Re: [avr-gcc-list] Re: What Happened to the sbi() and cbi()

[E. Weddington]

Pertti Kellomäki wrote:

> 1. helmi 2005 kello 01:00, Eric wrote:
>
> > So far there haven't been too many others complaining about it.
>
>
> Since you asked,  here's my data point. I am an academic
> planning to go small time into embedded systems, and I have
> written a data logging application for the AVR Butterfly.
> It uses sbi() and cbi() simply because the examples
> I based it on used them. If it ain't broke, don't fix it.
>
> I have been through a variety of languages, so I am perfectly
> capable of rewriting my code. However, in this case my fix would
> be to redefine sbi() and cbi() privately, as I fail to see the inherent
> evil in them. Frankly, to set a bit using a construct whose
> stated purpose is to set a bit sounds like a good idea to me. 

See, that's the problem.

Originally, those macros were written and used to *actually* generate 
the SBI and CBI instructions. Over time, that functionality was put into 
the compiler, where it does it automatically if optimization is turned 
on. So the macros were re-written to just simply set and clear bits. So 
the big problem is that those macros give a false impression. Hence, 
deprecated and removed.

Eric