#archlinux32 | Logs for 2021-04-30

Back
[00:03:00] <buildmaster> i686/gnome-shell is broken (says eurobuild6-6): https://archlinux32.org
[00:16:14] -!- sunshavi has joined #archlinux32
[01:42:54] -!- MrBIOS__ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
[02:01:45] <buildmaster> i486/croc is broken (says eurobuild6-8-i486): https://archlinux32.org
[02:44:50] <buildmaster> i686/evolution is broken (says rechenknecht): https://archlinux32.org
[02:45:33] <buildmaster> i486/onednn is broken (says eurobuild6-8-i486): https://archlinux32.org
[02:55:49] <buildmaster> i686/folks are broken (says rechenknecht): https://archlinux32.org
[03:03:41] <buildmaster> i686/gnome-todo is broken (says eurobuild6-3): https://archlinux32.org
[03:04:10] <nit-picker> The following packages are longer than 3 days in i686/community-staging:
[03:04:10] <nit-picker> lxappearance-obconf-0.2.3-3.1-i686.pkg.tar.zst: since 2021-04-26 (3 days)
[03:04:10] <nit-picker> lxappearance-obconf-gtk3-0.2.3-3.1-i686.pkg.tar.zst: since 2021-04-26 (3 days)
[03:04:10] <nit-picker> obconf-2.0.4-7.1-i686.pkg.tar.zst: since 2021-04-26 (3 days)
[03:04:10] <nit-picker> switchboard-plug-desktop-2.8.4-2.1-i686.pkg.tar.zst: since 2021-04-27 (3 days)
[03:04:10] <nit-picker> waybar-0.9.7-1.1-i686.pkg.tar.zst: since 2021-04-27 (3 days)
[03:04:10] <nit-picker> ... (6 total)
[03:04:40] <buildmaster> i686/gnome-notes are broken (says rechenknecht): https://archlinux32.org
[03:10:23] <buildmaster> i686/glabels are broken (says eurobuild6-3): https://archlinux32.org
[03:13:54] <buildmaster> i686/gnome-phone-manager is broken (says rechenknecht): https://archlinux32.org
[03:20:46] <buildmaster> i686/blender is broken (says eurobuild6-3): https://archlinux32.org
[03:23:42] <buildmaster> i686/openvkl is broken (says eurobuild6-3): https://archlinux32.org
[03:27:01] <buildmaster> pentium4/openvkl is broken (says eurobuild6-3): https://archlinux32.org
[03:28:31] <buildmaster> pentium4/blender is broken (says rechenknecht): https://archlinux32.org
[04:36:07] <buildmaster> i686/prusa-slicer is broken (says eurobuild6-3): https://archlinux32.org
[04:46:37] <KillerWasp> Hello, is there a normal procedure to follow every time a package is broken and does not follow the necessary dependencies?
[04:54:34] <bill-auger> KillerWasp: filing a bug report is the usual procedure
[04:54:45] <KillerWasp> I'm trying to follow the wikipedia instructions, but they don't seem to be correct. I am also checking the packages mentioned in the website list, but their values are wrong.
[04:54:51] <bill-auger> but i suppose that depends what you mean by "broken"
[04:55:34] <bill-auger> sry i dont understand what wikipedia has to do with arch32 packages
[04:56:00] <bill-auger> or what is "the website list"
[04:56:49] <KillerWasp> bill-auger# I mean dependency issues, do I have to report it as a normal bug? Even for every package with dependency bug you find? Even in the web database?
[04:57:21] <bill-auger> do you have a actual problem, or is it a hypothetical question?
[04:57:51] <bill-auger> it would help if you simply explain the exact problem
[04:57:59] <KillerWasp> bill-auger# I have several problems, and I am trying to help in some way, with reports or something. But I find something tedious.
[04:58:19] <bill-auger> when i: ____
[04:58:19] <bill-auger> i expect that ____ should happen
[04:58:19] <bill-auger> instead, ____ actually happens
[04:59:00] <bill-auger> for example: when i pacman -S foo
[04:59:11] <bill-auger> i expect that the foo package will install correctly
[04:59:43] <bill-auger> instead, i see this error message <THE_EXACT_ERRROR_MSG>
[05:00:18] <bill-auger> generally, anytime pacman gives an error, that is a bug deserving of a bug report
[05:01:00] <KillerWasp> bill-auger# Are you reading what I say? they are dependency issues, not pacman bugs.
[05:01:23] <bill-auger> i dont know what that means - pacman's job is to manage dependencies
[05:01:41] <bill-auger> if there were a dependency problem, pacman should catch it
[05:01:54] <bill-auger> again, it would help if you simply explain the exact problem
[05:02:42] <bill-auger> when pacman reports an error, it is almost never a bug with pacman - it is a dependency problem
[05:03:45] <bill-auger> "the web database" i beleive you are referring to the pacman database
[05:04:27] <bill-auger> the information on the packages website is entirely from pacman - you can get all the same information on the command line with pacman -Si foo
[05:09:15] <KillerWasp> bill-auger# i have 5 problems in my own list. And with the database of the web is the 3: https://0x0.st
[05:09:38] <KillerWasp> i must report it one by one in the bug reports?
