#archlinux-ports | Logs for 2017-10-11
Back
[00:05:15] -!- isacdaavid has joined #archlinux-ports
[00:13:56] -!- isacdaavid has quit [Quit: isacdaavid]
[00:14:12] -!- isacdaavid has joined #archlinux-ports
[01:19:32] -!- isacdaavid has quit [Read error: Connection reset by peer]
[01:19:48] -!- isacdaavid has joined #archlinux-ports
[02:35:42] -!- isacdaavid has quit [Quit: isacdaavid]
[03:03:03] <tyzoid> hey deep42thought, when you hop on, can you add me to the releng git repo?
[03:03:17] <tyzoid> the 'dev' team doesn't have access to it yet.
[03:17:28] -!- tyzoid has quit [Ping timeout: 240 seconds]
[03:19:47] -!- tyzoid has joined #archlinux-ports
[04:25:12] -!- tyzoid has quit [Ping timeout: 260 seconds]
[04:31:29] -!- isacdaavid has joined #archlinux-ports
[04:32:29] -!- tyzoid has joined #archlinux-ports
[04:34:52] -!- tyzoid has quit [Read error: Connection reset by peer]
[04:38:52] -!- tyzoid has joined #archlinux-ports
[05:26:47] -!- tyzoid has quit [Ping timeout: 255 seconds]
[05:28:56] -!- tyzoid has joined #archlinux-ports
[05:33:39] -!- isacdaavid_ has joined #archlinux-ports
[05:33:41] -!- isacdaavid has quit [Ping timeout: 240 seconds]
[06:00:05] -!- tyzoid has quit [Ping timeout: 240 seconds]
[07:33:50] -!- deep42thought has joined #archlinux-ports
[07:47:13] <deep42thought> tyzoid: done, I noticed it myself (your pull request turned my attention to it)
[08:49:21] -!- deep42thought has quit [Quit: Leaving.]
[08:54:12] -!- isacdaavid_ has quit [Quit: isacdaavid_]
[11:19:00] -!- deep42thought has joined #archlinux-ports
[14:14:54] -!- tyzoid has joined #archlinux-ports
[14:16:41] <tyzoid> deep42thought: So what's the mirror URL that includes x86?
[14:17:13] <tyzoid> x86_64, that is
[14:23:09] -!- Polichronucci has joined #archlinux-ports
[14:26:20] <deep42thought> tyzoid: I deleted it again
[14:26:24] <deep42thought> should I set it up again?
[14:27:19] <deep42thought> have a look at the current dual-boot and master branches of archiso32
[14:27:20] <tyzoid> If you want me to test with it... I think a better strategy is to alt-root pacman, and build a local copy of the repo
[14:28:30] <deep42thought> ah, right, we want on the iso our packages too, not just our mirror list
[14:28:31] <deep42thought> sry
[14:28:35] <deep42thought> give me a sec
[14:28:58] <deep42thought> done
[14:29:10] <deep42thought> https://mirror.archlinux32.org
[14:29:10] <phrik> Title: Index of /x86_64/ (at mirror.archlinux32.org)
[14:31:28] <deep42thought> the archiso32 and archiso-dual version 32.2 should include the patch I sent you
[14:31:38] <deep42thought> so they should set up the correct mirror lists inside the iso
[14:31:56] <deep42thought> that's why I thought we would not need the merged mirror, sry
[14:32:15] -!- tyzoid has quit [Ping timeout: 260 seconds]
[14:56:23] -!- deep42thought has quit [Quit: Leaving.]
[15:06:50] -!- tyzoid has joined #archlinux-ports
[15:09:46] <tyzoid> deep42thought: sorry, we had fire alarms go off in the building
[15:11:22] <tyzoid> Alright, I see the changes on github. Did you make a new version on your mirror?
[15:14:54] <tyzoid> Running a test to see if the updated mirrorlist is enabled.
[15:15:45] <tyzoid> So here's the big question: for i686, it seems pretty easy to change pacman.conf to just point to our repos
[15:15:54] <tyzoid> but that's not so easy for -dual
[15:16:29] <tyzoid> so, my proposal is to point both at a local folder
[15:16:52] <tyzoid> which would be a local repo generated by a script
[15:17:39] <tyzoid> and that script (could be a stage of build.sh, or external requirement) would download the latest version of all the packages in packages.both/packages.<arch> and repo-add them in the relevant locations
[15:33:33] <fsckd> tyzoid: sorry i didn't answer your question (re reproducible builds). yes, upstream has to be patched, especially build tools which set timestamps and stuff. seems patches have been pushed to get projects to use SOURCE_DATE_EPOCH when set. https://wiki.debian.org
[15:33:34] <phrik> Title: ReproducibleBuilds/TimestampsProposal - Debian Wiki (at wiki.debian.org)
[15:49:16] <tyzoid> fsckd: Thanks, I had seen that before. Is there a script/build tool being developed that allows users to take advantage of reproducible builds and verify the source themselves?
[15:59:15] <tyzoid> deep42thought: That patch didn't work, $work/$arch/airootfs still has mainline mirrorlist
[16:12:02] <brtln> so rebuilding a package with packages specified in .PKGINFO should give you the same hash
[16:22:45] <tyzoid> brtln: But that still means that I need to know the source used to do the build, which may not necessarily be available.
[16:24:02] <tyzoid> For ex, if they have a git repo, does the .PKGINFO store the commit head used to build?
[16:38:46] -!- AndrevS has joined #archlinux-ports
[16:39:03] <brtln> I'm not someone who knows, sorry
[16:39:16] <brtln> I'm the deprecation guy :P
[17:12:30] -!- guys has joined #archlinux-ports
[17:17:23] <eschwartz[m]> tyzoid: .PKGINFO can be extracted from the *package* being reproduced, and we should assume that the epoch:pkgver-pkgrel is always unique
[17:18:00] -!- deep42thought has joined #archlinux-ports
[17:18:19] <eschwartz[m]> If not, we're screwed anyway, but as it is we just need to trust that published PKGBUILDs are really the ones that were used. Then again, that's rather (part of) the point of reproducibility :p
[17:22:19] -!- eschwartz has joined #archlinux-ports
[17:22:25] <deep42thought> published PKGBUILDs are sometimes not the ones used - or at least, building is sometimes more involved than running "*-x86_64-build" in the right directory
[17:23:28] <deep42thought> tyzoid: which version do you use?
[17:23:30] <deep42thought> 32.2?
[17:24:37] <eschwartz> deep42thought: but that's usually just an issue of reproducing it with later versions of dependent packages. ;)
[17:25:08] <deep42thought> eschwartz: sometimes it's also an issue with local changes the maintainer did and forgot to publish
[17:25:21] <eschwartz> ohhh
[17:25:29] <eschwartz> that's bad and they should feel bad.
[17:25:56] <eschwartz> (which granted they probably do as most maintainers tend to be responsible and well-meaning people)
[17:26:12] <eschwartz> (alad excepted)
[17:29:23] <deep42thought> tyzoid: hopefully I have some time this evening (in approx. 4-5h) to have a look, what's going on
[17:47:51] -!- guys has quit [Ping timeout: 246 seconds]
[17:47:59] -!- eschwartz has quit [Read error: Connection reset by peer]
[17:48:26] -!- guys has joined #archlinux-ports
[17:51:28] -!- eschwartz has joined #archlinux-ports
[17:55:33] <tyzoid> deep42thought: It re-downloads the package from your repo each time, since each iteration I'm doing a whole vm-create, vm-run, vm-destroy cycle.
[17:55:48] <tyzoid> so it should be the latest in an absolutely spotless environment
[17:55:59] <tyzoid> but I'll double check in just a sec
[17:56:06] <deep42thought> ok, so I'll test with the most recent version then
[17:58:12] <tyzoid> deep42thought: I'm using that script I committed
[17:58:22] <tyzoid> so if you use that, it'll initialize the same VM I've got over here
[17:59:10] <deep42thought> I'll try that and look what it does
[17:59:20] <tyzoid> (just make sure to comment out the vagrant destroy -f line at the end, if you want to mess around with it after the iso is made
[18:00:19] <tyzoid> deep42thought: archiso32-32.2-1 is what it's installing
[18:00:38] <tyzoid> I verified that the line in build.sh did change too
[18:00:43] <deep42thought> ok
[18:00:51] <deep42thought> I'll try it outside of vagrant, then
[18:00:53] <tyzoid> but the actual mirrorlist found in $work/$arch/airootfs didn't change
[18:00:58] <deep42thought> this should yield the same behaviour
[18:01:08] <deep42thought> I'll investigate
[18:01:10] <tyzoid> should
[18:01:12] <tyzoid> thanks
[18:01:24] <deep42thought> later ... :-)
[18:01:27] <tyzoid> no need to rush on this, it's more of a nice-to-have before next month rolls around :P
[18:01:33] <tyzoid> cya, deep42thought
[18:01:37] <deep42thought> cu
[18:15:03] -!- isacdaavid has joined #archlinux-ports
[18:20:13] -!- isacdaavid has quit [Ping timeout: 258 seconds]
[18:26:45] -!- isacdaavid has joined #archlinux-ports
[18:30:13] -!- eschwartz has quit [Ping timeout: 255 seconds]
[18:30:14] -!- guys has quit [Read error: Connection reset by peer]
[18:40:16] -!- guys has joined #archlinux-ports
[18:49:07] -!- eschwartz has joined #archlinux-ports
[19:24:07] -!- guys has quit [Read error: Connection reset by peer]
[19:24:07] -!- eschwartz has quit [Read error: Connection reset by peer]
[19:24:18] -!- guys has joined #archlinux-ports
[19:26:04] -!- eschwartz has joined #archlinux-ports
[20:00:30] -!- guys has quit [Read error: Connection reset by peer]
[20:00:31] -!- eschwartz has quit [Read error: Connection reset by peer]
[20:00:43] -!- guys has joined #archlinux-ports
[20:02:20] -!- eschwartz has joined #archlinux-ports
[20:45:44] -!- guys has quit [Read error: Connection reset by peer]
[20:45:44] -!- eschwartz has quit [Read error: Connection reset by peer]
[20:46:36] -!- guys has joined #archlinux-ports
[20:48:13] -!- eschwartz has joined #archlinux-ports
[21:24:43] -!- eschwartz has quit [Read error: Connection reset by peer]
[21:24:57] -!- guys has quit [Read error: Connection reset by peer]
[21:25:18] -!- guys has joined #archlinux-ports
[21:27:54] -!- eschwartz has joined #archlinux-ports
[21:28:20] <deep42thought> tyzoid: the new archiso32 and archiso-dual should install the correct packages and mirror list
[21:28:23] <deep42thought> v32.3
[21:29:08] <deep42thought> it is necessary to create /var/cache/archbuild32, though, because the package caches are now separated, too
[21:29:19] <deep42thought> one possibility to do so easily is to install devtools32 :-/
[21:29:45] <deep42thought> ah no, it does not create that directory
[21:29:54] <deep42thought> it=devtools32
[21:30:32] <tyzoid> deep42thought: Did you test with the virtualmachine?
[21:30:39] <tyzoid> or should I run that test
[21:30:47] <deep42thought> no, but I saw your problem on a real machine, too
[21:31:02] <deep42thought> it would be nice if you could do the test
[21:31:14] <deep42thought> especially what happens if the cache dir does not exist (yet)
[21:31:32] <tyzoid> okay, running.
[21:31:43] <tyzoid> takes ~5min
[21:31:46] <deep42thought> ok
[21:31:49] <tyzoid> ping me if I don't get back to you within 10
[21:31:58] <deep42thought> I'll get ready for bed now and ask you when I'm done :-)
[21:32:10] <tyzoid> sounds good
[21:38:26] <tyzoid> the iso build routine exited with nonzero status during stage one
[21:38:33] <deep42thought> :-/
[21:38:33] <tyzoid> re-running and seeing if the problem happens again
[21:38:55] -!- kl1602 has joined #archlinux-ports
[21:40:29] <deep42thought> I had sometimes problems with poisoned caches
[21:40:50] <deep42thought> I hoped this would vanish with different caches for i686 and x86_64, though
[21:41:04] <tyzoid> yeah, it did the same thing a second time
[21:41:09] <tyzoid> something broke, there
[21:41:14] <deep42thought> any specific error?
[21:41:18] <tyzoid> since I'm instantiating a completely new vm each time
[21:41:26] <tyzoid> no errors outputted, just a nonzero status code
[21:41:29] <deep42thought> ok, not surprising, then
[21:41:37] <deep42thought> can you pass "-v" to build.sh?
[21:41:42] <tyzoid> sure, one sec
[21:41:48] <tyzoid> gotta destroy + rerun
[21:42:02] <deep42thought> I guess, it's missing /var/cache/archbuild32
[21:42:39] <tyzoid> running
[21:45:31] <tyzoid> 'error: failed to update core (no servers configured for repository)'
[21:45:39] <tyzoid> deep42thought: same error for the other repos too
[21:45:39] <deep42thought> gnaa
[21:45:47] <tyzoid> did you forget to uncomment a repo?
[21:45:56] <deep42thought> you need one mirror in /etc/pacman.d/mirrorlist32
[21:46:02] <deep42thought> which is installed with comments only
[21:46:27] <deep42thought> I think, it's more your job after installing the packages in the vm
[21:46:58] <deep42thought> e.g. I run this on some arbitrary box with devtools32 installed (which also build i686 packages) and therefor already has a mirror uncommented
[21:47:20] <deep42thought> thus I think, I shouldn't mess up /etc/pacman.d/mirrorlist32
[21:47:24] <tyzoid> okay, I can add that into the script
[21:47:30] <tyzoid> but the question is where this gets pulled in from
[21:47:39] <tyzoid> is it when installing the packages that archiso32 depends on it?
[21:47:46] <deep42thought> yes
[21:47:48] <tyzoid> or is it build.sh itself that downloads it?
[21:47:55] <deep42thought> archiso32 now depends on pacman-mirrorlist32
[21:48:27] <tyzoid> okay
[21:48:30] <deep42thought> you can simply place a /etc/pacman.d/mirrorlist32 with one mirror enabled inside your vm (even before installing the packages)
[21:48:47] <tyzoid> okay
[21:48:56] <tyzoid> I'll do that tonight, and get back to you.
[21:49:10] <deep42thought> ok, I'll go to bed now :-)
[21:49:15] <tyzoid> it looks like my ssh tunnel collapsed now
[21:49:17] <tyzoid> good night then
[21:49:58] <deep42thought> good night!
[21:50:01] -!- deep42thought has quit [Quit: Leaving.]
[22:02:33] -!- eschwartz has quit [Read error: Connection reset by peer]
[22:02:36] -!- guys has quit [Read error: Connection reset by peer]
[22:05:15] -!- guys has joined #archlinux-ports
[22:06:22] -!- eschwartz has joined #archlinux-ports
[22:32:56] phrik is now known as PHRIK
[22:38:11] -!- eschwartz has quit [Remote host closed the connection]
[22:39:38] <tyzoid> !ping
[22:39:38] <phrik> pong
[22:53:08] -!- kl1602 has quit [Quit: Leaving]
[22:58:35] -!- tyzoid has quit [Ping timeout: 260 seconds]
[23:13:22] -!- guys has quit [Read error: Connection reset by peer]
[23:13:34] -!- guys has joined #archlinux-ports
[23:15:42] -!- eschwartz has joined #archlinux-ports
[23:17:59] -!- isacdaavid has quit [Ping timeout: 246 seconds]
[23:21:44] -!- isacdaavid has joined #archlinux-ports
[23:26:21] -!- AndrevS has quit [Remote host closed the connection]
[23:51:48] -!- eschwartz has quit [Ping timeout: 258 seconds]
[23:57:15] -!- eschwartz has joined #archlinux-ports