This is the mail archive of the libstdc++@gcc.gnu.org mailing list for the libstdc++ project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Standard header format.


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Saturday 05 June 2004 20:59, Gabriel Dos Reis wrote:

> It is not a conjecture.  It is a fact.  What you state as your belief
> simply does not hold. At best, it is an insult targetted at him.

No, I merely assumed that he was making a suggestion he expected would apply 
to most implementation at the time he was writing. Rather than clutter the 
text with several qualification, he left it stand as an imperfect yet 
generally useful suggestion. But read what I wrote about my ability to follow 
that suggestion: "That presents a _minor_ problem for me."

I was not suggesting Stroustrup's statement indicated that the implementation 
is required by the Standard to provide header files.  If I thought that was 
the case I would have referenced the relevant sections of the Standard.

Look at 9.3.2 of TC++PL(SE). In particular his use of _impl.h verses the 
header file for users.  Explain to yourself why that concept fails to apply 
to the Standard Library.  Note that I did not say the Standard requires the 
implementation provide a user interface.  I am saying providing one is a good 
idea. You tell me I should not "look underneath".  If you cover it up with an 
interface, I don't have any reason to do so.
- -- 
Regards,
Steven
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)

iD8DBQFAwnd+wX61+IL0QsMRAiRGAJ4yFI8f5enCQeEgKxvfwhR38gy2XgCg8Ghb
AxxOc2X/+Q8aQ8kYSgW3JFI=
=gdK7
-----END PGP SIGNATURE-----


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]