#archlinux32 | Logs for 2024-03-03
Back
[00:18:16] -!- bdju has quit [Ping timeout: 268 seconds]
[00:20:07] -!- bdju has joined #archlinux32
[00:28:01] -!- bdju has quit [Ping timeout: 264 seconds]
[01:17:50] -!- drathir_tor has quit [Ping timeout: 260 seconds]
[01:24:36] -!- drathir_tor has joined #archlinux32
[02:36:56] -!- sL1pKn07 has joined #archlinux32
[02:53:59] -!- eschwartz has quit [Ping timeout: 256 seconds]
[02:58:34] -!- bdju has joined #archlinux32
[03:06:02] -!- eschwartz has joined #archlinux32
[03:26:35] -!- zxrom has quit [Quit: Leaving]
[03:51:22] -!- drathir_tor has quit [Ping timeout: 260 seconds]
[03:58:10] -!- drathir_tor has joined #archlinux32
[04:23:59] -!- drathir_tor has quit [Remote host closed the connection]
[04:24:32] -!- drathir_tor has joined #archlinux32
[04:26:01] -!- drathir_tor has quit [Remote host closed the connection]
[04:29:39] -!- drathir_tor has joined #archlinux32
[05:59:43] <KitsuWhooa> KillerWasp: try updating archlinux32-keyring first
[05:59:55] <KitsuWhooa> And the reason I mentioned SSE is because i686 requires SSE
[06:00:09] <KitsuWhooa> so if all CPUs that support SSE also support PAE then there won't be any regressions
[06:17:19] <KitsuWhooa> Right, buildmaster went insane. Just fixed it
[07:47:26] -!- abaumann has joined #archlinux32
[07:47:27] <buildmaster> Hi abaumann!
[07:47:27] <buildmaster> !rq abaumann
[07:47:27] <phrik> buildmaster: <abaumann> I never used zen at all :-)
[07:48:08] <abaumann> KitsuWhooa: the only purpose of having a linux-pae is so that the default kernel doesn't have PAE. There are many old CPUs around which don't do PAE, in the i686 range.
[07:48:19] <KitsuWhooa> yeah, okay
[07:48:33] <abaumann> Anyway: PAE is needed to build the kernel, and only the kernel.
[07:48:44] <abaumann> There is no reason why it has to be enabled for any other package.
[07:49:01] <KitsuWhooa> I thought if it was just i486 then we could get away with it
[07:49:09] <KitsuWhooa> "enabled for any other package" ?
[07:49:14] <KitsuWhooa> you can't enable it for packages though
[07:49:16] <KitsuWhooa> it's just a kernel thing
[07:49:29] <abaumann> you can, if you set some global cflags
[07:49:46] <KitsuWhooa> you can't make a 32 bit process use larger pointers though, right?
[07:49:49] <abaumann> it will hopefully not generate any code containing something PAEish :-)
[07:49:57] <KitsuWhooa> either way I was talking about just the kernel
[07:50:59] <KitsuWhooa> Also, I can't tell if it's something I did that broke things, but pahole segfaults when building the kernel
[07:51:18] <KitsuWhooa> I ran get-package-updates again so we'll see if it still fails after it tries again
[07:51:28] <KitsuWhooa> And I can't rebuild pahole because it wants python
[07:51:37] <KitsuWhooa> well, I could patch it, but I really should go back to trying to get python to work
[07:51:42] <abaumann> strange. I saw some packages of linux-pae on the buildmaster?
[07:51:56] <KitsuWhooa> tar: This does not look like a tar archive
[07:51:56] <KitsuWhooa> tar: extra/linux/PKGBUILD: Not found in archive
[07:51:58] <KitsuWhooa> what just happened
[07:52:09] <KitsuWhooa> my builder just spit that out
[07:52:10] <abaumann> -rw-r--r-- 2 http mirror 10866426 Jul 9 2023 linux-pae-headers-6.4.2-1.0-pentium4.pkg.tar.zst
[07:52:13] <abaumann> ah, old ones
[07:52:15] <KitsuWhooa> that is old
[07:52:57] <KitsuWhooa> Do the "upstream git revision" comments mean anything?
[07:53:02] <KitsuWhooa> Maybe that broke things?
[07:53:16] <KitsuWhooa> as in
[07:53:16] <KitsuWhooa> core/linux$ cat PKGBUILD
[07:53:16] <KitsuWhooa> # upstream git revision: 3ee8900562726ca75feddde13bef02b5d046938d
[07:53:41] <abaumann> no, that is just to know which upstream version you are referencing in the PKGBUILD, it's documentation
[07:53:45] <KitsuWhooa> okay
[07:53:46] <abaumann> ..and not up-to-date
[07:54:03] <KitsuWhooa> so why on earth did my builder fail to find the PKGBUILD
[07:54:03] <KitsuWhooa> wait
[07:54:05] <KitsuWhooa> extra/linux?
