| veebers | thomi: hey, I know that you've changed teams already, but can you do me a favor? Can you review/approve this MP so that I can unblock the current landing for AP? | 00:43 |
|---|---|---|
| thomi | sure, link me? | 00:43 |
| veebers | oops | 00:44 |
| veebers | https://code.launchpad.net/~veebers/autopilot/revert-ext-class-fix-for-now/+merge/251025 | 00:44 |
| thomi | the name makes me sad already | 00:44 |
| veebers | The current plan is to revert this, get the current stuff landed and getting the actual fix for this issue is on the backlog of the projects team up (so I hope to get that on the books next week) | 00:44 |
| veebers | thomi: aye, I knew that it would, I'm not happy about it either | 00:44 |
| thomi | so, you're going to release with the CPO extension class bug ? | 00:45 |
| veebers | thomi: right, the current fix doesn't actually ffix it in the wild | 00:46 |
| veebers | i.e. with the fix a _bunch_ of tests fail, without the fix they pass | 00:46 |
| veebers | then next week we'll be working on getting the proper fix in plae | 00:46 |
| veebers | place* | 00:46 |
| thomi | so it's not a fix, my point is, if you do that you're releasing a known bad autopilot | 00:46 |
| veebers | thomi: right, but it's an issue that already exists in autopilot | 00:47 |
| thomi | ok | 00:47 |
| veebers | we're releasing an autopilot that doesn't have a fix for that _yet_ | 00:47 |
| veebers | it's not the optimal position to be in, but I didn't realise that it was a bigger issue until this sprint was already underway | 00:48 |
| thomi | so why release at all? | 00:48 |
| thomi | (approved, BTW) | 00:48 |
| veebers | thomi: cool thanks. | 00:49 |
| veebers | thomi: To release the doc improvements (incl the enablement of the json for doc publishing) now | 00:49 |
| thomi | oh, is that work done? | 00:50 |
| veebers | heh, enablement :-P I like making words up | 00:50 |
| veebers | thomi: For the publishing? Not complete but needs these changes to go forward | 00:50 |
| veebers | thomi: see, you've only just left and the quality of the autopilot code is already going downhill ;-) | 00:53 |
| thomi | lol | 00:54 |
| thomi | anyway, that's inaccurate, it's just not going uphill :D | 00:54 |
| veebers | ^_^ | 00:57 |
| === MacSlow is now known as MacSlow|lunch | ||
| === MacSlow|lunch is now known as MacSlow | ||
| MacSlow | rhuddie, elopio: how can I make sure an AP-test is run against a locally compiled unity8 from within the source-tree? | 14:31 |
| rhuddie | MacSlow, I'm not entirely sure, but you might want to look at: http://bazaar.launchpad.net/~unity-team/unity8/trunk/view/head:/tests/autopilot/unity8/__init__.py#L75 | 14:36 |
| rhuddie | seems to look in a particular path for the unity8 binary in the source tree | 14:37 |
| MacSlow | rhuddie, if that's the only way, I'll use that | 14:37 |
| elopio | MacSlow: I think you have a make task in your project. | 14:44 |
| elopio | like make-autopilot. It's documented on the README or CODING file. | 14:45 |
| MacSlow | elopio, I know these bits of info... but I still don't seem to get the local unity8 being executed | 14:47 |
| MacSlow | elopio, that's the reason I was asking | 14:48 |
| elopio | MacSlow: I almost never change things in unity8, last time I did it was months ago. mzaanneti or saviq should know. | 14:56 |
| MacSlow | elopio, ok | 14:56 |
| elopio | there is a way, and I think it's to run the build script, get into the build dir, and make autopilot in there. | 14:56 |
| elopio | but I'm not really sure about the steps. | 14:56 |
| Saviq | MacSlow, you need "make install" | 14:57 |
| MacSlow | Saviq, I did | 14:57 |
| Saviq | MacSlow, then PYTHONPATH=tests/autopilot autopilot run foo | 14:57 |
| Saviq | MacSlow, should run the tests on the local build | 14:57 |
| MacSlow | Saviq, that's what I use | 14:57 |
| MacSlow | Saviq, but when I let it dump the qml-item tree the source: line for the shell is Shell.qml where it should be OrientedShell.qml | 14:58 |
| Saviq | MacSlow, what makes you say that it runs the wrong unity8 though? | 14:59 |
| Saviq | MacSlow, -vv will be verbose and say what is launched etc. | 14:59 |
| MacSlow | Saviq, no wonder I've problems trying to grab objectName="shell" | 14:59 |
| MacSlow | Saviq, -vv not for autopilot?! | 15:01 |
| Saviq | MacSlow, yes, for autopilot | 15:01 |
| MacSlow | Saviq, as a parameter for run?! | 15:02 |
| Saviq | MacSlow, autopilot -vv run foo | 15:02 |
| MacSlow | Saviq, autopilot and autopilot3 only spit out version information with this | 15:04 |
