In a message dated 01-02-25 13:56:40 EST, Ryan writes...
> I believe the solution to the first issue, is to tell the
> compiler not to issue those warnings.
As far as 'unused arguments' goes... there are really only 2
options other than --Wdont_report_this_one
One is to just make the arguments reference themselves
for any platform where they are really, truly unsued...
int ap_something( parm1, parm2, win32onlyparm )
{
#ifndef WIN32
win32onlyparm=win32onlyparm;
#endif
}
That will stop all non-WIN32 compiles that don't use the
argument from reporting it as 'unused' ( At least, it SHOULD ).
Second option...
#pragma argsused
int ap_something( parm1, parm2, win32onlyparm )
{
...
}
The '#pragma argsused' does just what it says.
It stops the compiler reporting any 'unused' agruments
but it is a 'one shot dea' and should only be added just
above any functions where you KNOW it's ok not to
report it. It is the best way to 'catch bugs' by not having
to tell the entire compile to ignore an error that MIGHT
actually be a problem for some functions.
> Does anybody have any thoughts for the other two issues?
Mike Abbot's SGI patches are not all 'typecast' solutions.
Might be good to look at those 'techniques'.
I don't see how you can pull off a true 32/64 bit single codebase
for all compilers without at least a 'few' casts but I guess
it's possible.
What is the real argument behing the whole typecasting
thing, anyway? It works.
Yours
Kevin Kiley
CTO, Remote Communications, Inc.
