| gary_poster | gmb, approved july 13-16 | 12:05 |
|---|---|---|
| gmb | gary_poster: Ah, thanks. | 12:05 |
| * benji returns from the coffee-in-a-tor. | 12:08 | |
| * bac tea, brb | 12:10 | |
| gary_poster | bac benji frankban gmb https://plus.google.com/hangouts/_/0fa5a03006b2f488f8c1c361642911ef26b87e68 | 12:10 |
| * bac owes benji a beverage of his choice | 12:33 | |
| benji | heh | 12:33 |
| benji | gmb: let me know when you're ready to pair | 13:53 |
| gmb | benji: Will do, just OTP ATM. Should be done soon. | 13:54 |
| benji | sounds good | 13:54 |
| gmb | benji, I'm in https://plus.google.com/hangouts/_/4d4738a1bd3c2d4fb104618ab69a5aeed54f6163?authuser=0&hl=en-GB# when you're ready | 14:17 |
| benji | gmb: cool, I'll be there in a sec | 14:17 |
| gmb | benji, and then it crashed | 14:24 |
| * gmb reloads | 14:24 | |
| benji | gmb: same here | 14:24 |
| bac | i figured out the unittest duplication problem. a testcase was defined and used in one module and then imported into another module and subclassed. the import caused it to be discovered again and re-run. i'm surprised this hasn't happened before. | 15:19 |
| bac | the fix was to make the shared parts a mixin that did not extend TestCase | 15:20 |
| gary_poster | huh, interesting. yeah | 15:20 |
| bac | seems obvious in retrospect | 15:20 |
| gary_poster | although...it would seem that the finding code could dedupe based on the module of a test case | 15:20 |
| gary_poster | maybe that's too much magic | 15:20 |
| bac | i guess most of our test subclassing happens within a given module | 15:20 |
| gary_poster | usually | 15:20 |
| gary_poster | yeah | 15:20 |
| bac | benji: you were kidding about the warhead part, no? | 15:21 |
| benji | bac: yeah :) | 15:21 |
| bac | was 95% sure | 15:21 |
| bac | that's the wackiest damn thing i've seen in launchpad | 15:22 |
| bac | never thought i'd read PBUH in a license | 15:22 |
| bac | benji: i'm making changes to zope.testing and am pretty close. do we need to coordinate or will you be a while? | 15:42 |
| bac | should i make p8 if i get there first and you'll do p9 | 15:42 |
| benji | bac: I have some tests running now in EC2, but it will likely be quite a while; I'll sync up with you before I get close to landing | 15:43 |
| benji | bac: yeah, the winner gets p8 | 15:43 |
| bac | ok | 15:43 |
| gary_poster | hard hangs :-( | 16:29 |
| frankban | benji: I think --shuffle-seed doesn't work because the initial order of tests changes depending of the options you use: with -t the order of initial list reflects how the tests are found in the suite; with --load-list the order of initial list is taken from the file | 16:48 |
| benji | this flaky internet connection is getting annoying | 16:50 |
| frankban | aargh, you were offline benji, I'll repeat | 16:51 |
| benji | if we end up buying a house here I'll have to think about getting two connections and setting up a fail-over router | 16:51 |
| frankban | benji: I think --shuffle-seed doesn't work because the initial order of tests changes depending of the options you use: with -t the order of initial list reflects how the tests are found in the suite; with --load-list the order of initial list is taken from the file | 16:51 |
| frankban | benji: a possible fix could be sorting the tests before shuffling them | 16:51 |
| frankban | heh | 16:54 |
| gary_poster | bac, you want a different merge target for your zope.testing branch | 17:09 |
| gary_poster | the MP is very unhappy | 17:09 |
| gary_poster | merge instead to...(looking) | 17:09 |
| gary_poster | bac, merge to lp:~launchpad/zope.testing/3.9.4-fork | 17:10 |
| bac | gary_poster: i thought that was where i targeted | 17:51 |
| bac | gary_poster: ah, if you use the chooser to pick the merge target it fails to change the radio button! | 17:52 |
