#archlinux-ports | Logs for 2017-06-29

[00:00:00] -!- deep42thought has quit [Quit: Leaving.]
[00:24:16] -!- tyzoid has joined #archlinux-ports
[00:24:20] <tyzoid> back
[03:51:45] -!- isacdaavid has quit [Quit: isacdaavid]
[04:32:36] -!- fhdrin has joined #archlinux-ports
[05:01:04] -!- tyler has joined #archlinux-ports
[05:01:27] tyler is now known as Guest13035
[05:01:35] -!- tyzoid has quit [Read error: Connection reset by peer]
[05:01:59] Guest13035 is now known as tyzoid
[06:05:17] -!- isacdaavid has joined #archlinux-ports
[06:07:15] tyzoid is now known as tyzoid|afk
[06:08:17] <fhdrin> what does the arch32 dual-boot mean? i thought it was only for i686
[06:19:29] tyzoid|afk is now known as tyzoid
[06:20:24] <tyzoid> fhdrin: The arch32 dual iso is an arch livecd with both i686 and x86_64 kernels
[06:20:44] <tyzoid> It's a continuation of the old -dual.iso files.
[06:21:05] <fhdrin> ah ok thanks
[06:21:11] <tyzoid> no problem
[06:21:19] <tyzoid> fhdrin: Are you running a 32bit system?
[06:21:22] <fhdrin> slightly confusing when they say arch32
[06:21:52] <fhdrin> just checking it out in virtualbox, but i do have an old core2duo that would be a good canidate for this
[06:21:52] <tyzoid> fhdrin: See https://www.archlinux.org
[06:21:54] <phrik> Title: Arch Linux - News: Phasing out i686 support (at www.archlinux.org)
[06:22:08] <tyzoid> arch32 is the community project that's picking up i686 support
[06:22:14] <tyzoid> so the isos are arch32 isos
[06:33:01] <tyzoid> So I'm looking at openqa
[06:33:28] <tyzoid> but the software is complicated to get even the web client built
[06:33:39] <tyzoid> I'm going to go to bed now, as I'm getting quite tired
[06:33:51] <tyzoid> but I'm working on writing a packagebuild file for openqa
[06:34:02] <tyzoid> so that we can get initial testing started
[06:34:15] <tyzoid> anyway, I'm off for now
[06:34:30] -!- tyzoid has quit [Quit: WeeChat 1.9]
[06:34:56] -!- deep42thought has joined #archlinux-ports
[06:39:24] <deep42thought> missed by 26 seconds ...
[06:57:46] -!- Faalagorn has quit [Remote host closed the connection]
[07:13:13] -!- deep42thought has quit [Remote host closed the connection]
[07:44:36] -!- eschwartz has quit [Ping timeout: 260 seconds]
[07:45:12] -!- eschwartz has joined #archlinux-ports
[07:47:25] -!- dmakeyev_ has joined #archlinux-ports
[07:59:40] -!- dmakeyev_ has quit [Remote host closed the connection]
[08:02:19] -!- isacdaavid has quit [Quit: isacdaavid]
[08:10:25] -!- wyre has quit [Read error: Connection reset by peer]
[08:11:47] -!- deep42thought has joined #archlinux-ports
[08:44:33] -!- fhdrin has quit [Quit: Leaving]
[11:23:37] -!- AndrevS has joined #archlinux-ports
[11:37:35] -!- Cthulu201 has quit [Quit: WeeChat 1.8]
[11:58:10] -!- Cthulu201 has joined #archlinux-ports
[12:30:50] <deep42thought> Is anyone in here familiar with haskell? A lot of haskel-* builds fail due to "missing package base-"
[12:34:03] <woffs> base should be included in ghc?
[12:34:41] <deep42thought> I can't find any reference to "haskell-base" via pacman, but I have absolutely no experience regarding haskell
[12:34:48] <deep42thought> but sounds right
[12:37:42] <woffs> base-4.9 should be part of ghc-8.x.y
[12:38:14] <woffs> is
[12:38:33] <woffs> it is just /usr/lib/ghc/base- or something like that
[12:41:44] <woffs> current ghc-8.0.2 contains /usr/lib/ghc-8.0.2/base- and not
[13:16:05] -!- tyzoid has joined #archlinux-ports
[13:16:11] <tyzoid> deep42thought: got your email
[13:16:14] <tyzoid> and I saw the logs
[13:16:20] <tyzoid> 26 seconds :/
[13:36:49] <tyzoid> be back in about an hour
[13:36:51] -!- tyzoid has quit [Quit: WeeChat 1.9]
[13:42:58] <deep42thought> woffs: hmm, it is - also in our "community-testing", some one is not using the right package (or something else is broken :-/)
[14:04:23] <woffs> deep42thought, can you send me some logfiles?
