1 |
Hi, |
2 |
|
3 |
Almost 2 years ago almost to the day I included a line blocking the |
4 |
installation of gmpy alongside sage in the sage ebuild. |
5 |
This was done because I had regular reports of people breaking |
6 |
sympy when both were installed at the same time - ultimately |
7 |
sage was broken too but the big symptom was an un-ability to |
8 |
build/reinstall sympy. |
9 |
|
10 |
The fundamental issue seems to have fixed in mpmath upstream |
11 |
so I lift the restriction today. |
12 |
|
13 |
But it will be back if new issue are discovered. |
14 |
|
15 |
A quick note for people wanting to use gmpy as the backend |
16 |
for mpmath/sympy. By default if both gmpy and sage are installed |
17 |
mpmath will use sage as its backend for arbitrary precision |
18 |
computation. This is a runtime behavior that cannot be altered |
19 |
by useflags. If you want to use gmpy over sage as the backend |
20 |
you need to define the variable MPMATH_NOSAGE (its value is not |
21 |
checked, just its existence). Similarly there is a MPMATH_NOGMPY |
22 |
check. |
23 |
People interested in the mechanics of the backend should look at |
24 |
/usr/lib64/python2.7/site-packages/mpmath/libmp/backend.py |
25 |
|
26 |
Cheers, |
27 |
Francois |