| -queuebot:#ubuntu-release- New binary: mythtv [s390x] (artful-proposed/multiverse) [2:29.0+fixes.20170728.696806310a-0ubuntu1] (mythbuntu) | 00:01 | |
| -queuebot:#ubuntu-release- New binary: mythtv [i386] (artful-proposed/multiverse) [2:29.0+fixes.20170728.696806310a-0ubuntu1] (mythbuntu) | 00:09 | |
| -queuebot:#ubuntu-release- New binary: mythtv [amd64] (artful-proposed/multiverse) [2:29.0+fixes.20170728.696806310a-0ubuntu1] (mythbuntu) | 00:22 | |
| -queuebot:#ubuntu-release- New binary: mythtv [arm64] (artful-proposed/multiverse) [2:29.0+fixes.20170728.696806310a-0ubuntu1] (mythbuntu) | 02:12 | |
| -queuebot:#ubuntu-release- New binary: mythtv [armhf] (artful-proposed/multiverse) [2:29.0+fixes.20170728.696806310a-0ubuntu1] (mythbuntu) | 02:23 | |
| flocculant | infinity: hi - not often I shout out, but we've now been trying to sort out our 'simple' Xubuntu since 16.04.1 https://code.launchpad.net/~xubuntu-dev/livecd-rootfs/xubuntu-base https://code.launchpad.net/~xubuntu-dev/debian-cd/xubuntu-base https://code.launchpad.net/~xubuntu-dev/ubuntu-cdimage/xubuntu-base | 02:45 |
|---|---|---|
| flocculant | not sure what we really need to be doing but, we would really like to see this sorted in one way or the other for the next lts | 02:46 |
| tsimonq2 | flocculant: Why don't you add something to the official infra (like with Kubuntu's (deprecated) other images and Lubuntu Next, for example)? Just curious. | 02:47 |
| flocculant | tsimonq2: iirc we tried all that 2 years ago - really just want offical answers here now thanks | 02:49 |
| tsimonq2 | flocculant: Sure, I just know Adam's a busy guy, wanted to see if I could give you a hand, but if you'd like to wait for him, that's your choice :) | 02:50 |
| tsimonq2 | Had a discussion in another channel, forget I said anything here. | 03:20 |
| flocculant | but you had to comment ... | 03:27 |
| flocculant | sigh | 03:27 |
| LocutusOfBorg | can you please remove libldap-2.4-2-dbg, slapd-dbg? | 07:04 |
| infinity | LocutusOfBorg: Wrong question. | 08:21 |
| infinity | LocutusOfBorg: Or rather, you wanted to add "in artful-proposed". | 08:21 |
| infinity | (done) | 08:22 |
| LocutusOfBorg | thanks, but the dbg package exists in release, not in proposed... sorry if I wasn't clear, but I don't get it :) | 09:18 |
| LocutusOfBorg | btw do you have any clue for my notmuch debug logs? with strace and valgrind | 09:18 |
| infinity | LocutusOfBorg: You were asking me to remove them because britney was listing them as "old binaries left..." | 09:47 |
| infinity | LocutusOfBorg: "old binaries left on amd64: libldap-2.4-2-dbg, slapd-dbg (from 2.4.44+dfsg-8ubuntu2)" | 09:47 |
| infinity | LocutusOfBorg: Note the version there. That's not the version in the release pocket, it was a previous upload that never made it out of proposed. | 09:47 |
| infinity | LocutusOfBorg: (Yes, they exist in the release pocket too, but they SHOULD, until migration happens) | 09:48 |
| LocutusOfBorg | ok thanks | 10:03 |
| -queuebot:#ubuntu-release- New binary: farmhash [amd64] (artful-proposed/universe) [0~20170626-g23eecfb-1] (no packageset) | 11:09 | |
| -queuebot:#ubuntu-release- New binary: farmhash [s390x] (artful-proposed/universe) [0~20170626-g23eecfb-1] (no packageset) | 11:09 | |
| -queuebot:#ubuntu-release- New binary: farmhash [ppc64el] (artful-proposed/universe) [0~20170626-g23eecfb-1] (no packageset) | 11:09 | |
| -queuebot:#ubuntu-release- New binary: farmhash [arm64] (artful-proposed/universe) [0~20170626-g23eecfb-1] (no packageset) | 11:11 | |
| -queuebot:#ubuntu-release- New binary: farmhash [i386] (artful-proposed/universe) [0~20170626-g23eecfb-1] (no packageset) | 11:11 | |
