Yes, and a challenge I wasn't expecting.  I was thinking the biggest problem 
was likely to be packaging the app for easy consumption in windows. I'd heard 
of potential issues around pickle, but had never needed to serialize anything 
more than simple data, I didn't think I'd be butting heads with it simply 
because of the need for some background processing. 

As I learn more about pickle, I do find myself wondering if it would be easier 
to change the basic program flow rather than try to figure out which pieces are 
pickleable, particularly if it turns out that some piece of some library just 
isn't built with pickling in mind.


> Date: Mon, 2 Jan 2012 19:06:06 -0800
> From: [email protected]
> To: [email protected]
> Subject: Re: [SEAPY] multiprocessing in win32: pickle limitations?
> 
> Hmm, sounds like a challenge!
> 
> What kind of objects do you need to pass around between processes?
> 
> If they are big blocks of data, numpy arrays may be helpful -- they
> are pickle, and there are also other ways to share the data without
> copying.
> 
> If you are sharing pyGTK object -- I suspect you have a challenge
> ahead of you, as they are wrappers around C structs (classes?), so
> pickling is a challenge. You may be better off passing the dta
> required to re-create your objects, rather than trying to pass the
> objects themselves.
> 
> HTH,
>    -Chris
> 
> 
> --
> 
> Christopher Barker, Ph.D.
> Oceanographer
> 
> Emergency Response Division
> NOAA/NOS/OR&R            (206) 526-6959   voice
> 7600 Sand Point Way NE   (206) 526-6329   fax
> Seattle, WA  98115       (206) 526-6317   main reception
> 
> [email protected]
                                          

Reply via email to