Re: [sage-devel] Build failure OS X sage-6.10.beta7 numpy-1.10

2015-12-10 Thread François Bissey
I suspect you are hit by http://trac.sagemath.org/19679 Francois On 12/11/15 09:41, Ian Hoffman wrote: You were correct, I added the SAGE_ATLAS_ARCH to my environment for the build. It works better now, but now I get a different failure when building some of the documentation. Appears to have t

Re: [sage-devel] Build failure OS X sage-6.10.beta7 numpy-1.10

2015-12-10 Thread Ian Hoffman
You were correct, I added the SAGE_ATLAS_ARCH to my environment for the build. It works better now, but now I get a different failure when building some of the documentation. Appears to have trouble with an obj file. If it help, the only environment I've set is to have it build a dmg file. Than

Re: [sage-devel] Build failure OS X sage-6.10.beta7 numpy-1.10

2015-12-09 Thread François Bissey
Works here, and your log definitely indicate ATLAS has been found. The error message seem to indicate the wrong options are passed to gfortran to link the shared object. This is probably due to some confusion in the numpy build system which doesn't expect ATLAS on OS X. Francois On 12/10/15 15:1

Re: [sage-devel] Build failure OS X sage-6.10.beta7 numpy-1.10

2015-12-09 Thread François Bissey
Have you set SAGE_ATLAS_ARCH? Doing so is definitely not supported. We prefer using the accelerate framework on OS X. I have just upgraded xcode so I will give it a try. Francois On 12/10/15 13:20, Ian Hoffman wrote: This is my first post (and first time building Sage). I'm using El Capitan wit