=== coreycb1 is now known as coreycb | ||
paride | @pilot in | 12:26 |
---|---|---|
=== ChanServ changed the topic of #ubuntu-devel to: Archive: open | Devel of Ubuntu (not support) | Build failures: http://qa.ubuntuwire.com/ftbfs/ | #ubuntu for support and discussion of Focal-Mantic | Patch Pilots: paride | ||
amikhalitsyn | Dear colleagues, can someone help with sponsoring upload for Noble https://bugs.launchpad.net/ubuntu/+source/criu/+bug/2066148 ? | 15:02 |
-ubottu:#ubuntu-devel- Launchpad bug 2066148 in criu (Ubuntu) "Ubuntu Noble lacks of CRIU package" [Undecided, New] | 15:02 | |
ahasenack | does anybody know what is wrong with software-properties' DEP8 tests in noble? | 15:18 |
ahasenack | 318s autopkgtest [06:54:20]: @@@@@@@@@@@@@@@@@@@@ apt-source software-properties | 15:18 |
ahasenack | 324s autopkgtest [06:54:26]: ERROR: erroneous package: rules extract failed with exit code 1 | 15:18 |
ahasenack | 324s blame: software-properties | 15:18 |
ahasenack | 324s badpkg: rules extract failed with exit code 1 | 15:18 |
ahasenack | between that and infra failures, it's been a bad week... | 15:19 |
paride | @pilot out | 15:48 |
=== ChanServ changed the topic of #ubuntu-devel to: Archive: open | Devel of Ubuntu (not support) | Build failures: http://qa.ubuntuwire.com/ftbfs/ | #ubuntu for support and discussion of Focal-Mantic | Patch Pilots: N/A | ||
paride | ahasenack, did that happen more than once? | 15:53 |
paride | ahasenack, not that once is ok | 15:53 |
ahasenack | yes, in multiple architectures | 15:53 |
ahasenack | https://autopkgtest.ubuntu.com/packages/s/software-properties/noble/amd64 for example | 15:54 |
paride | ahasenack, I just reproduced it locally | 15:54 |
ahasenack | 0.99.48 extracted fine with apt-get source here, noble | 15:56 |
paride | it tries to download a different version:\ | 15:56 |
paride | apt-get source -d -q --only-source software-properties=0.99.48.1 | 15:57 |
paride | and this fails: E: Can not find version '0.99.48.1' of package 'software-properties' | 15:57 |
ahasenack | where is 0.99.48.1 coming from? | 15:57 |
ahasenack | noble has 0.99.48, and noble-proposed has 0.99.49 | 15:57 |
paride | ahasenack, checking | 15:58 |
ahasenack | juliank: do you know something abouv this? I vaguely remember a comment from you in an SRU that mfo was checking | 15:59 |
juliank | paride, ahasenack So it gets confused by software-properties-qt temporarily existing in its own source package in a 0.99.48.1 | 16:01 |
juliank | Well I say temporarily but that is in the release pocket | 16:01 |
juliank | and it seems to use that to determine the version of software-properties source package to download | 16:01 |
dbungert1 | @pilot in | 16:01 |
=== ChanServ changed the topic of #ubuntu-devel to: Archive: open | Devel of Ubuntu (not support) | Build failures: http://qa.ubuntuwire.com/ftbfs/ | #ubuntu for support and discussion of Focal-Mantic | Patch Pilots: dbungert1 | ||
=== dbungert1 is now known as dbungert | ||
paride | I wonder if some changes I introduced locally to the "find source package to download" algorithm would work better | 16:01 |
ahasenack | why would a different source package interfere here? | 16:02 |
ahasenack | just because it's a substring match? | 16:02 |
juliank | ahasenack: autopkgtest tries to find out the source version to download but it looks at the wrong binary package | 16:02 |
paride | ahasenack, it is not that simple | 16:02 |
juliank | something like that | 16:02 |
paride | ahasenack, let me test from master, i.e. with https://salsa.debian.org/ci-team/autopkgtest/-/merge_requests/324 merged in | 16:03 |
-ubottu:#ubuntu-devel- Merge 324 in ci-team/autopkgtest "Truly honor pinning when looking for the source package to download" [Merged] | 16:03 | |
ahasenack | paride: ok | 16:03 |
paride | ahasenack, still fail, same failure mode | 16:04 |
paride | *fails | 16:04 |
ahasenack | paride: looks like we need an autopkgtest bug, were you about to file one? If not, I can, with the info we have so far | 16:11 |
paride | ahasenack, that's indeed a bug in autopkgtest and not an obvious one | 16:12 |
paride | ahasenack, file against the autopkgtest ubuntu package or debian, as you prefer. | 16:13 |
ahasenack | ok | 16:14 |
krytarik | dbungert: Btw, with that nick change after piloting in, I don't think ubottu is going to handle your piloting out later on. >_< | 16:15 |
ahasenack | paride: where did you see it trying to download the wrong version? Only locally, or also in the logs on the server? | 16:16 |
paride | ahasenack, only locally | 16:16 |
ahasenack | paride: what was the command, just run the tests from software-properties, nothing else? | 16:16 |
paride | autopkgtest --debug -U software-properties -- qemu ~/path/to/autopkgtest-noble-amd64.img | 16:17 |
=== dbungert is now known as dbungert1 | ||
ahasenack | ta | 16:17 |
dbungert1 | @pilot out | 16:17 |
=== ChanServ changed the topic of #ubuntu-devel to: Archive: open | Devel of Ubuntu (not support) | Build failures: http://qa.ubuntuwire.com/ftbfs/ | #ubuntu for support and discussion of Focal-Mantic | Patch Pilots: N/A | ||
