[02:44] <JonelethIrenicus> I have some remaining library I built myself and had to manually delete it from my system, but for some reason I think it still may be linked how can i get rid of it?
[07:13] <cpaelzer> ahasenack: thanks for spotting my wrong targetting of released on the spamassassin SRU
[07:14] <cpaelzer> I really think it is artful only
[07:43] <ricotz> chrisccoulson, hi, please push your thunderbird branches
[09:53] <juliank> rbasak, cpaelzer: I think my multipath-tools upload is almost ready, I dropped the uds importer merge because I was waiting for some other stuff. Can I open a new merge so one of you importer guys can merge that upload before the importer runs so we keep the log?
[09:53]  * juliank does not want to lose that, there's quite a bit of stuff in there :)
[09:54] <cpaelzer> juliank: yeah
[09:55] <cpaelzer> juliank: so you don't want an actual review now, but instead just to save your history ?
[09:55] <cpaelzer> did I get that right?
[09:55] <juliank> Hey, if you want to look at it, that's fine too. I'm just cleaning this up a bit now, and will push that out in a few mins.
[09:56] <cpaelzer> ok, ping me once it is up
[09:56] <cpaelzer> and yes, I can do the git magic to let the history be retained
[09:59] <cpaelzer> juliank: looking forward to the review - have you run tests on actual MP HW against the new code?
[10:00] <cpaelzer> if not would you mind providing a ppa along the MP so that I can do some (minor) tests?
[10:00] <juliank> cpaelzer: No, I don't have any.
[10:00] <juliank> Sure, I can do that
[10:00] <rbasak> I'm not sure I follow, but cpaelzer seems to have understood
[10:00] <cpaelzer> great, post the ppa in the MP comment
[10:01] <cpaelzer> juliank: and mid term get together with xnox and jfh to get you a z/VM guest with a bunch of scsi disks via multipath
[10:01] <cpaelzer> or an lpar if you have a spare one left in the team
[10:02] <juliank> I have no idea how to actually build a source package with git-ubuntu
[10:02] <juliank> I can manually run debuild with the correct options, but the integrations seems broken
[10:02] <juliank> $ git ubuntu build-source
[10:02] <juliank> 01/24/2018 11:01:54 - ERROR:Unable to automatically determine importer branch: Multiple candidate branches found and they do not target the same series: pkg/debian/sid, pkg/debian/buster. Please pass --target-branch.
[10:02] <cpaelzer> git ubuntu build-source --verbose --sign --for-merge
[10:03] <juliank> If I do pass a --target-branch argument it just says it does not understand --target-branch
[10:04] <cpaelzer> I've never seen the target-banch disambiguation on a build-source so far
[10:04] <juliank> Oh well, I'll just use debuild :)
[10:04] <cpaelzer> rbasak: ^^ did you?
[10:04] <cpaelzer> well let us find the issue and fix it :-)
[10:05] <cpaelzer> juliank: when you have pushed your MP I can try on it
[10:05] <cpaelzer> if I can reproduce I can file the issue as needed
[10:07] <juliank> cpaelzer: https://code.launchpad.net/~juliank/ubuntu/+source/multipath-tools/+git/multipath-tools/+merge/336526
[10:10] <rbasak> nacc: ^
[10:21] <juliank> cpaelzer: One thing to note is that I rebased on top of my for-debian branch, essentially, so the commit order might be a bit confusing :)
[10:21] <juliank> Hopefully that'll get merged soon.
[10:21] <juliank> Or rather, I'll just NMU it in delayed/7 or something
[10:21] <cpaelzer> soon able to take a look
[10:21] <cpaelzer> and be confused then :-)
[10:35] <cpaelzer> rbasak: could you check logs why ipxe 1.0.0+git-20161027.b991c67+really20150424.a25a16d-1ubuntu2 seems not to be imported?
