#archlinux32 | Logs for 2018-08-03

[00:03:21] -!- oaken-source has quit [Ping timeout: 264 seconds]
[02:02:44] -!- Alina-malina has quit [Ping timeout: 260 seconds]
[02:06:34] -!- Alina-malina has joined #archlinux32
[02:27:17] -!- isacdaavid has quit [Quit: Leaving.]
[02:54:33] -!- bill-auger has quit [Quit: No Ping reply in 60 seconds.]
[02:54:48] -!- bill-auger has joined #archlinux32
[04:51:40] -!- NoobAlice has joined #archlinux32
[07:33:34] -!- NoobAlice has quit [Quit: Leaving.]
[08:26:33] -!- eduardoeae has quit [Ping timeout: 268 seconds]
[08:27:24] -!- oaken-source has joined #archlinux32
[08:32:27] -!- deep42thought has joined #archlinux32
[08:32:27] <buildmaster> Hi deep42thought!
[09:11:17] -!- abaumann has joined #archlinux32
[09:11:18] <buildmaster> Hi abaumann!
[09:12:52] <abaumann> unknown error (_ssl.c:3630): looks like I have a i486-specific boo-hoo in python's ssl support..
[09:13:05] <abaumann> verifying certificates..
[09:22:05] -!- abaumann has quit [Quit: leaving]
[10:03:27] -!- oaken-source has quit [Ping timeout: 240 seconds]
[10:03:38] -!- abaumann has joined #archlinux32
[10:03:40] <buildmaster> Hi abaumann!
[10:04:44] <abaumann> deep42thought: you were right: python-pytest downloads setuptools-scm during installation, I added it do the makedepends. I fear, there is a whole bunch of modules upstream which download stuff from the internet while being packaged instead of using the already built Archlinux package
[10:05:48] -!- abaumann has quit [Client Quit]
[11:45:49] -!- ahuillet has joined #archlinux32
[11:45:51] <ahuillet> hi
[11:45:57] <deep42thought> Hi
[11:46:16] <ahuillet> rsync: send_files failed to open "/.bash_history" (in archlinux32): Permission denied (13)
[11:46:28] <deep42thought> gnaaa
[11:46:30] <deep42thought> not again
[11:46:34] <ahuillet> when rsyncing from rsync://mirror.archlinux32.org
[11:46:40] <deep42thought> try again
[11:46:44] <ahuillet> not sure what that is but I think it's not on my end :)
[11:46:57] <ahuillet> looks good now.
[11:47:21] <deep42thought> I su'ed to that user yesterday to correct some issues and that creates that file on the mirror
[12:13:16] -!- eduardoeae has joined #archlinux32
[12:55:14] -!- oaken-source has joined #archlinux32
[14:11:01] -!- woshty has joined #archlinux32
[14:27:05] <elibrokeit> abaumann, deep42thought: if they're missing makedepends/checkdepends, please give me a list and I'll try to fix this, it's unambiguously wrong
[14:27:32] <elibrokeit> ... though, because of the "download if missing" behavior, packagers might not notice :p
[14:27:55] <deep42thought> there should(?) be some feature to disable internet during build()
[14:28:06] <deep42thought> e.g. emulate some entry in /etc/hosts
[14:28:12] <deep42thought> that disconnects the machine completely
[14:29:50] <elibrokeit> too many things use the internet when running tests, let alone download their source code during build() like node, golang, java
[14:30:29] * deep42thought was not serious
[14:31:10] -!- ahuillet has parted #archlinux32
[14:31:19] <deep42thought> but there was once the concept of a source tar ball, that would contain all the necessary parts for building a package offline ....
[14:32:57] * elibrokeit shakes fist at nodejs
[14:45:01] -!- abaumann has joined #archlinux32
[14:45:04] <buildmaster> Hi abaumann!
[14:45:09] <deep42thought> Hi abaumann!
[14:45:14] <abaumann> hi there. :-)
[14:45:53] <abaumann> very soon only a machine learning algorithm will be able to devise a build plan for a Linux distribution..
[14:46:03] <deep42thought> !grab abaumann
[14:46:03] <phrik> deep42thought: Tada!
[14:46:34] <deep42thought> the question is rather if the number of necessary boot strap stages grows faster than you pass through them ;-)
[14:47:14] <elibrokeit> well, all I can fix for now is the missing makedepends, so if you can give me that list I will see what I can do about it :)
[14:47:26] <abaumann> I have already two python packages downloading sources while building: community/python-pytest downloads python-setuptools-scm and community/python-mock downloads community/python-pbr
[14:47:40] <abaumann> just those two so far.
[14:47:56] <deep42thought> is there an easy test to check this?
[14:48:09] <abaumann> use a broken i486 python package of mine :-)
[14:48:17] <abaumann> with broken ssl support.
