Hi Sandy,
Thank you so much for your reply.
So the problem I am facing is that I am trying to install protobuf with "no binary", so basically I am trying to do
"
pip install protobuf --no-binary protobuf" which would work in a requirements.txt, however our library is packaged with setup.py. From some googling I came across this solution which indicates that the workaround is to download the python tar.gz in the install_requires. (I can't add the --no-binary flag in the the install-requires)
Does that make sense?
Maybe to give a bit more background as to why I am even trying to install protobuf with the --no-binary flag, because there may be an alternative solution to my issue -
When trying to build my environment using your suggestion of install_requires=['protobuf>=3.17.3'] , I get the following error, which seems to be caused by a conflict in numpy versions being used in the build of protobuf vs my local requirements of numpy.
ValueError: numpy.ndarray size changed, may indicate binary incompatibility. Expected 88 from C header, got 80 from PyObject.
According to this solution which I found for the same error but using a different repo, the options are to - upgrade numpy to >=1.20 (which I can't because I have dependencies with numpy<1.20) or
- to use "minimum supported numpy" version in pyproject.toml (which may be an option from the protobuf side?)
- or to install protobuf with the --no-binary tag (which is the one I am attempting)..
So I either need "minimum supported numpy" version in pyproject.toml of protobuf to be set, or to install protobuf with the --no-binary tag?
Hope you can assist.
Kind regards,
Sonja