#archlinux32 | Logs for 2018-06-14

Back
[00:00:56] <tyzoid> One possible solution is to re-add ffmpeg2.8 and gegl02 to the repos so that packages can be run, or to focus on getting stuff in testing fixed up and pushed out
[00:14:55] -!- oaken-source has quit [Ping timeout: 268 seconds]
[02:34:25] -!- NoobAlice has joined #archlinux32
[02:51:38] -!- jetfrog28 has joined #archlinux32
[03:59:41] <elibrokeit> deep42thought: I committed fixes for several different haskell build cycles :)
[03:59:50] <elibrokeit> only in trunk though
[06:53:25] -!- abaumann has joined #archlinux32
[06:53:26] <buildmaster> Hi abaumann!
[06:53:32] -!- NoobAlice has quit [Quit: Leaving.]
[06:54:12] -!- deep42thought has joined #archlinux32
[06:54:13] <buildmaster> Hi deep42thought!
[06:54:19] <deep42thought> good morning, all!
[06:54:27] <deep42thought> thanks elibrokeit for the haskell fixes
[06:54:31] <deep42thought> this looks great
[06:54:46] <deep42thought> I'll have a look at the dependency mess soonish (not now)
[06:54:56] <deep42thought> I prefer moving testing -> stable
[06:56:59] -!- eduardoeae has quit [Ping timeout: 268 seconds]
[07:39:40] <abaumann> /build/qt5-webkit/src/qtwebkit-5.212.0-alpha2/Source/WebCore/bindings/js/JSDOMConstructor.h:195:35: error
[07:39:43] <abaumann> : invalid conversion from ‘JSC::EncodedJSValue (*)(JSC::ExecState*)’ {aka ‘long long int (*)(JSC::ExecSta
[07:39:46] <abaumann> te*)’} to ‘JSC::NativeFunction’ {aka ‘long long int (__attribute__((fastcall)) *)(JSC::ExecState*)’} [-fp
[07:39:49] <abaumann> ermissive]
[07:39:55] <abaumann> isn't alpha a little bit too bleeding edge, even for Archlinux? ;-)
[07:40:15] <deep42thought> why is this alpha?
[07:40:20] <abaumann> qtwebkit-5.212.0-alpha2
[07:40:25] <deep42thought> ah
[07:40:58] <deep42thought> btw: I moved gimp and vlc, but now they have different issues :-D
[07:41:15] <abaumann> oh.
[07:41:40] <abaumann> I see tons of errors in Xorg.
[07:42:00] <deep42thought> e.g. libgegl-0.4.so.0 and libx265.so.160 are not in stable (yet)
[07:42:10] <abaumann> modeswitching seems to be completly borked for old graphic cards.
[07:42:17] <deep42thought> buildmaster: wtf libgegl-0.4.so.0
[07:42:24] <deep42thought> buildmaster: wtf libx265.so.160
[07:42:24] <buildmaster> [testing] gegl (0.4.2+14+g190a81b4a-1.0): /usr/lib/libgegl-0.4.so.0
[07:43:19] <deep42thought> abaumann: apparently noone tests on such hardware anymore
[07:43:28] <deep42thought> buildmaster: wtf libx265.so.160
[07:43:30] <buildmaster> [testing] x265 (2.8-1.0): /usr/lib/libx265.so.160
[07:43:36] <abaumann> yes, apparently.
[07:43:53] <abaumann> but breaking standard VGA is a little bit too much.
[07:43:59] <deep42thought> yeah
[07:44:13] <abaumann> I cannot test anything anymore in libvirt vms this way.
[07:44:58] <abaumann> QXL in libvirt never worked for me, and other stuff is even more experimental.
[07:45:19] <abaumann> Cyrrus should really not be used anymore..
[07:45:19] <deep42thought> that's frustrating
[07:45:23] <abaumann> StandardVGA is borked
[07:45:49] <abaumann> And I suspect Intel-AVX optimization in KMS mesa softrendering.
[07:46:20] <abaumann> hopefully there is a switch when compiling the thing to disable such things.
[07:47:15] <abaumann> The future looks bright, no window managers working, Xorg not working, sound and video stuff "optimized", so not working. Gives us a nice text-only distribution. :-)
[07:48:50] <deep42thought> ah, I know "the flaw in the deletion logic": it ignores dependencies of packages which are already broken (e.g. if vlc's dependencies are broken, none of its dependencies are taken into account), but vlc (and gimp?) have strange always-broken "dependencies" in the database
[07:49:14] <deep42thought> !grab abaumann
[07:49:14] <phrik> deep42thought: Tada!
[07:49:24] <abaumann> thought so. :-)
[07:49:26] <deep42thought> my sound still works
[07:49:32] <deep42thought> ... still
[07:50:06] <abaumann> I have nothing against optimizations, but what's so hard to do a configure --enable-avx and then #ifdef HAVE_AVX in the code?
[07:52:29] <abaumann> OpenSW: "By default, the build will compile with support for AVX and AVX2 targets", --with-swr-archs=avx,avx2,skx,knl
[07:52:40] <abaumann> ok. maybe recompiling that beast is at hand.
[07:54:12] <abaumann> I remember the old saying in archlinux: "core" is for the things we need to boot the system and build the system. So meson is now part of core?
[07:54:28] <deep42thought> !grab abaumann
[07:54:28] <phrik> deep42thought: Tada!
[07:55:25] <abaumann> -D swr-arches=avx,avx2 in mesa PKGBUILD is not good.
[07:55:35] <abaumann> (for us, that is).
[07:56:47] <deep42thought> well, so we need to change that :-)
[07:56:56] <deep42thought> but I have to leave now
[07:56:59] <deep42thought> ... cu later!
[07:57:02] <abaumann> yes. I'll do that.
[07:57:04] <abaumann> cu
[07:57:08] -!- deep42thought has quit [Quit: Leaving.]
[07:59:09] <abaumann> Oh mesa uses meson.. must be the similarity in the name.. ;-)
[08:01:54] -!- abaumann has quit [Quit: leaving]
[08:32:29] <buildmaster> mesa is broken (says nlopc46).
[08:48:22] -!- abaumann has joined #archlinux32
[08:48:22] <buildmaster> Hi abaumann!
[08:48:34] <abaumann> mesa, that's me :-)
[09:02:19] <abaumann> mesa (source), that was not me..
[09:46:35] <buildmaster> deep42thought: Your buildslave "nlopc43" builds some outdated package.
[10:09:00] <buildmaster> libreoffice-still is broken (says tyzoid-srv0-bs0).
[10:29:58] -!- oaken-source has joined #archlinux32
[11:18:58] -!- oaken-source has quit [Quit: leaving]
[11:19:05] -!- oaken-source has joined #archlinux32
[11:23:20] <abaumann> doublecmd: found the bug, fpc was broken for 32-bit, so lazarus
[11:25:50] <abaumann> elibrokeit: sorry for misreporting it upstream. :-)
[11:26:58] <abaumann> buildmaster: wtf /usr/lib/xorg/modules/libglamoregl.so
[11:27:00] <buildmaster> Huh, I don't know that one.
[11:27:08] <abaumann> !wtf /usr/lib/xorg/modules/libglamoregl.so
[11:27:10] <phrik> abaumann: extra/xorg-server
[11:30:47] <abaumann> buildmaster: !wtf /usr/lib/xorg/modules/libglamoregl.so
[11:55:09] -!- eduardoeae has joined #archlinux32
[12:18:26] <buildmaster> lazarus is broken (says nlopc46).
[12:22:35] -!- Alina-malina has quit [Ping timeout: 268 seconds]
[12:33:41] -!- oaken-source has quit [Ping timeout: 268 seconds]
[13:25:57] <abaumann> lazarus was me :-)
[13:37:07] <elibrokeit> what lazarus thing
[13:42:23] <abaumann> buildmaster was unhappy about lazarus, a free pascal thingy
[13:42:38] <abaumann> which is then used by doublecmd..
[14:36:06] -!- Cthulu201 has quit [Quit: Nowhere special. I always wanted to go there.]
[14:38:54] -!- Cthulu201 has joined #archlinux32
[14:57:20] <tyzoid> abaumann: buildmaster's wtf doesn't accept a path
[14:59:25] <abaumann> ah: thanks. :-0
[14:59:29] <abaumann> * :-)
[14:59:46] <abaumann> buildmaster: !wtf libglamoregl.so
[14:59:53] <tyzoid> no bang
[15:00:05] <abaumann> buildmaster: wtf libglamoregl.so
[15:00:14] <buildmaster> [testing] xorg-server (1.20.0-7.0): /usr/lib/xorg/modules/libglamoregl.so
[15:00:14] <buildmaster> [extra] xorg-server (1.20.0-5.0): /usr/lib/xorg/modules/libglamoregl.so
[15:00:16] <abaumann> I should read the botcode :-)
[15:00:29] <tyzoid> it calls over to my package api
[15:00:55] <tyzoid> https://pkgapi.arch32.tyzoid.com
[15:01:17] <tyzoid> the .xml can be dropped off and it'll return json
[15:03:25] <abaumann> about cgit: nice. :-)
[15:03:50] <abaumann> there is a project-list parameter. maybe you can add descriptions there?
[15:04:24] <abaumann> there is also a 'readme' parameter of the file within the repo to be shown
[15:04:26] <abaumann> I think
[15:04:46] <abaumann> cgit is a miracle with tons of options :-)
[15:05:22] <tyzoid> yeah, I'm using repo-scan, though, so I don't need to change config if/when we add/remove repos
[15:05:40] <abaumann> ah. yes.
