You can tell BMS which partition set to load for a particular transaction by naming it in the PARTITIONSET option of the TRANSACTION definition. If you do this, and the named partition set is not already loaded at the terminal, BMS adds the partition definitions to your data on the first BMS SEND in the task.
You can also direct BMS not to change the partitions from their current state (PARTITIONSET=KEEP in the TRANSACTION definition) or indicate that you load the partitions yourself (PARTITIONSET=OWN). If you do not specify any PARTITIONSET value, BMS sets the terminal to base state (no partitions) at the time it initiates the transaction.
Whatever the PARTITIONSET value associated with the transaction, a task can establish new partitions at almost any time with a SEND PARTNSET command, except that you cannot issue the command while you are building a logical message.
SEND PARTNSET does not send anything to the terminal immediately. Instead, BMS remembers to send the partition information along with the next BMS command that sends data or control information, just as it sends a partition set named in the PARTITIONSET option of the TRANSACTION definition on the first BMS SEND. Consequently, you must issue a SEND MAP, SEND TEXT or SEND CONTROL command before you issue a RECEIVE or RECEIVE MAP that depends on the new partitions.
When BMS loads a partition set, it suffixes the name requested with the letter that represents your terminal type if device-dependent support is in effect, in order to load the one appropriate to your terminal. It takes suffix from the ALTSUFFIX option value of the TYPETERM definition associated with your terminal. Partition set suffixing is analogous to map set suffixing, and the same sequence of steps is taken if there is no partition set with the right suffix (see Device-dependent maps).
[[ Contents Previous Page | Next Page Index ]]