A viable solution for Python concurrency

Posted Oct 14, 2021 17:50 UTC (Thu) by azumanga (subscriber, #90158)
In reply to: A viable solution for Python concurrency by chris_se
Parent article: A viable solution for Python concurrency

I agree with this. I worked on another similar parallelization attempt (for the GAP language, www.gap-system.org). Once the parallelization "core" was done, it was clear every library was going to misbehave, or crash, forever.

We fixed many of the problems by having objects owned by a single thread at a time, but then all code needs to know about transfering objects around, which is very painful for highly recursive objects.