mcdona_m@wacsvax.OZ (McDonald) (10/09/89)
If we can't afford the space for old and new style function
definitions in the C compiler, I know which way I'd rather swing.
If Dr Tanenbaum's compiler supported new style declarations, cleanit.c
would not have had the problem it did. I'd much prefer new-style only to
old style only declarations. Apart from anything else they're simpler for
a compiler to deal with than the old style declarations.
The only real problem I can see is the mass of PD stuff which uses
old style declarations. Seeing however that something called protoize
was recently mentioned in the gnu.gcc group, I can't see this being a
major problem any longer.
Using old style declarations is begging for tricky bugs.
If anyone wants to post the protype thing from the gcc group here, Id be
really happy. Over here, we don't have ftp running properly.
I may be able to get it anyway, but I'm not counting on it.
PS: will you all drop this goto stuff, we sound like
a bunch of first year students. We all know they're bad
and to use them sparingly, if at all. Can we please
drop it. I've got an opinion on the subject too, but
it's just that, opinion. I'm not going to inflict it on
you, so don't inflict yours on me.
And to the person who told everyone to shut up about DMA,
I think that was pretty bloody rude. This is supposed to be
a forum devoted to a teacher's operating system. The people
in operating system courses should be aware of the issues involved
and I am sick of being told that DMA is the only way
to handle communication with disks. If discussion in this group
helps raise general awareness, it's a good thing. Besides that
it's a little trivial for discussion in an architecture forum.
Matt