#archlinux32 | Logs for 2018-06-27

Back
[00:18:06] <elibrokeit> why not?
[00:20:59] <tyzoid> because devtools doesn't work on i686 hosts
[00:26:48] <elibrokeit> It used to AFAIK, why do you say that?
[01:16:41] -!- isacdaavid has joined #archlinux32
[01:24:22] -!- isacdaavid has quit [Ping timeout: 264 seconds]
[04:21:50] -!- abaumann has joined #archlinux32
[04:21:50] <buildmaster> Hi abaumann!
[04:22:35] <abaumann> deep42thought: my slave had a stale ping-build-master.lock in the workdir. :-)
[04:23:10] <abaumann> *abaumann makes a mental note to grep first for error messages in the source code and then to ask stupid questions..
[04:30:36] -!- abaumann has quit [Quit: leaving]
[06:24:40] -!- deep42thought has joined #archlinux32
[06:24:41] <buildmaster> Hi deep42thought!
[06:25:08] <deep42thought> abaumann: the file is no problem - only not getting a flock on it is.
[06:25:44] <deep42thought> tyzoid: I think, as elibrokeit, that the running-on-i686 would be the least problem for devtools (e.g. should work out of the box)
[06:26:38] <elibrokeit> under the hood it should all be the same logic flows, except for pointing to the right configured repositories and hosts and things
[06:26:58] <deep42thought> but as said before: currently devtools32 is _not_ the original devtools, e.g. it's more/different than current devtools + some configuration and symlinks
[06:27:08] <deep42thought> elibrokeit: mainly, yes
[06:27:10] <elibrokeit> devtools on mainline used to work for i686 just fine, in fact
[06:27:15] <deep42thought> but there are some small but important changes
[06:27:19] <deep42thought> e.g. different cache dir
[06:27:59] <elibrokeit> well, this would be issueful then
[06:28:12] <elibrokeit> which cache dir do you use
[06:28:57] <deep42thought> /var/cache/archbuild32
[06:29:23] <deep42thought> we had massive problems with stale signatures with the old one
[06:30:07] <deep42thought> these are the commits atop devtools: https://ptpb.pw
[06:30:14] <elibrokeit> oh, because you rebuild instead of importing I guess
[06:30:16] <deep42thought> I think, some of them are suprafluid
[06:30:19] <deep42thought> but not all
[06:30:24] <deep42thought> elibrokeit: exactly
[06:31:31] <deep42thought> we also patched it to run namcap in a similar environment for already-built packages
[06:31:43] <deep42thought> so we could more easily compare i686 and x86_64 namcap logs
[06:31:52] * deep42thought wonders if anyone looks at those logs
[06:33:36] <deep42thought> other important point: different mirror layout
[06:36:04] <deep42thought> elibrokeit: what would be the benefit of making devtools32 an extension of devtools?
[06:36:24] <elibrokeit> less rebasing?
[06:38:10] <deep42thought> yeah, ok
[06:38:25] <elibrokeit> archbuild.in: delete old chroot more forcefully ???
[06:38:34] <deep42thought> yeah
[06:38:46] <deep42thought> there are conditions, when the temp directory gets not purged
[06:38:51] <elibrokeit> force correct architecture in pacman ==> pacman should detect this via setarch ==> uname ?
