<mattip>
still no support for deep learning frameworks (pytorch, tensorflow), but everything else should just work on linux64
<lesshaste>
mattip, could this be added to the website?
<lesshaste>
the info is very minimalist currently
<lesshaste>
the download page would ideally provide all the info you need to download the right thing
<lesshaste>
should
<mattip>
PR welcome
<lesshaste>
first I would need to be the expert :)
<lesshaste>
I am happy to copy whatever the experts tell me into a PR
<mattip>
the point of the information is to help people who are not experts
<mattip>
so it would be great if a non-expert would make a stab at contributing information that was helpful for them to get going
<mattip>
we can then iterate on that to fine tune it
<mattip>
I have little insight as to where a newcomers gaze falls when looking for information, since I don't use the site for that
<lesshaste>
sure. I just didn't know what conda provides for example
<mattip>
right, so it makes sense to write that after you try it out, hit the hurdles and figure out how to cross them, and then you can help others avoid them
<lesshaste>
mattip, sort of. If you or others who have extensive experience and knowledge already know the answers to all the main questions it does seem a little perverse to keep them secret and ask someone who doesn't know to write the relevant docs
tumbleweed has quit [*.net *.split]
tazle has quit [*.net *.split]
antocuni has quit [*.net *.split]
pjenvey has quit [*.net *.split]
ammar2 has quit [*.net *.split]
<lesshaste>
I am happy to help but I have very little knowledge except for sudo apt install pypy
<lesshaste>
this reminds me a little of a conversation I had recently with the main nuitka dev
<mattip>
there is no secret. I know to point people to that blog post. I know alot about conda and how to set up a channel, how to specify pypy
<lesshaste>
I guess you also know what should really be after "Standard library modules supported by PyPy. Note that large parts of python library are implemented in pure python, so they don't have to be listed there. Please just check if it imports. If it imports, it should work:"
<lesshaste>
that list in unexplained and includes numpypy
<lesshaste>
that list *is unexplained
<mattip>
IMO it would be less confusing to rewrite that whole section and remove that list.
<mattip>
does that list help at all?
<lesshaste>
ah ok
lesshaste has quit [Remote host closed the connection]
<lesshaste>
what is the point that they are trying to make?
lesshaste has joined #pypy
<mattip>
it is discussing stdlib modules, which are not all that interesting
<lesshaste>
but _collections is not the right name is it?
<lesshaste>
I mean with the "_"
<mattip>
maybe it was python2?
<lesshaste>
that was an underscore
<lesshaste>
and numpypy?
<lesshaste>
that is ancient isn't it?
<mattip>
dead
<mattip>
ahh, some of those are internal modules that support the stdlib modules
<lesshaste>
ok so deleting the list makes sense. There used to be a website where modules were tested for compatibility I think
<mattip>
it too is so out of date as to not be useful
<lesshaste>
ok
<mattip>
thanks, this was helpful
<Dejan>
few times i was wondering how would pure python numpy perform on top of PyPy...
antocuni has joined #pypy
tazle has joined #pypy
ammar2 has joined #pypy
tumbleweed has joined #pypy
pjenvey has joined #pypy
<Dejan>
once I even thought of starting a project called "pynum" that basically implements the same set of numpy modules but in pure python :)
<Dejan>
temporal insanity i would say...
ammar2 has quit [Ping timeout: 265 seconds]
ammar2 has joined #pypy
<mattip>
lesshaste: does this preview of a PR to fix up the page look more helpful?