Thursday, 2021-04-08

*** crobinso has quit IRC00:20
*** fuentess has quit IRC00:38
*** egernst has joined #kata-dev02:24
*** egernst has quit IRC03:31
*** egernst has joined #kata-dev03:33
*** egernst has joined #kata-dev03:37
*** egernst has quit IRC03:40
*** egernst has joined #kata-dev03:41
*** egernst has quit IRC04:03
*** egernst has joined #kata-dev04:03
*** egernst has quit IRC04:36
*** sameo has joined #kata-dev05:34
*** sgarzare has joined #kata-dev06:43
*** fgiudici has joined #kata-dev06:43
*** dklyle has quit IRC06:54
*** egernst has joined #kata-dev07:09
*** egernst has quit IRC07:17
*** jodh has joined #kata-dev07:29
kata-irc-bot<james.o.hunt> Thanks @fidencio!07:55
kata-irc-bot<fidencio> @james.o.hunt, you're in the project for quite some time, do you happen to know why we never went for having a `make dist` target and distributing the tarball generated rather than just using what GitHub automagically gives us as a tarball?07:57
kata-irc-bot<fidencio> I'm mostly asking because, for instance, we don't allow warnings on the rust code, which is amazing for development, but doesn't scale very well for people who'll be consuming our source from a tarball later on07:57
kata-irc-bot<fidencio> using, most likely, a different version of rust07:58
kata-irc-bot<james.o.hunt> Well, it's not an autotools project and GH was providing the tarballs for free ;) But also in the early days nobody apart from us was doing any distro packaging, so there wasn't an issue.08:00
kata-irc-bot<james.o.hunt> From what you're saying though, couldn't we just add a variable in `src/agent/Makefile` to allow packagers to override the current `--deny-warnings` setting for building the agent?08:01
*** davidgiluk has joined #kata-dev08:02
kata-irc-bot<james.o.hunt> I presume you mean that packagers may have issues building the code with different (newer) versions of `rustc`` ?08:02
kata-irc-bot<fidencio> Yep, I'm think both about packagers  and random people building the project from the released tarball.08:02
kata-irc-bot<fidencio> Having someone going through the process of "trying, failing, searching for info, finding out about what has to be replaced, trying again" is quite problematic, IMHO.08:03
kata-irc-bot<fidencio> Almost feels like we should invest some time and get some tooling working with the project, be it autotools, be it meson, but have something that actually can help us.08:04
kata-irc-bot<fidencio> Those warnings are one example, vendored code is another one.08:04
kata-irc-bot<fidencio> We should distribute the vendored code (for the agent) as the majority of the distros won't allow internet connection during the build of their packages08:05
kata-irc-bot<fidencio> But we can't do that right now as we rely on the GH tarball that we can't modify much.08:06
*** sgarzare has quit IRC08:13
kata-irc-bot<james.o.hunt> Sounds like a good topic for an AC meeting discussion, particularly if we can get some representation from various distro packagers.08:13
kata-irc-bot<james.o.hunt> And yeah, I agree that meson would involve much less pain than autotools ;)08:14
kata-irc-bot<fidencio> We need to get other distro packagers around to start with. :slightly_smiling_face:.08:15
*** sgarzare has joined #kata-dev08:52
kata-irc-bot<fidencio> @fupan, @tim, let me loop you here. I can't simply backport James changes to the stable-2.0 branch, they don't build. :slightly_smiling_face:  My I get some of your cycles to take a look at what's going on on stable-2.0 branch? A super easy way to reproduce the issue is `make` from the `kata-containers`  dir.09:08
kata-irc-bot<fidencio> Head's up, updating the crate version from `0.3.0`, used in the stable-2.0 branch to a newer one would generate a whole lot of other pieces to be udpated.09:17
kata-irc-bot<tim> OK09:20
*** egernst has joined #kata-dev09:23
kata-irc-bot<fidencio> Thanks, appreciated!09:24
