making conforming headers.

[Martin v. Loewis]

> Oops.  I dont know where I got that "only" (probably from
> 17.4.4.1/3).  Its the cname/name.h problem that I was looking at.
> The cname headers are, as most of them note, non-conforming and the
> name.h headers dont exist.  Adding the missing definitions to the
> cname headers would probably make them more or less conforming, but
> given the restriction in 17.4.4.1/3 and the requirements of annex
> D.5, its not possible to add the name.h headers without reworking
> the whats being included by the bits/std_cname.h headers.

The relationship between cname and name.h is more tricky,
indeed. libstdc++ supports a mechanism to wrap the system headers into
namespace std; you can ask Nathan Myers for details.

I believe the restriction in 17.4.4.1/3, as well as annex D.5, is
mostly a non-requirement, because of the as-if rule. Can you give an
example where using /usr/include C library name.h headers violates any
of the requirements? I'm asking for a violation that can be detected
with a C++ program; not one that is detected by looking at the sources
of the headers in an editor.

Regards,
Martin