On Tue, Jul 22, 2008 at 9:58 AM, Alan McIntyre <alan.mcintyre@gmail.com> wrote:
On Tue, Jul 22, 2008 at 11:44 AM, Charles R Harris <charlesr.harris@gmail.com> wrote:
OK, the only other change since 1.1.0 is using np.x in the doctests, which doesn't look like a big problem. I wonder what the status of that is in 1.1.x? Alan?
All the changes I made for that were in the trunk after 1.1.1 was tagged, so the 1.1.1 test behavior should be completely under the old framework's rules. I could probably backport the np.x context to rundocs() (off the top of my head I can't remember if there was any mechanism to run all the doctests in NumPy), but it might just be easier to add "import numpy as np" to the modules containing doctests that use np.x. I can help with that if need be, just let me know.
If we do a 1.1.2 it might be handy for future backports if rundocs() has the np.x context. For the moment I'm tempted to just leave them as they don't look any worse than than they were, i.e., assuming from numpy import * Thanks, Chuck