| vorlon | force-badtest autopkgtest/5.10ubuntu1/amd64 autopkgtest/5.10ubuntu1/i386 | 02:44 |
|---|---|---|
| vorlon | yay | 02:44 |
| -queuebot:#ubuntu-release- Unapproved: rejected ubuntu-release-upgrader [source] (disco-proposed) [1:19.04.16.7] | 07:44 | |
| Laney | vorlon: When removing packages per-architecture, can you start looking at reverse test-deps too please? | 08:50 |
| Laney | dbus-test-runner/ppc64el broken by bustle removal | 08:50 |
| -queuebot:#ubuntu-release- Unapproved: ubuntu-release-upgrader (disco-proposed/main) [1:19.04.16.6 => 1:19.04.16.7] (core) | 08:50 | |
| LocutusOfBorg | Suggests: bustle | 09:18 |
| LocutusOfBorg | Laney, ^^ you mean that? | 09:18 |
| LocutusOfBorg | its not even a real dependency... :/ | 09:18 |
| Laney | LocutusOfBorg: I said test-deps | 09:19 |
| LocutusOfBorg | so, the real build has no bustle support, and the test builds with it... | 09:19 |
| LocutusOfBorg | Laney, I was saying that the test is enabling features that nobody can use? | 09:19 |
| Laney | it's perfectly valid for a test to use its own dependencies | 09:19 |
| LocutusOfBorg | and I agree, but why test a feature that is not really built and used in the archive? | 09:20 |
| LocutusOfBorg | anyhow, I wish I could have a way with reverse-depends to pick up tests/control packages... | 09:20 |
| Laney | packages don't exist only to serve the Ubuntu archive | 09:21 |
| Laney | If a package offers a feature, it is legitimate to test that it works | 09:21 |
| LocutusOfBorg | it would be legitimate to build it too in the archive, unless I'm missing something obvious... | 09:22 |
| LocutusOfBorg | I'm not saying that "testing the matrix of various configurations" is something useless to do, but it is something that upstream should do, not Ubuntu :) | 09:23 |
| Laney | you're muddying the waters by attacking this case btw | 09:23 |
| LocutusOfBorg | this case is my fault, so I want to understand and possibly help to fix it | 09:24 |
| Laney | OK, you can use dbus-test-runner with bustle if you want to, it had a test to check that this functionality worked, bustle was removed so the test can't be run any more | 09:26 |
| LocutusOfBorg | ok, so we can change the test to not run with bustle? I wish I could fix bustle, I asked for help, but there is no upstream activity | 09:27 |
| LocutusOfBorg | even the new release is building a little bit more, but not up to the end | 09:28 |
| Laney | I mean that's what the test is, it's called 'with-bustle' ... | 09:28 |
| Laney | I'm not particularly saying that it was wrong to remove this binary (I don't know the case) | 09:29 |
| Laney | more asking for tests to be checked in the same way that in-archive reverse-build and reverse-binary deps are | 09:29 |
| LocutusOfBorg | [11:20:59] <LocutusOfBorg> anyhow, I wish I could have a way with reverse-depends to pick up tests/control packages... | 09:30 |
| LocutusOfBorg | this is why I said this ^^ I would like to avoid next time the same issue, to have a way to do it | 09:30 |
| Laney | laney@disco:~$ xzcat /srv/mirrors/ubuntu/dists/eoan/*/source/Sources.xz | grep-dctrl -FTestsuite-Triggers -sPackage bustle | 09:31 |
| Laney | Package: dbus-test-runner | 09:31 |
| LocutusOfBorg | thanks | 09:31 |
| LocutusOfBorg | now I have to find a way to download that file | 09:32 |
| Laney | you might want to use something like chdist | 09:34 |
| LocutusOfBorg | Laney, it might be looking bad, but it works https://paste.ubuntu.com/p/W9KVdRzhGQ/ | 09:51 |
| -queuebot:#ubuntu-release- New binary: webkit2gtk [amd64] (eoan-proposed/main) [2.26.0-1ubuntu1] (desktop-core) | 10:08 | |
