Fractran: Difference between revisions
Content added Content deleted
(→Revised Plan: Explain the revised output of N as a set of prime factor powers.) |
m (→The Code) |
||
| Line 745: | Line 745: | ||
===The Code=== |
===The Code=== |
||
The source style is F77 except for the use of the I0 format code, though not all F77 compilers will offer INTEGER*8. By not using the MODULE scheme, array parameters can't be declared via P(:) which implies a secret additional parameter giving the size of the array and which |
The source style is F77 except for the use of the I0 format code, though not all F77 compilers will offer INTEGER*8. By not using the MODULE scheme, array parameters can't be declared via P(:) which implies a secret additional parameter giving the size of the array and which can be accessed via the likes of <code>UBOUND(P, DIM = 1)</code> Instead, the old-style specification involves no additional parameters and can be given as P(*) meaning "no statement" as to the upper bound, or P(M) which ''may'' be interpreted as the upper bound being the value of M in the compilers that allow this. The actual upper bound of the parameter is unknown and unchecked, so the older style of P(12345) or similar might be used. Rather to my surprise, this compiler (Compaq F90/95) complained if parameter M was declared after the arrays P(M),Q(M) as it is my habit to declare parameters in the order of their appearance. <lang Fortran>C:\Nicky\RosettaCode\FRACTRAN\FRACTRAN.for(6) : Warning: This name has not been given an explicit type. [M] |
||
INTEGER P(M),Q(M)!The terms of the fractions.</lang> |
INTEGER P(M),Q(M)!The terms of the fractions.</lang> |
||
So much for multi-pass compilers! |
So much for multi-pass compilers! |
||
| Line 799: | Line 799: | ||
END !Whee! |
END !Whee! |
||
</lang> |
</lang> |
||
===The Results=== |
===The Results=== |
||
Output: |
Output: |
||