PMXBOT Log file Viewer

Help | Karma | Search:

#pypa-dev logs for Monday the 16th of November, 2015

(Back to #pypa-dev overview) (Back to channel listing) (Animate logs)
[04:12:30] <lifeless> dstufft: https://github.com/pypa/interoperability-peps/pull/56/files#r44730474 ready for final review I think
[05:36:31] <abadger1999> lifeless: Thanks. I wonder if what I need (for unittest2 on py3.5) is just to update unitest2/tests/test_loader.py with changes in python-3.5
[05:36:39] <abadger1999> will have to try that later.
[06:07:16] <lifeless> abadger1999: it should just work on python 3.5
[06:08:00] <abadger1999> lifeless: it doesn't. Or at least, the tests don't pass... which could be the tests' fault rather than the actual module's fault.
[07:02:44] <abadger1999> lifeless: Okay, it seems to be confirmed to me that the tests for unittest2 need to be updated to take into account a new format for traceback messages in python-3.5.
[07:04:57] <abadger1999> lifeless: I just did a naive substitution of lines from python-3.5's test_loader.py into unittest2's test/test_loader.py and changed "unittest" to "unittest2" and the test works on python-3.5... But of course, it then fails on 2.7 (probably on 3.4 as well). I'll have to adapt this a little more intelligently to make it work in both places.
[07:33:55] <lifeless> abadger1999: so
[07:34:12] <lifeless> abadger1999: if it is behaving differently on 2.7/3.4/3.5, do you have *any* other patches to unittest2?
[07:34:30] <lifeless> abadger1999: e.g. are you perhaps patching out the use of traceback2, which should be taking care of those differences
[07:35:41] <abadger1999> lifeless: nope. there's apparently a patch that gets rid of the argparse dep.
[07:35:58] <lifeless> abadger1999: upstream uses markers to make that only apply on 2.6
[07:36:03] <lifeless> abadger1999: what release are we talking about ?
[07:36:53] <abadger1999> lifeless: there was a patch that got rid of traceback2 supposedly for bootstrapping on a new python version. Once I changed that so the patch was no longer applied, the unittests started failing.
[07:37:26] <abadger1999> lifeless: so it looks like the unittests work when traceback2 is not being used.
[07:37:36] <lifeless> hmm
[07:37:36] <abadger1999> lifeless: version == 1.1.0
[07:38:07] <abadger1999> lifeless: but I haven't looked into the code in the patch, just the comments on when it is applied.
[07:38:35] <lifeless> abadger1999: ok; so I've just reproduced the failure here
[07:38:40] <lifeless> abadger1999: I'll poke at it this week
[07:38:47] <abadger1999> lifeless: Cool. Thanks.
[07:43:26] <lifeless> thank you for raising it :)
[20:26:32] <xafer> https://github.com/pypa/pip/pull/3173 still looking for review :)
[20:35:05] <lifeless> dstufft: morning
[20:35:41] <dstufft> lifeless: hello
[20:46:50] <lifeless> dstufft: so afaict the dependency specifier stuff is done
[20:47:10] <lifeless> dstufft: where is the process at in your opinion ?
[20:48:47] <dstufft> lifeless: I need to re-review it and then pronounce on it. Has your latest updated been posted to distutils-sig? If not it might be sane to start a final thread with it to let anyone else get one last look at it and I'll try to either get it reviewed tonight or tomorrow
[20:49:36] <dstufft> my truck broke down last night so at some point I'll have to run to the mechanic to pick it up (hopefully... unless it died completely D:) and I'm knocking some stuff out in warehouse ATM, but I think I can hit it tonight
[20:50:04] <lifeless> dstufft: I posted the diffs in the thread
[20:50:11] <dstufft> okay
[20:50:14] <lifeless> dstufft: would you like me to post a full final copy ?
[20:50:24] <dstufft> yea probably should
[20:50:46] <dstufft> I don't suspect anyone is going to have much to say, the fiddly bits most won't care about, and it's largely just standardizing wht already exists
[20:50:51] <dstufft> but should give them a chance anyways
[20:51:28] <lifeless> done
[20:51:47] <lifeless> I'll revise the bootstrap / abstract build interface pep today
[22:43:02] <lifeless> dstufft: btw
[22:43:21] <lifeless> dstufft: I don't know if you saw, but the tweak I did to identifier in the parsley grammar should work for pyparsing too
[22:47:34] <dstufft> lifeless: ok great
[22:47:59] <lifeless> dstufft: no lookahead or regex needed