[Buildroot] [PATCH 2025.02.x 1/3] package/python-scipy: fix build with cython >=3.1.0

[Buildroot] [PATCH 2025.02.x 1/3] package/python-scipy: fix build with cython >=3.1.0

[Thomas Perale via buildroot]

Since Cython version 3.1.0 [1] and buildroot commit [2] python-scipy
failed to build. This has been fixed upstream with [3].

This commit backports it to fix the build with the cython present in
2025.02.x.

[1] https://github.com/cython/cython/commit/ed7e95912b6ed7d20e190fbf1aecb9f2a997d479
[2] b536caaec0 package/python-cython: bump to version 3.1.2
[3] https://github.com/scipy/scipy/commit/217748476403a1e5726a921cedb674e7ffbc952e

Fixes: b536caaec0 package/python-cython: bump to version 3.1.2
Signed-off-by: Thomas Perale <thomas.perale@mind.be>
---
 ...e-int-instead-of-long-in-cython-code.patch | 37 +++++++++++++++++++
 1 file changed, 37 insertions(+)
 create mode 100644 package/python-scipy/0002-use-int-instead-of-long-in-cython-code.patch

diff --git a/package/python-scipy/0002-use-int-instead-of-long-in-cython-code.patch b/package/python-scipy/0002-use-int-instead-of-long-in-cython-code.patch
new file mode 100644
index 0000000000..6763c8cc5e
--- /dev/null
+++ b/package/python-scipy/0002-use-int-instead-of-long-in-cython-code.patch
@@ -0,0 +1,37 @@
+From 217748476403a1e5726a921cedb674e7ffbc952e Mon Sep 17 00:00:00 2001
+From: Thomas A Caswell <tcaswell@gmail.com>
+Date: Thu, 21 Dec 2023 14:20:52 -0500
+Subject: [PATCH] MNT: use int instead of long in cython code (#19725)
+
+https://github.com/cython/cython/pull/5830 /
+d7e95912b6ed7d20e190fbf1aecb9f2a997d479 in cython removed
+long as a valid type in favor of int.
+
+Upstream: https://github.com/scipy/scipy/commit/217748476403a1e5726a921cedb674e7ffbc952e
+Signed-off-by: Thomas Perale <thomas.perale@mind.be>
+---
+ scipy/_lib/_ccallback_c.pyx | 4 ++--
+ 1 file changed, 2 insertions(+), 2 deletions(-)
+
+diff --git a/scipy/_lib/_ccallback_c.pyx b/scipy/_lib/_ccallback_c.pyx
+index 59ba74369267..6c493dac7c9c 100644
+--- a/scipy/_lib/_ccallback_c.pyx
++++ b/scipy/_lib/_ccallback_c.pyx
+@@ -58,7 +58,7 @@ def get_raw_capsule(func_obj, name_obj, context_obj):
+     elif context_obj is None:
+         context = NULL
+     else:
+-        context = PyLong_AsVoidPtr(long(context_obj))
++        context = PyLong_AsVoidPtr(int(context_obj))
+ 
+     if PyCapsule_CheckExact(func_obj):
+         capsule_name = PyCapsule_GetName(func_obj)
+@@ -70,7 +70,7 @@ def get_raw_capsule(func_obj, name_obj, context_obj):
+         if name == NULL:
+             name = capsule_name
+     else:
+-        func = PyLong_AsVoidPtr(long(func_obj))
++        func = PyLong_AsVoidPtr(int(func_obj))
+ 
+     if name == NULL:
+         name_copy = name
-- 
2.56.0

_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

[Buildroot] [PATCH 2025.02.x 2/3] package/python-scipy: fix build with GCC14

[Thomas Perale via buildroot]

When building python-scipy package with target GCC14 the following error
would occure:

../scipy/interpolate/src/_fitpackmodule.c: In function ‘fitpack_bispev’:
../scipy/interpolate/src/_fitpackmodule.c:255:10: error: assignment to ‘npy_int32 *’ {aka ‘long int *’} from incompatible pointer type ‘int *’ [-Wincompatible-pointer-types]
255 |     iwrk = (int *)(wrk + lwrk);
    |          ^

This has been fixed upstream with the patch [1] introduced in scipy
version '1.13.0'. For more information see [2].

