[03:11] hmm, so the latest LXD upload effectively switches from 3 binary packages per-arch to a single arch-all package, is there any way to make proposed-migration happy enough with that so it starts running autopkgtests? [03:11] looking at docs, looks like the solution to unstick this is to remove those binaries from the release pocket, but that'd break people at this point... [03:12] (I'm fine doing the package removal from the release pocket once I know that the new package will migrate within the hour, but right now I specifically don't want it to migrate yet I want adt to do its thing) [05:07] -queuebot:#ubuntu-release- Unapproved: qemu (xenial-proposed/main) [1:2.5+dfsg-5ubuntu10.31 => 1:2.5+dfsg-5ubuntu10.32] (ubuntu-server, virt) [07:34] stgraber, do those break the originals ? [07:34] etc [07:40] stgraber, could you not make both of those going away binaries transitionals, and would this both allow this to migrate _and_ effectivly deinstall the now redundant packages [07:45] then they would be removable in -release in a subsequent upload [07:45] as nothing is in them [08:58] -queuebot:#ubuntu-release- New binary: linux-signed [amd64] (trusty-proposed/main) [3.13.0-159.209] (core, kernel) [09:00] -queuebot:#ubuntu-release- New binary: linux-signed-lts-xenial [amd64] (trusty-proposed/main) [4.4.0-136.162~14.04.1] (kernel) [09:11] -queuebot:#ubuntu-release- New: accepted linux-signed-lts-xenial [amd64] (trusty-proposed) [4.4.0-136.162~14.04.1] [09:11] -queuebot:#ubuntu-release- New: accepted linux-signed [amd64] (trusty-proposed) [3.13.0-159.209] [09:24] sil2100: you are on SRU duty right now correct? [09:24] I wanted to ask if you could cancel 1:2.11+dfsg-1ubuntu7.6 from bionic unapproved [09:25] it is good, but I have two more changes complete and want to minimize how often people have to update [09:25] so I'd want to bundle those two more changes and re-upload [09:26] but be careful, there also is an upload in xenial-unapproved, do not cancel that one (it is unrelated) [09:32] cpaelzer: ok! [09:33] thank you sil2100 [09:33] Done ;) [09:34] -queuebot:#ubuntu-release- Unapproved: rejected qemu [source] (bionic-proposed) [1:2.11+dfsg-1ubuntu7.6] [09:37] LocutusOfBorg: your automake 1.16 update broke dozens of builds ... [09:45] doko, can I see them? [09:49] e.g. newt. currently running the test rebuild [09:53] juliank, ^^ please merge it? [09:53] um [09:53] better cancel the automake update? [09:54] after it has been removed from the archive and the new package migrated? [09:55] Well, I don't know what you did [09:55] migrated 20 days ago... [09:55] juliank, followed the change from automake 1.15 to 1.16 [09:55] and some bad programs doing bad things, hardcoding stuff, breaks [09:55] sil2100, bdmurray: I didn't manage to get back to gnome-software in Xenial unapproved yesterday. I'll finish it off now. [09:56] juliank, I can steal newt if you want [09:56] the fix is two lines [09:57] I'd like to figure out why you merge automake 1.16 after the feature freeze [09:57] I asked here, and I got "ok to go" [09:57] because there was still time to fixup stuff [09:58] Well, ok [09:58] and there aren't new "features" in new automake, just bugfixes [09:58] stuff is not breaking because of actual changes, but because "export AUTOMAKE 1.15" and similar [09:58] Did you? I see you asking, but no explicit yes [09:59] So packages have missing Depends: automake (<< 1.16) [09:59] Well, it would have required somebody to process NEW, but that's a different headspace [09:59] (that's what they should have so you can actually notice them) [10:00] Can someone run a search for 1.15 in the archive? [10:00] so we know what breaks? [10:00] juliank, do you have any idea about how to do it? [10:01] cjwatson, actually automake has been uploaded in early august in debian, for some reasons really obscure to me, it didn't get autosyncd [10:01] and it was 24 hours after the freeze, yes [10:01] I think security team has a script for greeping source packages [10:01] oh... lets move there maybe [10:01] Doesn't seem obscure, you even pasted the reason [10:01] automake-1.16_1:1.16.1-1 is trying to override modified binary automake_1:1.15.1-3ubuntu2. OK (y/N)? n [10:02] Sources that would produce binary packages that currently have Ubuntu deltas don't get autosynced for what I would hope would be entirely clear reasons [10:02] ok so I don't remember anymore after 20 days, and my irc history doesn't go that far [10:02] cjwatson, probably I remember "why the hell the delta is a stupid 2 line autopkgtestsuite and nobody ever pushed it to debian in years" [10:03] and then I filed https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=906919 [10:03] Debian bug 906919 in automake-1.16 "automake-1.16: please enable testsuite [PATCH]" [Important,Open] [10:03] rbasak: ok! [10:03] rbasak: thanks for the heads up [10:04] LocutusOfBorg: FWIW; I don't see anybody responding to you, but someone did let it pass NEW [10:04] btw, the "dozen" might be actual 3 packages [10:04] https://codesearch.debian.net/search?q=1%5C.15+path%3Adebian%2Frules [10:05] Sounds like doko has more examples? [10:05] I hope so [10:05] and the package above results in the search [10:06] libjpeg9: export AUTOMAKE = automake-1.15 [10:06] ugh [10:06] Should add a lintian check "hardcoded autotools versions" [10:06] ekg2 [10:06] icecast2 are the others [10:07] newt "libjpeg9?" (not sure) ekg2 and icecast2 [10:07] yes, exactly [10:08] When bumping to 1.16, it would make sense to add proper Build-Depends: automake (<< 1.17) [10:08] so things get noticed more easily [10:08] Or just unhardcode [10:08] do you care to file a bug? [10:08] That would be nicer [10:09] right now I'm fixing ekg2 to stop hardcoding [10:09] How? [10:09] Are you using dpkg to query version of automake and then sed it correclty? [10:11] i.e. AM_VERSION=$(shell dpkg-query -f '\${source:Upstream-Version}' -W automake | egrep -o '^[0-9]+\.[0-9]+') [10:11] or just hardcode AUTOMAKE = automake [10:11] well, remove the vars [10:12] -queuebot:#ubuntu-release- Unapproved: rejected freshplayerplugin [source] (xenial-proposed) [0.3.9-0ubuntu0~16.04.1] [10:12] -queuebot:#ubuntu-release- Unapproved: freshplayerplugin (xenial-proposed/multiverse) [0.3.4-3ubuntu0.1 => 0.3.9-0ubuntu0.16.04.1] (no packageset) [10:15] -queuebot:#ubuntu-release- Unapproved: accepted freshplayerplugin [source] (xenial-proposed) [0.3.9-0ubuntu0.16.04.1] [10:19] -queuebot:#ubuntu-release- Unapproved: metacity (bionic-proposed/universe) [1:3.28.0-1 => 1:3.28.0-1ubuntu0.1] (ubuntu-desktop) [10:19] juliank, the latter [10:21] ekg2 is fixed for real now [10:21] I send the $(shell) invocation to Debian for newt [10:21] I did not check quoting, though [10:22] thanks! [10:22] icecast2 is good now [10:22] sil2100, bdmurray: gnome-software looks good to me, except that there's one change with no bug reference that enables a patch that was supposed to be added in a previous SRU but never functionally landed (missing from series). The changelog says "Add patch that seemed to get missed in 3.20.5-0ubuntu0.16.04.10". That previous upload listed two bugs that were marked verification-done, but AFAICT that [10:22] was false since the patch was never enabled. So should I ask for those two bugs to be reopened and the changelog to include them so they can be reverified? Opinions? [10:23] I don't see robert_ancell online - out of time zone I guess. [10:23] It's actually [10:23] AM_VERS:=$(shell dpkg-query -f '$${source:Upstream-Version}' -W automake | egrep -o '^[0-9]+\.