Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!swrinde!cs.utexas.edu!yale!cmcl2!lanl!cochiti.lanl.gov!jlg
From: jlg@cochiti.lanl.gov (Jim Giles)
Newsgroups: comp.lang.c
Subject: Re: Careful "for" Loops
Message-ID: <19064@lanl.gov>
Date: 25 Mar 91 18:43:48 GMT
References: <MCDANIEL.91Mar19124111@dolphin.adi.com> <4176@rwthinf.UUCP>
Sender: news@lanl.gov
Reply-To: jlg@cochiti.lanl.gov (Jim Giles)
Organization: Los Alamos National Laboratory
Lines: 42

In article <4176@rwthinf.UUCP>, berg@marvin.e17.physik.tu-muenchen.de
(Stephen R. van den Berg) writes:
|> Tim McDaniel writes:
|> >Case a)
|> >   semi-open intervals, like [low,high).
|> >Case b)
|> >   closed intervals, like [low,high].
|> [...]
|>  unsigned long i,low,high,incr; 
|>         for(i=low;i-high>=incr;i+=incr)
|> 
|> Restriction is now: high-low<=UNSIGNED_LONG_MAX-incr

Since this is a Case a) loop, the restriction can be further
narrowed (by just the value of incr):

  unsigned long i, j, low, high, incr, trip_count;
            ...
            trip_count = (high-low)/incr;
            for (i=low, j=0; j<trip_count; i+=incr, j++) ...

The restrictions are now: high-low <= UNSIGNED_LONG_MAX, incr != 0,
and high >= low.  (These last two conditions are obvious, but should
be bourne in mind anyway.)

Similarly, Case b) can be handled with the same restrictions:

  unsigned long i, j, low, high, incr, trip_count;
            ...
            trip_count = (high-low)/incr;
            i=low; j=0;
            while (1){                  /* loop forever (sort of) */
               ...
               if (j==trip_count) exit; /* exit condition in middle */
               i+=incr; j++;
            }

Note that in Case a) the trip_count variable is actually the number
of times the loop body is executed.  In case b), trip_count is one
less than the number of passes through the loop.

J. Giles