Backport this patch into the current 2025.02.x scipy version.

[1] https://github.com/scipy/scipy/commit/fdf3b908eeb93203c1847c6fb4efd12e88c37ce5
[2] https://github.com/scipy/scipy/pull/20135

Signed-off-by: Thomas Perale <thomas.perale@mind.be>
---
 ...e-F_INT-as-int-rather-than-npy_int32.patch | 50 +++++++++++++++++++
 1 file changed, 50 insertions(+)
 create mode 100644 package/python-scipy/0003-interpolate-define-F_INT-as-int-rather-than-npy_int32.patch

diff --git a/package/python-scipy/0003-interpolate-define-F_INT-as-int-rather-than-npy_int32.patch b/package/python-scipy/0003-interpolate-define-F_INT-as-int-rather-than-npy_int32.patch
new file mode 100644
index 0000000000..f722c4ca37
--- /dev/null
+++ b/package/python-scipy/0003-interpolate-define-F_INT-as-int-rather-than-npy_int32.patch
@@ -0,0 +1,50 @@
+From 53796772ed735c1564863fc7ca8d902acb628167 Mon Sep 17 00:00:00 2001
+From: Ralf Gommers <ralf.gommers@gmail.com>
+Date: Thu, 22 Feb 2024 09:10:46 +0100
+Subject: [PATCH] MAINT: interpolate: define `F_INT` as `int` rather than
+ `npy_int32`
+
+This fixes an incompatible pointer issue that shows up as a warning in
+Windows CI jobs, and is reported to break the build with GCC 14 on
+Fedora 40 in gh-19993.
+
+Using `#define F_INT int` is done in several other submodules; this
+was the only instance of using `npy_int32`.
+
+Closes gh-19993
+
+Upstream: https://github.com/scipy/scipy/commit/fdf3b908eeb93203c1847c6fb4efd12e88c37ce5
+Signed-off-by: Thomas Perale <thomas.perale@mind.be>
+---
+ scipy/interpolate/src/_fitpackmodule.c | 4 ++--
+ scipy/interpolate/src/fitpack.pyf      | 2 +-
+ 2 files changed, 3 insertions(+), 3 deletions(-)
+
+diff --git a/scipy/interpolate/src/_fitpackmodule.c b/scipy/interpolate/src/_fitpackmodule.c
+index 0e913d1c7e43..dc79b2b28833 100644
+--- a/scipy/interpolate/src/_fitpackmodule.c
++++ b/scipy/interpolate/src/_fitpackmodule.c
+@@ -28,8 +28,8 @@ static PyObject *fitpack_error;
+ 
+ #else
+ 
+-#define F_INT npy_int32
+-#define F_INT_NPY NPY_INT32
++#define F_INT int
++#define F_INT_NPY NPY_INT
+ #define F_INT_MAX NPY_MAX_INT32
+ #if NPY_BITSOF_SHORT == 32
+ #define F_INT_PYFMT   "h"
+diff --git a/scipy/interpolate/src/fitpack.pyf b/scipy/interpolate/src/fitpack.pyf
+index 08cb0c141c98..a9535a83f479 100644
+--- a/scipy/interpolate/src/fitpack.pyf
++++ b/scipy/interpolate/src/fitpack.pyf
+@@ -16,7 +16,7 @@ python module dfitpack ! in
+ #ifdef HAVE_ILP64
+ typedef npy_int64 F_INT;
+ #else
+-typedef npy_int32 F_INT;
++typedef int F_INT;
+ #endif
+ 
+ static double dmax(double* seq, F_INT len) {
-- 
2.56.0

_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

[Buildroot] [PATCH 2025.02.x 3/3] package/python-numpy: fix meson configuration with LAPACK

[Thomas Perale via buildroot]

Building numpy with the LAPACK package enabled (like when building
python-scipy) would result in the following error during configuration:

```
numpy/meson.build:203:49: ERROR: Could not get pkg-config variable
and no default provided for <PkgConfigDependency lapack: True None>
```

This is reproducible with the following defconfig:

cat >.config <<EOF
BR2_arm=y
BR2_cortex_a7=y
BR2_TOOLCHAIN_EXTERNAL=y
BR2_TOOLCHAIN_EXTERNAL_BOOTLIN=y
BR2_PACKAGE_PYTHON3=y
BR2_PACKAGE_PYTHON_NUMPY=y
BR2_PACKAGE_LAPACK=y
EOF
make oldefconfig
make python-numpy

The variable that triggered that error got assigned a default value in
upstream commit [1] which is available in python-numpy v2.0.0, which fix
this error.

