[00:32] -queuebot:#ubuntu-release- Unapproved: supertux (xenial-proposed/universe) [0.4.0-1 => 0.4.0-1ubuntu1] (no packageset) [09:16] -queuebot:#ubuntu-release- Unapproved: accepted grub2 [amd64] (yakkety-proposed) [2.02~beta2-36ubuntu11.1] [09:16] -queuebot:#ubuntu-release- Unapproved: accepted grub2 [arm64] (yakkety-proposed) [2.02~beta2-36ubuntu11.1] [09:23] if anyone is good at alignment issues [09:23] pdl/armhf... [09:27] xnox: could you look at libterm-readline-gnu-perl and libtext-bibtex-perl which both fail on s390x please? [09:27] even if to decide if they should be removed [09:34] robru,oSoMoN: you don't need to ask me whether it's in production yet, just watch the status of the Launchpad task on https://bugs.launchpad.net/launchpad/+bug/1633608 (though I don't recall whether this will apply to copies) [09:34] Ubuntu bug 1633608 in Launchpad itself "Bileto PPA upload rejections are lost to the ether" [High,In progress] [09:35] cjwatson, thanks, I subscribed to the bug [09:54] * cjwatson blacklists the GHC 8 transition for now [09:54] (this time in a less hacky way) [10:03] Laney, i want to say that those failures are normal, because terminals are weird on s390x. Let try building it on a possibly slightly more conventional installation, to see if e.g. the launchpad builder's configuration can be tweaked. [10:03] however these things did build before. somehow. [10:22] cjwatson, is there any way we could get the rejection messages to queuebot ? [10:23] cjwatson, or perhaps cc: the emails to -changes or something [10:28] a few years ago there was a bunch of work to store them for later auditing [10:28] but it was never quite completed [10:28] probably better to finish that than to add hacks [10:35] Laney, https://launchpad.net/ubuntu/+source/libtext-bibtex-perl/0.76-1ubuntu1 [10:35] and https://rt.cpan.org/Public/Bug/Display.html?id=96593 [10:36] xnox: nice, that's a handy find [10:43] * Laney might have a patch for pdl [10:43] * Laney is new to this alignment shizzle though [10:43] * Laney test builds [10:49] Laney, somehow readline state ends up as NONE(0) instead of INITIALIZED(2), there is INITIALIZING(1) in between as well. [10:49] enoclue at the moment why =) [10:51] found it, https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=840689 will get a better fix for it, but first coffee. [10:51] Debian bug 840689 in libterm-readline-gnu-perl "libterm-readline-gnu-perl: FTBFS on 64-bit big endian architectures" [Serious,Open] [11:09] xnox: yeah, saw that - the upstream dude says he will do a proper fix, whatever that means :-) [11:10] it means use unsigned, when unsigned is due [11:15] xnox, hopefully it means use long when long is due [11:15] always long your longs [11:16] yeah. [11:16] -queuebot:#ubuntu-release- Unapproved: webbrowser-app (xenial-proposed/main) [0.23+16.04.20160413-0ubuntu1 => 0.23+16.04.20161028-0ubuntu1] (ubuntu-desktop, ubuntu-qt-packages) (sync) [11:16] Gnu.xs:539:5: warning: initialization from incompatible pointer type [-Wincompatible-pointer-types] horum. I fail this morning [11:16] it was broken then moment they had a cast in there to anything other than void* [11:23] needs patching perl code as well. and magic bridge functions to/from C [11:23] argh [11:33] xnox, if there is a proper upstream fix coming ... could we not use the dirty hack until it comes ? [11:34] apw, well dirty hack is scary to me =) if 64 bit, and if big endian, point randomly to the next memory address, as that will be the "right" half of long. [11:34] I think i have it properly in a second. And if not, will upload dirty hack. [11:34] xnox, the layout of a long in BE is pretty well defined [11:34] true [11:35] what they are doing makes