No formal minutes were recorded, but here is the IRC log (all participants were happy for it to be posted). The various TODOs were integrated into the rest of the wiki.
Log timezone is KST.
20:41 <bremner> spwhitton: did you get my mail yesterday? about work session in 20 minutes?
20:53 <remiv> Hi, I'm here too.
20:56 <bremner> cool
21:01 <spwhitton> bremner: yup good evening
21:01 <spwhitton> Hello remiv
21:01 <bremner> hurray!
21:02 <bremner> what to start with?
21:03 <bremner> remiv: btw, I'm assuming you're not in Cape Town
21:03 <remiv> No, I'm waiting for my employer to tell me I'm not needed anymore here in France
21:03 <spwhitton> Well, we have an agenda.
21:04 <bremner> right. OK. issues about dh-elpa? I started looking at the autodebp8 branch, and we can discuss that if nothing else arises.
21:05 <spwhitton> Well, autodep8 support was the last thing on my wishlist.
21:05 <bremner> no bugs in the BTS. no recent screwups with maintainer scripts.
21:06 <bremner> remiv: what's your experience? sean and I are probably a bit close to the tool?
21:06 <remiv> In my experience it work mostly good.
21:07 <spwhitton> "mostly"?
21:07 <remiv> I do not always folow how it develop, and so I'm not sure that I use everything it could do
21:07 <remiv> Well recently, I had no problem
21:09 <spwhitton> Dmitry Bogotov is successfully churning out packages with dh_elpa over on email@example.com. Only problems he is having is getting test suites to run.
21:10 <remiv> I didn't try that
21:10 <bremner> spwhitton: is this bugs / missing features in dh_elpa_test?
21:11 <spwhitton> bremner: I don't think so. The issue is that evil's test suite fails under emacs -batch; it needs emacs -nw. But of course we can't do that as it will fail on a bulidd.
21:11 <spwhitton> We're still investigating but I think the answer will be to comment out those tests.
21:11 <bremner> spwhitton: for notmuch-emacs, we run emacs in dtach
21:12 <spwhitton> ooo.
21:12 <bremner> but yeah, that doesn't sound like a tooling issue.
21:12 <spwhitton> Well now that you say it, a DH_ELPA_TEST flag to automatically do the dtach wrapping would be nice
21:12 <bremner> heh. OK, if you think it would be generally useful.
21:13 <spwhitton> Well we just found at least two packages that could use it
21:13 <bremner> well, notmuch doesn't use ERT or buttercup, so...
21:14 <bremner> OK, well, if serious, maybe you or dimitri could file a wishlist bug on dh_elpa to start?
21:14 <spwhitton> Okay. I'll make sure this actually works for evil, and if it does, I'll file the bug.
21:15 <bremner> feel free to steal what you can from notmuch, obviously
21:15 <spwhitton> I was planning to just tell Dmitry to go look at it
21:15 <bremner> sounds good.
21:16 <bremner> he can find me here, or more people on #notmuch on freenode who understand that test suite
21:17 <bremner> since we have feature completeness (?) maybe we can bump to 0.1.0 for the next upload. Or even 0.1, not sure if that causes problems/confusion
21:18 <spwhitton> Yes, let's do one of those two.
21:18 <bremner> ok, I have a few codereview style questions about the autopkg8 stuff, but we can come back.
21:18 <bremner> so, dh-make-elpa?
21:19 <spwhitton> Right. I've used it to make about 5 packages now.
21:20 <spwhitton> I haven't yet found anything that I'd like it to be able to do that it doesn't do, and that I'm capable of implementing.
21:21 <spwhitton> I think the best thing is if other people make packages with it, report bugs saying what it got wrong and then the heuristics can be made steadily more sophisticated.
21:21 <bremner> spwhitton: do you want to do a lightning demo remotely about dh-make-elpa? I'll have to check the schedule, but I could walk through what you tell me to do.
21:21 <bremner> is dimitry using it, do you know?
21:21 <spwhitton> Unfortunately I can't persuade him to. He just keeps copy/pasting the paredit-el packaging.
21:22 <spwhitton> bremner: if a lightning demo is what I'm guessing, sounds cool, yeah. It could be effective recruitment.
21:22 <spwhitton> People who can upload to NEW might like the idea of dropping MELPA with just a few hours of reading over generated control files.
21:23 <bremner> spwhitton: 5 minutes, anything goes, video streaming
21:23 <remiv> I certainly would love to watch it.
