[10:15] just when you think you're done with the firewall things... it strikes back! [10:15] just spent 10 minutes wondering why my bzr push wasn't working... [10:26] tomwardill: was trying to think of something constructive to reply to that... came up with nothing :) I was pushing to the wrong repo yesterday, spent an hour wondering why things are not working as they should :D [10:29] computers: how do they work!? [10:29] They don't. [10:29] ... a valid point [10:31] I always drop in for the valuable contributions. [10:31] <3 [10:31] LOL [11:26] cjwatson: I'm getting the Mir thing on running tests now too [11:30] I haven't had time to debug it, so if you feel the urge then please do [11:30] It doesn't happen in the wrapper I normally use, only when I ssh into the container and run tests directly with bin/test [11:31] yeah, same [14:19] https://code.launchpad.net/~cjwatson/launchpad/+git/launchpad/+merge/378658 is another py3 dependency upgrade [14:19] full test suite passed [14:26] do we feel strongly about names for the migration endpoint ? /migrate or /inject ? [14:26] what are we leaning towards ? [14:27] ilasc: inject seems more explicit; I'd go with that [14:28] migrate leaves the direction unclear [14:28] cjwatson: excellent, thanks! agreed [14:28] +1 for inject. We could need this endpoint for other purposes, other than migration [14:28] and forgot to say in Standup: thank you for taking on the webhooks QA and test issue! [14:29] cool, we're all in agreement, thanks guys! [14:30] I was feeling slightly guilty about stealing that from you, but you did say you wanted to get onto lp-signing work [14:30] And as it happened writing target-generic tests for this turns out to be weirdly difficult [14:30] Something very odd going on in the innards of the navigation menu system [14:35] hmmm.... definitely no need to feel guilty, that would've been a 3 months time sync for me by the sound of it, very glad I didn't step on that particular landmine at this point - really appreciate you looking at it, you're right I did want to get to lp-signing as soon as possible [14:37] ok :) [14:53] ilasc: https://code.launchpad.net/~cjwatson/launchpad/+git/launchpad/+merge/378660 [14:54] wow! that was fast! looking now [14:54] Well, I was working on it most of the morning :) [14:55] Agreed re inject, fwiw [14:55] Including lots of "why on earth is the navigation menu not showing up at all?" [14:56] SpecialK|Canon: +1 [14:57] cjwatson: sounds like u had my kind of morning, however the first part did make laugh as in my world it would translate to "well I was working on it most of this spring" :) [14:58] but I see what you mean on the navigation topic [15:02] cjwatson: the rename to 'oci_project' in the DB patch, are you intending to backport that fix to 'OCIProjectSeries'? [15:03] tomwardill: Renaming DB columns is kind of unreasonably painful. I'd suggest it might be a good idea to paper over it at the app level though (i.e. oci_project_id = Int(name='ociproject', allow_none=False); oci_project = Reference(oci_project_id, "OCIProject.id")) [15:03] tomwardill: Sorry, I should have caught it before [15:04] well, the rename wasn't a thing before this patch :) [15:04] Yeah, but William is right that we've preferred this style in most new things for a while [15:25] Ran 48 tests with 0 failures, 0 errors, 0 skipped in 7.600 seconds [15:25] now to re-add all the tests I've taken out [15:32] Ugh, I really hate system-site-packages, so confusing [15:32] indeed [15:32] I've had a fair few disagreements with that [15:38] I'd love to stop using it, but python-apt, mostly [15:44] https://github.com/sivel/python-apt-wheel not quite the shape of thing I'd ideally like, but I wonder if it'd be worthwhile [15:51] that seems to be a lot of technology involved [15:53] Right, ideally we'd just unpack the .deb and slam it into a .whl [15:54] (We have a few other system Python dependencies too; it's never been worth solving all of them without a strategy for python-apt) [15:55] Also would need to think about how it interacts with OS upgrades [15:58] cjwatson: given the name uniqueness constraint on OCIRecipe, where is best to check for