[10:36] <cpaelzer> zesty-devel is also behind proposed, seems not recently imported at al
[10:36] <cpaelzer> but is in the whitelist
[10:41] <rbasak> I'm not sure the logs are working currently :(
[10:41] <rbasak> I didn't find any up-to-date ones
[10:41] <cpaelzer> ok, I'll kick a local import over lunchtime
[10:41] <cpaelzer> maybe it pops up something
[10:42] <rbasak> nacc: ^
[10:52] <tseliot> sil2100: hi, are you ok with the workaround in LP: #1728547 ?
[11:00] <sil2100> tseliot: oh, didn't see that, I'll take a look at it after lunch but from the comment it sounds okayish
[11:09] <tseliot> sil2100: ok, thanks
[11:26] <ahasenack> juliank: hi, saw in irc backlog that you were uploading multipath-tools, is that still the case?
[11:30] <ahasenack> request.cgi in autopkgtests is broken? I get an internal server error
[11:32] <cpaelzer> ahasenack: the MP is up of juliank
[11:32] <cpaelzer> I was just about to check if I can repro his build error on it
[11:33] <cpaelzer> rbasak: ipxe imported fine, new versions there now
[11:33] <cpaelzer> rbasak: but almost everything is a forced update now
[11:33] <cpaelzer> rbasak: did you commit the import breaking changes to beta already?
[11:33] <cpaelzer> (you or nacc)
[11:34] <ahasenack> I was going to mention an ibm bug
[11:34] <ahasenack> https://bugs.launchpad.net/ubuntu/+source/multipath-tools/+bug/1711749
[11:34] <cpaelzer> ahasenack: the multipath-tools buld of juliank fails with the equivs package just as samba
[11:34] <ahasenack> see if the patches they mention are in this upload, presumably to bionic
[11:34] <cpaelzer> it must be something generic
[11:35] <ahasenack> cpaelzer: ah, g-u
[11:35] <ahasenack> cpaelzer: ok, I have something for that maybe
[11:35] <ahasenack> let me see your paste again, just a sec
[11:35] <cpaelzer> ahasenack: interesting
[11:36] <ahasenack> cpaelzer: line 24: https://paste.ubuntu.com/26449720/
[11:36] <cpaelzer> on the multiapth-build it works on try 3/6
[11:36] <ahasenack> there is no fix-up patch to create
[11:36] <ahasenack> unless that returns empty and is just how the code checks for it
[11:36] <rbasak> cpaelzer: I've not touched the importer since nacc got back. I don't think we've bumped beta either.
[11:36] <rbasak> I should sync with nacc on this.
[11:36] <ahasenack> I noticed that because my build of your spamassassin backtraced with unicodedecode error near that area
[11:36] <ahasenack> I wonder if it's related
[11:37] <cpaelzer> weird - it seems reproducible for samba, but on multipath works on the integrated retry
[11:38] <cpaelzer> nacc: could you as a clean slate approach grab both recent merges and run build-source on them?
[11:38] <ahasenack> it worked for my clamav build
[11:39] <ahasenack> cpaelzer: doesn't line 61 confirm that it finally installed equivs?
[11:39] <ahasenack> 01/24/2018 08:59:11 - DEBUG:Executing: /snap/bin/lxc exec star-eagle -- apt-get install -y devscripts equivs sudo
[11:39] <ahasenack> I think in your case it was just a missing --for-merge, now?