| -queuebot:#ubuntu-release- New binary: farmhash [armhf] (artful-proposed/universe) [0~20170626-g23eecfb-1] (no packageset) | 11:11 | |
| LocutusOfBorg | infinity, sorry for being pesty, but an hint for notmuch will be so much appreciated, I'm really worried about miscompilation of our toolchain on armhf | 11:25 |
| infinity | LocutusOfBorg: Sorry, I haven't really been paying attention. What have you tried, and what have you managed to discover? Logs, notes? | 11:26 |
| LocutusOfBorg | I did a run of the test under strace gdb and another with valgrind gdb | 11:27 |
| LocutusOfBorg | I see SIGKILL being started | 11:27 |
| LocutusOfBorg | this one with valgrind https://launchpad.net/~costamagnagianfranco/+archive/ubuntu/locutusofborg-ppa/+build/13166232 | 11:27 |
| LocutusOfBorg | the valgrind output is between BEGIN LOG and END LOG | 11:28 |
| LocutusOfBorg | https://launchpad.net/~costamagnagianfranco/+archive/ubuntu/locutusofborg-ppa/+build/13161434 this one is with strace | 11:30 |
| LocutusOfBorg | (keyword is BEGIN OUTPUT= | 11:30 |
| LocutusOfBorg | +==15994== Invalid read of size 4 | 11:33 |
| LocutusOfBorg | +==15994== at 0x499D362: ??? (in /usr/lib/arm-linux-gnueabihf/libpython3.6m.so.1.0) | 11:33 |
| LocutusOfBorg | this in particular looks scary to me | 11:33 |
| -queuebot:#ubuntu-release- New: accepted farmhash [amd64] (artful-proposed) [0~20170626-g23eecfb-1] | 11:39 | |
| -queuebot:#ubuntu-release- New: accepted farmhash [armhf] (artful-proposed) [0~20170626-g23eecfb-1] | 11:39 | |
| -queuebot:#ubuntu-release- New: accepted farmhash [ppc64el] (artful-proposed) [0~20170626-g23eecfb-1] | 11:39 | |
| -queuebot:#ubuntu-release- New: accepted farmhash [arm64] (artful-proposed) [0~20170626-g23eecfb-1] | 11:39 | |
| -queuebot:#ubuntu-release- New: accepted farmhash [s390x] (artful-proposed) [0~20170626-g23eecfb-1] | 11:39 | |
| -queuebot:#ubuntu-release- New: accepted farmhash [i386] (artful-proposed) [0~20170626-g23eecfb-1] | 11:39 | |
| LocutusOfBorg | infinity, if you want I did both logs in a single run https://launchpadlibrarian.net/331134898/buildlog_ubuntu-artful-armhf.notmuch_0.25-2ubuntu1_BUILDING.txt.gz | 11:50 |
| infinity | LocutusOfBorg: So, this is curious. I just ran the build (with testsuite) on a machine basically identical to the builders, and it passed. | 12:05 |
| infinity | The only exception is that I didn't run it under linux32. I'll try that now. | 12:05 |
| infinity | LocutusOfBorg: Aaaand, went fine under linux32 as well. So, uhm. Wat. | 12:11 |
| infinity | I wonder if this is a qemu/kvm bug. :/ | 12:11 |
| infinity | LocutusOfBorg: I think it's fair to say that notmuch itself isn't being miscompiled here. However, that leaves open the question of WTF *is* going wrong with the testsuite on the buildds. | 12:20 |
| LocutusOfBorg | the question now is: did the qemu/kvm got upgraded in the meanwhile? | 12:36 |
| LocutusOfBorg | the latest successful build is here created on 2016-03-13 https://launchpad.net/ubuntu/+source/notmuch/0.21-3ubuntu2/+build/9344826 | 12:37 |
| LocutusOfBorg | and the first bad is this one: Started on 2016-08-31 https://launchpad.net/ubuntu/+source/notmuch/0.22.1-2ubuntu1/+build/10600002 | 12:37 |
| LocutusOfBorg | Kernel version: Linux kishi10 3.2.0-98-highbank #138-Ubuntu SMP PREEMPT Mon Jan 11 13:24:41 UTC 2016 armv7l | 12:37 |
| LocutusOfBorg | Kernel version: Linux bos01-arm64-024 4.2.0-42-generic #49-Ubuntu SMP Tue Jun 28 21:24:20 UTC 2016 aarch64 | 12:37 |
| LocutusOfBorg | so, in the first case armhf was ran on top of an armv7 kernel, in the other case it became an arm64 one | 12:38 |
