16:02:29 <quidam> #startmeeting 09/Aug/2019
16:02:29 <Trisnity> Meeting started Fri Aug  9 16:02:29 2019 UTC.  The chair is quidam. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:02:29 <Trisnity> Useful Commands: #action #agreed #help #info #idea #link #topic.
16:02:39 <quidam> hi everybody!
16:03:17 <quidam> I'm checking on the build logs
16:03:49 <quidam> I see a lot of build issues, but I guess it comes from the change in the watchdog script, now it launches anything not built
16:03:55 <dctrud> I think the etiona package helpers still aren't working due to the hkps://....443 not working ?
16:04:21 <dctrud> e.g. http://jenkins.trisquel.info:8085/job/signed-dsc/lastFailedBuild/console
16:04:48 <dctrud> '11:55:07 gpg: keyserver receive failed: g10err=127' which is the symptom of no hkps support
16:04:56 <quidam> yeah, I think you are right
16:05:16 <quidam> also I've been seeing ipv6 errors with apt
16:05:28 <quidam> that can be fixed by forcing ipv4
16:06:04 <quidam> dctrud: I had problems rebuilding the etiona jails
16:06:21 <quidam> I'll start checking those things
16:07:33 <bandali> i was supposed to bring this up last week, but i was away, so i’ll mention it now: Wioxjk was asking if there were any sysadmin/devops type tasks they could help out with
16:07:38 <dctrud> This week the only thing I've done is look at some helpers that can be removed - in etiona MR https://devel.trisquel.info/trisquel/package-helpers/merge_requests/230
16:08:41 <quidam> dctrud: and none of them did build, right?
16:09:02 <quidam> because if the changes need to be removed that requires manual action
16:10:01 <quidam> merged
16:10:44 <dctrud> quidam: correct those didn't build
16:11:16 <dctrud> I'm building all the helpers locally on my desktop... and trying to go through a few per week which fail
16:11:30 <dctrud> picked the easy ones to start to be honest ;-)
16:13:46 <quidam> it is still good help, thanks!
16:16:38 <quidam> building new etiona jails
16:19:31 <quidam> ok, that should solve some issues, now waiting (for it to crash...)
16:21:06 <quidam> a piece of news, I'll be taking vacation (going to spain) from the 15 to the 23, so I will be missing the next two meetings :/
16:21:37 <dctrud> enjoy the trip... I'd love a trip to Spain!
16:23:01 <dctrud> Have some family in Valladolid
16:24:52 <quidam> oh, nice
16:27:32 <dctrud> quidam: I just noticed (apologies) that gnupg-curl is *not* in bionic... it was in xenial, but not bionic
16:28:17 <dctrud> if the jail is purely packages sourced from bionic then I'm not sure what is needed for hkps now.
16:32:16 <quidam> I just realized that it failed to build
16:32:46 <quidam> I guess in bionic it just works? idk
16:32:52 <quidam> I'll try without that package
16:33:12 <dctrud> I think maybe in bionic the minimal `gpg` package for public key operations doesn't support hkps but the full `gnupg` does
16:33:24 <dctrud> (without needing a separate gnupg-curl)
16:34:02 <dctrud> looking at the CHANGELOG they reorganized the binary packages produced by the gnupg2 source package
16:34:38 <dctrud> > - new package 'gpg' is just for public key operations
16:34:47 <dctrud> > - 'gnupg' package is the full suite
16:35:56 <dctrud> not sure what your jail is pulling in? debootstrap maybe only goes for gpg
16:37:32 <quidam> in etiona it just works
16:37:44 <quidam> tried apt-key adv --keyserver hkps://keyserver.ubuntu.com:443 --recv-keys 8B48AD6246925553
16:38:42 <quidam> idk why it fails when it is not run manually
16:38:57 <dctrud> is that in the jail, or another etiona environment?
16:39:18 <quidam> jail
16:39:34 <quidam> but I just rebuilt the jail
16:39:44 <quidam> so idk what changed
16:40:11 <quidam> oh, wait
16:40:20 <quidam> the signed-dsc process is always run on flidas
16:40:22 <quidam> iirc
16:40:51 <dctrud> ah okay - on flidas it does need `gnupg-curl`
16:41:49 <dctrud> `make-gnupg2` is one of the failing etiona helpers... so I guess that wouldn't get far anyhow on an etiona jail
16:42:04 <quidam> ok, works in either
16:42:14 <quidam> I'll re-launch the one you linked
16:42:57 <quidam> ...nope
16:44:52 <quidam> fixed
16:45:46 <dctrud> great!
16:47:36 <quidam> but now I want to make some changes to the process...
16:47:36 <quidam> one min
16:54:54 <quidam> or several
16:55:12 <quidam> in the meantime... how is everybody doing?
16:58:59 <bandali> i just got done marking two exam questions of 100 students, so brain is pretty fried lol
16:59:19 <dctrud> ouch - I remember that feeling
16:59:55 <bandali> dctrud, :p did you do grad school too?
17:01:42 <dctrud> bandali: I did a PhD in CS, and then covered lecturing, dissertation marking et al. for a lecturer out on maternity leave for a semester.
