Closed tmsincomb closed 4 years ago
Hey, thanks for the report, at least good to know. Unfortunately I don't have access to any MacOS, so debugging will be hard.
Just curious, what clang version is there on MacOS Mojave? I wonder if it's a MacOS difference, or a clang difference.
Another option is to edit the setup.py
file and remove the -Werror
setting; I think the "linker input unused" is just a warning, but -Werror
makes it a hard error.
You were correct about the clang difference - thanks! The issue was luckily already solved.
My current default Clang that works without LLVM installed:
> clang -v
Apple clang version 11.0.3 (clang-1103.0.32.62)
Target: x86_64-apple-darwin19.5.0
Thread model: posix
InstalledDir: /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin
I was using LLVM where the Clang version read as:
> clang -v
clang version 10.0.0
Target: x86_64-apple-darwin19.5.0
Thread model: posix
InstalledDir: /Users/tmsincomb/anaconda3/envs/py38/bin
I was able to track down the specific error in building xattr. It would seem there is a default option -platform_version that is deprecated. The people working on LLVM mention it here https://reviews.llvm.org/D74784. Although LLVM was indicated as up-to-date on both brew and anaconda it would still crash for me. A forced reinstall for either being used would fix it. Although that Clang version printout for LLVM is the same with the reinstall there are no more errors/warnings.
There was someone else with the same issue here indicating this might be a lingering issue for the MacOS - LLVM users. https://stackoverflow.com/questions/62726754/mujoco-on-macos-catalina-ld-unknown-option-platform-version-when-importing/62746029#62746029.
Thanks for clarifying, and glad to hear it was solved.
pip install pyxattr worked on Catalina 10.15.5 with python3.7.7 (and still does), but when I updated to python3.8.1 I get this error.
However testing it with MacOS Mojave with python3.8.1 I do not get an error. From what I can tell is that this is a Catalina specific issue so please feel free to close this issue, it is very likely it will work itself out.
I just wanted to bring this to your attention incase of any usefulness that would come from it.