kata-irc-bot<tim>09:26
kata-irc-bot<fidencio> @tim, yep, exactly that!09:27
kata-irc-bot<tim> it seems that we have to update the ttrpc09:28
kata-irc-bot<fidencio> I wonder why, the agent has `0.3.0`, which matches the one in the agent-ctl09:28
*** egernst has quit IRC09:29
kata-irc-bot<fidencio> I'm fine about bumping the agent to `0.5.0`, but that will require changes on the `agent` and `agent-ctl` , I guess.  And I'm not familiar with any of those.09:29
kata-irc-bot<tim> I have updated the `ttrpc = { git = "https://github.com/containerd/ttrpc-rust", branch="0.3.0" }` to  `ttrpc = "0.3.0"` and it works09:30
kata-irc-bot<fidencio> Hmm. Just that?  Let me check it here09:32
kata-irc-bot<fidencio> Indeed.  Mind to submit the patch to the stable-2.0 branch then?  I'll review and get it merged ASAP.09:34
kata-irc-bot<tim> ok, will do that.09:35
kata-irc-bot<fidencio> Thanks!09:37
kata-irc-bot<tim> you are welcome09:38
*** egernst has joined #kata-dev09:53
*** egernst has quit IRC09:59
*** ailan has joined #kata-dev10:03
*** Yarboa has quit IRC10:09
*** Yarboa has joined #kata-dev10:09
*** egernst has joined #kata-dev11:56
*** egernst has quit IRC12:01
*** devimc has joined #kata-dev12:16
kata-irc-bot<fidencio> Folks if someone have some acks to spare: • https://github.com/kata-containers/kata-containers/pull/1649https://github.com/kata-containers/kata-containers/pull/1652 Those two backport PRs would be a good candidate to receive those ACKs. :-)12:21
*** khyr0n has quit IRC12:33
*** khyr0n has joined #kata-dev12:33
*** crobinso has joined #kata-dev12:46
*** egernst has joined #kata-dev14:34
*** dklyle has joined #kata-dev14:58
*** dklyle has quit IRC15:02
*** dklyle has joined #kata-dev15:02
kata-irc-bot<fidencio> Thanks!15:30
*** egernst has quit IRC15:49
*** egernst has joined #kata-dev15:50
*** crobinso has quit IRC16:43
*** fgiudici has quit IRC16:46
*** sgarzare has quit IRC16:49
*** jodh has quit IRC17:14
kata-irc-bot<eric.ernst> @gabriela.cervantes.te @chen.bo -- footprint issue discussion :thread:17:16
kata-irc-bot<eric.ernst> Can ya'll clarify - is this for just runtime repo, and sometime around beginning of March? Just master? or stable-1.12 as well?17:17
kata-irc-bot<gabriela.cervantes.te> metrics do not run on stable17:17
kata-irc-bot<gabriela.cervantes.te> I noticed the increase in master on the runtime repository around march 12th when somebody raised a PR17:18
kata-irc-bot<eric.ernst> not a lot of PRs: https://github.com/kata-containers/runtime/pulls?q=is%3Apr+is%3Aclosed++base%3Amaster+is%3Apr+merged%3A2021-02-14..2021-04-08+17:25
kata-irc-bot<eric.ernst> widening the search, doesn't show much more: https://github.com/pulls?q=-repo%3Akata-containers%2Fshim+-repo%3Akata-containers%[…]3Aclosed+base%3Amaster+is%3Apr+merged%3A2021-01-14..2021-03-13+17:31
kata-irc-bot<gabriela.cervantes.te> yes maybe also osbuilder17:32
kata-irc-bot<gabriela.cervantes.te> or agent17:32
kata-irc-bot<gabriela.cervantes.te> I have seen little activity in kata 1.x17:32
kata-irc-bot<gabriela.cervantes.te> so there should not be a lot of PRs17:32
*** ailan has quit IRC18:20
*** khyr0n has quit IRC19:09
*** khyr0n has joined #kata-dev19:09
*** davidgiluk has quit IRC19:25
*** devimc has quit IRC20:59
*** sameo has quit IRC21:11
kata-irc-bot<gmmaharaj> Thanks a  ton @tim and @fidencio. I am super new to rust, might i ask why that change allows the build to work? are the sources of ttrpc crate different from the one in the github path? is that it?21:25
*** Yarboa has quit IRC22:31
*** Yarboa has joined #kata-dev22:33
*** egernst has quit IRC22:49
*** egernst has joined #kata-dev22:53
*** egernst has quit IRC22:59
*** Yarboa has quit IRC23:21
*** Yarboa has joined #kata-dev23:24
*** egernst has joined #kata-dev23:34
*** egernst has quit IRC23:59

Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!