[00:13:01] ** Maniac has left IRC ("ChatZilla 0.9.63a [Mozilla rv:1.6/20040206]")
[02:04:45] ** gbay has joined us
[02:20:31] ** gpciceri has joined us
[04:03:49] [connected at Tue Jun  1 04:03:49 2004]
[04:03:49] <> *** Looking up your hostname...
[04:03:49] <> *** Checking ident
[04:03:49] <> *** Found your hostname
[04:04:20] <> *** No identd (auth) response
[04:04:20] <> *** Your host is wells.freenode.net[wells.freenode.net/6667], running version dancer-ircd-1.0.35
[04:04:20] [I have joined #peak]
[04:04:20] ** wells.freenode.net set the topic to peak - Python Enterprise Application Toolkit | http://peak.telecommunity.com | alpha3 is out !!!
[04:12:05] ** gpciceri has left IRC (Read error: 110 (Connection timed out))
[04:12:15] ** gpciceri has joined us
[06:14:54] ** gpciceri has left IRC ("Client exiting")
[06:39:13] ** gbay has left IRC ("It's not like I'm leaving or anything...")
[07:21:33] jack-e|away is now known as jack-e
[07:57:10] ** sremington_ has left IRC ("bye")
[08:45:23] ** rdmurray has joined us
[08:57:37] ** sremington has joined us
[09:00:24] ** rdmurray has left IRC ("User disconnected")
[09:07:20] ** rdmurray has joined us
[11:32:39] ** hazmat has joined us
[12:00:43] ** Maniac has joined us
[13:01:02] ** hazmat has left IRC (Read error: 110 (Connection timed out))
[13:59:20] jack-e is now known as jack-e|away
[14:07:22] ** gpciceri has joined us
[15:21:42] ** hazmat has joined us
[15:21:45] ** hazmat_ has joined us
[15:24:41] ** hazmat_ has left IRC (Client Quit)
[15:24:41] ** hazmat has left IRC (Client Quit)
[15:53:14] ** _jpl_ has joined us
[15:55:13] ** panix- has joined us
[15:55:18] <panix-> anyone here?
[15:57:29] <rdmurray> No.
[15:58:07] <chillywi1ly> nah
[15:59:00] chillywi1ly is now known as chillywilly
[16:01:41] <panix-> well, just a quick python question... does anyone know how to modify the contents of globals, *globally*
[16:01:46] <panix-> I know that question sounds a bit off...
[16:02:02] <panix-> but, basically, I want to set a __metaclass__ so that all future imported objects will use that metaclass.
[16:02:45] <rdmurray> Well, if you modify globals, I believe that is a global effect.
[16:02:53] <panix-> unfortunately, it seems not to.
[16:03:08] <rdmurray> Hmm.  I could have sworn that zope stuff things into globals that way (which bugged me).
[16:03:12] <Maniac> that doesn't make sense
[16:03:14] <panix-> if I, for example, do -->
[16:03:14] <rdmurray> s/stuff/stuffed/
[16:03:25] <panix-> globals()['__metaclass__'] = MyType
[16:03:41] <panix-> from somemodule import SomeClass
[16:03:48] <panix-> SomeClass will not have a __metaclass__
[16:04:02] <panix-> rdmurray: I am sure that Zope does it somehow :)
[16:04:06] <panix-> just don't know how.
[16:04:11] <Maniac> ask in #python or #zope :P
[16:04:16] <panix-> heh, sure.
[16:04:32] <Maniac> all the smart people are idling
[16:04:32] <rdmurray> I don't think you can set __metaclass__ that way.
[16:04:50] <Maniac> oh, except for rdmurray and chillywilly they're probably smart
[16:04:56] <panix-> :)
[16:04:57] <chillywilly> cause each type has its own __metalclass__
[16:05:07] <rdmurray> __metaclass__ is a module variable, not a global, if I understand right.
[16:05:09] <chillywilly> __metaclass__ too
[16:05:22] <panix-> well, if it is set in globals() it has an effect on local classes
[16:05:26] <panix-> so thats not true
[16:05:29] <chillywilly> what's so hard about having one extra line ;)
[16:05:45] <panix-> though, if I print globals() inside each module, then I see that each seems to have its own unique copy.
[16:05:55] <rdmurray> panix-: how do you know?  How do you get it set in globals? 
[16:06:02] <panix-> I set it just as above
[16:06:13] <panix-> globals()['__metaclass__'] = MyType
[16:06:17] <rdmurray> Um, but you said that didn't work.  I'm confused.
[16:06:32] <panix-> no, it works on local classes, not on classes from imported modules.
[16:06:43] <rdmurray> Well, of course not.
[16:06:58] <rdmurray> The imported class takes its metaclass from its module. 
[16:07:03] <rdmurray> imports don't change metaclass.
[16:07:04] <panix-> right, hence the problem
[16:07:52] <panix-> what seems to be happening is that the builtin 'globals' function gives me a copy, not the *true* globals dictionary
[16:08:02] <panix-> so that each module has its own globals
[16:08:21] <chillywilly> what if you use __import__ and pass in the globals?
[16:08:52] <chillywilly> or wrap it up in some function or what not
[16:08:56] <panix-> well, that works great.
[16:09:03] <panix-> but, I don't want to do it that way :)
[16:09:11] <chillywilly> sucks to be you ;)
[16:09:21] <panix-> I want existing python modules to automagically take on my metaclass
[16:09:22] <panix-> indeed :)
[16:09:35] <panix-> I am doing some pretty hefty AOP trickery here :)
[16:09:50] <chillywilly> hmmmm
[16:10:03] <panix-> I can set it on __main__ and see it inside the module
[16:10:11] <panix-> but, __metaclass__ doesn't take effect from there.
