[00:03] aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa if i use objcopy to make a copy of ld-linux-armhf.so.3 with no .note.stapsdt section it works [00:18] fun [00:23] trying to build -O0 ffmpeg now [00:26] ok so this build now execution gcc commands with -O2 -O0 and -O3 in that order === kklimonda_ is now known as Guest97104 === zumbi is now known as Guest91296 === juergh_ is now known as juergh [07:16] xnox: Are you around to bounce a systemd SRU off? [07:20] ricotz: https://people.canonical.com/~ubuntu-archive/component-mismatches-proposed.svg [07:22] doko, isn't marcustomlinson on it? [07:23] doko, https://git.launchpad.net/~libreoffice/ubuntu/+source/libreoffice/commit/?h=ubuntu-eoan-6.3&id=83f37add9deb614f5fa4f79faeb5d8f575effbe2 [07:24] ricotz: I don't know, just saw it [07:25] marcustomlinson, this is what happens when scrapping old changes which are actually needed ;) [07:37] ricotz: no pain no gain huh :) === lan3y is now known as Laney === Trevinho_ is now known as Trevinho === Nafallo_ is now known as Nafallo === Nafallo is now known as Nafallo_ === Nafallo_ is now known as Nafallo [09:09] cpaelzer, coreycb: can postgresql-11 be built with llvm-9 (or -8) instead of 7? [09:10] cpaelzer: o/ [09:10] cpaelzer: bareos turns out not to be in the Eoan release pocket, so it doesn't matter for the transition [09:33] doko: I haven't looked at newer llvm for postgresql at all so far [09:33] is that a "needs to be solved today for FF" question? [09:33] or does that have some time? [09:44] doko: the discussion around when https://git.postgresql.org/gitweb/?p=users%2Fandresfreund%2Fpostgres.git&a=shortlog&h=refs%2Fheads%2Fjit landed was about v9.0 and v9.1 https://www.postgresql.org/message-id/20180124072038.jviav7h3fgkv7hto%40alap3.anarazel.de [09:45] but I found no clear statement or working reference using anything newer to give you a definitive answer [09:54] doko: it is not past sayin >=6 in any other build I checked. I have asked Myon in case he knows [10:05] doko: per Myon 9 should work but he wanted to wait until 9 is in unstable to make one swicth (7->9) leaving out the work to go to 8 [10:07] cpaelzer: could you track that for eoan? [10:08] llvm-9 is now the default [10:12] it won't be in eoan [10:12] FF is today, Debian hasn't even moved [10:13] llvm-7-dev is still there and we will pick it up together with Debian toward 20.04 [10:14] then on v9 most likely [10:14] is there a hard stop to llvm-7 in Eoan - like dropping it - on the way? [11:03] is the iso image builder for bionic on halt? http://cdimage.ubuntu.com/bionic/ [11:04] Yes, all those cron jobs are commented out at the moment. (While I can see it, I no longer operate it, so I can't tell you why) [11:05] I see, not having an updated one seems unfortunate [11:16] cpaelzer: no, but I would like to avoid three llvm versions in main ... [12:17] good morning o/ === ricab is now known as ricab|lunch === ricab|lunch is now known as ricab [13:58] is there any known problem with the autopkgtest testing under armhf architecture in eoan ? It seems to take quite some time, and things seems starting to pile up : http://autopkgtest.ubuntu.com/running#queue-ubuntu-eoan-armhf [13:59] rbasak: sorry, just to confirm, are you waiting on me to remove libglusterfs support from bareos and change dbconfig-common dependencies/etc today ? (just confirming to prioritize it) [13:59] cpaelzer: ^ [14:00] slashd: yep.. not the first time this happens :( [14:04] tinoco, ok tks [14:10] slashd: it is just slow [14:10] slashd: look here http://autopkgtest.ubuntu.com/running [14:11] it's 11 pages tests for armhf [14:11] waiting in the queue [14:12] but the numbers are low 136+162+138 [14:12] we have seen numbers with high 4 digits at times [14:13] I wonder if the queue is high becuase only a few nodes are up testing [15:27] FYI, eoan builds are now using new chroots. I know of breakage in eoan-based recipes and am working on it; AFAIK everything else is OK, so let me know