Ethereal-dev: Re: [Ethereal-dev] packet-mount.c

Note: This archive is from the project's previous web site, ethereal.com. This list is no longer active.

From: Guy Harris <gharris@xxxxxxxxxxxx>
Date: Wed, 14 Mar 2001 22:20:15 -0800
On Wed, Mar 14, 2001 at 06:59:36AM +1100, Ronnie Sahlberg wrote:
> I gave them the same name on purpose.
> They are basically the same thing,
> however, according to my docs, the statuscodes only have a welldefined 
> meaning for v3.
> v1 and v2 have no statuscodes defined in mount.x

My mail addressed this question.

The mere fact that there's no official definition for the status codes
in the RFC or the rpcgen file doesn't mean that, in practice, "anything
goes"; the RFC says it's a "UNIX error number" and, as I said in my
mail:

1) any mount daemon that uses a UNIX errno that's not the same
   on all flavors of UNIX is likely to cause confusing error
   messages if you fail to mount from it;

2) EPERM, ENOENT, EIO, EACCES, ENOTDIR, and EINVAL will be the
   same on all UNIXes that ever had AT&T code in them (this
   includes the BSDs), and Linux also gives them those values on
   all architectures supported by the 2.2.18 kernel, at least
   (and I'd be *IMMENSELY* surprised if the new ones used
   different values);

3) ENAMETOOLONG is the same on BSD-derived UNIXes and Linux, and
   the AT&T UNIXes *might* have mapped it, in mount, to 63;

4) the other V3 ones are, I think, new in V3;

5) it gets shown numerically anyway, so if somebody sends a
   "non-standard" value, you at least get to see it.