#archlinux32 | Logs for 2024-09-05

Back
[02:41:38] -!- AtleoS has joined #archlinux32
[03:14:44] -!- KitsuWhooa has quit [Quit: Unable to handle kernel NULL pointer dereference at null]
[03:15:43] -!- KitsuWhooa has joined #archlinux32
[05:00:54] -!- AtleoS has quit [Read error: Connection reset by peer]
[06:25:55] -!- titus_livius has joined #archlinux32
[06:32:09] -!- buildmaster has quit [Remote host closed the connection]
[06:34:36] -!- abaumann has joined #archlinux32
[06:34:53] <abaumann> any architecture packages should not be built on i486 machines or chroots.
[06:35:22] -!- buildmaster has joined #archlinux32
[06:35:23] <abaumann> The reason is simple: if ruby itself fails to build on i486, then ruby packages may fail at random
[06:36:04] <abaumann> another reason is: 486 are VMs and use more resources, also, they are needed to build the big backlog of i486 failures.
[06:41:06] <KitsuWhooa> yes, but any should probably be built on i686/pentium4 chroots so that the versions are vaguely related
[06:42:27] <abaumann> exactly
[06:42:29] <KitsuWhooa> Oh I think I misread what you meant. I thought by chroot you meant i686/pentium4, not an i486 chroot
[06:42:36] <KitsuWhooa> pretty sure any gets built on pentium4 by default
[06:42:42] <abaumann> ah. ok.
[06:42:44] <KitsuWhooa> at least that's how I remember seeing it all this time
[06:42:55] <abaumann> then everything is fine. :-)
[06:42:58] <KitsuWhooa> if you see any packages being built on i486 please let me know :p
[06:43:04] <abaumann> ok.
[06:43:19] <abaumann> I just had to kick ruby in i486 lately (because of llvm14 libs or so)
[06:43:30] <KitsuWhooa> ruby has been a major pain in general :p
[06:43:34] <abaumann> So I thought the ruby packages failing you mentioned are caused by this.
[06:43:44] <KitsuWhooa> nah
[06:43:45] <abaumann> who is still using ruby anyway?
[06:43:49] <KitsuWhooa> lots of people apparently
[06:43:53] <abaumann> ah?
[06:43:54] <KitsuWhooa> and protobuf now has a dependency on ruby
[06:43:55] <abaumann> mmh.
[06:44:00] <KitsuWhooa> so no ruby -> no protobuf
[06:44:01] <abaumann> oh, handy
[06:44:15] <KitsuWhooa> as in, the upstream package has a dependency to generate the bindings
[06:44:26] <KitsuWhooa> anyway, what happened with ruby is packages got built in the wrong order
[06:44:31] <abaumann> and a subpackage protobuf-ruby-bindings as subpackage is not an option, I suppose.
[06:44:50] <KitsuWhooa> ruby-* packages got built before ruby itself, and now a whole bunch of rebuilds ended up being built for 3.0.0 and not 3.2.0
[06:44:56] <KitsuWhooa> so I had to find them and force rebuild them
[06:45:17] <abaumann> yep, this is the pain for all of ruby, python, perl, haskell and friends
[06:45:30] <KitsuWhooa> surely there's some way to get buildmaster to not schedule those
[06:45:45] <KitsuWhooa> like, "if a new version of python is in the build list, don't build anything that depends on python"
[06:45:58] <abaumann> or the packages to fail fast. but this means one would have to pin the interpreter version somehow into thedependendies or so
[06:45:59] <KitsuWhooa> it'll slow things down until someone goes and fixes python for example, but it won't result in additional hedaches afterwards
[06:46:09] <KitsuWhooa> they already are pinned
[06:46:14] <KitsuWhooa> the problem is the new versions that get added
[06:46:21] <KitsuWhooa> the pins are added after the package is built
[06:46:30] <KitsuWhooa> so, say, upstream releases python 3.13 and a whole bunch of python-
[06:46:37] <KitsuWhooa> python itself fails to build, but python-* builds against 3.12
[06:46:40] <KitsuWhooa> they are now pinned against 3.12
[06:46:43] <abaumann> so, a makedepend with python=3.2 or so?