[14:48:20] <deep42thought> heh
[14:48:39] <abaumann> no, not really, maybe a proxy not downloading stuff from pypi?
[14:48:57] <deep42thought> it's all coming from pypi?
[14:49:16] <abaumann> all = two packages, well, don't know yet.
[14:49:19] <deep42thought> then an entry in /etc/hosts should solve it
[14:49:26] <abaumann> it's not exactly a statistically relevant sample..
[14:49:31] <abaumann> good idea.
[14:50:20] <elibrokeit> wtf it doesn't even print messages during build() to say it did it
[14:50:23] <abaumann> forbidding downloadings generally is not a good idea, but having 'no-download' the default and then add git/svn/hg/web/ftp whatever proxies is maybe a way.
[14:50:33] <abaumann> yep.
[14:51:02] <abaumann> * abaumann suggests shaking fists also at pyhthon
[14:52:21] <deep42thought> let's see, what " pypi.org" in /etc/hosts does ...
[14:53:15] <elibrokeit> deep42thought: was just going to suggest this :D
[14:53:21] <deep42thought> changes nothing ...
[14:53:28] <deep42thought> at least not during build() ...
[14:53:32] <abaumann> mmh.
[14:53:44] <elibrokeit> abaumann: well, it "makes sense" for most uses where you're not distro packaging
[14:53:46] <elibrokeit> like pip
[14:54:10] <abaumann> true.
[14:54:10] <abaumann> [pbr] Reusing existing SOURCES.txt
[14:54:21] <abaumann> is there a cache somewhere?
[14:55:58] <abaumann> Ok. Let me check. It could also be that my i486 bootstrapped python causes such effects and they are not reproducible on i686 or 64-bit.
[14:56:35] <abaumann> ah.
[14:56:43] <abaumann> or the modules are already installed?
[14:56:58] <deep42thought> which package emits this message?
[14:57:08] <abaumann> python-mock
[14:57:08] <deep42thought> python-pbr?
[14:57:39] <abaumann> Couldn't find index page for 'setuptools-scm' (maybe misspelled?)
[14:57:40] <abaumann> Download error on https://pypi.python.org unknown error (_ssl.c:3630) -- So
[14:57:42] <abaumann> me packages may not be found!
[14:57:49] <abaumann> Just for pbr instead of setuptools-scm
[14:58:36] <deep42thought> "[pbr] Reusing existing SOURCES.txt" is also there on i686
[14:58:50] <abaumann> ah. so this is a message fro pbr..
[14:59:58] <abaumann> aha.
[15:00:14] <abaumann> python-mock has a checkdepends python-pbr, you have to disable checks
[15:00:31] <abaumann> in setup.py there is 'setup_requires: pbr'
[15:03:40] <deep42thought> running install
[15:03:40] <deep42thought> [pbr] Generating AUTHORS
[15:03:40] <deep42thought> [pbr] AUTHORS complete (0.0s)
[15:03:50] <deep42thought> these are the lines that fail for you, abaumann?
[15:04:27] <abaumann> no. it doesn't even start building. setuptools complains about not being able to download pbr because of my broken SSL support.
[15:04:39] <abaumann> I'm struggling to reproduce it..
[15:05:04] <deep42thought> me, too
[15:05:31] <abaumann> Can I use staging-xxx-build and disable checks?
[15:05:40] <abaumann> Without hacking a makepkg.conf?
[15:05:49] <deep42thought> stagin-i686-build -- -- --nocheck
[15:06:10] <abaumann> ah. parameters get passed down to makepkg. thanks.
[15:06:47] <deep42thought> https://git.archlinux32.org
[15:06:52] <abaumann> damn, not reproducible.
[15:06:54] <phrik> Title:archlinux32/builder: Tools for building 32-bit archlinux packages from archlinux.org's official, 64-bit tested PKGBUILDs et al. - Archlinux32 Gitea (at git.archlinux32.org)
[15:08:23] <abaumann> aha.
[15:08:48] <abaumann> I hacked the /etc/hosts file of /var/lib/archbuild/staging-x86/root before starting the script.
[15:08:53] <abaumann> then it happends
[15:08:57] <abaumann> *happens
[15:09:02] <deep42thought> what did you put there?
[15:09:09] <abaumann> pypi.org
[15:09:18] <deep42thought> that did not change anything for me
[15:09:30] <deep42thought> pypi.python.org did not change anything either ...
[15:09:48] <abaumann> maybe the root fs of archbuild already has a pbr installed?
[15:09:54] <abaumann> no, should not be..
[15:11:25] <abaumann> http://archlinux32.andreasbaumann.cc
[15:11:48] <abaumann> ah. cunning. this is a server with wrongly setup SSL certificates :-)
[15:12:16] <deep42thought> pypi.org???
