Is the data being corrupted by bad programming logic?
To find out whether data is being corrupted by bad programming logic in the application, consider the flow of data through the transaction.
You can determine the flow of data through your transaction by "desk checking", or by using the interactive tools and tracing techniques supplied by CICS®.
Desk checking your source code is sometimes best done with the help of another programmer who is not familiar with the program. Such a person might see weaknesses in the code that you have overlooked.
- CEDF is a powerful interactive tool for checking your programming logic. You can use it to follow the internal flow from one CICS command-level statement to another. If necessary, you can add CICS statements such as ASKTIME at critical points in your program, to see if certain paths are taken, and to check program storage values.
- You can use CECI to simulate CICS command statements. Try to make your test environment match the environment in which the error occurred as closely as possible. If you do not, you might find that your program works with CECI, but not otherwise.
- You can use CEBR to look at temporary storage and transient data
queues, and to put data into them. This can be useful when many different
programs use the queues to pass data.
When you use CEBR to look at a transient data queue, the records you retrieve are removed from the queue before they are displayed to you. This could alter the flow of control in the program you are testing. You can, however, use CEBR to copy transient data queues to and from temporary storage, as a way of preserving the queues if you need to.
For programming information about how you can invoke user tracing, see .
You can use CSFE storage freeze to freeze the storage associated with a terminal or a transaction so that it is not released by a freemain request at the end of processing. This can be a useful tool if, for example, you want to investigate possible storage violations. You need to get a transaction dump to look at the storage after you have run the task with storage freeze on.
For long-running tasks, there is a possibility that a large amount of storage may be consumed because it cannot be released by a freemain request while storage freeze is on. For short-running tasks, however, there should be no significant overhead.
If, after using these techniques, you can find no fault with the logic of the program, the fault either lies with the way data is mapped to the terminal, or you could have missed some important evidence.