This commit backport the part of the upstream commit that assign the
value by default.

[1] https://github.com/numpy/numpy/commit/34afbb97cc9915ad5976ddc869a539e897545722

Fixes: https://gitlab.com/buildroot.org/buildroot/-/jobs/17125191093
Signed-off-by: Thomas Perale <thomas.perale@mind.be>
---
 ...implement-BLAS-LAPACK-auto-detection.patch | 77 +++++++++++++++++++
 1 file changed, 77 insertions(+)
 create mode 100644 package/python-numpy/0002-meson-implement-BLAS-LAPACK-auto-detection.patch

diff --git a/package/python-numpy/0002-meson-implement-BLAS-LAPACK-auto-detection.patch b/package/python-numpy/0002-meson-implement-BLAS-LAPACK-auto-detection.patch
new file mode 100644
index 0000000000..f7822ef852
--- /dev/null
+++ b/package/python-numpy/0002-meson-implement-BLAS-LAPACK-auto-detection.patch
@@ -0,0 +1,77 @@
+From 34afbb97cc9915ad5976ddc869a539e897545722 Mon Sep 17 00:00:00 2001
+From: Ralf Gommers <ralf.gommers@gmail.com>
+Date: Tue, 10 Oct 2023 00:27:02 +0200
+Subject: [PATCH] ENH: meson: implement BLAS/LAPACK auto-detection and many CI jobs [wheel build]
+
+This reimplements the auto-detection and many of the switches that
+`numpy.distutils` offered. Beyond that, it implements several new
+features:
+
+- Auto-detect the symbol suffix for ILP64 OpenBLAS (can be none or
+  `64_`)
+- MKL ILP64 support, threading control, and use of the layered library
+  model for MKL >=2023.0
+- FlexiBLAS support (LP64 and ILP64)
+- Support for the upcoming standard in Reference LAPACK for `_64` ILP64
+  symbol suffix convention.
+- A test suite for BLAS/LAPACK libraries, covering:
+
+    - OpenBLAS: LP64, ILP64 detected via pkg-config and with a "system
+                dependency" (i.e., custom code inside Meson)
+    - MKL: LP64, ILP64 (layered) and LP64 (SDL)
+    - Accelerate: LP64, ILP64 on macOS >=13.3
+    - FlexiBLAS: LP64, ILP64 on Fedora
+    - ATLAS (LP64, via pkg-config only)
+    - BLIS (LP64, via pkg-config only)
+    - plain libblas/liblapack (Netlib, LP64 only)
+
+The list of libraries that is tried with the default 'auto' setting
+excludes a couple of libraries, because they're either no longer
+developed (ATLAS), not mature (libflame), or can't be tested and
+may be re-added later (ArmPL, ssl2). Those libraries can still be
+quite easily used via pkg-config.
+
+The new CI jobs are running by default right now. Once things settle
+down, the plan is to disable them by default and allow triggering
+them via a `[blas ci]` command in the commit message (just like for
+wheel builds).
+
+Docs will be included in a separate PR with the pending rewrite of
+all the build/install docs. For now, the CI jobs and the
+`meson_options.txt` file serve as guidance for how to use this.
+
+Note that the test suite contains a few hacks, because of packaging
+bugs for MKL on PyPI (broken .pc files) and BLIS (missing .pc file in
+Debian).
+
+Upstream: https://github.com/numpy/numpy/commit/34afbb97cc9915ad5976ddc869a539e897545722
+[thomas: backport subset of upstream commit.
+  Only the part that fix the 2025.02.x scipy version LAPACK detection.
+
+ Fixes:
+    numpy/meson.build:203:49: ERROR: Could not get pkg-config variable
+    and no default provided for <PkgConfigDependency lapack: True None>
+]
+Signed-off-by: Thomas Perale <thomas.perale@mind.be>
+---
+ numpy/meson.build | 8 ++++----
+ 1 file changed, 4 insertions(+), 4 deletions(-)
+
+diff --git a/numpy/meson.build b/numpy/meson.build
+--- a/numpy/meson.build
++++ b/numpy/meson.build
+@@ -198,10 +198,10 @@
+   if dep.found()
+     conf_data.set(name + '_VERSION', dep.version())
+     conf_data.set(name + '_TYPE_NAME', dep.type_name())
+-    conf_data.set(name + '_INCLUDEDIR', dep.get_variable('includedir'))
+-    conf_data.set(name + '_LIBDIR', dep.get_variable('libdir'))
+-    conf_data.set(name + '_OPENBLAS_CONFIG', dep.get_variable('openblas_config'))
+-    conf_data.set(name + '_PCFILEDIR', dep.get_variable('pcfiledir'))
++    conf_data.set(name + '_INCLUDEDIR', dep.get_variable('includedir', default_value: 'unknown'))
++    conf_data.set(name + '_LIBDIR', dep.get_variable('libdir', default_value: 'unknown'))
++    conf_data.set(name + '_OPENBLAS_CONFIG', dep.get_variable('openblas_config', default_value: 'unknown'))
++    conf_data.set(name + '_PCFILEDIR', dep.get_variable('pcfiledir', default_value: 'unknown'))
+   endif
+ endforeach
+ 
-- 
2.56.0