| -queuebot:#ubuntu-release- Unapproved: accepted s390-tools [s390x] (eoan-proposed) [2.11.0-0ubuntu1] | 10:39 | |
| -queuebot:#ubuntu-release- New source: salt-pylint (eoan-proposed/primary) [2019.6.7-1ubuntu1] | 10:54 | |
| LocutusOfBorg | ^^ this is a leaf new package, please accept it, I would like to finish the pylint sadness in the archive | 10:57 |
| LocutusOfBorg | (it might be not needed, but the Debian tracker shows it) | 10:58 |
| -queuebot:#ubuntu-release- New binary: linux-signed [amd64] (bionic-proposed/main) [4.15.0-63.72] (core, kernel) | 12:27 | |
| -queuebot:#ubuntu-release- New binary: linux-signed [ppc64el] (bionic-proposed/main) [4.15.0-63.72] (core, kernel) | 12:27 | |
| -queuebot:#ubuntu-release- New: accepted linux-signed [amd64] (bionic-proposed) [4.15.0-63.72] | 12:39 | |
| -queuebot:#ubuntu-release- New: accepted linux-signed [ppc64el] (bionic-proposed) [4.15.0-63.72] | 12:40 | |
| === jdstrand_ is now known as jdstrand | ||
| -queuebot:#ubuntu-release- New binary: linux-signed-oem-osp1 [amd64] (bionic-proposed/universe) [5.0.0-1021.23] (no packageset) | 13:31 | |
| -queuebot:#ubuntu-release- New: accepted linux-signed-oem-osp1 [amd64] (bionic-proposed) [5.0.0-1021.23] | 13:33 | |
| vorlon | Laney: is there an index of reverse test-dep that I can refer to? | 15:29 |
| Laney | vorlon: not that I know of | 15:30 |
| Laney | well, unless you count Sources.xz Testsuite-Triggers | 15:31 |
| vorlon | Laney: do you think we should extend reverse-depends to cover those? | 15:39 |
| Laney | vorlon: sounds reasonable, maybe reverse-depends -t or similar? | 15:40 |
| Laney | (britney is also triggering for direct reverse binary dependencies too, but you're already checking that as part of the usual process) | 15:41 |
| vorlon | tumbleweed: ^^ I forget, is there a way to report bugs on the reverse-depends service? | 15:41 |
| tumbleweed | not really. Should move it to a launchpad project | 15:42 |
| tumbleweed | there: https://bugs.launchpad.net/reverse-depends | 15:44 |
| -queuebot:#ubuntu-release- New: accepted webkit2gtk [amd64] (eoan-proposed) [2.26.0-1ubuntu1] | 16:04 | |
| vorlon | tumbleweed: cool, LP: #1843614 filed | 16:24 |
| ubot5 | Launchpad bug 1843614 in reverse-depends "reverse-depends service should report test-reverse-dependencies as well" [Undecided,New] https://launchpad.net/bugs/1843614 | 16:24 |
| vorlon | tumbleweed: interestingly, I see the latest commit on https://code.launchpad.net/~stefanor/reverse-depends/reverse-deps says "cover pre-depends", but my recent experience was that reverse-depends multiarch-support did not report the list of packages that needed fixing | 16:25 |
| vorlon | coreycb: currently looks like python-microversion-parse, python-pbr, stevedore are critical-path for removing another chunk of the graph | 16:29 |
| -queuebot:#ubuntu-release- New binary: virtualenvwrapper [amd64] (eoan-proposed/universe) [4.8.4-4] (no packageset) | 16:52 | |
| coreycb | vorlon: thanks i'll take a look | 16:59 |
| RikMills | Hi, could qtwebengine-opensource-src package 5.9.8 be removed from bionic-proposed please? | 17:21 |
| RikMills | the last comment on LP: #1830807 requests this | 17:21 |
| ubot5 | Launchpad bug 1830807 in qtwebengine-opensource-src (Ubuntu Bionic) "Update to bug-fix release Qt 5.9.8 to fix security issues in qtwebengine in Bionic" [Undecided,Fix committed] https://launchpad.net/bugs/1830807 | 17:21 |
| RikMills | its presence in proposed will cause some of the things I am getting ready for SRU to FTBFS | 17:22 |
| vorlon | RikMills: doing | 17:26 |
| vorlon | RikMills: done | 17:26 |
