13:23 <jean-paul[m]> is the built-in greenlet module on pypy a drop-in replacement for https://github.com/python-greenlet/greenlet ?

13:25 <cfbolz> the more interesting question is the reverse ;-)

13:25 <cfbolz> jean-paul[m]: (proper answer) it should be, but it's not unlikely that the two have drifted in some details

13:25 <jean-paul[m]> Heh

13:25 <jean-paul[m]> I would generally assume whatever PyPy has is "better"

13:25 <jean-paul[m]> Okay - but at least they were meant to be compatible at some point in time

13:26 <jean-paul[m]> That's a starting point, at least :)

13:26 <jean-paul[m]> I don't have any of the details, but it seems like if a package declares an `install_requires` on "greenlet" then on PyPy it just ends up with the built-in module instead of the third-party module (which it seems probably doesn't work at all on PyPy?)

13:26 <cfbolz> jean-paul[m]: so even the greenlet module C code comes "from pypy" originally, arigato wrote it. eventually somebody took over

13:27 <jean-paul[m]> So maybe they're still at least a _bit_ close, since any project using "greenlet" just gets the runtime-appropriate version without any extra effort (so hopefully someone would notice if they diverge too much)

13:27 <cfbolz> right

13:30 <jean-paul[m]> okay, with changes to use the builtin "greenlet" instead of the third-party version on pypy (plus other random stuff, like disabling tkinter for some weird stuff...), my build system is now thinking _very_ hard. Maybe it will at least get as far as trying to run the test suite now. :)

13:31 <jean-paul[m]> ah no that was far too optimistic.

13:32 <jean-paul[m]> gfortran? great...

14:25 <cfbolz> :-(