Yes every project eventually grows to do something computationally hard. You always choose threading because of shared memory. But then you learn about the GIL shortly after.
Then you look at multiprocessing and managers. ugh.
>get rid of GIL (PEP 703)
All good and dandy, but easier said than done im sure.
>better interpreter perf
Would have been done already.
>We shouldn’t feel like we have to accept PEP 703.
I wonder if you could abstract the GIL to be more intelligent and multiprocessed in of itself. Likely come at a overhead but in turn gain the performance everyone is seeking.