#archlinux32 | Logs for 2019-01-22

Back
[00:00:22] <deep42thought> anyway, for my packages I'm now just doing, what you suggested: they depend on the exact version of every dependency they link against (which is _not_ necessarily the version of the library!) and this forces me to rebuild the package for each update of a dependency
[00:00:56] <deep42thought> which is funnily creating quite the situation upstream wanted to avoid in the first place - just worse :-D
[00:00:59] <deep42thought> but it's ok for me
[00:02:14] <thePiGrepper> I suggested that? hmm, I might have also said that it made sense to try to optimize the process so that if possible you could avoid rebuilding endlessly :-)
[00:02:44] <deep42thought> yeah, you suggested that too
[00:02:48] <thePiGrepper> but yeah, no perfect answer I guess
[00:02:54] <deep42thought> but upstream prevents me from making it efficiently
[00:02:57] <thePiGrepper> and yeah, their position is understandable as well
[00:03:38] <thePiGrepper> one question though. if you made all the changes upstream, which is an absurd amount of work btw, would they take it then?
[00:04:01] <deep42thought> umm, what changes are you talking about?
[00:04:05] <deep42thought> what should I change?
[00:06:15] <thePiGrepper> adding the provide lines I guess.. or is that only for makepkg?
[00:06:46] <deep42thought> you would have to do it for each PKGBUILD that creates libraries
[00:07:07] <thePiGrepper> right
[00:07:22] <thePiGrepper> I meant that
[00:07:26] <deep42thought> and you would actually need to _know_ which *.so files will be needed by other packages and which are only unimportant ballast
[00:07:36] <thePiGrepper> that's what they are against, right?
[00:07:42] <deep42thought> which is similar to the exceptions-thing I was talking about earlier
[00:07:47] <deep42thought> yes
[00:08:14] <deep42thought> they are against massively adding versioned provides (and depends?) entries for libraries
[00:10:17] <thePiGrepper> hmm. ok, going to back to today's example with qbittorrent. today was an example of the typical breakage case, right?
[00:10:26] <thePiGrepper> at least one of the typical cases
[00:10:37] <deep42thought> yeah
[00:10:47] <deep42thought> it should not be broken in the first place
[00:10:50] <deep42thought> thought
[00:10:53] <deep42thought> *though
[00:11:02] <thePiGrepper> one [testing] package was needed by one which by some reason built against it and was send to stable first
[00:11:05] <thePiGrepper> *sent
[00:11:30] -!- yans has quit [Ping timeout: 250 seconds]
[00:11:49] <deep42thought> that was already step 2 of the problem
[00:12:17] <deep42thought> the first step was, that a lib in stable got replaced (by mistake) by a newer lib from testing, breaking two packages in stable
[00:12:26] <deep42thought> ... actually only one package
[00:12:31] <thePiGrepper> at the very least, that should happen. and should have a way to fix. the other issue with programs unable to build or work with latest libraries I think it's another issue which can be dealt independently
[00:13:34] <thePiGrepper> right, that's an issue as well. when trying to fix something, you can easily break 'old' packages by updating their dependencies to newest versions
[00:13:52] <deep42thought> that's, what I did :-/
[00:13:54] <thePiGrepper> I think the key is the testing->stable check
[00:14:11] <deep42thought> there are checks, but I did the update with "-f" which overrides the checks
[00:14:27] <deep42thought> I did just not have that case in mind
[00:14:46] <thePiGrepper> something needs to go there to check the whole image collective as a holistically
[00:14:57] <thePiGrepper> s/as a//
[00:15:20] <deep42thought> db-update will do that - but it will most probably not update what you are waiting for ...
[00:16:00] <thePiGrepper> sure, there are checks, but then what you're saying is that you broke qbittorrent by forcing something?