[06:47:00] <KitsuWhooa> yes, but we can't know ahead of time
[06:47:05] <abaumann> which would be a royal pain to change on every update of python, of course.
[06:47:08] <KitsuWhooa> yup
[06:47:24] <KitsuWhooa> that's why if package has makedepends in python and python in build list then don't build
[06:47:27] <KitsuWhooa> no matter what
[06:47:55] <KitsuWhooa> although I find python itself these days does build without issues
[06:48:03] <KitsuWhooa> ruby has broken on every release :p
[06:49:41] <abaumann> my eurobuild6 builder have a while to come up today. I had some filesystem issues on the machine. And the machine has some Terabytes of disks. :-)
[06:51:33] <abaumann> There was a big update on the buildmaster/archlinux32.org
[06:51:40] <abaumann> They are up again now and look fine.
[06:51:41] <KitsuWhooa> I stopped my builders last night because I got annoyed that they were building haskell packages
[06:51:50] <KitsuWhooa> I don't know if they added new ones that need blacklisting
[06:52:02] <abaumann> yeah, saw that too. I don't understand, why haskell is built again now..
[06:52:12] <KitsuWhooa> sometimes a few packages slip by while get-package-updates is running
[06:52:19] <abaumann> ah
[06:52:27] <KitsuWhooa> but when it finishes the builders are told to stop
[06:52:36] <KitsuWhooa> however I don't think that's entirely what happened
[06:52:49] <KitsuWhooa> because I got buildmaster PMing me on IRC about outdated packages yet there were still haskell packages being built afterwards
[06:54:10] <abaumann> AFK for 10 mins, have to move location.. :-)
[06:56:20] -!- abaumann has quit [Quit: leaving]
[07:08:52] -!- ssserpent has joined #archlinux32
[07:17:40] -!- abaumann has joined #archlinux32
[07:17:40] <buildmaster> Hi abaumann!
[07:17:40] <buildmaster> !rq abaumann
[07:17:41] <phrik> buildmaster: <abaumann> All the times I was debugging some network settings, just to find out, that I tripped a cable. :-)
[08:15:28] -!- dvzrv has quit [Quit: WeeChat 4.3.6]
[08:15:51] -!- dvzrv has joined #archlinux32
[09:45:20] <abaumann> BOOST_STACKTRACE_LIBCXX_RUNTIME_MAY_CAUSE_MEMORY_LEAK, nice :-)
[09:45:35] <abaumann> boost got stuck on all sub-archs causing some build blocking
[09:49:36] <KitsuWhooa> great
[10:15:46] <KitsuWhooa> > 13:07:48 <buildmaster> any/qt5-doc is broken (says eurobuild6-4): https://archlinux32.org
[10:15:48] <KitsuWhooa> not again
[10:17:16] -!- dmc has quit [Quit: WeeChat 4.3.4]
[10:19:13] -!- dmc has joined #archlinux32
[10:19:26] -!- drathir_tor has quit [Ping timeout: 260 seconds]
[10:29:18] <KitsuWhooa> > WARNING: QDoc will not be compiled, probably because libclang could not be located. This means that you cannot build the Qt documentation.
[11:31:36] <abaumann> Well, it might need clang static analysis funcions to read the C++ header files for generating the documenration.
[11:43:23] <KitsuWhooa> yeah, it's just that libclang got updated and it broke things
[11:43:37] <KitsuWhooa> we really need some kind of stricter dependency tracking
[12:02:26] <abaumann> mmh. no boost shim packages since boost180-libs, from 1.81 to 1.85, suspicios. So everythin just got stuck linked against 1.81 or so?
[12:02:59] <abaumann> yeah, I got pretty lazy lately in the project. :-)
[12:19:21] -!- drathir_tor has joined #archlinux32
[12:35:00] -!- drathir_tor has quit [Remote host closed the connection]
[12:35:25] -!- drathir_tor has joined #archlinux32
[12:46:47] -!- abaumann has quit [Quit: leaving]
[13:22:29] -!- ssserpent has quit [Ping timeout: 255 seconds]
[13:24:22] -!- ssserpent has joined #archlinux32
[17:59:23] -!- ssserpent has quit [Quit: WeeChat 4.4.1]
[23:34:14] -!- gehidore has quit [Quit: brb parkour]