This is a followup to variations in decoding the same file. You may recall that I was implementing a sink block that accepted AFSK tones and produced a PDU message. After messing with set_history() and set_output_multiple(), I learned that the scheduler will sometimes provide different amounts of noutput_items to work(). Because I was mishandling noutput_items, it looked like the decoding was non-deterministic, but it was actually variations in noutput_items.
Other things that may be useful for beginners: a) it's easy to print a few things to the output window using: GR_LOG_INFO(d_logger, "Message goes here"); b) you can define a separate output (or outputs) for debugging and assign an internal variable to it in work(); then use the QT Time plot to trace your internal variable against inputs or other streams. c) the argument noutput_items applies to both inputs AND outputs, for any variation of a sync block. This includes sink and source blocks. This is especially confusing for a sink block because there are no outputs. You might think that you would tell the scheduler how many items you consumed, but no, you declare a quantity produced, and that gets applied to both ins and outs. Eventually you can find this in the docs and examples, but even then it is not obvious. ________________________________ From: Gavin Jacobs <[email protected]> Sent: September 21, 2016 10:07 AM To: [email protected] Subject: Re: [Discuss-gnuradio] Module development and debugging Are you saying that on the same file, it gives non-deterministic results? If so, that might be a bug in the file, but could also be in GNU Radio. There are rumours that there's a randomness bug in the scheduler, but it's hard to reproduce. GBJ> I set the wav file to repeat, and it always decodes differently; but I should be able to provide more info once I learn more about the items below.
_______________________________________________ Discuss-gnuradio mailing list [email protected] https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
