#archlinux32 | Logs for 2019-01-20

Back
[00:26:55] -!- ofara__ has quit [Ping timeout: 244 seconds]
[00:29:52] -!- ofara__ has joined #archlinux32
[00:33:16] -!- oaken-source has quit [Ping timeout: 272 seconds]
[01:17:41] <buildmaster> i686/haskell-githash is broken (says eurobuild3).
[01:19:31] rcf is now known as rf
[01:24:38] rf is now known as rcf
[01:32:29] <buildmaster> i686/haskell-yesod-test is broken (says buildknecht2) - I rescheduled: haskell-yesod-form, haskell-yesod-persistent.
[01:34:35] <buildmaster> i686/haskell-yesod-static is broken (says rechenknecht) - I rescheduled: haskell-yesod-test.
[02:23:55] <buildmaster> i686/vault is broken (says buildknecht).
[03:02:18] -!- thePiGrepper has quit [Ping timeout: 250 seconds]
[03:22:34] -!- thePiGrepper has joined #archlinux32
[05:02:18] -!- thePiGrepper has quit [Ping timeout: 244 seconds]
[06:42:33] -!- thePiGrepper has joined #archlinux32
[07:38:41] -!- MrBIOS has joined #archlinux32
[07:42:03] -!- MrBIOS has quit [Client Quit]
[07:42:29] -!- MrBIOS has joined #archlinux32
[07:42:50] -!- MrBIOS has quit [Client Quit]
[07:43:10] -!- MrBIOS has joined #archlinux32
[07:43:36] -!- MrBIOS has quit [Client Quit]
[07:43:58] -!- MrBIOS has joined #archlinux32
[07:44:22] -!- MrBIOS has quit [Client Quit]
[07:44:50] -!- MrBIOS has joined #archlinux32
[07:45:08] -!- MrBIOS has quit [Client Quit]
[07:45:37] -!- MrBIOS has joined #archlinux32
[07:45:54] -!- MrBIOS has quit [Client Quit]
[07:46:15] -!- MrBIOS has joined #archlinux32
[07:46:40] -!- MrBIOS has quit [Client Quit]
[07:47:10] -!- MrBIOS has joined #archlinux32
[07:47:26] -!- MrBIOS has quit [Client Quit]
[07:47:45] -!- MrBIOS has joined #archlinux32
[07:48:12] -!- MrBIOS has quit [Client Quit]
[08:15:58] -!- woshty has joined #archlinux32
[09:16:53] -!- deep42thought has joined #archlinux32
[09:16:53] <buildmaster> Hi deep42thought!
[09:16:53] <buildmaster> !rq deep42thought
[09:16:53] <phrik> buildmaster: <deep42thought> because you're a pessimist with no hope - or you're a realist afraid of change ... pick your evil
[09:17:33] <deep42thought> tyzoid: mariadb does not start anymore on packages.archlinux32.org after the upgrad, although I did it identical to the upgrade on several of my servers which worked fine
[09:17:47] <deep42thought> I'll change dns back to my server until we find the culprit
[09:19:50] <deep42thought> yours is still available via packages-tyzoid.archlinux32.org
[09:36:53] -!- abaumann has joined #archlinux32
[09:36:53] <buildmaster> Hi abaumann!
[09:36:53] <buildmaster> !rq abaumann
[09:36:53] <phrik> buildmaster: <abaumann> So I prefer cheap bootstrapping methods over the correct ones :-)
[09:37:04] <abaumann> morning deep42thought :-)
[09:39:16] <deep42thought> morning abaumann!
[09:40:42] <abaumann> mysqld playing games again..
[09:40:51] <abaumann> on packages, that is.
[09:41:10] <abaumann> any usable error messages?
[09:43:09] -!- oaken-source has joined #archlinux32
[09:44:22] <deep42thought> yeah, some permissions error
[09:44:30] <deep42thought> one moment
[09:44:59] <deep42thought> uhm now I cannot login anymore O.o
[09:45:05] <deep42thought> gnaaaa
[09:45:11] <deep42thought> I broke it
[09:45:11] <abaumann> more permission problems..
[09:45:16] <abaumann> oups.
[09:45:28] <deep42thought> *mimimi* tyzoid *mimimimi*
[09:45:39] <abaumann> !grab deep42thought
[09:45:40] <phrik> abaumann: Tada!
[09:51:59] <deep42thought> ah, my mistake - I changed the dns :-D
[09:53:23] <deep42thought> mariadb.service: Failed to set up mount namespacing: Permission denied
[09:53:29] <deep42thought> mariadb.service: Failed at step NAMESPACE spawning /bin/sh: ermission denied
[09:53:49] <abaumann> wow. that sounds bad.