me physically sick, but it likely is "safe" [11:35] there is handing for charp vs pint, adding pulong there. [11:36] to keep up with their sickness [11:50] Laney, apw - https://rt.cpan.org/Public/Bug/Display.html?id=118371#txn-1679838 [11:50] uploaded [11:52] all is good https://launchpad.net/ubuntu/+source/libterm-readline-gnu-perl/1.34-1ubuntu1 [11:53] xnox: Ta [11:54] and my pdl seems to build so I'll copy that [11:55] apw, wanna shed some light into bus errors that started to appear on armhf in the galera-3 icinga2 and percona-galera-3 packages? http://people.canonical.com/~ubuntu-archive/transitions/html/boost1.61.html [11:55] i have a theory that the new arm64 builders are better at emitting bus errors, and they simply were there the whole time. [11:56] i.e. unaligned access or some such. [11:56] unaligned access indeed [11:56] also have you ever seen this before: https://launchpadlibrarian.net/291194249/buildlog_ubuntu-zesty-s390x.libbitcoin_2.11.0-1ubuntu1_BUILDING.txt.gz [11:56] devlibs error: There is no package matching [ld64-1-dev] and noone provides it, please report bug to d-shlibs maintainer [11:57] ah https://lists.debian.org/debian-mentors/2016/06/msg00648.html [12:00] xnox: the change to 64bit kernels triggered a few unaligned accesses for armhf, but I'm not aware of any triggered on arm64 [12:00] doko, right, i see three new failures / unagligned accesses triggered on armhf. [12:00] =( [12:01] * xnox is reading https://lists.debian.org/debian-mentors/2016/06/msg00652.html and does not understand [12:01] We're intentionally not putting any effort into that because it brings us closer to where we wanted to be anyway for phones [12:01] (Because the Android kernel does much the same thing; there was previously a problem whereby stuff would succeed on builders and then fail on devices) [12:02] So the unaligned accesses should be fixed in packages [12:02] yeah, i understand that the right thing to do is to fix the packages. === Pici is now known as ZarroBoogs [12:51] slangasek: I don't understand https://bazaar.launchpad.net/~ubuntu-release/britney/hints-ubuntu/revision/2074 - doesn't look like they were gone from proposed to me? [12:51] and they've moved back to release now [12:55] Laney, slangasek - i wonder if we can remove remaining broken / untransitioned binaries for perl to get it migrate. =) or are they non-leaf packages? [12:55] as in is all of uwsgi, or just the perl uwsgi? [12:59] xnox: They were demoted [12:59] -queuebot:#ubuntu-release- Unapproved: metacity (xenial-proposed/main) [1:3.18.7-0ubuntu0.1 => 1:3.18.7-0ubuntu0.2] (ubuntu-desktop) [13:00] I guess we'll see what britney has to say about it (if it's only these packages or if there are other problems too) [13:00] Laney, but you should remove binaries & demote src to -proposed. That way it will be stuck with failure to build and will not migrate by virtue of miracles. [13:00] or keep a block-proposed bug against them. [13:01] They were blocked [13:01] ok. /me goes back to finishing boost1.61 transition laggers =) [13:02] there's a few random failures @ proposed-migration [13:23] Laney, there appear to be a coulple of cases where the armhf builds are missing too [13:24] Laney, and indeed similarly for s390x, which is oddness [13:24] apw: I think those are ones that just got uploaded [13:25] Laney, ahh ok, then i'll wait a bit ... before listing them [13:26] pdl, some readline thing [13:26] and something else [13:26] pdl yay [13:37] slangasek, powerpc dropped as release arch in debian [14:01] -queuebot:#ubuntu-release- Unapproved: ido (xenial-proposed/main) [13.10.0+15.10.20151002-0ubuntu1 => 13.10.0+16.04.20161028-0ubuntu1] (ubuntu-desktop) (sync) [14:16] Just those