IRC log of #zope3-dev for Monday, 2006-08-21

*** povbot has joined #zope3-dev16:23
*** J1m has joined #zope3-dev16:34
*** baijum has joined #zope3-dev16:36
*** whit_ has quit IRC16:37
*** whit_ has joined #zope3-dev16:39
*** baijum has quit IRC16:56
*** zagy has quit IRC17:03
*** baijum has joined #zope3-dev17:04
* baijum just started
*** alecm has joined #zope3-dev17:04
einsbaijum wouldn't it be better to update
baijumeins: planning to borrow content from there too..17:07
baijumwill ask srichter first17:07
baijumsrichter, ayt?17:08
baijumsrichter: can I use Zope3Book as a base for this work17:09
mgedminthere's also benji's quickstart with a similar scope17:09
philiKONbaijum, sure you can17:09
philiKONbaijum, it's CC17:09
philiKON(creative commons)17:09
baijumbut no derivative works17:09
benjibaijum: you might want to check the trademark status of "Foo Distilled" as a book title17:10
baijumbenji: better I will change name :)17:10
* baijum leaving17:12
*** eins has quit IRC17:13
baijumbenji: I have used few text from your quick start guide17:14
benjithat's fine, the license allows that (with attribution)17:14
baijumbut Zope3Book and benji's QuickStart guide are different licences?17:15
baijumboth are CC17:16
philiKONthere are different CC licenses17:16
*** zagy has joined #zope3-dev17:16
baijumso I have to get special permission from srichter to use that work17:17
*** jinty has quit IRC17:22
*** baijum has quit IRC17:36
*** hdima has quit IRC17:40
*** romanofski has quit IRC17:42
*** romanofski has joined #zope3-dev17:43
*** jinty has joined #zope3-dev17:43
*** elro has joined #zope3-dev17:47
SmokeyDhi all. What is the best schema field which can contain one value, chosen from several values, but where I can define the actual values to choose from, somwhere in the code and not in the interface/schema?17:48
mgedminif you use a vocabulary factory, you can generate the list of choices in Python code17:49
philiKONSmokeyD, got my book:?17:49
mgedminand then register an IVocabularyFactory utility named 'Foo'17:49
philiKONSmokeyD, read the "Vocabularies" chapter17:49
mgedminthere's also this concept called 'sources' which, I think, is supposed to replace vocabularies17:49
SmokeyDok, thanks17:49
philiKONchapter 16 i think17:49
mgedminbut I know little about them17:49
SmokeyDoverlooked it17:49
guruzWebMaven: well anyway, contact me if you run one again17:50
philiKONSmokeyD, read it thoroughly17:50
philiKONit explains everything except sources17:50
SmokeyDok, thanks. I didn't realise vocabularies were what I was looking for :) Thanks17:50
philiKONyou should've read the chapter, then you would have known ;)17:51
SmokeyDyou're right17:51
*** alecm has quit IRC17:54
*** dobee has quit IRC18:00
*** roym has joined #zope3-dev18:02
jintyI've been trying to comment on a issue in the collector, but seem to have "insufficient prviliges". can anyone help?18:11
d2mjinty: you need to login18:12
jintyd2m: I was under the impression I was logged in...18:12
jintyanyway, I'll try again18:13
jintyd2m: nope, definitely logged in and still getting insufficient priviliges18:16
philiKONyou can always ask J1m to make you a manager or supporter or hwatever it's called :)18:18
J1mI dunno, maybe you have to be a supporter to comment on an issue you didn't submit.18:19
d2mjinty: policy is "You need to be logged into your account to add issues and comment on existing issues18:21
d2mjinty: which issue nuber ?18:21
jintythe account i'm logged into is "jinty" for obvious reasons;)18:23
d2mi see you loged in, but no error in error log18:23
d2mjinty: ok, now its there, you cannot add a file in a comment18:24
d2mYour user account does not have the required permission. Access to 'addFile' of (__FactoryDispatcher__ instance at 1d907cc8) denied. Your user account, jinty, exists at /acl_users. Access requires one of the following roles: ['Manager', 'Owner']. Your roles in this context are ['Anonymous', 'Authenticated', 'Member'].18:24
jintyok, then i'll guess I'll just inline the patch18:25
philiKONjinty, btw, just read your pypy thing18:25
philiKONway cool]18:25
jintyphiliKON: thanks, I wasn't sure I should send it to the list...18:26
philiKONsure why not18:26
philiKONwas that after EP?18:26
jintyyeah, there was no Zope sprint, so PyPy sounded much more interesting18:27
jintynot sure where the line is between distributing interesting infromation and self promotion is...18:27
philiKONthis is definitely interesting info :)18:28