[09:54:16] <deep42thought> maybe some lxc-interference?
[09:54:26] <abaumann> rather some systemd-resolved interference
[09:54:49] <abaumann> lxc and systemd and resolved/networking are not working for me anywhere or it is very shacky..
[09:54:52] <deep42thought> what's systemd-resolved's business there?
[09:54:59] <abaumann> interference
[09:55:02] <deep42thought> lol
[09:55:15] <abaumann> I had all kind of bugs like: "/etc/resolve.conf still in use"
[09:55:28] <abaumann> races between networking and resolved during booting.
[09:55:41] <abaumann> there my statement originates: they should start testing..
[09:56:13] <abaumann> I agree, that most likely nobody uses lxc.. why take the simple thing, when you can take the fancy thing - aka docker.
[09:56:29] <deep42thought> the upgrade was only for mariadb-libs libsasl libwebp mariadb-clients and mariadb
[09:56:57] <abaumann> *abaumann feelds Sunday morning sarcasm entering the room - not that it would be any different from weekday sarcasm.
[09:57:07] <deep42thought> !grab abaumann
[09:57:07] <phrik> deep42thought: Tada!
[09:57:24] <abaumann> mmh..
[09:57:25] <deep42thought> actually I wanted to grab the one before, but that one is also good ;-P
[09:57:33] <buildmaster> i686/prettier is broken (says buildknecht).
[09:59:07] <abaumann> ok. wget on i486 seems fixed.. libpsl still had a dependency on libidn2.so.0
[09:59:24] <deep42thought> thank you :-)
[09:59:32] <abaumann> no :-)
[09:59:35] <abaumann> *np
[09:59:51] <deep42thought> "thanks" -- "no"
[09:59:56] <deep42thought> :-P
[10:00:02] <abaumann> broken keyboard..
[10:00:06] <abaumann> and broken fingers.. ;-)
[10:00:21] <deep42thought> crumbs under the keys? ... that's not broken, just dirty
[10:01:19] <abaumann> actually. it's a recycled keyboard formerly used by a heavy smoker. So even washing the keyboard for days still left some stickiness under the keys..
[10:01:42] <deep42thought> :-D
[10:01:55] <deep42thought> passive smoking
[10:02:19] <abaumann> yes. smokers actually kill elecronics much faster..
[10:03:05] <deep42thought> computers ranked by dirt: smokers, car mechanics, bedroom, ...
[10:03:18] <abaumann> !grab deep42thought
[10:03:18] <phrik> abaumann: Tada!
[10:04:24] <deep42thought> looks like the problem is indeed lxc
[10:04:25] <deep42thought> :-/
[10:04:32] <deep42thought> there is not much I can do about that ...
[10:04:51] <abaumann> why should mysql be running inside a container?
[10:05:02] <deep42thought> the whole packages vm is running in one
[10:05:02] <abaumann> I know, this is almost a sin to ask: why containers?
[10:05:06] <abaumann> ah.
[10:05:17] <abaumann> and is isolated from other vms. ok.
[10:05:22] <deep42thought> it's tyzoid's way of virtualizing the os
[10:05:49] <abaumann> no, that's very ok. it isolates services, and using lxc is more lightweight than KVM virtualization.
[10:06:13] <abaumann> I thought, it's a mysqld container running ON an already virtualized system..
[10:06:17] <abaumann> .. my mistake.
[10:06:23] <deep42thought> :-D
[10:06:25] <deep42thought> np
[10:06:30] <deep42thought> (and "no")
[10:19:04] <abaumann> is /etc/resolve.conf a symlink to /run/systemd/resolve/stub-resolve.conf? when using systemd-resolved..
[10:19:14] <abaumann> because that was wrong on my LXC containers.
[10:33:32] -!- oaken-source has quit [Ping timeout: 246 seconds]
[10:51:09] <abaumann> spoken too soon: /etc/resolv.conf contains a nameserver 127.0.0.53, what's that exactly?
[10:51:19] <abaumann> after a while, DNS resolving just stops.
[10:59:29] <abaumann> aha. had DNS blocked on the server hosting LXC container.. stupid me. :-)
[11:15:18] <deep42thought> 127.0.0.53 is the dns server exposed by systemd-resolved
[11:15:32] <abaumann> aha.
[11:15:32] <deep42thought> so "legacy dns" clients can use systemd-resolved via that
[11:15:45] <abaumann> legacy dns client? :-)
[11:16:02] <abaumann> yeah. I had a mess with config lxc files, iptables on the host, etc.
[11:16:05] <abaumann> now it works.