_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

Re: [Buildroot] [PATCH 2025.02.x 1/3] package/python-scipy: fix build with cython >=3.1.0

[Raphaël Mélotte via buildroot]

In reply of:
> Since Cython version 3.1.0 [1] and buildroot commit [2] python-scipy
> failed to build. This has been fixed upstream with [3].
> 
> This commit backports it to fix the build with the cython present in
> 2025.02.x.
> 
> [1] https://github.com/cython/cython/commit/ed7e95912b6ed7d20e190fbf1aecb9f2a997d479
> [2] b536caaec0 package/python-cython: bump to version 3.1.2
> [3] https://github.com/scipy/scipy/commit/217748476403a1e5726a921cedb674e7ffbc952e
> 
> Fixes: b536caaec0 package/python-cython: bump to version 3.1.2
> Signed-off-by: Thomas Perale <thomas.perale@mind.be>

Applied to 2025.02.x. Thanks

> ---
>  ...e-int-instead-of-long-in-cython-code.patch | 37 +++++++++++++++++++
>  1 file changed, 37 insertions(+)
>  create mode 100644 package/python-scipy/0002-use-int-instead-of-long-in-cython-code.patch
> 
> diff --git a/package/python-scipy/0002-use-int-instead-of-long-in-cython-code.patch b/package/python-scipy/0002-use-int-instead-of-long-in-cython-code.patch
> new file mode 100644
> index 0000000000..6763c8cc5e
> --- /dev/null
> +++ b/package/python-scipy/0002-use-int-instead-of-long-in-cython-code.patch
> @@ -0,0 +1,37 @@
> +From 217748476403a1e5726a921cedb674e7ffbc952e Mon Sep 17 00:00:00 2001
> +From: Thomas A Caswell <tcaswell@gmail.com>
> +Date: Thu, 21 Dec 2023 14:20:52 -0500
> +Subject: [PATCH] MNT: use int instead of long in cython code (#19725)
> +
> +https://github.com/cython/cython/pull/5830 /
> +d7e95912b6ed7d20e190fbf1aecb9f2a997d479 in cython removed
> +long as a valid type in favor of int.
> +
> +Upstream: https://github.com/scipy/scipy/commit/217748476403a1e5726a921cedb674e7ffbc952e
> +Signed-off-by: Thomas Perale <thomas.perale@mind.be>
> +---
> + scipy/_lib/_ccallback_c.pyx | 4 ++--
> + 1 file changed, 2 insertions(+), 2 deletions(-)
> +
> +diff --git a/scipy/_lib/_ccallback_c.pyx b/scipy/_lib/_ccallback_c.pyx
> +index 59ba74369267..6c493dac7c9c 100644
> +--- a/scipy/_lib/_ccallback_c.pyx
> ++++ b/scipy/_lib/_ccallback_c.pyx
> +@@ -58,7 +58,7 @@ def get_raw_capsule(func_obj, name_obj, context_obj):
> +     elif context_obj is None:
> +         context = NULL
> +     else:
> +-        context = PyLong_AsVoidPtr(long(context_obj))
> ++        context = PyLong_AsVoidPtr(int(context_obj))
> + 
> +     if PyCapsule_CheckExact(func_obj):
> +         capsule_name = PyCapsule_GetName(func_obj)
> +@@ -70,7 +70,7 @@ def get_raw_capsule(func_obj, name_obj, context_obj):
> +         if name == NULL:
> +             name = capsule_name
> +     else:
> +-        func = PyLong_AsVoidPtr(long(func_obj))
> ++        func = PyLong_AsVoidPtr(int(func_obj))
> + 
> +     if name == NULL:
> +         name_copy = name
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

