| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
| |
* Minor improvements in Array API docs
* Directory restruture
|
| | |
|
| |
|
|
|
|
|
|
|
| |
Instances remain for NumpyVersion and Numpy.rec.fromarrays that are
references to code.
Release notes were left unchanged.
see issue #7986
|
| |
|
|
|
|
|
| |
Examples
:cdata: -> :c:data:
.. cfunction:: -> .. c:function::
|
| |
|
|
|
|
| |
Caused all long longs to be required native only in memoryviews.
NPY_SIZEOF_LONG_LONG never existed, replace documentation with existing
NPY_SIZEOF_LONGLONG and code with a sizeof(npy_longlong).
|
| |
|
|
| |
Now is as good a time as any with open PR's at a low.
|
| |
|
|
|
|
|
|
|
| |
I managed to screw them up: they did not actually mimic the gblic
endian.h behavior, that is NPY_BIG_ENDIAN and NPY_LITTLE_ENDIAN should
always be defined, and endianness should be detected by comparison with
NPY_BYTE_ORDER. This needs to be fixed because the behavior of the NPY_
macros was different depending on whether endian.h was available or not.
Let's hope nobody depended on it...
|
| | |
|
| | |
|
| | |
|
| |
|