[11:16:11] <deep42thought> well, programs that do not resolve via system calls but via dns themself
[11:16:40] <abaumann> ah, not using the C api functions for DNS resolving, right.
[11:16:43] <deep42thought> I don't quite understand, why I need systemd-resolved _at_all_
[11:17:02] <abaumann> well, either that, or a static /etc/resolv.conf
[11:17:30] <abaumann> also the dhcp clients overwrite /etc/resolv.conf with the nameserver, domain etc. values provided by the DHCP server.
[11:17:45] <deep42thought> yeah
[11:17:49] <deep42thought> that's how I run it
[11:17:57] <deep42thought> there's a dns server on the router
[11:17:59] <abaumann> systemd-resolved takes DNS settings from the network profile and exposes them in /run
[11:18:03] <deep42thought> which gets pushed via dhcp
[11:18:15] <abaumann> that's usually also my preferred way.
[11:18:30] <abaumann> for laptops moving around networks, this is maybe not such a good idea.
[11:18:41] <abaumann> they are more in for networkmanager.
[11:18:51] <abaumann> which is something completly useless on servers, IMHO. :-)
[11:19:40] <deep42thought> why not let dhcp push stuff onto your laptops, too?
[11:19:46] <deep42thought> that sounds absolutely reasonable to me
[11:19:58] <deep42thought> constant setting: "get everything from dhcp"
[11:20:09] <abaumann> *abaumann shrugs
[11:20:27] <abaumann> maybe DNS settings are not that consistently correct in some DHCP networks?
[11:20:36] <deep42thought> they *should* be
[11:20:58] <deep42thought> setting up yet another service to work around broken dhcp servers is imho a bad idea
[11:21:09] <deep42thought> one might go with static ips, then, too
[11:22:08] <abaumann> ok, it also servers as DNS cache, like dnsmasq or upbound
[11:22:28] <deep42thought> yeah, I don't need that either
[11:22:37] <deep42thought> because my router is quite idle
[11:22:40] <deep42thought> and caches itself
[11:23:31] <abaumann> my router is OpenBSD and has upbound on it, for caching.
[11:23:38] <abaumann> caching is shared in one location in the network.
[11:23:46] <deep42thought> yeah, here too
[11:23:55] <deep42thought> also makes debugging dns issues easier
[11:24:11] <abaumann> oh yeah. :-)
[11:24:45] <abaumann> LLMNR, oh my.
[11:24:56] <deep42thought> what's that?
[11:25:04] <abaumann> Link-Local Multicast Name Resolution
[11:25:38] <abaumann> looks like a Windows thing, originally.
[11:26:52] <abaumann> DNS resolution without DNS servers? huh?
[11:27:22] <abaumann> yeah. that's most likely something I don't want to happen in my network. :-)
[11:30:06] <abaumann> quiz question: what's wrong here: "/var/cache/pkg -> /usr/ports/var/cacke/pkg"
[11:30:38] <abaumann> FreeBSD pkg just hung forever with 100% CPU :-)
[11:54:13] -!- MrBIOS has joined #archlinux32
[12:03:27] -!- deep42thought has quit [Ping timeout: 252 seconds]
[12:33:33] -!- pulec has quit [Quit: fuuuuu]
[12:35:20] -!- pulec has joined #archlinux32
[12:49:15] <buildmaster> i686/haskell-wreq is broken (says eurobuild3) - I rescheduled: haskell-authenticate-oauth.
[12:57:33] <buildmaster> i686/haskell-mustache is broken (says buildknecht2) - I rescheduled: haskell-authenticate-oauth, haskell-wreq.
[13:06:53] <buildmaster> i686/haskell-yesod-test is broken (says buildknecht2) - I rescheduled: haskell-yesod-form, haskell-yesod-persistent.
[13:16:09] -!- deep42thought has joined #archlinux32
[13:16:09] <buildmaster> Hi deep42thought!
[13:16:09] <buildmaster> !rq deep42thought
[13:16:10] <phrik> buildmaster: <deep42thought> good night, python - cu tomorrow python2!
[13:16:27] <deep42thought> ... stupid ipv6 ....
[13:18:44] <buildmaster> i686/haskell-yesod-static is broken (says eurobuild3) - I rescheduled: haskell-yesod-test.
[13:49:11] -!- DepositePirate has quit [Remote host closed the connection]
[13:49:36] -!- DepositePirate has joined #archlinux32
[13:59:01] <buildmaster> i686/pandoc is broken (says buildknecht) - I rescheduled: haskell-hslua-module-text.
[14:14:26] <buildmaster> i686/shellcheck is broken (says buildknecht).
[14:50:53] <abaumann> ... stupid security ...