| RikMills | thanks! | 17:26 |
| coreycb | vorlon: python-pbr reverse depends opens a bag of worms | 17:29 |
| coreycb | vorlon: do we need to remove python-pbr in eoan? | 17:34 |
| vorlon | coreycb: if we don't remove it, it'll be ftbfs due to missing build-depends | 17:57 |
| vorlon | that's probably a reasonable tradeoff | 17:58 |
| coreycb | vorlon: ok i didn't really look to see if its (Build-)Depends have already been removed. so yeah i guess we need to do that work this cycle. | 18:00 |
| vorlon | coreycb: having one or two packages ftbfs due to missing build-deps is better than leaving NBS packages around at release time, so I am willing to draw a line there and we can burn down python-pbr revdeps if time allows | 18:05 |
| coreycb | vorlon: that sounds like a good plan then | 18:06 |
| ahasenack | hi, just letting you know I'm looking at the red ruby-ferret on arm64 now, the only red that is blocking ruby2.5 from migrating | 18:10 |
| ahasenack | "you" == "the collective you" | 18:10 |
| bryce | ahasenack, "all y'all" ;-) | 18:30 |
| ahasenack | works :) | 18:31 |
| -queuebot:#ubuntu-release- Unapproved: rejected asterisk [source] (bionic-proposed) [1:13.18.3~dfsg-1ubuntu4build1] | 18:41 | |
| -queuebot:#ubuntu-release- Unapproved: rejected mailsync [source] (bionic-proposed) [5.2.2-3.1build1.18.04.1] | 18:42 | |
| -queuebot:#ubuntu-release- Unapproved: rejected prayer [source] (bionic-proposed) [1.3.5-dfsg1-4build2] | 18:42 | |
| -queuebot:#ubuntu-release- Unapproved: rejected asterisk [source] (disco-proposed) [1:16.2.1~dfsg-1build1] | 18:43 | |
| -queuebot:#ubuntu-release- Unapproved: rejected mailsync [source] (disco-proposed) [5.2.2-3.1build1.19.04.1] | 18:44 | |
| -queuebot:#ubuntu-release- Unapproved: rejected prayer [source] (disco-proposed) [1.3.5-dfsg1-6build0.19.04.1] | 18:45 | |
| vorlon | coreycb: I uploaded python-diskimage-builder (revdep of stevedore), and this reveals that the openstack package build scripts are only half magic; the package has manual update-alternatives maintainer script code, but because we're no longer building the python module, the files in the packages are installed to different locations | 20:02 |
| vorlon | coreycb: Debian dealt with this by just dropping the maintainer scripts when dropping python2. I'm following suit, but this makes the package not sanely upgradealbe. | 20:03 |
| coreycb | vorlon: thanks. it seems like it should be ok. pkgos-dh_auto_install won't create the /usr/bin/python{2|3}- prefixed files unless it is installing for both py2 and py3. | 20:13 |
| coreycb | so in eoan it'll install /usr/bin/dib-block-device but in disco it would install /usr/bin/python{2|3}-dib-block-device and point the alternative /usr/bin/dib-block-device at one of those | 20:15 |
| vorlon | coreycb: the reason it won't upgrade cleanly is that after upgrade the alternatives will still be in the database, and best case those'll be floating around in the dpkg database forever, worst case their presence will prevent the package from being upgraded at all | 20:28 |
| coreycb | vorlon: will the prerm script from the old package get run before the package is upgraded? | 20:32 |
| vorlon | it will, but I believe I saw that the old prerm script only removed the alternatives if the package was being removed, not upgraded | 20:33 |
| coreycb | vorlon: alright yeah just noticed that | 20:36 |
| === michagogo_ is now known as michagogo | ||
| -queuebot:#ubuntu-release- New: accepted virtualenvwrapper [amd64] (eoan-proposed) [4.8.4-4] | 23:04 | |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!