[15:12:24] <abaumann> now localhost
[15:12:31] <deep42thought> ah
[15:12:42] <abaumann> It may not fail on 404 and just continue..
[15:13:02] <abaumann> so maybe not putting there but something clearly illegal?
[15:13:27] <deep42thought>
[15:13:47] <abaumann> yeah. nice :-)
[15:13:59] <deep42thought> hah, I can reproduce!
[15:14:02] <deep42thought> (the issue)
[15:14:16] <abaumann> No route to host or so.
[15:14:33] <abaumann> BRB
[15:14:38] <deep42thought> cu
[15:15:46] <deep42thought> https://ptpb.pw
[15:16:09] <deep42thought> this was: extra-x86_64-build -- -- --nocheck
[15:16:28] <deep42thought> with " pypi.org" and " pypi.python.org"
[15:16:36] <deep42thought> (I don't have https running here on that box)
[15:17:30] <abaumann> back
[15:17:59] <deep42thought> I'll give this box here something to do over the weekend
[15:18:13] <deep42thought> and let it compile all python* packages --nocheck and with pypi.org blocked
[15:18:23] <deep42thought> to see what downloads stuff during build()
[15:18:36] <abaumann> that's actually a good idea.
[15:25:37] * deep42thought scheduled 2630 package builds
[15:25:45] <abaumann> ouch :-)
[15:42:57] <deep42thought> one should probably do this as cronjob: try to rebuild all current packages to see what new issues pop up
[16:16:00] -!- deep42thought has quit [Quit: Leaving.]
[16:22:34] -!- abaumann has quit [Quit: leaving]
[16:51:01] <elibrokeit> oh, hmm, looks like mock has pbr as a checkdepends. :p
[17:14:27] -!- oaken-source has quit [Ping timeout: 240 seconds]
[17:37:54] <buildmaster> electron is broken (says nlopc46).
[17:47:05] -!- woshty has quit [Ping timeout: 240 seconds]
[18:00:03] -!- deep42thought has joined #archlinux32
[18:00:04] <buildmaster> Hi deep42thought!
[18:00:24] <deep42thought> ok, there are other python packages with similar issues, too:
[18:00:48] <deep42thought> python2-pylint downloads pytest-runner
[18:01:10] <deep42thought> python-astroid downloads pytest-runner
[18:01:48] <deep42thought> python-wstools downloads pytest-runner
[18:02:08] <deep42thought> python-pylint downloads pytest-runner
[18:06:34] -!- abaumann has joined #archlinux32
[18:06:34] <buildmaster> Hi abaumann!
[18:06:46] <abaumann> I found lockfile to pbr
[18:06:52] <deep42thought> Hi abaumann!
[18:07:02] <abaumann> hi.
[18:07:22] <abaumann> you cannot disable sphinx on python-pip. nice. :-)
[18:07:37] <abaumann> now I have to build all of sphinx first
[18:09:03] <deep42thought> and you're sure, you won't hit any dependency circles?
[18:09:11] <abaumann> so far not.
[18:09:23] <abaumann> but I'm just properly rebuilding the requirements for meson for now.
[18:09:45] <abaumann> maybe some 40-50 modules so far.
[18:27:45] -!- abaumann has quit [Quit: leaving]
[18:28:36] <tyzoid> hey deep42thought: any regions you'd like to have additional mirror bandwidth?
[18:29:01] <deep42thought> umm, I have no opinion on that
[18:29:10] * tyzoid still hasn't taken down the varnish caches
[18:29:42] <tyzoid> I was planning on taking them down now, but if there's somewhere we want to have additional capability, then I'll leave it up
[18:29:46] <tyzoid> apac, perhaps?
[18:39:34] <deep42thought> we have a mirror in singapore
[18:39:41] <deep42thought> and india
[18:39:44] <deep42thought> IIRC
[18:43:33] <deep42thought> yeah: India, Singapore, Japan
[19:15:16] <elibrokeit> https://git.archlinux.org
[19:15:17] <phrik> Title:svntogit/community.git - Git clone of the 'community' repository (at git.archlinux.org)
[19:15:38] <elibrokeit> mock is more interesting, because I also want to fix the broken testsuite :)
[19:15:58] <elibrokeit> https://github.com
[19:15:59] <phrik> Title:Fix tests/drop 2.6-compat via unittest2 by eli-schwartz · Pull Request #432 · testing-cabal/mock · GitHub (at github.com)
[22:08:39] -!- isacdaavid has joined #archlinux32
[22:18:55] -!- oaken-source has joined #archlinux32
[22:37:24] <buildmaster> git-lfs is broken (says nlopc46).
[23:15:27] -!- eduardoeae has quit [Ping timeout: 240 seconds]
[23:15:40] -!- oaken-source has quit [Ping timeout: 256 seconds]
[23:17:50] -!- eduardoeae has joined #archlinux32