ItzSwirlz | trying to figureout the cause of this: https://ubuntu-archive-team.ubuntu.com/cd-build-logs/ubuntucinnamon/mantic/daily-live-20230518.log | 00:10 |
---|---|---|
vorlon | ItzSwirlz: doesn't that show a successful build? | 00:14 |
vorlon | ah you're just looking at the message in the log; well, the code currently tries both bzr and git | 00:15 |
vorlon | because it doesn't have enough information to know which to use | 00:15 |
vorlon | also it looks for platform.mantic under ~ubuntucinnamon-dev first, then falls back to ~ubuntu-core-dev (where it actually lives) | 00:16 |
vorlon | this code could perhaps be improved to have better knowledge of where the seeds are and not have to try every combination with failures; but better would be to get all the flavors to migrate to git and drop the bzr code | 00:17 |
ItzSwirlz | where am i using bzr a tm | 00:21 |
ItzSwirlz | -meta uses regular vcs: https://github.com/ubuntucinnamon/ubuntucinnamon-meta/blob/aab8dd7ec216ad683b4cb8e37210597ea196be02/update.cfg#L16 | 00:22 |
vorlon | ItzSwirlz: this is the livecd-rootfs code, it's nothing specific to ubuntucinnamon | 01:33 |
=== guiverc2 is now known as guiverc | ||
slyon | jbicha: FYI: https://bugs.launchpad.net/ubuntu/+source/glib2.0/+bug/2026826 | 10:33 |
-ubottu:#ubuntu-devel- Launchpad bug 2026826 in glib2.0 (Ubuntu) "glib2.0 (2.77.0 ) breaks Netplan build" [Undecided, New] | 10:33 | |
seb128 | slyon, could you report the issue to glib upstream if you think it's a bug there? | 10:36 |
seb128 | slyon, and thanks for the report/tag. It wouldn't migrate anyway because of autopkgtests regressions but still better to be safe | 10:37 |
slyon | seb128: I'm not sure if it's an upstream issue or something the way our glib package is being compiled... I didn't do any deep investigation yet (and don't have time for it right now). I just confirmed that everything works fine with GLib from mantic-release and wanted to escalate to the destkop team for further investigation | 10:38 |
seb128 | slyon, ack, do you have a specific example for the keyfile line break issue described? | 10:39 |
seb128 | slyon, also https://autopkgtest.ubuntu.com/packages/n/netplan.io/mantic/amd64 ... any idea why the netplan autopkgtests aren't seeing the issue? | 10:40 |
slyon | seeing we're moving even ahead of Debian experimental, I thought it's worth bringing it up with Jeremy. | 10:40 |
seb128 | ack | 10:40 |
slyon | seb128: Yes, the broken keyfiles can be seen from netplan's build log. I will copy it to the bug report. | 10:40 |
seb128 | slyon, the issue is that the pkg-gnome team in Debian decided to keep unstable for bugfixes updates only until the incoming point release | 10:41 |
slyon | seb128: the autopkgtests are not using mantic-proposed, but rather mantic-release, I think? | 10:41 |
seb128 | slyon, the first entry on that page is 0.106.1-2 glib2.0/2.77.0-0ubuntu1 2023-07-10 23:13:20 UTC | 10:42 |
slyon | Interestingly, it seems to pass on riscv64, though, but fail on any other arch | 10:42 |
seb128 | slyon, riscv is building with notests | 10:43 |
slyon | oh, that explains the riscv situation! | 10:43 |
seb128 | slyon, so netplan -2 autopkgtests are green with glib 2.77.0, are the runtest not catching the issue or is only -3 failing? | 10:44 |
seb128 | let me trigger with -3 to see | 10:44 |
slyon | Right... I think 0.106.1-2 was still compiled with older glib. while 0.106.1-3 FTBFS with new glib. GLib is a shared library, so I'm not sure this really explains the autopkgtest pass.. :-/ | 10:44 |
seb128 | well -3 fails to build, ignore that :p | 10:45 |
slyon | seb128: rigth. | 10:45 |
seb128 | compiled with shouldn't make a difference though | 10:45 |
slyon | it's a build-time failure (dh_auto_test) | 10:45 |
seb128 | it's loading dynamically glib | 10:45 |
slyon | when I try to compile Netplan -2 with new glib, it fails, too. | 10:45 |
seb128 | ack, I would just expect that if glib changed in a way that break keyfiles then installed tests would also fail somehow | 10:46 |
seb128 | if that glib version is installed | 10:46 |
slyon | Yes, that's strange.. Maybe something else is in that build environment, which messes things up. But OTOH, just downgrading the GLib packages fixes the build | 10:47 |
seb128 | slyon, ack, we will investigate, thanks for the report | 10:49 |
slyon | thank you seb128 | 10:50 |
slyon | I've updated the bug report with an example keyfile | 10:50 |
seb128 | slyon, thanks! | 10:51 |
ginggs | there was a late (after FF?) glib2.0 update in lunar that also broke some builds | 11:12 |
ginggs | e.g. LP: #2019852 | 11:12 |
-ubottu:#ubuntu-devel- Launchpad bug 2019852 in nbd (Ubuntu Lunar) "nbd-server hangs after fork" [High, Fix Released] https://launchpad.net/bugs/2019852 | 11:12 | |