21:23 <spwhitton> I wouldn't actually be involved, I take it?
21:24 <bremner> spwhitton: the idea would be you'd tell me what to type, in IRC, and I'd run the actual demo, hilarity ensuing
21:24 <spwhitton> Oh I see, haha, okay let's do it
21:24 <spwhitton> We have to pick something to debianise I guess.
21:24 <bremner> yes.
21:25 <bremner> Friday July 8, 1500 SAST
21:25 <spwhitton> That's doable
21:27 <spwhitton> I guess we would do: dh-make-elpa --pkg-emacsen ; gbp buildpackage -us -uc -b ; sudo debi ; emacs ; M-x foo
21:28 <bremner> right. Ideally you could cheat and make sure it works beforehand
21:28 <spwhitton> I wasn't going to mention it :P
21:30 <spwhitton> Okay then. That demo might generate some TODOs. We could move onto our next agenda item.
21:31 <bremner> I _think_ the guy who packages s-el and dash-el will be at debconf. I can try to talk to him; I wonder if having converted package(s) would help?
21:32 <spwhitton> Sure. They will be trivial.
21:32 <bremner> is that the happy sound of a volunteer?
21:33 <spwhitton> yes okay
21:33 <bremner> thanks
21:33 <bremner> are there some others we should think about? maybe file wishlist bugs?
21:33 <spwhitton> yasnippet and org-mode
21:34 <spwhitton> However, they are leaf packages, so it's definitely wishlist.
21:34 <remiv> In the early time I was testing dh-elpa, I had made a conversion of dash-el, It must be somewhere on a hard-disk
21:34 <spwhitton> remiv: oh cool. I'll do s-el, then.
21:34 <bremner> remiv: if you could hunt it up, that would be helpful.
21:35 <spwhitton> I seem to recall that upstream org-mode are hostile to package.el. I don't know if it affects dh_elpa; it might just be the problem of bytecompiling in an unclean Emacs. But I'm not sure.
21:36 <bremner> spwhitton: ok, but it's not really a problem for "us" generally speaking
21:37 <spwhitton> bremner: it might be. I can't remember what their reason for not using it is.
21:37 <spwhitton> When I said "hostile" i meant "because it breaks org-mode"
21:37 <spwhitton> Anyway I'll file the wishlist bug.
21:37 <bremner> ok
21:37 <spwhitton> The other thing is dpkg-dev-el
21:38 <spwhitton> (well, one of the other things)
21:38 <spwhitton> Like emacs-goodies it's a huge, out-of-date thing.
21:38 <bremner> but does it have active upstreams?
21:38 <spwhitton> I had assumed that the stuff in there was written by Debian people and just uploaded.
21:39 <spwhitton> Ah, it comes from emacs-goodies.
21:39 <bremner> probably. But not sure how dh-elpa helps then; sounds like it needs actual "upstream" development.
21:39 <bremner> ok. emacs-goodies is a topic for us.
21:40 <spwhitton> Yeah. If they were under our umbrella we could easily commit fixes to them as we saw them in our own Debian work.
21:40 <spwhitton> I'll add it to the agenda on the wiki. Is it okay with you two if we leave it until after we've discussed the other things on there?
21:41 <bremner> yes
21:41 <bremner> (for me)
21:41 <remiv> okay
21:42 <bremner> so, backporting?
21:43 <spwhitton> yes. As we said on the list, the thing to discuss is what we think we should backport.
21:43 <bremner> I think remiv should backport magit, once we make that possible
21:43 <remiv> I've tested to just run package from unstable on my stable box, with no problem, so once it will be possible, it will be easy
21:43 <bremner> typically backports are on a "by request", basis.
21:44 <bremner> remiv: right, that's what I do also.
21:44 <spwhitton> bremner: indeed but we could get the ball rolling a bit.
21:44 <bremner> spwhitton: are there infrastructure packages that should be backported?
21:44 <bremner> that seems like a place to start.
21:45 <spwhitton> yes: s-el, f-el, dash-el, emacs-async. Maybe: let-alist-el, seq-el, shut-up
21:46 <spwhitton> helm might be nice.
21:46 <spwhitton> (not infrastructure)
21:46 <bremner> isn't helm a monster?
21:46 <bremner> in terms of deps
21:46 <remiv> A very usefull monster.
21:46 <spwhitton> oh yeah, it is.
