Initial Structure for updating to the NewPassManager

[tpatki]

WIP.
Fixes #162.
Fixes #176.

• Add a CMake check to check for Clang 14 and above
• Rename perfflow_weave.hpp/cpp to perfflow_weave_legacy_pass.hpp/cpp
• Add an example similar to the InjectFuncPass in the perfflow_weave_new_pass.hpp/cpp
• Test the simple example
• Migrate the functionality from the legacy pass to the new pass .If applicable, also migrate common functionality across both passes (parsing, insertBefore/insertAfter) to its own header that can be shared if no other changes are needed.
• Test on Lassen with clang10 and clang14.05 (legacy and new pass)
• Split perfflow_weave_common.hpp into .hpp/.cpp
• Cleanup the modifyAnnotatedFunctions() method, make it easier to read/explain
• Clean up commented code
• Test with various clang versions on different systems
• Format code

Testing:

• Ruby : clang versions 14
• Corona: clang versions 14
• Turing: clang versions 14
• Turing: clang versions 15
• Turing: clang versions 17
• Lassen: clang versions 10
• Lassen: clang versions 14, -gcc module
• Lassen: clang versions 16, -gcc module
• Lassen: clang versions 15, non-gcc module (several C++ errors that I have been unable to resolve. Note: we don't have a clang-<>gcc-<> style module for clang15 on lassen. These have nothing to do with our code, but some incompatibility with libstdc++)
• Lassen: clang versions 17, non-gcc module (Similar errors to clang15, note that here too we don't have the clang-<>gcc-<> style module for clang17 on lassen.)

Note: Adiak related testing that builds on top of this PR has been moved to a separate branch, see #172.

[slabasan]

Could we double check these version comparisons? Is PERFFLOWASPECT_CLANG_15_NEWER supposed to be >=clang15 or >clang15?

[tpatki]

>= from what I recall from my testing (returns true of clang 15). I will test again.