IRC log of #zope for Saturday, 2012-01-28

*** benji has quit IRC00:14
*** ccomb has quit IRC00:16
*** sunew has quit IRC00:36
*** J1m has quit IRC00:49
*** J1m has joined #zope00:51
*** J1m has quit IRC01:14
*** J1m has joined #zope01:17
*** J1m has quit IRC01:21
*** J1m has joined #zope01:25
*** supton_ has joined #zope01:46
*** supton has quit IRC01:50
*** supton_ has quit IRC01:51
*** _mup_ has quit IRC02:00
*** _mup_ has joined #zope02:00
*** J1m has quit IRC02:28
*** tiwula has quit IRC02:54
*** daMaestro has quit IRC02:58
*** tiwula has joined #zope03:35
*** J1m has joined #zope03:38
*** tiwula has quit IRC05:13
*** alga has quit IRC05:18
*** Spanktar has quit IRC05:58
*** ganesh_ariveguru has joined #zope06:40
ganesh_ariveguruhi....where i can learn zodb data retrieving concepts06:42
*** goschtl has joined #zope08:58
*** goschtl has quit IRC09:28
*** ganesh_ariveguru has quit IRC09:35
*** ganesh_ariveguru has joined #zope09:36
*** juh has joined #zope12:04
*** juh has quit IRC12:05
*** alexpilz has joined #zope12:22
*** mr_jolly has joined #zope12:35
*** agroszer has joined #zope12:42
*** alexpilz has quit IRC12:46
*** evilbungle has joined #zope13:18
*** evilbungle has quit IRC13:18
*** J1m has joined #zope13:24
*** m8 has joined #zope13:28
*** Wu has joined #zope13:56
*** eperez has joined #zope13:58
*** alga has joined #zope14:03
*** J1m has quit IRC14:10
*** thetet has joined #zope14:13
*** J1m has joined #zope14:59
*** thetet has quit IRC15:12
*** Wu has quit IRC15:41
*** TomBlockley has joined #zope15:43
*** TomBlockley has quit IRC15:53
*** juh has joined #zope15:53
*** agroszer has quit IRC17:03
*** J1m has quit IRC17:14
*** J1m has joined #zope17:15
*** TomBlockley has joined #zope17:32
*** J1m has quit IRC18:01
*** J1m has joined #zope18:03
*** Arfrever has quit IRC18:45
*** Arfrever has joined #zope18:54
*** juh has quit IRC18:56
J1mwtf, zope.testing claims Python 3 support, but it's tests don't even run w Python 3.19:02
*** tiwula has joined #zope19:14
koshJ1m: I have hit that with a few other things before also19:59
koshso far I see python 3 as a disaster that has horribly split the community20:00
koshwith many turning to other languages instead of upgrading20:00
J1mno kidding.20:00
koshin some ways I feel that zope should work towards pypy instead of zope 320:00
J1mdid you mean python 3?20:00
koshit would at least run faster, I think it has some kind of restricted access thing20:00
koshyeah20:01
koshzope should work towards pypy instead of python 320:01
koshit seems like it would give greater benefit20:01
J1mI don't care whether something supports Python 3 or not, but if it says it supports Python 3, it should actually support it.20:01
koshand much of the c could probably be redone in rpython to make it simpler and still keep the same speed20:01
J1mI wanted to make some small enhancements to zope.testing. Because it *claims* to support Python 3, I feel obliged not to break Python 3 support. But then the tests fricking don't run.20:03
J1mSo now, I'm tring to unscrew the tests.20:04
J1mtrying20:04
koshapparently the next version of python 3 adds more support for bytestring stuff back again with encodings so network stuff can be made to work again without insane hacks20:07
J1minsane hacks? like print_, like never using iterfoo, like always using sys.exc_info() to get exception values?20:09
J1moh well, I don't want to talk about it any more. It'll just make me cranky. :)20:09
J1mer20:09
koshthere is some probably in python 3 related to encoding of bytestrings20:17
koshthere is the one true belief in python 3 that unicode is everything20:17
koshbut network is not unicode and networking also has encodings like gzip20:18
koshor treating filesystems as all unicode when that does not actually work20:18
koshpython 2 allows bytestring to bytestring encodings20:18
koshbut something about that does not work right in python 320:18
koshto be honest I did not put a lot of effort into looking into it, it was just one more problem that I don't have time to deal with20:19
*** tiwula has quit IRC20:20
*** TomBlockley has quit IRC20:23
*** do3cc has joined #zope21:16
*** Arfrever has quit IRC21:19
*** Arfrever has joined #zope21:29
*** m8 has quit IRC21:43
*** J1m has quit IRC21:44
*** J1m has joined #zope21:49
*** m8 has joined #zope21:57
mgedminJ1m, I suspect that 'python3 setup.py test' might work with zope.testing22:30
mgedminif it uses 2to322:30
J1mIt didn22:31
mgedminah22:31
J1mIt didn't work for me.22:31
J1mMaybe I didn't invoke it right.22:31
mgedminI never tried22:31
mgedminbut I remember it was one of the first packages that people worked on to get it to run on Py3K22:31
J1mI'm guessing Lennart didn't either.22:32
mgedminmaybe22:32
J1mI suspect "run" is relative.22:32
mgedminit might have bitrotted since then22:32
J1mLike, maybe they got it working without getting it's tests to pass.22:32
mgedminwait, zope.testing or zope.testrunner?22:32
J1mzope.testing22:32
J1mThere's something about 2to3 in the setup script, but I don't think it's being run.22:33
J1mI've hacked it a fair bit today so the tests run under Python 3, both with the test runner and with setup.py test.22:34
J1mThis excludes tests for some modules that clearly haven't been ported.22:34
J1mI'll probably make a new release tomorrow.22:35
J1mI wanna make sure the some other changes work the way I expect them to first.22:35
*** davetoo has joined #zope22:44
davetooI'm having trouble finding the doc that explains exactly how zope.schema field constraints work.  I want to limit the length of a Text/TextLine field.22:46
davetoooh foo22:46
davetoon/mind it's simpler than that, but still would  be good to know22:46
davetoo.Text.max_loength just hit me in the face22:47
davetoolength22:47
*** davetoo has left #zope22:58
*** m8 has quit IRC22:58
*** ccomb has joined #zope23:17
*** fdrake has joined #zope23:27
fdrake(Resuming a conversation with J1m....)23:28
J1mfred objects to zope.testing.wait_until, because it has an _ in the name.23:28
fdrakewait_until is an ugly module name.23:28
* J1m objects to zope.testing.waituntil.wait_until23:29
fdrakeIf the module were zope.testing.wait, and the class Until, and the callable until, client code would be more readable.23:29
fdrakezope.testing.wait_until.wait_until is ... long.23:29
fdrake(And repetitive.)23:29
J1mgenerally, you'll: from zope.testing.wait_until import wait_until23:29
J1mor stick wait_until into test globs.23:30
J1mso it's usually just wait_until(cone)23:30
J1mor23:30
fdrakeNo I won't.23:30
fdrakeThat violates my good sense.23:30
J1m@wait_until23:30
J1mdef cond(): ...23:30
* fdrake doesn't like from...import any more than underscores.23:31
* J1m wonders if mgedmin is still around.23:31
J1many other opinions.....23:31
mgedminJ1m, I'm sort of busy having fun at a game jam (writing a MUD in Pyramid)23:32
J1mI can live with zope.testing.wait.until -- to make fdrake happy23:32
* fdrake likes being happy.23:32
J1mI'll at least remove the tag I just made. :)23:33
fdrakeCool; thanks!23:33
fdrakeThat'll give me time to proof your typing.   :-)23:33
* mgedmin is strictly against import q.w.e.r.t.y.u.i.o; q.w.e.r.t.y.u.i.o.p() style when the number of dots exceeds two -- zope.component.provideAdapter() is barely okay)23:33
J1mIn test code, I'd end up doing:23:34
J1mwait.until(cond)23:34
