On Sat, Apr 17, 2010 at 03:19:49PM -0400, William Chops Westfield wrote:
> 
> On Apr 17, 2010, at 11:53 AM, Byron Jeff wrote:
> 
> > it brings consistency to structure access.
> 
> That's rather neat.  Does it solve the "." vs "->" annoyance inside  
> nested structures as well?
> 
> struct a {
> 	struct b ab[1];
> 	struct c* ac;
> 	struct d ad;
> } *aa;
> 
> Normally you'd need to remember to say "aa->ac.acinsides" vs "aa- 
>  >ad.adinsides",
> but can you now say aa->ab->abinsides even though ab storage is part  
> of aa ?

Yep. It's totally self consistent. And exactly the headache it's designed
to get rid of.

> 
> 
> > But it's not very efficient on eight-bit PICs as far as code  
> > generation.
> 
> PICs aren't so great at handling pointers in general, but do you think  
> Byron's technique is any worse than the more traditional alternative  
> of a pointer and a structure explicitly defined for the same space?   
> Seems to me that for global structures most of the inefficient pieces  
> should be eliminated by the compiler (for either style)...

Actually one would hope that the compiler is smart enough to know that if
you are dereferencing a constant address, which an array name is defined to
be. there's no need to actually load that address into an FSR and do an
indirect access.

But if I really wanted an efficient use for an 8 bit PIC, I can write in
assembly. Any compiled language is going to add overhead unless it is
optimized.

BAJ
-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist