[01:36] lamont: So, a combination of user error and bad input validation. [01:38] infinity: root cause: perl5 is gone [01:38] (very long story) [01:38] and autoremove took out libnss-db [01:39] the fix was to fix things so that libnss-db came back [01:39] lamont: Oh, wow, you didn't have libnss-db marked manual? [01:39] Oops. [01:39] after rooting my desktop to figure it all out, of course. [01:39] it's a dep of ud-ldap [01:39] which used to depend: perl5 [01:39] :( [01:39] lamont: Anyhow, yeah, totally down with calling that user error for the first half of the bug report, but the segv is clearly crap even with that. [01:39] yeah - segv is never the right response from the progrma [01:39] * infinity nods. [01:42] http://paste.ubuntu.com/15606351/ <-- soo..... is that debmirror, or apt, that is totally failing to acknowledge a mirror of a ppa archive, simply because Translation-en isn't there? [01:49] The requested URL /lamont/prairie-net/ubuntu/dists/xenial/main/i18n/ was not found on this server. <-- infinity why does my xenial ppa not have a i18n directory? [01:51] meh [01:51] Surely it must. [01:51] nope [01:51] I see it in your var/lib/apt... [01:51] ppa.launchpad.net_lamont_prairie-net_ubuntu_dists_xenial_main_i18n_Translation-en [01:52] interesting [01:52] debmirror failed to fetch it, and then apt decided that the repo should be ignored and left in partial [01:53] Erm. Confused a bit. [01:54] InRelease references it, but it sure doesn't look to be there.. [01:54] wgrant: WTF? [01:54] wgrant: http://ppa.launchpad.net/lamont/prairie-net/ubuntu/dists/xenial/ [01:55] wgrant: y no main/i18n on disk? [01:55] http://paste.ubuntu.com/15606413/ [01:55] (Or why is it referenced in Release, but not on disk) [01:56] lamont: Short answer, I have no idea what's up. Long answer, ask wgrant or cjwatson. [01:56] http://paste.ubuntu.com/15606426/ <-- that's a find of the xenial tree after the debmirror [01:57] infinity: ack [01:57] I shall pester them [01:57] oh. wgrant might even be around [01:57] Timezone-wise, he should be. [01:57] Maybe this is fallout from mangling things to re-sign PPAs. [01:58] Or maybe i18n has always been busted, and you're the first person ever to notice. :P [02:00] lamont, infinity: It's there, just not listable, and some files within it are verboten. [02:00] Because of the apt probing silliness. [02:00] Colin and I discussed it last week and came to the conclusion that we may be able to eventually conclude that the Apache 404 rules can be removed. [02:00] So a mirror tool that follows Release should be able to get the files? [02:00] Yes. [02:00] wget -r won't. [02:00] * infinity wonders what debmirror's up to, then. [02:00] wgrant: debmirror doesn't [02:01] lamont: I wonder if debmirror is still hardcoded to Translation-*.bz2. [02:01] and when did translations files become critical? [02:01] lamont: Which version are you using? [02:01] uh, trusty's debmirror;... [02:01] We dropped bz2 for xenial indices last week. [02:01] Only gz and xz exist now. [02:01] sounds like time to SRU debmirror/trusty [02:02] ./debmirror:Additionally download FlangE.bz2> files, which contain [02:03] next if $filename !~ /bz2$/; [02:03] Brilliant. [02:03] Yeah, trusty's debmirror only supports bz2. [02:03] Well, easy fix at least. [02:03] lamont: SRU away, we can fasttrack it if you test it in proposed and the fix is obvious. [02:03] I assume it's a 1-liner. [02:04] Can probably just remove the next if !bz2 bit. [02:04] I think.' [02:04] Except you need to skip uncompressed files because apt. [02:04] next if $filename !~ /xz$|bz2$/; [02:04] did it for me [02:04] Except when they do exist. [02:04] For the SRU I'd allow gz as well [02:05] Might as well ensure that it copes with all three compressions for all types of files [02:05] gimme a package and I'll smashtest it [02:05] I see something without xz support just above. [02:06] oh debmirror [02:06] So cute [02:06] Attempts to compress the uncompressed file exactly the same way as the server. [02:07] lamont: Colin's upstream for debmirror now, so probably worth filing a bug and pointing him at it. [02:08] The changes are not entirely trivial, due to the !slow_cpu path. [02:08] Yeah, I see a few xz-related commits. [02:09] I'd fix it if it wasn't Perl, but ew Perl. [02:10] ok. /me files [02:10] https://anonscm.debian.org/cgit/collab-maint/debmirror.git/commit/?id=417f28d5396839a5dbe6ea9c3dd00a14dcf12ef7 might be enough? [02:10] ... [02:10] Might miss the translations bit, though. [02:10] why [02:10] why [02:10] That drops bz2 support :/ [02:10] can't SRU that [02:11] But yeah, that's basically what my test diff does, modulo the removing bz2 bit. [02:11] Am testing. [02:11] Colin's later commits clean it up a lot. [02:11] But that's less obviously reviewable/SRUable. [02:12] wgrant: Dropping bz2 support