--yLVHuoLXiP9kZBkt
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

Thor Lancelot Simon wrote:
> You misunderstood the other replies.

I might have misunderstood your one, I didn't quite look as a strong
objection and as you can guess from my mail address I'm neither a
citizen of the US or UK, so I probably don't know all idioms and
sayings or misinterpret them. Ok, two objected and two didn't find
it very useful.

> It's not the job of a Unix
> utility to prevent system administrators (who should, after all,
> know better -- and who could, easily enough, RTFM if they didn't)

That has *nothing* to do with RTFM. I've surely read the ifconfig
manual page more than once but neither do I remember every aspect
of it nor do I read it everytime before I use ifconfig. Further,
I don't use ifconfig quite often since I've configured my system
so that I don't have to. Granted I'm probably a moron for thinking
I know what "dot notation" means while I don't (I thought it was
a.[b.[c[.d]] - it's not). Even worse, I didn't understand that "the
dots are optional, if the user wishes to (carefully) count out long
strings of digits in network bytes order) means that 1500 or
0x12 or 0415473 is a valid address.

I would even think that "ifconfig rtk0 1500" matches neither the
usage information by ifconfig nor the syntax description in the
manual page - I must be wrong or someone else didn't RTFM.

> from hosing their systems with invalid input.

Too bad, we have memory protection. Why should the memory controller care
about writing buggy programs?
=20
> Unix utilities don't try to guess what you _really_ meant to type
> on the command line; that's your job.

Oh really? That's probably the reason to the discard all redudancy
in an IP address so you can enter almost anything which comes close
to a number and still have valid input.

> Unix gives you enough rope;
> tie knots or hang yourself, but the rope's all there.

Maybe it's just me, but a plain "I rely on this feature, don't change
it." was all I was aiming for (in the negative case). Your phrases
and cited principles do not quite convince me anybody needs this or
that changing one or maybe two lines of code would make anything worse,
I could and would simply update the manual page and throw a funny RTFM,
if anyone would ever complain.

--=20
Christian

--yLVHuoLXiP9kZBkt
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (NetBSD)

iD8DBQFAe5Zh0KQix3oyIMcRAubIAJ4qniX3T5GynLr4rVaopW2TND0YpQCgrq1j
2pP97jkFoC59rYLM2U9v40c=
=zONl
-----END PGP SIGNATURE-----

--yLVHuoLXiP9kZBkt--