Colin O'Flynn wrote:

Perhaps if the issue comes up again, referring to the ANSI C spec might at
least have some authority?

For those that understand the terminology, yes.

... any expression referring to such an object [with volatile-qualified
type] shall be evaluated strictly according to the rules of the abstract
machine...

Not everyone will understand what this means.  I have tried to provide an
interpretation of this in the context of bob's original post.

(114): A volatile declaration may be used to describe an object
corresponding to a memory-mapped input/output port ...

This is actually advice for how to use the volatile storage qualifier and
probably got into the specification due to just the confusion we're
experiencing now.

Which again sounds exactly as GCC behaved.

YES!

Graham.



_______________________________________________
AVR-chat mailing list
[email protected]
http://lists.nongnu.org/mailman/listinfo/avr-chat

Reply via email to