Forum Discussion

Fimbultyr's avatar
6 years ago

Running different Project flavors in NetworkSuite Jobs

Hi guys,


I am currently thinking about modifying my TestComplete Projects to run in several flavors like:

- nighlty=Smoke Test

- weekly=Full Regression

- RC=Full Regression+Edge Cases


However i am struggling a bit on how to set this up properly.


Currently my Project Suite looks like this:

Common Project - Commonly used scripts and tests

Product 1 Project - Product 1 specific Setup

Product 2 Project - Product 2 specific Setup

Product 3 Project - Product 3 specific Setup

Master Project - NetworkSuite Runner


The Product Projects cover one Product in its full scope.

The Product Test Items are configured as full Regression Tests like:

1. Deploy

2. Install

3. Tests organized in Folders (Those should be switchable!)

4. Deinstallation of the Product


At the Moment if you run the Master Project it Runs Network Suits of each Product (Full Regression Tests) against the Target Machines. This runs as expected but may take a long time.


Now i am thinking about a way to modify the Test Items in a Project automatically.

As far as i understood the Network Suite / Job configuration you can just run a full Project Test Item List that needs to be preconfigured, single scripts, single KeywordTests.

The unelegant way i found is cloning the Project Test Item List in a Keyword Test and let this one run in a NetworkSuite Job.

But the Results are than all glued together, which does not make traceability of each Test easy.


Do you have experience or suggestions how to handle this?

I would suspect, that you may run a script in a Job that manipulates the .mds Files of a Project. But this might me a bit dirty.




2 Replies

  • Marsha_R's avatar
    Champion Level 3

    What I've seen done is building your own test items list in an external file (Excel or .csv) and have a test runner that reads that file.


    You can have several versions of this file and just rename the one you want to run on a particular day or even have the test runner decide which one to use depending on the day of the week, etc.


  • tristaanogre's avatar
    Esteemed Contributor

    Your solution of combining the tests into a single KeywordTest is the best solution that I can give you.  As you noted, though, it does not make the tests explicitly separate in the log that way with each one passing or failing.  For that, you will need to add custom code to your test cases or to that "parent" keyword test to distinguish one test case from another.  This also means that your reporting may need to be altered to, instead of relying on TestItems to drive the report, use things like log folders (see Log.AppendFolder and Log.PopFolder) to structure your log.