[14:51:40] <abaumann> lightdrm-gtk-greeter is locking all memory with mlockall, just because it deals in 20 character somewhere with a password?
[14:51:56] <deep42thought> well, better safe than sorry ;-)
[14:53:02] <abaumann> Yes, but when threads each get 8MB (!!) of memory, this might be unlockabable, hence the mmap2 EAGAIN error (which is another error, I don't understand, EAGAIN in Posix means, the function call should succeed in the future). should I use a loop around that?
[14:53:57] <deep42thought> maybe you have more memory in the future?
[14:54:16] <abaumann> yes. Like auto-plugable..
[14:54:17] <deep42thought> just pack a loop around it and install more memory on-the-fly :-D
[14:54:27] <abaumann> ..actually, there is hot-swappable memory :-)
[14:55:17] <deep42thought> "hold on a sec, my oom-lamp is flashing, I need to insert a few more ram modules"
[14:55:32] <abaumann> OOM-killer could order memory on ebay..
[14:55:40] <deep42thought> :-D
[14:55:41] <deep42thought> !grab abaumann
[14:55:42] <phrik> deep42thought: Tada!
[14:56:04] <deep42thought> what's the hotkey for the ebay order?
[14:56:16] <abaumann> :-)
[14:58:20] <deep42thought> big haskell rebuild in progress, lets see what packages pass on the 2nd or 3rd trial :-)
[14:58:35] <abaumann> yeah. seen quite some Haskell-reschedules..
[14:59:05] <deep42thought> what is "yarn"?
[14:59:25] <abaumann> "Fast, reliable, and secure dependency management"
[14:59:25] <deep42thought> it segfaults in the build of bash-language-server ... whatever _that_ may be
[14:59:27] <abaumann> mmh.
[15:00:05] <abaumann> yet another package manager for nodejs.
[15:00:07] <abaumann> nice.
[15:00:14] <deep42thought> umm, "yarn" is of architecture "any", so I doubt it itself segfaults
[15:00:29] <abaumann> node. I bet, it's node itself.
[15:00:31] <deep42thought> shouldn't it be yapmfnj, then?
[15:00:40] <deep42thought> oh, my
[15:00:52] <abaumann> I remember a node related segfault some time ago in the mailing list..
[15:02:28] <abaumann> LimitMEMLOCK=infinity
[15:02:33] <abaumann> hah in /lib/systemd/system/lightdm.service
[15:04:47] <abaumann> I'll make a patch :-)
[15:05:13] <deep42thought> so this issue is an upstream one, right?
[15:06:02] <abaumann> not really.
[15:06:27] <abaumann> it happens, because default stack sizes are different in 64 and 32 bit and per platform
[15:06:35] <deep42thought> ah, ok
[15:06:40] <abaumann> lightm-gtk-greeter works without any problems on Archlinux 64.
[15:07:20] <abaumann> But I'll make a bug report upstream/upstream for the greeter itself, as setting the limit just for the greeter to fit in memory, feels like a workaround.
[15:29:38] <buildmaster> i686/hedgewars are broken (says rechenknecht).
[16:31:53] -!- finsternis has quit [Excess Flood]
[16:32:26] -!- finsternis has joined #archlinux32
[16:36:36] <thePiGrepper> deep42thought: hi, can you please do one more merge from asp upstream. it's a new bug fix.
[16:36:47] <deep42thought> ok, gimme a sec
[16:37:35] <deep42thought> ok, done
[16:37:41] <thePiGrepper> thx
[16:44:48] <thePiGrepper> btw, quick git question. if by error you created a branch named 'some/thing' and you already have a remote named 'some' containing a branch named 'thing', how can you delete the local 'some/thing' branch? git gets confused..
[16:45:11] <deep42thought> no idea
[16:46:13] <thePiGrepper> solved it :)
[16:46:18] <deep42thought> how?
[16:47:34] <thePiGrepper> by realizing I was being stupid and moving HEAD to another branch... \facepalm
[16:47:58] <thePiGrepper> however this did used to be an issue sometime in the past
[16:48:09] <thePiGrepper> apparently not anymore
[16:56:46] -!- DCyrax has joined #archlinux32
[17:36:14] -!- abaumann has quit [Quit: leaving]
[17:42:28] <deep42thought> thePiGrepper: I can also give you write access to the repo on gitea - as it seems you do much more work on asp32 than me :-)
[17:53:32] -!- AndrevS has joined #archlinux32
[18:25:21] <buildmaster> i686/haskell-hpack is broken (says rechenknecht) - I rescheduled: haskell-interpolate, haskell-src-meta.