if you see oddities that look like chroot problems [15:40] tinoco: no I'm not blocked on you currently thanks [15:43] rbasak: tks [16:24] cjwatson: thanks! === fenris is now known as ejat [17:57] is the excuses page stuck? Last update was ~55min ago [18:00] ok, it just updated [18:11] ahasenack: not stuck, just busy :) [18:14] I saw, the timestamp is :45, and it was live 15min later [18:39] vorlon: I see you uploaded a rebuild of ffmpeg. update_output.txt doesn't seem to be trying it though. Does that need hinting, or am I missing something? [18:39] Or is it not trying because of something else that I need to figure out? [18:40] rbasak: update_excuses shows there was a failing vlc/armhf test which I have now hinted past, so on the next refresh it should be included in update_output [18:42] rbasak: in fact the page regenerated 14m ago and I see ffmpeg included [18:42] Ah [18:42] * rbasak catches up [18:44] OK so next clamfs needs a no change rebuild for libpocofoundation62 [18:44] vorlon: ^ shall I start uploading the necessary rebuilds or are you actively doing it? [18:48] rbasak: I can batch upload them here in a minute [19:01] rbasak: k, clamfs ftbfs if you want to look into that [19:03] ack [19:05] rbasak: and this build failure is https://bugs.debian.org/925653 [19:05] Debian bug 925653 in src:clamfs "clamfs: ftbfs with GCC-9" [Serious,Open] [19:06] so I'm fine just removing the package now [19:06] Oh, OK. Sure! Thanks. [19:48] rbasak, cpaelzer: bareos tests are broken for debian as well, but service (mysql backend) does work and I could make all packages without glusterfs, can I upload it and solve the autopkgtest as a SRU after freeze ? (disabling the broken tests) ? [19:52] fixes can still be uploaded after freeze, if that's what you mean [19:52] to eoan [19:52] ahasenack: my question was about disabling the test [19:52] so migration passes [19:53] and I can fix it as SRU [19:53] what do you mean sru [19:53] like if it was sru =o) [19:53] it would be an sru only after eoan is released [19:53] err, fix after freeze, but for test [19:53] forget the word i used =) [19:53] ok :) [19:54] * ahasenack goes back into his tent :) [19:54] #) [20:02] There are some lingering rdepends on libdouble-conversion1 that I think we need to clear [20:03] From sources github-backup, haskell-bytestring-conversion, haskell-github, haskell-http-link-header, haskell-text-format [20:03] I'm not sure why haskell-bytestring-conversion hasn't been upload [20:03] ed [20:03] But some of the others have been and have build dependency issues it seems [20:05] https://bugs.launchpad.net/ubuntu/+source/bareos/+bug/1840485 [20:05] Launchpad bug 1840485 in bareos (Ubuntu) "Eoan autopkgtest regressions - Migration Excuses" [Medium,In progress] [20:06] if someone can review/dput my pkg/debdiff I would appreciate [20:13] DEP8 test (pgsql only) PASSES and changes are no-brainers [20:13] rbasak and or andrewsh[m] if you have sometime [20:13] in this busy day for you [20:30] rbasak: yes the ghc transition is not in particularly good shape currently. I'm removing most of these packages now [20:32] rbasak: what's the story on default-libmysqld-dev? [20:32] libmysqld is gone from 8.0 [20:33] I had just found that haskell-text-format in sid requires (effectively) ghc < 4.12 [20:33] Uh [20:34] Let me try that again [20:34] haskell-text-format in sid requires "base" < 4.12 [20:34] That's provided by libghc-base-dev- [20:34] That's a virtual package provided by ghc [20:35] We have 4.11 of that virtual package in the release pocket, and 4.12 in proposed [20:35] Hence it fails to build in proposed [20:36] sid's ghc provides 4.11 of that virtual package [20:36] at version 8.6.5+dfsg1-3 of ghc [20:36] eoan's ghc is also 8.6.5+dfsg1-3 [20:36] So it must be something determined at build time :-/ [20:37] vorlon: ah. I should probably have removed default-libmysqld-dev from src:mysql-defaults. [20:38] looks like it [20:38] rbasak: can you upload that now? I think that's