Updates:
Cc: [email protected] Vinzent.Steinberg [email protected]
[email protected] [email protected]
Comment #5 on issue 2160 by asmeurer: List of dependencies
http://code.google.com/p/sympy/issues/detail?id=2160
So I wasn't clear on this in me previous comment, but I actually agree that
we should remove pyglet from SymPy. I think that we need to rework the
plotting so that it works with matplotlib, pyglet, etc., all of which would
be optional external dependencies (it would automatically choose the best
installed one for the task, or you could override with a keyword
argument). Basically what you say is correct: plotting does not work for
me at all in Mac OS X (and based on some of the issues here, I would say it
doesn't work in Windows either). Unfortunately, this is a bit more work
than just typing "git rm -r sympy/thirdparty/pyglet", so it hasn't been
done yet.
Now, for mpmath, it is a different story. It is integrated pretty highly
in SymPy—virtually every SymPy function can be evalf'd. Also, I have to
agree with Ondrej in your second link there. I was the one who updated
mpmath the last two times. Last time, the update was pretty smooth. The
only thing that had to be changed (other than the standard stuff from
UpdateMpmath) was that I had to change some code in mpmath to be compatible
with Python 2.4 and 2.5 (see commit 5f4a2cf0a634f4). The previous time,
there was a fix that had to be made in Interval.evalf() that if I remember
correctly was because of an interface change to mpi in mpmath (see commit
1b86edabb9991). There's also fixes like 76f01a530bdfe which are just minor
AP
I changes which according to the update to mpmath 0.14, also had to be done
(see 9a72d817ee600). Based on Ondrej's comment in your second link, I
would say that this has probably been a common theme in the past: updates
to mpmath generally require some kind of fix either to mpmath or to SymPy.
This is just a result of the two being developed independently of one
another.
My point is that it is difficult to support more than just the most recent
version of mpmath, or rather, the most recent version of mpmath that has
bee installed in SymPy (for example, we can't even start to update to
mpmath 0.17 until we release because it drops Python 2.4 support). If we
stop bundling mpmath with SymPy, what versions do we support? Do we always
support the most recent version (which means that we have to make our
release cycles somewhat synchronized with mpmath's), or the past few
versions (which means several places in the code that are coded to work
with API changes and new features, and also a testing nightmare).
But mainly what I want to say is that you should be able to install and run
SymPy with nothing more than Python installed. You might not be able to do
advanced stuff like plotting, or have as nice of an interface if you don't
have IPython installed, but at least you can install it and use it.
By the way, I am CCing a bunch of other developers. Let's see what they
think about all this.
--
You received this message because you are subscribed to the Google Groups
"sympy-issues" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to
[email protected].
For more options, visit this group at
http://groups.google.com/group/sympy-issues?hl=en.
