RE: draft-ops-endpoint-mib-06.txt example syntax?

[Tim Jenkins <tjenkins@TimeStep.com>]

> -----Original Message-----
> From: Juergen Schoenwaelder [mailto:schoenw@ibr.cs.tu-bs.de]
> Sent: February 10, 2000 6:16 AM
> To: tjenkins@TimeStep.com
> Cc: mibs@ops.ietf.org
> Subject: Re: draft-ops-endpoint-mib-06.txt example syntax?
> 
> 
> 
> >>>>> Tim Jenkins writes:
> 
> Tim> I'm trying to use this MIB in the IPsec monitoring MIBs, and I'm
> Tim> running into a problem compiling when trying to indicate that
> Tim> support for DNS addresses is not required.
> 
> [...]
> 
> Tim> It compiles fine if I take out the repeated sections below:
> 
> Tim>         OBJECT  ipsecSaEspInAddressType (and the others as well)
> Tim>             SYNTAX  InetAddressType { ipv4(1), ipv6(2) }
> Tim>             DESCRIPTION
> Tim>                "An implementation is only required to 
> support IPv4
> Tim>                and IPv6 addresses."
> 
> Tim> Unless I've an earlier problem with ordering, I'd like to know if
> Tim> I'm doing something wrong, or the example text is not valid.
> 
> I think the definition is valid. Looks like a limitation of this
> particular compiler you are using.

mosy 8.1.0 (running on Linux) doesn't like it either (well, it stops at the
same line, anyway):

-->

ipsec-mon.mi2:3268: syntax error
last token read was "{"

<--

However, changing the statements to (substituting INTEGER for
InetAddressType in each)

         OBJECT  ipsecSaEspInAddressType (and the others as well)
             SYNTAX  INTEGER { ipv4(1), ipv6(2) }
             DESCRIPTION
                "An implementation is only required to support IPv4
                and IPv6 addresses."

works for both compilers.

I'm starting to wonder if the indirection of the SYNTAX from INTEGER via the
textual convention InetAddressType is going to be too confusing to
compilers, and it should be avoided.

Thanks,

Tim