[0-9]+') [10:24] what about dropping that variable? [10:25] cp /usr/share/automake-$(AM_VERS)/install-sh ./install-sh [10:25] seriosly because of this^ [10:25] ? [10:26] LocutusOfBorg: Well, it's clearly needed here [10:27] cp /usr/share/automake-*/install-sh ./install-sh [10:27] Replacing $(AM_VERS) with * would work sometime [10:27] but not if you have automake-1.11 installed I guess [10:27] yes I know it doesn't with automake 1.11 and 1.15 [10:27] I'm trying to find if --add-missing or whatever works [10:27] you know, there is some obscure automake stuff [10:27] for this [10:27] bug 908739 [10:27] bug 908739 in Odoo Addons (MOVED TO GITHUB) "web : compagin ugly screen" [Low,Fix released] https://launchpad.net/bugs/908739 [10:27] in debian [10:27] https://bugs.debian.org/908739 [10:28] Debian bug 908739 in src:newt "newt: Do not hardcode automake version" [Normal,Open] [10:28] there we go [10:28] nice! [10:28] I hope to find a better solution btw :) [10:28] I don't know why I had to use $(strip) around the dpkg-query | egrep [10:28] or why I did not use cut -1-2 -d. [10:29] oh well, this probably works better in case of an alpha or stuff [10:29] juliank, please [10:29] AUTOMAKE_FLAGS="--add-missing --copy" [10:29] end of the story [10:29] LocutusOfBorg: Tell the bug report [10:29] mmm didn't work :/ [10:29] :) [10:30] * LocutusOfBorg tries with a bigger hammer [10:31] LocutusOfBorg: Where does _FLAGS come from? [10:31] AUTOMAKE=automake --add-missing --copy [10:31] is what I'd do [10:31] that's what autoreconf(1) documents [10:32] LocutusOfBorg: Ah, duh, there's no automake in the build [10:33] only a Makefile.in, no Makefile.am [10:36] LocutusOfBorg: That said, it seems wrong to have the file in the first place [10:36] I mean, Makefile.in defines an install-sh target for install shared library [10:36] cpaelzer: online? pmdk [10:37] and the Makefile does not use install-sh [10:38] I mean, it uses AC_PROG_INSTALL in configure [10:38] but then it has to use $(INSTALL), not install [10:39] I feel so dirty [10:39] but doing this works [10:39] AUTOMAKE=/usr/bin/automake --add-missing --copy --foreign [10:39] $(AUTOMAKE) || true [10:40] in configure, because automake fails, but after having copied the file :) [10:40] FWIW, python-newt should be called python-snack, it provides a snack module, no newt module [10:40] LocutusOfBorg: The cp is cleaner than that, though [10:40] sure juliank :) I want to hide in a cave [10:41] LocutusOfBorg: Patching out the check for AC_PROG_INSTALL is even easier [10:41] Since Makefile.in does not use $(INSTALL)... [10:42] I think it's a bug in autoconf [10:42] autoconf should copy install-sh as needed [10:42] It can't rely on libtool or automake to do it [10:43] I confirm, removing PROG_INSTALL works... mind to update the bug report? :) [10:43] and do the upload if you can, so I can focus on libjpeg9 [10:43] (the last one for me) [10:45] LocutusOfBorg: I already did the newt merge so we're good for now, I'm not introducing a delta for this [10:45] s/introducing/increasing/ [10:46] ack but maybe adding a patch to the bug report? [10:47] Well, he can remove a line himself [10:48] I also reported a bug to merge the built -> build change in python{3,}-newt [10:49] :) [10:49] nice, indeed! [10:49] LocutusOfBorg: Probably file an RC bug in Debian then with the libjpeg9 fix [10:50] libjpeg9 is fine without anything of that hardcoding