few (re)removals and autopkgtest failures then [14:36] -queuebot:#ubuntu-release- Unapproved: apt (xenial-proposed/main) [1.2.14 => 1.2.15] (core) [15:32] -queuebot:#ubuntu-release- Unapproved: nova-lxd (xenial-proposed/main) [13.0.0-0ubuntu3 => 13.0.0-0ubuntu3.1] (ubuntu-server) [15:57] Laney: I did remove those packages from zesty-proposed; which ones do you see still there, at the perl 5.22 versions? [15:58] Laney: regardless, demoting them to -proposed was the wrong thing in the first place [16:00] Laney: it's possible there was a race between my removal being processed, and p-m re-copying them from -proposed after I dropped the hints but before the publisher had finished making them disappear from -proposed; this was all done during the window when publishing was really slow [16:00] Laney: (re-removed now from zesty) [16:00] slangasek: Riiiight - so what's your rule of thumb for demotion vs removal then? [16:01] Laney: the only time we ought to demote a package is if it depends on something else that's being removed, but itself is not buggy and requires no changes to become installable [16:02] slangasek: If you remove them, they won't come back via autosync, right? [16:02] Laney: they absolutely will [16:03] I thought it checked for an explicit removal [16:03] which is what we want to happen, if the package has gotten an upload in Debian that (hopefully) fixes the RC bug that got it removed from testing [16:03] if it does, that's news to me and behavior that I think should be corrected [16:03] it won't try to re-autosync at the same /version/ [16:04] I certainly see entries that look like that in http://people.canonical.com/~ubuntu-archive/auto-sync/current.log [16:04] Uh [16:04] salmon, sga [16:04] slangasek: Checking for removals was absolutely deliberate and not something I think should be corrected [16:04] (So that's why I was asking for demotion rather than removal) [16:05] cjwatson: really? I thought we had the sync-blacklist for things we were removing and didn't want back [16:05] slangasek: That was a nightmare [16:05] hmm [16:05] well, I didn't get the memo that this behavior had ever changed ;) [16:05] slangasek: It was around the time auto-sync moved to client side, and I banged on about it for a while :) [16:05] and, as I commented on ubuntu-release, leaving stuff in -proposed is also a bit of a nightmare [16:06] IMO and IME auto-syncing stuff that's been removed is just too error-prone [16:06] I'd much rather require AAs to maintain the sync-blacklist (which I still was!) than to bump stuff in -proposed and make update_excuses unusable [16:06] :/ [16:07] The auto-sync log shows what hasn't been synced due to a previous removal [16:07] I think it's sufficient to just take a brief glance through that from time to time [16:07] ok [16:07] (I used to do that) [16:08] but if my first inclination, when glancing at what's not been synced for this reason, is to sync them all and see what sticks? :) [16:08] anyway [16:08] gotta go chase dinner on the streets of Bucharest now [16:08] slangasek: the auto-sync log tries to give you an indication of why things were removed [16:09] IME it's a mix of short-term and long-term reasons [16:09] I don't at all object to things being added to the sync-blacklist when the decision is that they should never be shipped in Ubuntu [16:10] apw: Could you take care of my 4 new requests on https://bugs.launchpad.net/ubuntu/+source/kvirc/+bug/1636804 please? [16:10] Ubuntu bug 1636804 in nama (Ubuntu) "perl 5.24 demotions" [Undecided,New] [16:10] But quite often the next Debian upload after a removal doesn't in fact address the reason it was removed in Ubuntu, e.g. due to stricter build requirements; or the last