| Saviq | MacSlow, right, sorry, yeah, autopilot run -vv foo | 15:04 |
| MacSlow | Saviq, tried that in the meantime too :) | 15:05 |
| Saviq | MacSlow, and what makes you think it's trying to run the wrong binary? | 15:06 |
| MacSlow | Saviq, the fact that there's no single qml-item with objectName="shell", which is the one used for the OrientedShell.qml | 15:08 |
| Saviq | MacSlow, that doesn't sound like proof enough, what does the autopilot log say about which BINARY is used? | 15:09 |
| MacSlow | Saviq, which would be written where? | 15:10 |
| MacSlow | Saviq, I only know of the stuff in ~/.cache | 15:11 |
| Saviq | MacSlow, "Starting job unity8 with arguments..." | 15:11 |
| Saviq | MacSlow, on your console | 15:11 |
| Saviq | MacSlow, it's the output from autopilot run -v | 15:11 |
| MacSlow | Saviq, I don't have that in the output of autopilot (tried 1.4.0 and 1.5.0) | 15:18 |
| Saviq | MacSlow, pastebin your whole console output please | 15:18 |
| Saviq | with the command you used | 15:18 |
| MacSlow | Saviq, http://pastebin.ubuntu.com/10430561 | 15:22 |
| Saviq | MacSlow, and if you run any of the existing tests? I tried unity8.shell.tests.test_lock_screen.TestLockscreen.test_can_unlock_pin_screen | 15:23 |
| Saviq | MacSlow, most likely you're not using the base testcase that actually takes care of setting the right binary | 15:23 |
| * MacSlow tries... | 15:23 | |
| Saviq | MacSlow, tests/autopilot/unity8/shell/tests/__init__.py | 15:24 |
| Saviq | you want to inherit from tests.UnityTestCase | 15:24 |
| MacSlow | Saviq, doing that | 15:25 |
| MacSlow | Saviq, the issue seems to be the the args passed to restart_unity() from process_helpers.py | 15:41 |
| Saviq | MacSlow, is it working for the previous tests? | 15:42 |
| Saviq | like do you get the correct BINARY mentioned in the log if you're running one of the previous tests? | 15:42 |
| MacSlow | Saviq, yeah... for e.g. one of thre greeter-tests it spits out a very long option-list from process_helpers.py which has the "BINARY=..." bit in it | 15:43 |
| MacSlow | Saviq, and i nmy case it's only 'QT_LOAD_TESTABILITY=1', and nothing else | 15:43 |
| Saviq | MacSlow, so look at the other test and see what it's doing differently than you | 15:45 |
| MacSlow | Saviq, yup... fyi http://pastebin.ubuntu.com/10430933/ | 15:46 |
| veebers | thomi: are you keen to approve another ap mp for me? I need to find a replacement :-) | 21:27 |
| thomi | veebers: maybe, but you really need to start asking people in QA | 21:28 |
| veebers | thomi: fair enough, I'll do that now | 21:28 |
| thomi | veebers: maybe save me for when others don't have the necessary experience to reviwe the code? | 21:28 |
| thomi | veebers: like, introspection internals, for example | 21:28 |
| veebers | nuclearbob, alesage, elopio: Could I get a review please? Just an updated changelog for the current release (so we get something sensible out of it) | 21:28 |
| veebers | thomi: sure, makes sense to me. Sorry to bother you, it's muscle memory :-) | 21:29 |
| thomi | heh, no worries. | 21:29 |
| thomi | veebers: want to review my launchpad branch? :P | 21:29 |
| veebers | lol, point taken. But I would take a shufti just out of interest | 21:30 |
| thomi | will let you know when the diff is generated | 21:30 |
| nuclearbob | veebers: the 1.5.0+15.04.20141031-0ubuntu1 looks a little funny to me | 21:30 |
| veebers | sweet | 21:30 |
| veebers | nuclearbob: oh, so that'll be coming from 1.5 when I merged it back into trunk | 21:31 |
| nuclearbob | veebers: the same changelog is under five people, and elopio is named twice | 21:31 |
| nuclearbob | veebers: ah, okay | 21:31 |
| veebers | nuclearbob: to clarify I merged the changelog from 1.5 back into trunk then added the topmost entry | 21:31 |
| * veebers updated MP desc | 21:31 | |
| nuclearbob | veebers: ah, okay. Sounds reasonable to me, then | 21:31 |
| veebers | nuclearbob: if you're cool with that do you mind approving the mp? :-) | 21:36 |
| nuclearbob | veebers: done | 21:37 |
| veebers | sweet ,cheers | 21:37 |
| thomi | veebers: https://code.launchpad.net/~thomir/launchpad/devel-fix-editemails-link/+merge/251176 | 21:44 |
| veebers | thomi: very cool. Ah, what's it like writing non-pure python3 code again? | 21:51 |
| thomi | veebers: it's actually not too bad | 21:51 |
| thomi | veebers: but zope is mental | 21:51 |
| veebers | ^_^ | 22:03 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!