raulcd commented on issue #50648: URL: https://github.com/apache/arrow/issues/50648#issuecomment-5089062376
Mangling or other techniques would still load two full copies of Arrow in the process, so two memory pool registries, two compute function registries, etcetera. I think the only sensible solution for a use case like that is that a C++ library shipping Python bindings that interop with pyarrow *has* to load the same Arrow the wheels ships, via pyarrow.get_include() / get_library_dirs() / create_library_symlinks. I don't know the exact use case but this working previously seems to have been just a coincidence rather than something supported. Can't you avoid linking system libarrow.so (or not adding it into the wheel) and on import to load the libarrow.so provided? cc @pitrou -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
