Status: Accepted
Owner: asmeurer
Labels: Type-Defect Priority-Medium Integration Polynomial
New issue 2650 by asmeurer: non-determinism in integrate()
http://code.google.com/p/sympy/issues/detail?id=2650
Tomo and I just went through a lot of grief in
https://github.com/sympy/sympy/pull/573 over what appeared to be some
bizzare caching problem. In the end, it was caused by this part of
risch.py:
347 poly_part = Add(*[ poly_coeffs[i]*monomial
348 for i, monomial in enumerate(monoms) ])
The problem is that momons is a set, so the result of enumerate(monoms) is
completely arbitrary. In fact, I've see list(set) return a different
result for the same set within the *same* session before. This leads to
integrate() sometimes returning different (though still correct) results
than others. It may even lead to being able to do it in once case and not
in another (I'm not sure of this, though, and I haven't actually seen it).
In the case that I saw, it returned cos**2 in one case and -sin**2 in
another; both are valid because they differ by a constant.
We should order the monoms. Actually, earlier, I would have liked it if the
monoms were ordered in a way that made it easier to debug. I think it was
grlex that I found the most useful, though I could be wrong. But any
canonical sorting would be better than none. Will default_sort_key always
give the same thing even though the elements of monoms are Dummys?
--
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.