| LocutusOfBorg | this might not even be a regression in qemu/kvm, but rather a change in buildd system that spot a new bug | 12:38 |
| LocutusOfBorg | I'm doing a xenial build here https://launchpad.net/%7Ecostamagnagianfranco/+archive/ubuntu/costamagnagianfranco-ppa/+delete-packages | 12:40 |
| LocutusOfBorg | just to see if this is still building on xenial toolchain | 12:40 |
| LocutusOfBorg | something strange is happening on autopkgtest builders | 14:21 |
| LocutusOfBorg | nplan is building since days, being retried continuously (according to the webpage) | 14:21 |
| LocutusOfBorg | Triggers:['python3-defaults/3.6.1-0ubuntu2'] | 14:22 |
| LocutusOfBorg | Requester:costamagnagianfranco | 14:22 |
| LocutusOfBorg | the log seems to be however not showing it | 14:22 |
| LocutusOfBorg | Laney, can you please have a quick look? :) | 14:23 |
| infinity | LocutusOfBorg: The successful build you pointed at wasn't done under qemu. | 14:28 |
| infinity | LocutusOfBorg: kishi was bare metal armv7, bos01* are qemu/kvm instances on armv8 hardware. | 14:30 |
| infinity | LocutusOfBorg: My local test was on hardware nearly identical to bos01*, but not under qemu/kvm. | 14:30 |
| LocutusOfBorg | in fact, I retried a xenial build, and it failed now | 14:30 |
| infinity | LocutusOfBorg: Hence my blaming qemu/kvm, for now. | 14:30 |
| LocutusOfBorg | now, I tried a xenial build with strace, so I can see if the SIGILL is the same | 14:30 |
| LocutusOfBorg | in case, how to report such qemu/kvm issues? | 14:31 |
| infinity | Might take some more digging to make it a useful bug report. | 14:31 |
| LocutusOfBorg | yep, but I don't know how to easily reproduce, and I don't have access to the infra | 14:32 |
| infinity | But for now, "fails the testsuite under qemu/kvm, works fine on real hardware" filed against qemu *and* linux (could be kernel or qemu) would be a start. | 14:32 |
| * infinity goes to get some sleep. | 14:32 | |
| * LocutusOfBorg just woke up :p | 14:33 | |
| LocutusOfBorg | thanks infinity I will, and maybe subscribe you for knowledge if you want | 14:33 |
| LocutusOfBorg | since armhf is fine, can we please have a rebuild of perl stuff against proposed pocket? | 15:26 |
| LocutusOfBorg | can anybody please remove libjaxp1.3-java-gcj? it seems to be NBS libjaxp1.3-java (1.3.05-2ubuntu3 to 1.3.05-3) | 17:17 |
| slangasek | doko, apw: looks like binutils 2.29 is blocked for a bit in artful-proposed because linux-tools still has a dep on binutils (< 2.29); what would be the timeline for 4.11.0-12? | 17:24 |
| slangasek | LocutusOfBorg: buh, how does proposed-migration not get that case right? sorting | 17:29 |
| LocutusOfBorg | I don't know, sorry :) sometimes I'm not sure about what is automagic and what needs hammer | 17:30 |
| slangasek | LocutusOfBorg: yeah, my griping here is that this seems like something we *ought* to have made automagic before now but it clearly isn't | 17:31 |
| slangasek | this is the "source package previously did arch: any builds, now is arch: all only" | 17:31 |
| LocutusOfBorg | mmm I don't fully get it | 17:32 |
| LocutusOfBorg | ok the any-> all move, but two binaries disappeared too | 17:32 |
| LocutusOfBorg | (well, one because the dbgsym is not a real binary I guess for your system) | 17:33 |
| slangasek | something in the code assumes that if there are no arch-specific binaries and it previously saw some, this means the package isn't yet built :P | 17:33 |
| LocutusOfBorg | oh, interesting :) | 17:34 |
| LocutusOfBorg | meh, as long as you are around, my minions are happy to do more work | 17:34 |
| apw | slangasek, it ought to be copyable out to -proposed on monday | 17:42 |
| apw | slangasek, possibly before if we decide we don't care about the one that is there | 17:43 |