https://launchpad.net/~costamagnagianfranco/+archive/ubuntu/locutusofborg-ppa/+build/15453009 [10:50] sure, an RC is what I plan to do [10:55] doko: I was out for lunch [10:55] what is up with pmdk? [10:55] would be ahasenack btw, but he is out until the sprint [10:56] doko: saw your questions on the bug [10:57] all the questions are fine to be answered at the sprint next week I think [10:57] thanks for taking a look [10:58] jemalloc being bundled in there is interesting, didn't realize that before [10:58] I'm not trying to half-answer the questions, I'm sure it will be better if I do so together with ahasenack [11:27] juliank, debian bug: #908741 [11:27] Debian bug 908741 in src:libjpeg9 "libjpeg9: do not hardcode automake version" [Serious,Open] http://bugs.debian.org/908741 [11:39] http://people.canonical.com/~doko/ftbfs-report/test-rebuild-20180911-cosmic.html [11:39] LocutusOfBorg: ^^^ [11:39] why is pkg-stripfiles so slow since yesterday? https://launchpad.net/ubuntu/+source/ekg2/1:0.4~pre+20120506.1-14ubuntu1/+build/15452901 [11:39] hours to finish a job of seconds... [11:39] doko, ack [11:41] of course, not yet finished [11:48] I only found one possible, I looked at all main logs [11:49] (only logs where failure was in every architectures, since autoreconf is a problem everywhere) [12:00] juliank, dpkg fails for a similar reason, do you think it is sane to change aclocal.m4 into ./aclocal.m4:[am__api_version='1.16' [12:00] ./aclocal.m4:m4_if([$1], [1.16], [], [12:00] ? [12:00] aclocal should be regenerated [12:00] Is dpkg not running dh-autoreconf or overriding aclocal? [12:01] debian/rules: autoreconf -v -i [12:01] no aclocal overrides [12:02] nice, works in debian... [12:02] odd stuff [12:30] -queuebot:#ubuntu-release- Unapproved: gnome-flashback (bionic-proposed/universe) [3.28.0-1ubuntu1.1 => 3.28.0-1ubuntu1.2] (edubuntu) [12:31] jbicha, seb128: why isn't xdg-desktop-portal seeded anywhere? [12:32] doko: my guess is that we're waiting for bug 1750069 to be approved [12:32] bug 1750069 in xdg-desktop-portal-gtk (Ubuntu) "[MIR] xdg-desktop-portal-gtk" [Undecided,Confirmed] https://launchpad.net/bugs/1750069 [12:33] ahh, ok [12:37] jamespage: any update on https://bugs.launchpad.net/aodh/+bug/1737989 ? [12:37] Ubuntu bug 1737989 in ujson (Ubuntu) "[MIR] ujson?" [Undecided,Incomplete] [13:44] apw: so yeah, I can introduce two empty packages just to make britney happy, seems like a bit of a hack though as those wouldn't be transitional packages, they'd literally be empty packages [13:45] apw: they wouldn't be transitional because the "lxd" package does not actually provide the content of lxd-tools [13:47] stgraber, except in essence it does, because the installation of lxd installs the snap which brings you the contents, in some sense [13:51] only a temporary requirement anyway ^^ ? [13:54] -queuebot:#ubuntu-release- Unapproved: accepted nautilus [source] (bionic-proposed) [1:3.26.4-0~ubuntu18.04.1] [13:58] apw: well, that's true for the lxd-client package and lxd package, that isn't true for the lxd-tools package which has no equivalent post-deb [13:59] anyway, that's easy enough to do so I'll just add two transitional packages for now and then make a subsequent upload on Monday to remove them both from the archive [14:01] hi, do i need FFe for LP: #1317164 and LP: #1791931 ? [14:01] Launchpad bug 1317164 in update-manager (Ubuntu) "Show reason of TransactionFailed and let the user try again instead of crashing" [High,In progress] https://launchpad.net/bugs/1317164 [14:02] Launchpad bug 1791931 in update-manager (Ubuntu) "Update-manager crashes in _show_transaction due to packages being already removed" [High,In