#archlinux32 | Logs for 2023-09-29

Back
[00:19:44] -!- titus_livius has joined #archlinux32
[00:20:05] -!- titus_livius has joined #archlinux32
[04:52:38] -!- abaumann has joined #archlinux32
[04:52:39] <buildmaster> Hi abaumann!
[04:52:39] <buildmaster> !rq abaumann
[04:52:40] <phrik> buildmaster: <abaumann> singularity applied to software development - releases and changes get faster and faster.. till..
[04:53:50] <abaumann> KitsuWhooa: I probably started memtester first with a little bit too much memory :-)
[04:54:38] <abaumann> it run fine 5 loops afterwards with I think 10 GB or so.
[04:55:19] <abaumann> A server is by definition something which runs 24x7 in my brain, the hardware underneath is quite irrelevant. Some of my most important servers are running on old Raspberry Pis. :-)
[04:56:00] <abaumann> A good desktop board can beat a lousy server board in all aspects
[04:56:20] <abaumann> chinese-calendar d0e397786e0be69774957394e227fcc68d4a3da6 0000000000000000000000000000000000000000 extra "make_source_info" did not return a "pkgbase" - eh, what?
[04:56:23] <abaumann> >==> ERROR: Cannot find the bzr package needed to handle bzr sources.<
[04:56:26] <abaumann> ><
[04:56:28] <abaumann> ><
[04:56:31] <abaumann> ah, this is almost normal behaviour for get-package-updates
[04:57:06] <abaumann> upstream has a little bit - aeh - consistency problems if it comes to what belongs into git repos and what not.
[04:57:15] <abaumann> I personally would add git hooks agressively..
[04:57:46] <abaumann> ..or propagate pkgctl as the only tool to manipulate the git repos..
[04:58:59] <abaumann> ah 'bzr' was dropped to the AUR and we have local adaptions which don't find the master entry
[04:59:54] <abaumann> huh? no bzr local adaption?
[05:00:40] <abaumann> Ah, or does PKGBUILD use bzr to get resources? And bzr is dropped to the AUR?
[05:01:43] <abaumann> how I miss the big PKGBUILD repo with all packages where I could just do a grep over everything..
[05:02:23] <abaumann> I personally was never a big fan of the git repo split, if git cannot handle it, maybe it has a scalability issue?
[05:02:41] <abaumann> Handling 30'000 text files in a repo with history should really be not a big issue..
[05:02:50] <abaumann> *not be
[05:03:10] <abaumann> question of the day: what package(s) are using source+bzr :-)
[05:03:44] <abaumann> ah, the PKGINFO cache.. :->
[05:04:23] <abaumann> ah and /srv/http/upstream-packages
[05:04:33] <abaumann> though you have to unpack the packages first
[05:05:07] <abaumann> /srv/http/pkginfo: all requested PKGINFO files
[05:06:07] <abaumann> source=("$pkgname::bzr+$url/trunk#revision=58")
[05:06:12] <abaumann> chinese-calendar
[05:06:37] <abaumann> yeah, it was in the output of get-package-updates all along.. *blush*
[05:07:43] <abaumann> ah: "we have to restart firefox in order you can continue to use firefox". Loving that. We used to call it a lousy job when not doing memory management properly: restart the thing if it runs out of resources.. :-)
[05:10:25] <abaumann> Tricky: I canned do early source package I presume and ignoring a package also needs the PKGBUILD to do so.
[05:10:52] <abaumann> Maybe the pkgcache generator can patch the PKGBUILD, but that feels dirty.
[05:11:26] <abaumann> Yeah, we do to much when fetching the PKGBUILD, it should basically be a 'pkgctl repo clone', not more.
[05:11:56] <abaumann> I'll patch by hand for now in the cache..
[05:12:57] <abaumann> makepkg --printsrcinfo
[05:13:06] <abaumann> aha, that one says: > ERROR: Cannot find the bzr package needed to handle bzr sources.
[05:14:33] <abaumann> source = https://launchpad.net
[05:14:39] <abaumann> solves the issue
[05:36:43] <abaumann> gloobus-preview: more bzr
[05:37:05] <abaumann> again launchpad
[05:37:35] <abaumann> https://launchpad.net has no downloads, great
[05:37:36] <phrik> Title: gloobus-preview in Launchpad (at launchpad.net)
[05:39:05] mavica is now known as mavicaway
[05:41:14] <abaumann> mmh. this could get tedious.
[05:45:12] <abaumann> mmh. we should make the thing more robust and handle errors of makepkg --printsrcinfo somehow..
