Executing RFT Scripts from RQM Shared Resource

Hi,

I have a set of automated RFT scripts that run fine when I execute them from my local machine. I have one master script that calls all of the other scripts, so I just need to run that one in order to run them all.

When I try and run that master script off of the RQM shared resource via an adapter, something strange happens - not all of the .class files are downloaded, and the test script execution stops after only a subset of scripts are run. Even stranger is that RQM says the test passed, when it should throw some exceptions if I try to call a script that I do not have a .class file for.

Has anyone ever had a problem similar to this? I am going to try to adjust the ordering by which the scripts are called and see if that changes anything. Thanks for your help.

I am not sure what version of RFT you are running, but I spoke with a user a few days ago who was seeing some similar behavior with RFT 8.1. They updated to RFT 8.1.0.3 and the problem went away. If you're not already on the current RFT release (8.1.0.3), that might be worth trying.

Thanks for your reply - I had upgraded to 8.1.0.3 prior to this problem. I will try and adjust the ordering of the scripts to see if the problem is with one particular script, or if there is a limitation on the number of scripts that RQM/RFT can run in this way.

Thanks again.

Okay, it appears that there is a limit of 4 .class files that can be pulled down for any 1 test case. I adjusted the order in which I called the test scripts, and every time, only 4 .class files were brought in from the shared resource.

I adjusted my master script to try and bring in all of the .class files in one shot, but that did not work - now it does not execute any test scripts, since it needs to pull down all 8 of them in order to start the actual execution of a script.

I will keep working to find a solution to this problem. I have a feeling that there is a configuration somewhere that needs to be changed.

Okay, it appears that there is a limit of 4 .class files that can be pulled down for any 1 test case. I adjusted the order in which I called the test scripts, and every time, only 4 .class files were brought in from the shared resource.

I adjusted my master script to try and bring in all of the .class files in one shot, but that did not work - now it does not execute any test scripts, since it needs to pull down all 8 of them in order to start the actual execution of a script.

I will keep working to find a solution to this problem. I have a feeling that there is a configuration somewhere that needs to be changed.

Hi, I was wondering if you could clarify the issue you are seeing. Are you saying that the adapter can only download 4 class files total per test case, or that i can only download the resources for 4 test scripts per test case?

Our scripts require multiple imports of utility classes and 3rd party JARs, and one of my biggest concerns about using a shared resource is how the adapter would handle getting all of the resources required for a script to execute. Obviously, if the adapter is only able to download 4 class files from the shared resource, we would have a problem. I am hoping that the issue is actually the later of the two, but from what I have been able to gather so far, the adapter is pretty bare bones and kind of dumb.

Btw, we have seen the same issue with RQM showing that an executed script passed when it either did not actually run, or the script actually failed. From what I have gathered, the result in RQM is based completely on the RFT log and there is no actual communication between the adapter and the RQM server.

Okay, it appears that there is a limit of 4 .class files that can be pulled down for any 1 test case. I adjusted the order in which I called the test scripts, and every time, only 4 .class files were brought in from the shared resource.

I adjusted my master script to try and bring in all of the .class files in one shot, but that did not work - now it does not execute any test scripts, since it needs to pull down all 8 of them in order to start the actual execution of a script.

I will keep working to find a solution to this problem. I have a feeling that there is a configuration somewhere that needs to be changed.

Hi, I was wondering if you could clarify the issue you are seeing. Are you saying that the adapter can only download 4 class files total per test case, or that i can only download the resources for 4 test scripts per test case?

Our scripts require multiple imports of utility classes and 3rd party JARs, and one of my biggest concerns about using a shared resource is how the adapter would handle getting all of the resources required for a script to execute. Obviously, if the adapter is only able to download 4 class files from the shared resource, we would have a problem. I am hoping that the issue is actually the later of the two, but from what I have been able to gather so far, the adapter is pretty bare bones and kind of dumb.

