Re: Mistaken change in GCC (fwd)

[Richard Stallman <rms at gnu dot org>]

    In this case, it looks like it would cause problems for Emacs... and
    who else?

It is never easy to get the answer to such a question.  The users
one can find out about are a subset of the users there are.  The only
way you find out what breaks when a feature is removed is by removing it
and seeing.

That question would be relevant if we wanted to delete a feature and
only preexisting uses stood in the way.  That is not the situation
here.  We're talking about the combination of two useful features,
#cpu and --traditional.  The two should work together for simple
consistency.

Anyone writing a header file that does not need to support other
compilers, and that needs to check for certain computer types, ought
to use #cpu rather than the alternative.  Some header files that have
not used #cpu probably should use it.

    With respect, I submit that the affected audience is not nearly numerous
    enough to simply demand that an oddball extension be reinstated 

The term "oddball" appears to criticize an extension because it is not
yet popular; that indicates a basic misunderstanding of our design
goals.  One of the design goals of GCC is to add useful extensions and
*make* them popular.