanil Patel: the cogill conversations kind of got back burnered behind some other stuff, but I'll pick them up in the new year, I think. And then the cron ticketing I haven't done yet. So, let's see. Yeah, the thing I do want to talk about is the set of next maintenance items and kind of the approach we take there.  So one of the things that I am thinking is that we would take it up that work pretty soon even though in terms of Django it's actually not possible to upgrade to Django 52. what is possible is that you can start testing on Django 50 and 51 right now.

Feanil Patel: and that actually is where most of the backwards incompatible changes are happening. but I'd love to hear people's thoughts on that approach. So, essentially, Django 5.2 is not going to be released till April. but most of the issues that we're going to run into it when we try to upgrade to it are going to be issues in Django50.  which is a nonLTS. So does it make sense to take it up the work and just start on that now and encourage maintainers to sort of schedule in adding testing and dealing with deprecations that come out in 50 now rather than waiting till 52 is out and asking them to upgrade to all the way to 52 all at once.

Robert Raposa: I mean, sounds reasonable just so you don't get silence. and I also am unclear about RB Bomb and what they do and what their schedule is and how they'll help in terms of even just so we can simplify the process for maintainers and they're not like everyone's not rediscovering the same fixes and all of

00:05:00

Feanil Patel: Right. Right.

Feanil Patel: Yeah. I think some sort of centralized documentation would be helpful for common issues.

Robert Raposa: And then I know we have the like…

Robert Raposa: how we do upgrade docs.

Robert Raposa: I don't know if those are all open edex docks at this point or if there's two ducks. So I don't know.

Feanil Patel: Yeah. Yeah.

Feanil Patel: Those docs are very predicated on Arbomb sort of leading essentially all of the maintenance which I don't think is how any of the maintenance has gone I think in the last year. so I think they're kind of out of date, although there are pieces of them that I think are useful.

Robert Raposa: Or do we not? Is the documentation just unnecessary? Yeah.

Feanil Patel: Yeah, I'm not sure. it's probably worth reviewing it and having an opinion on it, but at the moment the last time I looked at them I think they were so wrong for the amount the way we need to coordinate that it was hard to reconcile them. and now I think that's still true. I think that what we need is a place for people to share information about sort of things they learn but we don't need the centralized creation of all the PRs and the centralized sort of delegation is different from how it was set up in that process.

Robert Raposa: need the new process documented even…

Robert Raposa: if it's not based on the old doc and it's simply based on reality or do we not need documentation for the new process?

Feanil Patel: I think we probably do need documentation.

Feanil Patel: If for no other reason than that somebody other than me could run it in the future.

Robert Raposa: Yeah, sounds good.

Feanil Patel: I think For the previous upgrades, I was a little bit reluctant to do it because I didn't know what it was that I was documenting. But I think now that we have I have a better idea of sort of what that process would look like. Yeah. Yeah.

Feanil Patel: And with that, I'm hoping that we have sort of giving maintainers enough time, we can sort of start finding issues much sooner and addressing them much sooner across the board. I think from a lot of our repos there won't be major issues and for platform I'm sure getting that out up front as much as possible will be useful. and to that end,…

Robert Raposa: One quick question is…

Feanil Patel: yeah.

Robert Raposa: if someone adds tests to 52 452 and we find that there are no issues. I guess one question is it worth leaving them enabled to avoid regressions that are probably not going to happen in said repo or would we disable until some future date and just be like okay this has been proven to be okay anyway something to consider or something.

Feanil Patel: Yeah. Yeah. Yeah. Yeah. and this is where I struggled was this if in a library the request would be to essentially enable 51 testing today right which would cover any 50 related upgrades that need to happen and 51 related upgrades. the end state if you did the work before April would be that you had a repo that had 42 and 51 being tested and then you would have to go back after April to enable 52 and disable 51. and that feels like a little bit of churn that we may want to avoid.  So I wonder if it is useful to just wait till April to start the Django testing.

00:10:00

Feanil Patel: It reduces our time frame, but it also reduces the amount of work that we need to do and have people go back to 200 repos to reenable tests and redisable tests at various versions.

Robert Raposa: And I guess this goes along with the other conversation of maybe there's some way to automate all the things whatever they happen to make that easier or…

Feanil Patel: Yes. Yeah.

Robert Raposa: least put together a wish list of automation for the future that be like here's how it becomes basically a breeze for a maintainer. just push a button, get everything you need.

Feanil Patel: Yeah. Yeah. Yeah. Right.

Robert Raposa: And yeah, because then you could add 52, make sure it passes. If it doesn't go down to 51. And if it doesn't pass down, …

Feanil Patel: Right. Yeah.

Robert Raposa: there's all kind

Feanil Patel: I think there's definitely all kinds of automation we could build here. I just don't think that anybody has capacity to build them.  So making the wish list feels like I don't know not useful yet but also I understand having a backlog is how you get work started so I have mixed feelings about it. but yeah, I think I'm convincing myself that it's probably better to hold off, but I'm curious to hear other people's opinions on it.

Robert Raposa: Hold off on 52 specifically or…

Robert Raposa: hold off on anything.

Feanil Patel: …

Feanil Patel: hold off on anything because it would be a shame…

Robert Raposa: Yeah. I guess one question is…

Feanil Patel: if people went in and did a bunch of work and then had to sort of go back and do more work later. Although the 51 to 52 jump would be non-breaking. So maybe it would be really easy, but it's still sort of churn. And for a lot of the repos where we don't currently have a maintainer or a small team maintains a lot of repos, it is sort of unnecessary turn.