[00:16:00] <deep42thought> that's why I used -f
[00:16:27] <thePiGrepper> that's not right. bittorrent was broken when it itself was updated alone
[00:16:28] <deep42thought> qbittorrent was broken, probably due to the bug which I removed a few days ago
[00:16:40] <thePiGrepper> without it updated dependency
[00:16:44] <thePiGrepper> *its
[00:16:56] <thePiGrepper> boost-libs
[00:16:56] <deep42thought> btw: qbittorrent was never broken, it was libtorrent-ra*something*
[00:17:09] <thePiGrepper> ok
[00:17:14] <thePiGrepper> point taken
[00:17:15] <deep42thought> which was broken, because it was _not_ updated
[00:17:28] <deep42thought> but its dependencies were
[00:17:40] <thePiGrepper> hmm, sure?
[00:17:43] <deep42thought> yes
[00:17:46] <thePiGrepper> because the issue didnt reflect that
[00:17:49] <thePiGrepper> give me a min
[00:17:52] <deep42thought> then I rebuilt that libtorrent thing
[00:18:04] <deep42thought> but unfortunately, there was already a newer version of the dependency out
[00:18:11] <thePiGrepper> qbittorrent: error while loading shared libraries: libboost_system.so.1.69.0: cannot open shared object file: No such file or directory
[00:18:15] <deep42thought> and of course, it was rebuilt against _that_
[00:18:34] <thePiGrepper> libtorrent-ra*blabla* asked for a newer version of boost-lib
[00:18:35] <deep42thought> yes, the error is from a library which is used by qbittorrent
[00:18:38] <thePiGrepper> not an older one
[00:18:42] <deep42thought> yes
[00:18:48] <deep42thought> oh, wait
[00:19:03] <deep42thought> that was after the rebuild of libtorrent-...
[00:19:14] <deep42thought> first it asked for an older version
[00:19:37] <deep42thought> ah, no
[00:19:40] <thePiGrepper> are you sure? this is the state of stable 20 hours ago
[00:19:42] <thePiGrepper> approx
[00:19:46] <deep42thought> looks like I read that wrong
[00:19:53] <deep42thought> oops :-)
[00:20:00] <deep42thought> I should go, get some sleep ...
[00:20:06] <thePiGrepper> dont worry
[00:20:08] <thePiGrepper> cu
[00:20:53] <deep42thought> nah, not so fast ;-)
[00:21:37] * thePiGrepper is slowing down..
[00:21:47] <deep42thought> so, ok: libtorrents dep was too new, then I (unneccessarily) rebuilt libtorrent :oops: and moved it to stable
[00:21:51] <deep42thought> which changed nothing :-D
[00:22:03] <deep42thought> and then I moved the dep to fix libtorrent - by force
[00:22:04] <thePiGrepper> right
[00:22:10] <thePiGrepper> correct
[00:22:17] <deep42thought> which broke of course all packages built against the old version of libboost
[00:22:24] <thePiGrepper> exactly
[00:22:28] <thePiGrepper> bingo
[00:22:36] <thePiGrepper> which points out something important
[00:22:49] <deep42thought> and then I moved all packages that did not break something which was not already broken (e.g. hopefully all packages built against the new libboost version)
[00:23:17] <thePiGrepper> that original issue. the one of bringing new libtorrents to stable without its newest dependency version wasnt caught by the checks
[00:23:21] <thePiGrepper> and that's an issue
[00:23:38] <thePiGrepper> hmm, I think that last statement is also correct, yes
[00:23:54] <deep42thought> damn, I cannot check anymore, when libtorrent was moved to stable
[00:24:09] <deep42thought> to see wether it was before I removed the bug from db-update or after
[00:24:57] <thePiGrepper> I think waiting for anything to break is the best approach currently
[00:25:06] <deep42thought> :-)
[00:25:06] <thePiGrepper> lets hope nothing does
[00:25:28] <thePiGrepper> but it would be good to think about that long term solution
[00:25:47] <thePiGrepper> that libtorrent went to stable alone
[00:25:53] <thePiGrepper> and noone stopped it
[00:25:54] <deep42thought> well, if db-update is bug-free *cough*, then nothing will break anymore :-D
[00:26:29] <deep42thought> with the old bug present it is totally reasonable to have random packages moved around
[00:26:37] <deep42thought> it was messing up the check for the correct architecture
[00:26:47] <thePiGrepper> because human-made projects reach that bugfree status eventually :-)
[00:27:03] <deep42thought> e.g. $a (depending on $b) would be moved if $b was available for any architecture at all in the target repositories' stabiliy
[00:27:30] <thePiGrepper> does that bug have to do with this particular situation?