existing recipes with the same name on creation? the .new method of OCIRecipeSet? [15:59] tomwardill: I think that's our usual strategy, ues [15:59] *yes [16:00] Compare SnapSet.new [16:01] cool, thought so [16:32] I've lost count of how many times I've been bitten by MoinMoin syntax not being MD/rST [16:33] I only have enough room in my brain for a small number of markup languages [17:11] tomwardill: Could you have a look at https://code.launchpad.net/~cjwatson/launchpad/+git/launchpad/+merge/378670 ? That's my system-site-packages nightmare for today [17:11] on it [17:11] And addresses the reason that qastaging isn't updating [17:12] Also staging in fact [17:12] dogfood was OK because it updates in place so had a cached wheel [17:12] But production would run into the same thing [17:13] well, it's kind of ergh-ish, but all fixes for this type of thing are :) [17:14] +1, lgtm, glad I didn't have to write it. [17:17] I would feel exactly that way if I hadn't had to write it [17:17] Thanks, and agreed [17:39] cjwatson: Storm is giving me "RuntimeError: Property used in an unknown class" [17:40] query looks like this: https://pastebin.canonical.com/p/tKWcSKWNTs/ in OCIRecipeSet [17:42] tomwardill: Can I see the OCIRecipe model? [17:42] Query itself looks OK [17:42] cjwatson: https://pastebin.canonical.com/p/kGHKKWdc2p/ [17:42] oh wait [17:42] that's the interface [17:43] sec [17:43] I was going to say [17:43] cjwatson: https://pastebin.canonical.com/p/mV2nHh8Wvh/ slightly better [17:46] tomwardill: That looks OK to me. Um. What's the traceback? [17:46] tomwardill: Or maybe push WIP to a temporary branch and I can try it [17:47] sure, one sec [17:48] cjwatson: https://code.launchpad.net/~twom/launchpad/+git/launchpad/+ref/runtime-errors [17:48] test is `lp.oci.tests.test_ocirecipe.TestOCIRecipeSet.test_new` [17:50] OK, give me a bit to build that [17:55] cjwatson: I've got it [17:55] Oh? [17:55] passing 'owner' to self.exists, rather than 'oci_project' [17:56] stupid copy and paste error [17:56] (found it by debugging at the point of raise and wondering why it was searching the Person class) [17:56] Ah right, wrong caller [17:56] Got it [18:03] pappacena: Audit trail stuff looks good on dogfood. The only comment I have is that the colspan="2" thing doesn't look quite right on all queues. On the Rejected queue the audit trail appears under the file list, which is reasonable enough. But on the Done queue there's no checkbox column and so it appears under the version instead, which looks odd [18:03] pappacena: Example of the latter is https://dogfood.paddev.net/ubuntu/xenial/+queue?queue_state=3&queue_text=httmock [18:04] pappacena: I think you need something conditional on view/availableActions there (see a bit further up in lib/lp/soyuz/templates/distroseries-queue.pt) [18:04] uhm... yes, it doesn't look alright this way. [18:04] Thanks for showing. I will try to reproduce locally and open a MP later today [18:04] It's deployable like this, just a little odd-looking [18:05] Thanks for this feature, I bet it will make some people very happy [18:05] I hope so! :-) [18:05] I'll work on a fix today... let me just finish something I'm in the middle here with the signing service... [18:15] Thanks [18:15] I've done most of the QA, but I need to wait for my fix-clean-build patch to finish landing so that (qa)staging come back up before I can get any further [18:16] So possibly a deployment tomorrow late morning ish, which is always nice for the snap store weekly report [18:19] Cool! I'll try to have it ready to review by the end of the day today [19:31] cjwatson: https://code.launchpad.net/~pappacena/launchpad/+git/launchpad/+merge/378679 [19:31] The fix for the extra colspan when there is no action [20:20] Thanks, looking [20:22] pappacena: one query [20:22] (in a comment there) [20:35] Sure. I will check in a minute [20:51] Replied there... I needed to double check and test something about your comment. [21:05] pappacena: of course you're right - sorry for my confusion. Go ahead and land at your convenience [21:06] No problem! I got a bit confused too, and had to check to make sure. ;-) [21:06] I'll top-approve now