[16:10:16] <chillywilly> you can repolace __import__ with your own implementation
[16:10:17] <chillywilly> I think
[16:10:22] <chillywilly> replace too
[16:10:22] <panix-> thats a decent idea.
[16:12:17] <chillywilly> http://docs.python.org/lib/built-in-funcs.html#l2h-6
[16:12:46] <panix-> I really would like to figure out where globals() is getting its variables from.
[16:13:30] <chillywilly> the module?
[16:13:49] <panix-> well, I mean the actual dictionary
[16:14:08] <panix-> the globals() builtin function is copying some dictionary and returning it
[16:14:14] <panix-> I want access to that dictionary :)
[16:14:26] <panix-> but, I think that overriding __import__ will probably work.
[16:14:41] <rdmurray> I found what zope does.  It modifies __builtin__.
[16:15:46] <rdmurray> Very evil :)
[16:15:49] <panix-> heh
[16:15:52] <panix-> of course :)
[16:15:55] <panix-> link to CVS?
[16:16:06] <rdmurray> Nah, I have it installed.
[16:16:12] <panix-> gotcha
[16:16:13] <rdmurray> ZODB/Transaction.py
[16:16:27] <rdmurray> import __builtin__
[16:16:40] <rdmurray> __builtin__.get_transaction = get_transaction
[16:16:51] <chillywilly> hmmm, is __metaclass__ in the module's __dict__ ?
[16:16:52] <rdmurray> del __builtin__
[16:18:40] <panix-> well, __metaclass__ can be located there.
[16:22:04] ** hazmat has joined us
[16:23:55] <_jpl_> Doesn't each module have it's own global scope?
[16:24:02] <panix-> _jpl_: yes.
[16:24:07] <_jpl_> Or is that considered module-level scope?
[16:24:19] <panix-> globals() == module level scope.
[16:24:39] <panix-> and, unless you are inside a function or class, locals() == globals()
[16:25:01] <_jpl_> Right, so then globals() in the importing scope will not be the same as that of the imported module.
[16:30:19] <panix-> hmm
[16:30:21] <panix-> interesting
[16:30:37] <panix-> well, I overwrote __builtins__.globals with my own function
[16:30:43] <panix-> and managed to shove the metaclass into there
[16:30:55] <panix-> but, now the module sees the metaclass, but doesn't actually use it :)
[16:34:42] ** Maniac has left IRC ("Client exiting")
[16:40:43] <_jpl_> I wonder if you could do some tricks with the 'new' module.
[16:41:47] <panix-> I am actually trying to override __import__ right now to see if that will do the trick
[16:42:30] <_jpl_> That might be the right direction to go, but then you should look at the 'imp' module.
[16:45:09] <panix-> I looked at it a little... looks kinda complicated :)
[16:53:20] <panix-> arg... overriding __import__ won't work.
[16:53:23] <panix-> I will have to use imp
[16:54:26] <_jpl_> What will you do if an object already has a custom metaclass?
[17:00:12] <panix-> overwrite it.
[17:00:28] <panix-> but, for my case, this should be okay
[17:00:52] <panix-> actually, it seems that the local one will hold.
[17:00:56] <panix-> it takes preference over mine.
[17:00:57] <panix-> which is good.
[17:07:19] <chillywilly> you can use imp to implement __import__ supposedly ;)
[17:10:37] <panix-> chillywilly: yes
[17:10:43] <panix-> that is what I am going to have to do.
[17:10:58] <chillywilly> ah well ;)
[17:11:00] <panix-> the problem is that I need it to be exactly like import so that I don't break anything...
[17:11:03] <chillywilly> such is life I guess
[17:11:04] <panix-> kinda scary
[17:11:06] <panix-> indeed.
[17:11:19] ** gpciceri has left IRC ("Client exiting")
[17:11:24] <panix-> I am considering just forcing things to explicitly import and reference the metaclass, but I don't like that very much...
[17:19:56] ** gbay has joined us
[17:26:58] ** sremington has left IRC ("bye")
[17:36:33] <panix-> ah hah!
[17:36:36] <panix-> got it to work :)
[17:36:46] <chillywilly> w00t
[17:37:02] <panix-> I did it without needing imp too
[17:37:29] <panix-> just overrode __import__
[17:37:41] <panix-> dynamically created a class that used my metaclass
[17:37:52] <panix-> then forced any imported objects to subclass that class
[17:39:15] <panix-> still not perfect.
[17:39:18] <panix-> but, good enough for now
[17:39:20] <panix-> time to go home!
[17:39:21] ** panix- has left IRC (Remote closed the connection)
[17:55:59] <rdmurray> That's how p
[17:55:59] <rdmurray> PEAK does it, AFAIK.
[17:56:06] <rdmurray> Oh, he's gone.
[17:56:25] <chillywilly> doh
[17:59:23] ** gpciceri has joined us
[18:14:18] ** gpciceri has left IRC ("Client exiting")
[18:30:29] ** sremington has joined us
[18:31:34] ** hazmat has left IRC (Read error: 110 (Connection timed out))
[18:50:19] ** gbay has left IRC ("It's not like I'm leaving or anything...")
[19:00:56] ** Maniac has joined us
[19:02:52] ** rdmurray has left IRC ("User disconnected")
[19:14:00] ** _jpl_ has left us
[19:28:46] ** Maniac has left IRC ("ChatZilla 0.9.63a [Mozilla rv:1.6/20040206]")
[19:38:05] ** philips has left us
[19:38:07] ** Maniac has joined us
[19:39:06] ** hazmat has joined us
[20:18:28] ** hazmat has left IRC (Read error: 110 (Connection timed out))
[21:55:35] ** rdmurray has joined us