Ralf W. Grosse-Kunstleve

> So we need to support them with fixes and not abandon them. Why slip into such radical language? I just want to wait a day or two to hopefully...

"Google doesn't care" and turning "abandon a release series to protect users" into "abandon users" will not get us anywhere good. > Our time matters too I know. I'll focus...

> We've now missed the Python 3.11 release party so I don't care anymore. My work was for nothing. Do what you want. I have to go back to teaching....

@whophil If I understand your situation correctly, you're building from 2.10, could you please try this tiny side-project, with everything else exactly as you have it already? Simply apply this...

I couldn't reproduce the `free(): invalid pointer` crash using the code here, but there is certainly a GIL issue that you can confirm by using PR #4146. The problem in...

Hi @wjakob, what do you think about the existing behavior? It's not very likely to be a problem in real life, but if it bites, it could be very surprising...

I think (but have not verified) that the existing behavior goes back to this https://github.com/pybind/pybind11/blob/5bbcba548a27baf8d6bfb7993e814da24b432dbf/include/pybind11/detail/type_caster_base.h#L524-L526 pre-empting this https://github.com/pybind/pybind11/blob/5bbcba548a27baf8d6bfb7993e814da24b432dbf/include/pybind11/detail/type_caster_base.h#L546-L562

Thanks Wenzel! That's very useful background information. > I don't do anything special for these other cases in nanobind, because I wasn't 100% sure what to do about them. What...

Did you try this with `-UNDEBUG` already?

(FYI: I just clicked the Approve & Run button for GitHub Actions without actually looking at the code. Will look asap.)