* benji is very excited about pypy18:28
philiKONjinty, did you pair with anyone from the pypy guys?18:29
jintyyeah, but it was more asking a huge mountain of questions;)18:30
philiKONwho did you pair with?18:31
jintyRestricted Python is a strange world18:31
philiKONit's basically the sewers underneath real python :)18:31
jintyAnders Chrigström18:31
philiKONhmm, dunno him18:31
jintyIt's like looking through a mirror18:32
jintykeeping straight what's a real python object and what's restricted python is a real task18:32
philiKONhence the hungarian notation18:33
jintybut I'm looking forward to when I can actually do something usable in RPython18:34
jintybut I think It'll take a long time18:34
philiKONtill then we can use pyrex ;)18:35
faassenjinty: why do you think it'll take a long time? i.e. what improvements need to be made?18:36
jintyfaassen: Because how do you begin to distribute a pypy extension module?18:37
faassenjinty: but I thought the idea was that you could build cpython extension modules, I thought you were talking about that.18:37
*** alecm has joined #zope3-dev18:37
jintyyes, it is compiled to a cpython module18:37
faassenjinty: you mean distribute the source?18:37
jintybut if you want to distribute it as a part of zope, thats difficult18:38
faassenjinty: (I'm not debating you, I'm just asking so I know what to tell the PyPy people what I want :)18:38
philiKONjinty, so, you can't just take the C source that pypy is spitting out and drop that into zope?18:38
faassenphiliKON: well, that wouldn't be maintainable.18:38
faassenphiliKON: the whole point over rpython + ctypes over pyrex would be that ideally you could run the same as Python code.18:38
jintyphiliKON: yep18:38
faassenphiliKON: which then hopefully would lead to increased maintainability.18:39
philiKONyou could pdb it18:39
jintyalso the incantations you have to do to connect the python and RPython world are far too magical and confusing18:39
faassenyeah, instead of maintaining a C implementation and a python implementation which are guaranteed to go out of sync.18:39
faassenyou'd just have a single implementation.18:40
philiKONbut you'd need pypy as a required interpreter18:40
faassenjinty: right.. I hope you give 'm plenty of feedback on how to improve things for you.18:40
philiKONjinty, perhaps helper (meta)classes will help there18:40
jintyphiliKON: no, just in the build process18:40
philiKONright. that's what i meant initially18:40
faassenphiliKON: no, the nice thing about the extension system of PyPy is that you can build CPython extensions, so just CPython would be sufficient. unless you're a developer and you want to maintain the modules, perhaps you mean that. :)18:41
philiKONbut if you wanted to use the python version, you'd need pypy18:41
philiKONwhat faassen says (the last part)18:41
philiKONso, we meant the same thing all along18:41
faassenphiliKON: no, the Python version should be Python + CTypes, which is core Python 2.518:41
faassenphiliKON: ideally.18:41
faassenphiliKON: so you can choose to run it as pure Python + CTypes, or rebuild it using RPython & RCTypes into a C module.18:42
jintyphiliKON: perhaps by organizing the code correctly the python version can be used in CPython unchanged18:42
faassenjinty: right, that would be the nice idea.18:42
philiKONhow's ctypes helping me to run pypy-dependent code?18:42
*** srichter has quit IRC18:42
faassenit's not pypy dependent code.18:42
faassenI mean, the idea would be...18:42
faassenI mean, whether this is reality is another question, but this is what they were promoting.18:42
faassenthat you'd write PYthon code using Ctypes.18:42
faassenthat happens to be RPython.18:43
faassenso follows some restrictions.18:43
*** _projekt01 has left #zope3-dev18:43
faassenand then you can just run it in plain python.18:43
faassenhowever if you want the performance, you'd transform it into a C module using the RPython + RCTypes stack with PyPy.18:43
philiKONclearly i need to read on ctypes again18:44
faassenso that means that instead of lxml in Pyrex, you'd have lxml in a restricted Python using CTypes that just runs.18:44
faassenin plain python.18:44
faassenbut I could also transform the whole thing into a C module.18:44
faassenusing the PyPy stack.18:44
philiKONi thought ctypes was all about calling foreign code in C from python18:44
jintythough CTypes or RCTypes are not a required part of the mix unless you want to access C extensions18:44