[15:05:53] <tyzoid> I could fix it quite easily by not doing that xD
[15:06:22] <abaumann> but repo-path and readme should work along
[15:06:58] <abaumann> well.. our repos have descriptive names. :-0
[15:07:02] <abaumann> * :-)
[15:07:20] <tyzoid> hmm, how would you config. readme?
[15:07:21] <abaumann> * abaumann swears at his much too big fingers compared to the much too small keyboard of an old eeepc
[15:07:58] <abaumann> The question is: do we have a description file inside the repos?
[15:08:05] <abaumann> readme.md might be too big
[15:08:14] <abaumann> readme=<list of files to try>
[15:08:17] <tyzoid> It's also a bare repo
[15:08:20] <abaumann> oh.
[15:08:49] <abaumann> I doubt cgitrc checks out a workspace just to get the files from the bare repo..
[15:10:11] <tyzoid> I'm just pointing it at the archlinux32/* repos from gitea
[15:10:16] <tyzoid> and gitea uses bare repos
[15:10:17] -!- oaken-source has joined #archlinux32
[15:11:22] <abaumann> I'm just curious: what was the cause of the segfault you had?
[15:13:41] <tyzoid> the cause of the segfault was the order of repo.url and repo.path
[15:13:51] <tyzoid> in cgitrc
[15:14:03] <abaumann> oh. good parser. :-)
[15:14:24] <tyzoid> Well, the issue is that in the code, repo.url will trigger the allocation of a new repo object
[15:14:35] <tyzoid> but putting repo.path first won't
[15:14:40] <tyzoid> but it'll still pass through
[15:14:52] <tyzoid> just that the repo will have no path set (rather, it'll be set at null)
[15:15:00] <tyzoid> then when it tries to access it, it's a null pointer
[15:15:33] <abaumann> repo { name=bla; path=blub }
[15:15:43] <abaumann> then you can allocate the repo object on the opening repo scope.
[15:15:56] <tyzoid> except that's not how cgitrc does it
[15:16:03] <abaumann> yes.
[15:16:04] <tyzoid> that's how it should be done
[15:16:08] <tyzoid> or even more simply
[15:16:13] <tyzoid> [repo]
[15:16:15] <tyzoid> path = blah
[15:16:17] <tyzoid> url = blah
[15:16:17] <abaumann> or a json or ini
[15:16:18] <abaumann> yes.
[15:16:27] <abaumann> same applies to the ini format.
[15:16:45] <abaumann> but it does the job well :-)
[15:20:53] -!- abaumann has quit [Remote host closed the connection]
[16:13:48] <buildmaster> logstash is broken (says nlopc46).
[16:57:34] -!- isacdaavid has joined #archlinux32
[17:50:05] -!- abaumann has joined #archlinux32
[17:50:06] <buildmaster> Hi abaumann!
[17:50:10] <tyzoid> wb
[17:50:14] <tyzoid> fixed cgit descriptions, btw
[17:50:21] <abaumann> tyzoid: I get wrongtoken when adding a comment to a bug.
[17:50:29] <tyzoid> :/ odd
[17:50:37] <tyzoid> Try signing out and back in again
[17:51:05] <abaumann> the equivalent of "did you try to turn it off and on again." ;-)
[17:51:36] <tyzoid> Yeah, except it suggests that the login cookie is outdated / unrecognized
[17:51:54] <abaumann> But I lost my comment *grmpf*
[17:53:43] <abaumann> Yes, worked.
[17:53:45] <abaumann> :-)
[17:53:50] <abaumann> https://bugs.archlinux32.org
[17:53:51] <phrik> Title:FS#39 : xorg modesetting module fails with illegal instruction (at bugs.archlinux32.org)
[17:53:59] <abaumann> so Xorg has AVX code in the software renderer.
[17:54:09] <abaumann> This could be the problem for many people..
[17:54:28] <tyzoid> Could that be the problem behind https://bugs.archlinux32.org ?
[17:54:29] <phrik> Title:FS#42 : [xorg-server] 1.20 / [xf86-video-intel] gives me problems with [tigervnc] vncviewer (at bugs.archlinux32.org)
[17:54:50] <abaumann> mmh.. not really.
[17:55:05] <abaumann> this is the Intel driver there, the one I have trouble with is the modesetting driver.
[17:55:16] <tyzoid> I see
[17:55:16] <abaumann> possible, but not liekly.
[17:55:21] <abaumann> *likely
[17:56:27] <abaumann> http://openswr.org
[17:56:29] <phrik> Title:OpenSWR (at openswr.org)
[17:56:40] <abaumann> mmh. should I disable OpenSWR from mesa alltogether?
[17:57:42] <tyzoid> no idea
[17:57:47] <buildmaster> deep42thought: Your buildslave "buildknecht" builds some outdated package.