Re: [Buildroot] [PATCH 2025.02.x 2/3] package/python-scipy: fix build with GCC14

[Raphaël Mélotte via buildroot]

In reply of:
> When building python-scipy package with target GCC14 the following error
> would occure:
> 
> ../scipy/interpolate/src/_fitpackmodule.c: In function ‘fitpack_bispev’:
> ../scipy/interpolate/src/_fitpackmodule.c:255:10: error: assignment to ‘npy_int32 *’ {aka ‘long int *’} from incompatible pointer type ‘int *’ [-Wincompatible-pointer-types]
> 255 |     iwrk = (int *)(wrk + lwrk);
>     |          ^
> 
> This has been fixed upstream with the patch [1] introduced in scipy
> version '1.13.0'. For more information see [2].
> 
> Backport this patch into the current 2025.02.x scipy version.
> 
> [1] https://github.com/scipy/scipy/commit/fdf3b908eeb93203c1847c6fb4efd12e88c37ce5
> [2] https://github.com/scipy/scipy/pull/20135
> 
> Signed-off-by: Thomas Perale <thomas.perale@mind.be>

Applied to 2025.02.x. Thanks

> ---
>  ...e-F_INT-as-int-rather-than-npy_int32.patch | 50 +++++++++++++++++++
>  1 file changed, 50 insertions(+)
>  create mode 100644 package/python-scipy/0003-interpolate-define-F_INT-as-int-rather-than-npy_int32.patch
> 
> diff --git a/package/python-scipy/0003-interpolate-define-F_INT-as-int-rather-than-npy_int32.patch b/package/python-scipy/0003-interpolate-define-F_INT-as-int-rather-than-npy_int32.patch
> new file mode 100644
> index 0000000000..f722c4ca37
> --- /dev/null
> +++ b/package/python-scipy/0003-interpolate-define-F_INT-as-int-rather-than-npy_int32.patch
> @@ -0,0 +1,50 @@
> +From 53796772ed735c1564863fc7ca8d902acb628167 Mon Sep 17 00:00:00 2001
> +From: Ralf Gommers <ralf.gommers@gmail.com>
> +Date: Thu, 22 Feb 2024 09:10:46 +0100
> +Subject: [PATCH] MAINT: interpolate: define `F_INT` as `int` rather than
> + `npy_int32`
> +
> +This fixes an incompatible pointer issue that shows up as a warning in
> +Windows CI jobs, and is reported to break the build with GCC 14 on
> +Fedora 40 in gh-19993.
> +
> +Using `#define F_INT int` is done in several other submodules; this
> +was the only instance of using `npy_int32`.
> +
> +Closes gh-19993
> +
> +Upstream: https://github.com/scipy/scipy/commit/fdf3b908eeb93203c1847c6fb4efd12e88c37ce5
> +Signed-off-by: Thomas Perale <thomas.perale@mind.be>
> +---
> + scipy/interpolate/src/_fitpackmodule.c | 4 ++--
> + scipy/interpolate/src/fitpack.pyf      | 2 +-
> + 2 files changed, 3 insertions(+), 3 deletions(-)
> +
> +diff --git a/scipy/interpolate/src/_fitpackmodule.c b/scipy/interpolate/src/_fitpackmodule.c
> +index 0e913d1c7e43..dc79b2b28833 100644
> +--- a/scipy/interpolate/src/_fitpackmodule.c
> ++++ b/scipy/interpolate/src/_fitpackmodule.c
> +@@ -28,8 +28,8 @@ static PyObject *fitpack_error;
> + 
> + #else
> + 
> +-#define F_INT npy_int32
> +-#define F_INT_NPY NPY_INT32
> ++#define F_INT int
> ++#define F_INT_NPY NPY_INT
> + #define F_INT_MAX NPY_MAX_INT32
> + #if NPY_BITSOF_SHORT == 32
> + #define F_INT_PYFMT   "h"
> +diff --git a/scipy/interpolate/src/fitpack.pyf b/scipy/interpolate/src/fitpack.pyf
> +index 08cb0c141c98..a9535a83f479 100644
> +--- a/scipy/interpolate/src/fitpack.pyf
> ++++ b/scipy/interpolate/src/fitpack.pyf
> +@@ -16,7 +16,7 @@ python module dfitpack ! in
> + #ifdef HAVE_ILP64
> + typedef npy_int64 F_INT;
> + #else
> +-typedef npy_int32 F_INT;
> ++typedef int F_INT;
> + #endif
> + 
> + static double dmax(double* seq, F_INT len) {
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

