Re: [Issue 7] - Build with VC++ under NT broken

[Derek R. Price]

Manfred Klug wrote:

> ----- Original Message -----
> From: Derek R. Price <bug@ccvs.cvshome.org>
>
> > Also, would you mind creating a new issue for your S_ISSOCK fix which you
> also attached to this issue with a more complete description of the
> > problem?
> About the Problem in general or about the patch for buffer.c

Both.  I remember this issue now, but someone else going through IZ#7 would
probably be confused.


> For client.c I use currently the following code, but I'm not satisfied with
> it.
>
>     if (fstat (tofd, &s) < 0)    /* If it is not a file it must be a socket
> */
>     {
> // error (1, errno, "Failed to stat file descriptor\n");
> //    }
> //    if (S_ISSOCK(s.st_mode))
> //    {
>  assert (tofd == fromfd);
>  *to_server = socket_buffer_initialize (tofd, 0,
>            (BUFMEMERRPROC) NULL);
>  *from_server = socket_buffer_initialize (tofd, 1,
>       (BUFMEMERRPROC) NULL);
>     }

Yeah.  I'm not happy about that either.  On most systems, it is an error if
the descriptor isn't a socket.

Also, why do you ignore the fstat error?

I think I'd rather have a function (probably aliased to S_ISSOCK
windows-NT/system.h or whatever the file is) which works under windows in
windows-NT/filesubr.c or the like.

It would probably be best (read: most efficient) to put your reply to these
allegations ( :) ) in the new issue.  :)

Derek

--
Derek Price                      CVS Solutions Architect ( http://CVSHome.org )
mailto:dprice@collab.net         CollabNet ( http://collab.net )
--
Goldfish don't bounce.
Goldfish don't bounce.
Goldfish don't bounce...

          - Bart Simpson on chalkboard, _The Simpsons_