#archlinux-ports | Logs for 2017-08-01

Back
[01:55:09] -!- eschwartz has quit [Remote host closed the connection]
[01:59:52] -!- eschwartz has joined #archlinux-ports
[03:14:08] -!- eschwartz has quit [Remote host closed the connection]
[03:14:58] -!- Faalagorn has quit [Remote host closed the connection]
[03:15:03] -!- isacdaavid has joined #archlinux-ports
[03:15:59] -!- eschwartz has joined #archlinux-ports
[04:40:08] -!- eschwartz has quit [Ping timeout: 240 seconds]
[04:44:37] -!- eschwartz has joined #archlinux-ports
[05:15:31] -!- https_GK1wmSU has joined #archlinux-ports
[05:17:28] -!- https_GK1wmSU has parted #archlinux-ports
[06:48:17] -!- deep42thought has joined #archlinux-ports
[07:08:08] -!- isacdaavid has quit [Quit: isacdaavid]
[07:28:28] -!- eschwartz has quit [Ping timeout: 260 seconds]
[07:32:05] -!- deep42thought has quit [Remote host closed the connection]
[07:39:55] -!- guys has quit [Ping timeout: 258 seconds]
[07:44:31] -!- guys has joined #archlinux-ports
[07:48:30] -!- eschwartz has joined #archlinux-ports
[07:50:28] -!- eschwartz has quit [Remote host closed the connection]
[07:52:29] -!- eschwartz has joined #archlinux-ports
[08:12:26] <tyzoid> Okay, so two things
[08:12:44] <tyzoid> https://news.archlinux32.org - New isos are built/verified
[08:12:47] <phrik> Title: Archlinux 32 News (at news.archlinux32.org)
[08:13:03] <tyzoid> And also, my mirror, http://32.arlm.tyzoid.com now has https available, https://32.arlm.tyzoid.com
[08:13:04] <phrik> Title: Index of / (at 32.arlm.tyzoid.com)
[08:13:17] <tyzoid> anyway, I'm heading off, since it's getting late.
[08:13:21] <tyzoid> g'night, everyone
[08:42:36] -!- deep42thought has joined #archlinux-ports
[09:10:48] -!- eschwartz has quit [Ping timeout: 240 seconds]
[09:14:09] -!- guys has quit [Ping timeout: 248 seconds]
[09:22:28] -!- guys has joined #archlinux-ports
[09:24:54] -!- Faalagorn has joined #archlinux-ports
[09:25:29] -!- eschwartz has joined #archlinux-ports
[09:29:49] -!- eschwartz has quit [Remote host closed the connection]
[09:32:46] -!- eschwartz has joined #archlinux-ports
[09:45:40] -!- phrik has quit [Remote host closed the connection]
[09:45:49] -!- phrik has joined #archlinux-ports
[09:47:57] -!- Faalagorn has quit [Ping timeout: 240 seconds]
[10:03:57] -!- Faalagorn has joined #archlinux-ports
[10:12:27] -!- eschwartz has quit [Ping timeout: 240 seconds]
[10:13:12] -!- guys has quit [Ping timeout: 260 seconds]
[10:22:50] -!- guys has joined #archlinux-ports
[10:25:27] -!- eschwartz has joined #archlinux-ports
[10:30:33] -!- rewbycraft has quit [Ping timeout: 255 seconds]
[10:33:06] -!- rewbycraft has joined #archlinux-ports
[11:14:37] -!- CalimeroTeknik has quit [Ping timeout: 255 seconds]
[11:28:39] -!- CalimeroTeknik has joined #archlinux-ports
[12:17:26] -!- piernov has quit [Ping timeout: 246 seconds]
[12:26:48] -!- piernov has joined #archlinux-ports
[14:12:28] -!- eschwartz has quit [Ping timeout: 240 seconds]
[14:12:45] -!- guys has quit [Ping timeout: 240 seconds]
[14:24:49] -!- guys has joined #archlinux-ports
[14:32:16] -!- eschwartz has joined #archlinux-ports
[14:43:07] -!- eschwartz has quit [Remote host closed the connection]
[14:43:37] -!- eschwartz has joined #archlinux-ports
[14:54:33] <tyzoid_> So deep42thought, we should consider when we want to start removing older isos from the repos
[14:54:45] <tyzoid_> and if we want to archive them somewhere
[14:54:47] <deep42thought> ah, yeah
[14:54:58] <tyzoid_> mainline arch has their archive repo that has a historic archive of all packages
[14:55:10] <tyzoid_> could be helpful, but not sure it's necessary right now
[14:56:10] <deep42thought> How much capacity would that require?
