Eclipse Community Forums: Jubula » Data Entry options

Help

Home

Home » Archived » Jubula » Data Entry options

Show: Today's Messages :: Show Polls :: Message Navigator

Data Entry options [message #774311]

Tue, 03 January 2012 17:16

Chris Missing name

Messages: 32
Registered: June 2011

Member

I have some concerns with using the standard data entry methods in Jubula for providing test suite level data entry to my tests. I am hoping to share my concerns as well as my approach, to get feedback and weed out any issues I may have in my hackish method.

My concerns with Jubula Central Data Sets is that they are difficult to organize, cannot be modified outside the IDE, and give no way to parameterize values that are logically the same across data sets. This forces multiple points of maintenance should a value change and overall seems clunky. Using Excel sheets addresses to an extent the organizational issues and external modification.

An approach I am considering is initialization of all test suite level data through environment variables set by our setup and teardown script. This would allow us to organize the data by test suite rather than test cases, external modification of test data, parameterize test data for a single point of maintenance, and near painless ant integration. My only concern with this approach is that it is somewhat hackish and not something considered standard but it does accomplish everything we need. Are there any obvious issues with this method that I am not aware of, or perhaps something I have gotten incorrectly with central data sets? Central data sets do not seem usable to me in their current form so I cant help but feel I may have missed something.

Report message to a moderator

Re: Data Entry options [message #774761 is a reply to message #774311]

Wed, 04 January 2012 14:48

Alexandra Schladebeck

Messages: 1613
Registered: July 2009

Senior Member

Hi Chris,

Thanks for your post.

You are right, variables that change depending on the environment (paths to directories, filenames etc) are not best handled in central data sets. These are more designed to handle repeating data (e.g. default customer details that come up again and again) at one central place.

For our platform-specific variables, we initialize these as environment variables before the test and just call them using $VARIABLE_NAME in Jubula. This has the nice effect that we only need to write the data once, but has the disadvantage that testers manually starting the automated tests e.g. during the day to reproduce bugs have to make sure that they have declared and initialized these variables (either as environment variables or as variables stored by Jubula itself). This can be time-consuming, as there is no way to find out whether a variable is referenced without being initialized until runtime. If you're going to go this way, I'd recommend having naming conventions for variables that have to be set (e.g. EXTERNAL_...) so you can search for them easily in a test to make sure you get them all first time.

There are a couple of ideas we've thrown around to combat this problem: one of these would be to have variables declared for AUT configurations in the Project settings. We'd still have to have some way of making sure that all variables are declared and initialized that are required in the test, or at least ensure that each AUT configuration is required to declare the same variables.

If that's the kind of direction that would be useful, feel free to enter an enhancement. If not, feel free to elaborate Smile

Regards,
Alex

Report message to a moderator

Previous Topic:	Waiting for end of Eclipse Job
Next Topic:	Add multiple sub folder under Test Suites folder

Goto Forum:

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

]

Current Time: Thu Apr 18 13:26:37 GMT 2024

.:: Contact :: Home ::.

Breadcrumbs

Sign up to our Newsletter