>>>>> On Fri, 29 Jun 2001 12:29:06 +0700,
>>>>> Robert Elz <[EMAIL PROTECTED]> said:
> | Please forget about this topic.
> Done... Apologies for sending lengthy comments on stuff that was
> old - I half thought I'd seen it before, but my memory just isn't that
> good...
No, that was my bad, I should've cited the related part only. Anyway,
thanks for the kind reply.
> | I don't think so. I believe the protocol stack (normally the kernel)
> | should prevent (at least) non-privileged users from sending invalid
> | packets according to the protocol specification.
> Sure. But that wasn't the point - the point is exactly what does get
> sent (if anything) when the user does something invalid. I think it is
> OK to say that that is implementation defined. No correct application will
> ever care. That broken applications behave differently on different stacks
> should be a feature, not a bug - it helps identify them as broken sooner.
> I certainly wasn't suggesting that the stack should be creating invalid
> packets out of invalid user input, and sending that.
Hmm, I see. So, "the API does not prohibit the bad behavior, but the
underlying kernel would (or might) prevent the bad packets from beiing
sent on the wire." Right? Then, I'm perhaps okay with it.
JINMEI, Tatuya
Communication Platform Lab.
Corporate R&D Center, Toshiba Corp.
[EMAIL PROTECTED]
--------------------------------------------------------------------
IETF IPng Working Group Mailing List
IPng Home Page: http://playground.sun.com/ipng
FTP archive: ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMAIL PROTECTED]
--------------------------------------------------------------------