Re: [Buildroot] [PATCH 2025.02.x 3/3] package/python-numpy: fix meson configuration with LAPACK

[Raphaël Mélotte via buildroot]

In reply of:
> Building numpy with the LAPACK package enabled (like when building
> python-scipy) would result in the following error during configuration:
> 
> ```
> numpy/meson.build:203:49: ERROR: Could not get pkg-config variable
> and no default provided for <PkgConfigDependency lapack: True None>
> ```
> 
> This is reproducible with the following defconfig:
> 
> cat >.config <<EOF
> BR2_arm=y
> BR2_cortex_a7=y
> BR2_TOOLCHAIN_EXTERNAL=y
> BR2_TOOLCHAIN_EXTERNAL_BOOTLIN=y
> BR2_PACKAGE_PYTHON3=y
> BR2_PACKAGE_PYTHON_NUMPY=y
> BR2_PACKAGE_LAPACK=y
> EOF
> make oldefconfig
> make python-numpy
> 
> The variable that triggered that error got assigned a default value in
> upstream commit [1] which is available in python-numpy v2.0.0, which fix
> this error.
> 
> This commit backport the part of the upstream commit that assign the
> value by default.
> 
> [1] https://github.com/numpy/numpy/commit/34afbb97cc9915ad5976ddc869a539e897545722
> 
> Fixes: https://gitlab.com/buildroot.org/buildroot/-/jobs/17125191093
> Signed-off-by: Thomas Perale <thomas.perale@mind.be>

