On Fri, 1 Sep 2000, sam th wrote:
> On Fri, 1 Sep 2000, Thomas Fletcher wrote:
>
> > This code has caused us enough porting grief that I'm
> > of the opinion that it should be moved to a platform
> > dependant layer ... all those in favour say eye!
> >
>
> This porting grief is *not* the code's fault, and therefore I think it
> should stay XP. wchar.h and its contents are, if the various websites I
> have read are to be believed, part of the C standard, and therefore
> something that should be available on every platform.
>
I agree that the greif is not the code's fault. The header
and the function are both standard, but as we can see from
our experience on a couple of platforms now ... it is not
by any means a fully implemented standard. The other
problem I have with this code is that it makes it a pain
to replace the function with the non-stateful version
(which seems to be present in most cases) because of the
name clash. This is why all the #defines are needed
for the code to make it work in a guaranteed manner.
In any case it is working for QNX/NTO now, but I'm
sure that we will have to address this issue again
later when some other platform needs to adjust it.
Thomas
