numpy 1.10: histo test failure about sherpa

[DougBurke]

Actually, it appears that this particular change isn't that much of a problem (or our tests are oblivious to it). However, I've edited in a note about the known test failure.

[DougBurke]

How is this fixed by #109 - which only involves matplotlib?

[olaurino]

The bug (and the fix) is in sherpa/sherpa-test-data@b06d88d

I had to include this fix as otherwise the combination matplotlib 1.5 numpy 1.10 would have triggered failures in the tests.

However, I forgot to merge the relative PR in sherpa-test-data, which may be a problem.

[olaurino]

Ok, that branch is merged, it should be all good.

[DougBurke]

Aha, that makes sense (re the problem with matplotlib version 1.10 actually being in the test code rather than sherpa code itself). I'll re-close this now you've merged in the sherpa-test-data changes.