almost seems like a feature to me, as long as there's gz support. :P [02:12] bzip2 was the worst thing to ever happen to apt. [02:12] cjwatson: https://bugs.launchpad.net/ubuntu/+source/debmirror/+bug/1565585 is for you, apparnetly [02:12] [ 0%] Getting: dists/xenial/main/i18n/Translation-en.gz... ok [02:12] Launchpad bug 1565585 in debmirror (Ubuntu) "fails to mirror translations, resulting in apt hating the archive" [Undecided,New] [02:12] [ 0%] Getting: dists/xenial/main/i18n/Translation-en.xz... ok [02:12] infinity: It'll break other clients. [02:13] wgrant: I know. I agree it's obvious not SRUable to drop the support. [02:13] Just sayin'. I can't wait for bz2 to die. [02:13] I might backport a few things to precise after this, since my mirror is still precise. [02:13] Fetched 115 kB in 0s (255 kB/s) [02:13] Reading package lists... Done [02:13] # [02:13] wgrant: heh [03:39] stgraber: ^-- Review? [03:40] looking [03:41] looks reasonable [03:41] kinda surprised the apparmor issue wasn't found with initial testing :) [03:42] stgraber: The bad includes, or the /dev/shm thing? [03:42] The bad includes should have shown up to anyone testing, yes. [03:42] The /dev/shm thing is more subtle, it seems. [03:42] was thinking of the bad includes [03:43] yeah, I can certainly see how only some squid configurations would need /dev/shm [03:43] Anyhow, I have a working squid-deb-proxy OOTB without mangling anything, using squid from -proposed, so I call that success. [03:43] (Well, using the one you just accepted) [03:45] We really should unleash this on the world so we can find the other 30 bugs I didn't hit. [03:49] The apparmor bits drove me nuts until I remembered that apparmor's a thing and thought to check dmesg. [03:49] That's either proof that we don't have enough confined daemons, or most of the profiles are actually correct. Not sure. [03:51] if you were using containers, you'd know to check dmesg before doing any further debugging of anything :) [03:52] apparmor blocking some random crap is the #1 source of problems [03:52] well, we got most of them sorted by now but still get a few new ones whenever Lennart feels like annexing something new for systemd to do [05:04] ^ that golang-1.6 makes it simple to fix the docker.io ftbfs on s390x [05:31] mwhudson: looking [06:51] slangasek: ta [07:18] yay https://launchpad.net/ubuntu/+source/docker.io/1.10.3-0ubuntu4 === s8321414_ is now known as s8321414 [09:12] infinity: mysql-workbench builds against mysql-5.7 if it's updated to 6.3.6 (current is 6.3.4), but this also requires updating libmysqlcppconn [09:12] to 1.1.7 (current 1.1.3) === seb128_ is now known as seb128 [11:38] any chance somebody can proceed to do the RMs at lp #1556226 lp #1556229 ? (— cjwatson ?) :) [11:38] Launchpad bug 1556226 in scribus-ng (Ubuntu) "please remove scribus-ng from the archive" [Undecided,New] https://launchpad.net/bugs/1556226 [11:38] Launchpad bug 1556229 in scribus-ng-doc (Ubuntu) "please remove scribus-ng-doc from the archive" [Undecided,New] https://launchpad.net/bugs/1556229 [11:40] I wish somebody else would take more responsibility for removals - it needs to not be me [11:40] mapreri: "sincerely hope", but apparently didn't check? [11:40] there are no reverse *dependencies*, but the recommends from edubuntu-desktop should be update [11:40] d [11:41] I'll go and do that now [11:43] cjwatson: let's just say I ran dak rm (and got it remove from debian too), but didn't check on ubuntu too, hence the hope. [11:43] mapreri: reverse-depends makes it pretty easy to check [11:44] please do that in future - archive admins should not be responsible for cleaning stuff up before the removal [11:45] umh. https://paste.ubuntu.com/15611023/ [11:45] WFM, maybe proxy trouble? [11:45] sorry, my proxy running wild [11:51] cjwatson: ok, sorry, I'll do such check next time. I should have run seeded-in-ubuntu too... === lan3y is now known as Laney === Laney is now known as Guest76158 === Guest76158 is now known as Laney [12:25] mapreri: done now [12:29] cjwatson: thank you. [13:34] ^^ I'd appreciate a quick review on mysql-5.7 please. We expect to start the transition very shortly. [13:34] (and this fixes some transition issues we picked up in a test PPA) [15:57] doko: ^ this fixes the FTBFS of "swedish" and presumably a few other similar dict packages [16:03] pitti, ta, but somebody else accepted these [16:04] * pitti looks at dict-*, these are due to something else (they use aspell, not ispell) [16:05] pitti, we should either update these dicts, or remove them (all the south african dicts). [16:06] when oh when will LP buildds default to C.UTF-8.. [16:06] doko: is there somewhere to update from, i. e. some nwer upstream? [16:06] pitti: err ... about a month ago [16:07] cjwatson: oh nice, it does that now? thanks [16:07] yes, as