14:59:16 <h01ger> #startmeeting
14:59:16 <MeetBot> Meeting started Tue Jan 25 14:59:16 2022 UTC.  The chair is h01ger. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:59:16 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
14:59:23 * h01ger moins
14:59:35 <h01ger> #topic agenda at https://pad.riseup.net/p/rb-irc-meetings-keep
15:00:01 <h01ger> #topic say hi or otherwise indicate your presence if you like (this meeting is publically logged)
15:00:22 * h01ger Holger Levsen says hi and welcome (again) to r-b in 2022 :-D
15:00:48 * vagrantc = Vagrant Cascadian
15:01:06 * danielsh waves
15:01:07 <h01ger> & thanks to whoever added the link to the november meeting to that agenda pad! <3
15:01:14 * aehlig is Klaus Aehlig
15:02:09 * h01ger will wait some more minutes to give more chance to say hi
15:02:56 <aparcar> Hi!
15:03:00 <rclobus> rclobus = Roland Clobus, working in reproducible live-build-based ISO images
15:03:30 <aparcar> aparcar = Paul Spooren, working (mostly) on reproducible OpenWrt builds
15:03:41 <saksophony> Hello! saksophony = Aditya Sirish, NYU, operating r-b.engineering.nyu.edu, mostly lurking
15:04:04 <bmwiedemann> bmwiedemann: = Bernhard M. Wiedemann, working on openSUSE reproducible builds, also contributing upstream
15:04:20 <Foxboron> Yo! Morten Linderud. Working on Arch Linux and misc upstreams and projects :)
15:05:30 <h01ger> yay. happy to see you all! :)
15:05:54 <h01ger> so lets start. late comers can still and always say hi :)
15:06:08 <h01ger> and please amend the agenda as needed..
15:06:21 <h01ger> #topic short time slot for checkins from various projects
15:06:40 <h01ger> #topic short time slot for:     Alpine Linux: status update (Ariadne)
15:07:12 <h01ger> i guess Ariadne is not here, so we'll move on in 15sec or so
15:07:35 <h01ger> #topic short time slot for:     Arch Linux: rebuilder status update (kpcyrd)
15:07:56 <h01ger> or Foxboron or saksophony maybe? :)
15:08:10 <Foxboron> I'm not aware of any progress :)
15:08:23 <saksophony> I have nothing to report either :)
15:08:36 <Foxboron> We have some regressions in our rebuilder infra because of new developer tooling. But that will iron itself out after a little while
15:08:42 <h01ger> so the rebuilder is happily rebuilding i assume :)
15:08:51 <Foxboron> Mostly :D
15:08:54 <h01ger> new developer tooling in arch?
15:09:24 <jelle> h01ger: we now record the used "buildtool" to handle changed CFLAGS
15:09:30 <jelle> buildtool and version
15:09:44 <h01ger> ah, nice. record in .buildinfo or where?
15:09:56 <jelle> it's recorded in BUILDINFO indeed
15:10:08 <h01ger> coolio
15:10:08 <danielsh> "buildtool" = gcc/clang ?
15:10:36 <jelle> danielsh: ah no the buildtool is like hmm dpkg-buildpackage, the package which provides the buildscripts
15:10:53 <Ariadne> i'm presently driving
15:10:53 * fepitre waves and is sorry to be late
15:11:02 <danielsh> jelle, *nod* ty
15:11:20 <Ariadne> (at a stop sign)
15:11:21 <h01ger> Ariadne: drive safely and give us a report later, eg in a month ;)
15:11:34 <h01ger> (or whenever you dont drive anymore)
15:11:34 <jelle> danielsh: https://github.com/archlinux/devtools/pull/74 for background info
15:11:39 <danielsh> MeetBot, go go gadget text-to-speech
15:11:39 <MeetBot> danielsh: Error: "go" is not a valid command.
15:11:58 <h01ger> fepitre: hi & just in time
15:12:25 <danielsh> jelle, tyvm
15:12:31 <h01ger> #info https://github.com/archlinux/devtools/pull/74 for background info on how arch records (in their .buildinfo files) the used "buildtool" to handle changed CFLAGS
15:12:56 <h01ger> thats it for arch / rebuilder now?