| slangasek | apw: Monday sounds fine to me. how long from copy-to-proposed to ready-to-migrate? | 17:44 |
| apw | normally testing is a day assuming it is good | 17:44 |
| * slangasek hits a bunch more perl autopkgtests with a hammer, then goes afk | 17:48 | |
| slangasek | apw: yeah, a couple more days is fine, thanks. I just wanted to understand whether we were looking at 3 weeks of blockage or something, given that the gcc-7 transition is due to start next week | 17:48 |
| LocutusOfBorg | thanks slangasek for perl :D | 17:49 |
| LocutusOfBorg | I did a no-change rebuild of src:stacks because of armhf and s390 were in some "new queue", before they were not built, and after they appeared after the package migrated | 17:55 |
| LocutusOfBorg | for this reason they ended up in some "limbo" in artful-proposed | 17:56 |
| LocutusOfBorg | stacks | 1.46-1 | artful/universe | source, amd64, arm64, i386, ppc64el | 17:56 |
| LocutusOfBorg | stacks | 1.46-1 | artful-proposed/universe | armhf, s390x | 17:56 |
| LocutusOfBorg | hopefully rebuilding should work | 17:56 |
| * LocutusOfBorg goes AFK, have a nice saturday to everybody | 17:56 | |
| slangasek | LocutusOfBorg: I don't know why you reuploaded samtools. samtools was already built against libhts2 and was already a candidate for migration; now the libhts transition is delayed further | 19:05 |
| slangasek | LocutusOfBorg: and stacks was also fine, the new binaries simply depended on the newer samtools migrating first because the samtools in artful had no s390x or armhf builds | 19:07 |
| slangasek | LocutusOfBorg: in fact, it looks like you reuploaded all the revdeps of htslib without checking whether this was required | 19:46 |
| LocutusOfBorg | I didn't get the stacks issue, I thought this was a "migrating in different timelines" issue | 21:31 |
| LocutusOfBorg | I saw that all of them were already built, and I even looked at build logs, but I thouth britney wasn't able to let them migrate because of same versions already in archive or something similar | 21:32 |
| LocutusOfBorg | I remember in the past some race-condition caused similar issues, and no-change rebuilding was the fix | 21:32 |
| LocutusOfBorg | sorry! | 21:32 |
| slangasek | LocutusOfBorg: did you look at update_output.txt? | 22:02 |
| slangasek | it should have shown that all of these packages were ready to migrate and that only blasr was holding up - they literally would've migrated in the next p-m run :) | 22:03 |
| slangasek | anyway, they've migrated now, with a bit of skiptesting of the redundant test runs | 22:03 |
| LocutusOfBorg | yes, I looked at it, and I saw samtools on s390x | 22:08 |
| LocutusOfBorg | I looked at runtime dependencies of samtools, to see if some perl was holding it | 22:08 |
| LocutusOfBorg | and I found nothing, literally nothing | 22:08 |
| LocutusOfBorg | and then I though about a race condition, nice to see it fixed :) | 22:09 |
| LocutusOfBorg | probably britney was showing s390x because it is the first architecture tested? | 22:09 |
| LocutusOfBorg | starting by amd64 would be better, at least "testable" in my pbuilder environment | 22:10 |
| LocutusOfBorg | and auto-trackers will save me for badly understanding britney! | 22:14 |
| LocutusOfBorg | anyway, g'night! | 22:14 |
| LocutusOfBorg | doko, lots of qt stuff will fail with gcc-7, so I presume I will upload 5.9 after gcc-7 switch, to avoid useless bug fixing | 22:15 |
| LocutusOfBorg | qt is already mostly ready-to-land in silo | 22:16 |
| LocutusOfBorg | tsimonq2, ^^ | 22:16 |
| tsimonq2 | Works for me. | 22:17 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!