#archlinux-ports | Logs for 2017-04-24

Back
[00:43:26] -!- shentey has quit [Quit: Konversation terminated!]
[00:57:41] -!- Polichronucci has quit [Ping timeout: 252 seconds]
[00:59:17] -!- Polichronucci has joined #archlinux-ports
[01:39:27] -!- Polichronucci has quit [Ping timeout: 240 seconds]
[01:41:18] -!- Polichronucci has joined #archlinux-ports
[02:05:48] -!- Polichronucci has quit [Ping timeout: 240 seconds]
[02:07:47] -!- Polichronucci has joined #archlinux-ports
[03:57:36] -!- Polichronucci has quit [Quit: ""]
[04:03:53] -!- Polichronucci has joined #archlinux-ports
[05:00:34] -!- bhart has joined #archlinux-ports
[05:28:44] <phillid> What sort of functionality are you looking for in the build slace?
[05:28:46] <phillid> slave*
[05:28:52] <phillid> Just looking at a build list and going at it?
[05:37:21] -!- isacdaavid has joined #archlinux-ports
[05:41:14] <phillid> Oh no, I just saw what you committed to the repo
[05:51:10] <phillid> What will you do to avoid throwing large package builds onto weaker hosts?
[05:59:10] <phillid> s/Oh no/Oh I see/
[06:07:46] * eschwartz preferred the original version lol
[06:21:58] <phillid> I think manually splitting the build list might be better
[06:24:24] -!- bhart has quit [Ping timeout: 260 seconds]
[06:28:17] <phillid> Or have arch32 maintainers adopt packages to build and maintain like archlinux currently does, then leave each maintainer to sort out how they build their bundle of packages
[06:30:07] -!- deep42thought has joined #archlinux-ports
[06:40:40] <deep42thought> phillid: I currently have no plans to keep away heavy package builds from smaller builders. I think, we should first get something running and later we might think about additions (e.g. a 'I do not want'-list on a per host basis or the like)
[06:47:32] <deep42thought> regarding your question what we need at the build-end: additionally to what I started writing, we'd need a fork of archbuild with additionally symlinks and config files for core, opt, ...
[06:48:26] <eschwartz> deep42thought: I see you have gotten it switched to tsort :)
[06:54:06] <deep42thought> eschwartz: yes, its _much_ faster now :-)
[06:54:19] <deep42thought> ... and much cleaner
[07:06:53] <deep42thought> btw: thanks for the hint
[07:12:10] -!- isacdaavid has quit [Ping timeout: 240 seconds]
[07:18:26] <phillid> Not sure why we'd need to fork archbuidl
[07:20:23] <phillid> archbuild*
[07:21:31] <deep42thought> phillid: for the additional configs and symlinks, only
[07:22:39] <phillid> Do you mean like /usr/share/devtools/makepkg-i686.conf ?
[07:23:59] <phillid> That is to say, to keep those files around when arch64 gets rid of them
[07:24:22] <deep42thought> that, and we'd need configs for core, opt, ..., too
[07:28:59] <deep42thought> or do we just use pacman-staging.conf and pacman-testing.conf for that? (I thought we'd skip testing and staging?)
[07:31:07] <deep42thought> ah, and of course, we need to add our own mirror to the configs at some point - this should happen in the packagee archlinux-mirrorlist, though
[07:32:34] <phillid> Yeah, that would have to be one of our custom packages
[07:35:30] <deep42thought> unfortunately, package customizations are not implemented yet :-/
[07:36:36] <phillid> The easiest way is to keep them as a separate repo which has higher precedence than the others
[07:36:51] <phillid> alarm takes this approach with their devtools package, but not their mirrorlist
[07:37:36] <deep42thought> yes, but in case we need some modifications to existing PKGBUILDs, City-busz's proposal of a modification-shell-script is better
[07:38:09] <deep42thought> I think of checking out the original package tree, copying all existent modifications on top of that, and executing all the relevant shell scripts
[07:38:48] <deep42thought> this way, one could simply replace the PKGBUILD (and other files) or put a shell-script in the modification-repo which does more sophisticated tasks
[07:39:06] <phillid> Sounds like it could get messy using shell scripts
[07:39:24] <deep42thought> the tree will be messy, if that's what you mean
[07:40:09] <phillid> That, and the scripts will be harder to maintain than a PKGBUILD
[07:40:26] <deep42thought> one could also skip the shell-scripts for now and only copy the modification-repo over the original ones
[07:41:00] <phillid> I wonder if we could have our own fork of the repo with i686 modifications to the PKGBUILDs etc
[07:41:31] <phillid> Then pull in changes from the arch64 repo
[07:42:15] <deep42thought> this works as long as there are no merging conflicts
[07:42:25] <phillid> Same goes for the shell scripts
[07:42:28] <deep42thought> but in that case we'd have to intervene manually anyway
[07:42:32] <deep42thought> yes
[07:42:38] <phillid> Hmmm
[07:42:58] <deep42thought> actually, no
[07:43:05] <phillid> I like that the shell scripts wouldn't be as "bloaty"
[07:43:09] <phillid> But they would be hard to maintain
[07:43:40] <deep42thought> if we replace a package completely (e.g. mirrorlist), there will be merge conflicts, but the replace-completely-script (or -overwrite) approach would work flawlessly, still
[07:44:11] <phillid> Fair point
[07:44:31] <phillid> Well packages are kept on the arch64 repo on their own branch, aren't they?