progress] https://launchpad.net/bugs/1791931 [14:04] the first one is a bit of a user interface improvement but prior crashes were not considered to be u-m issues per the bug's old text [14:15] i guess that's a no for requiring FFe :-) [14:39] -queuebot:#ubuntu-release- Unapproved: qemu (bionic-proposed/main) [1:2.11+dfsg-1ubuntu7.5 => 1:2.11+dfsg-1ubuntu7.6] (ubuntu-server, virt) [14:52] -queuebot:#ubuntu-release- New binary: linux-signed [amd64] (bionic-proposed/main) [4.15.0-35.38] (core, kernel) [15:15] sil2100: hi. is there any chance you could do the cosmic packageset update for Kubuntu? [15:17] acheronuk: hey! Yes, the scripts are running [15:17] Will take a moment I suppose [15:17] sil2100: great. thanks. sorry to ping more than once [15:50] sil2100: got your email. can I see the diff anywhere to check? [15:53] if not I've saved my current one and will diff when it updates [15:57] -queuebot:#ubuntu-release- Unapproved: accepted python3.6 [source] (bionic-proposed) [3.6.6-1~18.04] [16:00] acheronuk: http://paste.ubuntu.com/p/pkGrpwbDZk/ [16:01] -queuebot:#ubuntu-release- Packageset: 448 entries have been added or removed [16:03] -queuebot:#ubuntu-release- Unapproved: accepted python3.7 [source] (bionic-proposed) [3.7.0-1~18.04] [16:04] slangasek: thanks. what I needed is in, and nothing vital dropped :) [16:12] acheronuk: was that directed at sil2100 ? [16:13] slangasek: yup. apologies [16:37] slangasek, ogra: https://bugs.launchpad.net/ubuntu/+source/ubuntu-core-meta/+bug/1572539 where should that be seeded? [16:37] Ubuntu bug 1572539 in ubuntu-core-meta (Ubuntu) "[MIR] ubuntu-core-libs" [Critical,Fix committed] [16:38] doko, probably just in supported ... though i'm not sure the archive package has even the recent packagelist [16:38] not sure who worked on that [16:41] (the bug is 2 years old and i have moved teams twice since i filed it ... with each move moving me further away from any archive related work sadly) [16:46] -queuebot:#ubuntu-release- Unapproved: accepted binutils [source] (bionic-proposed) [2.30-21ubuntu1~18.04] [17:04] -queuebot:#ubuntu-release- New binary: linux-signed-hwe [ppc64el] (xenial-proposed/main) [4.15.0-35.38~16.04.1] (kernel) [17:05] stgraber: the snapd we rather badly need for fixing regressions in cosmic is blocking on lxd autopkgtest failures; "accept4: too many open files; retrying in 1s" --> timeout. Do you know what this is? [17:08] -queuebot:#ubuntu-release- New binary: linux-signed-hwe [amd64] (xenial-proposed/main) [4.15.0-35.38~16.04.1] (kernel) [18:09] infinity, slangasek: Could wxl please get livefs FTBFS nag emails as well? [18:09] Also, gilir left the project, I don't know if he still wants to be nagged. :P [18:09] slangasek: hi, sorry, was in a meeting, looking now [18:09] tsimonq2: email address? [18:09] slangasek: wxl@ubuntu.com [18:10] tsimonq2: added. and if gilir doesn't want to be nagged, he knows where to find us [18:10] slangasek: ack :) [18:10] Thanjs. [18:10] *Thanks [18:11] slangasek: all 3 failures appaear to be some kind of race in clustering code, amd64 was retried and is now green, just retried the other two [18:12] slangasek: so snapd isn't breaking LXD (not sure why it even triggered the lxd test in the first place), feel free to hint if urgent [18:12] o [18:12] ok [18:14] -queuebot:#ubuntu-release- New: accepted linux-signed-hwe [amd64] (xenial-proposed) [4.15.0-35.38~16.04.1] [18:14] -queuebot:#ubuntu-release- New: accepted linux-signed-hwe [ppc64el] (xenial-proposed) [4.15.0-35.38~16.04.1] [18:23] -queuebot:#ubuntu-release- New: accepted protobuf [amd64] (xenial-proposed) [2.6.1-1.3ubuntu1] [18:23] -queuebot:#ubuntu-release- New: accepted