Hello Rahul
If you have a reasonably complex application you will end up with 50% or less of business logic coding and the rest is used for error handling.
I will give you an example. Let us assume you want to read open items of a debitor from a remote system (via RFC). Now if the RFC connections fails how will your application react? Because you could not read any open items does it return to the user:
- Debitor xyz => 0 open items???
Obviously this would be wrong. Instead, you have to handle the RFC failure
CALL FUNCTION '<rfc function module>'
DESTINATION ld_target
EXPORTING
...
IMPORTING
...
EXCEPTIONS
system_failure = 1 MESSAGE ld_rfc_failure
connection_failure = 2 MESSAGE ld_rfc_failure
others = 99.
The application must inform the user that, due to an RFC failure, the open items for debitor xyz could not be read:
- Debitor xyz => reading open items failed (RFC error)
From this example you can immediately derive test cases:
(1) RFC connection ok & debitor has n open items => expected test result:
- Debitor xyz => n open items
(2) RFC connection ok & debitor has 0 open items => expected test result:
- Debitor xyz => 0 open items
(3) RFC connection failed & debitor has n open items => expected test result:
- Debitor xyz => reading open items failed (RFC error)
(4) RFC connection failed & debitor has 0 open items => expected test result:
- Debitor xyz => reading open items failed (RFC error)
I hope I could give you some input.
Regards
Uwe
Bluesky
