Testing: A brief summary of white box testing

We all know white box testing by many other names as well such as glass box testing, clear box testing, structural testing, transparent box testing etc. This testing methodology is used for testing the internal structure of the software or to be precise to test how the application actually works and whether it works as desired or not. This is just the opposite of black box testing that is responsible for testing the system’s functionality only, not getting into the internal structure. The test cases for the white box testing are designed based up on the programming skills and internal perspective from which the system is viewed. From the various inputs, some are selected for exercising all paths through the program and determining whether they are producing the desired outputs or not.
A number of errors can be discovered using this testing methodology. But it is not capable of detecting the parts of the specification that have not been implemented or are missing. The following testing techniques are included under the white box testing:

- Data flow testing
- Path testing
- Decision coverage
- Control flow testing
- Branch testing
- Statement coverage

This testing methodology is used for testing the source code of the application using the test cases that are derived using the above mentioned techniques. All these techniques are used as guidelines for the creation of an error free environment. Any fragile piece of code in the whole program can be examined using white box testing. These techniques act as the building blocks for white box testing and its essence lies in carefully testing the source code so that no errors occur later on. Using these different testing techniques all the paths through the source code that are visible can be exercised and an error free environment can be created. But the tester should have the ability to know which path of the code is being tested and what the output should be. Now we shall discuss about the various levels in white box testing:

- Unit testing: White box testing done during this phase ensures that code is working as desired before it is integrated with the units that have already been tested. This helps in catching the errors at an early stage so that later it does not becomes difficult to find them when the units are integrated and the system complexity increases.
- Integration testing: Test cases at this level are meant to test how the interface of various units interact with each other i.e., the system’s behavior is tested in an open environment. During such testing any interaction that is not familiar to the programmer comes to light.
- Regression testing: This phase uses the recycled test cases from white box testing.

For the white box testing to be effective the tester should understand the purpose of the source code so that they are tested well. The programmer must understand the application so that he/ she can create effective and appropriate test cases.
The programmer working on preparing the test cases needs to know the application very well in order to ensure that he / she can prepare cases to handle every visible path for testing purposes. With the source code having been understood, this understanding of the course code lets the team use this analysis of the source code for preparing test cases. And for making these test cases, these are the steps that need to happen in white-box testing:
- Input, that may include functional requirements, specifications, source code etc.
- Processing unit, that carries out the risk analysis, prepares a test plan, executes tests and gives results.
- Output, that provides the final result.