Re: Callout Protocol Core - Kept parameter

[Alex Rousskov <rousskov@xxxxxxxxxxxxxxxxxxxxxxx>]

On Tue, 2006-04-25 at 18:32 +0200, Martin Thelian wrote:

> Martin Stecher schrieb:
> > [..]
> > I am confused too and wonder why I didn't see that before.
> > Probably the idea was that the value of a named parameter can be a
> > space separated list of values and therefore the CRLF separation of
> > the named parameters has been introduced.
> > That means:
> >
> >   named-value = name ":" SP value *(SP value)
> >   
> From my point of view this would be the best solution to resolve this 
> conflict. The rule for named values could then be rewritten to:
> 
> named-value = name ":" SP anonym-parameters

I agree that the spec is buggy, but it seems to me that a better fix is
to replace

	[Kept: org-offset org-size ];

with

	[Kept: kept-info];

where "kept-info" is a new type, extending OCP structure with two
anonymous parameters org-offset and org-size. I think many other OCP
message parameters follow this scheme already.

BTW, the old Kept definition also contradicted the "Declaring a new
message with parameters" rules in Section 9!

I did not find other similar "value SP value" definitions that would
need changing. Did I miss any?

> I've used this example, because there is the following sentence in rfc4037:
> 
> |> For simplicity, preserved data is always a contiguous chunk of
> |> original data, described by an (offset, size) pair using a "Kept"
> |> parameter of a Data Use Mine (DUM) message.
> 
> In this sentence the Kept parameter parts where written as list. But 
> maybe it's only a strange coincidence.

The parenthesis are used to mark a pair of two parameters, nothing more.
It would have been better to use a different notation, apparently.

Sorry for the delay in responding. I was under a false impression that
the WG [list] is dead.

HTH,

Alex.
P.S. This is not the only bug in the OCP RFCs. I should publish the
errata.