About execution procedures
There are a total of ten operations possible on objects. For each
operation that is valid for an object class, the OSA handling
that class must provide an execution procedure to carry out the operation.
Eight of these operations are available through the object API.
The first five were introduced above, since they operate on attributes:
-
get
-
replace
-
replace with default
-
add
-
remove
Three others operate on an entire object:
create-
create a new object of the class
delete-
delete an object
action-
perform a class-specific action. For example, an action
available for floppy disks is the "format" action.
Two operations are called internally by the framework:
filter-
filter a list of objects based on the values of their attributes
list-
make a list of objects that are contained by another object
(containment is described in the next section)
The decisions that need to be made with respect to object
operations are as follows:
-
For the five attribute operations, you must implement a procedure for
the operation if it is a valid operation for any of the
attributes. You can choose to implement the procedures in two ways:
per attribute-
the procedure is called for each attribute on each
object of the operation. In our example, if someone wanted to get
all the attributes for two UUCP system objects, the get procedure
would be called twelve times (two objects multiplied by six
attributes per object).
per object-
the procedure is called only once for each object, and is
expected to handle all attributes for the object in that call. In
the example above, the get procedure would only be called twice, once
for each object, but would have to return all the requested
attributes in each call.
Next topic:
Step 6: Implement the execution procedures
Previous topic:
Execution Procedures screen
© 2004 The SCO Group, Inc. All rights reserved.
UnixWare 7 Release 7.1.4 - 27 April 2004