#archlinux32 | Logs for 2018-06-22

[06:09:57] <buildmaster> nvidia-390xx-lts is broken (says tyzoid-srv0-bs0).
[06:27:02] <buildmaster> nvidia-340xx is broken (says tyzoid-srv0-bs0).
[07:44:04] -!- Polichronucci has quit [Ping timeout: 265 seconds]
[08:21:49] -!- deep42thought has joined #archlinux32
[08:21:50] <buildmaster> Hi deep42thought!
[11:44:45] <buildmaster> cifs-utils is broken (says eurobuild3).
[12:23:08] <deep42thought> hmm, I added stunnel in front of the buildmaster's mysqld and now the replication works through the encrypted connection, but the `Slave_running` status "flickers" between "ON" and "OFF" on the slave - so 50% of the time, packages.archlinux32.org prints a warning about a desynchronized slave :-/
[15:01:32] -!- Polichronucci has joined #archlinux32
[15:06:08] <tyzoid> weird
[15:07:04] <deep42thought> nvm, found the issue (was not related): I set up another replication slave and used the same server-id for it
[15:07:14] <tyzoid> ah
[15:07:44] <deep42thought> in case we need more replicas of the database, there is now a script in releng to achieve this with only a few key strokes
[15:08:24] <tyzoid> sweet
[15:09:54] <tyzoid> uh
[15:10:07] <tyzoid> why is https://git.archlinux32.org minified?
[15:10:38] <deep42thought> I guess, I copied that from upstream :-/
[15:10:43] <tyzoid> lol
[15:11:10] <deep42thought> feel free to de-minimize it :-)
[15:11:41] <tyzoid> damn, no vim 8 on 16.04 :/
[15:12:19] <tyzoid> git.archlinux32.org been stable for you since the move?
[15:12:48] <deep42thought> yeah
[15:12:51] <deep42thought> has some latency
[15:12:55] <deep42thought> but no errors
[15:13:13] <deep42thought> e.g. sometimes pushing takes upto 5 or 10 seconds
[15:13:37] <tyzoid> Yeah, but the website seems solid
[15:13:44] <tyzoid> at least to me
[15:14:35] <deep42thought> it's no permanent issue, just something I noticed a couple times
[15:16:57] <tyzoid> Yeah, I think it's something to do with key auth on that box
[15:17:09] <tyzoid> I've had systems that just take a while to ssh to, even under no load
[15:17:14] <tyzoid> probably some config issue
[15:17:24] <deep42thought> haveged?
[15:17:37] <deep42thought> e.g.: do you have enough entropy?
[15:17:39] <tyzoid> worth a shot
[15:18:09] <tyzoid> aha
[15:18:13] <tyzoid> memory usage is 90%
[15:18:24] <deep42thought> oh
[15:18:30] <tyzoid> gonna power off the vm to boost that
[15:20:46] <tyzoid> alright, rebooted and installed / started haveged
[15:22:21] <tyzoid> lol whoops https://git.archlinux32.org
[15:22:22] <phrik> Title:archweb32.git - Archlinux32 package and build system information interface (at git.archlinux32.org)
[15:22:56] <deep42thought> dunno, it worked on packages.archlinux32.org, but not on my local clone
[15:23:22] <tyzoid> what did?
[15:23:50] <deep42thought> well, here it complained about already installed functions
[15:23:58] <deep42thought> but on your server, it simply worked
[15:24:05] <tyzoid> oh, yeah
[15:24:06] <deep42thought> (I mean _before_ that commit)
[15:24:09] <tyzoid> because it couldn't find the file
[15:24:16] <deep42thought> ah, ok
[15:24:24] <deep42thought> that explains it :-)
[15:24:31] <tyzoid> require will error on file missing, include will not
[15:24:51] <tyzoid> hence why I usually use require or require_once
[15:26:15] <deep42thought> yeah, makes sense
[16:09:29] -!- abaumann has joined #archlinux32
[16:09:30] <buildmaster> Hi abaumann!
[16:09:34] <deep42thought> Hi Andreas!