Btw, we have seen the same issue with RQM showing that an executed script passed when it either did not actually run, or the script actually failed. From what I have gathered, the result in RQM is based completely on the RFT log and there is no actual communication between the adapter and the RQM server.

Thanks for the reply. To answer your question, the adapter was only able to download 4 .class files per test case. I never found a solution for that particular problem.

I got around it by creating a test case for each test script, instead of having one "master" test case that called a large number of test scripts. Now, I am able to execute almost all of the test cases - there is one that runs just fine in RFT off of the local machine, but then fails when I run it off the shared resource. The cause of the failure, oddly enough, is that the verification point file for that test script does not get downloaded by the adapter.

I've also seen the problem that you are mentioning with RQM incorrectly showing that an executed script passed, when it actually failed or did not run. This sometimes happens when the adapter does not pull the latest version of the scripts off of the shared resource in RQM. For whatever reason, the adapter does not check RQM to see if there is a newer version of the source files available. To fix this, you will need to go find where the adapter is storing the source files and delete them. On the machine I used, the path was
C:\Documents and Settings\<your_username>\Local Settings\Temp\RQMRFTProjects\
Delete whatever is in that folder, and try executing your scripts again. The adapter should download the latest versions of your source files, and it should fix the problem that you are having.

Thanks for the reply. To answer your question, the adapter was only able to download 4 .class files per test case. I never found a solution for that particular problem.

I got around it by creating a test case for each test script, instead of having one "master" test case that called a large number of test scripts. Now, I am able to execute almost all of the test cases - there is one that runs just fine in RFT off of the local machine, but then fails when I run it off the shared resource. The cause of the failure, oddly enough, is that the verification point file for that test script does not get downloaded by the adapter.

I've also seen the problem that you are mentioning with RQM incorrectly showing that an executed script passed, when it actually failed or did not run. This sometimes happens when the adapter does not pull the latest version of the scripts off of the shared resource in RQM. For whatever reason, the adapter does not check RQM to see if there is a newer version of the source files available. To fix this, you will need to go find where the adapter is storing the source files and delete them. On the machine I used, the path was
C:\Documents and Settings\<your_username>\Local Settings\Temp\RQMRFTProjects\
Delete whatever is in that folder, and try executing your scripts again. The adapter should download the latest versions of your source files, and it should fix the problem that you are having.

Daniel, thank you for the quick response. From what you are saying it sounds like we're going to have to forgo the "shared resource" route unless we are willing to write a custom adapter that will be able to download all of the necessary class files and associated JAR files.

Also, I found your last comment concerning getting the latest version of a script or class particularly interesting. We actually ran into that same issue last year when we were first playing around with executing scripts from RQM 1.0. It is somewhat disappointing to see that the problem still exists. I'll have to do some research and see if a defect was ever written.

Hi,

Hi
Just put all the dependency in Folder Structure
C:\Documents and Settings\All Users\Application Data\IBM\RFT\customization
also make sure that you have not mentioned any thing under execution Argument. putting garbage data in Execution Argument make RQM show wrong status, which is a bug to be fixed in upcoming RFT build

you should be able to Run the Scripts.

Manish



I have a set of automated RFT scripts that run fine when I execute them from my local machine. I have one master script that calls all of the other scripts, so I just need to run that one in order to run them all.

When I try and run that master script off of the RQM shared resource via an adapter, something strange happens - not all of the .class files are downloaded, and the test script execution stops after only a subset of scripts are run. Even stranger is that RQM says the test passed, when it should throw some exceptions if I try to call a script that I do not have a .class file for.

Has anyone ever had a problem similar to this? I am going to try to adjust the ordering by which the scripts are called and see if that changes anything. Thanks for your help.

Hi
Just put all the dependency in Folder Structure
C:\Documents and Settings\All Users\Application Data\IBM\RFT\customization
also make sure that you have not mentioned any thing under execution Argument. putting garbage data in Execution Argument make RQM show wrong status, which is a bug to be fixed in upcoming RFT build

you should be able to Run the Scripts.

Manish