15:14:10 <h01ger> i guess so.
15:14:29 <h01ger> #topic short time slot for:     Debian: snapshot.d.o mirror status update (fepitre)
15:14:48 <fepitre> Everything looks fine. I've did few improvements for rclobus
15:15:03 <h01ger> :thumbsup:
15:15:14 <rclobus> Thanks!
15:15:28 <fepitre> I've a refactor that I've not merged yet because I lack time for validating it
15:15:39 <fepitre> but still, I would prefer to work on a better design
15:15:45 <h01ger> ah, cool!
15:16:03 <fepitre> notably: https://github.com/fepitre/debian-snapshot/projects
15:16:36 <fepitre> (you may click on the project for subtasks)
15:16:55 <fepitre> anyway, currently all looks good for rebuilding Debian ( kpcyrd can confirm it too )
15:17:16 * h01ger is very happy this project and subprojects exist (as a first step in getting it implemented :)
15:17:57 <fepitre> that's all for me for this part :)
15:17:58 * vagrantc cheers and dances
15:18:01 <h01ger> next topic i guess...
15:18:12 <h01ger> after some more time for dancing :)
15:18:29 <h01ger> #topic short time slot for:     Debian: rebuilder (beta.t.r-b.o) status update (h01ger)
15:19:02 <h01ger> little progress here over the years :/
15:19:28 <h01ger> though beta.t.r-b.o "looks good" as it still points to fepitre's server
15:19:47 <h01ger> and i'm happy the url has been shown around here and there
15:19:55 <fepitre> I owe you also some improvement for showing results
15:20:07 <fepitre> notably to split Debian suite
15:20:16 <fepitre> but more generally, I need to do this: https://github.com/fepitre/package-rebuilder/issues/7
15:20:37 <h01ger> my biggest concern is that loading of https://beta.tests.reproducible-builds.org/debian.html takes more than 10 sec even on a fast computer
15:20:51 <fepitre> h01ger, because it loads a big JSON
15:20:59 <h01ger> ah, cool i guess thats what you mean with 'split the debian suite'
15:21:04 <fepitre> yup
15:21:06 <h01ger> fepitre: i dont care why ;-D
15:21:15 <h01ger> (i assumed that also :)
15:21:27 <fepitre> this is what I meant indeed
15:21:37 <h01ger> :thumbsup:
15:22:00 <jelle> fepitre: does this use rebuilderd?
15:22:06 <fepitre> jelle, no
15:22:08 <jelle> aha
15:22:08 <h01ger> https://github.com/fepitre/package-rebuilder/issues/7 is not about the refactoring you've done and need to verify, or is it?
15:22:29 <fepitre> h01ger, this one is for package-rebuilder
15:22:31 <h01ger> #save
15:22:41 <fepitre> the other one I talked about earlier was for the snapshot service
15:22:46 <h01ger> ah
15:22:48 <h01ger> sorry
15:23:07 <fepitre> this one is to improve package-rebuilder and how you can manage it, how you get results (API) etc
15:23:15 <fepitre> no problem
15:23:16 <h01ger> raboof: there's backlog at http://meetbot.debian.net/reproducible-builds/2022/reproducible-builds.2022-01-25-14.59.log.html (not sure whether you have it or not)
15:23:43 <fepitre> jelle, it uses https://github.com/fepitre/package-rebuilder
15:23:48 <jelle> fepitre: aha!
15:23:59 <raboof> thanks! o/ latecomer, Arnout Engelen, NixOS, various upstreams and general enthousiast :)
15:24:12 <h01ger> \o
15:24:28 <fepitre> which internally uses https://github.com/fepitre/debrebuild (like rebuilderd) to rebuild Debian
15:24:49 <h01ger> rather btw & offtopic atm, but i really like this people indicating what they work on. now we all know whom to ask about nix :)
15:25:01 <h01ger> so, next topic?>
15:25:08 <fepitre> ok for me :)
15:25:34 <h01ger> #topic short time slot for:     Debian live-build (rclobus)
15:25:51 <rclobus> I've sent a summary in advance: https://lists.debian.org/debian-live/2022/01/msg00015.html
15:26:16 <rclobus> Most notably: I've placed all hooks that I need for generating reproducible image in the live-build repository.
