Fork relocatable-python for native packaging

[AkechiShiro]

@viferga to package natively for macOS, we need to fork and maintain our own relocatable Python, this repo : https://github.com/gregneagle/relocatable-python

Also merging these two PRs :

• Add Universal Python testing script gregneagle/relocatable-python#26
• Use correct URL for fetching prerelease builds gregneagle/relocatable-python#19

[brillard1]

On utilzing the relocatable-python for native packaging, it successfully creates the python framework object for the target version till 3.9.9 [Listed here]

To check if the framework is universal or not, run the universal script in terminal

./python_universal_tester.sh 3.x.x [enter python version]

It produces the following output:

Using Python 3.9.9
      90 libraries (*.so and *.dylib) found in the framework; only        0 are universal!
.
.
.
Python.framework/Versions/3.9.9/Resources/Python.app/Contents/MacOS/Python is not universal!
Python.framework/Versions/Current/Python is not universal!
Python.framework/Versions/Current/bin/python3.9.9 is not universal!

The above issues arise due since the current Python framework is not built as a universal binary, which means it only supports a single architecture. This poses a problem as our users may be using different architectures, and we want our framework to be compatible with all of them. The following issues have been identified:

• Lack of support for building universal Python frameworks
• Failure to compile libraries for both arm64 and x86_64 architectures
• Libraries compiled for a specific architecture may cause issues when used on another architecture

To address these issues and build a universal Python framework, the following can be potential solutions:

• Use the ./configure --enable-universalsdk flag when building Python to ensure it is built as a universal binary.
• Use the lipo tool to merge the architecture-specific libraries into a single universal library. This can be done by adding a script to our build process that calls lipo after each library is built. [here]
• Ensure that any third-party libraries used by our framework are also built as universal binaries or, alternatively, provide separate builds for each architecture.

By following these steps, we can create a universal Python framework that is compatible with multiple architectures and can be used by a wider range of users.