🚀 go-pugleaf

In article <1992Nov16.200843.5565@Princeton.EDU> rhl@grendel.Princeton.EDU (Robert Lupton the Good) writes:
>#include <stdio.h>
>   const int i = 0;
>   (int)i = 1;				/* Fails */
>   *(int *)&i = 1;			/* OK */
>Thinking about ROMs my guess is that gcc and the IBM xlc compilers are
>correct to reject the former assignment,

Thinking about the standard, where it says that the result of a const is
not an lvalue, you should know that every conforming implementation is
correct to reject the former assignment.

>and that their acceptance of the latter is permitted,

Yes.  The latter assignment is allowed to have any effect at all, without
even needing to be documented by an implementation.  Acceptance is legal.

>What do you think about the portability of the cast?

The cast is portable.  The attempted modification of a const object (i)
through use of a non-const lvalue (*(int *)&i) is allowed to succeed,
fail, dump core, or print some advice that you should buy a standard.

>(I want to set the variable once in a private place just after mallocing it,

malloc a non-const variable, set it once, and then copy the address to a
pointer-to-const-int.  Then pass around the const version.
--
Norman Diamond       diamond@jit081.enet.dec.com
If this were the company's opinion, I wouldn't be allowed to post it.
"It's been a lovely recession."