the last missing piece [20:38] Yes. [20:38] ta [20:38] well there's also libreoffice-style-hicontrast, not sure yet what's up with that [20:38] rbasak: I uploaded an updated mysql-defaults to the ppa [20:39] hmm libreoffice-style-hicontrast is built from libreoffice source and is uninstallable [20:41] Skuggen: sorry I had missed that from the PPA and constructed it myself. [20:42] Skuggen: my first is a little different though [20:43] Yeah, it's not a complicated change, anyway, so no need to use the one from the ppa :) [20:43] There are a couple of extra bits needed than from the PPA [20:43] Drop of default-libmysqld-dev from debian/control [20:44] And from the actual dh_gencontrol line [20:44] Yeah, it still builds it I see [20:44] And I've also stopped generating it for MariaDB, since it makes no sense there either. [20:44] so I think I'm going to force hint the lot of this right now [20:44] rbasak: therefore please hold off on the defaults upload for a bit until that goes through [20:45] Here is the new src:mysql-defaults against Debian: https://paste.ubuntu.com/p/q4cxqtDrYw/ [20:45] Skuggen: ^ please review [20:45] vorlon: ack. [20:46] rbasak: LGTM [20:46] Thanks! [20:59] blah, vokoscreen just migrated as a new package w/ autopkgtests, now qtbase-opensource-src is not a candidate. Overriding. [21:06] I'm curious - are you going to kick it back to proposed afterwards? [21:07] Though ubuntustudio-video recommends it [21:15] rbasak: kick which back? [21:16] O_o [21:17] vorlon: vokoscreen [21:17] vorlon: or are you intending just to skip autopkgtest for the greater good? [21:18] Perhaps I don't follow the situation exactly. [21:18] I'm ready to upload mysql-defaults when you wish, BTW, or feel free to sponsor my debdiff above if I'm not around. [21:22] rbasak: I'm making qtbase not block on the test results, which I am assuming are actually going to pass once they have a chance to run [21:23] I see, OK [21:24] and actually, between one britney run and the next, the tests have already run and passed [21:25] Nearly there! [22:05] vorlon: OK, want me to upload mysql-defaults now? [22:05] What are your plans with libreoffice-style-hicontrast? [22:06] rbasak: wait a bit more, please; I screwed up the hints, I can never remember which is 'force' and which is 'force-hint' [22:06] OK [22:06] you don't have to upload it tonight, it'll just hang around as an uninstallable binary in eoan until we clean it up [22:06] OK [22:07] Oh, and the force-hint will just ignore the remaining uninstallables? [22:08] yes [22:27] vorlon: once mysql-8.0 migrates, shall we start the PHP transition immediately? [22:51] rbasak: that would be fine [22:51] ack [22:52] vorlon: britney doesn't seem to like you? [22:52] FYI, britney additionally wants sitplus and freeradius vs. your hint in the latest output [22:53] rbasak: britney is basically running constantly; the run whose output you now see had already started when I fixed my hint [22:53] so the one that's running /now/, whose output you don't see yet, should use the fixed hint [22:53] I have mused that we should put a starting timestamp in the output as well [22:54] having packages missing from the force-hint shouldn't be a problem, because britney will try other packages for migration after applying the hint [23:03] Oooh, here comes a flood of emails [23:06] I had been trying to infer the start time from the available timestamps (earliest of excuses, output and the log mtime) [23:07] I guess it's earlier than that then. [23:55] Oh lovely. pull-debian-source is broken. [23:55] Seems related to my port of it to Python 3. [23:55] But in the build of ubuntu-dev-tools, the shebang gets reverted to python2. WTF? [23:57] iirc there's a dh_python that does those rewritings [23:58] AFAICT it's a setuptools feature? [23:58] "feature" [23:59] The problem is that if it's going to do this then we can't have some scripts on Python 2 and some on Python 3. [23:59] Without writing them to support both etc.