[14:56:22] <deep42thought> sounds like a huge pile of data ...
[14:57:05] <tyzoid_> deep42thought: It does steadily grow. Like I said, not sure if it's useful
[14:57:16] <tyzoid_> rephrase: It's useful, not sure it's necessary*
[14:57:35] <deep42thought> :-)
[14:57:50] <tyzoid_> The benefit is it lets users upgrade their system in stages if they've gone a long time without upgrading
[14:58:24] <deep42thought> btw: Did you test-boot the isos with aif-ng?
[14:58:54] <tyzoid_> deep42thought: Not yet, going to do that next.
[14:59:06] <brtln> it is, and mostly useless now
[14:59:06] <brtln> we will be cutting it
[14:59:06] <brtln> I'd love to preserve it but it eats too much
[14:59:06] <brtln> there will be use case when we get to make packages reproducible
[14:59:08] <tyzoid_> I did a test boot/install as usual from the i686 one
[14:59:49] <deep42thought> tyzoid_: I should move all packages to stable then, right?
[14:59:54] <tyzoid_> brtln: Would you be trimming the tail end off? like only keep package archives back 1 year or two years?
[14:59:57] <tyzoid_> deep42thought: I built out of stable.
[15:00:07] <tyzoid_> deep42thought: testing hasn't been tested by me.
[15:00:07] <deep42thought> ah, crap
[15:00:20] <jelle> 26
[15:00:31] <deep42thought> jelle: why not 42?
[15:01:03] <tyzoid_> brtln: Also, how much space does that require?
[15:01:37] <tyzoid_> brtln: If it's under a few terabytes, I might want to make an archive of it.
[15:03:13] -!- eschwartz[m] has quit [Ping timeout: 276 seconds]
[15:03:20] -!- brtln has quit [Ping timeout: 246 seconds]
[15:12:35] -!- Abbott has parted #archlinux-ports
[15:42:17] -!- brtln has joined #archlinux-ports
[15:52:33] -!- eschwartz[m] has joined #archlinux-ports
[16:17:31] -!- tyzoid has quit [Disconnected by services]
[16:17:33] tyzoid_ is now known as tyzoid
[16:18:00] -!- tyler has joined #archlinux-ports
[16:18:20] tyler is now known as Guest56557
[16:19:38] -!- deep42thought has quit [Quit: Leaving.]
[16:32:54] -!- eschwartz has quit [Remote host closed the connection]
[16:35:35] -!- eschwartz has joined #archlinux-ports
[17:04:16] -!- eschwartz has quit [Remote host closed the connection]
[17:06:03] -!- eschwartz has joined #archlinux-ports
[17:40:07] -!- deep42thought has joined #archlinux-ports
[17:55:09] -!- eschwartz has quit [Remote host closed the connection]
[17:57:00] -!- eschwartz has joined #archlinux-ports
[18:07:49] -!- isacdaavid has joined #archlinux-ports
[18:29:06] <deep42thought> FYI: I scheduled a rebuild of all base and base-devel packages - hopefully this will solve all build issues :-)
[18:29:36] <tyzoid> deep42thought: Cool, thanks. Did we ever reach a concensus on removing older isos?
[18:29:46] <tyzoid> deep42thought: keep it at 3 months, like mainline does?
