Welcome to the main channel on the development of MoarVM, a virtual machine for NQP and Rakudo (moarvm.org). This channel is being logged for historical purposes.
Set by lizmat on 24 May 2021.
nine Role currying wasn't actually that bad. Type captures look better now as well. During the runtime of the role body, both parameters and type captures are available and have sensible values. So the final piece that's missing is proper lexical capture and fixup. 12:52
We already have the p6capturelex calls and cloning of methods. But that apparently isn't enough.
Now it would help a lot if I actually understood how this lexcial fixup thingy works 13:02
lizmat understanding is a great help! :-) 13:03
nine Understanding is a three edged sword 13:12
lizmat I guess that makes it tricky to hold without getting cut 13:37
Nicholas MasterDuke: you mentioned lwn.net/Articles/893686/ (which is from one of the Pyston developers). Then we have pyfound.blogspot.com/2022/05/the-2...mit_2.html which is (core) python "faster Python" 17:38
but the author of the first talk (Kevin Modzelewski) "warned that his “unsubstantiated hunch” was that the Faster CPython team had already accomplished “greater than one-half” of the optimisations that could be achieved within the current constraints." -- pyfound.blogspot.com/2022/05/the-2...ummit.html 17:39
continuing "Modzelewski encouraged the attendees to consider future optimisations that might cause backwards-incompatible behaviour changes."
which leads to pyfound.blogspot.com/2022/05/the-2...on_11.html -- 3.9 had been chosen as a target for now, [...] There is still much broader support for Python 3.9 among third-party packages than for Python 3.10, and so Python 3.9 still made more sense as a base branch for now rather than 3.10 or main. 17:40
er, wait, so "much broader support" *with* what are considered not (yet) to be "backward *in*compatible behaviour changes"
(and it's also clear that the various GIL changes need C ABI changes, and C extensions might not work unchanged) 17:42
sigh, I like jnthn for deciding that there would be no C extensions
I also wonder (in the future) if the "faster python" stuff attempts a JIT, but it's going to hit the same fun as MoarVM and the JIT, where progressively more C code needs to be re-written in "slower" NQP so that it can actually go faster
(none of this stuff is easy, and I applaud people for attempting it) 17:44
nine Yeah, keeping internals internal was the best design decision ever 18:02
[Coke] +1 22:07

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!