[14:04:41] <deep42thought> they're online: buildmaster.archlinux32.org
[14:05:19] <deep42thought> e.g. https://buildmaster.archlinux32.org
[14:05:26] -!- Faalagorn has joined #archlinux-ports
[14:08:14] <deep42thought> The error might be trivial, but I have no plan about haskell at all.
[14:11:35] <woffs> where does haskell-ansi-wl-pprint package come from?
[14:12:21] <deep42thought> was built by our build systemm
[14:12:37] <deep42thought> 3 days ago
[14:13:02] <deep42thought> https://mirror.archlinux32.org
[14:13:09] <woffs> build log?
[14:13:19] <deep42thought> not saved :-/
[14:14:34] <woffs> was built with ghc-8.0.1, meanwhile ghc-8.0.2 popped in.
[14:14:51] <woffs> you need to rebuild all haskell-* again with ghc-8.0.2
[14:14:59] <deep42thought> ah, ok
[14:15:02] <deep42thought> thanks
[14:15:10] <woffs> :-)
[14:15:19] <deep42thought> the right ghc version is not mentioned in the dependencies?
[14:15:23] <woffs> timing issues in a rolling distribution
[14:16:32] <woffs> in most cases haskell packages depend just on other haskell packages, and in this case base- hides in ghc-8.0.1 while base- hides in ghc-8.0.2, I guess
[14:17:14] <deep42thought> but the clean (but tedious) solution would be to include depends=(ghc=8.0.2) in the respective package, right?
[14:17:15] <woffs> the dependency inside the raw package is just base >3 <5 or something like that ;)
[14:17:23] <deep42thought> hmm
[14:17:43] <deep42thought> That sounds like it will break the build system when the slightes desync happens ...
[14:18:07] <woffs> I'm not sure if depends=(ghc=8.0.2) is a good solution
[14:18:36] <deep42thought> it'd be tedious to update, but besides that I don't see any downside
[14:20:09] <woffs> I would try to be as clos as possible to the mainline
[14:20:16] <deep42thought> I do :-/
[14:20:30] <woffs> and ghc updates are not so often
[14:20:52] <woffs> just be sure to rebuild everything just when a new ghc comes in
[14:21:06] <deep42thought> at least, I now know what you-compiled-against-a-wrong-haskell-version-errors look like :-)
[14:21:07] <woffs> (or a new base)
[14:21:20] <deep42thought> same with python and perl ...
[14:21:26] <deep42thought> am I missing anything else?
[14:21:44] <woffs> I thinnk ghc version does not matter. the problem was the base lib.
[14:21:50] <woffs> ruby?
[14:21:55] <woffs> go?
[14:22:12] <deep42thought> oh no, too much to remember all of them
[14:29:12] -!- eschwartz has quit [Ping timeout: 260 seconds]
[14:29:18] -!- tyzoid has joined #archlinux-ports
[14:29:22] -!- guys has quit [Ping timeout: 240 seconds]
[14:50:26] -!- guys has joined #archlinux-ports
[14:52:24] <fsckd> ghc and Haskell packages are not friendly to distribution packagers. :/
[14:54:24] <woffs> web search for "cabal dependency hell"
[14:58:08] -!- eschwartz has joined #archlinux-ports
[15:39:40] <jelle> also ghc needs some install order :-)
[15:39:48] <jelle> well haskell-*
[15:40:14] <deep42thought> I was messing around with build order of packages lately
[15:40:34] <deep42thought> I'm now settled for "satisfy build and runtime dependencies"
[15:40:59] <deep42thought> which I had first, but which generates _a_lot_ of dependency cycles if you try to build all packages (what I did in the beginning)
[15:46:16] <deep42thought> tyzoid: Did you have a chance to look into running a build slave yet?
