At 08:47 01/05/1998 PST, Tom Hastings wrote:
>At 08:03 01/03/1998 PST, Ira Mcdonald x10962 wrote:
>>Hi Carl-Uno,
>>>>Do you envision conference calls (to help sort out our few
>>remaining issues and any edits that should have made it into
>>the most recent Model and Protocol specs but didn't, for
>>example the range of request-ID being '1..n' and not '0..n')?
>>Ira,
>The Protocol document was changed in section 3.6 to make the example value
>for clients that aren't using it be the constant 1, instead of 0,
>so that its value is a legal value as agreed to align with the Job MIB and
>the SNMP requirement not to use 0 as a table index value.
>>However, the ABNF fails to specify the syntax of the request-id token.
>(0 or 1). It should be SIGNED-INTEGER, as all four octet integers are,
>but with some restriction on the range to be 1 to 2**31-1.
>>Also I would think that section 3.6 should also include the range limits,
>as has been done for other fields.
>>Also the Model document doesn't seem to mention the request-id at all,
>that I could find. I'm not sure whether it should or not, since the
>request-id is more of a protocol mechanism.
I found where "request id" is specified in the Model document. Its in
section 3.1.1. In the protocol document, its called "request-id", but
is called "request id" in the Model document (which is why I didn't
find it when searching the Model document). However, the lower bound
is still specified as 0 to 2**31-1 in section 3.1.1 in the Model document
and needs to be changed to be: 1 to 2**31-1 as agreed.
Perhaps the protocol document section 3.6 should also be fixed to mention
the "request id" in the model document as mapping to the "request-id"
in the protocol document?
>>Thanks,
>Tom
>>snip...
>>>>