seb128 | ginggs, right, that was a different issue though (also that's one the reason we are landing the new serie earlier this cycle, avoiding issues just around feature freeze) | 11:38 |
ginggs | seb128: oh sure, not the same issue. but a test rebuild of reverse-build-deps before bumping glib2.0 would be nice | 11:39 |
seb128 | ginggs, that's something we could organize for doing if you/release team think there is value, I'm unsure we need a test rebuild specific for it so early in the cycle? | 11:52 |
seb128 | the build issues will get flagged with one of the regular test rebuilds no? | 11:53 |
ginggs | seb128: yes, they will | 11:55 |
ginggs | the FTBFS report only says the package FTBFS, it doesn't say glib2.0 broke this build | 12:04 |
ginggs | the 2nd lunar test rebuild was after the glib2.0 upload, and two failures that I know of (nbd and thrift) appear in that report | 12:06 |
ginggs | https://people.canonical.com/~ginggs/ftbfs-report/test-rebuild-20230324-lunar-lunar.html | 12:07 |
dbungert | @pilot in | 15:19 |
=== ChanServ changed the topic of #ubuntu-devel to: Archive: Mantic Open | Devel of Ubuntu (not support) | Build failures: http://qa.ubuntuwire.com/ftbfs/ | #ubuntu for support and discussion of Bionic-Lunar | Patch Pilots: dbungert | ||
wouter | ginggs: whoa, nbd was stuck for 18 hours?! Can I get the build log somewhere? | 17:47 |
wouter | oh, no, nvm -- that's the pre-fix version | 17:49 |
wouter | it's expected to do that, you just need the update which you helped me with a while back | 17:49 |
ginggs | wouter: exactly, yes | 17:50 |
wouter | ginggs: sorry for the noise -- I have a highlight configured on NBD, so I noticed ;) | 17:53 |
ginggs | wouter: no worries! | 17:53 |
ginggs | wouter: of course an autopkgtest in nbd would have caught that issue early ;) | 17:54 |
wouter | ginggs: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1036918 ;-) | 17:56 |
-ubottu:#ubuntu-devel- Debian bug 1036918 in debvm "debvm: manual mounting of root image" [Wishlist, Open] | 17:56 | |
ginggs | wouter: for a start, couldn't an autopkgtest just run the same tests that are run at build-time? | 17:58 |
wouter | "it's complicated". In theory yes, in practice that would require a wholesale refactoring of the test suite | 17:59 |
wouter | it's a good idea, for sure, but not as easy as it seems | 17:59 |
wouter | and yes, that refactoring is planned, but still | 17:59 |
dbungert | @pilot out | 19:19 |
=== ChanServ changed the topic of #ubuntu-devel to: Archive: Mantic Open | Devel of Ubuntu (not support) | Build failures: http://qa.ubuntuwire.com/ftbfs/ | #ubuntu for support and discussion of Bionic-Lunar | Patch Pilots: N/A | ||
ahasenack | dbungert: ocaml is quite a circular mess, did you conclude anything? | 20:59 |
dbungert | ahasenack: yep, making progress. 2 things needed retest and moved on, the next problem is LP: #2027333 | 21:00 |
-ubottu:#ubuntu-devel- Launchpad bug 2027333 in ocaml-dune (Debian) "ocaml-dune: Please add ocaml:Provides to d/control" [Undecided, New] https://launchpad.net/bugs/2027333 | 21:00 | |
dbungert | I sent a patch to Debian on the last part, we'll see if that gets picked up | 21:01 |
ahasenack | ok | 21:01 |
Unit193 | ahasenack: As promised, if it's not merged by the time a new upstream comes along I'll merge it, but I was looking over https://salsa.debian.org/debian/wireguard/-/merge_requests/6/diffs#diff-content-867caf2861d36d708e7ffe26f3783111183154cd again and it'd be more useful I think to echo those hash and "must be equal"/"must be different" lines into STDERR and only give them on error, to make it more | 21:44 |
Unit193 | obvious. I care less about the other autopkgtests since they won't run in Debian, but at the same time having the errors in STDERR makes sense there too. | 21:44 |
-ubottu:#ubuntu-devel- Merge 6 in debian/wireguard "New DEP8 and build-time tests" [Opened] | 21:44 | |
ahasenack | lemme switch context | 21:45 |
Unit193 | I was just trying to find a time when you were online, it's not time sensitive at all. :P | 21:45 |
ahasenack | oh, I had totally forgotten about that PR | 21:45 |
Unit193 | I hadn't! | 21:46 |
ahasenack | can you add that as a comment? Then I'll get an email and won't forget to act on it | 21:46 |
Unit193 | dkg seems very inactive there, otherwise I'd follow up about renaming the source package, dropping the meta, and dropping -modules and -dkms from recommends. | 21:46 |
Unit193 | Sure. | 21:47 |
ahasenack | thx | 21:47 |
ahasenack | I'm on +1 maintenance this week, so I will probably act on it next week | 21:47 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!