[19:09:36] -!- oaken-source has joined #archlinux32
[19:17:35] <thePiGrepper> hm. that might be a good idea.. one benefit would be a git history with less merges and me not bothering you so much ;-)
[19:18:06] <thePiGrepper> deep42thought: btw, I just push my changes for 'export' and 'checkout', maybe you can test it
[19:18:07] <deep42thought> I gave you write access
[19:18:34] <deep42thought> I will do (later, today)
[19:18:54] <thePiGrepper> I see. then, do I just need to try to push it directly to that remote right?
[19:19:02] <deep42thought> exactly
[19:19:15] <thePiGrepper> can I create a new branch there? like an experimental/feature branch?
[19:19:33] <deep42thought> sure, go ahead
[19:19:34] <thePiGrepper> to not pollute the master
[19:19:37] <thePiGrepper> ok then
[19:23:27] <thePiGrepper> done. I created a 'next' branch, and pushed the patches there. take a look when you can. :)
[19:24:58] -!- MrBIOS has quit [Quit: MrBIOS]
[19:25:58] <thePiGrepper> btw, I'd like to change the PKGBUILD for asp so that it starts updating from signed git commits. what do you think?
[19:26:13] <deep42thought> good idea
[19:28:04] <thePiGrepper> ok. to where do I send that patch? packages.git, right?(I think I just answered it myself..)
[19:31:39] <deep42thought> you did :-)
[19:31:57] <deep42thought> but probably, you need to open a pull request, because I only gave you write access for the asp32 repo
[19:32:08] <thePiGrepper> yeah, I know/assumed that. np
[19:33:09] <thePiGrepper> when/if the patches are accepted, the package version would change?
[19:34:06] <deep42thought> yes
[19:34:32] <deep42thought> the sub_pkgrel
[19:38:43] <thePiGrepper> also, one more random question regarding asp. do you think it's made in bash for a reason? like portability for example? or it was just because of convenience. I have the same question for all other arch util scripts, like makepkg.
[19:38:56] <thePiGrepper> what do you think?
[19:43:57] <deep42thought> I have no idea about the reason
[19:44:29] <deep42thought> I guess, it's also for readability, non-breakability and easy-to-modifyability
[20:17:14] -!- isacdaavid has joined #archlinux32
[20:18:05] <deep42thought> thePiGrepper: do I read the source correctly, that it should apply our patches?
[20:18:16] <deep42thought> I'm asking, because it looks like it does not apply them :-/
[20:20:38] <deep42thought> ah, stupid me
[20:20:51] <deep42thought> I should actually switch to the "next" branch before building :-D
[20:20:56] <buildmaster> i686/rust is broken (says rechenknecht).
[20:25:11] <deep42thought> ok, applying patches looks good besides that it does not copy the additional files
[20:33:45] -!- deep42thought has quit [Ping timeout: 252 seconds]
[20:53:52] <thePiGrepper> deep42thought: it does not copy the additional files? how so? with both commands, extract and checkout, asp copies the arch32 repo package files onto the local dir, and then proceeds to append PKGBUILD.
[20:54:03] -!- deep42thought has joined #archlinux32
[20:54:04] <buildmaster> Hi deep42thought!
[20:54:04] <buildmaster> !rq deep42thought
[20:54:04] <phrik> buildmaster: * deep42thought listens carefully, but doesn't hear any bells ringing over here
[20:54:22] <thePiGrepper> and if you want the original asp functionality, you need to add -u
[20:58:11] <deep42thought> I tried "asp32 export linux" and it did not copy the config.i686 and config.i486 file
[21:00:52] <deep42thought> -u works as expected
[21:01:03] <thePiGrepper> hm. it's weird. I'm not having that issue here
[21:01:27] <thePiGrepper> the only thing I can think is different between our systems is ~/.cache/asp32
[21:02:22] <thePiGrepper> when you tried your first command with the build, did it take some time more than usual?
[21:02:41] <thePiGrepper> it fetches the whole arch32 packages.git master branch
[21:02:55] <deep42thought> hard to tell
[21:03:01] <deep42thought> it took maybe a few seconds more
[21:03:25] <deep42thought> but that could be due to my sloppy internet connection (I'm accessing the box via ssh and have some backups running in parallel)
[21:03:35] <thePiGrepper> can you 'mv ~/.cache/asp32 ~/.cache/asp32.old' and try again?
[21:05:05] <deep42thought> let's just rm -rf it :-)
[21:05:43] <deep42thought> same
[21:05:54] <deep42thought> but this was too fast for a complete re-clone :-/
[21:06:06] <deep42thought> I do have ~/.cache/asp
[21:06:10] <deep42thought> maybe it's using that one?
