Re: support for i386-pc-nto-qnx* target

[Zack Weinberg <zack at codesourcery dot com>]

On Mon, Aug 26, 2002 at 04:52:12PM +0100, Neil Booth wrote:
> Zack Weinberg wrote:-
> 
> > Also, new ports should not create any assertions
> 
> I know you want to get rid of assertions, but are you aware that
> they are specified as part of the Itanium C++ ABI?

Ugh.  What are they doing there?

> They work as GCC assertions work, but they list
> 
> #assert cpu
> 
> as a valid syntax (a zero-token answer), but don't permit
> 
> #if #cpu
> 
> which admittedly always seemed kinda pointless to me.
> 
> I believe EDG follows these semantics.

It's a demonstration of how little-used these are, that no one has
objected to the semantic difference.

In principle, we should match EDG -- the whole point of this extension
is compatibility with other compilers -- but I wouldn't bother fixing
it unless and until someone actually trips over the difference.  There
are more important things to do.

zw