[06:38:54] <deep42thought> and this leads to problems on following builds
[06:39:09] <elibrokeit> what conditions and does x86_64 need this too
[06:39:16] <deep42thought> elibrokeit: some cleanup of these commits should land on my todo list
[06:39:33] <elibrokeit> err, archbuild: add --arch=$CARCH to pacman options ==> this is the one which should be detected
[06:39:55] <deep42thought> and this is the commit which is suprafluid, I guess
[06:40:25] <elibrokeit> what's nice about extensions, is things like "add repo build-support for i686" don't need to conflict at all
[06:40:48] <deep42thought> yes, totally
[06:41:09] <deep42thought> ah, another problem: our build process requires a different ordering of repositories
[06:41:43] <deep42thought> staging, community-staging, testing, community-testing, core, extra, community
[06:42:08] <deep42thought> instead of staging, testing, core, extra, community-staging, community-testing, community
[06:43:19] <elibrokeit> why do you need community-staging to override testing
[06:43:21] <deep42thought> ah, I was wrong: 7cc12bd is probably suprafluid, too - the part that was actually useful and doing something important was upstreamed :-)
[06:43:39] <deep42thought> because pacman orders by pacman.conf, not by version
[06:44:13] <deep42thought> and if a package moves from extra to community, the freshly-built in community-staging/testing needs to superssed the one in extra
[06:44:14] <elibrokeit> I guess, lib/archroot.sh: subvolume_delete_recursive: support arbitrary recursion
[06:45:02] <elibrokeit> Oh right, you mentioned you rebuild when it is db-moved
[06:45:23] <elibrokeit> since you cannot (yet?) detect cross-svnrepo moves
[06:45:54] <elibrokeit> I mentioned looking for the communit2extra tag in the svntogit history
[06:46:21] <deep42thought> yes, I know
[06:48:37] <deep42thought> ah, found the condition: if archbuild fails during the download phase, the directory is not purged
[06:49:04] <deep42thought> https://git.archlinux32.org
[06:49:05] <phrik> Title:remove empty tree if "--verifysource" failed · 9810581c7d - Archlinux32 Gitea (at git.archlinux32.org)
[06:49:11] -!- deep42thought has parted #archlinux32
[06:49:14] -!- deep42thought has joined #archlinux32
[06:49:14] <buildmaster> Hi deep42thought!
[06:50:01] <girls> I think, what we should definitely do is to clean up the changes (e.g. remove what is unnecessary)
[06:50:11] <girls> then rebasing will be simpler
[06:50:30] <girls> but I think, making devtools32 an extension to devtools will not work in the current form
[06:51:16] <deep42thought> it's too early here, I need my coffee ...
[06:56:53] -!- NoobAlice has quit [Quit: Leaving.]
[06:58:24] <elibrokeit> deep42thought: ah, that commit is different
[06:58:36] <elibrokeit> Istr seeing it on arch-projects too
[07:00:32] <deep42thought> I think, I opened a bug report for that one once
[07:00:36] <deep42thought> but I'm not sure
[07:00:43] <deep42thought> and I'm afk for breakfast now
[07:12:33] <buildmaster> cifs-utils is broken (says nlopc46).
[07:27:40] -!- deep42thought has quit [Quit: Leaving.]
[07:47:29] -!- vboxler has joined #archlinux32
[07:49:50] <vboxler> https://pastebin.com
[07:49:51] <phrik> Title:stag, 26. Juni 2018] [08:49:10 CEST] <deep42thought> it's there already [Dienst - Pastebin.com (at pastebin.com)
[07:50:08] <vboxler> Hello
[07:50:48] <vboxler> deep42thouthg, what did you mean?
[07:53:22] -!- oaken-source has joined #archlinux32
[08:30:32] -!- deep42thought has joined #archlinux32
[08:30:32] <buildmaster> Hi deep42thought!
[08:31:28] <deep42thought> vboxler: I meant, that the maintainer of the aur package knows about this issue - which is also present on x86_64
[08:40:31] -!- vboxler has quit [Quit: Konversation terminated!]
[09:00:37] -!- eduardoeae has quit [Ping timeout: 245 seconds]
[09:56:17] -!- oaken-source has quit [Ping timeout: 248 seconds]
[10:02:53] <buildmaster> qcad is broken (says rechenknecht).
[10:27:55] -!- oaken-source has joined #archlinux32
[10:36:50] -!- oaken-source has quit [Remote host closed the connection]
[10:37:40] -!- oaken-source has joined #archlinux32
[10:57:32] -!- oaken-source has quit [Remote host closed the connection]
[11:16:26] -!- oaken-source has joined #archlinux32
[11:57:12] -!- eduardoeae has joined #archlinux32
[12:40:58] -!- lexxx82 has joined #archlinux32
[13:33:56] <buildmaster> ktorrent is broken (says nlopc43).
[14:29:46] -!- sahasatvik__ has quit [Ping timeout: 264 seconds]
[14:32:07] -!- sahasatvik__ has joined #archlinux32
[14:37:40] <buildmaster> llpp is broken (says rechenknecht).
