multi version application automation - TOSCA

Question

Is there a way to manage multiple versions of TOSCA modules with same test cases. Please consider scenario below:

There are 3 versions of applications that are in production. For each version there are small to mid level changes in modules scanned for automation. We have to maintain and execute test cases on each version using TOSCA. We know that we can create multiple copies of test cases and modules to manage this but that would create multiple working copies of automation test steps and modules.

Is there any efficient way to avoid duplicity of test cases and efforts?

score 2 · Accepted Answer · answered Jul 19 '18 at 14:04

Tricentis Tosca 11.2 introduced Branching and Merging to support testing different versions of the system under test.

This feature is similar to version control you would use for the application. In that way you could have a branch for every version of your application.

Creating and Managing Branches in Tricentis Tosca

score 0 · Answer 2 · answered Sep 05 '18 at 14:02

If you're not running 11.2 as mentioned in MartinThe's answer, you can try this...

So, I manage my test scripts such that they can run in multiple test environments - Dev, SIT, UAT, etc. I start each script with a folder called "Prepare Test Data", and within that folder is usually just multiple Set Buffer steps. This folder is the key to driving the differences between environments.

To start with, I use a Test Configuration Parameter called "TestEnvironment". Because this is a test configuration parameter, it can be attached directly to test cases or test case folders, and it cascades down onto any test cases contained within... So, I set the TCP to "SIT" in my blue section because I develop all my automation scripts using my SIT environment to play around in. When Executing, I also use TCP "TestEnvironment" on my execution lists(Green section), because TCPs found there override TCP's from the blue section. So, although my test scripts are set to "SIT", my execution lists can be set to "DEV" or "UAT", and while executing via execution list, they run with the TCP set to that.

Next, within the "Prepare Test Data" block, I set up my switching buffers. The plan here is for all of the test steps within the script to just use simple basic buffers like "Username", "Instrument Ref#", etc. So, the Prepare Test Data block preloads those buffers with the data for the test, but it's switched depending on which environment I'm running the script in. I'll use the buffer "Username" as an example...

In the first Set Buffer block, I create three buffers like this:

Username-DEV = "Fred"
Username-SIT = "TestUser1"
Username-UAT = "TestDude"

{(*advanced note) Sometimes my scripts are more advanced and I pull data from Test Case Design test sheets... In that case, the above would look something more like this: Username-SIT = "{XL[Test Data.Users.Clerk User.SIT]}", where the datasheet would have all the different users, usually stored in TCD classes... I could discuss that in a different topic...}

In the next Set Buffer step, I then switch the buffer to it's final value that the test steps will use like this:

Username = "{B[Username-{CP[TestEnvironment]}]}"

So, in this example, if TestEnvironment is set to SIT, then Username gets set to the value of Username-SIT, which is "TestUser1" . This switching happens live during the test run.

(*Advanced note:) You have to use separate Set Buffer steps because Tosca seems to execute all parts of a single "step" sequentially-simultaneously. What I mean by that is that the buffer for "Username-SIT" won't exist until the first Set Buffer step completes, so if you combine the assignment for Username buffer in the same step, it will crash.(or use an old, possibly wrong value)

Below this, all of my test steps can use basic buffers like "Username" instead of "Username-SIT" because they've all been "switched".

Using this technique, and to specifically answer your question, I would also use Tosca IF blocks to separate modules being used. For example, in my test enviroments, one of my login screens between SIT and UAT is so different that I can't use the same module for both. So, a simple IF block can switch it.

IF
   Condition
      TBox Evaluation tool
         Expression = "{CP[TestEnvironment]} == "SIT"
   Then
      SIT-version of module
         Username = "{B[Username]}"
   Else
      UAT-version of module
         Username - "{B[Username]}"    <-- can use same buffer because its' already switched

Naturally, you can stack "IF-ELSE-IF" blocks if you need more than two environments/versions. I hope you get the idea.

multi version application automation - TOSCA

2 Answers2