[00:27:40] <deep42thought> it created it
[00:27:45] <thePiGrepper> for certain?
[00:27:50] <deep42thought> pretty sure
[00:27:59] <deep42thought> it has created lots of similar situations in the past
[00:28:05] <thePiGrepper> I see. then it's good news, let's hope it doesnt happen again :-)
[00:28:10] <deep42thought> yeah
[00:28:18] <thePiGrepper> is that database a git project as well?
[00:28:27] <deep42thought> no, it's a mysql database
[00:28:31] <thePiGrepper> oh
[00:28:34] <deep42thought> or rather a mariadb
[00:28:39] <thePiGrepper> isnt that the same?
[00:28:47] <deep42thought> yeah, somewhat
[00:28:53] <thePiGrepper> oracle-less mysql, right?
[00:28:57] <deep42thought> yeah
[00:29:27] <thePiGrepper> ok. well. you better go rest, what time is it there?
[00:29:36] <deep42thought> 0:29
[00:29:48] <thePiGrepper> oh
[00:30:11] <thePiGrepper> tomorrow I need to ask you something regarding building for atom netbooks
[00:30:22] <thePiGrepper> might be an interesing conversation
[00:30:26] <thePiGrepper> :-)
[00:30:41] <deep42thought> ok, cu tomorrow!
[00:30:44] <deep42thought> (or today)
[00:30:44] <thePiGrepper> cu
[00:30:46] <deep42thought> :-D
[00:30:48] -!- deep42thought has quit [Quit: Leaving.]
[00:45:44] -!- woshty has quit [Ping timeout: 250 seconds]
[01:05:40] -!- oaken-source has quit [Ping timeout: 250 seconds]
[01:46:21] -!- MrBIOS has quit [Quit: MrBIOS]
[02:21:35] -!- isacdaavid has quit [Quit: Leaving.]
[02:22:14] <finsternis> after the ruby upgrade, I get http://sprunge.us
[03:12:33] <bill-auger> thanks for reminding me - that was noticed on parabola recently https://labs.parabola.nu
[03:12:34] <phrik> Title: Bug #2148: [ruby]: broken x86_64 and i686 - Packages - Parabola Issue Tracker (at labs.parabola.nu)
[04:27:27] <buildmaster> i686/gpgme is broken (says eurobuild3).
[04:33:23] -!- MrBIOS has joined #archlinux32
[05:19:13] * buildmaster goes insane.
[05:42:27] -!- guys has quit [Quit: A random quit message]
[05:56:59] * rcf offers buildmaster a bottle of haldol
[05:58:36] -!- guys has joined #archlinux32
[06:05:10] <girls> get well soon, buildmaster
[06:09:00] * buildmaster resumes sanity.
[06:09:12] <girls> thanks, buildmaster
[06:09:12] <buildmaster> np, girls
[06:22:39] -!- deep42thought has joined #archlinux32
[06:22:39] <buildmaster> Hi deep42thought!
[06:22:39] <buildmaster> !rq deep42thought
[06:22:40] <phrik> buildmaster: <deep42thought> computers ranked by dirt: smokers, car mechanics, bedroom, ...
[06:23:21] <deep42thought> bill-auger: what do you think the fix for the ruby errors is?
[06:23:50] <deep42thought> just moving packages around or do we (and you) need to rebuild ruby-* packages?