Robert Raposa: if we had docs and it was clear what you need to do then couldn't we allow maintainers to do it whenever it works for them up to a point at some point…

Feanil Patel: Yeah, that's fair.

Robert Raposa: if you require

Robert Raposa: But it's like here, if you want to get a jump on this, then you have capacity now and not some future time. here's what you're going to want to do on your reput.

Feanil Patel: Right. Yeah,…

Feanil Patel: that's a fair point. We could just say note that before April whatever you will only be able to upgrade all through to 5.1. This is still valuable if you have capacity now, but after April one you should aim for 52 and you'll have to come back and update to 52 anyway.

Feanil Patel: and then okay I think I've convinced myself back in the other direction thank you yeah I think you're right I think we just have to make it very clear sort of what the tradeoffs are and then let the maintainers decide based in their personal capacity

Feanil Patel: requirements and leave it to them and then after April I'll startounding everybody who hasn't done anything yet to start work and then the maintainers at large the update I wanted to give there is that I think I have now sort of based on the conversations previously I think we've sort of arrived at a single  team rather than sort of domain specific teams and I'll create the team and sort of create all the tooling around it and then seek nominations for people to be added to that hopefully in January. I nominated some people today and I realized I had to basically make the nomination period a month because the week of Christmas and the week of New Year's are in between there.

00:15:00

Feanil Patel: So, two weeks that most people are likely to be working includes next week and then the second week of January. so I think if you're nominating people in the next couple of days or in the next week, keep that in mind and give people extra time to review things because many people are sort of dropping off as early as tomorrow.  And then Robert, you had this idea about an IDA for forums v2 and sumac.

Feanil Patel: Is that a thing you want that you're going to lead that you want Dave Ormsby to lead as the technical point person on the project? I think this is not maintenance.

Robert Raposa: Yeah, I could, and maybe it's I guess my first real question is this something that's not really a topic for this group and only for BTR? where does this or does it belong anywhere and everywhere that people want to talk about?

Feanil Patel: So, I would say this is actually …

Robert Raposa: Yeah. Yeah.

Feanil Patel: but also I'm happy to be a clearing house for when people don't know where things go. yeah.

Robert Raposa: But I'm also happy to just bring this up with Dave directly. And so it was really just

Adolfo Brandes: It's also not BTR.

Feanil Patel: Yeah. I think it's Dave and…

Adolfo Brandes: I wouldn't suggest this be a BTR thing. this also fell on BTR's lap pretty much. so it does make me think that we don't have a recurring meeting for new developments from an engineering perspective.

Feanil Patel: Rajis are sort of the leads here and the people the worth talk like that that is the team that spun up around this that is worth talking to.

Robert Raposa: Yeah. Yeah.

Adolfo Brandes: We have product working group.

Feanil Patel: Right. Amen. Yes.

Adolfo Brandes: but just something that Yeah.

Robert Raposa: And I don't know that another recurring meeting is required necessarily, but I do think the output of whatever was discussed if there is a mini RCA could be shared in multiple places whether it's discussed and across channels just hey here's what we learned.

Feanil Patel: Yeah. Yeah. I think this have a retro with Dave and…

Feanil Patel: Reis and share the learnings from that everywhere so that we can improve whatever communication processes broke down at product and…

Robert Raposa: Yeah. Yeah.

Feanil Patel: at BTR and probably in terms of cool.

Robert Raposa: That sounds great. Okay. We don't have to discuss further.

Feanil Patel: right, And then I think we're done unless somebody has other topics.  We're kind of in this quiet period right now because I think not just because of the holidays, but I think ve finished a lot of the major we've done a lot of maintenance this year and we've landed a lot of it and the platform is actually in a pretty good shape from third-party changes coming down the pike that we need to respond to. which is really awesome. but it also means that these meetings are very short and quiet.

Feanil Patel: So, I appreciate you guys sort of bearing with me as we get through this part. I think the major things that we're going to sort of want to focus on is finding maintainers for things in the early part of next year. and so maybe we can sort of refocus in January around sort of what the list is, who we can talk to, and how to get people to start sort of thinking about those projects and repos.  I'll have some new data to share along those lines hopefully by then as well.

Feanil Patel: currently working on sort of a mapping of the people who are listed as having organizational commitments to the open edex project and people who are providers partners of the open edex project they get listed and get referrals from us and those are people that have sort of promised to provide  support and maintenance and I think we need to sort of lean on them further. So I think that is a project that I would love people's help on in the new year that's hopefully a bit smaller than running the maintenance for Django and a little bit more easy to split up. but yeah, I think honestly it's been a really good year and we're in really great shape on the technical side for maintenance.

00:20:00

Feanil Patel: So, it's just a matter of getting the people side to be a little bit smoother so that there are maintainers for more of our systems. And I will probably cancel all of the meetings until the week of January like 5th or 6th unless anybody has objections. Okay. All right, that's all I got. Thanks for listening.

Michelle Philbrick: Bye everyone.

Feanil Patel: Bye, everybody.

Meeting ended after 00:21:44 👋

This editable transcript was computer generated and might contain errors. People can also change the text after it was created.