[05:09:41] <bill-auger> ok thats better
[05:09:52] <bill-auger> yes, if they are separe problems
[05:10:45] <bill-auger> the first one would be a bug report against the 'blender' package
[05:11:21] <bill-auger> the second one would be a bug report against the 'gedit' package
[05:11:52] <bill-auger> the third one looks like you have not upgraded your system - which may fix the first two also
[05:12:29] <bill-auger> the fourth one does not looks like any problem - it is just a helpful note
[05:13:34] <bill-auger> the fifth one would be a bug report against the 'ffmpeg' package - though i installed that same package yesterday
[05:13:55] <bill-auger> do you have the [testing] repos enabled?
[05:14:24] <bill-auger> i think you can fix problenm 5 by disabling the [testing] repos , and running pacman -Syuu
[05:14:42] <KillerWasp> bill-auger# no. i have disabled for testings.
[05:14:59] <bill-auger> i would run pacman -Syuu - just to be sure
[05:15:15] <bill-auger> and also disable [staging] first
[05:16:27] <bill-auger> again take note, that regarding your "3) PROBLEM In Database of packages in the web:" - Database of packages in the web is exactly the same database as your local pacman
[05:17:20] <bill-auger> if your local pacman does not report the same packages as the website, then your local pacman database is out-of-sync - pacman -Sy would put it in sync
[05:18:11] <bill-auger> after you run pacman -Syuu - then run pacman -Ss gedit
[05:18:56] <bill-auger> you will probalby see something like: 40.1-1 (installed 3.38.1-1.4)
[05:19:42] <bill-auger> this is making sense now?
[05:20:49] <bill-auger> if your local pacman still do not show the same packages as the website, that could be an indication that your chosen mirror is out-of-sync
[05:21:46] <bill-auger> in which case, you could change mirrors, then run `pacman -Syyuu`
[05:23:15] <nit-picker> The following packages are longer than 3 days in i686/staging:
[05:23:15] <nit-picker> tepl-6.00.0-2.1-i686.pkg.tar.zst: since 2021-04-27 (3 days)
[05:32:34] <KillerWasp> mmmm, i don't have permission for report the bug in the web (blender specific). I only can send from "Add new task" for all packages.
[05:33:19] <KillerWasp> https://0x0.st
[05:33:55] <KillerWasp> this show late enter in 'Add New Bug' from https://www.archlinux32.org
[05:33:57] <phrik> Title: Arch Linux 32 - blender 17:2.83.5-2.0 (pentium4) (at www.archlinux32.org)
[05:39:53] <bill-auger> "add new task" is correct - just put the package name in the subject
[05:40:03] <KillerWasp> ok
[05:41:29] <bill-auger> a good subject would be like:
[05:41:29] <bill-auger> [blender]: unsatisfied dependency on 'opensubdiv'
[05:42:31] <KillerWasp> bill-auger# Should I do something with that report or do I leave it as is?
[05:44:47] <bill-auger> i dont think you can change it now
[05:50:05] <KillerWasp> well, very thanks :)
[05:50:54] <KillerWasp> for 'gedit' i forget that's stable package
[05:51:57] <KillerWasp> I think that instead of trying with 'testing', I better try 'stable'.
[05:53:04] <KillerWasp> Also, about the number of versions of gedit, it was just a false mistake. Everything's okay.
[05:55:14] <bill-auger> 'stable' is not the right word for a rolling distro
[05:57:07] <bill-auger> 'stable' is not a matter of reliability or bugginess, like radioactivity - it means that the software does not change very often
[05:58:19] <bill-auger> on a rolling distro, the software changes very often - it is by definition 'unstable' - in fact, that is the name debian gave to their rolling-release repo
[06:02:36] <KillerWasp> I guess it is difficult to establish something stable for a small community.
[06:02:59] <bill-auger> a better term for [core], [extra], and [community] would be "standard" or "main-line"
[06:03:20] <bill-auger> no, that is the completely wrong idea
[06:03:45] <bill-auger> stable does not mean "reliable" or "it wont explode" - it means "it wont change very often"
[06:04:22] <bill-auger> when "it wont change very often", that is like an LTS distro - it is actually much much easier to maintain an LTS distro
[06:04:46] <KillerWasp> i see...
[06:04:54] <bill-auger> preceisely because the software does not change very often - a rolling distro is like an order of magnitude more work to maintain
[06:08:21] -!- torv has quit [Ping timeout: 240 seconds]
[06:09:01] -!- torv has joined #archlinux32
[06:12:11] <buildmaster> pentium4/prusa-slicer is broken (says rechenknecht): https://archlinux32.org
[06:12:55] <buildmaster> pentium4/performous are broken (says eurobuild6-4): https://archlinux32.org
[06:43:25] <trotz> 2021/04/30 06:43 CRIT buildmaster Current Load CRITICAL - load average: 8.25, 9.55, 8.19
[06:48:25] <trotz> 2021/04/30 06:48 OK buildmaster Current Load OK - load average: 4.70, 6.16, 7.06
[06:51:40] <buildmaster> i686/onednn is broken (says eurobuild6-3): https://archlinux32.org
[06:52:04] -!- bill-auger has quit [Remote host closed the connection]
[06:52:40] <buildmaster> pentium4/onednn is broken (says eurobuild6-6): https://archlinux32.org
[06:53:03] -!- bill-auger has joined #archlinux32
[06:56:41] <buildmaster> any/python-jsonpatch is broken (says eurobuild6-6): https://archlinux32.org
[07:00:13] <nit-picker> deep42thought: your slave nlopc46 builds i686/chromium for more than a day, now (1 day(s) 09:17:26)
[07:23:43] -!- abaumann has joined #archlinux32
[07:23:43] <buildmaster> Hi abaumann!
