b3m2a1

I'm going to expunge the "-q" from the `sys.argv` with the hope that it allows things to be cleaner for people with issues in the future. Do you mind testing...

@bgatessucks I *think* I got a patch for this path as well as the others. The path finding is a bit more centralized now. Can you give it another try?...

@bgatessucks Okay I think I got a fix in place for the bad directory finding. Thanks for all your help with this. I also have a system where you can...

@bgatessucks do you have a library called MLi4 in /opt/Wolfram/Mathematica/11.3/SystemFiles/Links/MathLink/DeveloperKit/Linux-x86-64/CompilerAdditions? For me it's MLi4.a, but setup.py doesn't care about the extension. The fact that it got through the gcc call...

I tried to rewrite the dependency finder so that it doesn't require a file called MLi4 but instead looks for the highest-version library files in the directory. Let me know...

Ah that could entirely be the issue then. Any chance the current version (1.0.6) builds? I added more intelligent path-finding so it ought to have found either `libML64i4` or `libML32i4`...

That's very odd...so the library compiled but gave that warning? Could you show me the compiler output and what's in ~/.Mathematica/Paclets/Repository/PJLink-1.0.6/PJLink/PJLinkNativeLibrary? I think it must be linking against the .so...

Implementation of this has begun but isn't done and the design isn't finalized. See the `PythonObject` system, though.

That's a resource finding bug that I'll patch when I have the time and energy. On the other hand this can be used cleanly and easily and should be considered...

This is a hard one. I will need to think about how to handle it...