#archlinux32 | Logs for 2018-11-05

[01:30:35] -!- DCyrax has joined #archlinux32
[01:35:37] -!- guys has quit [Ping timeout: 246 seconds]
[01:49:25] -!- guys has joined #archlinux32
[01:49:26] -!- DCyrax has quit [Read error: Connection reset by peer]
[01:57:00] -!- woshty has quit [Ping timeout: 244 seconds]
[02:38:03] -!- suudo has joined #archlinux32
[02:39:48] <suudo> hi, i've got a shortage of usb sticks around my office but i do have a pxe server, attempting to copy the working arch pxe setup to include archlinux32 but i need archiso.img and vmlinuz and i'm not sure how to produce those files, if anyone's done this before assistance would be greatly appreciated
[02:42:28] -!- thePiGrepper has quit [Ping timeout: 244 seconds]
[02:54:26] -!- thePiGrepper has joined #archlinux32
[03:36:22] -!- guys has quit [Ping timeout: 268 seconds]
[03:50:25] -!- guys has joined #archlinux32
[04:08:48] <suudo> i've made a forum thread for this, it's not urgent, looking to replace ubuntu on an ancient netbook and arch is preferable to dsl or something similar. love the work, have a good one
[04:08:52] -!- suudo has quit []
[05:15:35] -!- DCyrax has joined #archlinux32
[05:36:37] -!- guys has quit [Ping timeout: 268 seconds]
[05:36:43] -!- thePiGrepper has quit [Ping timeout: 245 seconds]
[05:50:35] -!- guys has joined #archlinux32
[06:42:00] -!- davor_ has joined #archlinux32
[06:42:30] -!- davor has quit [Ping timeout: 264 seconds]
[06:42:30] davor_ is now known as davor
[07:32:35] -!- abaumann has joined #archlinux32
[07:32:35] <buildmaster> Hi abaumann!
[07:32:35] <buildmaster> !rq abaumann
[07:32:36] <phrik> buildmaster: <abaumann> Maybe we should convert Rust to C using a Go to Vala transpiler written in Ocaml?
[07:33:26] <abaumann> deep42thought: nah. As it looks like only my virtual test machines have been hacked, the slave runs on the main host in a chroot, I just rebuilt it. I hope my keys were not compromised.
[07:37:29] -!- guys has quit [Ping timeout: 268 seconds]
[07:48:32] -!- guys has joined #archlinux32
[07:50:09] -!- abaumann has quit []
[09:05:22] -!- deep42thought has joined #archlinux32
[09:05:22] <buildmaster> Hi deep42thought!
[09:05:22] <buildmaster> !rq deep42thought
[09:05:23] <phrik> buildmaster: <deep42thought> one day, some genius will prove, that all programming languages are isomorphic and there is actually only _one_ compiler
[09:16:19] -!- woshty has joined #archlinux32
[09:17:21] <elibrokeit> deep42thought: yes, the build server could/should be fully provisioned to run makepkg
[09:17:51] <deep42thought> yeah, I was just noting, that there _is_ a use case of running printsrcinfo, but not actually building
[09:21:52] <elibrokeit> we unconditionally check for the strip binary too
[09:22:46] <elibrokeit> <elibrokeit> this should instead be e.g. (( EUID > 0 )) && (( SOURCEONLY || ( !VERIFYSOURCE && !PRINTSRCINFO && !PACKAGELIST ))
[09:22:46] <elibrokeit> <elibrokeit> allanbrokeit: what level should we be enforcing? FWIW I guess back in the day we didn't care if you only ran verifysource or something, we still demanded you install fakeroot
[09:22:46] <elibrokeit> <allanbrokeit> I'm happy with it just being enforced - can't see why you would want to use makepkg for the other things but not run packaging
[09:22:48] <elibrokeit> FWIW
[09:24:07] * elibrokeit goes back to pondering the use of emojis in bash function names
[09:24:58] <deep42thought> this would be useful for local routines, so they won't be used elsewhere: https://xkcd.com
[09:24:59] <phrik> Title:xkcd: Vomiting Emoji (at xkcd.com)
[09:25:54] -!- dwilc has quit [Ping timeout: 272 seconds]
[09:26:14] <elibrokeit> !emojicode
[09:26:15] <phrik> ☁() { curl -k "https://wttr.in/$*"; }
[09:26:18] <elibrokeit> it's a thing
[09:37:36] -!- guys has quit [Ping timeout: 252 seconds]
[10:32:14] -!- thePiGrepper has joined #archlinux32
[10:39:24] -!- woshty has quit [Ping timeout: 250 seconds]
[11:10:10] <thePiGrepper> deep42thought: hi
[11:12:22] <deep42thought> Hi thePiGrepper
[11:12:25] <thePiGrepper> I was wondering about asp32. besides the name 'asp32', does it do anything different than asp? I've read the commit log, and you have written several commits already, however the PKGBUILDs been downloaded are from the arch repo still.
