Re: comedi python wrappers

From: John Conner <conner_at_empiredi.com> · Date: Thu, 13 Nov 2003 11:32:47 -0700

Bryan,

Have you been able to test the sample code that is in comedilib 0.7.21?

I'm not able to get it to run on my box with the PC104-DAS16 or the
PCI-DAS1602. I know for sure both are now happy with comedi/comedilib.
Similar problems with the code Luc Lefebvre put out. Only his  version 
of info.c really works.  

What I need is to get the comedi_read_data function working. I get errors
saying that on of the parameters needs to be a _p_lsample_t pointer. I'm
probably missing something obvious but so far I've not found the right 
approach. Any ideas? 

The swig wrappers are a good idea but not always as simple as it seems.

John Conner

PS. You should update the url for your PythonDAq page. 
	http://www.cole.uklinux.net/PythonDAq
Took a bit to find and I remembered it was changed.

> Yes, I've looked at the existing "ComPy" module in comedi CVS. As far as
> I can see, ComPy does not include streaming/commands support.
> 
> In fact, it would be good to liase with the authors of the ComPy module
> (or anyone else with an interest in python/comedi wrappers), since it
> would make sense to write a more robust,"pythonic" ComPy module on top
> of the SWIG-generated wrappers. SWIG maps all the comedilib functions to
> python functions in a 1-to-1 fashion, with no error checking or other
> "idiot-proofing". When I next get a Comedi-supported card installed,
> I'll have a go at this.
> 
> The SWIG-generated code provides a simple mechanism for creating and
> manipulating all the various comedi structs as python objects. Since
> python already supports file-descriptor I/O, no special code is required
> for streaming acquisition. The SWIG generated wrappers have an
> additional advantage of easy maintainability: as Comedilib evolves, just
> re-run SWIG to re-generate the code. SWIG generates the entire wrapper
> c-code from an interface file under 30 lines long!
> 
> Bryan
> 
> 
> 
> 
> _______________________________________________
> comedi mailing list
> comedi_at_comedi.org
> https://cvs.comedi.org/cgi-bin/mailman/listinfo/comedi