[05:45:39] <abaumann> ..but the problem really is: you cannot simply ignore the package if the PKGBUILD has a bug
[06:05:47] <abaumann> aha. maybe something more works now and we even have the new pkgbuilds :-)
[06:08:41] <abaumann> https://buildmaster-status.archlinux32.org
[06:09:03] <abaumann> shows huge blue get-assignments, now we are back to normal small dots (green/red) on the bottom.
[06:09:07] <abaumann> I like graphs :-)
[06:15:23] <abaumann> libappindicator 7a77634d2eb4e285d52987a15210451ebb93a2e9 0000000000000000000000000000000000000000 extra "make_source_info" did not return a "pkgbase" - eh, what?
[06:15:26] <abaumann> >==> ERROR: Cannot find the bzr package needed to handle bzr sources.<
[06:15:33] <abaumann> after half an hour. lovely..
[06:15:48] <abaumann> speed, speed, speed :-)
[06:16:33] <abaumann> mmh, we could first fetch and compute all pkginfos, then do the database operations and checks.
[06:16:52] <abaumann> So get-package-updates would fail much faster in case of errors in pkginfos
[06:19:15] <abaumann> opencv=extra=88a014cf6597eeae081b2462feced39db88cf057=7288ebe7786463250e117edfa9491ed63b612201:==> ERROR: depends is not allowed to start with a hyphen.
[06:20:26] <abaumann> This sounds like sedfu going wrong
[06:28:17] <abaumann> every year the same blues: arch32 goes basically in hiatus during summer time (European summer time), gets almost abandoned till autumn, then weather gets nasty and people bored, so it gets revitalized. :-)
[06:37:35] <KitsuWhooa> abaumann: https://archlinux.org
[06:37:37] <phrik> Title: Arch Linux - breezy 3.3.4-1 (x86_64) (at archlinux.org)
[06:38:21] <KitsuWhooa> > 07:55:19 <abaumann> A server is by definition something which runs 24x7 in my brain, the hardware underneath is quite irrelevant. <-- until you have an intel board and want to use ecc :p
[06:39:57] <abaumann> aeh, yeah. ecc is nice, I admit that.
[06:40:04] <abaumann> raspberry pis don't have ecc :-)
[06:40:08] <abaumann> libdbusmenu: more bzr
[06:40:21] <KitsuWhooa> can you not use breezy?
[06:40:35] <KitsuWhooa> I assume the reason bazaar was removed is because breezy replaced it
[06:42:14] <abaumann> aha. good point. So it should actually have a provides=bzr and thus this is not an upstream issue but us having an old makepkg in devtools.
[06:42:32] <KitsuWhooa> it does
[06:42:34] <abaumann> Though in my brain in this case it should work because makepkg --printsrcinfo happens on the host, not in the 32-bit chroot
[06:42:47] <KitsuWhooa> that I don't know
[06:42:57] <KitsuWhooa> but it's probably better to figure out than try to patch every single package
[06:42:58] <KitsuWhooa> :p
[06:42:58] <abaumann> buildmaster still has a very old version of devtools32, this is not good
[06:45:29] <abaumann> So, merging devtools.. later, no time now..
[06:46:07] <abaumann> but the builds are working again, maybe with older pkgbuilds till get-package-updates runs to the end..
[06:46:10] <abaumann> cu
[06:46:12] -!- abaumann has quit [Quit: leaving]
[07:22:55] -!- zxrom has joined #archlinux32
[08:05:53] -!- Tampliera has joined #archlinux32
[08:06:17] <Tampliera> Hello all
[08:22:14] <T`aZ> sir
[09:12:56] -!- Tampliera has quit [Quit: Konversation terminated!]
[10:27:51] -!- drathir_tor has quit [Ping timeout: 252 seconds]
[10:29:55] -!- drathir_tor has joined #archlinux32
[16:22:10] -!- pulec has quit [Ping timeout: 258 seconds]
[16:26:25] -!- pulec has joined #archlinux32
[18:05:05] -!- sL1pKn07 has quit [Quit: Client Excited...]
[18:05:17] -!- sL1pKn07 has joined #archlinux32
[18:41:29] -!- Tampliera has joined #archlinux32
[19:04:06] -!- Tampliera has quit [Quit: Konversation terminated!]
[19:31:32] -!- swimm3r has quit [Ping timeout: 258 seconds]
[22:15:50] -!- kiska3 has quit [Quit: The Lounge - https://thelounge.chat]
[22:16:11] -!- kiska3 has joined #archlinux32
[22:21:39] -!- swimm3r has joined #archlinux32