Re: [IronPython] IronPython 1.1.1 wish-list?

2008-01-02 Thread Dino Viehland
This is easy enough to fix for 1.1.1. The issue here (I think) is that somewhere in your sys.path you have a file called Oracle.exe that contains exactly 1 type (which, not surprisingly, is not a subtype of CompiledModule). Apparently we can create an instance of that type but because it's

Re: [IronPython] Difference in 1.1 and 2.0 with Dialog Result

2008-01-02 Thread Dino Viehland
I think this is bug - .NET conventions say that enum's should generally have a zero value (FxCop CA1008 :)). Because of that we can safely define __nonzero__ on enums and have that return true/false based upon if the value is zero/non-zero. Otherwise Python says all expressions otherwise

Re: [IronPython] IronPython 1.1.1 wish-list?

2008-01-02 Thread Dino Viehland
FYI this one is already fixed in the v2.0 tree, but I'll look at fixing it in v1.1.1 as well - it's a little more invasive than I'd usually want but it shouldn't be too bad. But if you want to work around it for the time being you can replace __init__ w/ __new__: class TESTINGwrite(file):

Re: [IronPython] [python] Re: IronPython 1.1.1 wish-list?

2008-01-02 Thread Dino Viehland
I think all of these can be fixed w/ the exception of the first two - or at least I have fixes for them. Now I just need to make sure I haven't broken anything :). The 1st one is probably too invasive to fix in v1.x (and is fixed in v2.0). The 2nd one I'm not sure I understand. I seem to be

Re: [IronPython] IronPython 1.1.1 wish-list?

2008-01-02 Thread Dino Viehland
On quit(): quit is actually not defined by the runtime implementation of Python, rather it's part of the standard library (defined in site.py and installed w/ the setquit function). Running w/ the standard library set (or just from CPython's Lib dir) will result in that functionality being