[18:29:48] -!- brtln_logger has quit [Ping timeout: 240 seconds]
[18:30:33] <deep42thought> well, my personal opinion on isos is: remove all but one stable iso
[18:30:44] <deep42thought> but I think, it might be worth keeping old packages (at some point)
[18:30:54] -!- brtln_logger has joined #archlinux-ports
[18:31:13] <deep42thought> however, currently, there is plenty of space on the build master, so I don't see need to rush
[18:31:28] <deep42thought> maybe we should harvest some opinions on the mailing list about that?
[18:31:32] <tyzoid> <sarcasm>lol, having a *wonderful* conversation on tabs-vs-spaces over in ##php</sarcasm>
[18:31:39] <deep42thought> s/buildmaster/master mirror/
[18:32:01] <tyzoid> deep42thought: Couldn't hurt. We should probably allocate space to make an archive of these packages.
[18:32:27] <tyzoid> deep42thought: If we set a timeframe, like "only keeping a public archive of 1year of packages"
[18:32:34] <tyzoid> then we can limit the space used in the future
[18:33:38] <deep42thought> hmm, ok
[18:34:43] <deep42thought> where would that space be?
[18:34:48] <deep42thought> on the master mirror?
[18:42:50] <deep42thought> I'm thinking simply of a separate folder "archive" next to "i686", "irc-logs", ... which would hold all '*.pkg.tar.xz's and '*.pkg.tar.xz.sig's (without an appropriate database file)
[18:44:23] <tyzoid> deep42thought: I would sugest not puting that archive on the main mirror, since that would dramatically increase the spare requirements for other mirrors
[18:44:33] <tyzoid> deep42thought: Possibly archive.archlinux32.org?
[18:44:46] <deep42thought> good point
[18:45:16] <deep42thought> do you have enough space for something like that somewhere?
[18:45:54] <tyzoid> deep42thought: Not on my cloud servers at the moment.
[18:45:58] <deep42thought> ok
[18:46:10] <tyzoid> deep42thought: In a few months, I might be able to do that.
[18:46:24] <tyzoid> deep42thought: I'd be fine maintaining an offline archive of those packages.
[18:46:55] <tyzoid> deep42thought: it's just an rsync without a delete.
[18:47:12] <deep42thought> Actually, I think, an offline archive is fine for now
[18:47:39] <tyzoid> deep42thought: I probably won't get to making it until mid august (aug. 15ish)
[18:47:57] <tyzoid> since I'll be away from the 4th to the 12th at a conference
[18:48:05] <deep42thought> The most probable scenario will be, that we break something serious (e.g. delete the gcc package from the master mirror) and need to get back to an older working configuration
[18:48:08] <tyzoid> I'll have internet, but I won't be able to do much in the way of projects
[18:48:21] <deep42thought> tyzoid: no problem
[18:48:36] <deep42thought> currently, everything is redundand anyway, because upstream still has i686 ;-)
[18:48:59] <deep42thought> in other words: You got time until November :-)
[18:51:36] -!- eschwartz has quit [Remote host closed the connection]
[18:53:29] -!- eschwartz has joined #archlinux-ports
[19:04:37] <tyzoid> deep42thought: Also, you may be interested in this: http://stats.arlm.tyzoid.com
[19:05:05] <tyzoid> I'll need to split those between x86_64 and i686, but that shows all packages sorted by download count from my mirror.
[19:05:34] <deep42thought> ah, nice :-)
[19:05:49] <deep42thought> I don't keep that kind of statistics
[19:06:00] <tyzoid> deep42thought: I've got it in json, for automatic consumption: http://stats.arlm.tyzoid.com
[19:06:07] <tyzoid> deep42thought: That's just crawling my apache logs
[19:06:20] <tyzoid> deep42thought: Figured I'd use it since I've got 'em
[19:06:54] <deep42thought> so it does not account for rsync downloads, then
[19:07:02] <deep42thought> but these should be minor, anyway
[19:07:05] <tyzoid> deep42thought: Correct.
[19:09:47] <deep42thought> Do you have any idea, why cryptsetup has the highest count by such a large margin?
