=== Ursinha_ is now known as Ursinha
[19:40] <highvoltage> stgraber: when should trusty alpha 1 show up on the iso tracker? I just see all the saucy's there atm
[19:41] <highvoltage> (I guess I could use the trusty-daily entry but I'm testing the current image as an alpha so would be nice if I could use that)
[19:46] <stgraber> highvoltage: I'm not the one running the show this time around since I'll be travelling in the midle of the week.
[19:46] <stgraber> highvoltage: but if that helps you, I can already turn off cron for Edubuntu and add it to the alpha-1 milestone
[19:46] <highvoltage> stgraber: that would be nice
[19:48] <stgraber> highvoltage: done. I have setup the milestone (but not enabled auto-publishing at this point), disabled dailies for Edubuntu and copied the current daily to the alpha-1 milestone on the tracker
[19:49] <stgraber> highvoltage: if you need any rebuilds, you can request them from the tracker, however as auto-publishing is off, they'll only show up on the daily milestone. If that happens, ping me and I'll copy the new build over to alpha-1.
[19:50] <highvoltage> stgraber: ok, so far amd64 and i386 both seem more or less ok
[19:51] <stgraber> highvoltage: cool. I believe I tested LTSP live a couple of weeks back to make sure I didn't break the world with the latest ldm and ltsp upload, but it looked reasonably good back then.
[19:51] <highvoltage> stgraber: yep, passed on i386 and amd64
[19:52] <highvoltage> stgraber: I'm going ahead and doing an edubuntu-server deploy too, should probably become part of the required tests on the iso tracker. last night I ran out of disk space on my VM so re-trying in a bigger VM...
[19:53] <stgraber> highvoltage: I'd expect it to fail because I haven't updated it since the samba4 => samba source transition, but that'd still be interesting to confirm :)
[22:13] <rsalveti> can anyone review/approve the android-headers package? the i386 (all) package is still in new, iirc didrocks approved the src package
[22:34] <infinity> rsalveti: What's the rationale for splitting that from hybris?  Shouldn't the headers match the library?
[22:37] <infinity> rsalveti: Also, you probably want a conflict, not a break, as your goal is to only have one of the packages installed when you're done.
[22:38] <infinity> rsalveti: Did you actually try this upgrade path?
[22:38] <rsalveti> infinity: upstream is not maintaining the headers as part of the same src package/tree
[22:38] <infinity> As a general rule, if you have an unversioned Breaks, you've probably done something wrong.
[22:39] <infinity> And, in this case, the part where it's unversioned is correct, which means the Breaks itself is wrong.
[22:39] <rsalveti> yup, with tested with newer hybris (depending on the new package)
[22:39] <infinity> rsalveti: And did android-platform-headers actually get forced off the system?
[22:39] <infinity> Cause it should have just been deconfigured, given that it's a Break, not a Conflict.  Which wouldn't be what you were looking for.
[22:39] <rsalveti> let me test in a sec
[22:41] <infinity> rsalveti: To force it right off, you probably want P/C/R: android-platform-headers, and no B.
[22:41] <infinity> rsalveti: See Policy 7.6.2
[22:42] <rsalveti> infinity: right, but it's not replacement (not 100% compatible, some internal changes done in the previous version were removed)
[22:43] <infinity> rsalveti: That's unfortunate, but still what you're looking for, since they can't both be installed together, and you want people upgraded to the new one.
[22:47] <infinity> rsalveti: Actually, Policy 7.4 implies that C/R is enough to force the old on off, if you're not comfy with the Provides, since it's not 100% compat.
[22:48] <infinity> rsalveti: Conflict+Replaces means "Replaces the package", while Replaces alone (or with Breaks) means "overwrites some files".
[22:48] <rsalveti> yup, then should probably only change breaks to conflicts
[22:48] <rsalveti> still updating here to test though
[22:49] <infinity> In theory, this should even happen at the dpkg level, no higher level tools required.
[22:50] <infinity> So, I'll tear your deb apart, s/Breaks/Conflicts/, and dpkg -i it to a system with android-platform-headers installed and see if reality matches policy. :P
[22:50] <rsalveti> infinity: yup, old package removed
[22:50] <infinity> rsalveti: With a conflict, you mean?  Cause it sure doesn't work with the breaks, just tried that. :)
[22:50] <rsalveti> infinity: with breaks
[22:51] <infinity> I disagree. :P
[22:51] <rsalveti> but not testing with dpkg -i
[22:51] <infinity> I assume you were using apt, which uses --auto-deconfigure.
[22:51] <rsalveti> yeah
[22:51] <rsalveti> let me change and test with dpkg -i
[22:53] <infinity> rsalveti: Yeah, with Conflicts it does exactly what I think it should.
[22:53] <infinity> Conflicts+Replaces, that is.
[22:53] <infinity> I just did s/Breaks/Conflicts/ on DEBIAN/control and rebuilt the deb.
[22:54] <infinity> (trusty-amd64)root@cthulhu:/home/adconrad/android# dpkg -i android-headers.deb
[22:54] <rsalveti> awesome, let upload the change then
[22:54] <infinity> Selecting previously unselected package android-headers.
[22:54] <infinity> dpkg: considering removing android-platform-headers in favour of android-headers ...
[22:54] <infinity> dpkg: yes, will remove android-platform-headers in favour of android-headers
[22:54] <infinity> (Reading database ... 11923 files and directories currently installed.)
[22:54] <infinity> Unpacking android-headers (from android-headers.deb) ...
[22:54] <infinity> Setting up android-headers (4.2.2-1-0ubuntu1) ...
[22:55] <infinity> rsalveti: For the record, this is what one gets with Breaks + auto-deconfigure:
[22:56] <infinity> http://paste.ubuntu.com/6580781/
[22:56] <infinity> rsalveti: So, apt may pull more tricks on top of that to DTRT, but I don't consider abusing apt's fragile resolver a win, so doing what works for raw dpkg is definitely saner.
[22:56] <rsalveti> yeah, makes sense
[23:00] <rsalveti> infinity: just uploaded newer version, thanks for the review
[23:00] <infinity> rsalveti: NP.  Looked sane otherwise.
[23:00] <infinity> Well, as sane as hybrid/android on Ubuntu can ever look. ;)
[23:00] <infinity> hybris*
[23:00] <rsalveti> yeah :-)