[14:53:26] <tyzoid> deep42thought: Does the homepage have PHP?
[14:58:10] <deep42thought> not yet
[14:58:24] <deep42thought> I asket Polichronucci, but he hasn't got to it yet
[14:58:39] <deep42thought> what feature would you like to put there?
[15:15:09] <tyzoid> I was thinking of loading the rss feed from the news page
[15:15:17] <tyzoid> so we could have news on the homepage
[15:16:42] <deep42thought> sounds nice
[15:16:49] <bill-auger> i thought that was standard archweb
[15:16:56] <tyzoid> We're not using standard archweb
[15:16:59] <deep42thought> we don't run archweb
[15:17:05] * deep42thought is too slow
[15:17:06] <tyzoid> since everything is scattered between several servers
[15:17:37] <tyzoid> and even the stuff I run are on different virtual servers
[15:18:05] <tyzoid> news/forum, bugtracker, packagedb, homepage are all different servers.
[15:18:16] <tyzoid> SSO/auth is its own too
[15:18:29] <deep42thought> abaumann: we're having other qt5-* include issues (like with obs-studio) for ktorrent, but there I could reproduce the issue with x86_64 and simply reported upstream :-)
[15:18:31] <tyzoid> but that's still indev
[15:18:46] <tyzoid> deep42thought: Ahh, feels like the old days
[15:19:47] <tyzoid> deep42thought: Any opposition to a javascript API to grab newsitems, with an iframe in a noscript as a backup?
[15:20:19] <deep42thought> elibrokeit: if you want to improve your llpp PKGBUILD, you can add libglvnd as makedepends - it does not build with other libgl providers
[15:20:43] <deep42thought> tyzoid: I'm usually against js :-)
[15:20:51] <tyzoid> hence the iframe noscript fallback
[15:21:03] <deep42thought> yeah, ok
[15:21:11] <deep42thought> I'll heavily test the no-js part :-D
[15:21:11] <tyzoid> I'm usually against iframes, so I think this is a decent comprimise
[15:21:53] <tyzoid> probably won't get to it today, since I'm heading off for vacation in a few days
[15:22:21] <deep42thought> np
[15:22:24] <tyzoid> Speaking of, I'm going to make the ISO tomorrow, so I don't have to deal with remotely creating the iso on vacation
[15:22:33] <deep42thought> ok
[15:22:41] <buildmaster> calligra is broken (says rechenknecht).
[15:23:23] <deep42thought> .. also qt5
[15:26:32] <deep42thought> the iso coming a bit earlier is also not bad, because it will have a working keyring :-)
[15:26:54] <deep42thought> it will only break the mirror-checker ...
[15:26:57] <deep42thought> hmmm
[15:30:45] -!- lexxx82 has quit [Ping timeout: 260 seconds]
[15:32:16] <tyzoid> deep42thought: Btw, just set up `rsync://archive.archlinux32.org/archlinux32-archive`
[15:32:47] <tyzoid> if it's wanted for anything
[15:33:16] <deep42thought> doesn't hurt, at least :-)
[15:33:30] <deep42thought> how big is the archive currently?
[15:34:18] <tyzoid> running du, one sec
[15:35:55] <tyzoid> 119GiB of packages and 220MiB of signatures
[15:36:07] <tyzoid> It's currently counting the size of all the folders and their symlinks
[15:36:20] <tyzoid> but IIRC it comes out to be ~9GiB of symlinks
[15:36:31] <deep42thought> whoah O.o
[15:36:52] <tyzoid> two symlinks for each package every day?
[15:37:08] <deep42thought> ah, right
[15:37:12] <deep42thought> that's a lot of inodes ...
[15:37:24] <deep42thought> I think I'll set up a (offline) backup of the archive somewhere here ...
