veebers | thomi: to confirm, this bug is _long_ finished right? or is there a subtlety that I'm missing: https://bugs.launchpad.net/autopilot/+bug/1242835 | 01:33 |
---|---|---|
ubot5 | Launchpad bug 1242835 in Autopilot "Port autopilot script to Python 3" [Undecided,New] | 01:33 |
thomi | heh | 01:33 |
thomi | yeah, That's done | 01:33 |
veebers | cool, thanks | 01:33 |
thomi | +1 for you doing bug triage - good work :D | 01:33 |
veebers | thomi: heh, it's taken me long enough to actually do it. Will prob. still have questions for you | 01:34 |
veebers | thomi: i.e. I can't delete a bug, I've just marked them as invalid (if they are, well, no longer valid) | 01:34 |
thomi | seeif you can get the list of open bugs to < 50 | 01:34 |
thomi | veebers: yeah, even invalid of fix released | 01:34 |
veebers | either or? | 01:35 |
thomi | well, if we actually fixed it, set it to fix released | 01:36 |
thomi | if the conditions that led to the bug are no longer around, set it to invalid or wontfix | 01:37 |
veebers | sweet, that's what I've been doing so far | 01:37 |
veebers | thomi: what's your suggestion for this bug. (currently invalid, but the code is fix released just not by the branch linked to that bug)_: https://bugs.launchpad.net/autopilot/+bug/1274139 | 02:31 |
ubot5 | Launchpad bug 1274139 in autopilot (Ubuntu) "autopilot should rather use upstart to stop an application rather than sending sigkill" [Undecided,Invalid] | 02:32 |
thomi | veebers: unlink the branch and set it to fix released | 02:32 |
veebers | thomi: sweet, cheers | 02:32 |
thomi | veebers: maybe add a note saying it was fixed in trunk, unknown version | 02:32 |
veebers | thomi: the duplicate scenario names is a testtools bug, correct? | 02:35 |
thomi | veebers: yes, Feel free to assign to me | 02:35 |
veebers | thomi: I was going to mark as won't fix or something as it's not APs fault. Would you prefer I leave as is & assign you? | 02:36 |
thomi | veebers: can you add a bug task to testtools, assign the testtools task to me, and make the AP one as wontfix? | 02:37 |
veebers | thomi: will do | 02:37 |
veebers | thomi: Created, I can't assign (not on the team) | 02:39 |
veebers | https://bugs.launchpad.net/testtools/+bug/1372726 | 02:39 |
ubot5 | Launchpad bug 1372726 in testtools "A test run should fail if scenarios have duplicate names" [Undecided,New] | 02:39 |
thomi | done | 02:40 |
thomi | thanks | 02:40 |
thomi | odd though - I don't see the AP task? | 02:40 |
thomi | did you report a new bug? | 02:40 |
thomi | it looks like you did - you really don't need to do that (and in fact shouldn't) - just click the 'also affects project' link, to add a new 'bug task' (bug task != nug) | 02:41 |
thomi | *bug | 02:41 |
veebers | thomi: ugh of course, that's a better way to do it. Sorry that slipped my mind | 02:46 |
veebers | thomi: do you want to delete that one I created and I'll do it properly | 02:46 |
thomi | nah, it's OK | 02:48 |
veebers | thomi: is there a reason not to have '_poll_time' (in DBusIntrospectionObject) an argument to wait_select_single? that would be a quick solution for bug lp:1297780 | 04:24 |
veebers | err, easier link: https://bugs.launchpad.net/autopilot/+bug/1297780 | 04:25 |
ubot5 | Launchpad bug 1297780 in Autopilot "No way to extend timeout period for wait_select_single()" [Medium,Triaged] | 04:25 |
thomi | veebers: uhh... | 04:25 |
thomi | veebers: from memory it uses Timeout.default? | 04:26 |
veebers | thomi: no, a DBusIntrospectionObject has self._poll_time = 10 set in it's __init__ which is used in wait_select . . . | 04:27 |
thomi | veebers: really? huh | 04:27 |
thomi | that seems wrong | 04:27 |
thomi | to me | 04:27 |
veebers | thomi: It would be good to have it changable per select (as we might expect a select to take a while, but any others taking longer should be a failure) | 04:28 |
thomi | veebers: maybe, but you'd need to keep the API simple | 04:29 |
thomi | and I think we already used posargs and kwargs there :D | 04:29 |
thomi | which is probably why it's an ivar - so we could add an API to change it at a later date | 04:29 |
thomi | (a context manager perhaps) | 04:30 |
thomi | anyway, I have a call - bbs | 04:30 |
veebers | thomi: ack that makes sense. aight have fun | 04:30 |
thomi | veebers: I'm back now - was there annything else? Otherwise I'll EOD now | 04:54 |
veebers | thomi: nothing pressing | 04:55 |
thomi | ok then - talk to you tomorrow! | 04:56 |
=== zbenjamin_ is now known as zbenjamin | ||
thomi | hey barry, how's it going? | 19:46 |
barry | thomi: good! how's it with you? | 19:47 |
thomi | barry: good, but I have a packaging question for you | 19:48 |
barry | thomi: sure thing | 19:48 |
thomi | I'm trying to package trv, which is on pypi. I have a bzr branch with just a debian/ folder in it | 19:49 |
thomi | I have d/watch set up correctly, so running 'uscan' downloads the new tarball into .. | 19:49 |
thomi | I then run 'uupdate ../trv_1.1.0.orig.tar.gz', expecting it to create a new directory which is the new version + my debian/ dir, but I get this error message: | 19:50 |