[19:10:02] <deep42thought> Are there significantly more updates to cryptsetup than to everything else?
[19:10:27] <tyzoid> deep42thought: Not sure. That was my thought, but I have no idea.
[19:10:33] <tyzoid> deep42thought: I'd need to do more analysis on it.
[19:21:39] <deep42thought> lol
[19:22:01] <deep42thought> "makepkg --printsrcinfo" run on the linux kernel does not show 'groups=base'
[19:22:27] <brtln> tyzoid: I guess that it's because it became a dep of systemd recently
[19:23:20] <tyzoid> brtln: The question is less of why is it so high up, but why does it have an order of magnitude more downloads than the 'linux' package
[19:23:31] <deep42thought> or any other package
[19:23:36] <deep42thought> even systemd
[19:23:51] <deep42thought> (you don't need linux inside a container)
[19:23:52] <brtln> well, that's exactly why?
[19:24:15] <brtln> you can assume that almost every arch user has systemd installed
[19:24:17] <tyzoid> deep42thought: pacman -Si linux shows the base group
[19:24:22] <deep42thought> tyzoid: what time interval did you evaluate for the numbers?
[19:24:26] <brtln> if it got a new dep, there you go
[19:24:36] <tyzoid> deep42thought: It's over a period of almost a year, now
[19:24:36] <deep42thought> tyzoid: yes, but "makepkg --printsrcinfo" does not
[19:24:59] <deep42thought> and we can't rely on 'pacman -Qi' for the build master
[19:25:21] <tyzoid> deep42thought: This is for my mainline mirror, not for the archlinux32 mirror
[19:25:46] <tyzoid> I can compile similar stats there, eventually, but we don't have enough users afaik for that yet
[19:26:16] <deep42thought> don't compile these stats unless we get three digit numbers - otherwise it's demotivating ;-)
[19:26:17] <tyzoid> My logs reach back to 13/Jun/2016 when I launched the mirror
[19:27:58] <tyzoid> brtln: The thing is that cryptsetup is in the base group as well, but I wouldn't imagine it having 10x more dl's than linux, unless there were 10x more updates for cryptsetup
[19:28:54] <deep42thought> it might get significant more dl's if there are a) very few updates at all in the past year and b) it entered base (became a dependency of a base package) in the last year
[19:36:57] -!- eschwartz has quit [Remote host closed the connection]
[19:53:28] -!- piernov has quit [Ping timeout: 240 seconds]
[19:54:45] -!- guys has quit [Quit: A random quit message]
[19:56:29] <tyzoid> deep42thought: Potentially, but if I have the base group installed, that doesn't (afaik) track to the future. It's just a shortcut for installing a group of packages together
[19:57:00] -!- guys has joined #archlinux-ports
[20:01:29] -!- guys has quit [Client Quit]
[20:03:07] -!- guys has joined #archlinux-ports
[20:05:23] -!- eschwartz[m] has quit [Changing host]
[20:05:23] -!- eschwartz[m] has joined #archlinux-ports
[20:05:23] -!- eschwartz[m] has quit [Changing host]
[20:05:23] -!- eschwartz[m] has joined #archlinux-ports
[20:08:40] <deep42thought> tyzoid: you're right, but when it becomes a dependency of systemd, it will be installed even in the future
[20:09:50] -!- eschwartz has joined #archlinux-ports
[21:12:07] -!- Guest56557 has quit [Ping timeout: 246 seconds]
[21:13:28] -!- Guest56557 has joined #archlinux-ports
[21:14:21] <brtln> tyzoid: fair point
[21:14:42] <brtln> I didn't use base group for a few years but yeah, I guess that makes my hypothesis stupid
[21:15:12] <brtln> and I guess not that many people review what base consists of afterwards
[21:15:42] <tyzoid> brtln: The only stupid thing is forgetting to cover a potential reason when trying to figure something out. Better to give all hypothesis - including the bad - than miss the right one
[21:16:04] <brtln> deep42thought: by any chance, can you e-mail me that list of bugs to bpiotrowski@archlinux.org? I promise I will handle it tomorrow morning
[21:16:18] <deep42thought> will do
[21:16:35] <tyzoid> brtln: Sounds like a bold statement.