[15:37:36] <deep42thought> jsut the important, irreplaceable stuff
[15:37:40] <tyzoid> ok. Rsyncing those symlinks is probably going to take the longest time xD
[15:37:50] <tyzoid> btw, you can archive just pool and sigs if you want
[15:37:51] <deep42thought> e.g. pool/ (and probably sigs/)
[15:37:56] <tyzoid> yeah
[15:38:03] <tyzoid> and potentially web/isos
[15:38:03] * deep42thought is too slow again
[15:39:04] * tyzoid is sitting here drinking caffeine listening to some new (to him) music.
[15:39:56] <tyzoid> stopped du before it reported the web dir
[15:40:07] <deep42thought> 21922457851
[15:40:09] <tyzoid> but isos is 20 gigs
[15:40:26] <deep42thought> echo $[$(rsync rsync://archive.archlinux32.org | awk '{print $2}' | tr '\n' '+' | tr -d ',')0]
[15:40:27] <deep42thought> :-D
[15:40:47] <buildmaster> llpp is broken (says rechenknecht).
[15:40:55] <deep42thought> grml
[15:41:03] <tyzoid> nice
[15:41:19] <tyzoid> I usually pipe that through bc
[15:42:23] <tyzoid> bc -l <<< "scale=2;($(rsync ... | cut -f2 -d' ' | tr '\n' '+')0)/(2^30)"
[15:43:08] <deep42thought> does cut ommit empty columns?
[15:43:21] <tyzoid> yeah, that doesn't work :/
[15:43:45] <tyzoid> anyway
[15:44:04] <tyzoid> I like BC since I can divide by 2^<whatever> and get MiB, GiB, etc.
[15:44:13] <deep42thought> yeah
[15:46:56] <deep42thought> when you create the iso, can you name it 2018.07.01 anyway?
[15:50:41] <tyzoid> That's the plan
[15:50:47] <deep42thought> perfect :-)
[15:57:04] <deep42thought> cu later, I gotta go
[15:57:07] -!- deep42thought has quit [Quit: Leaving.]
[16:06:05] -!- oaken-source has quit [Ping timeout: 245 seconds]
[16:06:52] <tyzoid> deep42thought: Just saw your DMs. When I moved the bbs over to the container, I added you as a user just like the other boxes, so you should have a deep42thought account
[16:07:41] <tyzoid> just double checked, and that is the case
[16:14:33] -!- rcf has quit [Quit: WeeChat 2.1]
[16:27:58] -!- rcf has joined #archlinux32
[16:37:46] -!- abaumann has joined #archlinux32
[16:37:47] <buildmaster> Hi abaumann!
[16:38:09] <abaumann> deep42thought: ok, so I remove the patch for obs-studio in our PKGBUILD..
[16:40:14] -!- abaumann has quit [Client Quit]
[16:53:00] <elibrokeit> deep42thought: omg really?
[16:55:59] <elibrokeit> Hmm, wait, the only other provider is Nvidia-340xx-utils
[16:56:07] <elibrokeit> Not sure this counts
[17:09:35] <buildmaster> obs-studio is broken (says nlopc46).
[17:21:02] -!- lexxx82 has joined #archlinux32
[17:46:57] -!- oaken-source has joined #archlinux32
[17:56:56] -!- deep42thought has joined #archlinux32
[17:56:56] <buildmaster> Hi deep42thought!
[18:03:07] <girls> tyzoid: ah, right - I got fooled by some old entries in my /etc/fstab
[18:03:18] <tyzoid> :)
[18:29:43] <elibrokeit> deep42thought: is this just an issue with Nvidia?
[19:36:20] <deep42thought> elibrokeit: it looks, like it works with libglvnd and does not work with nvidia-340xx-utils
[19:36:25] -!- deep42thought has parted #archlinux32
[19:36:28] -!- deep42thought has joined #archlinux32
[19:36:28] <buildmaster> Hi deep42thought!
[19:37:17] <deep42thought> and because nvidia-340xx-utils is in testing on archlinux32, it superseeds libglvnd
[19:37:50] <deep42thought> I have seen this issue multiple times in the past for libgl-dependent packages
[20:08:39] -!- isacdaavid has joined #archlinux32
[20:14:39] <deep42thought> hmm, we could need a --build-with-trunk switch for build-packages
[20:16:25] <deep42thought> abaumann: apparently the correct fix is to do exactly what you did: add the missing include :-D
[20:25:21] -!- sahasatvik__ has quit [Quit: ZNC 1.8.x-git-99-eec5c5a - https://znc.in]
[20:28:31] -!- sahasatvik__ has joined #archlinux32
[20:43:14] -!- abaumann has joined #archlinux32
[20:43:14] <buildmaster> Hi abaumann!