[07:23:43] <buildmaster> !rq abaumann
[07:23:44] <phrik> buildmaster: <abaumann> *abaumann does some hackeroo on his rsync-woodo-script..
[07:28:26] <abaumann> KillerWasp, bill-auger: https://0x0.st this is actually fine with me to just drop it here in IRC. As those are simple bugs caused by dependencies. bug reports are better when things are expected to take more time to fix. I'll have a look..
[07:32:41] -!- Nact has joined #archlinux32
[07:34:25] <abaumann> normal users should always use stable, this is Arch, not Debian. :-)
[07:34:37] <abaumann> testing is for, well, when you want to test something
[07:34:56] <abaumann> staging is for the build system itself in order to handle cycles of packages
[07:37:19] <abaumann> there are things which don't build (or don't build on certain architectures): https://archlinux32.org
[07:37:20] <phrik> Title: Blacklisted packages (at archlinux32.org)
[07:38:11] <abaumann> https://archlinux32.org
[07:38:28] <abaumann> this is a more internal view on broken dependencies, more for the devs. :-)
[07:39:33] <abaumann> we have little resources, especially there is little automatic testing in place, so expect things to ripe at the customer..
[07:42:31] <abaumann> The following packages are longer than 3 days in i686/staging: tepl-6.00.0-2.1-i686.pkg.tar.zst: since 2021-04-27 (3 days) points to bug 2)
[07:44:19] <abaumann> btw: IRCS logs are logged and preserved at http://archlinux32.andreasbaumann.cc so you can read this later..
[07:54:26] <abaumann> KillerWasp: please report your architecture (pentium4 or i686), as the packages might differ, the following page explains what the differences are: https://www.archlinux32.org
[07:54:28] <phrik> Title: Arch Linux 32 - Architecture Overview (at www.archlinux32.org)
[08:01:37] <buildmaster> i486/tepl is broken (says eurobuild6-7-i486): https://archlinux32.org
[08:03:39] <abaumann> 2) gedit tepl fixed
[08:29:24] <buildmaster> i686/blender is broken (says eurobuild6-4): https://archlinux32.org
[08:31:18] -!- deep42thought has joined #archlinux32
[08:31:19] <buildmaster> Hi deep42thought!
[08:31:19] <buildmaster> !rq deep42thought
[08:31:20] <phrik> buildmaster: <deep42thought> abstraction, abstraction, abstraction! - in other words: for i in {1..3}; do echo 'abstraction'; done
[08:31:23] <abaumann> hi deep42thought
[08:31:27] <deep42thought> good morning
[08:31:57] <deep42thought> i pushed ffmpeg and the other lib (already forgot the name ...)
[08:32:10] <abaumann> aom, heif?
[08:32:15] <deep42thought> aom, I think
[08:32:35] <deep42thought> btw: I'm fine with minor bugs being reported on the bug tracker - this way, we can simply answer, that we're looking into it and don't do double-work :-)
[08:32:37] <abaumann> 5) fixed :-)
[08:32:42] <deep42thought> but IRC is also fine
[08:33:43] <abaumann> 3) gedit was 3.38 in stable and 40.1 in testing, now both are 40.1
[08:33:51] <buildmaster> pentium4/blender is broken (says rechenknecht): https://archlinux32.org
[08:34:03] <abaumann> 4) not a problem
[08:34:18] <abaumann> 1) blender refuses to build currently
[08:35:01] <deep42thought> I have trouble following your enumeration - but I think, I can guess the respective issues :-)
[08:35:23] <abaumann> ah, I'm referring to https://0x0.st
[08:38:13] <deep42thought> ah, I see
[08:42:47] <KillerWasp> i install gedit and ffmpeg from testing (from something hours), and work fine now. my architecture is pentium4.
[08:43:08] <KillerWasp> abaumann# ^
[08:43:20] <deep42thought> buildmaster: wtp ffmpeg
[08:43:21] <buildmaster> deep42thought: i486/ffmpeg: i486/extra (2:4.3.2-6.4)
[08:43:21] <buildmaster> i686/ffmpeg: i686/extra (2:4.3.2-6.4), i686/staging (2:4.4-1.0)
[08:43:21] <buildmaster> pentium4/ffmpeg: pentium4/extra (2:4.3.2-6.4), pentium4/staging (2:4.4-1.0)
[08:43:28] <deep42thought> buildmaster: wtp gedit
[08:43:29] <buildmaster> deep42thought: i686/gedit: i686/extra (40.1-1.0)
[08:43:29] <buildmaster> pentium4/gedit: pentium4/extra (40.1-1.0)
[08:43:42] <deep42thought> ffmpeg and gedit arrived in stable
[08:43:55] <deep42thought> sry, bill-auger, I think, I'll stick with the name "stable" for now
[08:44:24] <buildmaster> i486/blender is broken (says eurobuild6-7-i486): https://archlinux32.org
[08:44:35] <KillerWasp> abaumann# forget about of 3) and 4)
[08:45:02] <abaumann> I did :-)
[08:45:28] <abaumann> the real remaining issue is with blender..
