Was the implementer actually ignoring octets
in the PDU or was the implementer in fact ignoring octets in values taken from
the PDU? It would seem strange, nay impossible, to ignore octets in a PDU. This
means you would read the ID, read the length, take the first 20 bytes then ..
what? Stop? If you stop taking bytes then you may not be at the next tag, so the
decode would fail. If you continue taking bytes but don't store them, then you
do regard the bytes in the PDU as significant.
Therefore, I don't believe you can EVER
ignore bytes IN a PDU. You may ignore them after slecting the value from the
PDU, but not IN the PDU.
Ron
I think you're right, it doesn't cross the typical mind that one can
silently ignore sections of values. Apparently it did cross the mind of at
least one implementor, as it was pointed out in the IETF57
LDAPBIS meeting that at least one implementor ignores some octets in
certain PDU values.
I believe there is text in X.680 that specifies that all octets are
significant--this clarification would really just be a repetition of that
statement. Maybe another way of looking at it is: this is an item called
out in a reference document that we need to carry into the referring document
because people aren't fully reading or understanding the reference
document.
Jim
>>> "Ramsay, Ron" <Ron.Ramsay@ca.com> 7/17/03
9:48:20 AM >>>
Hi Jim,
Is there a point to the 'clarification'?
Isn't it like saying that all octets of a JPEG file are significant? Sorry, I
don't see the point.
Ron
All,
At this point it seems clear that it would be a good thing to make a
clarification to [Protocols]. That clarification would state something
like: all octets of a ber-encoded LDAP PDU are significant. No octets
may be ignored unless they are part of extended elements that are themselves
ignored.
Whether or not we want the TS to enforce mandatory minimums for
element lengths (such as attribute type names) is an issue in the data model
(and/or schema) doc(s).
Jim
|