RE: [Discuss-gnuradio] Building own module

[Tom Rondeau]

> -----Original Message-----
> From: Eric Blossom [mailto:address@hidden
> Sent: Tuesday, December 05, 2006 1:03 AM
> To: Tom Rondeau
> Cc: 'Daniel Garcia'; address@hidden
> Subject: Re: [Discuss-gnuradio] Building own module
> 
> On Tue, Dec 05, 2006 at 12:13:40AM -0500, Tom Rondeau wrote:
> >
> > > -----Original Message-----
> > > From: address@hidden
> [mailto:discuss-
> > > address@hidden On Behalf Of Daniel Garcia
> > > Sent: Tuesday, December 05, 2006 12:06 AM
> > > To: address@hidden
> > > Subject: [Discuss-gnuradio] Building own module
> > >
> > >
> > > I am trying to make my own module. I used gr-how-to-write-a-block as a
> > > skeleton for my module. The module compiles and passes a simple test
> in
> > > the python test directory. I then do "make install" and it installs
> the
> > > module. When I try to run a python script with that module I get an
> error
> > > that says "ImportError: No module named xxxxxxx". This happens for the
> gr-
> > > how-to-write-a-block as well as my own module.
> > >
> > > Does anyone have a clue as to why my module does not load outside of
> the
> > > "test" environment?
> > >
> > > Regards,
> > > Daniel Garcia
> > >
> >
> > Make sure all references are properly listed in Makefile.am, general.i
> (this
> > is in the gnuradio-core/src; not sure if you have one from the how-
> to...,
> > haven't looked at that in quite some time), and your *.i file is written
> > correctly.
> >
> > I'd also do a "make clean && make && make install" in case SWIG doesn't
> want
> > to look at the new block (basically, I think the problem is that SWIG
> isn't
> > compiling it into the Python module properly).
> >
> > Tom
> 
> Tom, I think it's a PYTHONPATH problem.

I see; makes sense. I had just screwed myself up last night working on the
GNU Radio module itself because of a typo in general.i. It produces a
similar problem, so I shot that post out. I'm so used to working in that
context that I never consider PYTHONPATH.

 
> It passes "make check", thus I think it's built OK.  The difference
> between the make check environment and the normal environment is the
> path where the new module is searched.

Ok, thanks.

Tom