[16:09:48] <abaumann> Hi. :-)
[16:10:43] <abaumann> mmh. https://packages.archlinux32.org shows ospray as being on the deletion-list.
[16:10:45] <phrik> Title:Arch Linux 32 - embree 3.2.0-1.0 (i686) (at packages.archlinux32.org)
[16:11:01] <abaumann> embree2 the same.
[16:11:11] <abaumann> Whouldn't they be built now?
[16:11:26] <deep42thought> 'i686 ospray "Your Embree build needs to support at least one ISA >= SSE4.1!"'
[16:11:29] <deep42thought> it is blacklisted
[16:11:32] <abaumann> oh.
[16:11:33] <abaumann> :-)
[16:12:07] <abaumann> aha. embree and embree are available now, but none of the packages above can make use of them. :-)
[16:12:13] <abaumann> *embree and embree2
[16:12:16] <deep42thought> idk
[16:12:24] <deep42thought> maybe they _can_ be built now ...
[16:13:15] <abaumann> I could try to increase the ISA, but find a 32-bit CPU with SSE4.1 :-)
[16:13:32] <deep42thought> well, if you run on 64bit hardware?
[16:13:56] <abaumann> running an archlinux32 on 64-bit? well. who would do such a thing?
[16:14:09] <abaumann> maybe on a weak atom or so..
[16:14:15] <deep42thought> actually, I do
[16:14:19] <abaumann> oh. :-)
[16:14:19] <deep42thought> but mostly due to laziness
[16:14:24] <abaumann> lol
[16:14:26] <deep42thought> I assumed it was 32 bit
[16:14:37] <abaumann> ah. that one. I remember..
[16:14:48] <deep42thought> ... happened on two boxes!
[16:14:55] <abaumann> well. 64-bit cpu and only 2GB RAM would also be a reason to run 32-bit.
[16:16:37] <abaumann> i can try embree and embree with higher isa's. I admin SSE2 was just the lowest I could safely choose.. but I didn't test with more..
[16:16:42] <abaumann> *admit
[16:17:19] <deep42thought> well, I guess it's just a corner case
[16:17:27] <deep42thought> don't waste too much time on it :-)
[16:18:07] <abaumann> it's just a makepkg and a flag. and I can add comments, so we remember laster..
[16:18:34] <deep42thought> ok
[16:38:27] -!- deep42thought has quit [Quit: Leaving.]
[16:40:19] -!- abaumann has quit [Quit: leaving]
[17:02:46] -!- rcf has quit [Read error: Connection reset by peer]
[17:03:35] -!- rcf has joined #archlinux32
[17:36:44] -!- isacdaavid has joined #archlinux32
[18:10:36] <buildmaster> girls, my database is dirty again ...
[19:06:40] -!- musty26069 has joined #archlinux32
[19:19:30] * buildmaster resumes sanity.
[19:58:32] <buildmaster> haskell-hopenpgp is broken (says eurobuild3). - I rescheduled: haskell-binary-conduit.
[20:21:54] <buildmaster> embree2 is broken (says tyzoid-srv0-bs0).
[20:47:15] -!- deep42thought has joined #archlinux32
[20:47:16] <buildmaster> Hi deep42thought!
[20:47:50] <deep42thought> just in case anyone wonders: I'm force-moving (hopefully only) haskell-* packages around to satisfy some dependencies :-)
[21:01:31] <tyzoid> sounds like a plan
[21:02:11] <tyzoid> deep42thought: I'm planning on working on this a bit tonight, so I would like your thoughts on it
[21:02:19] <tyzoid> for https://accounts.archlinux32.org
[21:02:20] <phrik> Title:Arch Linux 32 (at accounts.archlinux32.org)
[21:03:16] <tyzoid> Essentially, for services that need authentication, I'm thinking you can redirect to accounts.archlinux32.org/?service=<subdomain>&url=/path/to/resource
[21:03:28] <tyzoid> mostly to try to avoid redirecting out of domain
[21:03:56] <tyzoid> it'll set the user cookie, then redirect them
[21:39:25] <deep42thought> I'm not sure, I fully understand - what is "/path/to/resource"?