=== dbungert1 is now known as dbungert | ||
dbungert | @pilot in | 16:17 |
=== ChanServ changed the topic of #ubuntu-devel to: Archive: open | Devel of Ubuntu (not support) | Build failures: http://qa.ubuntuwire.com/ftbfs/ | #ubuntu for support and discussion of Focal-Mantic | Patch Pilots: dbungert | ||
dbungert | krytarik: thanks! | 16:17 |
krytarik | XD | 16:18 |
paride | ahasenack, I believe the fundamental inconsistency that tricks autopkgtes is that `apt-cache showsrc software-properties` has Package-List: software-properties-qt, but no bin:software-properties-qt package *from that source* is actually available | 16:21 |
ahasenack | d/control has it | 16:22 |
paride | fine, but autopkgtest looks at what apt knows | 16:22 |
ahasenack | and something built on amd64: https://launchpad.net/ubuntu/+source/software-properties-qt/0.99.48.1 | 16:22 |
paride | so if on Noble you: `apt install software-properties-qt=0.99.48` it will error out: E: Version '0.99.48' for 'software-properties-qt' was not found | 16:23 |
ahasenack | well, yes, it only exists at .1 | 16:23 |
paride | ahasenack, I have half of an idea on how to fix it | 16:27 |
ahasenack | and I haven't finished filing the bug yet even :) | 16:27 |
ahasenack | paride: here: https://bugs.launchpad.net/ubuntu/+source/autopkgtest/+bug/2066290 | 16:28 |
-ubottu:#ubuntu-devel- Launchpad bug 2066290 in autopkgtest (Ubuntu) "Attempts to fetch wrong version for package" [Undecided, New] | 16:28 | |
paride | ahasenack, thanks | 16:28 |
ahasenack | so wait a second, src:software-properties in noble still builds bin:software-properties-qt, and noble also has a src:software-properties-qt that ALSO builds bin:software-properties-qt? | 16:30 |
ahasenack | two different source packages building a binary package each with the same name | 16:31 |
paride | ahasenack, yes, and apparently the one with the highest version makes the other disappear | 16:36 |
ahasenack | brb, reboot | 16:43 |
paride | ahasenack, upstream ci is still partially broken :( see https://salsa.debian.org/paride/autopkgtest/-/pipelines/680532 | 16:59 |
ahasenack | paride: is that debian/sid or what? | 17:00 |
ahasenack | bookworm | 17:01 |
paride | ahasenack, what is broken is the image used in the CI jobs. Then the jobs themselves test autopkgtest in different scenarios. for example test-lxc-old-testbed test on a trusty testbed | 17:01 |
paride | as trusty is the oldest thing around | 17:01 |
paride | ahasenack, I already pinged around about the broken ci images, but got no reply | 17:02 |
ahasenack | would an apt -f install fix that, temporariluy? | 17:02 |
paride | hmm, I don't think so | 17:03 |
paride | ahasenack, you can try yourself by running: docker run -it registry.salsa.debian.org/salsa-ci-team/pipeline/autopkgtest | 17:03 |
paride | and then apt-get update && apt-get install libc6-dev | 17:03 |
ahasenack | hah, I just uninstalled docker | 17:03 |
ahasenack | will get a vm | 17:03 |
=== mfo_ is now known as mfo | ||
=== coreycb1 is now known as coreycb | ||
arraybolt3 | juliank: random question for you if you're around and have time - do you have any idea what generates the file /boot/efi/EFI/ubuntu/grub.cfg? Is this something that an installer generates or is it something that a package has control over and creates via a postinst script or something? | 19:44 |
juliank | arraybolt3: update-grub / grub-mkconfig creates thst | 19:45 |
arraybolt3 | not /boot/grub/grub.cfg, but /boot/efi/EFI/ubuntu/grub.cfg. | 19:45 |
arraybolt3 | It's a weird three-or-four line file that points GRUB to the real grub.cfg. | 19:45 |
arraybolt3 | and it does not seem to be updated when I run update-grub. | 19:46 |
arraybolt3 | just confirmed, I can move it to grub.cfg.bak and then run update-grub, and it is **not** regenerated. | 19:47 |
juliank | Ah sorry | 19:53 |
juliank | Grub-install creates that I believe when installing grub to the esp | 19:53 |
arraybolt3 | juliank: ah ok. Nothing ever updates it right? | 19:54 |
arraybolt3 | Asking because I'm designing a tool at my workplace that modifies that file for reasons, and I'm wanting to confirm that my changes won't be clobbered later | 19:54 |
arraybolt3 | though it sounds like a GRUB update might clobber it | 19:54 |
arraybolt3 | ah, `grub-install` does update it BUT it also regenerates it correctly. Fantastic. Thank you! | 19:55 |
dbungert | @pilot out | 21:41 |
=== ChanServ changed the topic of #ubuntu-devel to: Archive: open | Devel of Ubuntu (not support) | Build failures: http://qa.ubuntuwire.com/ftbfs/ | #ubuntu for support and discussion of Focal-Mantic | Patch Pilots: N/A | ||
=== bdmurray changed the topic of #ubuntu-devel to: Archive: open | Devel of Ubuntu (not support) | Build failures: http://qa.ubuntuwire.com/ftbfs/ | #ubuntu for support and discussion of Focal-Noble | Patch Pilots: N/A |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!