kurt

I am just looking for a better writing method to improve myself. So, the situation is this (i'm writing a test script to catch errors, where the situation is that I am expecting an exception to occur. It must):

1. I hide the implementation, but the idea is there
2. Is there a better and more elegant way in C#?

ScottyB

With exception handling in C# the try block is where you put the code that you expect may produce an exception.

The catch clauses are used to handle exceptions that are generated from within the try block.

The catch clause section can contain multiple exception handlers. When an exception occurs the catch clauses are searched in order from first to last. As a result when writing your catch clauses you should put the most specific exception handlers first followed by the more general ones.

In the example below if the specific handlers cannot handle the exception it is caught by the last handler. In this example I have caught the exception message. You can choose what to do with this message after the calling block.

Whether or not an exception occurs the finally block is always run. It is usually where your clean up code goes. You can put your code to close database connections or files that you have open in the try statement.


string errorMessage = String.Empty;

int d = 0;

try
{
// Generated Exception for Demonstration Purposes
float f = 1 / d;
}
// Catch specific exception firsts and work your way to the more general ones
catch (DivideByZeroException ze)
{
errorMessage = ze.Message;
}
catch (IOException ie)
{
errorMessage = ie.Message;
}
catch (SqlException se)
{
errorMessage = se.Message;
}
catch (Exception ge)
{
errorMessage = ge.Message;
}

finally
{
// Clean up code here
}

return errorMessage;


I hope this helps

Regards

Scottyb

kurt

Yeah, it sort of helps. Thanks Scottyb.

But my real intention is the see whether my command returns an exception or not? Like i said, i am writing a test function, and that particular command must return an exception as it is no longer supported. If it did not, i render the test as "failed".

My method works, but I want to learn to write elegant code. I'm trying my best. But I feel learn by example is the best.

xavier

how about throwing some code our way ? maybe we'll have better suggestions.

__________________
Don't take life too seriously, it's not permanent !

mbd

if this is a test method you should be using NUnit. are you? if so, on the test method you can place an attribute called ExpectedException. NUnit does the rest.

kurt

Yes, indeed. I am using NUnit.

Initially I thought that [ExpectedException] just makes sure if that exception occurs, the test still passed. I didn't know it does check if the exception did not occur, the test fails -- this is what I wanted !!

Thanks for pointing out, else I would not have looked into the manual to check.

Thanks.

kurt

Although I get a clean elegant way with NUnit, I still would like to know how to write it better if the environment was not NUnit. So here's an example of my code

xavier

Well, you could use this from nunit : http://www.nunit.org/index.php?p=exception&r=2.4.3
in order to make the code smaller :

If the method does not throw, the test fails. However , you should be careful when using the ExpectedException attribute. It's ok for a one line test, but if you had several methods and another method would throw that exception, the test would still pass.

An improvement to that has been made in xUnit where you can do : Assert.Throws(method).
Best of luck.

__________________
Don't take life too seriously, it's not permanent !

mbd

beauty is in the eye of the beholder