2 ms·

This is a delicate issue. As guido stated (https://discuss.python.org/t/pep-703-making-the-global-interpreter-lock-optional-3-12-updates/26503/19 https://discus

This is a delicate issue. As guido stated (https://discuss.python.org/t/pep-703-making-the-global-interpreter-lock-optional-3-12-updates/26503/19 https://discuss.python.org/t/pep-703-making-the-global-inter...): > If there’s one lesson we’ve learned from the Python 2 to 3 transition, it’s that it would have been very beneficial if Python 2 and 3 code could coexist in the same Python interpreter. We blew it that time, and it set us back by about a decade. So it's not so easy to decide to go with again 2 different versions of Python. I understand how caution they are. However, this commitment from Meta, complemented with MS financing Guido, and the quality of Sam Gross proposal makes me optimistic.

dist-epoch 3y ago
> However, this commitment from Meta, One core dev for 3 years. Hardly earth-shattering, if compared to the billions poured into JavaScript.
- manfre 3y ago
  > three engineer-years (from engineers experienced working in CPython internals) between the acceptance of PEP 703 and the end of 2025 This means multiple devs over a period of less than 3 years.
- BiteCode_dev 3y ago
  True, but it's way better than what we had before.
ignoramous 3y ago
> PEP-703: Concurrent collection requires write barriers (or read barriers). The author is not aware of a way to add write barriers to CPython without substantially breaking the C-API. Jesus. Imagine a language runtime being this hamstrung.
- BiteCode_dev 3y ago
  Well, CPython is 4 years older than Java. That's very old. When it came out, this was in store: https://img.redbull.com/images/w_3000/q_auto,f_auto/redbullcom/2016/01/06/1331768922880_2/game-boy-nintendo-console-portable-jeux-video https://img.redbull.com/images/w_3000/q_auto,f_auto/redbullc... What was not in store were: - multi-core CPU - wide spread internet access So people solve concurrency with a simple but efficient and sufficient method at the time: the GIL. And nobody planned for the scientific and ML stack.

dist-epoch 3y ago

> However, this commitment from Meta, One core dev for 3 years. Hardly earth-shattering, if compared to the billions poured into JavaScript.

manfre 3y ago
> three engineer-years (from engineers experienced working in CPython internals) between the acceptance of PEP 703 and the end of 2025 This means multiple devs over a period of less than 3 years.
BiteCode_dev 3y ago
True, but it's way better than what we had before.

ignoramous 3y ago

> PEP-703: Concurrent collection requires write barriers (or read barriers). The author is not aware of a way to add write barriers to CPython without substantially breaking the C-API. Jesus. Imagine a language runtime being this hamstrung.

BiteCode_dev 3y ago
Well, CPython is 4 years older than Java. That's very old. When it came out, this was in store: https://img.redbull.com/images/w_3000/q_auto,f_auto/redbullcom/2016/01/06/1331768922880_2/game-boy-nintendo-console-portable-jeux-video https://img.redbull.com/images/w_3000/q_auto,f_auto/redbullc... What was not in store were: - multi-core CPU - wide spread internet access So people solve concurrency with a simple but efficient and sufficient method at the time: the GIL. And nobody planned for the scientific and ML stack.