agentgen handling of AUGMENTS clause

[Alex Finogenov]

Frank,

You are right. In the paragraph 7.8 SMIv2 indeed says that "AUGMENTS clause
names the object corresponding to the base conceptual row".

Since we are discussing the AUGMENTS clause, I also noticed that agentgen
generated the ilen value of 1 (the last argument to table entry's
constructor in .cpp files) when, according to my understanding of the
agentgen, it should have generated a larger value reflecting the number of
indexes for the tables and the size of the index_info array. At the same
time it generated the index_info arrays of the correct sizes and with the
correct number of entries in each case.

All of the entries in question contain AUGMENTS clause that refers to a base
entry that contains a few INDEX entries. In every case one of the INDEX
entries is local to the table and the rest are not local to it, i.e. either
imported from other modules or are defined in other tables of the same
module. For example, when an entry AUGMENTS a base entry with  one INDEX
defined in that table, two INDEXes imported from two different modules, and
one INDEX defined in the considered module in some other table, the agentgen
generated the ilen value of 1, index_info array of size 4 that contained 4
entries.

Thanks,
Alex



-----Original Message-----
From: Frank.Fock____t-online.de [mailto:Frank.Fock____t-online.de]
Sent: Wednesday, July 11, 2001 4:11 PM
To: Alex Finogenov
Cc: 'agentpp-dl____agentpp.com'
Subject: Re: agentgen handling of AUGMENTS clause


Hi Alex,

It *is* illegal to define an augmenting table that is at the same time
base table to another table. So, sequences of augmentation is
not allowed in SMIv2. AgentGen / MIB Designer should reject
such constructs during parsing. I will add such a check for the
next releases.

Best regards,
Frank

Alex Finogenov wrote:

> Hi,
>
> I have found that if a table entry AUGMENTS another table's entry, which
in
> its turn AUGMENTS yet another table's entry, the agentgen generates in the
> .cpp file an EMPTY
> index_info "ind...Entry" array with the size 1.
>
> It does not seem to be illegal to do this "double AUGMENTation", so is
this
> a bug in the agentgen?
>
> Thanks,
> Alex

--
Frank Fock - AGENT++
Email: fock____agentpp.com
Fax: +49 7195 177108