#archlinux32 | Logs for 2019-08-06
Back
[01:40:19] -!- MrBIOS has quit [Read error: Connection reset by peer]
[01:57:11] -!- MrBIOS has joined #archlinux32
[02:20:35] -!- MrBIOS has quit [Quit: MrBIOS]
[02:33:22] -!- bill-auger has quit [Ping timeout: 245 seconds]
[02:38:20] -!- bill-auger has joined #archlinux32
[05:22:45] -!- renodr_ has joined #archlinux32
[05:24:29] <renodr_> Hi folks. At Linux From Scratch, we have a report in Trac regarding 32-bit compilation for Firefox. One of our users, tnut, suggested adding ac_add_options --disable-debug-symbols to our mozconfig to bypass the OOM error from rust when compiling gkrust. I don't know if that will help you guys at all, but I wanted to at least drop it here for y'all!
[05:32:47] -!- renodr_ has quit [Remote host closed the connection]
[05:55:18] -!- buildmaster has quit [Remote host closed the connection]
[05:56:45] -!- buildmaster has joined #archlinux32
[05:56:45] <buildmaster> !rq buildmaster
[05:56:46] <phrik> buildmaster: <buildmaster> I might be insane, but never confused ... ;-)
[06:04:28] -!- titus_livius has joined #archlinux32
[06:08:32] * buildmaster resumes sanity.
[07:02:34] -!- thePiGrepper has quit [Ping timeout: 246 seconds]
[07:46:39] -!- nit-picker has joined #archlinux32
[07:48:34] -!- deep42thought has joined #archlinux32
[07:48:35] <buildmaster> Hi deep42thought!
[07:48:35] <buildmaster> !rq deep42thought
[07:48:36] <phrik> buildmaster: <deep42thought> because you're a pessimist with no hope - or you're a realist afraid of change ... pick your evil
[07:50:12] -!- nit-picker has quit [Remote host closed the connection]
[08:08:03] -!- davor has quit [Ping timeout: 248 seconds]
[08:09:29] -!- davor has joined #archlinux32
[08:15:55] <buildmaster> i686/libraw is broken (says rechenknecht): https://archlinux32.org
[08:44:26] <deep42thought> blacklisting simply takes too long ...
[08:44:52] <deep42thought> needing 30 Minutes to determine which packages should be blacklisted and which should be kept is not acceptable
[08:45:18] <deep42thought> especially when you consider, that this is run each time we pull updates from upstream
[09:06:10] <deep42thought> I propose, we put all dependent packages on the blacklist directly, too
[09:06:46] <deep42thought> to keep the blacklist readable / managable, we could change to some directory layout
[09:07:20] <deep42thought> suggestion: blacklist/$architecture/$package are files which blacklist $package for $architecture and the content of the file is the reason for doing so
[09:07:42] <deep42thought> however, I'm open to suggestions for a better layout
[09:07:57] <deep42thought> (e.g. one, where we can easily group packages - like haskell-* for i486)
[09:17:17] <buildmaster> i686/shellharden is broken (says nlopc46): https://archlinux32.org
[09:22:31] <buildmaster> i486/scons are broken (says nlopc46-i486bs1): https://archlinux32.org
[09:25:42] -!- infides has joined #archlinux32
[09:27:13] <buildmaster> i486/aarch64-linux-gnu-glibc is broken (says eurobuild6-7-i486): https://archlinux32.org
[09:36:35] -!- bill-auger has quit [Ping timeout: 248 seconds]
[09:37:00] -!- bill-auger has joined #archlinux32
[09:45:46] <buildmaster> i686/vault is broken (says rechenknecht): https://archlinux32.org
[09:46:45] <buildmaster> pentium4/virtualbox-modules-arch is broken (says eurobuild6-5): https://archlinux32.org
[09:48:00] <buildmaster> pentium4/vault is broken (says buildknecht): https://archlinux32.org
[09:49:09] <buildmaster> i686/virtualbox-modules-arch is broken (says eurobuild3): https://archlinux32.org
[09:51:10] -!- buildmaster has quit [Remote host closed the connection]
[09:51:22] -!- buildmaster has joined #archlinux32
[09:51:22] <buildmaster> !rq buildmaster
[09:51:23] <phrik> buildmaster: <buildmaster> I might be insane, but never confused ... ;-)
[09:53:56] -!- infides has quit [Ping timeout: 268 seconds]
[10:16:13] -!- eschwartz has quit [Ping timeout: 250 seconds]
[10:34:42] -!- eschwartz has joined #archlinux32
[10:44:05] -!- infides has joined #archlinux32
[11:41:48] -!- infides has quit [Quit: leaving]
[12:07:46] -!- infides has joined #archlinux32
[12:31:37] -!- thePiGrepper has joined #archlinux32
[13:27:58] <buildmaster> i686/nvidia-390xx is broken (says rechenknecht): https://archlinux32.org
[13:34:51] <buildmaster> pentium4/broadcom-wl is broken (says eurobuild6-6): https://archlinux32.org
[13:39:13] <buildmaster> pentium4/nvidia-390xx is broken (says buildknecht): https://archlinux32.org
[14:04:12] -!- deep42thought has quit [Ping timeout: 245 seconds]
[14:34:38] -!- thePiGrepper has quit [Ping timeout: 245 seconds]
[14:35:45] -!- thePiGrepper has joined #archlinux32
[14:51:27] -!- slacka123 has joined #archlinux32
[14:54:26] -!- nit-picker has joined #archlinux32
[14:55:17] <buildmaster> i486/ffmpeg is broken (says eurobuild6-7-i486): https://archlinux32.org
[14:55:48] -!- deep42thought has joined #archlinux32
[14:55:48] <buildmaster> Hi deep42thought!
