Re: Raspberry Pi image with GNU Radio 3.10, SDRplay, and gr-radio_astro

[Jeff Long]

Great! Moving to 3.10 will make your life much easier.

On Thu, Dec 2, 2021 at 12:47 PM Glen Langston <glen.i.langston@gmail.com> wrote:

Hello everyone,

Sorry to have been a gnuradio "problem child" for so long.

Basically the only way we could overcome the various gnuradio coding

improvements was to beg for help.   Franco Venturi has been

marvelous!

Thanks to Franco pointing to links to useful webpages, our

troubled python code seems to be working correctly now.

Our Radio Astronomy group ended up jumping from gnuradio 3.7 to 3.10,

At least for the Raspberry PI 4 astronomy versions, it looks like we 

may get everything running very soon.

 

For our little group, the relevant graph, vectordemo.grc, works fine now.

This design only uses python code I wrote and no C++ programs.

I pushed ra_vmedian.py and ra_vave.py back to our gr310 branch.

These are in the new python/radio_astro subdirectory.

Our updates probably should also be pushed back to earlier branches.

(How did this ever work?)   I’ll try to fix our WVURAIL/gr-radio_astro

3.8 main branch.

Hopefully, we’ll get to the bottom of the remaining code issues.

Cheers

Glen

On Dec 1, 2021, at 9:26 PM, Franco VENTURI <fventuri@comcast.net> wrote:


Lastly the way the function 'forecast' is supposed to work in a Python module (see here: https://github.com/gnuradio/gnuradio/blob/master/gnuradio-runtime/python/gnuradio/gr/gateway.py#L153-L163) is that it takes two arguments, the number of expected output values and the number of inputs (as in input ports, not input values), and it should return a vector of integers (one for each input port), that answers more or less this question 'how many values of the n-th input do you need in order to be able to give me n_output values out?'; in a simple case of a block with just one input (port), i.e. ninputs=1, the returned value should be a vector of length 1, not a scalar.

Franco