News for hackers. Stuff that matters.
Re:Ask HN: Production Lisp in 2020? by elcomet on Tuesday May 19, 2020 @09:53AM (#23234867)
> and is not really standardized

What do you mean by this ?

Re:Ask HN: Production Lisp in 2020? by justinmeiners on Tuesday May 19, 2020 @11:04AM (#23235612)
My understanding is the Python standard is essentially whatever C Python chooses to do. They do write the changes to a specification separately, but essentially its one moving implementation, with alternatives that have compatibility limitations.
Re:Ask HN: Production Lisp in 2020? by schmudde on Tuesday May 19, 2020 @01:30PM (#23237576)
Indeed. I'm curious what Python 4 will do to Python 3.
Re:Ask HN: Production Lisp in 2020? by p_l on Thursday May 21, 2020 @03:57PM (#23264197)
"standard" is "whatever CPython managed to miss a segfault for today".

And PyEvalFrameEx is the stuff of nightmares that can't be really fixed because it hardcoded spaghetti that, if rewritten, could subtly break python in myriad ways that would make 2->3 migration a cakewalk.

Oh, and there's no real test suite to verify if you didn't break something subtle in implementation.

While parsing Ruby code is well known to be problematic, there is a very good reason why Ruby can have interchangeable implementation while Python pretty much can't.