PC> Anybody see any obvious problems with the following code?  Calling
 PC> free(Track_Buffer) under OS/2 causes SYS3175, under DOS there are no
 PC> apparent problems.
 
   Works fine when I build it under C/Set++ 2.0 after removing the
reference to dos.h, and making the fopen mode "r" rather than
"rt" (IBM doesn't accept the t, but it's the default anyway).
 
   A few comments though.. why bother with
 
#define ONE_HALF(x)       (x*50/100)
 
   when putting x/2 or x>>1 is already quite readable? Also, while I
don't know what this program is leading to, your calls of tslice() worry me
- why are you sleeping after each and every character? A program that has
real work to do should just do it - OS/2 will grab the CPU when it's needed
elsewhere.
   Lastly, just a style comment - the use of gloabls makes it easy to break
things since on reading over a function that uses them, you don't know from
the context where they've been and what other code might have done to them.
Better as a habit to pass into a function everything that it's going to be
using. That way, once you have the function working, if on using it
elsewhere, something breaks, you know to look at what you're passing it
rather than having to step your way through the function again to figure
out what it's using.

--- Maximus/2 2.02p1