Problem determination in the Presentation Manager environment is similar to that for other application environments, and proceeds through a number of phases:
a)
The documentation phase is normally part of the application testing cycle, and is performed by those responsible for such testing. No particular technical or programming skills are required for this phase.
Proper documentation of the failure usually allows a developer to determine the window procedure and message which caused the failure. This provides a useful starting point at which to search for the underlying problem.
The isolation phase is normally employed only for those failures that occur every time a particular action is performed, and involves the use of a symbolic debugging tool to single-step through the processing of the failing message, in order to determine the statement in the source code at which the error occurs.
Problems that occur only after many repetitions of a particular user action normally indicate that an operating system or Presentation Manager resource limit has been exceeded. This is usually the result of an application acquiring resources and failing to release them. In such cases, the resolution process may be expedited by immediately checking the processing of the offending message for resource allocation statements, and ensuring that each of these is matched by a corresponding statement that releases the resource.
Once the problem is narrowed down to a single application statement, the identification phase determines the cause of the problem and makes appropriate corrections to the source code. This phase requires familiarity with the OS/2 and Presentation Manager environments.
When the application has been corrected and submitted once more for testing, the problem and its resolution should be documented and this information made available for future problem determination activities. The availability of such information may be used to more quickly determine likely causes of similar problems in the future.