[15:46:59] <tyzoid> I haven't yet
[15:47:06] <tyzoid> I was busy trying to get a package for openqa
[15:47:26] <deep42thought> ah, right, that's more important currently
[15:47:33] <tyzoid> my head started hurting, and I got tired
[15:47:52] <deep42thought> what I've read so far looks promising
[15:47:57] <tyzoid> turns out there's only so many nights you can get 4hrs of sleep in a row before your performance drops :)
[15:48:02] <deep42thought> but like an enormous effort
[15:48:04] <tyzoid> deep42thought: I agree too
[15:48:14] <deep42thought> :-)
[15:48:14] <tyzoid> deep42thought: The web package shouldn't be too difficult
[15:48:27] <deep42thought> "shouldn't" or "isn't"?
[15:48:27] <tyzoid> but separating the codebase out into individual packages doesn't look like it'll be as easy
[15:48:35] <tyzoid> most definately shouldn't
[15:48:41] <deep42thought> :-)
[15:49:09] <tyzoid> deep42thought: Any objection to only making an x86_64 package from openqa?
[15:49:19] <deep42thought> no, that's perfect
[15:49:25] <tyzoid> I don't think I'll be running it on any 32bit machines :)
[15:49:32] <deep42thought> I also only test the devtools32 for x86_64
[15:49:46] <tyzoid> sounds good
[15:49:55] <tyzoid> I'll be heading off in a few minutes
[15:50:03] <tyzoid> I've got a 2hr meeting coming up
[15:50:06] <deep42thought> I was thinking if we want to put up a archlinux32 repository with x86_64 packages related to archlinux32
[15:50:26] <tyzoid> deep42thought: That's not a bad idea
[15:50:45] <deep42thought> but I would keep it as manual as possible
[15:51:16] <deep42thought> Have fun in your meeting!
[15:52:04] <tyzoid> thanks
[16:15:00] -!- deep42thought has quit [Quit: Leaving.]
[17:33:13] -!- deep42thought has joined #archlinux-ports
[18:28:18] -!- AndrevS has quit [Remote host closed the connection]
[19:54:49] -!- fsckd has quit [Quit: WeeChat 1.9]
[19:56:00] -!- fsckd has joined #archlinux-ports
[20:14:27] -!- WarheadsSE has quit [Quit: WeeChat 1.6]
[20:15:10] -!- WarheadsSE has joined #archlinux-ports
[22:22:54] -!- isacdaavid has joined #archlinux-ports
[23:15:56] -!- isacdaavid has quit [Read error: Connection reset by peer]
[23:16:12] -!- isacdaavid has joined #archlinux-ports
[23:33:11] -!- tyzoid has quit [Disconnected by services]
[23:33:21] -!- tyzoid has joined #archlinux-ports
[23:33:37] -!- tyler1 has joined #archlinux-ports
[23:35:09] <deep42thought> wb tyzoid
[23:39:30] -!- titus_livius has joined #archlinux-ports
[23:39:42] <tyzoid> ${DAYJOB} pays the bills/
[23:39:59] <tyzoid> anyway, deep42thought: I'm thinking of creating separate packages from the one large package they provide
[23:40:07] <tyzoid> like openqa-web, openqa-*
[23:40:15] <tyzoid> to separate out the services
[23:40:15] <deep42thought> what's the benefit of that?
[23:40:20] <deep42thought> ah, ok
[23:40:25] <tyzoid> because the web package is just the master
[23:40:37] <tyzoid> so we can have the web interface run on a different server than the build/test slaves
[23:40:51] <deep42thought> yh, you mean "two packages" - that sound's reasonable
[23:40:59] <tyzoid> I'm thinking three
[23:40:59] <deep42thought> don't they even have distinct code bases?
[23:41:02] <tyzoid> nope
[23:41:04] <tyzoid> not afaik
[23:41:46] <tyzoid> deep42thought: They have some other sources
[23:41:53] <tyzoid> openqa is just the web interface, it looks like
[23:46:11] <deep42thought> Which three packages would this be? (just curious)
[23:46:25] <deep42thought> openqa, one for the worker and on for os-autoinst?
[23:47:38] <tyzoid> yeah
[23:47:45] <tyzoid> I'm working on openqa first
[23:47:52] <tyzoid> since I'm most familiar with web infrastructure
[23:47:53] <deep42thought> ok
[23:53:25] -!- tyzoid has quit [Ping timeout: 276 seconds]
[23:55:10] -!- tyzoid has joined #archlinux-ports
[23:57:37] -!- deep42thought has quit [Quit: Leaving.]