#archlinux-ports | Logs for 2017-10-05

Back
[04:00:01] -!- ianaca has quit [Ping timeout: 240 seconds]
[04:02:06] -!- ianaca has joined #archlinux-ports
[04:27:13] -!- Faalagorn has quit [Quit: Leaving.]
[06:16:24] -!- deep42thought has joined #archlinux-ports
[07:21:48] -!- deep42thought has quit [Quit: Leaving.]
[07:27:27] -!- isacdaavid has quit [Quit: isacdaavid]
[08:29:01] -!- deep42thought has joined #archlinux-ports
[11:18:53] -!- dantob has joined #archlinux-ports
[11:56:36] -!- dantob has quit [Quit: dantob]
[13:25:34] -!- Faalagorn has joined #archlinux-ports
[13:51:50] <deep42thought> brtln: I just stumbled over the google-api-keys in firefox' PKGBUILD - am I right in assuming, that we have permission to use the archlinux api keys?
[14:09:42] -!- tyzoid_ has joined #archlinux-ports
[14:09:59] -!- tyzoid has quit [Disconnected by services]
[14:10:03] tyzoid_ is now known as tyzoid
[14:10:08] <tyzoid> hey deep42thought
[14:10:21] <deep42thought> Hi tyzoid
[14:10:24] -!- tyler has joined #archlinux-ports
[14:10:39] <tyzoid> I'm having trouble finding your public key for the archlinuxewe repo
[14:10:47] tyler is now known as Guest6223
[14:10:58] <deep42thought> https://arch.eckner.net
[14:12:06] <tyzoid> hmm
[14:12:15] <tyzoid> Can't believe I missed that again :/
[14:12:19] <tyzoid> thanks!
[14:12:20] <deep42thought> np :-)
[14:14:24] <tyzoid> deep42thought: I'm going through the process of automating the building of fresh iso build-vms, so that I have a clean environment to automate the build procedure
[14:14:39] <deep42thought> nice
[14:14:53] <deep42thought> you can even include the update of the website in that process :-)
[14:15:05] <tyzoid> at some point, yeah
[14:15:11] <tyzoid> I just found out that Archlinux mainline now provides x86_64 vagrant machines
[14:15:15] <tyzoid> so I'm basing it off of those
[14:16:48] <deep42thought> ah, you forked archbox to gtihub.com/archlinux32 ?
[14:17:15] <deep42thought> arch-boxes
[14:18:49] <tyzoid> Yeah
[14:18:54] <tyzoid> that builds it from the iso
[14:19:22] <tyzoid> but I'm going to eventually modify that to build i686 boxes from our iso
[14:19:31] <deep42thought> :-)
[14:20:51] <tyzoid> Is there a pacman flag that automatically accepts the prompts?
[14:21:01] <deep42thought> --noconfirm
[14:22:06] <tyzoid> thanks. Hard to navigate the manpage on my phone :P
[14:22:12] <deep42thought> lol
[14:22:18] <deep42thought> too few keys ;-)
[14:22:37] <tyzoid> Nah, just the screen size, since it's in portrait
[14:22:46] <tyzoid> once you gave it, I could slash search
[14:22:59] <tyzoid> I've got a full qwerty keyboard installed, which has arrow keys, control keys, etc
[14:23:11] <tyzoid> but it's still difficult to see stuff sometimes
[14:30:10] <tyzoid> deep42thought: So another question. Do you have a decent solution to the need-to-change-pacman.conf-in-middle-of-iso-build issue?
[14:30:20] <tyzoid> specifically on the -dual builds
[14:30:42] <deep42thought> sry, forgot about that - remind me: why is it needed?
[14:31:00] <tyzoid> archiso-dual assumes all packages are available from one mirror/source
[14:31:12] <tyzoid> but i686 and x86_64 are from different mirrors
[14:31:15] <deep42thought> hmm
[14:31:32] <deep42thought> you can pass different pacman.confs via command line options
[14:31:44] <deep42thought> maybe, we should just hard-define them in the build script?