[21:06:21] <thePiGrepper> I'll take a look
[21:06:24] <thePiGrepper> give me a min
[21:06:45] <deep42thought> ok, I removed both, but still it's a) fast, b) not copying the other configs
[21:06:54] <deep42thought> arrrgh
[21:06:56] <deep42thought> wait
[21:07:09] <deep42thought> it works
[21:07:27] <deep42thought> I relied on tab-auto-complete and the fact, that asp32 and asp exlude each other
[21:07:39] <deep42thought> but it appears, I have asp installed and asp32 via "make install"
[21:07:47] <deep42thought> and I accidentally ran asp instead of asp32
[21:08:14] <deep42thought> yup, it works
[21:08:29] <deep42thought> one thing, though: it would be nice if you applied the patch to trunk/, too
[21:08:38] <deep42thought> although that's technically not 100% correct
[21:09:48] <thePiGrepper> hm, ok I guess. you're talking about checkout, right? checkout only applies the changes to x86-64
[21:10:44] <deep42thought> yes
[21:10:58] <deep42thought> export works as I would expect
[21:11:26] <thePiGrepper> ok, I will do that then. it's quite simple.
[21:11:29] <deep42thought> ah, one todo are packages that have no upstream
[21:11:39] <deep42thought> e.g. archlinux32-keyring
[21:12:04] <thePiGrepper> oh, that's right. I didnt think of those
[21:12:22] <deep42thought> checkout certainly does not make much sense for those, though
[21:13:15] <thePiGrepper> hmm, yeah, however it would be weird if checkout wouldnt work at all, and they only worked with export..
[21:13:34] <thePiGrepper> not very user-friendly..
[21:14:43] <deep42thought> or just do a "git init" before applying our stuff?
[21:14:46] <thePiGrepper> can you add an issue for that? so that I remember later. also, to add some extra things to do, like trying to add bash-completion for the arch32 repo
[21:15:29] <deep42thought> will do
[21:16:41] <thePiGrepper> also, before I forget, one of the patches adds zsh completion for -u, however Im not so sure if it's ok because I dont use zsh, let's hope someone who does use it finds if it's ok
[21:16:57] <deep42thought> well it's better than nothing, I suppose
[21:17:33] <thePiGrepper> let's see: asp checkout usually creates a single-branch clone from the correct remote/package.
[21:18:18] <thePiGrepper> in our case, we dont have a multiple-branch repo, instead we have a multiple directory, single-branch repo
[21:18:52] <deep42thought> don't create a git repo of our stuff
[21:18:59] <deep42thought> just create an _empty_ git repo
[21:19:02] <deep42thought> (upstream)
[21:19:04] <deep42thought> add the dirs
[21:19:08] <deep42thought> (not in git anyways)
[21:19:21] <deep42thought> and add our "patches", e.g. the whole package source
[21:19:35] <deep42thought> the last point should be as usual
[21:19:40] <thePiGrepper> yeah, it doesnt make sense. nvm. I will just add an empty git repo and copy the patches.
[21:19:45] <deep42thought> and the first two points are the fail-safe if no upstream exists
[21:20:27] <deep42thought> umm, what's actually to do with the bash completion?
[21:20:28] <thePiGrepper> how many of those arch32-only are there?
[21:20:31] <deep42thought> add the new options or what?
[21:20:39] <deep42thought> <10
[21:21:08] <thePiGrepper> the original asp have auto completion for packages, it actually auto complete the available packages in the repo
[21:21:21] <deep42thought> ah, ok
[21:21:22] <deep42thought> nice
[21:21:25] <thePiGrepper> so, that should be extended for the ones in arch32.
[21:21:58] <thePiGrepper> I think they use a cache file for that with all the repos, and keep updating it once per hour per call
[21:22:15] <thePiGrepper> I'll try to do something like that
[21:22:35] <deep42thought> pcmciautils, linux-pae, archlinux32-keyring, archlinux32-keyring-transition are the packages w/o upstream
[21:22:56] <deep42thought> yeah, sounds good
[21:23:12] <thePiGrepper> zsh auto completion I probably wont do, because Id need to take a look at how that works.. If I end up not doing it, I'll at the very least add the issue
[21:23:24] <deep42thought> ok
[21:23:47] <thePiGrepper> also, a couple things regarding the other commands
[21:24:22] <thePiGrepper> update works fine. no changes needed, it only needed the master branch to be added to work fine
[21:24:56] <deep42thought> yeah, that's the advantage of registering our repo with the other two :-)
[21:26:10] <thePiGrepper> the list-repos, list-arches, list-local commands. there's a pending issue for this
[21:26:33] <thePiGrepper> let's see: list-local is fine. no changes
[21:27:44] <thePiGrepper> list-arches would be a little weird. in theory it would need to show the 'upstream' arches *and* the ones supported by arch32, which are 3 always. would that need to be added, and if so, it would be hardcoded right?