protobuf [ppc64el] (xenial-proposed) [2.6.1-1.3ubuntu1] [18:23] -queuebot:#ubuntu-release- New: accepted protobuf [arm64] (xenial-proposed) [2.6.1-1.3ubuntu1] [18:23] -queuebot:#ubuntu-release- New: accepted protobuf [s390x] (xenial-proposed) [2.6.1-1.3ubuntu1] [18:23] slangasek: is that something another user could do on their own? (sign up for livefs failure notices) [18:23] it is not [18:24] the configuration currently lives in the private 'production' branch for cdimage [18:24] this bit probably doesn't need to be private, but otoh you also probably don't want notification emails from any other random people who try to run cdimage [18:25] right. sounds good [18:25] in other news it seems that the lubuntu bionic alternate images haven't built since april.. uhhh? [18:29] and the amd64 died due to unmet depends with linux. i'm assuming this will resolve itself once 4.15.0-35.38 migrates? https://launchpadlibrarian.net/388239221/buildlog_ubuntu_bionic_amd64_lubuntu_BUILDING.txt.gz [18:29] alternate> are they supposed to be building? there are no emails about build failures, so that appears to be intentional [18:29] (amd64 desktop) [18:29] it resolves itself when linux+linux-signed+linux-meta are all built and published in -proposed [18:30] slangasek: i would expect the alternates to continue to build. we're certainly on the hook to continue to support them for bionic :) [18:30] yet no one mentioned them being absent for the point release? [18:31] weird. they were in the original release. let me look into that further [18:32] slangasek: oh, just thought of something for the magic lxd deb-to-snap package, I need a Pre-Depends on snapd don't I? For the unlikely case where snapd isn't already installed. [18:32] the history is in lp:ubuntu-cdimage; infinity dropped lubuntu alternate builds on 2 May, and enabled bionic daily builds in a later commit the same day [18:33] we have a Depends right now, but that won't be good enough since we do things from preinst [18:33] stgraber: you need a pre-depends only if you need to talk to snapd from the deb's preinst package; do you? [18:33] ok then yes [18:33] like i said, i'll investigate further. no worries on that front, slangasek. [18:33] thanks for your help :) [18:51] wxl: Look back, we don't do the alternates for point releases, because representing the changes in the seeded sets for the HWE stack is somewhere between difficult and impossible. [18:51] wxl: This isn't new. :P [18:52] infinity: tsimonq2 reminded me. i disabled them. he should have done that, but he's a bad boy, as i'm sure you already know [18:53] But I thought infinity had access to do that. :P [18:54] wxl: So don't blame me, blame Adam. :P [18:54] oh no, you had access, too, mr. release manager :) [18:55] You're supposed to regularly looking at the tracker, Mr. QA Lead. ;) [18:55] i'm too busy sorting through all the bugs you make :) [18:55] * wxl *slap fight* [18:55] * tsimonq2 *slap fight* [18:56] Love you too, wxl. [18:57] https://v.gd/zMrzy3 [18:58] You're the short one. :P [18:58] whatever. that hat looks ridiculous. [18:58] hahahahahaha [18:59] wxl: I would propose we have #ubuntu-release-offtopic but I don't think anyone would join. :P [18:59] XD [18:59] it would give adam something else to roll his eyes at [19:00] ¯\_(ツ)_/¯ [19:02] *rolls eyes* [19:07] * tsimonq2 throws teward into /dev/null [19:17] Laney: do you know why amd64,arm64,i386 are taking so long to clear the autopkgtest backlog? [19:18] they're now a full day behind the other archs, which is exceptional [19:38] slangasek: Thoughts on vnstat's arm* tests? Seems like somethingbadtestable. [19:38] *something badtestable [19:45] -queuebot:#ubuntu-release- Unapproved: fonts-liberation2 (bionic-proposed/main) [2.00.1-5 => 2.00.1-7~18.04.1] (kubuntu, personal-gunnarhj, ubuntu-desktop) [19:45] -queuebot:#ubuntu-release- Unapproved: fonts-liberation (bionic-proposed/main) [1:1.07.4-5 => 1:1.07.4-7~18.04.1] (desktop-core, personal-gunnarhj) [20:02] -queuebot:#ubuntu-release- Unapproved: accepted fonts-liberation [source] (bionic-proposed) [1:1.07.4-7~18.04.1] [20:05] -queuebot:#ubuntu-release- Unapproved: accepted fonts-liberation2 [source] (bionic-proposed) [2.00.1-7~18.04.1] [20:05] slangasek, sil2100: hey, tomorrow could one of you please review the nodejs openssl change for cosmic? (the issue of abi breakage w/ upstream nodejs) there's a debdiff in LP #1779863, which fixes that plus the test-case failure seen in all archs in the version currently in cosmic-proposed. thanks! [20:05] Launchpad bug 1779863 in nodejs (Ubuntu Cosmic) "Ubuntu nodejs package isn't ABI compatible with mainline nodejs." [Medium,In progress] https://launchpad.net/bugs/1779863 [20:19] tsimonq2: it didn't look badtestable to me, it looked like a version in -proposed that had regressed the tests on these archs. What do you see that suggests otherwise? [20:21] mfo: done [20:21] slangasek, thanks! :) that was fast. [20:51] mfo: ftbfs on s390x [20:51] possibly going to ftbfs elsewhere and s390x is just first [20:51] slangasek, hm. i'll go check it. [20:52] slangasek: vnstat's arm* tests> I say that because the tests have regressed in the same way as all other arches. [21:01] tsimonq2: from what I see, we previously had a test, that made buggy assumptions about the host interface names so only passed on some architectures; now we have a different test, which makes different buggy assumptions, and is broken on all archs. [21:01] "test is now equally crappy on all architectures" is not in this case a reason to badtest, I think [21:05] slangasek, just checked the nodejs ftbfs on s390x. it is s390x specific and not a regression. looking at the build logs of the *previous* version (...ubuntu1), that test-case failure already happened in s390x and only in s390x. the other test-case failure in all archs has been fixed in this upload (ubuntu2). i'm happy to debug it on s390x and see if i can fix that too.. do you know where/whom can i get access to s390x within Canonical? [21:07] slangasek: ack, makes sense. [21:07] mfo: when you say 'not a regression', nodejs 8.11.2~dfsg-1 certainly built on s390x. as for access, let me see what I can do; we unfortunately don't have a porter box set up the way we do for other archs [21:07] slangasek, yes, that is the previous-previous version. i mean it's not a regression caused by this upload. :) [21:09] slangasek, it turns out that 8.11.2~dfsg-1 was built back on May iirc, then 8.11.2~dfsg-1ubuntu1 is recent (and had these test-case failures). then 8.11.2~dfsg-1ubuntu2 (this upload) should resolve the test-case failure that is common to all archs (e.g., ppc64el built successfully), but this s390x specific test-case failure I hadn't seen yet, but it is present in the build log of 8.11.2~dfsg-1ubuntu1 (previous version). that's what I mean :) [21:10] mfo: fair enough; still keeps the package from reaching the release pocket anyway [21:10] slangasek, right, i undertand. that's why i'm offering to debug it. :) and thanks for checking on s390x access, i understand it might be hard to get, unfrotunately. [21:12] slangasek, i have to leave for the day, but can continue working on this tomorrow. thanks again for your help. [21:36] -queuebot:#ubuntu-release- Unapproved: accepted plymouth [source] (bionic-proposed) [0.9.3-1ubuntu7.18.04.1]