Marcus Müller

Thanks! This also means we should probably have a unit test that covers this.

We don't have in-band signalling for that yet, and it would break existing functionality – we sadly can't fix that on the 3.x side of things. We'd need to come...

Can't reproduce on main with ```python #!/usr/bin/env python3 from gnuradio import gr from gnuradio import blocks from time import sleep tb = gr.top_block() src = blocks.null_source(1) sink = blocks.null_sink(1) tb.connect(src,...

Can't reproduce on v3.10.11.0 either. @EliteWeapon that puts the onus of giving us something minimal that can reproduce the issue, and providing us with more information on your platform, on...

Thanks for the extensive look and proposed solution. Interestingly I still can't make the thing crash, but your reasoning does seem valid.

marking as medium (because impact is limited in number of users, but still seems to be serious enough)

really wonder why this is not making my machine crash but yours; my guess is that on my machine, the thread lingers on? Or I have a different libc? Can...

@tbzatek I'm probably the reason @MestreLion showed up, as I [asked around](https://unix.stackexchange.com/questions/775488/udisksctl-get-loop-device-and-mount-point-without-resorting-to-parsing-localize) how to use udisks from a shell script. So, I honestly can say that, yes, I've tried to...

A viable solution would be to kill the `#defines` and finally getting around writing templated blocks that take `float (*func)(float)`, or more generally `typename T`, `T (*func)(T)`, and maybe `T...

this was, looking back, a feature request, and as such is still valid.