Hi Daniel, On 2014-11-02 19:39, Stefan Schwarzer wrote: >> There are certianly some rough edges with FiPy and >> Python 3 that we need to address. > > The first thing I'd like to see
Oops. What I wanted to say: The first thing I'd like to see is to be able to install FiPy directly from PyPI in a plain virtualenv. That would already remove quite a few hurdles. I suppose many people will try the installation, see some `SyntaxError` (as I did) and assume Python 3 isn't supported at all. `INSTALLATION.txt` seems to indicate the same, as the information on Python 3 support is kind of "hidden" in `documentation/USAGE.txt`. _If_ it's an option to drop support for Python 2.4 and 2.5, I'd recommend using the same source code for Python 2.x and 3.x, without involving `2to3` (or rather once). Since you can already create a working version of FiPy with `2to3`, a one-time translation and after that maintaining the resulting source code shouldn't be a problem. You might be able to support older Python versions (2.4 and 2.5) with the same source code if you use the `six` library, but this can be rather awkward. You can also find a few tips on the common source code approach in my talk from EuroPython 2014: http://www.youtube.com/watch?v=9vNr_ZzZZAk http://sschwarzer.com/download/python_2_3_europython2014.pdf :-) Aside from the Python 3 compatibility, there seem to be some issues with the setup script. See the start of the thread. ;-) Best regards, Stefan _______________________________________________ fipy mailing list [email protected] http://www.ctcms.nist.gov/fipy [ NIST internal ONLY: https://email.nist.gov/mailman/listinfo/fipy ]