[21:16:48] <tyzoid> deep42thought: Give all bugs at one time
[21:16:56] <tyzoid> (jk)
[21:17:18] <deep42thought> I'll just execute 'bin/find-bugs | sendmail'
[21:17:26] <brtln> well as long as they're related to what you do…
[21:17:58] <tyzoid> brtln: Thanks again for the help ;)
[21:17:59] <deep42thought> btw, FS#54956 is fixed in trunk, but not in repos
[21:18:22] <tyzoid> brtln: It'd be impossible to do this without some TUs helping sort stuff out on our behalf
[21:19:31] <brtln> deep42thought: send it as well… I disagree on calling it a bug or issue of sort
[21:19:35] <brtln> but well, it is an issue for you
[21:19:48] <deep42thought> yeah, I'll collect all open "issues" :-)
[21:19:57] <deep42thought> thanks
[21:20:34] <brtln> see, I'm having problems to form a strong opinion on this myself, so I can't even try to convince other people :)
[21:20:42] <brtln> tyzoid: bypass the queue, become TU yourself :P
[21:21:59] <brtln> although at some point I will have to bring up some tiering system and discuss granting you and arm team access to our svn/git as well
[21:22:09] <brtln> too busy lately though
[21:22:30] <deep42thought> I'm not sure I won't do big harm :-/
[21:23:17] <tyzoid> brtln: I wish. I'm not yet ready for the responsibility of TU. Not sure I could put the time in, given my other commitments.
[21:24:54] <deep42thought> brtln: tiering for what?
[21:25:43] <brtln> you overestimate how much it takes to be a TU or dev, I guess
[21:26:23] <brtln> architecture support tiering, x86_64 being formally tier 1, and i686 formally tier 2
[21:27:11] <brtln> but my idea is one thing, another is that our bug tracker sucks and git migration got stuck in limbo apparently, and both would help a lot with this
[21:27:20] <tyzoid> brtln: Perhaps. But I've got ~6 projects on the backburner that I haven't started yet since I haven't had enough time.
[21:27:38] <tyzoid> brtln: I'd like to get to those at some point before next year!
[21:28:13] <brtln> sure, I also have a list like this, it never got shorter so far :P
[21:28:19] <deep42thought> :-)
[21:28:23] <deep42thought> known issue
[21:29:13] <tyzoid> brtln: So what would the time commitment for a TU be?
[21:30:32] <brtln> find a package or set of specific packages to maintain and you are good to go
[21:30:53] <brtln> I pretty much handle just toolchain these days
[21:31:34] <deep42thought> so any package from aur which I like is fine?
[21:31:36] <tyzoid> brtln: And most builds are manually triggered by the maintainer, right?
[21:31:38] <brtln> also as you are handling i686, the vote shouldn't be an issue
[21:33:09] <brtln> deep42thought: I think specific packages "sell" better, this way some people got in by maintaining scientific libraries and I expect the D guy to pass the vote as well
[21:33:23] <brtln> but as I said, handling i686 is already something you can bring up instead
[21:33:46] <brtln> tyzoid: triggered is too big word, you build stuff yourself, although we have a server for building
[21:34:28] <brtln> and at least I have a script that starts the build remotely and just scp's built packages
[21:34:55] <tyzoid> brtln: But do you have to monitor upstream for new versions? Or did you write the script that does that for you?
[21:35:03] <brtln> current svn workflow is rather bad but we will move to git one day…
[21:35:13] <brtln> tyzoid: I use urlwatch
[21:35:19] <brtln> integration with Fedora's anitya in archweb would be nice though
[21:37:25] <tyzoid> brtln: Hmm. I may focus more on infrastructure then.
[21:37:43] <tyzoid> brtln: I'm still working on getting openQA up - just need to get some time to sit down and get it packaged.