[20:43:27] <deep42thought> Hi abaumann!
[20:43:37] <abaumann> deep42thought: yeah. but the question is, why is software published upstream with so simple missing includes?
[20:43:40] <abaumann> hi.
[20:43:41] <abaumann> :-)
[20:43:56] <deep42thought> this worked with qt5-5.1.0
[20:44:07] <deep42thought> e.g. qt5-5.1.1 broke it
[20:44:21] <deep42thought> I'm putting qt-5.1.0 into [build-support] for now
[20:44:28] <abaumann> aha. so Qt cleaned unnecessary includes which people above relied on (implicitly).
[20:44:35] <deep42thought> 5.11.x
[20:44:38] <deep42thought> not 5.1.x
[20:44:39] <deep42thought> sry
[20:44:50] <deep42thought> yes, looks so
[20:44:57] <abaumann> good idea.
[20:45:32] <deep42thought> I have to find a bug in copy-to-build-support first, though :-)
[20:46:01] <abaumann> oh :-)
[20:48:09] <deep42thought> found it :-)(
[20:48:14] <deep42thought> s/($//
[20:49:00] <deep42thought> https://packages.archlinux32.org :-)
[20:49:01] <phrik> Title:Arch Linux 32 - qt5-base 5.11.0-1.0 (i686) (at packages.archlinux32.org)
[20:49:11] <deep42thought> let's see what incidentally can be built now :-D
[20:51:43] <abaumann> the links 'archlinux32 sourfile files and changes' give me a 404 there :-)
[20:52:14] <deep42thought> upsi
[20:52:30] <deep42thought> this was introduced by a fix from me ...
[20:52:33] <deep42thought> gimme a sec
[20:53:38] <deep42thought> 0cb52e4d to be precise ...
[20:53:43] <abaumann> no hurry :-)
[20:55:38] <deep42thought> ah, that commit just replaced a 404 link with another 404 one :-D
[20:55:59] <deep42thought> hrmm
[20:56:08] <deep42thought> where should that link lead for build-support packages??
[20:56:27] <deep42thought> we don't have build-support specific modifications
[20:56:45] <deep42thought> the packages in build-support were previously in some other repo and got copied over from there
[20:59:34] <abaumann> if( build-support) then remove link?
[20:59:48] <deep42thought> would be incosistent
[21:00:06] <deep42thought> I just thought about a mysql query which would find the right repo
[21:10:48] <deep42thought> should be corrected now
[21:11:28] <abaumann> yes. thanks. :-)
[21:22:32] <buildmaster> firefox-developer-edition is broken (says nlopc46).
[21:23:32] <deep42thought> hmm, calligra fails to build with qt5-1.11.0-1.0 with the same error :-/
[21:24:12] <deep42thought> maybe it was a different qt5-* package dropping the include :-/
[21:27:14] <abaumann> should be in qt5 base?
[21:27:22] <deep42thought> I thought so, too
[21:27:25] <abaumann> mmh.
[21:37:16] <deep42thought> bed-time
[21:37:19] <deep42thought> good night!
[21:37:28] -!- deep42thought has quit [Quit: Leaving.]
[21:39:04] <abaumann> yeah. cu. :-)
[21:39:05] -!- abaumann has quit [Quit: leaving]
[22:22:14] <buildmaster> mame is broken (says eurobuild3).
[22:47:40] <buildmaster> firefox is broken (says rechenknecht).
[23:06:51] <buildmaster> kube is broken (says rechenknecht).
[23:08:04] <buildmaster> wiznote is broken (says eurobuild3).
[23:12:01] <buildmaster> qt5-webview is broken (says eurobuild3).
[23:14:22] -!- belanthor has joined #archlinux32
[23:26:55] <buildmaster> psi is broken (says buildknecht).