[21:28:01] <thePiGrepper> does that make sense? it might be a bit futile
[21:28:07] <deep42thought> well, packages can be blacklisted
[21:28:16] <deep42thought> but that is pretty much only known to the mysql database
[21:28:32] <deep42thought> or you would need to parse the blacklist from the packages repository
[21:28:38] <thePiGrepper> there's a blacklist file in the repo, is that what you're talking about?
[21:28:54] <deep42thought> I'd say: we keep list-arches as-is, open an issue and solve it whenever we are horribly bored
[21:28:59] <thePiGrepper> I can do that easily, the file is already in the branch after all
[21:29:01] <deep42thought> yes, exactly
[21:29:35] <thePiGrepper> ok, let's do that. let's leave list-arches for later then and open an issue ;-)
[21:29:36] <deep42thought> but be aware, that blacklisting a package for one architecture also blacklists it for all more mighty architectures
[21:29:52] <deep42thought> (or was it vice versa)
[21:29:57] <deep42thought> yeah, vice versa!
[21:30:09] <thePiGrepper> what is?
[21:30:21] <deep42thought> if you blacklist something for i486, it is blacklisted for that only
[21:30:32] <deep42thought> and if you blacklist something for i686, it's also blacklisted for i486
[21:31:27] <thePiGrepper> I see. I just took a look at the blacklist file, it would be trivial to parse with awk
[21:31:42] <deep42thought> that's how we do it :-D
[21:32:21] <deep42thought> list-repos is hard
[21:32:39] <deep42thought> because it's not in the source, wether the package is in staging,testing or stable
[21:32:47] <deep42thought> only if it will end up in extra,core or community
[21:33:22] <thePiGrepper> nice then. well, if asp32 checkout/export doesnt find files in the arch32 to apply/copy, it wont do it, leaving the cloned repo basically the same as it would be if you used -u
[21:33:23] <deep42thought> I would say, we resort to giving only _that_ info - e.g. a package which is currently in [community-testing] should show up as "community"
[21:33:46] -!- ofara__ has quit [Quit: ofara__]
[21:34:26] <thePiGrepper> hmm, yeah, we dont have that kind of info in the arch32 repo, better to just leave it with the info from upstream
[21:34:43] <thePiGrepper> it wont make much difference either way, I think
[21:34:53] <deep42thought> yes
[21:35:03] -!- ofara_ has joined #archlinux32
[21:35:06] <deep42thought> and packages w/o upstream should just be displayed as core/extra/community
[21:37:42] <thePiGrepper> ohh, right. well, probably after I add the functionality to checkout/export arch32-only packages, I'll have the piece of code needed to know the repos for those packages, and with that I'll modify list-repos.
[21:38:03] <deep42thought> yeah, one step after another :-)
[21:38:31] <thePiGrepper> there's a more generic issue for list-*, I guess that could be closed now that we have more specific ones
[21:38:47] <deep42thought> I added comments about the specific issues to that one
[21:38:59] <thePiGrepper> oh, nice. that works too :)
[21:39:50] <deep42thought> but we can split it in several issues, too - if you like to see the progress by closing issues ;-)
[21:43:17] <thePiGrepper> yeah, that might be better, if they arent too small that is.
[21:44:12] <thePiGrepper> also, there're the *log commands. for now no changes there. what do you think? no much we can do I think..
[21:44:25] <deep42thought> we could add a log32
[21:44:35] <deep42thought> but that's the maximum we can do easily
[21:45:33] <thePiGrepper> right. and it could show only the commits that change files inside ${repo}/${pkgname}/
[21:45:43] <deep42thought> exactly
[21:45:59] <deep42thought> all that goes beyond that, is imho insane
[21:46:09] <thePiGrepper> ok, let's do that.
[21:46:29] <thePiGrepper> yeah. I've been thinking many insane things lately regarding that
[21:46:42] <deep42thought> on-the-fly merges/rebases?
[21:46:46] <thePiGrepper> ie: recreating branches arch-upstream
[21:46:47] <thePiGrepper> lol
[21:46:49] <thePiGrepper> yeah
[21:46:58] <thePiGrepper> arch-upstream style
[21:47:57] <thePiGrepper> as long as there arent single commits dealing with multiples packages, it shouldnt be too difficult
[21:48:21] <thePiGrepper> and all operations would be local..
[21:48:31] <thePiGrepper> as I said, insane..
