ENH: Build Python3.10.0rc1 64 bit wheels on Intel. #129
Conversation
| AZURE_PYTHON_VERSION: "3.9" | ||
| MB_PYTHON_VERSION: "3.10" | ||
| MB_ML_VER: "2014" | ||
| CONFIG_PATH: "config_ilp64.sh" |
There was a problem hiding this comment.
I uploaded focal images, let's give it a try. They will not be available for 32-bit linux
| CONFIG_PATH: "config_ilp64.sh" | |
| CONFIG_PATH: "config_ilp64.sh" | |
| DOCKER_TEST_IMAGE: "multibuild/focal_{PLAT}" |
azure-pipelines.yml
Outdated
| PLAT: "i686" | ||
| MB_ML_VER: "2014" | ||
| ENV_VARS_PATH: "env_vars_32.sh" | ||
| DOCKER_TEST_IMAGE: "multibuild/xenial_{PLAT}" |
There was a problem hiding this comment.
Since 32-bit is not really supported anymore in Ubuntu, just test with the manylinux image
| DOCKER_TEST_IMAGE: "multibuild/xenial_{PLAT}" |
There was a problem hiding this comment.
AFAICT, this doesn't actually test with the manylinux image.
There was a problem hiding this comment.
The tests are currently run via this line
+ docker run --rm -e PYTHON_VERSION=3.7 -e MB_PYTHON_VERSION=3.7 \
-e UNICODE_WIDTH=32 -e CONFIG_PATH= -e WHEEL_SDIR=wheelhouse \
-e MANYLINUX_URL= -e 'TEST_DEPENDS=pytest hypothesis cffi pytz' \
-v /home/vsts/work/1/s:/io multibuild/xenial_i686 /io/multibuild/docker_test_wrap.sh
which uses multibuild/xenial_i686. I was intending for it to use the manylinux image instead by removing the line.
There was a problem hiding this comment.
Without the line the build still worked, but there was no 3.10 for the test, the docker image defaults to the multibuild xenial version for testing. I just commented out the 32 bit and ARM64 builds. I assume both will get fixed at some point.
There was a problem hiding this comment.
I assume both will get fixed at some point.
The arm64 image for focal now exists. On i686 we are stuck testing on the manylinux image since 32-bit images do not exist for current Ubuntu distributions.
499130b to
2264bc9
Compare
|
Merging. All the tests have passed, but github had problems and missed the notifications. |
* mimic the approach used by NumPy wheels in: MacPython/numpy-wheels#129 * what could go wrong.. (skip Travis for now) [skip travis] [skip travis ci] [skip travis-ci] [travis skip]
* mimic the approach used by NumPy wheels in: MacPython/numpy-wheels#129 * what could go wrong.. (skip Travis for now) [skip travis] [skip travis ci] [skip travis-ci] [travis skip]
* mimic the approach used by NumPy wheels in: MacPython/numpy-wheels#129 * what could go wrong.. (skip Travis for now) [skip travis] [skip travis ci] [skip travis-ci] [travis skip]
* mimic the approach used by NumPy wheels in: MacPython/numpy-wheels#129 * what could go wrong.. (skip Travis for now) [skip travis] [skip travis ci] [skip travis-ci] [travis skip]
No description provided.