Character Set Selection

Based on the character set used on the platform where the agent and subagent are running, you will encounter one of the following three scenarios:

• Both run on an ASCII based platform.

  In reality a lot of platforms use the ASCII character set. For those platforms, just use the native character set. In that case, the native character set is ASCII.

• Both run on the same non-ASCII based platform.

  It is expected that the agent and the subagent normally run on the same machine or at least on the same platform. In that case, it is easiest to use the native character set for data that is represented as strings. If such native character set is not the ASCII character set, the agent must translate from ASCII to the native character set (and vice versa) as needed.

• One runs on ASCII based platform, the other on a non-ASCII based platform.

  If the agent and subagent each run on their own platform and those platforms use different native character sets; for example, IBM OS/2 uses ASCII and IBM MVS uses EBCDIC, you must select the ASCII character set, so that you both know exactly how to represent string-based data that is being send back and forth. The entity that is not ASCII based must do the translation from ASCII to the native character set (and vice versa) as needed.

When the DPI subagent sends a DPI OPEN packet, it must specify the character set that it wants to use. The subagent here needs to know or determine in an implementation dependent manner if the agent is running on a system with the same character set as the subagent. If you connect to the agent at loopback, localhost, or your own machine, you might assume that you are using the same character set. As long as you are just using OS/2 on an Intel based processor, it does not matter. Always use the native character set, which is ASCII.

The subagent has two choices:

DPI_NATIVE_CSET

Specifies that you want to use the native character set of the platform on which the agent that you connect to is running.

Specifies that you want to use the ASCII character set. The agent will translate between ASCII and the native character set as required. If the subagent is on a non-ASCII based platform, it may have to translate also.

The DPI packets have a number of fields that are represented as strings. The fields that must be represented in the selected character set are:

• The null terminated string pointed to by the description_p, enterprise_p, group_p, instance_p, and oid_p parameters in the various mkDPIxxxx(...) functions.

• The string pointed to by the value_p parameter in the mkDPIset(...) function, that is if the value_type parameter specifies that the value is an SNMP_TYPE_DisplayString or an SNMP_TYPE_OBJECT_IDENTIFIER.

• The null terminated string pointed to by the description_p, enterprise_p, group_p, instance_p, and oid_p pointers in the various snmp_dpi_xxxx_packet structures.

• The string pointed to by the value_p pointer in the snmp_dpi_set_packet structure, that is if the value_type field specifies that the value is an SNMP_TYPE_DisplayString or an SNMP_TYPE_OBJECT_IDENTIFIER.

Related Information

The mkDPIopen() Function