17:02:36 <bandali> damn! i’m doing a masters in CS, hopefully wrapping it up in fall
17:02:44 <bandali> and am considering a phd too :p
17:03:05 <bandali> though i’m not *that* excited about it, it’s just it may be my best option if i don’t want to “go back where i came from”
17:03:30 <dctrud> My PhD was in neural networks just before they become cool with the GPU stuff... heh. I had bad timing on the job front there :-)
17:04:20 <dctrud> I went over to Bioinformatics research rather than into machine-learning for work... then HPC in academia... now a startup
17:07:10 <bandali> ah, sorry to hear that
17:07:22 <bandali> but cool! that’s a cool journey
17:07:35 <bandali> my research is in formal methods
17:08:14 <bandali> i’m more on the model checking and “lightweight” side now, but am hoping to do my phd more in the interactive theorem proving side
17:10:07 <dctrud> cool - that's interesting. There's a lot of reliance on purely empirical things these days, glad to hear of people interested in formal method.
17:10:39 <bandali> indeed
17:11:44 <dctrud> I can vaguely recall having to write a formal specification in Z for a coursework project... too long ago
17:11:56 <dctrud> makes me feel old thinking back now
17:13:56 <bandali> haha :) yeah Z was/is? pretty big
17:14:11 <bandali> these days thankfully there’s much more tooling/automation for the various methods
17:15:33 <dctrud> time to go bury my head in Go code again.. lunchtime over :-)
17:15:39 <dctrud> quidam: enjoy your vacation
17:17:54 <bandali> nice chatting with you dctrud
17:18:04 <bandali> :)
17:24:28 <quidam> hi valessio
17:24:42 <quidam> and two more? what happened there
17:44:55 <dctrud> well - I guess I'm stuck waiting for things so back lurking here :-)
17:45:29 <dctrud> quidam: is there any use to trying to fix the current trisquel-gtk-theme to be the same on etiona as it is on flidas? Or would there be a new theme?
17:47:52 <quidam> dctrud: the theme is based on https://www.deviantart.com/lassekongo83/art/Zukitwo-203936861
17:48:21 <quidam> it can be updated if the new version works in etiona, or replaced by something that looks similar
17:48:37 <quidam> https://github.com/lassekongo83/zuki-themes
17:48:41 <quidam> seems to be maintained
17:54:05 <quidam> dctrud: I still see issues with some gpg stuff
17:54:10 <quidam> try make-ufw
17:55:05 <dctrud> quidam - oh I thought it was based on bluementa from seeing: https://devel.trisquel.info/trisquel/trisquel-packages/blob/master/8.0/trisquel-gtk-theme/data/usr/share/themes/Trisquel/gtk-3.0/Makefile.am?
17:56:55 <dctrud> that make-ufw is failing on flidas too, not just etiona right?
17:58:33 <quidam> you are right about the menta thing
17:58:39 <quidam> I lose track
17:58:46 <quidam> make-ufw fails on flidas
17:58:56 <quidam> there is no etiona helper for that, it is a backport
17:59:47 <dctrud> I did bring across most of the newer bluementa into a trisquel-gtk-theme I have such that the basic GTK stuff is okay on mate... but it doesn't look good in gnome-shell, and the dark transparent mate panel isn't right
17:59:57 <dctrud> I'll likely get around to trying to fix that
18:04:25 <quidam> genome-sell part can be taken from the default theme
18:04:33 <quidam> and the panel always takes work...
18:04:35 <quidam> every time
18:14:28 <dctrud> okay I'll take a go on that. I'll also try and work out some of the stuff to make it possible to install `gnome-shell` or a `vanilla-gnome-desktop`
18:15:16 <dctrud> I have gnome shell working only through some very hairy ignoring of missing ubuntu packages, and manually bringing in some stuff that I noted doesn't contain ubuntu customization any more
18:16:18 <dctrud> it's rather annoying how many things are dependent on 'snap' stuff... they have that tied up with a lot of gnome pieces in 18.04. I gather it's been unpicked a bit to plugins going toward 20.04
18:21:26 <Ark74> seems like packages related to ffmpeg are broken for etiona
18:22:12 <Ark74> ffmpeg or trisquel-codecs not sure
18:22:46 <quidam> dctrud: we can make a placeholder for snap
18:23:04 <quidam> like having the trisquel-minimal package provide snap
18:23:11 <quidam> ugly hack though
18:29:22 <quidam> took me a while to figure out the gpg problem
18:29:35 <quidam> the script is fine, the upstream key for the package was revoked
18:29:50 <quidam> but it built before, so I assume there was a problem in our verification?
18:30:04 <quidam> or it was revoked recently
18:30:19 <quidam> I contacted upstream about those revocations
18:44:06 <dctrud> Ark74 - what particularly isn't working? I'm able to install ffmpeg okay
18:45:57 <Ark74> dctrud, after upgrading from flidas to etiona a bunch of ffmpeg tools got orphaned
18:46:30 <Ark74> and now some ffmpeg commands that were working aren't anymore
18:47:22 <Ark74> is more related with codecs that just installing ffmpeg, i'll try to investigate further
18:47:46 <Ark74> (I guess) ^
19:13:02 <quidam> ok, the gpg stuff seems to be working for as far as we can control, and I'm waiting for upstream info on the revokes
19:13:09 <quidam> any other topics?
19:18:42 <dctrud> I'm off into meetings in a second. Will see what I get done with helper stuff over next 2 weeks. Enjoy your trip.
20:13:56 <quidam> thanks everybody!
20:14:01 <quidam> #endmeeting