Add functional tests against regular cargo behavior

[CohenArthur]

I think this should get split up in multiple PRs in order to better the review process. This issue shall become the tracking issue for the subtasks of the final PR.

After discussing it with @flip1995, it seems that running functional tests within the rust code would be best. This is what clippy does, for example here

This should be split into multiple sub-issues:

• Add valid tests application that gets run when executing cargo test Add basic test harness #26
• Add abstraction around checking the validity of a test (exit code 0) and reporting this result to the developer Add basic test harness #26
• The test harness shall only be compiled when running tests Add basic test harness #26
• Checking that the objects created by both rustc and gccrs have the same filename Check correct filename between rustc and gccrs #33
• Check for deps in target/[release|debug]/deps Check correct filename between rustc and gccrs #33
• Check for final output in target/[release|debug] #41
• Check that static libraries are not empty Check correct ELF type  #37
• Check that static libraries contain the compilation unit #45
• Check that the objects created are of the correct file type: Executable, Shared object, Static object etc using readelf for example. Check correct ELF type  #37

Once this test harness framework is done, add tests for the project

• Add functional testing to the CI Add basic test harness #26
• Test the basic projects included in the repo as is (tests/binary_project, test/static_lib, etc...) Add basic test harness #26
• Figure out how to handle testing interesting crates/repos located at a specific URL or as submodules #43
• Add test case when passing extra environment variables (tests/warning_project) #44

Since all of these PRs wouldn't make sense by themselves, I think it would be best to create them against a specific branch such as feature/functional-testing and then merge this branch into master once the test harness is complete.

[CohenArthur]

I started work on checking the correct file type for each binary produced, but noticed that gccrs and rustc have a different behavior when producing binaries #35