[PythonCE] PyCrypto binaries

[alexandre.delattre at enst-bretagne.fr]

"d at tenmaulwurf" <lists at nanl.de> a écrit :

> alexandre.delattre at enst-bretagne.fr wrote:
>> Johnny deBris <johnny at debris.demon.nl> a écrit :
>>
>>
>>> Matt S. wrote:
>>>
>>>> since I've not yet noticed the wxPython and ctypes binaries available
>>>> for 2.5).
>>>>
>>>>
>>> ctypes seems to be part of the build... Works for me on 2.5 at least. ;)
>>>
>>> Cheers,
>>>
>>> Guido
>>>
>>>
>>>
>>
>> Yes, ctypes is now a part of Python standard library thanks to   
>> Thomas Heller.
>> As for wxPython, I maybe gonna look in compiling it for 2.5.
> It would be _REALLY_ cool. http://wxpyce.wikispaces.com/ is a try, but
> the binaries do not work for PythonCE 2.5 / wm5/6.
> I ever wanted to try following the Build-process-instructions with the
> latest sdk's (out of the hope, it will then run under 2.5 / wm5/6) but
> I have no x86-machine and no windows.
> Using wxwidgets with latest versions of python / wm would be very, very cool!
>>
>> Regards,
>>
>> Alexandre
>>
> Thanks,
>
> Mirko
>>
>> _______________________________________________
>> PythonCE mailing list
>> PythonCE at python.org
>> http://mail.python.org/mailman/listinfo/pythonce
>>

Hello Mirko,

I will surely compile wxPyCe for 2.5 these days, as many people should  
be interested in. However, by experience, don't expect to be able to  
run wx+pycrypto+numpy+whatever c extension at the same time unless you  
have a very recent WM 6.0 device (it's a known problem of lack of  
virtual adressing space for dlls, that has been expanded from 32 MB to  
1 GB in WM 6.0 kernel).
That's the reason why I have ported Venster (and now secretly working  
on something FAR more pythonic & high level for embedded gui  
development), and not because I dislike wx or I wanted to reivent the  
wheel.

It'll be cool if PythonCE people could have the choice between  
portability (TKinter/wx) at the cost of non-native rendering or high  
memory consumption or something more specific but that delivers native  
rendering with a small footprint at the cost of learning something new  
(VensterCE that may be phased out by my new toolkit).

My opinion is that splitting backend & frontend of an application  
helps portability, and rewriting the frontend for an handeld with its  
small screen size and its own design guidelines should not be a big  
deal if the toolkit is easy to learn (and better, pythonic), the  
biggest advantage of the new toolkit over venster.

This is only my opinion, and people are, of course, free to see  
otherwise and choose what they think best for their own development.

That's why I will compile wxPyCe for 2.5 and still support ctypes gui  
toolkit at the same time.

Regards,

Alexandre.

PS :
If someone want to give a try, help, give suggestions, or learn more  
about the new toolkit, I'll be glad to send you my current work with  
an example application. The reason why I call the new toolkit "new  
toolkit" is because I've wanted to call it first pycegui, but realized  
that it was the bindings of CEGUI for pygame ! Reminds me something  
about Pyro ;)