[06:30:05] -!- thePiGrepper has quit [Ping timeout: 246 seconds]
[06:35:50] <bill-auger> probably just needs rebuilding
[06:36:15] <bill-auger> arch upgraded ruby to 2.6 and added a new ruby-2.5 package
[06:36:32] <bill-auger> the ruby-2.5 package is required only by gitlab
[06:37:16] <bill-auger> i think the errors are because some binaries were moved out of the main ruby package into eparate packages (irb and rubygems)
[06:38:28] <deep42thought> ok, I'll reschedule ruby-*, then
[06:38:40] <bill-auger> i havenot actually put any of that to the test yet - that is what i grokked from reading the arch git logs
[06:39:08] <bill-auger> ruby needs upgrade though - not just rebuild
[06:40:07] <deep42thought> ah, hmm
[06:40:08] <bill-auger> i thnk the new rubygems and ruby-irb packages caused a conflict with the existing ruby package
[06:40:44] <bill-auger> that was ruby 2.5
[06:50:58] <deep42thought> hmm, I'll need to take a look at that when I have more time than now ...
[06:51:01] -!- deep42thought has quit [Quit: Leaving.]
[07:07:06] <bill-auger> i have not dealt with these yet myself, but it is likely there could be a dependency cascade with the ruby-* packages - those are mainly re-packaged ruby "gems" and ruby gems often specify a strict ruby version when installed with rubygems package manager
[07:08:57] <bill-auger> of course we would hope the PKGBUILD captures those deps if they are strict requirements
[07:46:11] -!- woshty has joined #archlinux32
[08:22:23] -!- deep42thought has joined #archlinux32
[08:22:23] <buildmaster> Hi deep42thought!
[08:22:23] <buildmaster> !rq deep42thought
[08:22:24] <phrik> buildmaster: <deep42thought> brtln: you're "the arch linux", we're "the other arch linux" and alarm is "arches linux", then?
[08:22:35] <deep42thought> bill-auger: hoping that seems futile
[08:37:25] <bill-auger> packaging gems and pip eggs seems fultile also, even though it should be the sane thing to do - but the upstreams never intend for their libs or "apps" to be packaged by a distro - its isolated nests of versioned eggs inside of versioned eggs - turtle eggs all the way down - it is amazing any two of them can fit into the same distro
[08:39:18] <bill-auger> hence the special ruby package that exists only for gitlab
[09:05:50] <buildmaster> i686/aubio is broken (says rechenknecht).
[09:21:54] -!- oaken-source has joined #archlinux32
[09:38:18] -!- bill-auger has quit [Ping timeout: 250 seconds]
[09:38:41] -!- bill-auger has joined #archlinux32
[09:50:58] -!- davor has quit [Ping timeout: 268 seconds]
[09:55:35] <buildmaster> i686/python-h5py-openmpi is broken (says nlopc43).
[10:16:58] -!- oaken-source has quit [Ping timeout: 272 seconds]
[10:25:13] <buildmaster> i686/alacritty is broken (says buildknecht).
[10:25:17] -!- davor has joined #archlinux32
[10:58:02] -!- oaken-source has joined #archlinux32
[11:54:05] <buildmaster> i686/haskell-hspec-meta is broken (says buildknecht).
[12:03:24] <buildmaster> i686/haskell-data-dword is broken (says nlopc46) - I rescheduled: haskell-tasty-quickcheck.
[12:05:43] <buildmaster> i686/haskell-ghc-typelits-knownnat is broken (says rechenknecht) - I rescheduled: haskell-tasty-quickcheck.
[12:15:14] <buildmaster> i686/haskell-spdx is broken (says rechenknecht) - I rescheduled: haskell-tasty-quickcheck.
[12:23:00] <buildmaster> i686/haskell-base64-bytestring is broken (says rechenknecht) - I rescheduled: haskell-test-framework-quickcheck2.
[12:27:49] <buildmaster> i686/haskell-decimal is broken (says nlopc43) - I rescheduled: haskell-test-framework-quickcheck2.
[12:31:19] <buildmaster> i686/haskell-blaze-html is broken (says nlopc46) - I rescheduled: haskell-test-framework-quickcheck2.
[12:35:39] <buildmaster> i686/haskell-fingertree is broken (says nlopc43) - I rescheduled: haskell-test-framework-quickcheck2.