| bac | gary_poster: https://code.launchpad.net/~bac/zope.testing/list-tests/+merge/109007 | 17:53 |
| * benji reboots | 18:09 | |
| gary_poster | bac approved | 18:13 |
| bac | thanks gary_poster. did you look at the other? https://code.launchpad.net/~bac/launchpad/bug-682772/+merge/108983 | 18:29 |
| gary_poster | no, looking | 18:30 |
| gary_poster | bac, why is this necessary? | 18:33 |
| gary_poster | 70- for key in sorted(special): | 18:33 |
| gary_poster | 171- special_suite = special[key] | 18:33 |
| gary_poster | 172+ for key, special_suite in sorted(special.items()): | 18:33 |
| gary_poster | sorry first line 170 | 18:33 |
| bac | easier to read | 18:33 |
| bac | LOC reduction. :) | 18:33 |
| === benji is now known as Guest68587 | ||
| gary_poster | bac, heh, ok. You are sorting on key rather than suite order (whatever that is) now too. | 18:35 |
| bac | yeah, the key is still the primary sort | 18:35 |
| gary_poster | cool | 18:35 |
| bac | i think sorted(dict) sorts by key | 18:36 |
| gary_poster | huh | 18:36 |
| gary_poster | oh sure | 18:37 |
| gary_poster | I was mis-thinking | 18:37 |
| gary_poster | nice change' | 18:37 |
| bac | yep: | 18:38 |
| bac | >>> d = dict(a=3,b=5,z=5,w=4,q=1) | 18:38 |
| bac | >>> sorted(d) | 18:38 |
| bac | ['a', 'b', 'q', 'w', 'z'] | 18:38 |
| gary_poster | yeah, did same | 18:38 |
| gary_poster | bac, why is id_extensions a list, rather that id_extension=None? Doesn't seem to be used anywhere as a multiple | 18:39 |
| bac | gary_poster: yeah, but paths could be | 18:45 |
| bac | i thought it nice to keep that option, even though no one was using it | 18:46 |
| benji | ok, I bought a wifi tethering app ($15 bucks!) so maybe I'll actually be online for a while | 18:46 |
| bac | i could've done the same 'string or list' trick i did with paths | 18:46 |
| bac | gary_poster: i'll make that change if you think it is a good idea | 18:46 |
| bac | benji: gary approved my change to the 3.9.4-fork branch, so i'm claiming p8 and merging | 18:47 |
| gary_poster | bac, can you imagine a scenario in which you would actually want to supply multiple id extensions, and this would do what you expect? You don't have to tell me what it is | 18:48 |
| benji | bac: cool; my zope.testing change is in rough waters again | 18:48 |
| bac | gary_poster: if someone called the factory with multiple paths then there would need to a 1:1 correspondence to id_extensions | 18:48 |
| gary_poster | ok | 18:48 |
| gary_poster | leave as is then bac. approving as is. thank you | 18:48 |
| bac | gary_poster: thanks! | 18:49 |
| gary_poster | done, bac | 18:50 |
| bac | ugh, our zope.testing for had 4 test failures. | 19:01 |
| bac | i added 10 more failures. doh. | 19:01 |
| gary_poster | :-/ | 19:11 |
| benji | aaaaand the teathering app doesn't work very well | 19:28 |
| gary_poster | I'm up to eight or so hangs in one day now :-( | 20:20 |
| benji | :( as irritating as it is, I'll take connectivity problems over machine hangs any day | 20:22 |
| gary_poster | yeah :-( | 20:22 |
| gary_poster | I'm a day or two away from trying to return this and buying another Mac TBH | 20:23 |
| bac | benji: you still around? | 22:57 |
| benji | bac: yep | 22:57 |
| bac | benji: my changes to 3.9.4-fork have introduced some new test failures due to the way the test names are being formatted. i've reverted my changes and will deal with them on monday. seems safer than mucking stuff up before disappearing for four days. | 22:58 |
| bac | benji: it is worth noting that our fork has 4 test failures as is. | 22:59 |
| benji | bac: makes sense | 22:59 |
| bac | benji: so you can have p8! :) | 22:59 |
| benji | My life is complete. | 22:59 |
| bac | it never got built | 22:59 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!