15:26:23 <h01ger> #info https://lists.debian.org/debian-live/2022/01/msg00015.html
15:26:30 <rclobus> That saves me some duplication work.
15:26:37 <h01ger> rclobus: thats super cool
15:26:41 <rclobus> #info https://wiki.debian.org/ReproducibleInstalls/LiveImages
15:26:44 <h01ger> upstream all the things
15:27:10 <h01ger> rclobus: i guess thats all for here+now?
15:27:30 <rclobus> Yes, that's the main point.
15:27:37 <h01ger> :thumbsup:
15:28:11 <vagrantc> excellently short and high density :)
15:28:23 <h01ger> #topic short time slot for:     F-Droid (obfusk)
15:29:22 <h01ger> i dont think we've seen obfusk or any other fdroid person recently
15:29:38 <h01ger> #topic short time slot for:     rebuilderd: status update (kpcyrd)
15:30:06 <h01ger> nor kpcyrd and we've covered some rebuilderd news before already too, so..
15:30:30 <h01ger> we are done with the short time slots :)
15:30:41 <h01ger> not sure if the upcoming ones will be much longer though
15:30:43 <h01ger> anyway
15:30:49 <h01ger> #topic r-b summit 2022 (mapreri)
15:30:58 * fepitre is happy that debrebuild.py is useful!
15:31:16 <h01ger> not seen mapreri here yet..
15:31:55 <h01ger> mapreri wrote a mail on Oct 21 2021 asking for all our thoughs on a summit in 2022
15:33:06 <aparcar> were there any responses? I didn't see any
15:33:08 <h01ger> not sure what else to say about this now. it also doesnt feel sensible to even ask what you'd think about an online event :/
15:33:27 <aparcar> why doesn't it feel sensible?
15:33:36 <h01ger> aparcar: i see 5 replies to that mail
15:33:37 <danielsh> aparcar, yes there have been
15:33:45 <h01ger> aparcar: because we are all sick of online meetings
15:33:57 <h01ger> (i assume, knowing i will be wrong on this :)
15:33:58 <aparcar> I missed it, will check the archives sorry
15:34:32 <h01ger> https://lists.reproducible-builds.org/pipermail/rb-general/2021-October/002404.html
15:34:37 <h01ger> and followups
15:34:41 <vagrantc> might not hurt to at least have an online social if not a meeting
15:34:51 * h01ger would be happy to hear peoples though on a 2022 summit
15:35:00 <h01ger> offline or online or hybrid or whatever you think
15:35:08 <danielsh> I think there's two separate questions.  1. Where (online / offline)   2. What will it comprise
15:35:18 <danielsh> Whether it's talks, or hacking, or anything else
15:35:40 <vagrantc> three questions!
15:35:41 <h01ger> danielsh: the previous summits were 'something else
15:35:42 <h01ger> '
15:36:08 <vagrantc> a bit of hacking but mostly something else, yeah
15:36:13 <danielsh> Yeah, that's good
15:36:33 <danielsh> synchronous events are few and far between; better to use them for discussions
15:36:45 <danielsh> but discussions don't strictly need to be in person :)
15:37:23 <vagrantc> while i too weary of online meetings, mailing list discussions seem to trail off pretty quickly also
15:38:37 <vagrantc> different mediums sometimes trigger different dynamics
15:39:14 <h01ger> yeah
15:39:55 <vagrantc> but i think for the topic at hand it's fair to say we have no real idea when we'll be able to meet in person again :/
15:39:57 <h01ger> i'd very much prefer an on-site outside event (and somehow i have not given up hope on this happening in 2022 on the 25th day of 2022 yet ;)
15:40:12 <h01ger> & what vagrantc just said
15:40:15 <danielsh> vagrantc, +1
15:40:19 <h01ger> so i'd say, move on to the next topic now
15:40:27 <danielsh> one moment
15:40:34 <h01ger> ok
15:40:37 * vagrantc also hopes and hopes :)
15:40:47 <danielsh> we can't have an in-person event... so we should think of how to get the same benefits in the given situation
15:41:06 <danielsh> e.g., if in-person means more chances for interaction/feedback, we should try to make that happen online. on IRC, lists, voice chat, whatever.