[07:44:40] <deep42thought> not sure
[07:44:57] <phillid> Oh no I am mistake
[07:45:02] <phillid> mistaken*
[07:45:33] <deep42thought> there are a lot of branches
[07:46:17] <deep42thought> however, there seem to be no merges ...
[07:46:25] <deep42thought> strange
[07:46:58] <phillid> On my git clone there is only one branch
[07:47:11] <deep42thought> "git branch -a" shows a lot of remote branches
[07:47:42] <phillid> Oh wait
[07:47:48] <phillid> is it git or svn remote?
[07:48:01] <eschwartz> If you need to override a PKGBUILD only partially, just append to the file, later variable/function definitions will replace earlier ones
[07:48:33] <phillid> Oh of course
[07:48:44] <phillid> I wonder how alarm does it
[07:48:49] * phillid looks
[07:49:16] <eschwartz> No idea, but appending changed vars/functions is totally how I'd do it. ;)
[07:50:44] <deep42thought> sound's like a good idea
[07:50:59] <phillid> It would mean we have only partial PKGBUILDs hosted, but that's fine
[07:51:11] <phillid> It'd look funny to the untrained eye is all
[07:51:16] <phillid> I have no objection to that
[07:51:21] <deep42thought> think of it as patches
[07:51:32] <eschwartz> That's the point, why duplicate all the logic and have to maintain the forked PKGBUILD?
[07:51:46] <phillid> Very trye
[07:52:04] <eschwartz> Mind you, I don't know how often you'll need to modify a PKGBUILD for i686...
[07:52:18] <phillid> I have my doubts that it will be often
[07:52:26] <deep42thought> haven't seen one, so far ... (but haven't looked at too many, either)
[07:52:50] <deep42thought> so, guys, I'll have to leave now. cya later!
[07:53:01] <phillid> I imagine most of our changes will be for our own environment changes rather than the CPU architecture itself
[07:53:11] -!- deep42thought has quit [Quit: Leaving.]
[08:52:54] <barthalion> I don't really read backlogs here, but I still fail to see why you didn't take archlinuxarm tool for building
[08:53:11] <barthalion> don't take it as complaining, but looked like perfect fit to me
[09:21:02] <phillid> Neither. I'm not too sure how closely archlinuxarm tracks arch64 though
[09:21:25] <phillid> Additionally, they have a lot of extra functionality with their scope for many different arm arches
[11:09:05] -!- alexandre9099 has quit [Ping timeout: 248 seconds]
[11:10:02] -!- alexandre9099 has joined #archlinux-ports
[11:42:32] -!- alexandre9099 has quit [Quit: ZNC - http://znc.in]
[11:42:48] -!- alexandre9099 has joined #archlinux-ports
[11:44:28] -!- alexandre9099 has quit [Client Quit]
[11:44:33] -!- woffs has quit [Remote host closed the connection]
[11:44:39] -!- alexandre9099 has joined #archlinux-ports
[11:44:50] -!- woffs has joined #archlinux-ports
[11:45:15] -!- woffs has quit [Remote host closed the connection]
[11:45:35] -!- woffs has joined #archlinux-ports
[11:45:42] -!- woffs has quit [Remote host closed the connection]
[11:46:07] -!- alexandre9099 has quit [Client Quit]
[11:46:19] -!- alexandre9099 has joined #archlinux-ports
[11:50:12] -!- woffs has joined #archlinux-ports
[16:36:01] -!- isacdaavid has joined #archlinux-ports
[16:41:48] -!- isacdaavid_ has joined #archlinux-ports
[16:45:43] -!- isacdaavid has quit [Ping timeout: 260 seconds]
[16:53:10] -!- isacdaavid_ has quit [Quit: isacdaavid_]
[16:53:26] -!- isacdaavid_ has joined #archlinux-ports
[17:52:01] -!- isacdaavid has joined #archlinux-ports
[17:55:27] -!- isacdaavid_ has quit [Ping timeout: 240 seconds]
[18:38:53] -!- isacdaavid has quit [Remote host closed the connection]
[18:39:05] -!- isacdaavid has joined #archlinux-ports
[18:44:45] -!- isacdaavid has quit [Quit: isacdaavid]
[19:49:26] -!- deep42thought has joined #archlinux-ports
[21:41:34] -!- bhart has joined #archlinux-ports
[22:38:08] -!- Polichronucci has quit [Ping timeout: 240 seconds]
[22:40:38] -!- Polichronucci has joined #archlinux-ports
[23:00:54] -!- Polichronucci_ has joined #archlinux-ports
[23:01:32] -!- Polichronucci has quit [Ping timeout: 252 seconds]
[23:01:33] Polichronucci_ is now known as Polichronucci
[23:02:19] -!- bhart has quit [Ping timeout: 260 seconds]
[23:05:25] -!- tsnake41_ has joined #archlinux-ports
[23:05:29] -!- tsnake41_ has quit [Client Quit]
[23:49:11] -!- piernov has quit [Ping timeout: 255 seconds]
[23:52:41] -!- deep42thought has quit [Remote host closed the connection]
[23:54:49] -!- piernov has joined #archlinux-ports