[sage-devel] Re: Brainstorming about Sage dependencies from system packages

2017-07-31 Thread vdelecroix
Hi Eric, Currently at a workshop in Leiden [1] we figured out one another possible use case for your proposal. Some people does develop PARI/GP in parallel of Sage. One simple way to have a testing environment would be to have: * a git repo for PARI/GP * a git repo for SAGE * telling SAGE to

[sage-devel] Re: Brainstorming about Sage dependencies from system packages

2017-05-31 Thread David Leach
Hi, A lot of issues here, but can I suggest maybe you take a look at using the Guix functional package manager as a way of keeping track of your dependency tree and for example using different versions of python concurrently etc. It just might be a great help in development with a lot

[sage-devel] Re: Brainstorming about Sage dependencies from system packages

2017-05-27 Thread Simon King
Hi, On 2017-05-27, Jeroen Demeyer wrote: > Personally, I find it more intuitive if "sage -i PKGNAME" would > unconditionally install the Sage package PKGNAME, even if PKGNAME was > detected as system package. I find it more intuitive if "sage -f PKGNAME" would install

[sage-devel] Re: Brainstorming about Sage dependencies from system packages

2017-05-26 Thread Kwankyu Lee
I wonder if the recent discussion of possible schemes about how to build Sage for python 2 and 3 (one may propose (or object to) that Sage should be built to be runnable both for python2 and python3) could be related with your proposal. Is it simply orthogonal with your proposal or somehow