On Fri, 16 Feb 2001 23:14:23 +0900, Martin Buchholz said:
Everyone knows size_t is the same as unsigned long on real systems.
Yes, the problem is when you codify that and then try to port some time
later to a system that doesn't behave that way.
How many of the bugs we've seen porting XEmacs to 64-bit chipsets have been
caused by underlying "everybody knows..." assumptions caused by "all real
systems are 32 bit, the toy ones are 16 bit"?
As soon as the KL-20 CPU died off and it seemed that "8 bits comprise a byte"
was a safe assumption, we started seeing code that used 16-bit w_char's ;)
--
Valdis Kletnieks
Operating Systems Analyst
Virginia Tech