[11:39] <ahasenack> because samba 4.7.4 is not yet in ubuntu, just in debian
[11:39] <ahasenack> 01/24/2018 09:00:21 - ERROR:stderr: error: open /var/lib/snapd/hostfs/tmp/review/samba_4.7.4+dfsg.orig.tar.gz: no such file or directory
[11:40] <cpaelzer> didn't work with or without
[11:41] <cpaelzer> but I agree that the equivs might be a red herring
[11:41] <cpaelzer> there also is the " Multiple candidate branches found" that juliank reported
[11:41] <cpaelzer> I see that for the samba build as well
[11:42] <cpaelzer> line 26
[11:42] <ahasenack> hm, that path does look odd
[11:42] <ahasenack> it's looking in /var/lib/snapd
[11:42] <ahasenack> that might be a recent change in the snap indeed, there was some shuffling around with the tmp dir where the build happens
[11:58] <cpaelzer> TL;DR several things are odd - @ nacc if you could (once around) try building the recent samba and multipath merge - and not related to that see if you'd see why ipxe wasn't continued to be imported
[12:15] <cpaelzer> juliank: now I found what you meant with the commit order :-)
[12:19] <ahasenack> cpaelzer: where is this multipath-tools mp, out of curiosity?
[12:20] <cpaelzer> https://code.launchpad.net/~juliank/ubuntu/+source/multipath-tools/+git/multipath-tools/+merge/336526
[12:20] <cpaelzer> I already mentioned the bug you pinged above
[12:20] <cpaelzer> but I have much more and write up the answer atm
[12:20] <ahasenack> it's not listed in https://code.launchpad.net/~usd-import-team/+activereviews
[12:20] <ahasenack> let me check against what the mp is
[12:20] <cpaelzer> ~10 minutes maybe to submit to juliank
[12:20] <cpaelzer> against me :-)
[12:21] <ahasenack> that's a big delta
[12:22] <cpaelzer> oh yeah - I did th elast few merges
[12:22] <cpaelzer> with cyphermox
[12:22] <cpaelzer> I knew what I was up to :-)
[12:26] <cpaelzer> juliank: review done, some cleanups requested
[12:27] <cpaelzer> juliank: if you want to overrule my review and upload right away let me know so I can make it known to at least retain history
[12:27] <juliank> damn, I missed notifications
[12:28] <juliank> cpaelzer: Thanks for the review, I'll fix that stuff first
[12:28] <juliank> Totally forgot the patch headers, ugh.
[12:30] <cpaelzer> well I spotted one of them which I added - so you are in (?good?) company :-)
[12:32] <juliank> cpaelzer: I'm not sure why git format-patch adds the weird date from 2001.
[12:32] <juliank> but that's what it does...
[12:34] <cjwatson> it's a fixed magic timestamp used by git format-patch
[12:34] <cjwatson> documented in git-format-patch(1)
[12:34] <juliank> ah
[12:34] <cjwatson> not what I'd have gone for but there you go ...
[12:35] <juliank> cjwatson is like an encyclopedia :D
[12:42] <cpaelzer> well if it is magic then keep it juliank
[12:43] <cpaelzer> I didn't know it was a special unicorn :-)
[12:43] <Unit193> I use format patch all the time, quite useful for IRC and pastebins.
[12:44] <cpaelzer> I guess we all do, but who of us have realized that the date after the hash is a magic thing
[12:44] <cjwatson> I do wonder why it's that particular date rather than 681200 seconds earlier
[12:47] <cjwatson> no explanation in git history AFAICS
[12:48] <cpaelzer> xnox: so many other things of the chrony change are moving fast (I didn't expect it to go that fast actually), but that makes me ask if you tnhk you'll have some time for bug 1744328?
[12:48] <cjwatson> https://stackoverflow.com/questions/15790120/what-is-the-first-line-of-git-format-patch-output
[15:53] <ricotz> chrisccoulson, you can find tb 52.6 tarball here https://launchpad.net/~mozillateam/+archive/ubuntu/ppa/+sourcepub/8739132/+listing-archive-extra
[16:51] <jibel> could someone review the patch attached to bug 1744722
[16:51] <jibel> ?
[17:07] <nacc> juliank: cpaelzer: that is due to an issue with the fetch_orig logic, i believe. Feel free to open a bug (--target-branch).