[14:55:48] <buildmaster> !rq deep42thought
[14:55:49] <phrik> buildmaster: <deep42thought> so you're a garbage collector?
[15:10:04] -!- nit-picker has quit [Remote host closed the connection]
[15:12:26] -!- nit-picker has joined #archlinux32
[15:17:18] -!- nit-picker has quit [Remote host closed the connection]
[15:18:04] -!- nit-picker has joined #archlinux32
[15:56:00] -!- deep42thought has quit [Quit: Leaving.]
[16:22:12] -!- abaumann has joined #archlinux32
[16:22:12] <buildmaster> Hi abaumann!
[16:22:12] <buildmaster> !rq abaumann
[16:22:13] <phrik> buildmaster: <abaumann> YASMSBOP: yet another silly make system based on python
[16:32:06] -!- infides has quit [Ping timeout: 258 seconds]
[16:32:36] <abaumann> elibrokeit: yep, sorry. configure indeed belong into build()
[16:35:18] <abaumann> deep42thought: I'm fine with the new blacklist format :-)
[16:35:39] <abaumann> deep42thought: on the time_t 32 to 64 bit migration, I'm not sure what to think (yet).
[16:36:07] <abaumann> I have to read and think more.. :-)
[16:36:08] -!- abaumann has quit [Quit: leaving]
[17:08:43] -!- MrBIOS has joined #archlinux32
[18:17:13] -!- tirsek has quit [Ping timeout: 250 seconds]
[18:22:58] -!- infides has joined #archlinux32
[19:12:22] -!- tirsek has joined #archlinux32
[19:48:00] -!- isacdaavid has joined #archlinux32
[20:22:19] -!- abaumann has joined #archlinux32
[20:22:20] <buildmaster> Hi abaumann!
[20:22:20] <buildmaster> !rq abaumann
[20:22:21] <phrik> buildmaster: <abaumann> "anywhere is anywhere for all values of anywhere"
[20:22:23] -!- deep42thought has joined #archlinux32
[20:22:24] <buildmaster> Hi deep42thought!
[20:22:24] <buildmaster> !rq deep42thought
[20:22:25] <phrik> buildmaster: <deep42thought> rcf you just detuned and are now receiving a different program - please execute /join #debian
[20:22:42] <abaumann> renodr_: thanks for dropping the hint about firefox, will try as soon rust runs again..
[20:22:46] <abaumann> hi deep42thought
[20:22:50] <deep42thought> Hi abaumann!
[20:23:17] <abaumann> how are the blacklists? :-)
[20:23:53] <deep42thought> it's now a directory
[20:23:58] <deep42thought> all dependency logic is remove
[20:23:59] <deep42thought> d
[20:24:09] <deep42thought> so we need to blacklist all packages explicitely now
[20:24:16] <abaumann> you see a positive effect?
[20:24:31] <deep42thought> blacklist/$arch/$arbitrary-identifier/$pkgbase contains descriptions why it is blacklisted
[20:24:42] <deep42thought> sure, the 30min query is no longer executed :-)
[20:24:53] <abaumann> ah. that's cool :-)
[20:25:05] <deep42thought> "CALL `blacklist_packages`"
[20:25:15] <abaumann> I don't see a i686 in the blacklist directory
[20:25:25] <deep42thought> the architecture logic is still in place
[20:25:33] <abaumann> ah. ok
[20:25:34] <deep42thought> only the dependency-tracking logic is removed
[20:25:49] <deep42thought> e.g. blacklisting ghc will not blacklist (almost) all of haskell-*
[20:25:53] <abaumann> btw. where do you keep the create sql index xxx scripts?
[20:26:00] <deep42thought> but blacklisting in i686 will blacklist in i686 and i486, too
[20:26:00] <abaumann> I fiddled with a combined index on ssh_log
[20:26:15] <deep42thought> create index?
[20:26:17] <deep42thought> errr
[20:26:24] <deep42thought> there is a database dump in misc
[20:26:26] <abaumann> so the dump is the reference :-)
[20:26:32] <abaumann> ah
[20:26:39] <deep42thought> and probably something in bin/bootstrap-mysql
[20:26:52] <deep42thought> though, that is mostly for stored procedures nowadays
[20:27:52] <abaumann> err. there is no ssh_log?
[20:28:09] <abaumann> ah.
[20:28:12] <abaumann> no. found it. :-)
[20:28:34] <abaumann> KEY `ssh_log_build_slave_date_idx` (`build_slave`,`date`),
[20:28:38] <abaumann> mmh. there it is.