[14:31:45] <tyzoid> My initial thought is to create pacman.64.conf and pacman.32.conf, and modify build.sh to switch between the two
[14:31:53] <deep42thought> right
[14:32:11] <tyzoid> That's the best I could come up with
[14:32:13] <deep42thought> what would be the difference between those? the cache dir and the mirror list, right?
[14:32:32] <tyzoid> how I've been doing it, is that the pacman.conf actually specifies the mirror itself
[14:32:42] <tyzoid> so instead of referring to a mirrorlist, it's got lines like
[14:32:43] <deep42thought> although cache would be not that critical, problably - as long as x86_64 packages are used for "any"
[14:32:45] <tyzoid> [core]
[14:32:51] <tyzoid> Server = https://...
[14:32:54] <deep42thought> yeah
[14:33:05] <deep42thought> you can uste /etc/pacman.d/mirrorlist32 for that
[14:33:16] <deep42thought> (it's what devtools32 use)
[14:33:21] <tyzoid> It doesn't matter much to me, since I use mine (which is geographically close)
[14:33:33] <tyzoid> but I'll do that for a modification I submit back
[14:34:03] <tyzoid> Unless you can think of a better alternative, I'll work on this for next month's iso release
[14:34:10] <deep42thought> maybe we should really set up one or two "mixed" mirrors
[14:34:32] <tyzoid> I think it's more effort than it's worth
[14:34:43] <tyzoid> since the way mainline organizes it's packages are different from ours
[14:34:48] <deep42thought> I could just include an x86_64 directory on the main mirror with symlinks only to the original repositories
[14:35:06] <deep42thought> one directory with like 8 symlinks inside is not much effort
[14:35:18] <tyzoid> yeah, but you'd also have to do a repo-add again, since the paths change
[14:35:32] <deep42thought> no
[14:35:41] <deep42thought> ?
[14:36:16] <tyzoid> if you move around those files, won't the indexes be out of date?
[14:36:35] <deep42thought> the indexes will be symlinked the same as the packages
[14:36:45] <deep42thought> e.g. the whole repository folder will be symlinked
[14:37:12] <deep42thought> e.g. x86_64/core -> /path/to/my/upstream/mirror/core/os/x86_64
[14:37:29] <deep42thought> I do a similar thing with archlinuxarm on my upstream mirror
[14:38:05] <deep42thought> core/os/armv7h -> /srv/arch-mirror/arch/arch/archlinuxarm/armv7h/core
[14:38:26] <brtln> deep42thought: better ask heftig
[14:38:31] <tyzoid> right, but x86_64/core is actually core/x86_64
[14:38:38] <tyzoid> because the structure is inverted from ours
[14:38:41] <deep42thought> brtln: will do
[14:38:57] <tyzoid> that's where I think a problem lies
[14:38:59] <deep42thought> tyzoid: no problem, just one symlink per repo
[14:39:05] <tyzoid> if you don't see this as a problem, then we can try it
[14:39:14] <deep42thought> it's the same issue between x86_64 and armv7h
[14:39:20] <deep42thought> yeah
[14:39:21] <deep42thought> will do
[14:39:25] <tyzoid> I'd rather not put it on the master, of course.
[14:39:55] <deep42thought> one could set up a directive in nginx, so it is served as it was symlinked
[14:39:57] <deep42thought> hmmm
[14:40:19] <deep42thought> ah no, then all mirrors syncing by http(s) will copy x86_64, too :-(
[14:41:22] <tyzoid> I have mine set so that rsync won't copy files outside of the mirror path
[14:41:26] <tyzoid> but you can serve it outside
[14:41:35] <tyzoid> iirc that's the --safe-links option
[14:41:55] <tyzoid> but not sure if everyone else is using the same recommended setting
[14:42:04] <tyzoid> and of course, syncing by http is an issue
[14:42:30] <tyzoid> deep42thought: Can you add a CNAME from status.archlinux32.org to stats.uptimerobot.com?