[17:08] <juliank> jibel: I can only say that it sounds ok, but I don't have any close insight into that part
[17:10] <juliank> jibel: If I understand things correctly we could just upload it to bionic and see if upgrades from zesty to it break
[17:12] <juliank> I mean, the worst thing that could happen is a few updates to bionic failing for a day or so
[17:48] <nacc> rbasak: pign
[17:48] <nacc> *ping, rather :)
[18:01] <nacc> cpaelzer: ahasenack: samba will onnly build with -propsoed enabled, afaict?
[18:01] <ahasenack> nacc: yes, it needs libldb that is in proposed
[18:01] <nacc> we don't currently expose that option (you could it manually)
[18:02] <nacc> or pass the correct flags to dpkg-buildpackage (-- -d)
[18:02] <ahasenack> which option?
[18:02] <nacc> ahasenack: is there something else i['m missing in the ping on that?
[18:02] <ahasenack> I think the error was the file-not-found when fetching the orig file, was it not?
[18:03] <nacc> ahasenack: what file not found?
[18:03] <ahasenack> the orig tarball
[18:03] <nacc> ahasenack: you need to use edge to work aroudn that
[18:03] <ahasenack> let me scroll up
[18:03] <nacc> ahasenack: i got too many pings to follow, and no bugs to read, so it's a bit tricky :)
 ahasenack: the multipath-tools buld of juliank fails with the equivs package just as samba
[18:03] <ahasenack> did you also check multipath-tools?
[18:03] <rbasak> nacc: pong
[18:04] <nacc> multipath-tools fails for a different reason
[18:04] <ahasenack> https://code.launchpad.net/~juliank/ubuntu/+source/multipath-tools/+git/multipath-tools/+merge/336526
[18:04] <nacc> rbasak: do you have time to sync still today? or want to put something on the cal?
[18:04] <rbasak> nacc: can sync now
[18:04] <rbasak> standup HO?
[18:04] <nacc> rbasak: ok, joining yep
[18:32] <nacc> ahasenack: oh you could build samba with a lxd profile probably
[18:34] <ahasenack> nacc: I wasn't trying to, this was cpaelzer reporting to me a problem he saw when he tried it
[18:34] <ahasenack> which didn't seem to come from the missing build-dep
[19:49] <nacc> cpaelzer: juliank: filed LP: #1745213 for the extraneous warning
[19:49] <nacc> warning(s)
[19:49] <nacc> cpaelzer: juliank: the multipath-toosl merge builds if i pass --for-merge
[19:49] <juliank> nacc: strange
[19:50] <nacc> (well, it's working now, let me see if it finishes)
[19:50] <nacc> juliank: where did it fail for you?
[19:50] <juliank> exporting a package
[19:50] <nacc> juliank: oh were you using the edge snap?
[19:50] <nacc> juliank: or stable?
[19:50] <juliank> installed:   0.2.2+git11.9fa9149 (291) 110MB classic
[19:50] <juliank> stable apparently
[19:51] <nacc> juliank: ah ok, we just landed the orig tarball fix to edge yesterday
[19:51] <juliank> that one looks old
[19:51] <nacc> i ened to do a release (i was out for > month last year)
[19:51]  * juliank switches to edge
[19:52] <nacc> juliank: yeah it works with edge (just finnished)
[19:55] <juliank> I'll be fixing the needs fixing stuff tomorrow I think. I have to rebase the patches for debian and then on top of that. It's a bit annoying right now, but hopefully this gets merged there eventually.
[19:56] <juliank> And I test both states, the for-debian one and the one supposed for merging in autopkgtest. Doing it right :)
[19:59] <juliank> The worst part is that these are two separate git repos :)
[19:59] <nacc> heh
[20:01] <juliank> LocutusOfBorg: you on curl merge? Seems trivial. If not, I'd be happy to.
[20:01] <juliank> It's silly merge season!
