[02:05] <imgbot> [03:30] <imgbot> [03:30] <imgbot> [06:49] <Mirv> kdub: camako: robru: mir is still stuck in proposed
[06:49] <robru> Mirv: ugh. Valid candidate my ass!
[06:51] <robru> Mirv: http://people.canonical.com/~ubuntu-archive/proposed-migration/update_output.txt i dont even
[06:51] <Mirv> robru: yeah that update_output looks like "the world will fail if trying to update mir"
[06:52] <robru> Mirv: and yet it passed silo verification...
[06:53] <robru> Mirv: lol, abiword to zenity. ZENITY. HOW DO YOU BREAK ZENITY.
[06:53] <Mirv> :D
[06:54] <robru> Mirv: but seriously I don't have a clue how to fix that
[06:55] <Mirv> yeah, I'm trying to update manually and see if it makes any sense, as soon as apt update would succeed
[06:58] <Mirv> so, someone has capped my bandwidth to archive.ubuntu.com to like 512kbit/s. hello 90's!
[06:59] <Mirv> oh, actually, in 1999 I had 100/10..
[06:59] <Mirv> hello 80's!
[07:00] <Mirv> in 80's I had ZX Spectrum and loaded from cassette at around 1500 bit/s, or sometimes the crazy hackers made it to 2600 bit/s turbo if you had really high quality cassette
[07:30] <Mirv> kdub: camako: so far we've only understood that update_output claims Mir would break everything, while in practice everything seems and upgrades fine (I've now installed all binary packages from source mir, from -proposed, on my laptop). we might need a wizard like cjwatson to check it eventually.
[09:02] <cjwatson> Mirv: you probably have a remaining dependency on a removed binary.  zenity breaking suggests that's probably in the vicinity of gtk.
[09:14] <cjwatson> Mirv: new libmirclient8 still Depends: mir-client-platform-mesa | mir-client-platform-android, both of which have been removed in favour of *2.  Furthermore, new libmirserver30 Depends: mir-platform-graphics-mesa | mir-platform-graphics-android, both of which have been removed in favour of *2.
[09:15] <cjwatson> kdub,camako: ^- mir bug
[09:16] <cjwatson> Mirv: when you're simulating the install for this kind of thing, check closely to make sure that it's possible to remove all the binary packages that the source package in question no longer builds.  proposed-migration pretends that they no longer exist when doing its calculations
[09:16] <infinity> Mirv: The simplest way to test that sometimes is just to hop in a chroot and "apt-get install (new packages) (broken package) remove1- remove2-"
[09:17] <infinity> Mirv: ie: tell apt "no, you can't install the ones we're removing" and see what happens.
[09:18] <infinity> In this case, "apt-get install libmirclient8 mir-client-platform-mesa- mir-client-platform-android-" was more than enough to show the issue.
[09:18] <cjwatson> Mirv: Oh, also, ubuntu-touch depends on mir-graphics-drivers-android.
[09:19] <cjwatson> Somebody should update the seeds, but the automatic update script probably won't work in this situation where the new thing is stuck in -proposed, so it's probably best to edit the metapackage directly.
[09:20] <cjwatson> Fortunately that's not rocket science.
[09:20] <infinity> cjwatson: It would be awfully nice if the try/pass/fail output gave a "would remove: bin1 bin2" line under the list of source package it's attempting an upgrade on, though.  Would be much easier to plug "apt-get install (list of fail) remove1- remove2-" into a console.
[09:21] <cjwatson> Yeah, I guess ...
[09:21] <cjwatson> patches welcome :)
[09:21] <infinity> cjwatson: I mean, I assume uploaders know what packages they're removing, but when I'm debugging someone else's fail, I need to look at diffs. :P
[09:25] <infinity> cjwatson: Who maintains britney upstream these days?  I feel like little formatting improvements like that should probably just go to Debian if I'm bothering to do them.
[09:25] <cjwatson> Niels Thykier mostly, I think
[09:25] <cjwatson> debian-release anyway
[09:26] <cjwatson> but: the NBS removal thing is a bit of an Ubuntuism to start with IIRC
[09:26] <infinity> cjwatson: My other consistent complaint (again, coming from doing "apt-get install (fail list)" is the unnecessary commas in the list, leading to the inevitable "echo list | sed 's/,//g' | xargs" which is silly.
[09:26] <cjwatson> I don't remember whether something like that would apply on either side
[09:26] <cjwatson> you'll have to see how closely it ties into the partial unstable handling
[09:27] <infinity> Though, with the complete illegality of commas as input to apt, maybe I can just make Michael accept them in the package list. :P
[09:27] <cjwatson> see also dpkg-checkbuilddeps
[09:27] <infinity> Yup.
[09:28] <infinity> checkbuilddeps has the excuse that it's control format, though.
[09:28] <infinity> And "foo (1.0) bar" is confusing without the commas.
[09:29]  * infinity takes what's left of his flu back to bed.
[10:06] <Mirv> cjwatson: infinity: thanks for all the advice. so indeed I didn't notice this "these packages are still in the archive but should be removed, and something depends on those"
[10:11] <Mirv> mir-platform-graphics-mesa/android has been replaced with *1, while mir-client-platform-mesa/android is at *2
[10:36] <Mirv> rsalveti: ^ if you're around, feel free to publish now that it does not seem anything scary, after it has built and some sanity testing. https://launchpadlibrarian.net/198975118/mir_0.12.0%2B15.04.20150227-0ubuntu1_0.12.0%2B15.04.20150228-0ubuntu1.diff.gz
[19:46]  * rsalveti waves
[22:08] <charles> rsalveti, want to approve the MP? :)
[22:08] <rsalveti> charles: yup, testing, reviewing and landing :-)
[22:08] <rsalveti> if all goes well
[22:08] <rsalveti> charles: can land the other one we got there as well
[22:08] <rsalveti> for indicator-datetime
[22:17] <rsalveti> mir finally landed :-)
[22:46] <robru> Sweet jeebus!