PSF grant for "Shed Skin" to python 3.x?

Dan Stromberg drsalists at gmail.com
Thu Jan 11 16:49:08 EST 2018


Hi folks.

Is anyone on the list familiar with the process of getting a Python
Software Foundation grant awarded to a worthy project?  And with
what's generally considered "worthy"?  I found
https://www.python.org/psf/records/board/resolutions/ , but almost all
of the grants seem to be for conferences, and other "peopley"
recipients.

But ISTR hearing that PSF grants are available for porting code to Python 3.

I find Shed Skin, the implicitly-static Python2 -> C++ transpiler
pretty interesting. Shedskin is at
https://github.com/shedskin/shedskin

ISTR that shed skin can be used for entire programs, but probably
where it really shines is in taking your CPython critical section and
turning it into an extension module that runs at near-C++ speed with
minimal tweaking.

I realize microbenchmarks aren't good indicators of the performance of
a large, complex software system, but still, the fact that shed skin
did so well in:
http://strombrg.blogspot.com/2015/05/python-compared-to-c-machine.html
...suggests to me that we don't want it to be a casualty of the move
from 2.x to 3.x.

On that microbenchmark at least, shed skin outperformed pypy, cython
(both manifestly typed and implicitly typed) and numba, and was just
slightly slower than hand-written C++. Furthermore it appears to be
the most practical of the handful of Python -> C++ transpilers
available, other than it still being 2.x-only.

So my question is: With a suitably written grant proposal and
reasonable budget, might a PSF grant be available to help the (quite
busy) author of shedskin, to get it contending with 3.x syntax, 3.x
semantics and 3.x extension module API?

Thanks!

PS: I know we could just write a grant and see how it goes, but I'm
hoping to get a "yes, that's the sort of thing PSF grants are for"
before we put the time into writing the grant proposal.



More information about the Python-list mailing list