Hello,
I've recently put some tests on Jenkins to run after a project builds.  The odd thing is that even if 1 test fails the build simply continues to move on to the next test regardless of the results of the previous.

Is there a way for the build to immediately stop if any failure is detected?  So that any issues can be dealt with at a quicker pace.

My project has a shared test-as-step that each of my other tests call.  This test-as-step has the following override that is attempting to use properties of the testresult parent/calling test name to add to the custom log we've created:

public override void OnAfterTestCompleted(TestResult result)
{
base.OnAfterTestCompleted(result);

string parentName = Convert.ToString(result.Parent.Name); //line 103
string parentId = Convert.ToString(result.Parent.Id);
string parentRootName = Convert.ToString(result.Parent.RootName);
string fileName = Convert.ToString(result.Parent.FileName);

\\custom code to write to the log
}

Each line in the above override that references the testresult.parent.* properties test trigger the following error:
System.NullReferenceException: Object reference not set to an instance of an object.
   at CIS.TestStudio.WebAccess.Shared.Handler.OnAfterTestCompleted(TestResult result) in c:\tfs\CoreCM\Root\CIS.TestStudio.WebAccess\Shared\Handler.tstest.cs:line 103
   at ArtOfTest.WebAii.Design.Execution.ExecutionEngine.InternalExecuteTestIteration(Object codeBehindInstance, Boolean isDataDriven)
   at ArtOfTest.WebAii.Design.Execution.ExecutionEngine.ExecuteDataDrivenTest(Test test, BaseWebAiiTest codeInstance)
   at ArtOfTest.WebAii.Design.Execution.ExecutionEngine.InternalExecuteTest(Test test, TestResult initializationResult)
   at ArtOfTest.WebAii.Design.Execution.ExecutionEngine.ExecuteTestInCurrentContext(Test test)

If this is not the correct way to access the name of the parent/calling test properties, is there a better way to do so through code?

Last year, I saw there was a beta release.  I did not personally try the beta, but I would have if I had known the final release gave me several headaches after its release. 

I am working on a project that is requiring lots of coded steps for various reason.  The coded steps are failing intermittently because they are processing too quickly.  I have added numerous wait.forexists() and wait.fornomotion() throughout, but they don't seem to do what I need.  If I add sleeps throughout, it works, but this is a bad solution.  

The specific scenario I'm having trouble with is after logging in, a grid is created with dynamic header values.  I'm selecting a value from the grid based on a value in one of the dynamic columns.  First step is to find location of the proper column by searching through all headercells to find the matching column index.  The issue I'm running into is the following lines of code executing before the radgridview is fully loaded resulting in a zero column count being returned:

public int globalTimeout = 10000;

public void login()
{
Pages.TestPage.SilverlightApp.LoginButton.User.Click();
Pages.TestPage.SilverlightApp.LoginPage.Wait.ForExistsNot(globalTimeout);
Pages.TestPage.SilverlightApp.HomePage.Wait.ForExists(globalTimeout);
Pages.TestPage.SilverlightApp.HomePage.Wait.ForNoMotion(globalTimeout);
}
public int findColumn(string value)
{
RadGridView grid = new RadGridView();
//if I add a System.Threading.Thread.Sleep(10000) before this grid is initialized, it works every time, if I don't, I randomly get a zero count of headerCells even though they exist
grid = Pages.TestPage.SilverLightApp.Test_RadGridView;

//If I add a grid.Wait.ForExists(10000) here, it still moves to the next line almost instantly as the grid exists but the data doesn't necessarily exist yet
 //If I add a grid.Wait.ForNoMotion(10000) here, it still doesn't seem to matter

int cols = grid.HeaderRow.HeaderCells.Count;

Assert.IsTrue(cols != 0, "No header columns were found.");
//code code code
return columnNumber;
}

Any suggestions on how to better handle (through code!) the wait for radgridview to populate/ready for indexing? 

Hi !!

First of all I have to say I don't speak english very well so sorry if you don't understand exactly what I want to say. I'll try to explain myself as well as I can.

I have seen in other discussion something about load testing, and now I'm looking for some information about that for my job. I would like to know if there is an ideal response time for logging tests. For example, I'm trying to logging in an app with 100 users at the same time, and I don't know if one second of response for each user could be considered a good response time. Is there any 'rule' for that?

Thanks and I hope you could understand me.

Regards.

Chris.

I am farily new to using Test Studio and am curious as to how I can validate when a new row of data is added to a grid in my Java application.

The application consists of two panes.  The left pane is a data grid and the right pane is for data entry.  I enter data in the right pane and click "Save."  This creates a new row in the left pane.  How do I validate that the new row has been added to the grid in the left pane and then validate the values in the cells of that new row?

Any help would be great appreciated :)

Thanks.

Chris

I am relatively new to test studio and i am trying to store a collection of links that i get from Find.AllByAttributes("id=
~linktext");

Then the idea is to click one from the list. Currently i am trying this code to get the last:

HtmlAnchor dellink= alllinks.Last();
dellink.Click();

I am having issues on creating the list containing the elements and then casting the link into the
HtmlAnchor... Any ideas?

Thanks and Regards.

I have a test the goes through and cleans-up all the data I generated during my testing.  Deleting test users and accounts is a major part of this.  I have these tests scheduled to run overnight so we get our results first thing in the morning.  The test run however gets to the delete user part and fails to handle the dialog. One would think after the wait time experiences and the element is still not found the test would fail that case and move on.  This is not what is happening. 

When I come in in the morning and launch the execution machine the test is just chilling with the Dialog on screen. When I manually click it the test continues.  I then have to wait for the entire schedule queue to process before I can schedule another run to try and debug it.  

On top of this when I get my results email it tells me 0 of 0 tests run.  But I just watched a bunch of them run.  All the tests in queue have the same results even though I am watching them run.  


Seems like there is a bug in the execution server that does not end and fail a test after the time to find or interact has passed.  Also another bug where when a scheduled test does not run when it was scheduled and eventually runs when it comes up in the queue why are the results saying 0 of 0 run when you can plainly see they did run.   If you are queing scheduled runs then the results of the run regardless of the time should be accurate and meaningful. 

I am using the latest internal build do to the issue with edits to elements or test cases not propagating in Test Lists.  A new issue has crept up for us when working with frames. 
In order for our tests to run in our various environments and to get around framebox IDs we edit the properties of the framebox.  We typically remove the frame name and add a "~" to the href as our base URL changes slightly depending on which environment (DEV,QA,Stage,Prod).  This works great we can run it in all environments as well as across our different clients personalized URLs without having to keep a bunch of one off test cases.  

The issue I am experiencing is when I make these edits to the properties they do not stick even if I save the project before I try to run it again.  I have found that I can make my changes and then save it but I have to refresh the entire element library in order for my edits to take hold.  We have a large element library so it takes time to refresh before we can confirm if the modification to the test worked.  This adds to much overhead having to refresh and wait and hope it works so we dont have to try again.    There are also times when doing these edits crashes the entire Studio and we have to restart which is very frustration for anyone.

Has anyone experianced this?  Is there a workaround that is possible? 

Hi all,

We have a web application that uses Frames and uses a Flash map for navigating around different regions which I wish to setup for testing. The region links are controlled via an XML file stored on the server.

I understand Test Studio does not fully support Flash objects and the best way around this is the blind clicking method. I have tried this however, and unless i'm doing something wrong, it doesn't work.

I've tried to put the link from the XML directly into the test step but because of the way the application works, the header is not loaded, only the body of the page.

Is there any way around this?