faassenoh, yeah, sorry, I was also pulling that in.18:44
faassensorry, jinty's right.18:44
philiKONah ook18:44
faassenso if you don't do that, you don't need to worry about it.18:44
philiKONso, you would just code zope.interface in Python18:44
faassenif you just want it for the performance.18:44
philiKONthat happens to qualify for rpython18:45
*** dobee has joined #zope3-dev18:45
philiKONright. i still see pypy dependencies there18:45
philiKON(in jinty's code)18:45
philiKONi presume those would have to go away18:45
philiKONthen it'd be runnable in cpython18:45
faassenyeah, I don't think they're there yet.18:45
philiKONand, for performance, compilable to C18:45
faassenjinty, you should give loud feedback about your experiences to the PyPy crew, from the perspective of an extension writer.18:45
jintyyeah, really I tried to keep it separate, but it was just too much for 2 days...18:45
philiKONi saw the separation18:46
philiKONmakes sense18:46
philiKONi also see faassen's idea about lxml18:46
philiKONsounds like a plan for lxml 2.0 ;)18:46
faassenI'm waiting for the pypy toolset to mature a lot more first. :)18:46
faassenthat's why I'm pushing Jinty to give them feedback. :)18:46
* jinty was complaining loudly during the sprint18:46
faassenanyway, lxml 2.0 would then be able to run in CPython, as a CPython extension.18:47
faassenjinty: excellent! :)18:47
faassenand as a PyPy extension.18:47
faassenso theoretically even on the JVM, .NET, Squeak platform, Lisp, etc. :)18:47
faassenas long as the PyPy crew makes it possible to transform RCTypes to code on those platforms.18:47
faassenthough Javascript might be a tad tricky.18:48
faassena native C call from javascript..18:48
faassenanyway if they can make this work well, then you'd have a nice migration path.18:48
faassenas basically it'd give people reasons to make their extensions work with PyPy even if they're not using PyPy yet.18:48
* jinty just wants the oppertunity to optimize things without learning C18:48
* philiKON is with jinty18:48
faassenI already know C. :)18:49
philiKONi do too, but oh, the pitfalls18:49
faassenI just don't want the pain.18:49
faassenanyway pyrex is pretty useful for the time being.18:49
philiKONwhich reminds me. perhaps next weekend i can get back to my Acquisition branch...18:49
faassenthat'd be cool. :)18:49
faassenZope 2.11 here we come.18:49
philiKONi managed to compile python in debug mode laready ;)18:49
* benji wants to be able to translate Python to JavaScript [insert evil laugh here]18:49
jintyphiliKON: please tell me this makes Acquisition not required....18:50
faassenbenji: well, we saw a demo of that at EuroPython18:50
philiKONwould be fun to translate test browser tests directly to javascript :)18:50
philiKONjinty, yes, in many ways18:50
faassenbenji: I saw a bub & bros client written in Python running in javascript.18:50
* jinty applauds18:50
faassenbenji: it was pretty amazing.18:50
philiKONfaassen, wow. too bad i missed that lightning talk18:50
philiKONfaassen, so he had ported pygame to JS?18:51
faassenI think it was in a PyPy talk.18:51
philiKONjinty, it will make explicit acquisition and __parent__ pointers equivalent18:51
philiKONjinty, i have it working in my sandbox18:51
faassenno, I think they haxored around all that. :)18:51
philiKONjinty, but i'm probably not counting refs right18:51
philiKONjinty, Acquisition tests pass, Zope tests segfault :(18:51
benjigood thing you posted that link, I had no idea what you were talking about... :)18:51
*** gumpa has joined #zope3-dev18:51
faassenbenji: that's another thing flowing from Armin Rigo's mind. :)18:51
philiKONhis bub'n'bros presentationw as pretty cool, too18:51
philiKONi think he's using generators and something like py.execnet for this multiplayer online game18:52
philiKONintriguingly simple architecture18:52
faassengenerators, not py.execnet18:52
faassenI think..18:52
faassenhe's using a compression algorithm stolen from video compression just to send the state of the game to the clients. :)18:52
philiKONbut something that pushed the new values over the net18:52
faassenand make it take a bit less bandwidth.18:53
philiKONriiiight, i remember18:53
philiKONbut didn't he have something where he "connected" two lists on two python interpreters and they would be like one object?18:53
faassenanyway, Armin Rigo's a programming genius.18:53
faassenyeah, but that's a pypy thing.18:53
philiKONthat's undoubted18:53
philiKONfaassen, ah. of course ;)18:53
