[patch] fixincl.c vs solaris 2.6

[DJ Delorie]

> Curious:  Is the 'n' argument the problem in that it gets substituted
> in the "\n" string?  That would be pretty weird, though, because the
> else clause is for  __STDC__ being defined.  How can you claim __STDC__
> and do that??  Oh, well.  It's still approved.

I don't know.  The original PR was for Solaris 2.5.1, but in our case
it fails for 2.6 but not 2.5.1 or 2.7 or 2.8.  It also fails for the
unbundled compiler but not the bundled compiler.  Go figure.