15:41:08 <danielsh> .
15:42:05 <h01ger> i'm not sure about "we can't have an in-person event", it always depends whom to exclude :/ (and online events also exclude people.)
15:42:23 <aparcar> shouldn't we table that and see how corona develops over the next month?
15:42:36 <h01ger> the world has sooo many regions..
15:42:56 <vagrantc> to some degree, i feel like we should just do more types of engagement and not necessarily have to be involved in all of them
15:43:04 <h01ger> but yes, we should move on *for now* and *here*. and have this topic again in a month here, and anytime on the list and also on irc, when there's no formal irc meeting
15:43:18 <h01ger> +1 on what vagrantc just said
15:43:28 <danielsh> +1
15:44:22 <h01ger> next topic then, though its a "no show"
15:44:25 <h01ger> #topic     r-b summit 2022 (mapreri)
15:44:35 <h01ger> #topic r-b.o/docs/rebuilders and conflict with r-b.o/tools (h01ger)
15:44:41 <h01ger> (sorry about the noise)
15:45:02 <h01ger> nothing happening on improving rebuilder/tools documentation :(
15:45:57 <fepitre> yeah...problably I'm part of this failure
15:46:04 <h01ger> we all are.
15:46:15 <h01ger> and this is not a failure but slow progress ;)
15:46:24 <h01ger> #topic     https://pad.sfconservancy.org/p/grow-r-b-debian - collecting ideas where to improve debian
15:46:28 <danielsh> is there a bug filed for this somewhere?
15:46:31 <danielsh> the previous topic
15:46:55 <h01ger> danielsh: no :( i think it would be great if you'd file one (as issue against the website.git)
15:47:31 <danielsh> I would, but I don't know what it should say exactly
15:47:52 <h01ger> those two pages should be merged and/or better split
15:47:58 <h01ger> read them and you should see
15:48:23 <h01ger> or lets discuss after the meeting? or should i go back to that topic formally?
15:48:34 <danielsh> after
15:48:41 <h01ger> ok, coolio
15:48:53 <h01ger> about 'grow your ideas for debian project'
15:49:05 <h01ger> https://salsa.debian.org/debian/grow-your-ideas
15:49:27 <h01ger> is an initiative to collect ideas how to grow debian in all kinds of areas
15:49:37 <h01ger> so i created https://pad.sfconservancy.org/p/grow-r-b-debian
15:49:46 <h01ger> to collect our ideas regarding improving r-b for debian
15:50:24 <h01ger> there's a timeline involved, so in about 10 days the pad should be moved into that salsa repo (i plan to do that)
15:50:53 <h01ger> so this topic is to make all of you (debian people) to go to https://pad.sfconservancy.org/p/grow-r-b-debian and write down your ideas there
15:51:01 <h01ger> .
15:51:22 <h01ger> \o/
15:51:27 <h01ger> danielsh: ^ :)
15:52:01 <danielsh> for the log - I just filed https://salsa.debian.org/reproducible-builds/reproducible-website/-/issues/35 for the previous topic
15:53:20 <h01ger> so i guess thats it for grow your ideas for now too..
15:54:00 <h01ger> #topic any other business
15:54:12 <h01ger> rclobus: Contacts with texlive?
15:54:35 <rclobus> I've seen that the Cinnamon image nowadays relies on texlive-base.
15:55:14 <rclobus> Wasn't there in the past something about the tex-world solving issues differently than SOURCE_DATE_EPOCH?
15:56:10 <vagrantc> they originally implemented their own variable name, which defeated the point for the most part
15:56:33 <vagrantc> we convinced them to use SOURCE_DATE_EPOCH, but they also require a second variable set, FORCE_SOURCE_DATE=1
15:57:10 <rclobus> Thanks, I'll keep that in mind when I propose new patches for them :-)
15:57:39 <danielsh> Why?  They thing S_D_E could be set unintentionally?
