lnblnk

[gregr]

I finally found lnblnk in libpepcf90mt.lib. When i tried to use it it promptly blew up. So i pasted the "interface" from the intel manual:

INTERFACE

INTEGER(4) FUNCTION LNBLNK(STRING)

!MS$ATTRIBUTES ALIAS:'lnblnk_'::LNBLNK

CHARACTER(LEN=*), INTENT(IN) :: STRING

END FUNCTION

END INTERFACE

Into my code. It doesn't blow anymore but it returns the length of the string instead of the position of the last non-blank character. I looked at the string in hex, there is no null at the end, it's definitely got blanks from character 12 to the end of the string.

Please help, i don't want to have to go back to writing my own routines for stuff that ought to just work. I really don't understand why i had to find these in a library, they are *inline* functions, no?

Thanks,

Greg

PS i'm using Intel fortran 7.0.

Message Edited by [email protected] on 07-28-2004 09:30 AM

[Steven_L_Intel1]

No - LNBLNK is a "Portability Library" routine, it is not a compiler intrinsic.

I tried the following code in 7.1 and 8.0 and got 7 returned in both cases:

Code:

use iflport
character*10 string
integer i
string = 'A B C D'
i = lnblnk(string)
write (*,*) i
end

I note that in 7.1, I needed to build with /4Yportlib to pull in the portability library - that is not needed in 8.0.

[sgongola]

You meant:
use ifport

[Steven_L_Intel1]

No, I didn't. In 7.1, one had to say "use iflport". 8.0 accepts either.

[gregr]

Ok, thanks, i'll keep thrashin with it.

Greg

[gregr]

Ok, here's what the deal was. The string i was lnblnking came from another of my C routines. I had a get_environment routine which called the C getenv which filled in that string i was lnblnking. I swear, there was no null (that the debugger showed me, i looked at it in hex) in that string, but something about it confused the Fortran support lib.