Comment #33 on issue 2482 by vlada.pe...@gmail.com: Stop bundling mpmath
http://code.google.com/p/sympy/issues/detail?id=2482
Some minor points:
* py.test: I was under the impression that our testing framework is a fork
of an old version of py.test. Anyway, the argument is that if we had used
the upstream py.test we would've had more features: it's much, much more
informative on errors and the above-mentioned JUnitXML would allow Jenkins
to display results in a nicer fashion (I don't know how exactly this looks,
though, so it might not be THAT important). Anyway, you can try py.test
with "git checkout rlamy/pybench", though the support isn't perfect yet.
* Testing mpmath: You are right; mpmath gets a lot more test coverage with
us, though there's nothing really stopping us from running mpmath tests
even if it's unbundled (they'll be there anyway). They are also quite
short, so I think we could also easily add them to Jenkins. And it's stable
in the sense that there have been a total of 13 commits since the latest
release (mostly docs and compatibility with gmpy2). This is the first
argument in favor of bundling that I agree with, though it can be
sidestepped quite easily (more easily than patching Distribute, that's for
sure). [I checked the logs too, the last mpmath error was more than a year
ago, when you ported to 0.15]
* Porting advantages: We could also specify our optional dependencies
(pyglet, matplotlib?) if using Distribute, I don't think Distutils supports
this. Minor, but eh. The other major advantage of Distribute (per Lennart's
book, Python 3 Porting) is easier testing, but again, I suppose we already
cover that with Tox.
* Distutils2: Yes, it will be the eventual standard. However, it's just
going to be a third party module for py25-32, so that's still an
install-time dependency, just like Distribute would be now.
---
All that said though, unless one of you wants to raise this discussion on
the mailing list (not that I expect different results), I give up. :)
I'll write a bin/2to3 script that will copy the source to a sympy-py3k
directory, run 2to3 on it, and then copy over the mpmath code (note that
this also implies a doc-py3k dir, as some stuff there needs conversion
too). This is going to be fragile by definition, but that's just the way it
is.
--
You received this message because you are subscribed to the Google Groups
"sympy-issues" group.
To post to this group, send email to sympy-issues@googlegroups.com.
To unsubscribe from this group, send email to
sympy-issues+unsubscr...@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/sympy-issues?hl=en.