fdrakemgedmin: That's why we should use short, readable names for things.23:34
J1mwhich I find a bit hinky23:34
mgedminshort, readable, *unique* names what we can bravely from somewhere import :)23:35
* fdrake suspects mgedmin of be an as advocate.23:35
J1mIn normal code, I like using dots so I don't have to think hard about where things are defined.23:35
J1mIn test code, I still often prefer dots for application code, but when using a test helper, I don't want a lot of clutter in the helper.23:36
J1mI wonder of there's a single word that says: "wait for a condition to hold"23:38
J1mmaybe just wait23:38
fdrakeThat would be it.23:39
J1muntil is sort of redundant.23:39
fdrakeThough I'm still not convinced that twowords shouldn't be considered a single word for the purpose of creating function names.23:39
J1mYou must be German.23:41
fdrake:-)23:41
J1mI'm still not convinced that modules shouldn't have _s.23:42
fdrakeI'm ok with you not being convinced.23:42
fdrakeThat wasn't my goal.23:42
J1mBut really, until is redundant.23:43
fdrakezope.testing.wait() (a function) would be fine with me as well23:43
J1mso I think I'll go w just wait.  But I'll wait until tomorrow to decide for sure.  :)23:43
fdrakeHeh.  :-)23:44
* mgedmin finally understands what fdrake meant with that advocacy statement!23:44
mgedminthere's no evil worse than 'from z3c.form import form as baseform'23:44
J1mI don't like it when modules write into package namespaces.23:44
mgedminor just 'from z3c.formui import form' in some modules mixed with 'from z3c.form import form' in other modules23:44
fdrakeJ1m: Hmm?  Why not include the class & callable in zope.testing/__init__.py?  No one clobbering other folks' namespaces.23:46
* mgedmin recently acquired a distaste for code in __init__.py23:46
* fdrake has become surprisingly accepting of code in __init__.py.23:46
J1mThat seems a bit presumptuous .23:46
fdrakeHa!23:46
mgedmin(after realizing that I pay the cost of importing useless stuff like lxml because some helpers in the parent package use it)23:47
fdrakeReally?  For a package not to control it's namespace?23:47
fdrakeI wouldn't want to poke things into *other* packages, or into namespace packages.23:47
J1mI wish Python would let me give modules special methos, like __call__.23:48
J1mmethods23:48
J1mThen I could have zope.testing.wait be both the module and the callable.23:48
fdrakeYou're not alone there.  Even I'd prefer that.23:49
J1mMaybe in Python 4.23:49
fdrakeAn J1m knows I don't prefer much.23:49
fdrake*And23:49
J1mI really regret what I did with ZODB.DB (the class replaces the module) and what Jeremy did with ZODB.FileStorage.FileStorage (likewise)23:51
J1mThe Wait(/WaitUntil) class is important at least for the case of gevent and similar things where you need to use a special sleep, so the class is an important part of the API.23:53
* J1m doesn't expect anyone to say anything. :)23:54
fdrakeJ1m: Nobody is expecting you to have made totally unreasonable decisions regarding what the code does.23:55
J1mMy point was just that the API has 2 bits, so the module is important too.23:56
J1muntil both wait and Wait were jammed into zope.testing, which doesn't feel right at this point.23:57
fdrakeSounds like the class is important, but the module is 'cause we're tend to put things in them.23:57
J1ms/until/unless23:57
J1mYes, that's what modules are for.23:58
J1mAlso, zope.testing isn't a module, it's a collection of modules.23:59
fdrakeIndeed.  And the package module works just fine for me.  :-)23:59
fdrakeYeah, well.23:59
*** dayne has joined #zope23:59
J1mzope.testing is more like a namespace module than a regular module.23:59

Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!