Travis adding py3.5 #292
Travis adding py3.5 #292
Conversation
|
I expect some fails as there isn't numpy 1.10 on conda for python 3.3 and 2.6. I'll add a temporary commit to have these tested against 1.9 instead, and we'll update again once there will be builds available in |
|
To be honest, I think we can probably drop support for Python 2.6 - I know that in the core package, people want to do this pretty slowly (i.e. announce it one release ahead, etc.) but in affiliated packages like this, I think we can just support when we decide. |
…icter type cast rules of numpy 1.10
|
+1 for dropping python 2.6. |
| @@ -94,7 +94,7 @@ apertures: | |||
| from photutils import datasets, daofind, aperture_photometry, CircularAperture | |||
| hdu = datasets.load_star_image() | |||
| image = hdu.data[500:700, 500:700] | |||
| image -= np.median(image) | |||
| image = image.astype(float) - np.median(image) | |||
There was a problem hiding this comment.
I wonder whether it would clearer to leave this line as before and instead do image = hdu.data[...].astype(float)
… available for it in conda. Revoke this commit once it becames available
bsipocz
force-pushed
the
travis_adding_py3.5
branch
from
efe25b9
to
bdcdac0
Compare
…ced bug fix section up to 0.2 [skip ci]
|
OK, so I've removed 2.6 support and updated CHANGES, too. However updating the appveyor was a bit over ambitious, and now it's failing with python 3.5&np 1.10. |
This PR updates the matrix to test against py 3.5 and numpy 1.10, too.