=== lionel626 is now known as lionel62 | ||
RikMills | MoM has been broken for a while it seems | 05:36 |
---|---|---|
RikMills | Generated at 2024-08-24 17:51:54 UTC. | 05:36 |
RikMills | bdmurray: is that in your purview? I need to write down who has admins over what things :/ | 05:37 |
Skia | RikMills: I'll have a look | 08:51 |
Skia | RikMills: there was a stall process, it should be running in the next half-hour | 08:59 |
RikMills | Skia: great. thanks! | 09:40 |
=== mfo_ is now known as mfo | ||
RikMills | jbicha: you synced fuse from debian which drops the fuse package. yet many packages depend on the fuse. so as fuse is unistallable now we have iso build failures and packages with broken dependencies | 13:32 |
RikMills | kubuntu livefs fail | 13:32 |
RikMills | The following packages have unmet dependencies: | 13:32 |
RikMills | fuse : Depends: libfuse2t64 (= 2.9.9-8.2) but it is not going to be installed | 13:32 |
RikMills | E: Unable to correct problems, you have held broken packages. | 13:32 |
RikMills | Install failed | 13:32 |
RikMills | and trying to manually install | 13:33 |
RikMills | The following packages have unmet dependencies: | 13:33 |
RikMills | fuse : Depends: libfuse2t64 (= 2.9.9-8.2) but 2.9.9-9 is to be installed | 13:33 |
RikMills | libfuse2t64 : Conflicts: fuse (< 2.9.9-9) | 13:33 |
jbicha | RikMills: the fuse3 binary package Provides: fuse | 13:38 |
jbicha | I'm going to try to find an Archive Admin to remove the binary package fuse to see if that helps | 13:40 |
jbicha | see bug 1978310 for why this helps that this was done for oracular | 13:40 |
-ubottu:#ubuntu-devel- Bug 1978310 in fuse (Ubuntu) "jammy,noble: Installing fuse package removes other important packages" [High, Triaged] https://launchpad.net/bugs/1978310 | 13:40 | |
RikMills | jbicha: ah, didn't see that provides as I only checked universe. thanks! | 13:43 |
jbicha | yes, the fuse situation is a mess, it's bad that it affects both jammy & noble :( | 13:44 |
RikMills | indeed :/ | 13:44 |
jbicha | maybe adding the original Provides was a bad idea? | 13:45 |
jbicha | actually fuse3 Provides, Conflicts, and Breaks fuse which sounds non-sensical to me. I think we need to do a fuse3 upload, perhaps making the Conflicts & Breaks versioned | 13:46 |
jbicha | hmm | 13:47 |
jbicha | https://www.debian.org/doc/debian-policy/ch-relationships.html#s-virtual says that we can have Provides & Conflicts simultaneously so maybe it works as is once fuse is removed | 13:52 |
RikMills | one way to find out | 13:53 |
jbicha | ok, if fuse is installed, apt then doesn't want to upgrade libfuse2t64 unless we specifically command it to, cc/ @juliank | 13:55 |
jbicha | converting real package to virtual (Provides) without a transitional package may be a bad idea | 14:00 |
juliank | jbicha: I'm not sure I follow | 14:18 |
juliank | Ah libfuse2t64 cannot be upgraded to 2.9.9-9 because fuse is installed at 2.9.9-8.2 | 14:20 |
juliank | That's to be expected of sorts, it really dpeends on who is more important when resolving the conflict | 14:21 |
juliank | fuse3 doesn't really play into that at all | 14:21 |
juliank | But yes it is true that you should have a transitional Package: fuse Depends: fuse3 | 14:22 |
juliank | in Section: oldlibs | 14:22 |
juliank | APT does not follow renames, so to speak | 14:22 |
jbicha | juliank: ok, I'll restore the fuse binary package, this time as a transitional package and ask Debian to do the same | 14:24 |
juliank | jbicha: You need to carefully replace the Conflicts in fuse3 with a versioned Breaks | 14:25 |
juliank | So if we re-add fuse in 2.9.9-10 it would be fuse3 Breaks: fuse (<< 2.9.9-10~) | 14:25 |
juliank | jbicha: Alternatively of course you can just ship the transitional package from the fuse3 source package, this may be easier | 14:25 |
juliank | Easier to understand for the user anyhow, because it has the correct version :) | 14:26 |
jbicha | we'd still want to change Conflicts to versioned Breaks if we do it all in the fuse3 source, right? | 14:26 |
juliank | Yes | 14:26 |
juliank | but you can then use Breaks: fuse (<< ${source:Version}) really | 14:27 |
juliank | it will be tigher than needed but less thinking | 14:27 |
jbicha | juliank: I think once we have a transitional package, the versioned Provides aren't helpful | 14:57 |
juliank | jbicha: We still want the versioned provides such that we can drop the transitional package for 25.04 | 15:06 |
juliank | jbicha: the usual way is to introduce the new package with provides+replaces+breaks old-package (<< last-version-of-it), and then introduce old-package transitional package | 15:07 |
jbicha | juliank: could you verify that the patch I sent to Debian & CC'd you on looks good then? | 15:09 |
jbicha | RikMills: would you like to handle packaging the snapd-glib 1.66 update? | 15:21 |
RikMills | jbicha: is there an urgency in the next few days? I'm fire fighting a few issues in the run up to the beta freeze | 15:24 |
RikMills | so beta ISO is good for testing | 15:25 |
RikMills | I would like to do it if possible | 15:26 |
jbicha | RikMills: wasn't snapd-glib urgent for Kubuntu? or did y'all do a workaround? | 15:42 |
RikMills | jbicha: I have a fixed 1.65 in a landing ppa. Just waiting on a core dev to ./copy-package it in an hr of so | 15:44 |
RikMills | then I could look at 1.66 | 15:45 |
jbicha | RikMills: oh I see, well if you patch 1.65, we could just leave 1.66 to be done for 25.04 instead | 15:47 |
RikMills | jbicha: yeah. I'll still have a look at the upstream changes to see if there is a good reason to take 1.66 though | 15:49 |
RikMills | did not know it was out until you mentioned it | 15:49 |
=== mateus-morais6 is now known as mateus-morais | ||
=== beto_ is now known as beto | ||
=== klebers_ is now known as klebers | ||
Skia | RikMills: just to let you know that MoM is still running fine. Didn't update the HTML yet, but has been making consistent progress throughout the day | 16:11 |
=== sforshee_ is now known as sforshee | ||
=== matttbe9 is now known as matttbe | ||
ahasenack | doko: if around, and since you just uploaded python3-defaults, would you also be able to fix https://bugs.launchpad.net/ubuntu/+source/python3-defaults/+bug/2078356/comments/62 ? I worry about the beta freeze on monday, perhaps it's not doable. The fix seems simple enough (define "errors" as local, or just set it to empty in the beginning of the script) | 18:24 |
-ubottu:#ubuntu-devel- Launchpad bug 2078356 in python3-defaults (Ubuntu Oracular) "python3 preinst failure to run pre-rtupdate hooks" [High, Confirmed] | 18:24 | |
ahasenack | the comment in the link has the root cause analysis | 18:24 |
=== john-cabaj1 is now known as john-cabaj | ||
doko | ahasenack: committed to the VCS (and overrrode tumbleweed's change). I'm traveling until Wednesday. Feel free to upload, however please contact the QA team to cancel the currently running autopkg tests. I honestly don't think this is important enough to fix ... however waveform is preparing a python3-defaults update to noble, so please contact him to include that change | 19:31 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!