[12:42:40] <buildmaster> i686/haskell-base-orphans are broken (says nlopc43) - I rescheduled: haskell-hspec, haskell-hspec-core.
[12:43:58] <buildmaster> i686/haskell-hspec-core is broken (says nlopc43).
[12:44:12] -!- ofara has quit [Ping timeout: 250 seconds]
[12:45:12] <buildmaster> i686/haskell-base-compat-batteries are broken (says rechenknecht) - I rescheduled: haskell-hspec, haskell-hspec-core.
[12:45:17] -!- ofara has joined #archlinux32
[12:49:51] <buildmaster> i686/haskell-logging-facade is broken (says nlopc43) - I rescheduled: haskell-hspec, haskell-hspec-core.
[12:50:31] <buildmaster> i686/haskell-constraints are broken (says nlopc46).
[12:52:58] <buildmaster> i686/haskell-generic-deriving is broken (says rechenknecht) - I rescheduled: haskell-hspec, haskell-hspec-core.
[12:55:53] <buildmaster> i686/haskell-loop is broken (says nlopc43) - I rescheduled: haskell-hspec, haskell-hspec-core.
[13:00:08] <buildmaster> i686/haskell-megaparsec6 is broken (says nlopc43) - I rescheduled: haskell-hspec, haskell-hspec-core.
[13:02:57] <buildmaster> i686/haskell-megaparsec is broken (says rechenknecht) - I rescheduled: haskell-hspec, haskell-hspec-core.
[13:05:34] <buildmaster> i686/haskell-reinterpret-cast is broken (says nlopc43) - I rescheduled: haskell-hspec, haskell-hspec-core.
[13:08:09] <buildmaster> i686/haskell-newtype-generics are broken (says nlopc46) - I rescheduled: haskell-hspec, haskell-hspec-core.
[13:10:07] <buildmaster> i686/haskell-safe-exceptions are broken (says nlopc43) - I rescheduled: haskell-hspec, haskell-hspec-core.
[13:11:53] <buildmaster> i686/haskell-resourcet is broken (says rechenknecht) - I rescheduled: haskell-hspec, haskell-hspec-core.
[13:12:01] <deep42thought> buildmaster: shut up
[13:12:01] <buildmaster> Sorry, I will do.
[13:59:10] -!- woshty has quit [Ping timeout: 250 seconds]
[14:14:46] <buildmaster> i686/haskell-infer-license is broken (says nlopc46) - I rescheduled: haskell-hspec, haskell-hspec-core.
[14:16:46] <buildmaster> i686/haskell-th-orphans are broken (says rechenknecht) - I rescheduled: haskell-hspec, haskell-hspec-core.
[14:17:53] <deep42thought> one might think, one hour of silence is enough to reschedule all haskell packages, but apparently it's not
[14:21:37] <buildmaster> i686/haskell-doctest is broken (says nlopc43) - I rescheduled: haskell-hspec, haskell-hspec-core.
[14:27:40] -!- woshty has joined #archlinux32
[14:39:20] <buildmaster> i686/haskell-network is broken (says nlopc43) - I rescheduled: haskell-hspec, haskell-hspec-core.
[14:45:35] -!- yans has joined #archlinux32
[14:49:35] <buildmaster> i686/haskell-hspec-core is broken (says rechenknecht) - I rescheduled: haskell-hspec-meta.
[14:57:58] -!- thePiGrepper has joined #archlinux32
[15:00:00] <buildmaster> i686/haskell-http2 is broken (says nlopc46) - I rescheduled: haskell-aeson, haskell-hspec, haskell-hspec-core.
[15:02:00] <buildmaster> i686/haskell-hspec is broken (says buildknecht).
[15:07:59] <buildmaster> i686/haskell-hspec-smallcheck is broken (says nlopc43) - I rescheduled: haskell-hspec.
[15:10:55] <buildmaster> i686/haskell-incremental-parser is broken (says rechenknecht) - I rescheduled: haskell-checkers.
[15:13:15] <buildmaster> i686/haskell-hspec-contrib is broken (says nlopc46) - I rescheduled: haskell-hspec.
