Hi Kurt.

23-Sep-03 01:59:13, Kurt Kuzba wrote to All


 KK> From: kkuzba{at}centurytel.net To: c_echo{at}yahoogroups.com

 KK> * Author: Jasen Betts
 JB>> You're testing string[0] against 0 twice, and zeroing space and
 JB>> iterator twice)...  (put the last if inside the first if)

 KK> The integer variable "space" is incremented only if the char value
 KK> is not a space in the first if.  In the second if, the value of
 KK> the integer value "iterator" is compared to
"space" to see if
 KK> there are trailing spaces, which will be indicated by a difference
 KK> in the two values.  If the second if were placed inside the first,
 KK> then it would provide a NUL terminator at the location of "space"
 KK> on each iteration in which a space had been encountered, even if
 KK> there were subsequent nonspace characters remaining in the
 KK> unparsed portion of the string.  This is a one-pass solution.

I meant do this:

 static char *notrail(char *string)
 {
     if (string)
     {
     int space = -1, iterator = 0;
         for(; *(string + iterator); iterator++)
             if(!isspace(*(string + iterator))) space = iterator;
     if( ++space != iterator)  *(string + space)  = NUL;
     }
     return string;
 }

 KK> Had I thought of it, I would have added a line testing the value
 KK> of "space".

 KK> if(!space) *string = NUL;

 KK> This is simple and quick, and "space" will only be equal to zero
 KK> if the only character remaining is a space.

are you sure?
that would'nt work with things like "a" and "b   "

 JB>> For long strings determining the length first and then working
 JB>> backwards may be more efficient like Bob Stout's code.

 KK> Maybe, but you are stepping through the string and then backing up
 KK> in that case, and I wanted to just go through it one time and call
 KK> it done.

yeah, that was my first take on the problem too. but in many cases each
isspace() is extra an array lookup (check ctype.h to see how isspace is
implemented by your compiler(s))

If you're trimming less than half of the string Bob's backwards-seek is
more efficient, and the code which uses strlen() is in many cases already
using optimised assembler to do most of the work.

 -=> Bye <=-

---