=== Acerio_ is now known as Acerio
[05:58] <cimbakahn> I can't find anyone at any of the lubuntu rooms.  Can i get help here?
=== dkessel_ is now known as dkessel
=== attente_ is now known as attente
=== Mirv_ is now known as Mirv
=== G__ is now known as G
[14:41] <hjd> Someone around who can retrigger failed builds for Xenial? It looks like felix-main, guice, httpcomponents-core and jackrabbit were synced at the same time as newer versions of some of their dependencies. I tried with httpcomponents-core and that built successfully locally at least. Could be argued that they might need stricter version requirements for these dependencies, but I don't know whether that's an issue or not...
[15:15] <ginggs> hjd: rebuilding soon
[15:21] <hjd> ginggs: ty :)
[15:22] <hjd> I also found the curious case of bug 1518701, but I don't know how issues like that are resolved in practice...
[15:22] <ubottu> bug 1518701 in phpunit (Ubuntu) "Circular dependency: unable to install phpunit from -proposed" [Undecided,New] https://launchpad.net/bugs/1518701
[15:51] <ginggs> hjd: have you filed a bug in debian for that circular dependency?
[16:35] <hjd> ginggs: No, I haven't.
[16:38] <hjd> I was mainly looking into what the root cause could be, but I wasn't sure how to proceed.
[16:39] <hjd> Any suggestions/tips on whether I should include more that what's in the original report, if I end up forwarding it to Debian? (For instance, I haven't found a good way to mark multiple packages as affected there like on Launchpad, but I might be missing something obvious) :)
[16:57] <JanC> seems like Debian knows about that issue: http://metadata.ftp-master.debian.org/changelogs//main/p/php-codecoverage/php-codecoverage_3.0.2+dfsg-1_changelog
[17:23] <mapreri> that's really not a debian problem.  some kind of bootstrap is needed, like uploading a binary as done in debian, but this really requires super-super powers in ubuntu
[17:24] <mapreri> well, unless it's actually possibile to drop the dep, which i fear is not, but worth a try maybe
[17:25] <JanC> maybe just changing the dependency versions can fix it?
[17:25] <JanC> if one can work with an older version of the other
[19:32] <cjwatson> hjd,ginggs,mapreri,JanC: I'll bootstrap it.
[19:34] <mapreri> cjwatson: cool.  OOI are you going to upload the binary of php-codecoverage ?
[19:35] <cjwatson> mapreri: Only to a bootstrap archive used only for build-deps.
[19:35] <cjwatson> mapreri: Even I can't upload binaries directly :-)
[19:35] <cjwatson> mapreri: But I do have superpowers to inject build-deps.
[19:35] <mapreri> wow
[19:36] <mapreri> cjwatson: but are there people who can push binaries directly to the ubuntu archive nowadays?
[19:36] <cjwatson> mapreri: No.
[19:36] <mapreri> oh, that's cool, actually
[19:36] <cjwatson> Quite deliberate :-)
[19:36] <mapreri> eheh
[19:36] <cjwatson> Hasn't been possible for a long time.
[19:37] <cjwatson> (If it ever was, which I slightly doubt.)
[19:37] <cjwatson> I guess maybe pre-Launchpad.
[19:38] <mapreri> maybe, but those details are hard to grasp from the outside...
[19:38] <cjwatson> infinity: Oh hai.  The xenial chroots don't include the bootstrap archive right now.
[19:38] <cjwatson> OK, so I can't do this easily right now, can do it once infinity deals with the above.
[19:38] <cjwatson> (Could in principle, but I don't really want to upload new chroots ...)
[19:45] <infinity> cjwatson: It was possible even post-LP, I used to inject binaries into the upload queues to bootstrap waaaay back when.
[19:46] <infinity> cjwatson: But I haven't done that in years, and I *think* the shell access required to do so is something none of us have anymore anyway.
[19:46] <infinity> cjwatson: Anyhow, I can refresh and fix the chroots.  They're still copies of wily-release chroots right now.
[19:53]  * infinity glares at his Unity launcher, which seems to have decided that auto-hiding isn't hip and cool anymore, despite its settings.
[19:56] <cjwatson> infinity: Ah, yes, now you mention it I think that was one of my motivations for rewriting the archive admin scripts. :P
[19:57] <infinity> cjwatson: Yeah.  We went from bootstrapping via binary-into-archive injection (eeeevil) to chroot injection (less evil, but still gross) to the bootstrap archive (mostly not gross, ish).
[19:57] <infinity> cjwatson: But I distincly remember perpetrating the evil method in the early LP days, since it was just a carryover of the status quo from jackass.
[19:58] <Unit193> cjwatson: Do you have a nice page somewhere on using reposurgeon to convert bzr repos to git?
[20:01] <infinity> cjwatson: After I'm done my lazy afternoon TV watching, I'll refresh the xenial chroots and turn the bootstrap back on.
[20:04]  * infinity is finding it oddly satisfying to watch the s390x stage1 bootstrap autobuild.
[20:04] <infinity> Well, "auto"build.  Shell loop of doom, but whatever. :P
[20:10] <hjd> cjwatson: Thank you :)
[20:10] <hjd> infinity: s390x for Ubuntu?
[20:20] <infinity> hjd: http://mainframeinsights.com/ubuntu-distribution-announced-for-linuxone-and-ibm-z-systems/
[20:25] <mapreri> now, why does http://archive.ubuntu.com/ubuntu/dists/xenial/Release lists armhf, arm64, ppc64el, powerpc which are not in archive.u.c but on ports.u.c, and conversely ports.u.c's Release lists amd64 and i386, which are only on archive.u.c
[20:25] <mapreri> -.-'
[20:25] <mapreri> that's ain't funny, guys
[20:27] <infinity> mapreri: Because the archive split is done on a machine other than the one that signs Releases, so it can't perform surgery and re-sign.
[20:28] <infinity> mapreri: A bug, perhaps, but hardly one that causes any real issues, since it's been like that for a decade.
[20:29] <mapreri> infinity: usually how can a person figure whether an architecture is in ports or the main archive without trying and hitting 404?
[20:29] <infinity> mapreri: archive = amd64/i386, ports = everything else?
[20:30] <JanC> well, that could change of course
[20:30] <mapreri> infinity: will it ever be that way, written in a rock?  (ok, not that i expect something so stable, but listing archs is evil)
[20:31] <mapreri> well, it has been for the past decade...
[20:31] <infinity> It *could* change, but it's based on traffic/demand, as long as x86 so firmly eclipses all the other ports, it's not likely we'll change the list.
[20:32] <mapreri> isn't ports mirrored?
[20:32] <infinity> If it did change, I don't think it would be a hard cutover, we'd do something like publish ppc64el to both archive and ports, so the world isn't disrupted.
[20:32] <infinity> ports has ccTLD DNS in place, but it's only mirrored in the UK and US right now.  Would be nice to expand that, but we don't have people beating down our door to do so.
[20:33] <mapreri> eheh
[20:37] <hjd> infinity: aaah, I see
[20:37] <JanC> you could also parse e.g. http://archive.ubuntu.com/ubuntu/dists/xenial/ & http://ports.ubuntu.com/dists/xenial/ (or ftp:same) for Contents-*.gz  :)
[20:39] <Unit193> infinity: What're you doing active on the weekends for?! :P
[20:39] <infinity> Unit193: Being a derpaholic.
[20:40] <Unit193> Haha, alright. :D