Applied to 2025.02.x. Thanks

> ---
>  ...implement-BLAS-LAPACK-auto-detection.patch | 77 +++++++++++++++++++
>  1 file changed, 77 insertions(+)
>  create mode 100644 package/python-numpy/0002-meson-implement-BLAS-LAPACK-auto-detection.patch
> 
> diff --git a/package/python-numpy/0002-meson-implement-BLAS-LAPACK-auto-detection.patch b/package/python-numpy/0002-meson-implement-BLAS-LAPACK-auto-detection.patch
> new file mode 100644
> index 0000000000..f7822ef852
> --- /dev/null
> +++ b/package/python-numpy/0002-meson-implement-BLAS-LAPACK-auto-detection.patch
> @@ -0,0 +1,77 @@
> +From 34afbb97cc9915ad5976ddc869a539e897545722 Mon Sep 17 00:00:00 2001
> +From: Ralf Gommers <ralf.gommers@gmail.com>
> +Date: Tue, 10 Oct 2023 00:27:02 +0200
> +Subject: [PATCH] ENH: meson: implement BLAS/LAPACK auto-detection and many CI jobs [wheel build]
> +
> +This reimplements the auto-detection and many of the switches that
> +`numpy.distutils` offered. Beyond that, it implements several new
> +features:
> +
> +- Auto-detect the symbol suffix for ILP64 OpenBLAS (can be none or
> +  `64_`)
> +- MKL ILP64 support, threading control, and use of the layered library
> +  model for MKL >=2023.0
> +- FlexiBLAS support (LP64 and ILP64)
> +- Support for the upcoming standard in Reference LAPACK for `_64` ILP64
> +  symbol suffix convention.
> +- A test suite for BLAS/LAPACK libraries, covering:
> +
> +    - OpenBLAS: LP64, ILP64 detected via pkg-config and with a "system
> +                dependency" (i.e., custom code inside Meson)
> +    - MKL: LP64, ILP64 (layered) and LP64 (SDL)
> +    - Accelerate: LP64, ILP64 on macOS >=13.3
> +    - FlexiBLAS: LP64, ILP64 on Fedora
> +    - ATLAS (LP64, via pkg-config only)
> +    - BLIS (LP64, via pkg-config only)
> +    - plain libblas/liblapack (Netlib, LP64 only)
> +
> +The list of libraries that is tried with the default 'auto' setting
> +excludes a couple of libraries, because they're either no longer
> +developed (ATLAS), not mature (libflame), or can't be tested and
> +may be re-added later (ArmPL, ssl2). Those libraries can still be
> +quite easily used via pkg-config.
> +
> +The new CI jobs are running by default right now. Once things settle
> +down, the plan is to disable them by default and allow triggering
> +them via a `[blas ci]` command in the commit message (just like for
> +wheel builds).
> +
> +Docs will be included in a separate PR with the pending rewrite of
> +all the build/install docs. For now, the CI jobs and the
> +`meson_options.txt` file serve as guidance for how to use this.
> +
> +Note that the test suite contains a few hacks, because of packaging
> +bugs for MKL on PyPI (broken .pc files) and BLIS (missing .pc file in
> +Debian).
> +
> +Upstream: https://github.com/numpy/numpy/commit/34afbb97cc9915ad5976ddc869a539e897545722
> +[thomas: backport subset of upstream commit.
> +  Only the part that fix the 2025.02.x scipy version LAPACK detection.
> +
> + Fixes:
> +    numpy/meson.build:203:49: ERROR: Could not get pkg-config variable
> +    and no default provided for <PkgConfigDependency lapack: True None>
> +]
> +Signed-off-by: Thomas Perale <thomas.perale@mind.be>
> +---
> + numpy/meson.build | 8 ++++----
> + 1 file changed, 4 insertions(+), 4 deletions(-)
> +
> +diff --git a/numpy/meson.build b/numpy/meson.build
> +--- a/numpy/meson.build
> ++++ b/numpy/meson.build
> +@@ -198,10 +198,10 @@
> +   if dep.found()
> +     conf_data.set(name + '_VERSION', dep.version())
> +     conf_data.set(name + '_TYPE_NAME', dep.type_name())
> +-    conf_data.set(name + '_INCLUDEDIR', dep.get_variable('includedir'))
> +-    conf_data.set(name + '_LIBDIR', dep.get_variable('libdir'))
> +-    conf_data.set(name + '_OPENBLAS_CONFIG', dep.get_variable('openblas_config'))
> +-    conf_data.set(name + '_PCFILEDIR', dep.get_variable('pcfiledir'))
> ++    conf_data.set(name + '_INCLUDEDIR', dep.get_variable('includedir', default_value: 'unknown'))
> ++    conf_data.set(name + '_LIBDIR', dep.get_variable('libdir', default_value: 'unknown'))
> ++    conf_data.set(name + '_OPENBLAS_CONFIG', dep.get_variable('openblas_config', default_value: 'unknown'))
> ++    conf_data.set(name + '_PCFILEDIR', dep.get_variable('pcfiledir', default_value: 'unknown'))
> +   endif
> + endforeach
> + 
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot