libpython3.8m.so should not have m suffix#2278
libpython3.8m.so should not have m suffix#2278AndreMiras merged 1 commit intokivy:developfrom davidhewitt:patch-1
libpython3.8m.so should not have m suffix#2278Conversation
|
Thanks, and sorry for the delay getting back to you - it seems the p4a developers have been otherwise occupied recently. The basic change looks good to me, but I think it will also be necessary to update the reference to libpython3.8m in pythonforandroid/bootstraps/common/build/src/main/java/org/kivy/android/PythonUtil.java. That's the only other thing I can think of offhand though. |
|
Thanks, I've fixed that and a few other places in the codebase which appeared relevant. CI failures seem unrelated - download error is presumably spurious and the travis issue looks to be kivy/buildozer#1144 |
|
ping @inclement is there anything I can do to help see this merged? |
AndreMiras
left a comment
There was a problem hiding this comment.
LGTM thanks for looking into it, addressing inclement comments and squashing everything 💪
Yeah dev have been pretty busy lately so thanks for your efforts and patience
Hi,
I'm helping a user to diagnose a downstream issue PyO3/pyo3#1077 when compiling a Rust application using p4a. In the linked issue is a lot of background and the OP also contains a link to a minimal example.
Their build with p4a targeting Python 3.8 was failing with link errors. Their workaround was to manually link to
libpython3.8m.so.However, I understood the
msuffix was removed with Python 3.8 after pymalloc builds became ABI-compatible with the mainline: https://docs.python.org/3/whatsnew/3.8.html#build-and-c-api-changesSo I think that
libpython3.8m.sois a non-standard name for libpython, and propose here themsuffix is dropped by p4a.I am very new to the p4a codebase, so please give me a heads-up for any other sections of the code that should be adjusted for this and I'm happy to round off this PR.