I have an accounting & amp; Payroll client / server application where there are many input forms with complex data validation rules. I'm getting an effective way of testing the user interface unit tests.
For complex verification rules I mean:
- "Disable the X button if I enter a value in textfield y" If I include a value in the textfield, then enable a combobox "...... ......
The most promising pattern I found was M. Suggestion by Fowler () has been suggested.
Do you have any experience with the user interface unit testing? As a Technology Stack I am using: .NET 3.5 & amp; Windows Farm Widget Library
I will not say this at all "unit testing", but I have automated testing against a WinForms UI With a little success, and also use WatiN in web UTI.
Assuming that you can handle the window of the application that you want to test, you should be able to scrip a lot of code to test the functionality of the user interface < / P>
Many people condemn the idea of trying to run an automated test against the UI, because it is enough that you can not test that way. For example, an automated test does not appear that a font is bad or some text is confusing or a button is a little too far, for these types of things you definitely need an intelligent person Who are looking at the screen.
However, isolation of that type of test, definitely a large array of repeated trials can be done automatically and performed regularly. One of the biggest applications is a complete batch of regression testing scripts, which must be performed manually whenever a new release door will go out. This test is usually something you can train a monkey, just click on this link to enter a list of instructions, enter some text, click this button, check out the resultant message etc. These things are terrible waste of time in your QA tester, and make them unhappy, so if they can be automated, great tests of these types should be enabled to run automatically by your build server every day, and somebody It can also be prepared more thoroughly than manual testing.
Again, this strange unexpected thing will not be found, but it will definitely give you a certain level of confidence that your small change has not broken any other screen which you have never heard of.
Valid, it works more for developers, because small changes in the application can break the test of stupid reasons, such as any automatic test, but it will give you a ton of time It should be avoided in testing and debugging whether it is desirable for you to decide, but I think it is an idea that you would not normally want to waste any time soon. E.
Comments
Post a Comment