[11:12:35] <thePiGrepper> do you have plans to change this in the future?
[11:12:46] <deep42thought> as I said, it's still a stub
[11:13:00] <deep42thought> we merely changes all the names from asp to asp32 and from archlinux to archlinux32
[11:13:05] <deep42thought> but besides that, it does nothing
[11:13:28] <deep42thought> if you're interested in implementing some modifications-features in there, we would be really thankful!
[11:13:32] <thePiGrepper> I see. I ask because I'm getting kinda interested in getting asp32 to actually download and merge the changes
[11:13:43] <deep42thought> yeah, that would be great
[11:13:43] <thePiGrepper> yeah, I think I am
[11:14:34] <thePiGrepper> I just forked it yesterday (after I registered to the gitea server you have there, dont know if it's ok for me to create an account there though..)
[11:15:06] <deep42thought> if you're planning to contribute in some way (even minor) it's totally ok to create an account there
[11:15:41] <thePiGrepper> I also sent a couple of PR to you. one is a quick fix that enables auto completion (it currently does not work). Im still testing it though
[11:15:55] <deep42thought> I just saw it
[11:15:59] <thePiGrepper> and the second is a merge
[11:16:04] <deep42thought> (I don't log in to gitea so often)
[11:16:04] <thePiGrepper> to v4 of the upstream one
[11:16:20] <thePiGrepper> it has two interesting commits
[11:16:42] <thePiGrepper> specially one fix to the 'checkout to different filesystem from ASPROOT' issue
[11:17:38] <deep42thought> we had a rebasing work flow for asp32 so far, but I'm also ok with merging
[11:17:46] <thePiGrepper> I think that one is good to go, the other one I need to test it further because I never dealt with autocompletion scripts before, so Im not sure if all the modifications are there yet
[11:19:32] <thePiGrepper> oh, I see. well, if you want to re do the patches as a rebase, that's fine as well. anyway I had some time yesterday so I worked on that briefly
[11:19:33] -!- dwilc has joined #archlinux32
[11:20:02] <deep42thought> I think merging is fine
[11:20:19] <thePiGrepper> I would like to know what exactly would be the best use case for 'asp checkout' for arch32
[11:20:32] <thePiGrepper> it can be the exact same function as asp
[11:20:44] <thePiGrepper> and just merge everything and send it
[11:20:56] <thePiGrepper> or it can do also both
[11:21:02] <thePiGrepper> with an additional flag maybe
[11:21:13] <deep42thought> yeah, a flag is probably a good idea
[11:21:21] <thePiGrepper> '--nocombine' or something like that
[11:21:25] <deep42thought> --upstream
[11:21:29] <thePiGrepper> better
[11:21:29] <deep42thought> or --upstream-only
[11:21:56] <deep42thought> also the architecture-mangling could be made optional
[11:22:02] <thePiGrepper> that'd be preferable.
[11:22:28] <deep42thought> I think, addition of sub_pkgrel should not happen
[11:22:39] <deep42thought> ... because: what would be added there? ;-)
[11:25:00] <thePiGrepper> that is the 'set PKGBUILD arch variable as system's $arch' right?
[11:25:07] <deep42thought> one thing, at which I got stuck thinking about is: How do we handle all the other commands, asp has? e.g. what should "asp32 log linux" show?
[11:25:13] <thePiGrepper> is one of the things the build system does automatically
[11:25:35] <deep42thought> 'package.inc.sh: mangle the arch=() entry of the PKGBUILD'
[11:27:34] <thePiGrepper> I just read it. mangle_pkgbuild() right? from the builder
[11:27:41] <deep42thought> exactly
[11:38:40] <thePiGrepper> and instead of just adding the supported arch, what would you prefer to do?