[21:39:56] <deep42thought> don't we have maximal one login service per subdomain?
[21:40:16] <tyzoid> let's say you have an authenticated resource (page, action, etc.) at buildmaster.archlinux32.org/admin/some_admin_page.php
[21:40:33] <tyzoid> service would be 'buildmaster' and url would be /admin/some_admin_page.php
[21:40:49] <tyzoid> so it knows where to redirect back to once auth is sucessful
[21:40:50] <deep42thought> ah, ok
[21:42:00] <deep42thought> I think, that sounds good
[21:42:53] <deep42thought> so this php would check for the cookie and would redirect if it is not yet present - right?
[21:43:06] <tyzoid> essentially.
[21:43:52] <tyzoid> just redirect the user to accounts.archlinux32 with service and url, and it'll prompt them to log in
[21:44:05] <tyzoid> then redirect the user back to you.
[21:44:18] <tyzoid> On your end, you can call the API and see what rights they have
[21:44:22] <tyzoid> oh - that's the other thing
[21:44:30] <tyzoid> I need to add permissions to the system
[21:46:09] <deep42thought> I suppose, you checked that this can be (easily) integrated with the bug tracker and forum
[21:46:37] <tyzoid> Bugtracker I can deal with, and it's using the forum's auth system for now
[21:46:47] <tyzoid> so it's essentially taking the forum data and extending it
[21:47:01] <deep42thought> that's not what I meant
[21:47:23] <deep42thought> question is if the forum and bugtracker can query your api
[21:47:40] <tyzoid> forum doesn't need to, and I can get the bugtracker to without too much difficulty
[21:48:04] <deep42thought> the forum does not need to, because I get the usual forum session cookies, too?
[21:48:25] <tyzoid> the forum session cookie _is_ the archlinux32.org auth cookie xD
[21:48:36] <tyzoid> hence 'extending the forum'
[21:48:45] <deep42thought> ah, ok :-)
[21:48:59] <tyzoid> Forum already has all the users already
[21:49:11] <deep42thought> plus a few spammers ;-)
[21:49:16] <tyzoid> Yup
[21:49:37] <tyzoid> that bit doesn't matter since all other permissions will be manuallly approved.
[21:50:15] <deep42thought> yeah, np :-)
[21:52:45] <elibrokeit> deep42thought: is haskell-hopenpgp broken on another circular dep?
[21:54:06] <deep42thought> elibrokeit: I don't know
[21:54:27] <deep42thought> I just tried to move haskell packages around until less dependency issues were in the database
[21:54:36] <deep42thought> and also rebuilt some haskell packages in the progress
[21:54:44] <deep42thought> maybe this is part of the fallout :-/
[21:56:41] <deep42thought> looks like it just needs to be rebuilt
[22:00:09] <deep42thought> elibrokeit: solved :-)
[22:01:59] <elibrokeit> oh, good
[22:30:19] <deep42thought> whoah, fascinating - removing bugs from db-update helps solving the haskell dependency issues :-)
[22:40:57] -!- musty26069 has quit [Ping timeout: 240 seconds]
[22:50:33] -!- isacdaavid has quit [Quit: Leaving.]
[22:59:17] <tyzoid> deep42thought: Consider it a test case :P
[23:07:23] <deep42thought> testing in production :-)
[23:08:14] <tyzoid> https://i.imgur.com
[23:08:21] <deep42thought> yeah
[23:08:23] <tyzoid> "Stay Buggy, My Friends"
[23:08:28] <deep42thought> didn't find a good one of that :-)
[23:08:56] <tyzoid> Don't know if you've seen them, but that's a screenshot from the "The Most Interesting Man In The World" commercials
[23:10:10] <deep42thought> no, that's new to me
[23:11:40] <tyzoid> The commercial ends with "I don't always drink beer, but when I do, I drink <brand>. Stay Thirsty My Friends."
[23:12:10] * deep42thought is reading its wikipedia article at the moment
[23:18:18] <deep42thought> I'm not very educated with commercials - not even the german ones ;-)
[23:40:28] -!- deep42thought has quit [Quit: Leaving.]