Closed snowman2 closed 12 months ago
Merging #1330 (8d82230) into main (c3e8aca) will not change coverage. Report is 2 commits behind head on main. The diff coverage is
n/a
.
@@ Coverage Diff @@
## main #1330 +/- ##
=======================================
Coverage 96.41% 96.41%
=======================================
Files 20 20
Lines 1812 1812
=======================================
Hits 1747 1747
Misses 65 65
pypy failure:
/meson.build:207:4: ERROR: Problem encountered: No BLAS library detected! Install one, or use the `allow-noblas` build option (note, this may be up to 100x slower for some linear algebra operations).
win32 failure:
Host machine cpu: x86_64
Program python found: YES (C:\Users\runneradmin\AppData\Local\Temp\cibw-run-dbhjql3e\cp39-win32\venv-test\Scripts\python.exe)
Need python for x86_64, but found x86
Run-time dependency python found: NO (tried sysconfig)
..\..\meson.build:43:12: ERROR: Python dependency not found
I think that this is a bug with numpy.
Do you want to build NumPy from source? If not, since NumPy has not released python 3.12 wheels, you need to add --pre
to the pip install
command to use the pre-release rc1 wheels from https://pypi.org/project/numpy/1.26.0rc1/#files
Although that will not solve the 3.12 32-bit windows story, you will have to wait a few days until the wheels are uploaded. I am not sure why PyPy is not finding a pre-built wheel, can you point to the log file with the failure?
And if I am already here, you should add a pin numpy<2.0
to the requirements-test.txt
to prevent inadvertantly testing with NumPy2.0 before you are ready to do so. You can get a taste of the coming changes by using the development wheels from https://anaconda.org/scientific-python-nightly-wheels/numpy/, again --pre
will be needed.
I am not sure why PyPy is not finding a pre-built wheel, can you point to the log file with the failure?
Looks like it is the PyPy 3.10 wheels failing. The 3.9 wheels pass just fine. I think it is due to numpy not providing PyPy 3.10 wheels.
For the PyPy 3.10 and i686 architectures, it seems like the build option is not working.
Action Run: https://github.com/pyproj4/pyproj/actions/runs/6211319703/job/16860392551
Relevant logs:
../../numpy/meson.build:207:4: ERROR: Problem encountered: No BLAS library detected! Install one, or use the `allow-noblas` build option (note, this may be up to 100x slower for some linear algebra operations).
...
+ /tmp/tmp.k5EpChUOv5/venv/bin/python /tmp/pip-install-h1to5_h2/numpy_4ff366ec076b4ce98ca00bd1cf8e5315/vendored-meson/meson/meson.py setup /tmp/pip-install-h1to5_h2/numpy_4ff366ec076b4ce98ca00bd1cf8e5315 /tmp/pip-install-h1to5_h2/numpy_4ff366ec076b4ce98ca00bd1cf8e5315/.mesonpy-b2pey0k1/build -Dbuildtype=release -Db_ndebug=if-release -Db_vscrt=md "-Dallow-noblas=true" --native-file=/tmp/pip-install-h1to5_h2/numpy_4ff366ec076b4ce98ca00bd1cf8e5315/.mesonpy-b2pey0k1/build/meson-python-native-file.ini
usage: meson [-h]
{setup,configure,dist,install,introspect,init,test,wrap,subprojects,rewrite,compile,devenv,env2mfile,help}
...
meson: error: unrecognized arguments: "-Dallow-noblas=true"
Due to this, disabling the builds that don't provide numpy wheels.
Looks like it is the PyPy 3.10 wheels failing. The 3.9 wheels pass just fine. I think it is due to numpy not providing PyPy 3.10 wheels.
Is that a bug or CI hiccup on numpy's end? Or intentional?
Is that a bug or CI hiccup on numpy's end? Or intentional?
Not sure at the moment ...
Is that a bug or CI hiccup on numpy's end? Or intentional?
Related #1325