Support for Fortran

[mikebentley15]

This is a requested feature.

We can try to have at least limited support for Fortran. I am not too good at Fortran, and would need to learn significantly in order to convert the testing framework over. However, this is a necessary feature for many potential users of FLiT.

[jjgarzella]

I'll be working on this issue as we discussed, by generalizing the concept of "compiler" to "compiler collection"

[mikebentley15]

Thank you JJ :smile:

[mikebentley15]

Today we had a meeting. Here are the notes:

Current Progress By JJ

• [#93]: Support for Fortran
• Per-compiler LD flags and switches
• Has tested the Fortran stuff
• Some problems:
  • Both changes created lots of indirection into the Makefile
  • Makefile is harder to read
    • A couple of places where makefile rules are copied and pasted (3 copies of each rule). When we add in another language, do we just resort to copying and pasting?
    • C++ rules were copied and pasted for Fortran
    • Fixed form (F77) fortran is different from free form fortran (current stuff)
    • Needs to be different compiler options for fixed form and for free form
  • Bisect is totally broken, probably

Itemize Work

• [ ] Pull request fork into PRUNERS feature branch
  • [ ] (JJ) squash some commits
  • [ ] (JJ) initiate pull request
• [ ] (Mike or Ian) Look to see if we can solve some of the ugliness problems (investigation)
• [ ] (JJ - but Mike would like a demo) Manually test the current implementation
• [ ] (Ian) Fix bisect for C++
• [ ] (Mike + Ian) fix automated tests
• [ ] (Mike) Get bisect working for Fortran as well
  • [ ] Does Fortran compilation have -fPIC flag?
• [ ] (volunteers?) Documentation
• [ ] (kick the can?) Create automated tests for Fortran support