[08:47:27] <deep42thought> I wonder, how it ended up built in the repositories with these unavailable dependencies
[09:01:03] <bill-auger> i was able to build blender with an ugly hack - it works just fine;and i asked the upstreanm about the implications - the response was unenthusiastic; but im pretty sure its fine
[09:01:52] <bill-auger> thats https://developer.blender.org
[09:01:54] <phrik> Title: rB4a179e8e3e0f (at developer.blender.org)
[09:02:32] <bill-auger> it appeared that one person was trying to restore 32bit support properly, but it is clearly not a prioirty
[09:04:31] <deep42thought> I think, we can live with an increased risk of UUID collissions - can't we?
[09:05:08] <bill-auger> AFAICTm the deal is that my hack will cause object hashes to be 32bit rather than 64bit; which potentially increases the chance of collision (from astromincally low, to astromincally less low)
[09:05:21] <bill-auger> yea that was my conclusion too
[09:05:57] <abaumann> I'm not that far yet, still opensubdiv is a missing dependency for me in blender
[09:06:16] <bill-auger> that change is LOC 95 https://git.parabola.nu
[09:06:20] <deep42thought> abaumann: far from what?
[09:06:49] <bill-auger> if you can patch out cuda from opendibdiv (somehow), i dont think it will be a problem - i just commented on that tocket
[09:07:32] <deep42thought> what's the implication of defining this function, but making it non-atomic?
[09:07:48] <deep42thought> or we add some global critical section (how's that called in C again?)
[09:08:07] <abaumann> Why people insist of writing their own atomic intrinsics is beyond me..
[09:08:16] <abaumann> ..there is support for it in the compiler or libatomic
[09:08:39] <abaumann> most likely some unique ids are then not so unique
[09:08:52] <abaumann> dependening on how threaded blender is, everything can happen. :-)
[09:09:53] <deep42thought> hmm? what do you mean by "writing their own atomic intrinsics"?
[09:10:11] <deep42thought> I thought, the problem was, that on 32 bit, there is *no* 64bit adding atomic intrinsic?
[09:10:42] <abaumann> ah, I thought there is one in atomic..
[09:10:48] <abaumann> ..so I'm wrong there.
[09:10:57] <deep42thought> not sure, I'm just assuming, here :-)
[09:11:07] * deep42thought dives back into `db-update -p`
[09:11:11] <abaumann> even if not, it should be added to the compiler and compiler support libraries
[09:11:18] <abaumann> and not to individual projects..
[09:11:44] <abaumann> maybe upcoming C standards could deal with atomics properly..?
[09:12:09] <abaumann> what about introducing an 'atomic' keyword? :-)
[09:12:29] <deep42thought> atomic main();
[09:13:40] <deep42thought> this sounds like a really nice idea, actually
[09:14:05] <buildmaster> i486/opencv is broken (says eurobuild6-7-i486): https://archlinux32.org
[09:14:12] <abaumann> https://git.archlinux32.org
[09:14:13] <phrik> Title: PKGBUILD « opensubdiv « community - packages - Archlinux32 package modifications (at git.archlinux32.org)
[09:14:22] <abaumann> ah, no cuda there.
[09:15:00] <deep42thought> this is all, that is necessary?
[09:15:06] <abaumann> btw, I'm aware tons of build errors happen on i486 now.. I removed the rust blacklist for i486
[09:15:11] <abaumann> seems so.
[09:15:40] <deep42thought> you try to build all rust-dependent packages on i486? Well, if some of them pass, then that's good :-)
[09:16:03] <abaumann> most of them where not blocked by rust.. librsvg-og helped a lot
[09:16:12] <deep42thought> ah, yes
[09:16:14] <abaumann> and there was a huge mesa messup blocking lots of things.
[09:16:46] <abaumann> https://archlinux32.org
[09:17:02] <abaumann> I see high 'Blocks' counts.. they are computed transitively up the tree?
[09:17:09] <deep42thought> yes
[09:17:14] <abaumann> that's good. :-)
[09:17:41] <deep42thought> some special case are circular dependencies, I think
[09:17:50] <abaumann> tons of them.
[09:17:56] <deep42thought> those are only counted as "1" no matter, how many packages they are actually blocking
[09:18:07] <abaumann> I really urge upstream to once in a year or to rebuild everything from scratch
[09:18:09] <deep42thought> because mathematically, it would be "infinity"
[09:19:22] <abaumann> otoh, once a year somebody does a port to another platform and complains a lot in the bug reporting tool :-)
[09:19:37] <deep42thought> and gets banned by scimmia >:-)
[09:19:47] <abaumann> for instance. :-)
[09:20:01] <abaumann> no hard feelings there..