[14:42:42] <deep42thought> ok
[14:43:08] * jelle wonders if there is a foss uptimerobot.com variant
[14:43:48] <tyzoid> I don't think so, since it's a service.
[14:44:09] <tyzoid> There is open source software to do it, but then you need to maintain and guarantee uptime of that
[14:44:17] <tyzoid> which defeats the whole purpose of a status monitor
[14:46:27] <deep42thought> done, you need a valid certificate, though
[14:47:01] <tyzoid> nope
[14:47:22] <deep42thought> I was talking about status.archlinux32.org
[14:47:30] <tyzoid> yeah, it automatically generates it
[14:47:36] <tyzoid> did you set up the cname correctly?
[14:47:54] <deep42thought> status.archlinux32.org. 486 IN CNAME stats.uptimerobot.com.
[14:48:02] <tyzoid> hmm, looks right. one sec
[14:49:25] <tyzoid> lol, it's a caching issue on my end
[14:49:33] <tyzoid> deep42thought: https://status.archlinux32.org seems to work
[14:49:49] <tyzoid> uptimerobot gets a cert from letsencrypt automatically
[14:50:01] <deep42thought> ah, I was too impatient :-)
[14:50:47] <tyzoid> deep42thought: Any other services we should monitor there?
[14:50:53] <deep42thought> master mirror
[14:50:56] <deep42thought> buildmaster
[14:51:09] <tyzoid> http(s) url for that?
[14:51:12] <deep42thought> maybe sources.archlinux32.org
[14:51:14] <tyzoid> or is it a host to ping
[14:51:23] <deep42thought> buildmaster.archlinux32.org
[14:51:31] <deep42thought> mirror.archlinux32.org
[14:51:36] <deep42thought> sources.archlinux32.org
[14:51:39] <deep42thought> :-)
[14:51:50] <deep42thought> but the latter is really optional
[14:52:02] <deep42thought> if that one is down, I have black screens in front of me ;-)
[14:52:48] <tyzoid> is https://buildmaster.archlinux32.org run off of the buildmaster?
[14:52:48] <phrik> Title: Buildmaster for Archlinux32 packages (at buildmaster.archlinux32.org)
[14:52:53] <deep42thought> yes
[14:54:27] <deep42thought> you can even check for the content of https://buildmaster.archlinux32.org
[14:54:44] <deep42thought> ... all lines should end with ' ... passed.'
[14:54:59] <tyzoid> uptimerobot doesn't allow me to do that fine grained of a check
[14:55:04] <deep42thought> ah, ok
[14:55:05] <deep42thought> np
[14:55:07] <tyzoid> it only verifies that it's up and records response time
[15:01:59] <deep42thought> does it help if I provide a webpage which errors if the sanity-check fails?
[15:02:46] <tyzoid> actually...
[15:02:53] <tyzoid> what will it say if there's a broken one?
[15:03:19] <deep42thought> who is "it" and who is "one"?
[15:04:00] <tyzoid> it: https://buildmaster.archlinux32.org
[15:04:06] <tyzoid> one = santiy check
[15:04:12] <tyzoid> sanity*
[15:04:31] <tyzoid> deep42thought: It says 'passed' if it succeeded, what's the failed message?
[15:04:38] <deep42thought> instead of "passed" it will say "failed" and then a long list of thinks which it considers wrong
[15:05:04] <tyzoid> So apparantly, I can set it up to verify that the word 'failed' doesn't appear at all
[15:05:27] <deep42thought> wait, let me check, that that's actually the case
[15:06:00] <deep42thought> ah, no, it just reports the errors :-/
[15:06:42] <deep42thought> https://buildmaster.archlinux32.org
[15:07:04] <deep42thought> but inserting "FAILED" is trivial :-)
[15:07:29] <tyzoid> sounds good. All caps?
[15:07:35] <tyzoid> the keyword is case sensitive
[15:07:41] <deep42thought> yes
[15:08:21] <tyzoid> just to test it, can you add FAILED somewhere on the page?
