FutureWarning for np.full(..., non-float). #6382
Conversation
cf. discussion in numpy#6366.
|
Lgtm. @charris, what do you think? |
FutureWarning for np.full(..., non-float).
|
LGTM also. Thanks @anntzer . |
|
Doh, only looking at the backport go by do I realize that this is missing a test. Don't really want to delay 1.10rc2 over it, but @anntzer if you have a few minutes it'd be good to have (in master if nothing else). |
|
in test_numeric or in test_deprecations? |
|
test_deprecations Unless you'd rather do it the other way. It really doesn't matter much :-)
|
|
This behavior is confusing. http://stackoverflow.com/questions/33070449/why-does-np-full-raise-a-futurewarning In the case where the user explicitly specifies the dtype, the behavior will not be different in the future, so there should be no warning. It's only when the user does not explicitly specify the dtype as something other than |
See @rkern's comment in numpy#6382.
|
Apologies, see #6449. |
See @rkern's comment in numpy#6382.
See @rkern's comment in numpy#6382.
cf. discussion in #6366.
@seberg @njsmith @charris
Please consider whether this can make it to 1.10.rc2 to get at least some testing before the actual release.