[21:38:29] <brtln> replacing Pierre with some CI and openqa is also a great idea 8)
[21:41:27] <tyzoid> also, deep42thought, we should look at the backup infrastructure again at some point.
[21:41:34] <deep42thought> yes
[21:41:39] <tyzoid> not sure if rewbycraft still wanted to take lead of that effort
[21:42:30] <deep42thought> my priority is currently to get all packages built
[21:42:40] <rewbycraft> tyzoid: I'm afraid not right now.
[21:42:53] <rewbycraft> I actually got a job
[21:42:58] <rewbycraft> (which is great)
[21:43:02] <rewbycraft> But also means I have less spare time
[21:43:16] <rewbycraft> I hope you understand
[21:43:22] <deep42thought> yes, sure
[21:44:57] <WarheadsSE> brtln: what 'arm team' where you referring to?
[21:44:58] <tyzoid> rewbycraft: Sure. Just let us know if there's something you can/would like to tackle
[21:45:05] <tyzoid> WarheadsSE: #archlinux-arm
[21:45:16] <rewbycraft> tyzoid: Will let you know.
[21:45:25] <WarheadsSE> Just wanted to check if that was meant for me
[21:45:33] <rewbycraft> I'll make sure the systems I'm providing keep running at the very least.
[21:45:40] <tyzoid> rewbycraft: I assume you're still good to keep buildmaster up?
[21:45:44] <tyzoid> rewbycraft: Ah, was just asking about that.
[21:45:46] <tyzoid> Thanks
[21:45:54] <deep42thought> great
[21:46:36] <rewbycraft> Yeah, sorry guys. But I've got to reevaluate how to spend my time due to also having to work now.
[21:46:49] <tyzoid> rewbycraft: It's no problem; we understand.
[21:46:49] <WarheadsSE> you folks can ping me directly when you have something about ARM brtln tyzoid
[21:47:07] <tyzoid> WarheadsSE: Sounds good. And yeah, if we can get that clarified, that'd be good.
[22:32:42] tyzoid is now known as tyz
[22:40:39] <eschwartz> deep42thought: `makepkg --printsrcinfo` failing to print the group for linux is a makepkg bug, not a core/linux bug. util/pkgbuild.sh provides extract_function_variable, and the attr_regex used does not account for valid bash like `conditional && groups=('base')`
[22:41:11] <deep42thought> hmm, was a 50:50 chance for me
[22:42:43] <deep42thought> is it possible to move the bugreport to the correct location?
[22:42:51] <deep42thought> or should I file a new one?
[22:43:01] <eschwartz> I will move it
[22:43:06] <deep42thought> thank you
[22:46:29] <deep42thought> where do I find the maekpkg source?
[22:46:47] <deep42thought> It is part of pacman, but I seem unable to find it in the pacman git repo
[22:46:48] <eschwartz> It's part of the pacman package :)
[22:47:18] <eschwartz> Actually, your suggested patch is a perfectly valid approach to fixing it for this one package, technically...
[22:47:26] <eschwartz> as a workaround.
[22:47:29] <deep42thought> yes
[22:47:29] <eschwartz> ...
[22:47:56] <deep42thought> but I guess, you prefer to have makepkg understand bash syntax in PKGBUILDs
[22:48:19] <eschwartz> I'll move it anyway, since --printsrcinfo isn't something we care about for repo PKGBUILDs whereas this is definitely something makepkg fails at.
[22:49:11] <deep42thought> ok
[22:49:48] <deep42thought> can you point me to the file you mentioned (util/pkgbuild.sh) - I can't find it in the pacman source tree
[22:52:43] <eschwartz> scripts/libmakepkg/ contains all the *.sh.in files
[22:53:44] <eschwartz> scripts/ is also where /usr/bin/makepkg comes from, along with pacman-key, repo-add, etc.
[22:59:03] <deep42thought> ah, stupid me
[22:59:11] <deep42thought> I forgot the .in ending ...
[23:11:17] -!- deep42thought has quit [Quit: Leaving.]