[15:08:23] <deep42thought> have a look: https://buildmaster.archlinux32.org
[15:09:00] <tyzoid> cool. It takes up to 5 minutes (when the next check happens)
[15:09:08] <tyzoid> so I'll let you know if it works or didn't
[15:14:26] <tyzoid> deep42thought: Yup, the keyword check seems to be working now
[15:14:34] <deep42thought> :-)
[15:14:39] <tyzoid> did you get the email?
[15:14:41] <deep42thought> no
[15:14:50] <tyzoid> hmm
[15:14:53] <tyzoid> odd
[15:14:55] <deep42thought> did you?
[15:14:57] <tyzoid> yeah
[15:15:11] <deep42thought> maybe another grey listing?
[15:15:15] <tyzoid> could be
[15:15:18] <deep42thought> is the sender the same?
[15:15:25] <tyzoid> yeah
[15:15:31] <tyzoid> or maybe not
[15:15:34] <deep42thought> then it shouldn't be greylisted again
[15:15:38] <tyzoid> it comes from alert@uptimerobot.com
[15:15:39] <deep42thought> ah, too slow
[15:15:43] <tyzoid> so if that's not the same
[15:15:48] <deep42thought> it's not the same
[15:15:48] <tyzoid> it came in?
[15:15:51] <tyzoid> ah
[15:15:51] <deep42thought> no
[15:15:52] <tyzoid> that's why
[15:15:56] <deep42thought> i was too slow typing
[15:16:24] <tyzoid> anyway, you can bring the sanity check back to the 'correct' listing, as the test was successful
[15:16:32] <deep42thought> I'm on it :-)
[15:17:07] <tyzoid> yup, just got email saying it's back up
[15:17:14] <deep42thought> I'm a little paranoid about such things, but the sanity-check already safed the master mirror a few times ...
[15:17:29] <tyzoid> having an email alert about it will be nice too
[15:17:35] <deep42thought> absolutely
[15:20:50] <tyzoid> deep42thought: Also, I can generate read-only API keys for each monitor
[15:20:58] <tyzoid> so if you want to play around with that, let me know
[15:22:00] <deep42thought> so I could set up my own alert for a subset of the infra=
[15:22:01] <deep42thought> ?
[15:22:34] <tyzoid> sort of
[15:22:46] <tyzoid> I have you set up to get emails for all the monitors
[15:22:49] <tyzoid> I can change that if you want
[15:22:55] <deep42thought> no, it's ok
[15:22:56] <tyzoid> but the API key would be for more stats on it
[15:23:03] <deep42thought> ah, ok
[15:23:13] <deep42thought> I think, I don't need that
[15:23:16] <deep42thought> thanks
[15:23:17] <deep42thought> :-)
[16:19:38] -!- deep42thought has quit [Quit: Leaving.]
[16:41:45] -!- p71 has quit [Ping timeout: 258 seconds]
[16:46:31] -!- p71 has joined #archlinux-ports
[17:36:34] -!- deep42thought has joined #archlinux-ports
[18:27:11] -!- isacdaavid has joined #archlinux-ports
[18:41:06] -!- p71 has quit [Read error: Connection reset by peer]
[19:06:56] -!- deep42thought has quit [Quit: Leaving.]
[19:21:37] -!- Faalagorn has quit [Ping timeout: 248 seconds]
[19:34:52] -!- vboxler has joined #archlinux-ports
[19:35:18] -!- vboxler has quit [Client Quit]
[19:54:52] -!- vboxler has joined #archlinux-ports
[19:55:06] -!- vboxler has quit [Client Quit]
[21:41:51] -!- shentey has joined #archlinux-ports
[21:47:24] -!- deep42thought has joined #archlinux-ports
[21:50:33] -!- shentey has quit [Quit: Konversation terminated!]
[21:52:15] -!- shentey has joined #archlinux-ports
[22:07:56] -!- tyzoid has quit [Quit: Page closed]
[23:02:00] -!- shentey has quit [Quit: Konversation terminated!]