faassenthe connected the two or three pythons.18:53
philiKONyup. makes parallel computing easier ;)18:54
*** dobee has quit IRC18:54
*** jukart has quit IRC18:55
mgedminand there's also CrackAJAX
*** volvox_ has joined #zope3-dev19:01
benjimgedmin: yep, I've written something very similar19:01
*** volvox has quit IRC19:02
SmokeyDwhat field types get visualised with a single selection widget? If I get it right, a LIst field can contain multiple values and thus gets displayed as a multiple selection widget. Does can a Collection field only contain one value, and thus gets displayed as a single selection widget?19:05
philiKONSmokeyD, Choice19:05
SmokeyDOk, I though choice was only used inside other fields19:06
SmokeyDin the value_type parameter of other fields19:06
philiKONi don't think any field is every designed to be just as part of another field19:07
SmokeyDok, thanks. I get the vocabularies. Very usefull.19:07
*** gumpa has quit IRC19:09
*** gumpa has joined #zope3-dev19:10
*** philiKON has quit IRC19:10
d2mmgedmin: have you got any more info on than the source ?19:14
mgedminno, I just found it on google19:14
mgedminian bicking provided the link19:14
d2mok, might be related to the pyjamas project19:14
benjid2m: if you're really interested, I have a slightly more advanced py2js converter with doctests19:18
*** alecm has quit IRC19:20
*** romanofski has quit IRC19:20
*** alecm has joined #zope3-dev19:20
*** gumpa has quit IRC19:29
*** gumpa has joined #zope3-dev19:29
*** kamalgill has joined #zope3-dev19:32
*** elro has quit IRC19:45
*** niemeyer has quit IRC19:48
*** niemeyer has joined #zope3-dev19:48
*** philiKON has joined #zope3-dev19:57
*** kobold has quit IRC19:58
*** elro has joined #zope3-dev20:05
*** alga has quit IRC20:18
*** d2m has quit IRC20:22
*** volvox_ has quit IRC20:24
*** RaFromBRC has joined #zope3-dev20:27
*** dobee has joined #zope3-dev20:27
*** RaFromBRC is now known as RaAtBRC20:29
*** flox has quit IRC20:34
*** romanofski has joined #zope3-dev20:35
*** faassen has quit IRC20:35
*** mkerrin has quit IRC20:37
*** d2m has joined #zope3-dev20:39
*** replicant has joined #zope3-dev20:43
*** regebro has joined #zope3-dev20:48
*** dobee has quit IRC20:48
*** Aiste has quit IRC20:50
*** dobee has joined #zope3-dev20:56
*** zbir has quit IRC21:00
*** srichter has joined #zope3-dev21:05
*** ChanServ sets mode: +o srichter21:10
*** rocky is now known as rocky|away21:20
*** jinty has quit IRC21:23
*** Aiste has joined #zope3-dev21:25
*** dobee has quit IRC21:38
*** kamalgill is now known as kamal-away21:39
*** RaAtBRC is now known as RaAtBRC|away21:42
*** regebro has quit IRC21:51
*** dobee has joined #zope3-dev21:52
*** jkakar has quit IRC21:57
*** dunny has joined #zope3-dev22:01
*** dobee has quit IRC22:07
*** alga has joined #zope3-dev22:07
srichterJ1m: if zope.component is used, is the __conform__ method in an object still valued?22:12
srichterTo me it looks like that the adapter_hook function in zope.component._api suggests that it only looks at the registry to get an answer22:13
J1mI'm not really sure what you are asking.22:14
J1mInterface __call__ methods honor it.22:14
srichterthough I cannot get my __conform__ method to be called22:14
srichterhow can I turn off the C optimization?22:15
J1mremove the .so files22:15
srichterok, thanks22:16
*** Aiste has quit IRC22:27
*** volvox has joined #zope3-dev22:30
mgedminsrichter: zope.component.getAdapter never calls __conform__22:31
mgedminonly the ISomeInterface(foo) spelling checks for __conform__22:31
*** flox has joined #zope3-dev22:32
srichtermgedmin: ok; that's what I am using22:33
srichtermgedmin: I think I figured out one bit; the error message was very cryptic since the conform call is wrapped by a TypeError exception handling, which masks errors22:33
*** flox is now known as flox|away22:33
srichter(at least in my case)22:33
*** stub has quit IRC22:34
*** mgedmin has quit IRC22:40
*** kamal-away has quit IRC22:43
*** philiKON has quit IRC22:54
*** philiKON has joined #zope3-dev22:56
*** Londo_ has joined #zope3-dev23:23
srichtermgedmin: J1m: Now that I have thought about it a bit, shouldn't zope.component.getAdapter also honor __conform__ or is this an expected difference to the IMyIface(self) call?23:23
J1mno yes23:24
*** replicant has quit IRC23:27
*** replicant has joined #zope3-dev23:28
*** zagy has quit IRC23:32
*** RaAtBRC|away is now known as RaAtBRC23:50

Generated by 2.15.1 by Marius Gedminas - find it at!