[ILUG] EU Rejects Patent Software Law - For Now

[Colm MacCarthaigh]

On Fri, Feb 18, 2005 at 08:27:14PM +0100, Brian Foster wrote:
>  2nd pedantic quibble --- that is not a valid (as in standard
>  conformat) definition of main(), albeit it does work in most
>  (all?) *ix systems.   from memory, the two valid forms are
>  int main(void)  and  int main(int, char **)  and equivalents.
>  and NO, it is not true that  foo()  and  foo(void)  mean the
>  same thing; in C, they do not.  (b.t.w., the first form, foo(),
>  is also deprecated.)  see K&R-II, where that first form, foo(),
>  is called ???old-style???.

Well, I didn't say that it was the same thing, I was talking about return
types. *scratches head* 

>  I have not read the patent in question, nor have I read this
>  thread very closely.  but it did strike me that all(?) of the
>  code posted so far seems to assume a location in memory has
>  exactly one address.  i.e., the possibility of the memory
>  being double-mapped (as one example, to two different virtual
>  addresses in the same process) is not handled.

Hmmmmm, I'm not so certain a language-level operator could ever
genuinely cope with this, a system-call might be the only way. Unless
the language itself managed its own memory (ala Java) or simply
prevented double-mapping. It would be very interesting if MS's
operator did do this, because that really does give it some novel
complexity.

-- 
Colm MacCárthaigh                        Public Key: colm+pgp at stdlib.net