On Fri, May 20, 2016 at 8:04 PM, William Stein <wst...@gmail.com> wrote: > Hi, > > Somebody with possibly substantial grant funds asked me today: "How > much might it cost to get Sage to work with Python3?" > > I don't know. Anybody have any thoughts? Is there anybody reading > this who wishes they could spend several months getting Sage to fully > work with Python3 instead of teaching Calculus?
Based on previous experience with porting large code-bases I'd wager something around 4 person-months. With any luck it could be less. So much of Sage is Cython-oriented it would get a lot of Python 3 support "for free", but even Cython modules would be best ported over to use Python 3-isms too. The other good news is that so much of Sage is numerically oriented and not as likely to have string-related issues in as many places. Some of the trickier I/O related code is bound up in pexpect which already works with Python 3. So my bet is that the majority of effort will be slogging through the test suite to make sure everything is passing. Doctests are notoriously tricky to make fully Python 2/3 compatible. But many of the issues can be addressed through custom doctest flags / output checkers. I did some work like this to get all of Astropy's doctests working across Python versions, but Astropy has orders of magnitude fewer doctests than Sage (since it relies more on unit tests). Erik -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