[09:20:10] <deep42thought> hard? or heart?
[09:20:12] <abaumann> and I try to make only good bug reports now :-)
[09:20:46] <deep42thought> please do! So I don't have to ban you from our bug tracker ;-)
[09:21:31] <abaumann> opensubdiv builds fine
[09:21:40] <deep42thought> \o/
[09:25:09] <bill-auger> yea, i was just looking at the opensibdix PLGBUILD history - it was built without cuda for the first 5 years, though the software apparently supported it - it was only build with cuda since about a year ago
[09:25:35] <abaumann> ah, ok.
[09:25:55] <bill-auger> so my guess was simply dont have cuda installed in the build env, and it will probably build without it
[09:26:31] <abaumann> seems to be auto-probed.. thought adding an explicit --disable-cuda or so is usually prefereable (if the package supports it)
[09:26:43] <abaumann> *though
[09:26:57] <deep42thought> as long as cuda is not in base-devel, the autoprobing will work ,too :-D
[09:27:17] <deep42thought> but I agree, that --disable-cuda would be cleaner
[09:28:47] <bill-auger> its cmake, so probably more like -DWITH_CUDA=OFF, (whatever the CMakeList accepts)
[09:29:09] <abaumann> option(NO_CUDA "Disable CUDA backend" OFF)
[09:29:26] <abaumann> if(NOT NO_CUDA) find_package(CUDA 4.0)
[09:29:26] <abaumann> endif()
[09:29:27] <deep42thought> -DNO_CUDA=ON ?
[09:29:44] <bill-auger> yea thats probably it
[09:29:45] <abaumann> well, double negation ;-)
[09:29:47] <deep42thought> -DNO_SKIP_CUDA_IGNORING=ON
[09:29:53] <abaumann> lol
[09:30:29] <bill-auger> -DNO_SKIP_CUDA_IGNORING=NOT_OFF
[09:30:37] <deep42thought> even better^
[09:30:41] <T`aZ> you guys need to take a break :o
[09:30:51] * deep42thought takes a sip of tea
[09:31:49] * abaumann takes a sip of coffee (number 3 today)
[09:32:20] <deep42thought> I wouldn't count liquids in numbers - I would measure the volume :-D
[09:32:33] <deep42thought> but then again, I'm a physicist
[09:32:40] <abaumann> 3 x one standard coffee unit
[09:32:55] <deep42thought> "coffee pot" = 750ml
[09:33:07] <deep42thought> umm, I meant "coffee cup"
[09:33:30] <abaumann> 25 fluid ounces
[09:33:49] <deep42thought> don't get me started on non-SI units ...
[09:33:53] <bill-auger> micrograms of caffine would be the appropriate unit
[09:34:03] <deep42thought> there's the pro :-)
[09:34:22] <deep42thought> is it micro, not milli?
[09:36:20] <abaumann> yea, 50 micrograms per "normal pot"
[09:36:53] <abaumann> 40 grams a year! oh my ;-)
[09:50:57] <KitsuWhooa> So, I have some patches for openal, however it looks like for i486 and pentium4 we can just do the same thing with cmake parameters
[09:51:06] <KitsuWhooa> for i686 we still need a patch to separate sse and sse2
[09:51:32] <KitsuWhooa> I assume it's preferred to patch the code as less as possible
[09:51:40] <deep42thought> yes :-)
[09:52:08] <deep42thought> it's always preferred to upstrem the patches
[09:52:20] <deep42thought> and if that is rejected, then to patch as less as possible
[09:52:31] <deep42thought> due to maintenance work
[09:52:35] <KitsuWhooa> considering they now bundle see and sse2, I feel like they wouldn't accept a patch splitting them
[09:52:51] <deep42thought> possibly
[09:52:52] <KitsuWhooa> they had them separate before
[09:52:54] <KitsuWhooa> but shrug
[09:53:07] <deep42thought> but I would still ask, if I was you
[09:57:03] <KitsuWhooa> Sure. I'll write a patch and also try to submit it upstream
[09:57:19] <KitsuWhooa> but first, how would it be preferred to add the extra cmake arguments to the pkgbuild?
[09:57:25] <deep42thought> so they'll at least notice, that someone would like sse and sse2 separate :-)
[09:57:47] <deep42thought> with some sed-fu
[09:57:57] <KitsuWhooa> excellent
[09:59:25] <deep42thought> /cmake .*openal-soft/ s/;$/ -D...;/
[09:59:29] <deep42thought> or something similar
[09:59:33] <deep42thought> whatever you like :-)
[10:01:42] <KitsuWhooa> Yeah, that's what I was thinking of just now
[10:02:18] <KitsuWhooa> although I have no idea how that works
[10:02:41] <deep42thought> eval "$(declare -f build | sed '/cmake .*openal-soft/ s/;$/ -D...;/')"
[10:02:47] <deep42thought> this would be the full thing^
[10:03:11] <KitsuWhooa> oh I just realised it's two separate sed expressions
[10:03:16] <KitsuWhooa> I thought it was the same one and I was confused
[10:03:20] <KitsuWhooa> yeah okay, fair enough
[10:03:21] <KitsuWhooa> thanks
[10:03:43] * abaumann is always confused about his own sed-fu ;-)
[10:03:48] <deep42thought> it's an regex filter to match the relevant line and some s/../../ to insert the actual option
[10:04:13] <KitsuWhooa> I'm pretty good at pcre, but I don't have too much experience with sed itself other than simple sed -i s/// :p
[10:46:55] <buildmaster> i486/picard is broken (says eurobuild6-8-i486): https://archlinux32.org
[10:47:06] <abaumann> so, that was the problem, opensubdiv was on the CUDA blacklist for pentium4, so that's why blender "lost" it's dependency..