15:57:42 <danielsh> think*
15:57:43 <vagrantc> which is mentioned in https://reproducible-builds.org/docs/source-date-epoch/ and i think the spec as well
15:57:53 <vagrantc> danielsh: somehow, yes.
15:58:15 <danielsh> ack
15:58:23 <vagrantc> oops, cat walked across the keyboard and now SOURCE_DATE_EPOCH=-25
15:58:32 <vagrantc> seems implausible, but ... who knows.
15:58:35 <danielsh> said cat could also set FORCE_*
15:58:39 <h01ger> lol
15:58:46 <vagrantc> ah, true, we need a third variable
15:58:52 <danielsh> it's like a euphemism treadmill (q.v.)
15:59:06 <danielsh> CATS_ARE_UGLY=1 ?
15:59:10 <danielsh> no cat will set that.
15:59:15 <rclobus> MEOW=1
15:59:28 * h01ger thinks we'll just needs to wait until texlive fixes this better
15:59:30 <rclobus> Anyway, thanks... Next topic?
15:59:43 <h01ger> sure
15:59:54 <h01ger> any other business is rather freeflowing
16:00:13 <vagrantc> there might also be issues with timezones and SOURCE_DATE_EPOCH if the implementation is only partly done correctly
16:00:14 <rclobus> h01ger is always asking me to ask about snapshot.reproducible-builds.org...
16:00:14 <danielsh> do you want to #info the issue 35 for the meeting logs ?
16:00:16 <h01ger> rclobus: but the topic you added to the agenda was already covered as short time slot
16:00:26 <h01ger> danielsh: if in doubt, always
16:00:36 <h01ger> #save
16:01:07 <h01ger> rclobus: short time slot for: Debian: rebuilder (beta.t.r-b.o) status update (h01ger) (h01ger, 15:18:29)
16:01:10 <danielsh> link's not there
16:01:23 <rclobus> Ah, I didn't recognize it as such.
16:01:38 <danielsh> #info filed https://salsa.debian.org/reproducible-builds/reproducible-website/-/issues/35 (see item 13 above)
16:01:40 <rclobus> Then this topic just disappeared...
16:03:15 <rclobus> Topic done :-)
16:03:26 <h01ger> :)
16:03:45 <h01ger> surely we can/could also continue topics here..
16:04:44 <fepitre> do you know why ubuntu does not publish .buildinfo?
16:05:32 <vagrantc> debian doesn't either :)
16:05:35 <h01ger> lol
16:05:42 <h01ger> debian does
16:05:57 <vagrantc> oh, yeah, i guess we're debian :)
16:06:00 <h01ger> just via single volunteers, just like eg the amd64 port was started as well
16:06:09 <vagrantc> right, not part of infrastructure
16:06:14 <h01ger> who would have thought amd64 would become a mainstream arch???
16:06:23 <fepitre> Just wondering because more and more people are using Ubuntu (server) because of security fixes that are done much quicker than Debian
16:06:55 <vagrantc> fepitre: it's a good question ... no idea if they're even saving the .buildinfo files at all
16:06:56 <h01ger> fepitre: i have some guesses but i'm not sure i want to share them here.
16:07:08 <fepitre> h01ger, sure no problem
16:09:13 <h01ger> just this: i believe its more a lawyer reasons then technical. AIUI you need a licence to distribute (modified) ubuntu or some such. AIUI its a bit similar with redhat, where they publish the sources but. - and then: i have no clue, this is no advice whatsover and random guessing.
16:09:47 <h01ger> my other guess is that they are waiting til debian has done the technical work and then they will just enable it and sell it as how cool they are.
16:09:50 * h01ger shrugs
16:10:17 <vagrantc> i've definitely heard from some folks at canonical that they're waiting for Debian to *solve* reproducible builds
16:10:23 * h01ger is happy for all the ubuntu users
16:12:46 * h01ger is happy for all the free software users
16:12:50 <h01ger> any other business?
16:14:53 <h01ger> #topic next meeting: here, last Tuesday of the month, 15 UTC, so 20220222:15oo
16:15:22 <h01ger> & thank you all for attending & sharing your cool works!
16:15:33 <vagrantc> indeed, thanks everyone!
16:16:31 <h01ger> #endmeeting