Re: [Discuss-gnuradio] Recent change in destructor behavior?

[Tom Rondeau]

On Mon, Oct 5, 2015 at 2:52 PM, Tom McDermott <address@hidden> wrote:

My OOT module prints some statistics upon closing of the flowgraph from GRC.

This is implemented with printf(stderr) in the destructor for my module.

GRC --> IMPL constructed by Make --> constructs one of my objects.

This worked well through all versions of gnuradio until about last week.

Now the printf() never shows in the GRC console after a rebuild of 3.7.9.git,

the version that introduced per-flowgraph QSS support (which appeared

broken as of last weekend).

printf(stderr) prior to the start of flowgraph shutown still works fine. It's just

the one in the destructor that seems no longer to print.

Is there a recent change that would explain this, or is printf() in a destructor

bad practice?

-- Tom, N5EG

There's nothing that I can think of that would have changed this behavior from the GNU Radio scheduler.

Sebastian, could this be due to a change in the structure of the flowgraph template for how Python shuts things down?

Tom