[11:39:41] <deep42thought> well, the goal should be to mangle the arch and append the modification PKGBUILD and overwrite the other files from the modifications repository, too
[11:40:17] <deep42thought> e.g.: everything the build master/slaves do (minus some "experimental" stuff, like removing "lib32" prefixes)
[11:42:42] <deep42thought> afk, lunch
[11:43:38] <thePiGrepper> I see. I could(and I surely will) look at the commits to see what kind of modifications you made(besides the s/asp/asp32/ ones, but right now there's nothing made to the functionality itself of the script, right?
[11:43:43] <thePiGrepper> oh, ok. np
[12:00:30] -!- Cthulu201 has quit [Read error: Connection reset by peer]
[12:18:09] -!- Cthulu201 has joined #archlinux32
[12:30:33] <deep42thought> re
[12:31:12] <deep42thought> yeah, functionality should be identical - although I already implemented tracking of our modifications repository
[12:31:23] <deep42thought> but nothing besides cloning it ;-)
[13:00:07] -!- thePiGrepper has quit [Ping timeout: 240 seconds]
[13:28:11] -!- DCyrax has quit [Remote host closed the connection]
[13:28:35] -!- DCyrax has joined #archlinux32
[13:44:58] -!- guys has joined #archlinux32
[13:55:04] -!- guys has quit [Ping timeout: 272 seconds]
[14:14:45] -!- dwilc has quit [Quit: WeeChat 2.3]
[14:33:39] <buildmaster> i686/thunderbird is broken (says buildknecht2).
[14:56:34] -!- dwilc has joined #archlinux32
[14:56:59] -!- thePiGrepper has joined #archlinux32
[14:59:06] <thePiGrepper> Also, I wanted to know something regarding the PKGBUILD of asp itself. it gets the source from a tar on the arch32 repo. does it have to be this way for a particular reason? maybe getting the source from the git repo would be easier to maintain and also to play with
[14:59:16] <thePiGrepper> idk, what do you think?
[14:59:32] <deep42thought> it's that way, so we can sign the tar
[14:59:42] <deep42thought> ... not especially important for asp, but for the keyring
[15:00:08] <thePiGrepper> oh, I see. we cant do that using a git repo then? sorry, I dont really know how signing works..
[15:00:29] <deep42thought> yeah, you can't do _exactly_ that with git
[15:00:34] <deep42thought> you can sign tags in git
[15:00:45] <deep42thought> but then you need to clone the repo (partially) to verify the signature
[15:01:07] <deep42thought> you cannot create a signed tarball from a signed git tag without re-using your private key
[15:01:38] <thePiGrepper> during the pacman -S?
[15:01:40] <deep42thought> ... summa summarum: the release cycle currently still needs me to sign (and upload) the tar balls
[15:01:43] <deep42thought> no
[15:01:54] <deep42thought> during the "makepkg" step
[15:02:13] <deep42thought> uh, one moment
[15:02:21] <deep42thought> _what_ tar balling are you referring to?
[15:07:15] <thePiGrepper> here: https://sources.archlinux32.org
[15:07:16] <phrik> Title:Index of /sources/ (at sources.archlinux32.org)
[15:07:35] <thePiGrepper> that's the current source for these PKGBUILDs, right?
[15:07:39] <deep42thought> yes
[15:07:40] <deep42thought> exactly
[15:07:57] <deep42thought> maintenance of that is pretty straight-forward - there is a "dist" and an "upload" target for make which create/sign and upload the tarballs
[15:09:06] <deep42thought> but you can simply use make install (look into the upstream PKGBUILD) to install a test version of asp32 somewhere if you like
[15:10:09] <thePiGrepper> oh, that's true. I just check asp's Makefile. you do that by hand I assume
[15:10:17] <deep42thought> yes
[15:11:00] <thePiGrepper> so, let me see if I get this right. that's the reason why you only see PKGBUILDs with sources from git repositories on the AUR ?
