FORTRAN (was Re: indentation)

[William B. Clodius]

Dennis Lee Bieber wrote:
> <snip>
> 
>         I believe even in the full F77, there were no real "varying"
> CHARACTER types -- the closest being "character*(*)" which only applied
> to arguments passed to a subroutine/function. The actual dimensions (*)
> were still fixed when the actual variable is declared, and the bounds
> were then passed along to the subroutine/function as part of the calling
> mechanism (IE, not just the address of the variable, but a structure
> containing length and pointer to storage).
> <snip>

Manipulations involving character*(*), i.e., character indexing and
concatenation, are easy to implement on a stack or a heap, but I believe
cannot be implemented statically without either a system dependent
maximum size of character entities or a detailed interprocedural
data-flow analysis.