[20:28:43] <deep42thought> the structures of all tables should be in the dump, plus the content of some "static" tables (e.g. `architectures`)
[20:28:44] <abaumann> maybe it got lost in a migration?
[20:28:57] <deep42thought> I updated that file today
[20:29:00] <abaumann> ah.
[20:29:08] <deep42thought> this is a write-only storage
[20:29:16] <deep42thought> noone receives the index from that file
[20:29:31] <deep42thought> it's just for me/us to check that the db is still in the state it should be
[20:29:40] <abaumann> good. it was a 10-20 second penalty on every ssh_log insert.
[20:29:45] <deep42thought> e.g. no architecture is missing, dependency_types did not change indices, etc.
[20:30:05] <buildmaster> i486/x264 is broken (says nlopc46-i486bs1): https://archlinux32.org
[20:30:06] <deep42thought> yeah, I read that. Good hunting, thanks!
[20:30:19] <abaumann> I knew it was something simple. :-)
[20:30:20] <deep42thought> ah, right, I enabled complaining about i486 now, too
[20:30:40] <abaumann> yeah. now that blacklisting is in effect, that's a good thing.
[20:30:43] <deep42thought> ... because there was literally only one package in the current i486 build list!
[20:30:54] <abaumann> uh.
[20:31:19] <abaumann> yeah. there is something wrong there.
[20:31:27] <deep42thought> you think so?
[20:31:30] <abaumann> the 5.2.x kernel should really have ended in the build list for i486
[20:32:03] <deep42thought> hmmm, indeed
[20:32:15] <deep42thought> I was happy that the build list was so short :-/
[20:32:40] <abaumann> also systemd, iptables
[20:33:11] <deep42thought> maybe the breaking get-package-updates dropped those packages?
[20:33:26] <abaumann> yeah. the deadlocks you mean?
[20:33:30] <deep42thought> yes
[20:33:42] <abaumann> I would go for more locking.
[20:33:59] <abaumann> db-update and get-package-updates + a mysqldump in parallel don't play nice, I'm afraid.
[20:34:18] <abaumann> maybe we should add sort of a global lock there
[20:34:32] <deep42thought> yeah
[20:34:41] <deep42thought> but we should also kill all the slow-running queries
[20:34:52] <abaumann> both, yes.
[20:35:14] <abaumann> I'm still debugging the get-assignment query.
[20:35:30] <abaumann> It is very fast (<1 second) for some parts, then suddenly not anymore..
[20:35:39] <abaumann> but the statement is..well..big :-)
[20:35:47] <abaumann> hard to debug in individual parts.
[20:35:55] <deep42thought> the part which determines which dependencies are available and which are not is the slow part
[20:36:15] <abaumann> yeah. this part cannot be avoided.
[20:36:23] <abaumann> I'm thinking about something simple, or..
[20:36:41] <abaumann> ..that the join with the dependencies happens with a too big of a temporary result set (like 100000 packages)
[20:37:01] <abaumann> as we need only one assignment per build slave, maybe a LIMIT 1 somewhere helps
[20:37:07] <buildmaster> i486/homebank is broken (says nlopc46-i486bs0): https://archlinux32.org
[20:37:25] <abaumann> i486/moneybank is broke(n) ;-)
[20:37:34] <deep42thought> !grab abaumann
[20:37:34] <phrik> deep42thought: Tada!
[20:37:40] <abaumann> pun of the day :-)
[20:38:05] <deep42thought> time to leave for me, cu tomorrow
[20:38:15] <abaumann> ok. no problem. cu.
[20:39:33] -!- deep42thought has quit [Quit: Leaving.]
[20:49:13] -!- isacdaavid has quit [Ping timeout: 245 seconds]
[20:49:26] <buildmaster> i486/lollypop is broken (says nlopc46-i486bs0): https://archlinux32.org
[20:49:38] <buildmaster> i486/syncthing is broken (says eurobuild6-7-i486): https://archlinux32.org
[20:49:50] -!- Dimtree has quit []
[20:56:40] -!- abaumann has quit [Quit: leaving]
[21:41:05] -!- Alina-malina has quit [Ping timeout: 258 seconds]
[22:24:18] -!- MrBIOS has quit [Quit: MrBIOS]
[23:23:32] <buildmaster> i486/xmonad is broken (says nlopc46-i486bs0): https://archlinux32.org
[23:25:49] <buildmaster> i486/ffmpeg is broken (says eurobuild6-7-i486): https://archlinux32.org
[23:32:30] <finsternis> upgrading I get: warning: cannot resolve "libwapcaplet>=0.4.2", a dependency of "libcss"
[23:32:33] <finsternis> :: The following package cannot be upgraded due to unresolvable dependencies:
[23:32:54] <finsternis> libcss
[23:32:55] <finsternis> Do you want to skip the above package for this upgrade? [y/N]
[23:40:36] -!- MrBIOS has joined #archlinux32
[23:52:07] <buildmaster> i686/haskell-aeson is broken (says eurobuild3): https://archlinux32.org