2026-08-02 06:45:57 <mps> is it possible to extend time on gitlab CI for MR which is already there
2026-08-02 18:14:14 <mps> today I tested armbian on musebook in gui mode. alpine is substantially faster or to better say armbian is too slow
2026-08-02 20:47:45 <shadow> I'd like any rust coders running Alpine on riscv64 to verify some upstream promotion of official rust tools
2026-08-02 20:48:27 <shadow> it's scheduled for rust 1.99 release and is probably okay but more testing is always more better
2026-08-03 04:46:22 <ikke> The overlap of people working actively with rust and running Alpine riscv64 is probably not that big
2026-08-03 06:08:12 <mps> rscv64 builder whole weekend trying to build one package (testing/php86-pecl-xlswriter) in loop
2026-08-04 07:16:19 <mps> there is new https://github.com/OpenKermit/ckermit/releases, maybe upgrade can fix builds
2026-08-04 08:55:21 <ikke> Changelog contains no indication tests were fixed
2026-08-05 10:03:06 <mps> I'd like to merge !106159 but not sure is it ok
2026-08-05 10:04:13 <mps> have fear that it will block builders for few days and will need lot fixes
2026-08-05 10:11:54 <ncopa> have you run a local rebuild of all of it?
2026-08-05 10:12:22 <ikke> ncopa: do you want to make an edge snapshot first? (#18378)
2026-08-05 10:12:59 <ncopa> yeah, we probably should
2026-08-05 10:18:08 <ikke> Need to wait for the builders to get idle
2026-08-05 11:03:21 <mps> ncopa: is it possible somehow to run build of merge request which have multiple commits and more than one package
2026-08-05 11:03:55 <mps> i mean locally
2026-08-05 11:04:40 <mps> because CI timeouts with many packages to build, time limit
2026-08-05 11:06:19 <mps> this MR have 309 pkgs to rebuild
2026-08-05 11:09:16 <ncopa> yeah. you checkout the branch and build the modified packages in build order
2026-08-05 11:10:24 <ncopa> I rebased the branch against origin/master and then:
2026-08-05 11:10:29 <mps> running one-by-one with abuild -r
2026-08-05 11:10:33 <ncopa> ncopa-edge-x86-64:~/aports/main (perl-5.44.0)$ for i in $(git diff --name-only origin/master -- . | cut -d/ -f2 | xargs ap builddirs); do abuild -C "$i" -rk || break; done
2026-08-05 11:11:07 <mps> ah
2026-08-05 11:11:58 <mps> will look and try
2026-08-05 11:23:20 <ncopa> i think you can just merge it once you can confirm that all the 309 packages builds locally on one architecture
2026-08-05 11:44:26 <mps> aha, I thought same
2026-08-05 12:06:22 <mps> ncopa: in which shell you run above 'script'
2026-08-05 12:09:25 <mps> in bash and ash it gives me this https://tpaste.us/EX4W and exits
2026-08-05 12:31:43 <mps> ok, i have script by which all these commits are added one-by-one in this MR. will adapt it to rebuild all these packages
2026-08-05 15:56:05 <ikke> mps: that's not a shell issue
2026-08-05 15:56:20 <ikke> you just already built perl and abuild exits with an error code, which trips the || break
2026-08-05 15:56:30 <ikke> s/error/exit
2026-08-05 16:29:41 <ncopa> maybe add `-d ~/aports/main ` to the ap command so the cwd does not matter
2026-08-05 17:11:39 <mps> i'm building now with script https://tpaste.us/aWBb
2026-08-05 17:12:33 <mps> main and community finished. testing is building now
2026-08-05 17:25:25 <mps> failed pkgs in my test https://tpaste.us/qbk4 on aarch64
2026-08-05 17:49:48 <mps> >>> freeradius: No update needed for ./config.sub
2026-08-05 17:55:00 <mps> git fails on check
2026-08-05 17:55:26 <mps> pcsc-lite fails 'curl: (22) The requested URL returned error: 404'
2026-08-05 17:57:38 <mps> and running `abuild -r` by hand for texinfo pass
2026-08-05 17:59:57 <mps> looks like not much work to fix these failed pkgs
2026-08-05 18:09:05 <ikke> builder is idle
2026-08-05 18:11:06 <ncopa> oh! let me tag
2026-08-05 19:06:10 <mps> 'ready to merge'. merge it or not?
2026-08-05 19:45:13 <mps> merged
2026-08-05 19:46:27 <ikke> Limiting output to 20 commits out of 302.
2026-08-05 19:54:27 <mps> build-edge-armv7 is already over 20
2026-08-05 19:55:30 <ikke> First build failure
2026-08-05 19:55:36 <ikke> pcsc-lite
2026-08-05 19:55:50 <mps> can't fetch source
2026-08-05 19:56:26 <ikke> 494
2026-08-05 19:56:28 <ikke> 404
2026-08-05 19:58:05 <mps> oh, new version is on site and old version is deleted, so needs upgrade
2026-08-05 19:59:44 <ikke> sometimes they are moved
2026-08-05 20:01:07 <mps> i upgraded it and pushed
2026-08-05 20:05:17 <ikke> freeradius no longer needs update_config_sub
2026-08-05 20:15:42 <mps> yes, i noticed it
2026-08-05 20:17:26 <mps> testing it now
2026-08-05 20:20:38 <mps> pushed update
2026-08-05 20:22:27 <mps> didn't bumped pkgrel for not finished pkgs, not sure it is needed
2026-08-05 20:22:39 <ikke> it's not
2026-08-05 20:24:00 <mps> good
2026-08-05 20:25:31 <mps> armv7 is fastest
2026-08-05 20:27:38 <ikke> git has test failures :(
2026-08-05 20:28:08 <ikke> 3 in total
2026-08-05 20:30:00 <mps> git passed on aarch64 if i correctly have seen
2026-08-05 20:47:53 <ikke> yet, armhf finished main first :P
2026-08-05 20:52:29 <mps> :)
2026-08-05 21:22:31 <iggy> more stuff disabled for that arch?
2026-08-05 21:27:30 <mps> time to sleep, gn
2026-08-05 21:27:49 <mps> ikke: thank you very much for help
2026-08-06 03:41:26 <ikke> iggy: yes
2026-08-06 06:48:50 <mps> mio: thank you for the help with perl upgrade
2026-08-06 06:50:35 <mps> ikke: testing/perl-net-curl hangs on all arches so I pushed commit to disable check temporary for it
2026-08-06 06:51:26 <mps> builders should be restarted because now they are blocked
2026-08-06 10:32:30 <mps> can someone restart edge builders
2026-08-06 13:52:37 <mio> mps: you're welcome
2026-08-07 06:18:16 <mps> edge builders still stuck