[SNMP4J] SNMP4J Agen Pro Generation

[Frank Fock]

Hello Ron,

Ron Fulkerson wrote:
> Hello All,
> 
> I have a number of questions about using snmp4j along with the Agen Pro 
> code generator.  I hope someone here will not mind giving a bit of their 
> time to enlighten me of some good resources to help me get started.
> 
Just a small info beforehand: AgenPro 2.5 which will be released
on 12. June will include a instrumentation How-To for the stub
code generated for SNMP4J-Agent.

> I am looking for a resource which explains how to work with the code 
> generated by Agen Pro v 2.1.  Specifically, I would like to see an 
> example of working with dynamic tables and values with read-write access.
> 
> The code generated by Agen Pro produces stubs to override, including one 
> for isValueOK() and a ValidationListener.  These two stubs seem to be 
> redundant in their purpose.  The generated code also produces a stub for 
> get(), yet I do not see the rational for overriding the default 
> implementation.

In v2.5 AgenPro will provide an option where you can choose to
overwrite isValueOK or to use a ValidationListener. These two
concepts can be used alternatively. The second approach is
recommended when the ManagedObject can be implemented without
extending SNMP4J-Agent MO* classes.

The rational to overwrite the get() method is to be able to
update the internal value of a ManagedObject when its value
is being requested.

> 
> Also, I am not sure I understand the rational behind making each OID in 
> the generated code private.  These values seem like they are needed in 

AgenPro v2.5 will generate "public" constants by default, but that
can be configured for each object/sub-tree individually.

> creating a notification originator.  I would think referencing a public 
> instance of these values would be the best approach when creating this 
> class.
> 
> I understand that I can change the generation templates to produce 
> whatever I would like, but I assume these templates were modeled after 
> SNMP best practices.  It is my desire to understand these practices and 
> follow their guidelines.  Any pointers in this area would be great.

It depends on how such a MIB module implementation is used.
When it is used in combination with other MIB modules a public
constant is useful, otherwise it might be better to hide
the (then) internal constants.

> 
> Finally, I have not yet figured out how to populate the tables the are 
> created in the generated code.  Any pointers in this area would be much 
> appreciated.

To populate a MOTable you can use the following scheme:

MOTable table = new DefaultMOTable(...);
OID newIndex = indexVariable.toSubIndex(false);
MOTableRow newRow = table.createRow(newIndex);
// probably change some column values
...
table.addRow(newRow);

Hope that helps.

Best regards,
Frank

-- 
AGENT++
http://www.agentpp.com
http://www.mibexplorer.com
http://www.mibdesigner.com