The anaconda staging repositories have changed layout and now only
25 files are showing per page. This update uses a different URL to
grab all of the names, thus avoiding the need for paging.

There are also some cleanups and added comments to make future
maintenance easier.

[skip azp] [skip cirrus] [skip actions]

MAINT: Update download-wheels for multiple pages

We've been carrying a variant of this patch in our local copy of NumPy
for some time. The alloc cache reduces the effectiveness of
use-after-free or use-of-uninitialized memory warnings because caching
of allocations means they may be reused for a different allocation,
potentially hiding real problems.

A thin wrapper over Google's Highway SIMD library to simplify its interface.
This commit provides the implementation of that wrapper, consisting of:

- simd.hpp: Main header defining the SIMD namespaces and configuration
- simd.inc.hpp: Template header included multiple times with different namespaces

The wrapper eliminates Highway's class tags by:
- Using lane types directly which can be deduced from arguments
- Leveraging namespaces (np::simd and np::simd128) for different register widths

A README is included to guide usage and document design decisions.

  - Fix hardware/platform terminology in documentation for clarity
  - Add support for long double in template specializations
  - Add kMaxLanes constant to expose maximum vector width information
  - Follows clang formatting style for consistency with NumPy codebase.

  - Add anonymous namespace around implementation to ensure each
    translation unit gets its own constants based on local flags
  - Use HWY_LANES_CONSTEXPR for Lanes function to ensure proper
    constexpr evaluation across platforms

…size

  Replace hn::Lanes(f64) with hn::MaxLanes(f64) when defining the index array
  size to fix error C2131: "expression did not evaluate to a constant". This error
  occurs because Lanes() isn't always constexpr compatible, especially with
  scalable vector extensions. MaxLanes() provides a compile-time constant value
  suitable for static array allocation and should be used with non-scalable SIMD
  extensions when defining fixed-size arrays.

Rename Highway wrapper macros for clarity:
- NPY_SIMDX → NPY_HWY
- NPY_SIMDX_F16 → NPY_HWY_F16
- NPY_SIMDX_F64 → NPY_HWY_F64
- NPY_SIMDX_FMA → NPY_HWY_FMA

To avoids confusion with legacy SIMD macros.

Skip Highway's EMU128 in favor of NumPy's scalar implementations for
due to strict IEEE 754 floating-point compliance requirements

MAINT: Update main after 2.2.6 release.

BUG: allclose does not warn for invalid value encountered in multiply or alternativly checks for validity of tolerances

TYP: Gradual shape type defaults

…64-idiv

TYP: allow inplace division of ``NDArray[timedelta64]`` by floats

ENH: Disable the alloc cache under address and memory sanitizers

numpy#28986)

MNT: fix CI issues on main

…8629) (numpy#28918)

Bumps [larsoner/circleci-artifacts-redirector-action](https://github.com/larsoner/circleci-artifacts-redirector-action) from 1.0.0 to 1.1.0.
- [Release notes](https://github.com/larsoner/circleci-artifacts-redirector-action/releases)
- [Commits](scientific-python/circleci-artifacts-redirector-action@4e13a10...7eafdb6)

---
updated-dependencies:
- dependency-name: larsoner/circleci-artifacts-redirector-action
  dependency-version: 1.1.0
  dependency-type: direct:production
  update-type: version-update:semver-minor
...

Signed-off-by: dependabot[bot] <support@github.com>

…arsoner/circleci-artifacts-redirector-action-1.1.0

MAINT: Bump larsoner/circleci-artifacts-redirector-action from 1.0.0 to 1.1.0

- Create 2.4.0-notes.rst
- Update release.rst
- Update pavement.py
- Update pyproject.toml
- Update cversions.txt
- Update numpyconfig.h
- Delete release fragments

BEG, MAINT: Begin NumPy 2.4.0 development.

BUG: Fix workflow bug

Co-Authored-By: Daniel Bertalan <dani@danielbertalan.dev>

BUG: Fix version check in blas_utils.c

MAINT: Enable linting with ruff E501

Bumps [github/codeql-action](https://github.com/github/codeql-action) from 3.29.1 to 3.29.2.
- [Release notes](https://github.com/github/codeql-action/releases)
- [Changelog](https://github.com/github/codeql-action/blob/main/CHANGELOG.md)
- [Commits](github/codeql-action@39edc49...181d5ee)

---
updated-dependencies:
- dependency-name: github/codeql-action
  dependency-version: 3.29.2
  dependency-type: direct:production
  update-type: version-update:semver-patch
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

This seems generally confusing. I would like to make it keyword only,
but this already gives a warning when loading _very_ old pickles, meaning
I am not quite sure we should change away from a warning quickly.

We should fix things around pickling and start pickling in a way that
makes it easier to move to keyword only arguments.  (I suppose one could
detect the case of `np.dtype(obj, False, True)` and assume it is via
unpickling, but...

I am assuming that it is OK to (eventually) break unpickling these 10+ year
old files, but I am not in a rush to actually do so and go through with
the deprecation.

Signed-off-by: Sebastian Berg <sebastianb@nvidia.com>

* MAINT: Enforce ruff E501

* fix merge

…onzero`, simplify `MaskedArray.nonzero` return type (numpy#29303)

DOCS: Remove incorrect "Returns" section from `MaskedArray.sort`

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

Fix incorrect grammar in TypeError message for ufunc argument count mismatch

DEP: Give a visible warning when `align=` to dtype is a non-bool

…#29310)