version in Ubuntu had a delta that should be reapplied if the package is reintroduced [16:10] I don't know if this conversation establishes a new rule about whether you should remove or demote, sorry :( [16:10] And there was an antipattern I observed where people added stuff to the sync-blacklist that wasn't actually for a long-term reason [16:11] For instance, people were blacklisting things that were removed because they currently failed to build [16:11] That was really undiscoverable and so a bunch of packages were inadvertently not shipped when they could have been [16:11] So I tried to push for blacklisting only for long-term reasons [16:12] I like(d) using demotion of in-sync things to mean "this will probably be fixed by the next autosync" [16:12] I grant that it does cause excuses to become cluttered if that takes a while [16:12] I think it would be good to surface auto-sync's log output in a less plaintexty way [16:33] slangasek, imho removing binaries + demote src to -proposed is fine. and let those things stuck failing to build. [16:33] and auto-sync back in, if and when debian fixes things. [16:34] we need to mimic "removed from testing, broken in sid" somehow on our side. [16:40] -queuebot:#ubuntu-release- New binary: libcgicc [ppc64el] (zesty-proposed/universe) [3.2.16-0.1~build1] (no packageset) [16:40] -queuebot:#ubuntu-release- New binary: libcgicc [amd64] (zesty-proposed/universe) [3.2.16-0.1~build1] (no packageset) [16:40] -queuebot:#ubuntu-release- New binary: libcgicc [armhf] (zesty-proposed/universe) [3.2.16-0.1~build1] (no packageset) [16:40] -queuebot:#ubuntu-release- New binary: libcgicc [arm64] (zesty-proposed/universe) [3.2.16-0.1~build1] (no packageset) [16:42] -queuebot:#ubuntu-release- New binary: libcgicc [i386] (zesty-proposed/universe) [3.2.16-0.1~build1] (no packageset) [16:43] -queuebot:#ubuntu-release- New binary: libcgicc [powerpc] (zesty-proposed/universe) [3.2.16-0.1~build1] (no packageset) [16:43] -queuebot:#ubuntu-release- New binary: libcgicc [s390x] (zesty-proposed/universe) [3.2.16-0.1~build1] (no packageset) [17:26] so ppc has been dropped as a release architecture from debian. does that mean we're going to be canning the support in ubuntu? (note this does NOT apply to ppc64el) https://lists.debian.org/debian-powerpc/2016/10/msg00125.html [17:30] wxl: infinity's main workstation is a powerpc ;) [17:39] mdeslaur: so i guess i shouldn't expect a response from him any time soon XD [17:44] -queuebot:#ubuntu-release- New: accepted fwupd [amd64] (zesty-proposed) [0.7.4-2] [17:46] cyphermox, slangasek, I have confirmed $DPKG_MAINTSCRIPT_ARCH is blank during initial installation. This causes the debconf message about invalid arch to appear on valid architecture systems. How would you like to fix the logic? [19:02] -queuebot:#ubuntu-release- Unapproved: isc-dhcp (precise-proposed/main) [4.1.ESV-R4-0ubuntu5.11 => 4.1.ESV-R4-0ubuntu5.12] (core) [19:03] -queuebot:#ubuntu-release- Unapproved: isc-dhcp (trusty-proposed/main) [4.2.4-7ubuntu12.7 => 4.2.4-7ubuntu12.8] (core) [19:04] -queuebot:#ubuntu-release- Unapproved: isc-dhcp (xenial-proposed/main) [4.3.3-5ubuntu12.3 => 4.3.3-5ubuntu12.4] (core) [19:04] -queuebot:#ubuntu-release- Unapproved: isc-dhcp (yakkety-proposed/main) [4.3.3-5ubuntu15 => 4.3.3-5ubuntu15.1] (core) [19:06] -queuebot:#ubuntu-release- Unapproved: nova (xenial-proposed/main) [2:13.1.1-0ubuntu1.1 => 2:13.1.2-0ubuntu2] (openstack, ubuntu-server) [19:25] -queuebot:#ubuntu-release- Unapproved: rejected nova [source] (xenial-proposed) [2:13.1.2-0ubuntu2] [19:38] -queuebot:#ubuntu-release- Unapproved: openstack-trove (xenial-proposed/universe) [1:5.1.1-0ubuntu1 => 1:5.1.1-0ubuntu2] (no packageset) [19:48] balloons: I think we want not DPKG_MAINTSCRIPT_ARCH in the .config file; I shipped an attempt at fixing this in my PPA friday, I will look in a few seconds [19:48] can someone please reject shim-signed in yakkety queue? [19:48] cyphermox, ack. I switched it out to replace the arch by using deb_host_arch in rules, and writing out the config.in file [19:49] what do you think of that approach? [19:49] yuck [19:49] but hey, if it works... [19:49] I'd need to see the diff [19:52] cyphermox, heh. Well I can say just pushing back the dpkg --print-architecture version does work [19:52] would you prefer that? [19:53] yea, I would [19:54] want to get me a debdiff against the juju currently in proposed, I could sponsor that now? :) [19:54] sure thing. I have one for yak/zest too [19:58] what for? that one wouldn't prompt [19:58] but we also need the xenial [19:59] cyphermox, ohh, for the missing dh_install and missing dependency in the first upload [19:59] oh, right [20:01] jderose: hey [20:01] cyphermox: hey back :) [20:16] -queuebot:#ubuntu-release- Unapproved: accepted nova [source] (xenial-proposed) [2:13.1.2-0ubuntu2] [20:19] cyphermox, xenial: http://paste.ubuntu.com/23408418/ [20:20] -queuebot:#ubuntu-release- Unapproved: accepted nova-lxd [source] (xenial-proposed) [13.0.0-0ubuntu3.1] [20:26] cyphermox, I hope this works for yakkety/zesty: http://paste.ubuntu.com/23408442/ [20:31] can we close a bug that describes the regression on yakkety/zesty please? [20:32] and is there one for the prompting on xenial? [20:34] cyphermox, ohh, new bugs? This is follow-up imho on the current SRU bugs [20:35] I need to put verification failed in there, or however you wish [20:35] I think we might want to track them separately on account of having to upload on top of something already in proposed [20:35] or maybe re-close, yeah, [20:36] I would typically see it as verification failed, new upload, re-verify [20:36] wfm [20:36] I guess we should re-mention the specific bugs? [20:36] yes [20:37] cyphermox, bug 1631038 for sysctl. We don't technically have a bug for the arch issues in the conf script. I guess I would use bug 1617440 [20:37] bug 1631038 in juju-release-tools "Need /etc/sysctl.d/10-juju.conf" [High,Fix committed] https://launchpad.net/bugs/1631038 [20:37] -queuebot:#ubuntu-release- Unapproved: shim-signed (yakkety-proposed/main) [1.21.3 => 1.21.4] (core) [20:37] bug 1617440 in juju-core (Ubuntu Yakkety) "[SRU] Juju 2 GA" [High,Fix committed] https://launchpad.net/bugs/1617440 [20:38] isn't there something for the no-longer-supports-32bit? [20:39] cjwatson: +1 for only blacklisting due to long-term issues; -1 for pushing to -proposed, which leaves the bottom of update_excuses full of things that are not actionable :) [20:39] cyphermox, bug 1614969 [20:39] bug 1614969 in juju-core-1 (Ubuntu) "Juju packaging allows builds for unsupported architectures" [Undecided,New] https://launchpad.net/bugs/1614969 [20:40] balloons: "blank during initial installation" - er it should always be set, whether initial install or upgrade.... oh, are you checking this variable from the .config script? [20:40] right, cyphermox responded. ok [20:40] slangasek, yep. Absolutely confirmed nothing is set when run during install [20:41] balloons: it's not 'during install', it's "when used from the .config script" [20:41] because the .config script is called from apt initially, /not/ via dpkg [20:41] slangasek, it IS set if I run a dpkg-reconfigure. sure ... [20:43] -queuebot:#ubuntu-release- Unapproved: rejected shim-signed [source] (yakkety-proposed) [1.21.4] [20:43] slangasek, are you comfortable then just using dpkg --print-architecture for the config? [20:43] I converted the config