Assertion testing here refers to use of assert() calls.
For a portion of code to work correctly, certain conditions or values must be true.
Replace these assumptions with specific assertion checks.
Say that a portion of code assumes something about, for example, the current condition of an object or value of a parameter or local variable. Usually this assumption will always hold true except in the event of an error.
Make these assumptions obvious by adding corresponding assertions. As with type hinting in method parameters, these assertions can act as live documentation for your code.
As a guideline to see where your code needs assertions, check for comments that describe the conditions under which a particular method will work.
- If an assumption is not true and the code therefore gives the wrong result, it is better to stop execution before this causes fatal consequences and data corruption. This also means that you neglected to write a necessary test when devising ways to perform testing of the program.
Sometimes an exception is more appropriate than a simple assertion. You can select the necessary class of the exception and let the remaining code handle it correctly.
When is an exception better than a simple assertion? If the exception can be caused by actions of the user or system and you can handle the exception. On the other hand, ordinary unnamed and unhandled exceptions are basically equivalent to simple assertions – you do not handle them and they are caused exclusively as the result of a program bug that never should have occurred.
How to Refactor
When you see that a condition is assumed, add an assertion for this condition in order to make sure.
Adding the assertion should not change the program's behavior.
Do not overdo it with use of assertions for everything in your code. Check for only the conditions that are necessary for correct functioning of the code. If your code is working normally even when a particular assertion is false, you can safely remove the assertion.