On 2024/08/27 17:03, Stuart Henderson wrote:
> The extra patches below fix things so tests can run on 3.11 (remove
> NO_TEST from Makefile too). "make test" on amd64 Py 3.11 results below
> the diff (summary: 31 failed, 35649 passed, 2372 skipped, 1305
> deselected, 33 xfailed, 1 xpassed, 21 warnings in 207.79s"; I'll run
> it on aarch64 when my machine's free and reply if there are many
> differences. Maybe i386 too when the current bulk finishes (a day or
> two).

Build fails early on aarch64. The meson full log file referenced does
not exist.


===> Building from scratch py3-numpy-1.26.4
===> py3-numpy-1.26.4 depends on: gas-* -> gas-2.31.1p0
===> py3-numpy-1.26.4 depends on: meson-python-* -> meson-python-0.16.0p1
===> py3-numpy-1.26.4 depends on: py3-cython-* -> py3-cython-3.0.10v0
===> py3-numpy-1.26.4 depends on: g95->=8,<9 -> g95-8.4.0p23
===> py3-numpy-1.26.4 depends on: python->=3.11,<3.12 -> python-3.11.9p3
===> py3-numpy-1.26.4 depends on: py3-build-* -> py3-build-1.2.1p1
===> py3-numpy-1.26.4 depends on: py3-installer-* -> py3-installer-0.7.0p0
===> py3-numpy-1.26.4 depends on: gcc->=8,<9 -> gcc-8.4.0p23
===> py3-numpy-1.26.4 depends on: gcc-libs->=8,<9 -> gcc-libs-8.4.0p23
===> py3-numpy-1.26.4 depends on: cblas-* -> cblas-1.0p7
===> py3-numpy-1.26.4 depends on: lapack-* -> lapack-3.8.0p2
===>  Verifying specs:  python3.11 m pthread gfortran>=8 blas cblas lapack gfortran>=8
===>  found python3.11.0.0 m.10.1 pthread.27.1 gfortran.8.0 blas.2.1 cblas.1.1 lapack.7.1
===>  Checking files for py3-numpy-1.26.4
`/usr/ports/distfiles/numpy-1.26.4.tar.gz' is up to date.
>> (SHA256) numpy-1.26.4.tar.gz: OK
===>  Extracting for py3-numpy-1.26.4
===>  Patching for py3-numpy-1.26.4
===>   Applying OpenBSD patch patch-numpy__pytesttester_py
Hmm...  Looks like a unified diff to me...
The text leading up to this was:
--------------------------
|1.26.4 has removed numpy.distutils but tests still reference it
|
|Index: numpy/_pytesttester.py
|--- numpy/_pytesttester.py.orig
|+++ numpy/_pytesttester.py
--------------------------
Patching file numpy/_pytesttester.py using Plan A...
Hunk #1 succeeded at 135.
done
===>   Applying OpenBSD patch patch-numpy_core_setup_common_py
Hmm...  Looks like a unified diff to me...
The text leading up to this was:
--------------------------
|disable backtrace for now. maybe possible to enable now that we have
|libexecinfo in base, but needs checking to make sure we don't end up
|with unresolved symbols.
|
|Index: numpy/core/setup_common.py
|--- numpy/core/setup_common.py.orig
|+++ numpy/core/setup_common.py
--------------------------
Patching file numpy/core/setup_common.py using Plan A...
Hunk #1 succeeded at 147.
done
===>   Ignoring patchfile patch-numpy_core_src_common_simd_neon_memory_h.orig
===>   Applying OpenBSD patch patch-numpy_distutils_command_build_src_py
Hmm...  Looks like a unified diff to me...
The text leading up to this was:
--------------------------
|fix build of other packages (e.g. py-scipy) in some cases (e.g. when
|WRKOBJDIR has a trailing slash)
|
|Index: numpy/distutils/command/build_src.py
|--- numpy/distutils/command/build_src.py.orig
|+++ numpy/distutils/command/build_src.py
--------------------------
Patching file numpy/distutils/command/build_src.py using Plan A...
Hunk #1 succeeded at 366.
done
===>   Applying OpenBSD patch patch-numpy_distutils_site_cfg
Hmm...  Looks like a unified diff to me...
The text leading up to this was:
--------------------------
|--- numpy/distutils/site.cfg.orig	Fri Feb 13 15:41:03 2009
|+++ numpy/distutils/site.cfg	Fri Feb 13 15:41:47 2009
--------------------------
(Creating file numpy/distutils/site.cfg...)
Patching file numpy/distutils/site.cfg using Plan A...
Empty context always matches.
Hunk #1 succeeded at 1.
done
===>   Applying OpenBSD patch patch-numpy_tests_test_public_api_py