file to use a .in file, and set it based on DEB_HOST_ARCH in debian rules and was playing with that, but it wasn't as desired, so, we're back to as it was [20:44] balloons: yes, in that context it's probably better [20:56] slangasek: all good, I'll do the sponsorings for balloons now [21:05] slangasek: could you please also review that new shim-singed? :) [21:07] shim-singed as in sung or slightly burnt? [21:08] cyphermox: probably not tonight yet, sorry [21:08] ginggs: as in slightly burnt [21:08] permanently-renamed, too [21:08] maybe I'll consider shim-sung for the next upload [21:09] slangasek: ack [21:24] -queuebot:#ubuntu-release- Unapproved: juju-core (xenial-proposed/main) [2.0.0-0ubuntu0.16.04.1 => 2.0.0-0ubuntu0.16.04.2] (ubuntu-server) [21:24] -queuebot:#ubuntu-release- Unapproved: juju-core (yakkety-proposed/main) [2.0.0-0ubuntu0.16.10.1 => 1:2.0.0-0ubuntu1.16.10.2] (ubuntu-server) [21:24] please reject juju-core in yakkety [21:24] crap :/ [21:26] ^ epoch is very bad, shouldn't have been in that debdiff [21:26] bad ballons ;) [21:26] * tsimonq2 gives balloons the epoch hat [21:26] done [21:26] * balloons feels ashamed [21:27] -queuebot:#ubuntu-release- Unapproved: rejected juju-core [source] (yakkety-proposed) [1:2.0.0-0ubuntu1.16.10.2] [21:27] stgraber: ta [21:27] balloons: Don't worry, you don't have to wear it as long as the Ducky Tie. :P [21:28] balloons: we'll find you a corner to sit in next week :) [21:28] balloons: there was another subtle issue too, see if you can spot it ;) [21:28] yea. I wonder how many minutes (days?!) I'll have to do that [21:28] stgraber: You guys going to a sprint next week? Make him an actual physical epoch hat. XD [21:29] heheheheheheheheh ;) [21:29] yeah, bunch of us sprinting in Budapest next week :) [21:29] balloons: that would be scary, if there were seconds involved. [21:29] stgraber: Oh nice! :D [21:29] cyphermox: already in Santa Fe I take it? [21:29] cyphermox: I'm on a plane to Denver here [21:30] yes, already there, about to leave for dinner [21:30] isn't 3pm a bit early for dinner? :) [21:30] he's stuck fixing my epochs [21:30] (basically, as soon as I'm done with this, then I'll be back after) [21:30] balloons: XD [21:30] 3pm US dinner. [21:31] wait wat [21:31] -queuebot:#ubuntu-release- Unapproved: juju-core (yakkety-proposed/main) [2.0.0-0ubuntu0.16.10.1 => 2.0.0-0ubuntu0.16.10.2] (ubuntu-server) [21:31] annoying clocks. [21:31] I'm handing out halloween candy, so I can munch on that. :P [21:34] stgraber: feels like it's way beyond 3pm here, and even way past 5pm cyphermox-standard-time. perhaps I've been up for too long. [21:36] that's why I booked late flights, get to sleep in before leaving and not see the timezone change :) [21:36] though that's just 2 hours west, the 9 hours east I'll have next weekend will be a bit harder to deal with I suspect ;) [21:36] pfft [23:53] -queuebot:#ubuntu-release- New binary: libcgicc [amd64] (zesty-proposed/universe) [3.2.16-0.1] (no packageset) [23:53] -queuebot:#ubuntu-release- New binary: libcgicc [ppc64el] (zesty-proposed/universe) [3.2.16-0.1] (no packageset) [23:54] -queuebot:#ubuntu-release- New binary: libcgicc [arm64] (zesty-proposed/universe) [3.2.16-0.1] (no packageset) [23:54] -queuebot:#ubuntu-release- New binary: libcgicc [i386] (zesty-proposed/universe) [3.2.16-0.1] (no packageset) [23:54] -queuebot:#ubuntu-release- New binary: libcgicc [armhf] (zesty-proposed/universe) [3.2.16-0.1] (no packageset) [23:54] -queuebot:#ubuntu-release- New binary: libcgicc [s390x] (zesty-proposed/universe) [3.2.16-0.1] (no packageset) [23:54] -queuebot:#ubuntu-release- New binary: libcgicc [powerpc] (zesty-proposed/universe) [3.2.16-0.1] (no packageset)