[10:52:10] <deep42thought> and blender got built before opensubdiv entered the blacklist?
[10:52:23] <buildmaster> i486/opensubdiv is broken (says eurobuild6-7-i486): https://archlinux32.org
[10:52:29] <KitsuWhooa> Here you go https://tasossah.com
[10:52:35] <KitsuWhooa> https://tasossah.com
[10:52:36] <phrik> Title: openal new sse patch (at tasossah.com)
[10:53:33] <KitsuWhooa> I'll also submit the separation patch upstream. If it's accepted, you can just get rid of it from the pkgbuild
[10:56:00] <KitsuWhooa> ah fuck
[10:56:05] <KitsuWhooa> set(CORE_OBJS ${CORE_OBJS} core/mixer/mixer_sse.cpp core/mixer/mixer_sse2.cpp)
[10:56:09] <KitsuWhooa> forgot to remove sse2 from there
[10:56:13] <KitsuWhooa> here we go all over
[10:56:56] <KitsuWhooa> *sse
[10:57:07] <KitsuWhooa> realistically the file just gets added twice on the list so it shouldn't harm anything, but it's bugging me
[11:02:37] <KitsuWhooa> https://tasossah.com there we go
[11:07:09] <abaumann> mmh, binary git patch?
[11:07:20] <abaumann> didn't even know this exists.
[11:07:50] <KitsuWhooa> Me neither, but for some reason git thinks the patches are binaries
[11:07:53] <KitsuWhooa> and I don't think I can do anything about it
[11:08:09] <KitsuWhooa> https://tasossah.com
[11:08:17] <KitsuWhooa> this is the patch inside the patch
[11:08:20] <abaumann> as long as it applies.
[11:08:31] <abaumann> yeah, that might be the reason, git sees the patch files as binary.
[11:13:09] <KitsuWhooa> ...
[11:13:14] <KitsuWhooa> I think I messed up
[11:13:18] <KitsuWhooa> -- Installing: /tmp/openal/pkg/openal/usr/lib-DALSOFT_CPUEXT_NEON=0/libopenal.so.1.21.1
[11:13:23] <KitsuWhooa> what the hell
[11:13:38] <KitsuWhooa> how did this even happen?
[11:13:50] <KitsuWhooa> *sigh*
[11:14:16] <T`aZ> laugh of the morning, thank you :p
[11:14:55] <deep42thought> heh, you missed a space :-)
[11:16:19] <KitsuWhooa> I swear it worked properly at first...
[11:16:21] <KitsuWhooa> time for v3!
[11:17:46] <KitsuWhooa> cmake -S openal-soft -B build -DCMAKE_INSTALL_PREFIX=/usr -DCMAKE_BUILD_TYPE=None -DCMAKE_INSTALL_LIBDIR=lib-DALSOFT_CPUEXT_NEON=0 -DALSOFT_CPUEXT_SSE4_1=0 -DALSOFT_CPUEXT_SSE3=0 -DALSOFT_CPUEXT_SSE2=0;
[11:17:47] <KitsuWhooa> yup
[11:20:03] <KitsuWhooa> apparently -DALSOFT_CPUEXT_NEON=0 isn't required, which makes sense I guess since we're not compiling on arm
[11:20:19] <KitsuWhooa> I didn't expect cmake to complain when I set it though. Oh well
[11:20:31] <deep42thought> as long as you install your libs into /usr/lib-DALSOFT_CPUEXT_NEON=0
[11:20:34] <KitsuWhooa> XD
[11:22:27] <KitsuWhooa> Okay this should be the final one :p https://tasossah.com
[11:27:42] <KitsuWhooa> why am I still getting a SIGILL
[11:30:11] <KitsuWhooa> maybe that's why they got rid of SSE I guess
[11:32:38] <KitsuWhooa> nevermind, I was loading the old library :p
[11:34:33] <KitsuWhooa> ...it broke later. *sigh*
[11:37:15] <abaumann> argh. speaking of failing patches, the SSE2 patch of blender is also broken..
[11:37:19] <KitsuWhooa> >/* SSE2 is required for any SSE support. */
[11:37:26] <abaumann> huh?
[11:37:31] <abaumann> ah. in openal?
[11:37:36] <KitsuWhooa> yeah
[11:38:05] <abaumann> just got a cvttss2si 0x40(%esp),%esi in SDL2 *sigh*
[11:38:13] <KitsuWhooa> ouch
[11:38:28] <abaumann> 486 is painful :-)
[11:39:33] <buildmaster> i686/cheese is broken (says rechenknecht): https://archlinux32.org
[11:40:11] -!- deep42thought has quit [Quit: Leaving.]
