<cfbolz>
Nice, but why is pyston almost exactly the same speed as pypy on many of these?
<Corbin>
Hmm, that is suspicious. And it's not just IO-bound tasks, but CPU-bound tasks too.
<cfbolz>
Yes
<cfbolz>
I would suspect a typo in a script somewhere
<Corbin>
Would you like me to be the one who reaches out? I know Michael from when I did GPU stuff. But also my technique would be pretty basic: Make an account on the comment-forum and politely ask why the numbers are so similar.
mannerism has quit [Quit: leaving]
mannerism has joined #pypy
lritter has joined #pypy
shimst3r has quit [Ping timeout: 256 seconds]
shimst3r has joined #pypy
shimst3r has quit [Ping timeout: 248 seconds]
shimst3r has joined #pypy
dmkoenig has joined #pypy
greedom has joined #pypy
derpydoo has quit [Quit: derpydoo]
<dmkoenig>
Hey folks, I'm getting a weird one where new("unsigned char **") (which should work) is being treated like new("char(*)[]") (which fails), but only does this sometimes. It's happening under PyPy 3.6 7.33 on Linux amd64
<dmkoenig>
Any idea even where to start looking? It seems like that should never rewrite like that with core types like 'char'
<dmkoenig>
(This is with CFFI, which said this was a valid channel to ask questions in too)
<Dejan>
I just realised, Fedora guys made pypy packages so that we can install pypy3.7, pypy3.8, and pypy3.9 if we want
<Dejan>
actually, that is how they do CPython too lately...
<Dejan>
dmkoenig, i wonder would it be worth it trying latest PyPy and see if that problem is still there?