[21:48:31] <deep42thought> still ugly :-D
[21:48:34] <thePiGrepper> yeah
[21:49:19] <thePiGrepper> between that and a complete rewrite in go just for the sake of it, many insane things indeed
[21:49:59] <deep42thought> why go - brainfuck!
[21:50:10] <thePiGrepper> I even began questioning that if I did that, it would start looking as a AUR helper soon.
[21:50:25] <deep42thought> yeah, track aur, too
[21:50:26] <deep42thought> :-D
[21:50:44] <deep42thought> that's how feature-creep starts
[21:51:10] <thePiGrepper> yeah. exactly. no one needs emacs-asp, (until one does...)
[21:51:21] <deep42thought> !grab thePiGrepper
[21:51:22] <phrik> deep42thought: Tada!
[21:51:25] <thePiGrepper> hahaha
[21:51:59] <thePiGrepper> I always wonder about that grab. that means that know, because I dont have any other, that one will pop up *always*??!
[21:52:13] <deep42thought> yes
[21:52:20] <deep42thought> !rq thePiGrepper
[21:52:21] <phrik> deep42thought: <thePiGrepper> yeah. exactly. no one needs emacs-asp, (until one does...)
[21:52:25] <deep42thought> !rq thePiGrepper
[21:52:26] <thePiGrepper> hahahaah
[21:52:26] <phrik> deep42thought: <thePiGrepper> yeah. exactly. no one needs emacs-asp, (until one does...)
[21:52:40] <deep42thought> !rq deep42thought
[21:52:41] <phrik> deep42thought: <deep42thought> good night, python - cu tomorrow python2!
[21:52:57] <thePiGrepper> good! :-)
[21:53:39] <thePiGrepper> well then. then log32 it is
[21:53:59] <deep42thought> yeah, and difflog32 and shortlog32
[21:54:14] <deep42thought> or do you have any objections?
[21:55:07] <thePiGrepper> hmm. let's see, CLOC-wise writing *log -u would be less lines than writing *log32 commands
[21:55:53] <thePiGrepper> as in *log = *log32, *log -u = *log(regular old one)
[21:56:02] <deep42thought> ah, you mean, we could say "log" is only for our stuff and "-u log" shows the upstream changes
[21:56:06] <thePiGrepper> right
[21:56:17] <deep42thought> yeah, why not
[21:56:22] <deep42thought> we have -u anyways
[21:56:33] <thePiGrepper> right, we better use it.
[21:56:38] <deep42thought> although, I bet abaumann trips over this the first couple times ;-)
[21:56:48] <deep42thought> (no offense)
[21:57:05] <deep42thought> I will probably also trip over that ...
[21:57:10] <thePiGrepper> you can always alias asp32='asp32 -u' I guess
[21:57:33] <deep42thought> yeah, but I'll need the w/o -u version for usual checkout, etc. much more often
[21:57:34] <deep42thought> anyways
[21:57:44] <deep42thought> tripping the first couple times is no big deal, I think
[21:59:25] <thePiGrepper> well, (I think) one uses more checkout/export than log. hmm, it probably depends..
[21:59:42] <deep42thought> I like the idea with the -u
[22:00:21] <thePiGrepper> ok then, let's keep it that way, and see what happens.
[22:00:27] <thePiGrepper> can you add that issue?
[22:00:45] <deep42thought> I'm on it
[22:00:59] <thePiGrepper> I think that leaves us with show and ls-files
[22:01:22] <thePiGrepper> the issue with those is that they need to display the 'new' files that are only in the arch32 patches
[22:01:44] <thePiGrepper> and in the case of show, the autocomplete needs to work for arch32 as well
[22:02:10] <deep42thought> the autocomplete of show uses the same data which ls-files displays, right?
[22:02:22] <thePiGrepper> right
[22:02:50] <thePiGrepper> so, it might(I need to check the code) be mostly the same solution
[22:03:10] <deep42thought> yep
[22:03:13] <deep42thought> I hope so :-)
[22:03:30] <thePiGrepper> I'll be afk for an hour or so now..
[22:03:43] <deep42thought> ok, I'll probably go to bed until then
[22:04:00] <thePiGrepper> cu later then. I'll take a look at the issues later
[22:04:23] <deep42thought> ok, cu later!
[22:04:44] <thePiGrepper> and Ill try to post a PR for the new asp PKGBUILD
[22:05:15] <deep42thought> thank you for your work!
[22:05:29] <thePiGrepper> if you have any other question/issue/suggestion/etc leave it here or in the issue itself.
[22:05:41] <thePiGrepper> hehe, np, it's fun :-)
[22:28:33] -!- deep42thought has quit [Quit: Leaving.]
[23:35:56] -!- autofsckk has joined #archlinux32