[11:43:50] <Cif> painful but beautiful :3
[11:46:33] <abaumann> :-)
[11:52:25] <KitsuWhooa> okay, I got it working
[11:52:31] <abaumann> cool :-)
[11:53:03] <KitsuWhooa> I had to comment this out https://github.com
[11:53:11] <KitsuWhooa> I'm not sure why this was added
[11:53:17] <KitsuWhooa> unless I'm missing something, I don't see any SSE2 in there
[11:53:58] <buildmaster> i686/grilo-plugins are broken (says eurobuild6-3): https://archlinux32.org
[11:56:33] <KitsuWhooa> the sad news is that something broke and extreme tux racer doesn't load any stages now
[12:04:15] -!- JuniorJPDJ has parted #archlinux32
[12:04:32] -!- JuniorJPDJ has joined #archlinux32
[12:04:36] <KitsuWhooa> okay, time for patch v4
[12:08:36] <abaumann> afk, lunch and fetching a fixed bike..
[12:09:01] <KitsuWhooa> have fun :p
[12:09:09] <abaumann> thanks :-)
[12:37:25] <KitsuWhooa> Okay, I'm done https://tasossah.com
[12:37:39] <KitsuWhooa> Do you want me to email it, or just linking it here enough? :p
[12:41:49] <abaumann> Just a link here is fine. :-)
[12:42:31] <KitsuWhooa> awesome, thanks
[12:42:39] <abaumann> thank you for fixing. :-)
[12:42:40] <KitsuWhooa> https://github.com I also opened this
[12:43:01] <abaumann> perfect!
[13:05:54] <buildmaster> pentium4/opencv is broken (says eurobuild6-4): https://archlinux32.org
[13:19:15] <buildmaster> i686/calibre is broken (says eurobuild6-4): https://archlinux32.org
[13:41:09] <buildmaster> i686/glom is broken (says eurobuild6-4): https://archlinux32.org
[14:01:05] <buildmaster> pentium4/kmplayer is broken (says eurobuild6-3): https://archlinux32.org
[14:03:27] <buildmaster> pentium4/frameworkintegration is broken (says rechenknecht): https://archlinux32.org
[14:05:52] <buildmaster> pentium4/soundkonverter is broken (says eurobuild6-4): https://archlinux32.org
[14:27:53] -!- deep42thought has joined #archlinux32
[14:27:53] <buildmaster> Hi deep42thought!
[14:27:53] <buildmaster> !rq deep42thought
[14:27:54] <phrik> buildmaster: <deep42thought> dependencies tend to be either trivial or circular
[14:46:10] <buildmaster> i686/digikam is broken (says eurobuild6-6): https://archlinux32.org
[14:55:38] <abaumann> blender has patches for OpenEXR 3.x, but they seem not to be complete..
[14:56:36] <abaumann> /usr/include/openvdb/Types.h:10:10: fatal error: OpenEXR/half.h: No such file or directory
[14:56:41] <abaumann> ah, hang on..
[14:58:06] <abaumann> upstream builds so..
[14:58:21] <deep42thought> we're missing include headers?
[14:58:51] <abaumann> openvdb
[14:59:07] <abaumann> usr/include/openvdb/Types.h is owned by community/openvdb 7.1.0-1
[14:59:09] <deep42thought> I meant openvdb on arch32 vs. on arch64
[14:59:23] <abaumann> Let me check first it's age..
[15:01:00] <abaumann> ah, there is a version 8..
[15:02:07] <buildmaster> pentium4/vamp-aubio-plugins are broken (says rechenknecht): https://archlinux32.org
[15:03:24] <abaumann> openvdb: Could NOT find NumPy (Required is at least version "1.14.0")
[15:03:34] <deep42thought> we don't have numpy?
[15:04:45] <buildmaster> pentium4/ardour is broken (says eurobuild6-4): https://archlinux32.org
[15:05:16] <abaumann> yes, we should..
[15:05:53] <abaumann> "Running cythonize failed"
[15:05:57] <abaumann> doesn't reawlly build.
[15:06:33] <abaumann> broken cython?
[15:07:06] <buildmaster> pentium4/sonic-pi is broken (says eurobuild6-6): https://archlinux32.org
[15:10:58] <abaumann> cython is fine (at least it built, and it's quite recent in the repos)
[15:11:25] <abaumann> subprocess.CalledProcessError: Command '['/usr/bin/python', '-m', 'cython', '-3', '--fast-fail', '-o', 'bit_generator.c', 'bit_generator.pyx']' died with <Signals.SIGSEGV: 11>.
[15:11:29] <abaumann> segfault. nice.
[15:18:00] <abaumann> ok, blender goes onto the buglist
[15:28:21] <abaumann> btw: major upgrade on my build machines and mirror: LTO-4, now with a proper U320 SCSI host card.. :-)
[15:28:26] <abaumann> expect some downtime..
[15:28:36] <deep42thought> from you or your machines?
