Re: Question about #pragma removal

[Paul Koning <pkoning at equallogic dot com>]

>>>>> "mike" == mike stump <mrs@windriver.com> writes:

 mike> My take on it is that pragma isn't really unacceptable, it is
 mike> just what happens to be in the manual currently.  Some ports
 mike> have various pragmas they support.  Some ports don't support
 mike> any pragmas.  All ports support attribute.  This is more just
 mike> an observation of what is.  At some point in time, I suspect
 mike> someone will contribute work to smuggle the full power of
 mike> attributes into pragmas.  And then, at some point, all of our
 mike> examples will change to use that form, and then at some point
 mike> attribute may go away.

That final step sounds like a very bad idea.  I would think that it is
GCC policy that its extensions are permanent features (at least if you
use the __keyword__ form which assures no clash with future C/C++
keywords).  Any suggestion that GCC extensions might disappear in the
future, especially one as critical as __attribute__, will undoubtedly
meet with howls of protest.

(Note that retiring a keyword on the grounds that its features are
available via a more recently added keyword is, for legacy code,
equivalent to removing the feature entirely.)

     paul