doc/source/release/1.17.4-notes.rst
.. currentmodule:: numpy
This release contains fixes for bugs reported against NumPy 1.17.3 along with some build improvements. The Python versions supported in this release are 3.5-3.8.
Downstream developers should use Cython >= 0.29.13 for Python 3.8 support and OpenBLAS >= 3.7 to avoid errors on the Skylake architecture.
random.random_integers biased generation of 8 and 16 bit integers.np.einsum regression on Power9 and z/Linux.A total of 5 people contributed to this release. People with a "+" by their names contributed a patch for the first time.
A total of 8 pull requests were merged for this release.
#14758 <https://github.com/numpy/numpy/pull/14758>__: BLD: declare support for python 3.8#14781 <https://github.com/numpy/numpy/pull/14781>__: BUG: random: biased samples from integers() with 8 or 16 bit...#14851 <https://github.com/numpy/numpy/pull/14851>__: BUG: Fix _ctypes class circular reference. (#13808)#14852 <https://github.com/numpy/numpy/pull/14852>__: BLD: add 'apt update' to shippable#14855 <https://github.com/numpy/numpy/pull/14855>__: BUG: Fix np.einsum errors on Power9 Linux and z/Linux#14857 <https://github.com/numpy/numpy/pull/14857>__: BUG: lib: Fix histogram problem with signed integer arrays.#14858 <https://github.com/numpy/numpy/pull/14858>__: BLD: Prevent -flto from optimising long double representation...#14866 <https://github.com/numpy/numpy/pull/14866>__: MAINT: move buffer.h -> npy_buffer.h to avoid conflicts