thomi | uupdate: could not find {diff|debian.tar}.{gz|bz2|lzma|xz} from version 1.1-0ubuntu1 to apply! | 19:50 |
thomi | So... it looks like it's looking for the debian/ directory as a tarball in .. - do I need to create that myself? | 19:50 |
barry | thomi: i've never really used uupdate tbh. this setup seems very similar to how packages are done in dpmt with svn containing debian/ only | 19:52 |
barry | in those cases, i just unpack the upstream tarball, cd into it, then symlink in the debian/ | 19:52 |
thomi | hmm, ok, I'll try that | 19:52 |
thomi | uupdate does the d/changelog entry for you as well | 19:53 |
thomi | or at least, creates the stub | 19:53 |
barry | yep, i just use plain ol' dch for that | 19:53 |
thomi | what should the new version be, if upstream version is 1.1.0? 1.1.0-0? | 19:54 |
barry | is this only in ubuntu? | 19:54 |
thomi | right now it's nowhere. I'd like to get it into at least ubuntu... debian as well if they'll have it | 19:54 |
barry | 1.1.0-0ubuntu1 | 19:55 |
barry | that way when it does show up in debian, it'll be 1.1.0-1 and that will be higher than the ubuntu version so it can replace ubuntu | 19:55 |
barry | and, if you need to rev the ubuntu version in the meantime, 1.1.0-0ubuntu2 -0ubuntu3 etc. will still sort lower than the first debian version | 19:56 |
balloons | thomi, ohh packaging trv, cool | 19:57 |
thomi | barry: the latter is just for packaging changes tho, right? | 19:58 |
thomi | if I change upstream, then the upstream part would change | 19:58 |
barry | thomi: right, then you'd have 1.2.0-0ubuntu1 etc | 19:59 |
barry | {upstream}-XubuntuY | 19:59 |
barry | where X is the debian version number, in this case 0 because there is no debian version | 20:00 |
thomi | ahh, ok | 20:01 |
thomi | cool | 20:01 |
barry | that's why you sometimes see 4.5-3ubuntu9. that's upstream 4.5, debian version 3, ubuntu delta 9 | 20:01 |
thomi | barry: ok, I'm getting closer, but now I get this: http://pastebin.ubuntu.com/8413229/ | 20:02 |
barry | you also sometimes see -XbuildY it's a similar idea in that X is the debian version but "build" does not prevent autosyncs | 20:02 |
barry | with "ubuntu" you have to manually resync the package with debian | 20:02 |
thomi | ahh ok, so 'ubuntu' has some semantic meaning - it's not just a name | 20:02 |
barry | correct | 20:02 |
barry | thomi: okay, that is dpkg-buildpackage -S being stupid. instead of symlinking debian/ into the directory, you have to physically move it in there (hard links *might* work). | 20:03 |
thomi | ahhh | 20:03 |
thomi | except you can't hard link a directory :D | 20:04 |
barry | then what i would do is once you have the .dsc, get a sane layout by `bzr import-dsc foo.dsc` in a new bzr repo, and then use source-full branch | 20:04 |
barry | thomi: oh yeah, duh | 20:04 |
thomi | yay! I made a source pacage | 20:04 |
* thomi fires up sbuild | 20:04 | |
barry | \o/ | 20:04 |
thomi | barry: what's the best way to get this into utopic, given that I'm not a UD? | 20:05 |
barry | thomi: well, you have a problem now that we're in feature freeze, so you need to first fill out an FFe, then once that gets approved (if it does), you need to find a sponsor. let me throw you some links | 20:06 |
thomi | ahh crap | 20:07 |
barry | https://wiki.ubuntu.com/FreezeExceptionProcess | 20:07 |
thomi | I forgot about FF | 20:07 |
barry | https://wiki.ubuntu.com/SponsorshipProcess | 20:07 |
barry | yeah | 20:07 |
barry | you can always throw it in a ppa of course | 20:07 |
thomi | yeah, that's what I've done till now | 20:07 |
barry | the other good approach is to try to get it into debian and then it will at least sync into voluptuous vulture | 20:08 |
barry | unless you really need it utopic, that's the approach i'd take | 20:08 |
barry | (no guarantees of course that the ffe would be granted) | 20:08 |
veebers | balloons: are you still around perchance? | 20:43 |
balloons | veebers, indeed | 20:49 |
veebers | balloons: sweet, I was hoping yourself and thomi had 5 minutes to hangout and discuss the datetime work and figure out what's left to do etc. | 20:51 |
veebers | thomi: would you have 5 minutes? | 20:51 |
thomi | sorry, I have back to back calls | 20:51 |
veebers | thomi: nw | 20:52 |
veebers | balloons: well, do you have 5 minutes to chat? | 20:52 |
balloons | veebers, sure I guess | 20:55 |
thomi | I have 3 minutes right noiw | 20:57 |
thomi | but maybe you don't need me to be in the hangout | 20:57 |
thomi | my requirements for this landing haven't changed since last time - I'd like an automated test suite that checks that the time in a Qml file is the same as the time in the proxy object, tested for ofour dates at various times of the year | 20:58 |
thomi | and, obviously, it must work in both automation, and on everyone's laptops | 20:58 |
balloons | thomi, veebers https://bugs.launchpad.net/ubuntu-calendar-app/+bug/1328600/comments/13 | 21:10 |
ubot5 | Launchpad bug 1328600 in Autopilot "Autopilot lacks support for large timestamps" [High,In progress] | 21:10 |
balloons | I added a comment that I hope lays out everyone's thoughts on this and what needs to happen. | 21:11 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!