Name already in use

TYP: Spelling alignment for array flag literal

DOC: Adopt a harmonic color scheme with the dark mode of pydata-sphinx-theme

Co-authored-by: Inessa Pawson <inessapawson@gmail.com>
Co-authored-by: Pamphile Roy <roy.pamphile@gmail.com>

MAINT: Update versioneer 0.19 → 0.26

MAINT,DOC: Revert "MAINT: fix typo in f2c_lapack.c"

This reverts commit 5d6b8b0.

python-versioneer/python-versioneer#342

We patch the LICENSE file for both sdist and wheel releases, making them
all "dirty", i.e., containing files that have not been committed. Having
"dirty" in the product name is bad marketing and the versioneer tool
does not have an option or style that will omit that bit of information,
so patch the versioneer files to make that tag an empty string.

Keep classic vendored mode.

This update will shut up this codespell warning:
	unparseable ==> unparsable

MAINT: fix typo in f2c_lapack.c

inifinity -> infinity

MAINT: always use sys.base_prefix, never use sys.real_prefix

This pull request add the changelog for #22046.

MAINT: Update main after 1.23.4 release.

`sys.base_prefix` was introduced in Python 3.3, so it will always be
available in supported versions of Python >= 3.3:
	https://docs.python.org/3/library/sys.html#sys.base_prefix

As a result, `sys.real_prefix` will never be used in supported versions
of Python >= 3.3. Besides, it has been removed from recent versions of
virtualenv:
	pypa/virtualenv#1622

DOC: Remove documentation specific to Python 2

MAINT: Remove PyCObject from the SWIG interface

TST, BLD: Fix failing aarch64 wheel builds.

…nversion

DEP: Deprecate conversion of out-of-bound Python integers

The aarch64 wheel build tests are failing with OOM. The new test for
complex128 dot for huge vectors is responsible as the useable memory
is incorrectly determined and the check for sufficient memory fails.
The fix here is to define the `NPY_AVAILABLE_MEM="4 GB"` environment
variable before the test call in `cibw_test_command.sh`.

This PR is partially addressing #22269
Added examples for ma.round and ma.round_

Co-authored-by: Matti Picus <matti.picus@gmail.com>

Addresses #22069. As discussed in the existing issue @seberg stated that it would be best to support explicitly defined like=None when creating arange objects.

Extended to all other array_creation methods which did not currently support like=None.

Closes gh-22069

DOC: Replace CObject with Capsule consistently

Co-authored-by: Matti Picus <matti.picus@gmail.com>

DOC: Fix title level for release note improvements

…x-theme (#22408)