char* foo(const char*) standard functions

[Joe Buck]

On Tue, May 30, 2006 at 11:11:15PM +0200, Paolo Carlini wrote:
> Marc Glisse wrote:
> 
> >Ok thanks, if this discussion is the reference on the subject, I will 
> >consider getting these 2 additional unexpected failures on solaris 
> >not-a-bug (or even a good thing), although I still find it arguable 
> >that the testsuite considers it a bug.
> 
> The testsuite considers it "a bug" because, 'til today we worked under 
> the assumption that the supported targets don't change the C headers 
> depending on the C++ compiler being used ...

This assumption is, I believe, violated by Sun, HP, and Microsoft,
all of which ship headers with "if C++" constructs in them.  If anything,
glibc is the outlier.