-
-
Notifications
You must be signed in to change notification settings - Fork 10.9k
pypy3: package import fails in 1.16.0 where it passed in 1.15 #12740
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Comments
Yes, this is showing up in SciPy master branch for PyPy test matrix entry as well. |
You need to either use a nightly build of PyPy from http://buildbot.pypy.org/nightly/py3.5/pypy-c-jit-latest-linux64.tar.bz2 or pin NumPy to 1.15.4. The last released pypy3 has a bug that makes |
Please refer to numpy/numpy#12740
FYI I adapted the code from the linked NumPy PR for SciPy and things look green again over there -- thanks for the info everyone. |
* + generalize set + add generic collection + update unicodes to work with six in python 2/3 * + add missing import * + add fallback across python versions * + screwed up the except * + fix string formatting * + fix automated tests issues * + add missing types for python 2 * + add datetime_tz contract to check for datetime with timezones * + bump version * + fix issue * + fix import * + might fix imports * + add deques to collections for python 2 * + move python 3.2 to allowed failure * + trying a fix to pypy numpy/numpy#12740
This seems to be addressed, closing. |
Had to use SciPy 1.2.2 for PyPy due to lack of Python 2.7 support in SciPy 1.3.0. Also had to use NumPy 1.5.4 due to PyPy issues described here: numpy/numpy#12740
The latest Cython version is intended for `numpy >= 1.17.0` but on PyPy that version of NumPy doesn't work because of numpy/numpy#12740
This is because of numpy/numpy#12740 https://stackoverflow.com/a/54381076/1068170 Also setting the `WHEELHOUSE` env. var. directly in the container; it was previously set in Circle CI but that isn't necessary.
Uh oh!
There was an error while loading. Please reload this page.
Scikit-learn was successfully testing on pypy3 with numpy 1.15 (https://circleci.com/gh/scikit-learn/scikit-learn/43203) but fails during installation when numpy is imported with version 1.16 (https://circleci.com/gh/scikit-learn/scikit-learn/43335).
This issue was introduced in 4d24bbd
The text was updated successfully, but these errors were encountered: