Mistral Contrastin

This is a problem with the parsing dependency https://github.com/camfort/fortran-src. It doesn't yet handle preprocessor statements (apart from native Fortran INCLUDE statements). One way to get around the problem is to...

@dorchard So we had this discussion before. The question boiled down to (a) how do we report meaningful results to the user if all macros are expanded? (b) preprocessing in...

We probably want to do interprocedural analysis instead? (@dorchard)

Yeah, that's what he expects. `yval` is only a field name. Consider the following program ```fortran program hey type type1 integer, dimension(10) :: array end type type1 type type2 integer,...

@GuiltyDolphin Why is our parsing wrong? Regardless the parsing is wrong or not, this bug is not related to parsing. It is related to grabbing wrong AST node (too deep)...

Hmm, yes. I see your point. You might find it difficult to unambiguously parse such expressions, but fingers crossed. Though that should be made an issue in fortran-src not in...

This probably relates to #83.

Yeah, but this should be very low priority in list of things to do.

@dorchard what do you think?

That's what I initially thought but I think this is actually a reasonably organic access pattern (notice the diagonal progression, which is plausible for matrix operations).