[20:01] <juliank> Except well, it's not silly
[20:02] <juliank> And it got so much easier now that the tarballs grab-merge fetches do not have partially applied patches in them
[20:17] <jbicha> nacc: could you try importing moon-buggy into git-ubuntu? I'm curious to see how it handles the brokenness from Debian bug 887740
[20:18] <jbicha> there's a new reply on the bug that hasn't been posted yet where he says it's only a LP bug that the pkgs is not syncable
[21:07] <smoser> does every upload of grub2 require a grub2-signed upload ?
[21:26] <LocutusOfBorg> juliank, feel free to steal :)
[21:31] <juliank> :)
[21:39] <juliank> I wonder if we should have build profile ubuntu for stuff like curl
[21:39] <juliank> libssh2-1-dev <!ubuntu>
[21:39] <juliank> build-depends would allow us to sync that stuff
[21:39] <juliank> infinity: ^
[21:40]  * juliank thinks curl has a weirfd test suite
[21:41] <juliank> but still better than (IIRC) coreutils
[21:41] <juliank> that one fails if you have a bind mount on your system
[21:47] <Unit193> juliank: Last time I thought that'd be a cool idea, someone said "Build profiles aren't gentoo use flags" (which, is fun because at the time I'd never used gentoo. :P )
[21:47] <infinity> juliank: See above.
[21:47] <juliank> They sort of are
[21:47] <infinity> juliank: The default behaviour of buildds should be to build without profiles, it's mind-numbingly confusing if we make that not true.
[21:50] <Unit193> I believe I had asked either in terms of the indicators specifically, or whether PPAs could parse the *.changes file and build with the appropriate profile.
[21:50] <juliank> infinity: Well, allright, I just thought it would save some work :D
[21:50] <Unit193> (And someone = cjw of course. :P )
[21:51] <Unit193> juliank: Yes, as far as I could tell the only way to do it was hacks that'd have #
[21:51] <Unit193> $universe package | package-in-main-but-not-installed, but now you can do build-depends inj uni.
[21:52] <infinity> juliank: I don't think it actually does save work.  If we built Ubuntu with a different profile, people would take advantage of that to fork packaging more often (and, sure, push that fork into more confusing rules/control in Debian, but ew) instead of either communicating with Debian about if certain deps are needed or MIRing where appropriate.
[21:52] <infinity> juliank: libssh2 is a bit of an outlier there (and it may also be time to revisit if we still hate it)
[21:53] <juliank> It would be interesting to see curl with libssh-2 and how that works
[21:53] <juliank> I tried to use curl for dput sftp but obviously failed since we have no ssh support.
[21:54] <Unit193> Yes the biggest issue for depends is annoying main packages, can say.  And while syncing would certainly be a nice gain, I do agree that having "hidden delta", that is where the version is the same but the build results aren't, isn't really my favorite idea.
[21:55] <juliank> curl tests are almost complete!!!
[21:56] <juliank> oh crap, it's doing the third variant now
[21:56] <juliank> totally forgot about nss
[22:04] <juliank> why do we spent 15 or 20 or 30 minutes running tests for all three variants and ignore the result. seems a bit pointless.
[22:04] <juliank> but oh well, not going to improve that today
[22:05] <juliank> merged
[22:07] <juliank> I still have what, 25 merges or so stuck in autopkgtests. oh the fun.
[22:07]  * juliank did not count, only guessed :)
[22:10]  * juliank did 40 uploads in 2 weeks and 3 days. that's insane.
[22:10] <Unit193> juliank: Merge dput?
[22:11] <juliank> Unit193: dput 1.0.1 is in -proposed
[22:11] <Unit193> Woah, good job.
[23:48] <cjwatson> infinity: I'm not opposed to maybe having a way to configure PPAs to build with a profile (could be handy for bootstrapping, say), but I agree that it would be best not to build Ubuntu proper with one.
[23:49] <cjwatson> Probably not even that hard these days; just needs changes at three or four different levels of the stack.