Hmm...  Looks like a unified diff to me...
The text leading up to this was:
--------------------------
|1.26.4 has removed numpy.distutils but tests still reference it
|
|Index: numpy/tests/test_public_api.py
|--- numpy/tests/test_public_api.py.orig
|+++ numpy/tests/test_public_api.py
--------------------------
Patching file numpy/tests/test_public_api.py using Plan A...
Hunk #1 succeeded at 159.
Hunk #2 succeeded at 240.
Hunk #3 succeeded at 335.
Hunk #4 succeeded at 380.
done
===>   Applying OpenBSD patch patch-numpy_typing_tests_test_isfile_py
Hmm...  Looks like a unified diff to me...
The text leading up to this was:
--------------------------
|1.26.4 has removed numpy.distutils but tests still reference it
|
|Index: numpy/typing/tests/test_isfile.py
|--- numpy/typing/tests/test_isfile.py.orig
|+++ numpy/typing/tests/test_isfile.py
--------------------------
Patching file numpy/typing/tests/test_isfile.py using Plan A...
Hunk #1 succeeded at 21.
done
===>   Applying OpenBSD patch patch-pyproject_toml
Hmm...  Looks like a unified diff to me...
The text leading up to this was:
--------------------------
|Index: pyproject.toml
|--- pyproject.toml.orig
|+++ pyproject.toml
--------------------------
Patching file pyproject.toml using Plan A...
Hunk #1 succeeded at 2.
done
===>   Ignoring patchfile patch-pyproject_toml.orig
===>  Compiler link: gcc -> env  CCACHE_DIR=/usr/obj/ports/.ccache  CCACHE_SLOPPINESS=pch_defines,time_macros  ccache /usr/local/bin/egcc
===>  Compiler link: cc -> env  CCACHE_DIR=/usr/obj/ports/.ccache  CCACHE_SLOPPINESS=pch_defines,time_macros  ccache /usr/local/bin/egcc
===>  Compiler link: gfortran -> env  CCACHE_DIR=/usr/obj/ports/.ccache  CCACHE_SLOPPINESS=pch_defines,time_macros  ccache /usr/local/bin/egfortran
===>  Compiler link: clang -> env  CCACHE_DIR=/usr/obj/ports/.ccache  CCACHE_SLOPPINESS=pch_defines,time_macros  ccache /usr/bin/clang
===>  Compiler link: clang++ -> env  CCACHE_DIR=/usr/obj/ports/.ccache  CCACHE_SLOPPINESS=pch_defines,time_macros  ccache /usr/bin/clang++
===>  Compiler link: c++ -> env  CCACHE_DIR=/usr/obj/ports/.ccache  CCACHE_SLOPPINESS=pch_defines,time_macros  ccache /usr/bin/c++
===>  Generating configure for py3-numpy-1.26.4
===>  Configuring for py3-numpy-1.26.4
===>  Building for py3-numpy-1.26.4
cp -f /pobj/py-numpy-1.26.4-python3/numpy-1.26.4/numpy/distutils/site.cfg /pobj/py-numpy-1.26.4-python3/numpy-1.26.4/site.cfg
ln -sf /usr/local/bin/gas /pobj/py-numpy-1.26.4-python3/bin/as
* Getting build dependencies for wheel...
* Building wheel...
+ /usr/local/bin/python3.11 /usr/obj/ports/py-numpy-1.26.4-python3/numpy-1.26.4/vendored-meson/meson/meson.py setup /usr/obj/ports/py-numpy-1.26.4-python3/numpy-1.26.4 /usr/obj/ports/py-numpy-1.26.4-python3/numpy-1.26.4/.mesonpy-4p6cmqhr -Dbuildtype=release -Db_ndebug=if-release -Db_vscrt=md --native-file=/usr/obj/ports/py-numpy-1.26.4-python3/numpy-1.26.4/.mesonpy-4p6cmqhr/meson-python-native-file.ini
The Meson build system
Version: 1.2.99
Source dir: /usr/obj/ports/py-numpy-1.26.4-python3/numpy-1.26.4
Build dir: /usr/obj/ports/py-numpy-1.26.4-python3/numpy-1.26.4/.mesonpy-4p6cmqhr
Build type: native build
Project name: NumPy
Project version: 1.26.4

../meson.build:1:0: ERROR: Unknown compiler(s): [['cc']]

A full log can be found at /usr/obj/ports/py-numpy-1.26.4-python3/numpy-1.26.4/.mesonpy-4p6cmqhr/meson-logs/meson-log.txt

ERROR Backend subprocess exited when trying to invoke build_wheel
*** Error 1 in . (/usr/ports/lang/python/python.port.mk:343 'do-build': @: ; if [ -e /pobj/py-numpy-1.26.4-python3/numpy-1.26.4/pyproject.to...)
*** Error 2 in . (/usr/ports/infrastructure/mk/bsd.port.mk:3057 '/pobj/py-numpy-1.26.4-python3/.build_done': @cd /usr/ports/math/py-numpy &&...)
*** Error 2 in /usr/ports/math/py-numpy (/usr/ports/infrastructure/mk/bsd.port.mk:2704 'all': @lock=py3-numpy-1.26.4;  export _LOCKS_HELD=" ...)