[15:11:15] <deep42thought> not really
[15:11:27] <deep42thought> there _are_ PKGBUILD with git sources in archlinux
[15:11:36] <deep42thought> but nothing, which requires signing
[15:11:54] <deep42thought> and since I set up that stuff anyway, I started signing all of archlinux32's source tar balls
[15:11:56] <thePiGrepper> as in, for example? one would think that most sources require signing
[15:12:39] <deep42thought> all sources need check-summing
[15:13:02] <deep42thought> but signing is only necessary for "bootstrap" sources - like the keyring or maybe a compiler or pacman itself
[15:13:21] <deep42thought> and the git hash is kind-of-a checksum
[15:14:45] <thePiGrepper> hm, I see. well, I think I get it know. thanks. ;-)
[15:14:51] <deep42thought> :-)
[15:16:21] <thePiGrepper> back to the asp port. basically the requirement would be: 1) do the same as it always do with current arch repos, but under a new flag. and 2) do the merge with the PKGBUILD patches and file substitution, according to $arch
[15:16:29] <thePiGrepper> is that it? I think Im missing something
[15:17:13] <thePiGrepper> maybe writing it down at issues would be a nice thing to do
[15:17:25] <deep42thought> good idea
[15:17:51] <thePiGrepper> do you have the time? that way we can chat about the written requirements if needed
[15:18:09] <deep42thought> can you open issues on asp32
[15:18:15] <deep42thought> or do you need permission for that?
[15:18:18] <deep42thought> I have (some) time
[15:18:35] <thePiGrepper> I think I can. I dont know if its different permission-wise than sending PRs
[15:18:51] <deep42thought> I need to enable the issue tracker :-D
[15:19:07] <deep42thought> there you go
[15:19:18] <thePiGrepper> but because you're the one with the overall requirements maybe you can start it with a brief overview requirement and start from there
[15:19:37] <thePiGrepper> I can add any specific stuff in the comments of the Issue
[15:19:52] <thePiGrepper> I think gitea has the same overall functionality that github at least in that regard
[15:21:09] <thePiGrepper> yeah, confirmed, I have access to the issue tracker(I dont know if I didnt before though lol)
[15:21:29] <deep42thought> you didn't, I didn't, everyone didn't ;-)
[15:21:34] <thePiGrepper> LOL
[15:22:06] <thePiGrepper> is that enabled by project? or it's a gitea setting?
[15:22:18] <deep42thought> neither: per repository
[15:22:21] <thePiGrepper> I see
[15:23:14] <thePiGrepper> if you can do that post,I'd prefer it honestly. in the meantime, I'll test that autocompletion fix so that you can merge it when you prefer
[15:23:30] <deep42thought> yeah, I'm writing something up
[15:23:37] <thePiGrepper> thx
[15:24:26] <elibrokeit> deep42thought: there's no difference between verifying a signed tag in git, and a signature file
[15:24:43] <deep42thought> no difference for whom?
[15:24:44] <elibrokeit> If you're using git repos, you need to clone irrespective of checking the sig
[15:24:58] <elibrokeit> In terms of "correctness" and security
[15:25:13] <deep42thought> and makepkg can check the signature?
[15:25:21] <deep42thought> ah, you're right
[15:25:25] <elibrokeit> Yes -- I wrote the code
[15:25:33] <deep42thought> :_D
[15:25:34] <deep42thought> :-D
[15:25:46] <deep42thought> the problem was downloading the github-created tar ball - which of course lacks the signature
[15:25:55] <elibrokeit> And there are packages with upstream signed tags, and no signed tar balls
[15:26:07] <elibrokeit> Right, that
[15:28:26] <thePiGrepper> but that wouldnt be the case with something in-house like asp, right?
[15:28:45] <thePiGrepper> we could add signed tags for it
[15:28:49] <deep42thought> yeah
[15:28:57] <deep42thought> in fact, I think, the tags _are_ signed
[15:29:06] <elibrokeit> The asp developer stopped uploading tarballs due to preferring to use git with signed tags
[15:29:07] <thePiGrepper> the tags *up to now* that is.
[15:29:20] <elibrokeit> ;)
[15:29:23] <deep42thought> elibrokeit: we're talking about asp32 here
[15:29:59] <elibrokeit> My point still stands ...
[15:30:42] <deep42thought> yeah, you're right
[15:31:10] <deep42thought> we can switch to built-from-git packages
[15:31:16] <deep42thought> I have no problem with that
[15:31:41] <elibrokeit> It causes additional makedepends though
[15:31:54] <deep42thought> and makes bootstrapping harder ;-)
[15:31:56] <elibrokeit> And the git tag is only secured by sha1
[15:33:02] <elibrokeit> I prefer using tarballs over git, and using signatures over either one
[15:33:38] <deep42thought> so your favourite solution are signatures w/o tar balls and w/o git????
[15:34:42] * deep42thought is confused by the english language
[15:58:12] -!- woshty has joined #archlinux32
[16:01:06] -!- NoobAlice has joined #archlinux32
[16:01:26] <deep42thought> abaumann: your pacman-staging-with-build-support-32.conf is broken on tyzoid-srv0-vm486
[16:04:22] -!- guys has joined #archlinux32
[16:04:27] <deep42thought> but more importantly, there seems to be some error packaging gcc on i486 :-(
[16:04:32] <deep42thought> but I have to leave now
[16:04:37] -!- deep42thought has quit [Quit: Leaving.]
[17:10:38] -!- woshty_ has joined #archlinux32
[17:11:54] -!- woshty has quit [Ping timeout: 252 seconds]
[18:08:00] -!- NoobAlice has quit [Ping timeout: 252 seconds]
[18:14:33] <buildmaster> i686/teeworlds are broken (says nlopc46).
[18:17:20] <buildmaster> i686/din is broken (says rechenknecht).
[18:42:31] -!- NoobAlice1 has joined #archlinux32
[19:04:54] -!- abaumann has joined #archlinux32
[19:04:54] <buildmaster> Hi abaumann!
[19:04:55] <buildmaster> !rq abaumann
[19:04:55] <phrik> buildmaster: <abaumann> some packages in Archlinux suffer from 'featureitis'.
[19:05:18] <abaumann> deep42thought: I have various issues obviously: eurobuild3 reports 'Packager error(s):
[19:05:21] <abaumann> displaycal- misses a valid packager.
[19:05:23] <abaumann> 'return-assignment' reports a signature error.
[19:05:33] <abaumann> I have pgp keys installed and configured in slave.conf, I'm puzzled.
[19:05:48] -!- deep42thought has joined #archlinux32
[19:05:49] <buildmaster> Hi deep42thought!
[19:05:49] <buildmaster> !rq deep42thought
[19:05:49] <phrik> buildmaster: <deep42thought> ahm, what is actually in the package "vulkan-headers" on i486? I thought, we disabled all vulcans?
[19:05:52] <abaumann> and what's exactly wrong in pacman-staging-with-build-support-i486.conf on my i486 vm?
[19:06:01] <deep42thought> "invalid packager" means you don't have one in /etc/makepkg.conf
[19:06:19] <deep42thought> I see errors that it cannot download stuff from your mirror
[19:06:27] <deep42thought> so I guess, it tries to download i686 again
[19:06:51] <abaumann> PACKER of the host or the chroot?
[19:06:58] <deep42thought> host
[19:07:07] <abaumann> ok.
[19:07:22] <abaumann> aha. bootstrapping stuff.
[19:07:40] <abaumann> We can clean up there anyway, I think only the main repo is still needed..
[19:07:57] <abaumann> [bootstrap]
[19:07:57] <abaumann> Server = http://archlinux32.andreasbaumann.cc
[19:08:01] <abaumann> The orhters seem all empty
[19:08:37] <deep42thought> yeah, but you should check, that the other (successfully) downloaded packages are for the right architecture
[19:08:47] <deep42thought> my guess is, that someone downloads i686 packages
[19:08:52] <deep42thought> which works for the normal repos
[19:08:57] <deep42thought> but fails for bootstrap repos
[19:09:02] <abaumann> mmh..
[19:09:11] <abaumann> 'return-assignment' reports a signature error.
[19:09:20] <abaumann> that one means I'm not signing the packages correctly?
[19:09:25] <deep42thought> this means, the buildmaster doesn't trust your signature
[19:09:44] <abaumann> displaycal-
[19:09:50] <deep42thought> there should be some lines preceding the error (I think)
[19:10:02] <abaumann> In my brain I have configured PGP keys correnctly.
[19:10:07] <deep42thought> can you verify, that you're using the correct key?
[19:10:08] <abaumann> ony "Packager error(s):
[19:10:08] <abaumann> displaycal- misses a valid packager.
[19:10:11] <abaumann> *only
[19:10:23] <abaumann> package_key='0xC8E8F5A0AF9BA7E7'
[19:10:26] <abaumann> in slave.conf
[19:10:50] <abaumann> pub rsa2048 2017-11-02 [SC] [expires: 2019-11-02]
[19:10:50] <abaumann> 16194A82231E9EF823562181C8E8F5A0AF9BA7E7
[19:10:50] <abaumann> uid [ultimate] Andreas Baumann (sign) <mail@andreasbaumann.cc>
[19:10:50] <abaumann> sub rsa2048 2017-11-02 [E] [expires: 2019-11-02]
[19:10:53] <abaumann> seems ok.
[19:11:09] <deep42thought> "misses a valid packager" means, you didn't configure a packager in your /etc/makepkg.conf (on the host)
[19:11:09] <abaumann> now, this is in the 'build' user PGP chain.. let's see for the hosts root one.
[19:11:28] <abaumann> I would see other errors, if the signing went wrong, right?
[19:11:53] <abaumann> I also have the same keys for root on the host.
[19:12:02] <abaumann> ok, fixed the makepkg.conf PACKAGER.
[19:12:07] <abaumann> let's see now :-)
[19:12:32] <abaumann> sorry for bugging. But my infrastructre really went down the drain after the attack..
[19:12:40] <deep42thought> np
[19:13:27] <deep42thought> it's a nice test what is needed to set up a slave ;-)
[19:14:42] <abaumann> Architecture = auto
[19:14:47] <abaumann> in staging makepkg.
[19:14:56] <deep42thought> that won't work
[19:14:57] <abaumann> actually, it should see a i486 on a i486 VM, but.
[19:15:01] <deep42thought> no
[19:15:08] <deep42thought> setarch i686 uname -m
[19:15:08] <abaumann> I don't trust those flags anymore now. :-)
[19:15:16] <abaumann> yeah. we habe been there.
[19:15:21] <abaumann> glad I found it :-)
[19:15:49] <abaumann> or 'ln -s i486/boostrap i686' ;-)
[19:16:36] <abaumann> BTW. I will have a power outage tomorrow between 9:00 and 12:00 PM
[19:16:40] <abaumann> AM
[19:16:54] <abaumann> not much I can do about it, I'm afraid.
[19:17:01] <deep42thought> np
[19:17:04] <abaumann> some power cables and counters get moved around in the house.
[19:18:04] <deep42thought> you can power the alix from a battery ;-)
[19:21:05] <abaumann> the alix and the optical switch will come up automatically, the web proxy I will keep alive with a Powergorilla battery and the webserver I can awake with a WOL package, beatiful new world of technoloy :-)
[19:23:43] <abaumann> aha: displaycal- build und uploaded correctly.
[19:24:02] <deep42thought> \o/
[19:53:54] -!- abaumann has quit [Quit: leaving]
[19:54:11] -!- abaumann has joined #archlinux32
[19:54:12] <buildmaster> Hi abaumann!
[19:54:12] <buildmaster> !rq abaumann
[19:54:13] <phrik> buildmaster: <abaumann> "anywhere is anywhere for all values of anywhere"
[20:13:11] -!- abaumann has quit [Quit: leaving]
[22:05:29] -!- isacdaavid has joined #archlinux32
[22:44:04] -!- deep42thought has quit [Quit: Leaving.]
[22:51:27] -!- DepositePirate has quit [Remote host closed the connection]
[22:51:56] -!- DepositePirate has joined #archlinux32
[23:01:18] -!- isacdaavid has quit [Ping timeout: 245 seconds]
[23:39:27] <buildmaster> i686/postgrest is broken (says buildknecht) - I rescheduled: haskell-connection, haskell-tls, haskell-x509-validation.
[23:41:20] <buildmaster> i686/shellcheck is broken (says rechenknecht).
[23:43:20] -!- buildmaster has quit [Remote host closed the connection]
[23:43:35] -!- buildmaster has joined #archlinux32