Re: [avr-libc-dev] (Fwd) WinAVR 20030424

[E. Weddington]

On 5 May 2003 at 9:46, Theodore A. Roth wrote:

> 
> 
> On Mon, 5 May 2003, E. Weddington wrote:
> 
> :) > > {
> :) > > #if defined(EIMSK)
> :) > >  EIMSK = ints;
> :) > > #elif defined(GIMSK)
> :) > >  GIMSK = ints;
> :) > > #elif defined(GICR)
> :) > >  GICR = ints;
> :) > > #else
> :) > > # error "No valid external interrupt control register defined."
> :) > > #endif :) > > }
> 
> <snip>
> 
> :) But in the meanwhile, as a stopgap, can we change the #error to :)
> #warning? Or just drop that branch and let it compile to nothing; :)
> it's harmless if those registers aren't defined.
> 
> It's an error since the header isn't handling something correctly.
> 
> Doing nothing is wrong. The user expects something to happen when they
> use the interface and silently doing nothing breaks their program. I
> made it an #error to help flush out unhandled cases (and it seems to
> have worked).
> 


Do you have an alternative?

Since there's no external interrupt control register defined for the 
part, the function is meaningless.

We could just make it unavailable such as:

#ifndef _AVR_IO86RF401_H_
extern inline void enable_external_int (unsigned char ints)
{
#if defined(EIMSK)
 EIMSK = ints;
#elif defined(GIMSK)
 GIMSK = ints;
#elif defined(GICR)
 GICR = ints;
#else
# error "No valid external interrupt control register defined."
#endif
}
#endif


Or, (shudder), yet another AVR architecture class to handle these 
oddball application-specific + AVR parts and conditionally compile on 
the architecture.

Eric