Bancroft Scott <baos(_at_)oss(_dot_)com> 02/03 4:16 pm >>>
On Tue, 3 Feb 1998, Darren Harter wrote:
e.g.
Test ::= SEQUENCE {
genName [0] IMPLICIT GeneralName OPTIONAL }
This example violates X.680 clause 28.8, (X.208 clause 26.10) which says
that you cannot use IMPLICIT when tagging a CHOICE type.
Notice that when CHOICE types and open types are defined as tagged types
(such as in the example above) they are always explicitly tagged. Further,
any reference to them must include a tag if the tagging would otherwise be
ambiguous. Since in the above example GeneralName, though OPTIONAL,
cannot be confused with any other type, the [0] tag is not required.
Yes, sorry it was a silly example. The reason I put that in was to show
that even if an IMPLICIT keyword were accidentally placed with an object
derived from CHOICE, it would be encoded explicitly.
The important thing to note in this is that the EXPLCIT for CHOICE rule
is not recursive, so the rfc822Name in GeneralName is NOT explcitly
tagged.
I am not sure what is meant by "not recursive", but the conclusion is
correct.
What I meant by "recursive" is that just because GeneralName is explicitly
encoded doesn't mean that any other tags in subordinate structure will be
explcitly tagged as well. Hence rfc822 being implicit not explicit.
But what does X.208 say?
It says the same thing as X.680,
Good, I ask this ask there are often emails saying things like: "Were using
X.208 not X.680 so that doesn't apply". I just thought I'd check in this
case as I tend not to use X.208.
Regards,
Darren Harter
p.s. I wonder how long it will be before the Named Bit List encoding queries
surface again? ;-)