[15:36:21] <abaumann> aeh. :-)
[15:36:26] <abaumann> hopefully the later
[15:36:33] <deep42thought> ok :-)
[15:41:03] <abaumann> -- SSEMATH (Wanted: ON): OFF
[15:41:04] <abaumann> mmh
[15:41:20] <abaumann> or is fpupath=sse since sss2
[15:41:33] <abaumann> well, SDL2 never really build correctly
[15:42:13] <abaumann> maybe no sse math is set globally (glibc), so it's switched off automatically?
[15:42:23] <abaumann> I have to go through all those things once.. :-)
[16:26:36] <buildmaster> i686/ruby-gdk3 is broken (says rechenknecht): https://archlinux32.org
[16:28:55] -!- GNUtoo has joined #archlinux32
[16:31:51] <GNUtoo> Hi, is it a known issue that somehow wget caps files it downloads at 2GiB while archlinux's wget works fine for files > 2GiB ?
[16:32:19] <GNUtoo> When downloading a file on my local network through HTTP, I have "Length: 2147483647 (2.0G) [application/x-xz]" whereas the file is about 6GiB
[16:33:12] <GNUtoo> On the same machine I've tried Guix's wget and it works fine (Length: 6222995512 (5.8G) [application/x-xz])
[16:33:34] <GNUtoo> I've tried rebuilding the package but I get the same issue
[16:35:12] <GNUtoo> When comparing both wget the build flags are the same (-cares +digest -gpgme +https +ipv6 +iri +large-file -metalink +nls +ntlm +opie +psl +ssl/gnutls)
[16:36:41] <abaumann> no, this is not known, but sounds like wget uses a non-64-off_t file position pointer.
[16:36:58] <abaumann> or something wget uses to do the download
[16:37:22] <buildmaster> pentium4/ruby-gdk3 is broken (says rechenknecht): https://archlinux32.org
[16:37:29] <GNUtoo> ok
[16:55:06] -!- abaumann has quit [Quit: leaving]
[17:04:45] -!- deep42thought has quit [Quit: Leaving.]
[17:15:45] -!- MrBIOS_ has joined #archlinux32
[17:33:02] -!- torv has quit [Remote host closed the connection]
[17:33:22] -!- torv has joined #archlinux32
[18:00:46] <buildmaster> pentium4/blender is broken (says eurobuild6-6): https://archlinux32.org
[18:01:46] <buildmaster> i686/blender is broken (says eurobuild6-4): https://archlinux32.org
[18:07:38] <buildmaster> pentium4/calibre is broken (says eurobuild6-1): https://archlinux32.org
[18:41:13] <buildmaster> pentium4/watchexec is broken (says eurobuild6-1): https://archlinux32.org
[18:41:14] <buildmaster> i686/watchexec is broken (says eurobuild6-4): https://archlinux32.org
[20:04:28] -!- Cif has quit [*.net *.split]
[20:04:29] -!- SamantazFox_ has quit [*.net *.split]
[20:04:29] -!- thePiGrepper_ has quit [*.net *.split]
[20:08:24] -!- Cif has joined #archlinux32
[20:08:24] -!- SamantazFox_ has joined #archlinux32
[20:08:24] -!- thePiGrepper_ has joined #archlinux32
[20:08:24] -!- titus_livius has quit [Ping timeout: 265 seconds]
[20:09:42] -!- titus_livius has joined #archlinux32
[20:33:38] <buildmaster> i486/gitlab-runner is broken (says eurobuild3-i486): https://archlinux32.org
[20:47:20] -!- sunshavi has quit [Remote host closed the connection]
[20:49:31] -!- fuselage has quit [Read error: No route to host]
[21:28:00] -!- Nact has quit [Quit: Konversation terminated!]
[21:39:44] -!- sunshavi has joined #archlinux32
[22:18:07] <buildmaster> i486/gnome-shell is broken (says eurobuild6-7-i486): https://archlinux32.org
[22:29:38] <buildmaster> i686/gpaste is broken (says rechenknecht): https://archlinux32.org
[22:30:19] <buildmaster> i686/evolution-data-server is broken (says eurobuild6-2): https://archlinux32.org
[22:31:40] <buildmaster> i686/gnome-terminal is broken (says eurobuild6-6): https://archlinux32.org
[22:35:56] <buildmaster> pentium4/gpaste is broken (says eurobuild6-5): https://archlinux32.org
[22:50:13] <buildmaster> i686/gnome-shell is broken (says eurobuild6-6): https://archlinux32.org
[22:55:21] <buildmaster> pentium4/mutter is broken (says eurobuild6-1): https://archlinux32.org
[23:01:59] -!- City-busz has quit [Quit: http://quassel-irc.org - Csevegjen kényelmesen. Bárhol.]
[23:02:00] -!- elibrokeit has quit [Quit: A random quit message]
[23:02:36] -!- City-busz has joined #archlinux32
[23:02:39] -!- elibrokeit has joined #archlinux32
[23:04:02] <nit-picker> The following packages are longer than 3 days in i486/community-staging:
[23:04:02] <nit-picker> ccache-4.2.1-1.0-i486.pkg.tar.zst: since 2021-04-27 (3 days)
[23:49:11] -!- oaken-source has quit [Ping timeout: 240 seconds]
[23:56:21] -!- oaken-source has joined #archlinux32