21:47 <bremner> ok, I'm not using it, but of course I'm fine with the idea.
21:47 <bremner> spwhitton: you should ask to be in the backports ACL https://backports.debian.org/Contribute/#index3h3
21:48 <bremner> yay more hoops.
21:48 <spwhitton> bremner: can I do that in advance of actually have any sponsored backports?
21:48 <bremner> spwhitton: yes, there some request tracker thingy that takes a little time
21:48 <spwhitton> okay will do it
21:49 <bremner> should we have a todo list of some kind in the team wiki?
21:49 <bremner> (for backports specifically)
21:50 <spwhitton> I guess so yeah
21:50 <remiv> Agreed
21:51 <bremner> OK, I'll try to start a page
21:52 <spwhitton> you'll probably succeed
21:52 <bremner> never know, the whole ikiwiki setup might have bitrotted
21:53 <spwhitton> okay then. Next on the agenda is increasing team participation. We already discussed that a bit.
21:53 <bremner> yeah. well, demo counts as positive action I guess.
21:53 <bremner> one thing the perl team does for packages that are undermaintained is offer to adopt them
21:54 <spwhitton> Yeah, I said that in the dh-elpa wishlist bugs I filed so far.
21:54 <remiv> Are there known undermaintained package?
21:55 <bremner> well, emacs-goodies-el, to pick a nasty example
21:55 <spwhitton> emacs-goodies-el is a bit of a mountain. The very task of making the task approachable is the task.
21:55 <bremner> generally I don't know, it would be interesting to understand the state of -el packages
21:55 <bremner> maybe some udd search of bugs?
21:56 <spwhitton> Maybe we could use codesearch to generate a list of all the packages.
21:57 <spwhitton> It would be good to have a list of all non-dh-elpa source packages on the wiki. Not sure what the best tool to get that with would be.
21:57 <bremner> yeah, it would have to be autogenerated
21:59 <spwhitton> Possibly just apt-file.
22:00 <spwhitton> Okay. I'll find somewhere on the wiki to put that as a TODO.
22:02 <bremner> ok. Next topic? I might have to leave for a bit, but I plan to be available same time tommorow
22:02 <spwhitton> Sure. Next thing is Spacemacs
22:03 <spwhitton> The particular thing I want to discuss here is how many of the spacemacs dependencies we think we need to package before doing spacemacs itself. Many of the dependencies are optional.
22:03 <bremner> I guess the reason people like spacemacs is the "batteries included" aspect
22:03 <spwhitton> RIght. We certainly need to package the ones that are turned on by default (which is quite a lot of them).
22:04 <bremner> and those will be regular elpa-foo packages?
22:05 <spwhitton> As far as I know, yes. And useful to non-spacemacs users.
22:05 <bremner> ok, that's a lot more attractive than the alternative ;).
22:05 <spwhitton> Yeah. I don't want to package spacemacs myself, but I do see it as a useful team milestone.
22:06 <spwhitton> (mainly because I don't want to hack it to stop it from downloading from MELPA when an apt package is available..)
22:07 <bremner> I see there's already a page to fill in with packages.
22:07 <spwhitton> yes when someone figures out the regexp
22:07 <spwhitton> The question is whether we're going to split that list into "prerequisitse" and "can be packaged post-spacemacs", or not.
22:07 <bremner> that would be ideal, since there's a limited amount of human effort available.
22:08 <spwhitton> So you think it would be okay to package spacemacs even if it potentially wants to pull stuff from MELPA?
22:08 <bremner> I'd not be very enthusiastic about that.
22:09 <spwhitton> Oh, when you said "that would be idael" you were perhaps referring to the regexp
22:09 <bremner> Yeah, I mean, even if we don't want to package spacemacs, it's prereqs are probably more important
22:10 <spwhitton> Ah right I see what you mean. Okay.
22:10 <bremner> OK, I should probably run, same time tommorow? I'll try to finish looking at the dh_elpa changes before
22:10 <spwhitton> yup see you tomorrow
22:10 <spwhitton> Enjoy debcamp!
22:11 <spwhitton> Wish I was there. Maybe next year.
22:11 <bremner> thanks to both of you for showing up on IRC
22:11 <bremner> bye for now
22:11 <spwhitton> yes thanks to both of you from me too. Several useful ideas.
22:11 <remiv> I will try to be here tomorow. Can't know for now.
22:11 <remiv> bye