[07:54:05] <KitsuWhooa> oh god
[07:54:15] <abaumann> should be core/linux
[07:54:18] <KitsuWhooa> Yeah
[07:54:28] <KitsuWhooa> I did a stupid earlier when trying to get extra/pae-linux working
[07:54:41] <KitsuWhooa> sorry I mean extra/linux-pae
[07:54:51] <KitsuWhooa> I didn't fully rename it, and it thought the PKGBUILD for extra/pae-linux would produce extra/linux
[07:55:01] <KitsuWhooa> so I guess we can just forget about it and ignore it :p
[07:55:03] <abaumann> you have to be careful here.
[07:55:22] <KitsuWhooa> I mean, I tried. I didn't notice it despite building it locally
[07:55:23] <abaumann> the patching mechanism kicks in later, patching a diff pkgname could cause things like this.
[07:55:29] <KitsuWhooa> oh I did much more cursed things
[07:55:33] <KitsuWhooa> (which didn't work)
[07:55:53] <abaumann> if a diff from linux to linux-pae would be possible, I have the feeling, it would be one :-)
[07:56:08] <abaumann> might be a reason why it was a copy-paste of an earlier linux package
[07:56:09] <KitsuWhooa> What I did was, take the upstream linux PKGBUILD, and try to patch it, to get linux-pae working
[07:56:40] <abaumann> so, you have to patch pkgname in a certain place, and if the system sees that too late, it builds for 'linux' instead of 'linux-pae'
[07:56:44] <KitsuWhooa> Yeah
[07:56:47] <KitsuWhooa> that's exactly what happened
[07:57:02] <KitsuWhooa> I figured it out after buildmaster had picked up on it
[07:57:18] <KitsuWhooa> is there something I can to do to make it forget about extra/linux, or do we just leave it like this?
[07:57:19] <abaumann> no problem, core/linux takes precedence in the repos
[07:57:27] <KitsuWhooa> it didn't build anything, anyway
[07:57:29] <KitsuWhooa> ahj
[07:57:30] <KitsuWhooa> *ah
[07:57:39] <abaumann> nope, I don't think so.
[07:57:47] <abaumann> Unless manipulating the database and the repos by hand.
[07:58:52] <KitsuWhooa> The insanity from earlier was unrelated, by the way
[07:59:08] <KitsuWhooa> it was three outdated pentium4 packages in the pool that weren't in the db, so I just deleted them
[07:59:38] <abaumann> that's a race, yes. has nothing to do with what you put into the packages, it's just random
[08:01:40] <abaumann> mmh. I have a wiped backup tape of the buildmaster, bacula just choked on it.
[08:01:52] <abaumann> nice. and the last full backup is from October 2023..
[08:02:05] <abaumann> ..because the full backup in January 2024 also failed.
[08:02:10] <KitsuWhooa> mmmm
[08:02:24] <abaumann> this is completely unrelated. :->
[08:02:33] <abaumann> I have a rsync on another machine, so no worries. :-)
[08:12:42] mavica is now known as mavicaway
[08:27:20] -!- mvchtz has joined #archlinux32
[08:49:58] <KitsuWhooa> 2024-03-03 08:49:47: building package "pahole" (revisions 8c18fc9029d94a4924eadc8c749cca52a5584783 0000000000000000000000000000000000000000, repository extra, straw :mirrored_source_by_hash:) for pentium4 ...
[08:50:00] <KitsuWhooa> good luck little builder
[08:50:33] <KitsuWhooa> ...oh wow it built
[09:14:06] <KitsuWhooa> okay, buildmaster thinks every time I touch linux-pae that linux has been modified as well
[09:14:09] <KitsuWhooa> this is an issue
[09:28:21] -!- mvchtz has quit [Quit: WeeChat 3.5]
[09:31:32] <KitsuWhooa> what's also an issue is that linux fails to build
[09:51:02] -!- titus_livius has joined #archlinux32
[11:02:33] -!- zxrom has joined #archlinux32
[11:42:16] -!- n0tiz has quit [Quit: Bye]
[11:43:54] -!- n0tiz has joined #archlinux32
[13:56:58] -!- abaumann has quit [Quit: leaving]
[14:52:51] -!- zxrom has quit [Remote host closed the connection]
[14:57:46] -!- zxrom has joined #archlinux32
[18:20:43] <KitsuWhooa> This is exhausting
[18:20:49] <KitsuWhooa> is pahole broken? Who knows
[20:12:25] mavicaway is now known as mavica
[23:19:11] -!- bdju has quit [Ping timeout: 264 seconds]
[23:41:10] -!- bdju has joined #archlinux32
[23:58:44] -!- sunshavi has quit [Ping timeout: 260 seconds]
[23:58:47] -!- bdju has quit [Ping timeout: 264 seconds]