of launchpad-buildd 138 [16:10] rbasak: Did you catch Skuggen's note about workbench and libmysqlcppconn? [16:10] infinity: I chatted with him earlier. I've looked at the upstream changelogs. They don't seem particularly risky to me, apart from that we're late in the cycle. [16:11] infinity: does bumping them seem OK to you? [16:11] infinity: for myodbc, Skuggen tells me that they're working on it as a top priority. We're expecting some patching to fix it, and will need to work on the packaging. [16:11] (update to cmake) [16:11] pitti, well, I packaged these in 2007 ... from upstream repos [16:12] Their current ETA for the patching is midweek this week. Skuggen is working on the packaging update concurrently. [16:13] infinity: finally, they consider mysql-proxy abandoned, and suggest a replacement product instead (mysql router?) that I presume isn't packaged. So mysql-proxy isn't planned to work in the future, so their proposal is to delete it from Xenial. [16:13] infinity: I think that's all the major ones. We have some lingering FTBFS that we haven't dug into yet, but they mostly look unrelated to the update to 5.7. [16:13] rbasak: +1 on the workbench/libmysqlcppconn bump. We can talk about the rest when I'm more awake. :P [16:13] OK [16:14] * rbasak goes back to MongoDB fun [16:14] I woke up to two geese getting their spring fever on on my balcony. [16:15] Not a good way to start the morning. [16:15] around here it's generally pigeons shouting VERY LOUDLY outside the window about how great said spring fever is [16:15] COO COOOO COO to you too [16:16] cjwatson: Canadian geese tend to be a bit louder than the average pigeon, though. :) [16:16] We have a particularly persistent owl near us; friends of ours used to live right around the corner and went full schadenfreude when I mentioned it to them. [16:16] * rbasak only hears the occasional bit of birdsong [16:17] My new computer has no spinning fans or disks, which is nice. Complete silence. [16:17] Odd_Bloke: reminds me of https://storify.com/lilrongal/seanan-s-epic-owl-adventure [16:18] infinity: Canadian Canada geese? [16:18] slangasek: Yeah, half the honking is apology for the other half. [16:20] infinity: have you petitioned your government to introduce a Rob Ford commemorative Looney? [16:20] cjwatson: <3 [16:35] slangasek: if infinity doesn't, I will [16:41] pitti: ^^ [16:44] cyphermox: a good citizen! [16:45] slangasek: regardless of the outcome of the elections, will you petition for Trump on $bill ? :) [16:46] cyphermox: none of our currency is nicknamed the "small-handed megalomaniac", so probably not [16:46] ok, pick a coin then. that should fix nicely in small hands? [16:46] *fit [16:47] slangasek: I mean, we did put Spock on some of our currency, at least unofficially ;) [16:47] hah [16:47] (showing that anything can be done) [16:47] let's see, the nickel is worth less than the base metal it's made out of [16:48] slangasek: there you go. [17:01] "And now I know why my balcony door has a lock," he said, slightly shaken, after a goose OPENED THE DOOR and sat there calmly gazing into his living room. [17:03] infinity, rbasak: router isn't strictly speaking a replacement of proxy, since I don't think it has every feature from proxy [17:03] Skuggen: Then why deprecate proxy? :P [17:04] infinity, >_< better that than a fox =) [17:04] xnox: I dunno, man, geese can be mean. [17:04] infinity: Before my time, I'm afraid :) [17:04] though geese do pinch quite hard [17:04] infinity: proxy never got past the alpha stage [17:04] Mmkay. [17:04] Mind, I've had slightly conflicting reports on this from different people. Someone said that router started at 2.0 because it's the new proxy [17:05] Skuggen, ryeng: If the intent is to kill proxy with fire, please file a removal request with Debian, and we'll thwack it from Ubuntu too. [17:05] It would be nice if there was a migration path, but I'm guessing they're not compatible enough implementations to do that? [17:05] infinity: OK, will do. There won't be any future releases, basically EOL upstream. [17:05] No, there's no migration path [17:06] Looks like mysql-proxy was last uploaded to Debian by Norbert in 2011. [17:06] It's under team maintenance so we can do a ROM. [17:41] doko: ^ re-uploaded with added section for term/*. Sorry about that, looks like the automated go packaging script needs a bit more smarts :) [17:51] thanks [19:26] slangasek: Oh man, I don't want to touch that cups-filters upload with a 20 ft pole. Feel the urge? [19:26] (Cause, hey, random postinsts symlinking loaders all over is totally sane) [20:37] new cleared except for these new python-* uploads [21:52] can i get juju-core-1.25 rejected, we've uploaded a new package juju-core-1 based on feedback [22:49] infinity: Fixed trusty debmirror uploaded. [23:11] wgrant: Ta.