[17:58:08] <abaumann> me neither :-)
[17:58:15] <abaumann> so, I'll dig trough some code..
[17:58:20] <tyzoid> abaumann: Question about cgit. If using repo scanning, is it possible to tell it to ignore / not match certain repos?
[17:58:30] <tyzoid> I'm trying to figure out how to exclude the .wiki.git repos
[17:59:10] <tyzoid> Didn't see much in cgitrc about it, but it's possible I missed something
[17:59:57] <abaumann> I can't see an option.
[18:00:12] <abaumann> Of course you can have an explicit list, but that's not what you want in the dymanic repo case
[18:00:55] <tyzoid> Yeah, I mean I can write a cron job to generate said list too
[18:01:06] <tyzoid> that's what I'm doing with the descriptions anyway
[18:01:13] <abaumann> ah. then..
[18:04:27] <abaumann> mmh. maybe I have to specify '-Dswr-arches=', just empty.
[18:04:31] <abaumann> let's see :-)
[18:05:25] <abaumann> gallium-drivers=svga or swrast, swr
[18:05:35] <abaumann> maybe there I have to set a gallium driver when starting X
[19:47:46] -!- Alina-malina has joined #archlinux32
[19:52:01] <abaumann> tyzoid: this wrongtoken is really annoying, it's the 3rd time now I'm commenting something and I loose everything I'm commenting as the session expired again..
[19:53:17] <tyzoid> abaumann: Are you changing IP addresses?
[19:53:23] <abaumann> nope.
[19:53:28] <abaumann> I hope not, at least. :-)
[19:53:51] <abaumann> How long is the session? Or did I forget to tick "Remember me"?
[19:53:54] <abaumann> most likely.
[19:54:06] <tyzoid> not sure what the timeout is off the top of my head
[19:54:29] <tyzoid> but I've been signed in for over 48 hours now
[19:55:13] <abaumann> mmh.
[19:55:29] <abaumann> mmh, maybe seamonkey has some issues.
[19:55:33] <tyzoid> btw, if you hit the back button in your browser, you should be able to recover the comment
[19:55:40] <tyzoid> copy it, then log out/in again
[19:55:51] <abaumann> nope. I got the old message wihtout a comment.
[19:55:59] <abaumann> another sign, that this is a seamonkey issue maybe..
[19:56:09] <tyzoid> are you browsing from multiple tabs?
[19:56:16] <abaumann> yep.
[19:56:20] <tyzoid> That could be it.
[19:56:27] <abaumann> mmh. ok.
[19:56:28] <tyzoid> I just checked the code, it's a CSRF token
[19:56:42] <tyzoid> and it checks against the session csrf variable
[19:56:58] <tyzoid> annoying, but I can see why they did it
[19:57:05] <abaumann> ah. and I have multiple tabs open, one per bug report.
[19:57:16] <abaumann> ok. no problem. now I know it. :-)
[20:01:17] <tyzoid> abaumann: Ok, this is really weird. Tabbed browsing works for me here
[20:01:34] <tyzoid> and I'm looking at the code, and it doesn't reset it on page views, it's a session login token
[20:01:44] <tyzoid> which kinda defeats the whole purpose
[20:02:19] <tyzoid> but as far as I can tell, it's not interfering with tabbed browsing
[20:03:57] <tyzoid> and I logged in without the 'remember me' so it shouldn't be that
[20:04:55] <tyzoid> abaumann: Ok, so when you do this, is there a long delay from when you load the page to when you add a comment on it?
[20:05:17] <tyzoid> "long" is relative, meaning approx over an hour
[20:08:42] <abaumann> yeah. could be.
[20:08:56] <abaumann> I was watching socker, so it could be over one hour. ;-)
[20:09:08] <abaumann> And the rest I blame on seamonkey.
[20:12:03] <abaumann> so. I'm trying to build a mesa without gallium and with classic DRI only, no openSWR. maybe I get some working xorg this way.
[20:22:00] <tyzoid> abaumann: I might have fixed it, if my assumption of what was causing it is correct
[20:22:13] <tyzoid> I set the serverside session timeout to 10 days from 30 minutes
[20:52:37] <abaumann> tyzoid: 10 days, that's a lot :-)
[21:03:46] <buildmaster> nvidia-340xx-utils is broken (says eurobuild3).
[21:37:46] <abaumann> moi? :-)
[21:37:50] -!- abaumann has quit [Quit: leaving]
[22:04:21] -!- isacdaavid has quit [Ping timeout: 240 seconds]
[22:38:22] -!- deep42thought has joined #archlinux32
[22:38:22] <buildmaster> Hi deep42thought!
[23:00:10] -!- oaken-source has quit [Ping timeout: 260 seconds]
[23:48:21] -!- deep42thought has quit [Quit: Leaving.]