#archlinux32 | Logs for 2023-12-22
Back
[02:46:31] -!- drathir_tor has quit [Ping timeout: 240 seconds]
[03:44:50] -!- drathir_tor has joined #archlinux32
[04:49:19] -!- GNUtoo has quit [Ping timeout: 240 seconds]
[04:54:57] -!- GNUtoo has joined #archlinux32
[05:15:00] -!- morriset has joined #archlinux32
[07:31:24] -!- abaumann has joined #archlinux32
[07:31:24] <buildmaster> Hi abaumann!
[07:31:24] <buildmaster> !rq abaumann
[07:31:26] <phrik> buildmaster: <abaumann> This will make guys happy, I actually checked the knowledge base of bugs first.. ;-)
[07:32:12] <abaumann> KitsuWhooa, girls: that's also why I stopped mine, basically everything fails with something "pythonesque" (or "ruby-ish", or "npm-ish", ...) :-)
[07:32:30] <KitsuWhooa> Which is why I'm trying to fix it
[07:32:46] <KitsuWhooa> If I can get the builders to stop deleting those packages, it'll work :p
[07:33:10] <abaumann> We have to agree on a manual build way. It's simply not feasable to feed the buildmaster then wait till it fails or succeeds to build some bootstrapping package.
[07:33:11] <KitsuWhooa> I don't suppose you know where the list of expected packages comes from in buildmaster :p
[07:33:28] <abaumann> no. I don't know the buildmaster at all.
[07:33:47] -!- carloratm has parted #archlinux32
[07:33:50] <KitsuWhooa> Mmm
[07:34:21] <KitsuWhooa> Yeah a way to manually add packages in build support will make life easier
[07:34:26] <abaumann> I did some reverse engineering on the upstream new git state building way, and I get more and more confused too.
[07:34:47] <abaumann> I was thinking about a manual build, add it to the buildmaster database and to the repository
[07:34:58] <KitsuWhooa> Yeah, that'd work
[07:35:06] <KitsuWhooa> But I'd rather it not get to end users
[07:35:19] <abaumann> yes, it should be strictly for build-support and staging
[07:35:22] <KitsuWhooa> This the only repo it should be done in should be build supper
[07:35:23] <KitsuWhooa> Support
[07:35:25] <KitsuWhooa> Yup
[07:35:43] <abaumann> though forcing packages from staging->testing->stable will always be necessary
[07:35:47] <abaumann> and is possible
[07:36:38] <abaumann> another option is to have build slaves with python and base modules built and maintained by hand without using the packages. This makes things easier, but also less reproducable (if reproducable at all).
[07:37:02] <abaumann> that's what I did in the past on major Python updates.
[07:37:27] <abaumann> just forget about the archlinux packages, because you just invite problems when packaging thousands of them.
[07:38:19] <abaumann> otoh, having specific depends sedfu to remove everything but the most important dependencies during bootstrap is also an option, so things are documented.
[07:39:04] <abaumann> I want to understand the downgrade of packages first (like 'filesystem'), this is an issue with ambiguities in the upstream state git repos.
[07:39:50] <abaumann> let's see.. me go to train into snowy weather now.. :-)
[07:40:29] <abaumann> cu
[07:40:30] -!- abaumann has quit [Quit: leaving]
[07:48:49] -!- titus_livius has joined #archlinux32
[07:50:56] -!- buildmaster has quit [Remote host closed the connection]
[07:53:49] -!- buildmaster has joined #archlinux32
[08:00:48] <KitsuWhooa> Late, but see ya
[08:50:18] <KitsuWhooa> So, is buildmaster upstream arch software? I guess I need to find out and bug them if so
[09:20:31] -!- drathir_tor has quit [Ping timeout: 240 seconds]
[09:21:11] -!- drathir_tor has joined #archlinux32
[16:41:51] -!- epony has quit [Remote host closed the connection]
[19:59:29] -!- morriset has quit [Quit: Leaving]
[22:30:17] -!- epony has joined #archlinux32