[15:20:27] <buildmaster> i686/haskell-unix-time is broken (says nlopc43) - I rescheduled: haskell-hspec.
[15:23:07] <buildmaster> i686/haskell-mono-traversable is broken (says nlopc46) - I rescheduled: haskell-hspec.
[15:27:29] <buildmaster> i686/haskell-haddock-library is broken (says rechenknecht) - I rescheduled: haskell-hspec.
[15:37:32] -!- davor has quit [Ping timeout: 250 seconds]
[15:38:42] -!- davor has joined #archlinux32
[15:59:55] -!- woshty has quit [Ping timeout: 244 seconds]
[16:07:03] -!- deep42thought has quit [Quit: Leaving.]
[16:31:01] -!- phrik has quit [Remote host closed the connection]
[16:31:15] -!- phrik has joined #archlinux32
[16:31:16] <buildmaster> Hi phrik!
[16:31:16] <buildmaster> !rq phrik
[16:35:19] <phrik> buildmaster: <phrik> i am bot
[17:06:27] -!- MrBIOS has quit [Quit: MrBIOS]
[17:37:08] -!- oaken-source has quit [Ping timeout: 245 seconds]
[17:38:43] -!- davor_ has joined #archlinux32
[17:41:28] -!- davor has quit [Ping timeout: 250 seconds]
[17:41:28] davor_ is now known as davor
[17:45:49] -!- MrBIOS has joined #archlinux32
[17:59:59] -!- thePiGrepper has quit [Quit: leaving]
[18:03:41] -!- T`aZ has quit [Ping timeout: 268 seconds]
[18:05:21] -!- T`aZ has joined #archlinux32
[18:16:29] -!- woshty has joined #archlinux32
[18:42:38] -!- abaumann has joined #archlinux32
[18:42:38] <buildmaster> Hi abaumann!
[18:42:38] <buildmaster> !rq abaumann
[18:42:39] <phrik> buildmaster: <abaumann> So I prefer cheap bootstrapping methods over the correct ones :-)
[18:50:23] <abaumann> mmh. firefox uses now /tmp, this is nice, so you get an 'out of disk space error' due to a tmpfs which again resides in memory?
[18:51:18] <abaumann> I would actually prefer to add a switch to staging-i686 and create a normal /tmp on a platter, though this will result in high I/O and build time, but at least it would succeed.
[18:52:01] <abaumann> df -k /tmp shows, it is on the platter, mmh..
[18:55:58] <abaumann> 804G!! this should really be sufficient event for ld to link libxul.so
[18:56:05] <abaumann> *even
[18:57:15] <abaumann> 225:22.78 lto-wrapper: fatal error: write: No space left on device
[18:57:22] <abaumann> *sigh*
[19:20:25] -!- abaumann has quit [Quit: leaving]
[19:27:45] -!- noctambulo has joined #archlinux32
[19:29:47] <noctambulo> I'm getting "^@" over the text after booting the archiso
[19:30:12] <noctambulo> and i can't even open cfdisk since it closes by itself
[19:53:28] <buildmaster> i686/firefox is broken (says eurobuild3).
[21:18:51] -!- noctambulo has quit [Ping timeout: 256 seconds]
[21:28:15] -!- DCyrax has quit [Remote host closed the connection]
[21:33:07] -!- oaken-source has joined #archlinux32
[21:45:13] -!- thePiGrepper has joined #archlinux32
[21:56:42] <buildmaster> i686/python-faker is broken (says nlopc43).
[21:57:25] <buildmaster> i686/prettier is broken (says nlopc43).
[22:39:14] <T`aZ> omfg, im trying to run 'rake' which seems to be a ruby software
[22:39:27] <T`aZ> and im hitting that issue : https://labs.parabola.nu
[22:39:28] <phrik> Title: Bug #2148: [ruby]: broken x86_64 and i686 - Packages - Parabola Issue Tracker (at labs.parabola.nu)
[22:39:58] <T`aZ> i have no idea what they are talking about, and how/if it applies to arch