Re: [avr-gcc-list] Re: [avr-libc-dev] RFD: more avr-libc API changes

[Erik Christiansen]

On Thu, Sep 08, 2005 at 10:52:40PM +0200, Joerg Wunsch wrote:
> As Wojtek Kaniewski wrote:
> 
> > How about...
> 
> >    #define VECTOR(signame) \
> >    void SIG_ ## signame (void) __attribute__ ((interrupt)); \
> >    void SIG_ ## signame (void)
> 
> I don't know.  I'm more inclined to use ISR(), but I'd rather like to
> see other people's opinions on this.

Yes, ISR (UART_DATA) { .... } has the advantage of descriptive accuracy.
Invisible setting of the vector table entry is incidental surely, and
insufficient reason for less than accurately describing the interrupt
handler during definition. It would be a pity to cause another cycle of
confusion.

> Obtw., of course, it's __attribute__((signal)). :-)

Hmmm ... I might just continue doing 'em in assembler. :-)
It does have the advantage that the ISR name does not magically change
to __vector_N in the ELF file.

Erik