Home > Archive > Software Testing > June 2007 > Automation and RTL builds
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
| Author |
Automation and RTL builds
|
|
| Bennet 2007-06-05, 8:15 am |
| Hi,
I am currently thinking of building a test harness (data driven) using
QTP for automating certain aspects of an application. The harness can
handle normal UI and localized builds. However, I am unsure what i
need to take care of, if i need to use the same test cases and data to
test out RTL or Mirrored builds.
Experts out there - help!
--
- =DF
(http://www.bennet.co.nr)
| |
| Vladimir Trushkin 2007-06-05, 8:15 am |
| On Jun 5, 12:22 pm, Bennet <xben...@gmail.com> wrote:
> Hi,
> I am currently thinking of building a test harness (data driven) using
> QTP for automating certain aspects of an application. The harness can
> handle normal UI and localized builds. However, I am unsure what i
> need to take care of, if i need to use the same test cases and data to
> test out RTL or Mirrored builds.
First of all, avoid referring to UI controls by associated text
visible to the user. It will be changed in the localized version thus
invalidating tests. Inner text, used for naming objects within the
application, which are not an option for the translation, can still be
used. Targeting controls by IDs is the best choice in the case when
IDs are stable from release to release.
Secondly, when it is not possible not to use associated texts, make
sure not to hardcode strings used to address UI objects. Instead try
using resource files like in programming languages.
Thirdly, use data driven test design for the tests which use input
strings. You will most probably need to update test data for a
localized version to supply input values in another language. Data
driven approach allows easy update of test data and expected results.
----
Best Wishes,
Vladimir
| |
| Bennet 2007-06-06, 4:23 am |
| Vladimir, thanx!
I have pretty much followed what you have mentioned - however, i
wasn't sure if this is good enough for a RTL build as well. I know for
a fact that my harness model would work with any localized unicode web
based application. The harness is data driven, the scripts refer to
common functions which have been abstracted into the framework
libraries and refer to control IDs rather than labels for navigation
and actions. However the problem i foresee is that applications like
QTP and WinRunner relay on the GUI map which ofcourse will be
different with a RTL build (for eg: frames from the left would shift
to the right and so on....). Would you have any more insote to offer
to this and if so, what problems did you face and how did you metigate
the same. Basically, i am looking for a best practice when it comes to
automation harness :-).
Thanx in advance!
..bennet
| |
| Vladimir Trushkin 2007-06-06, 8:13 am |
| On Jun 6, 11:31 am, Bennet <xben...@gmail.com> wrote:
> Vladimir, thanx!
You are welcome!
> I have pretty much followed what you have mentioned - however, i
> wasn't sure if this is good enough for a RTL build as well. I know for
> a fact that my harness model would work with any localized unicode web
> based application. The harness is data driven, the scripts refer to
> common functions which have been abstracted into the framework
> libraries and refer to control IDs rather than labels for navigation
> and actions. However the problem i foresee is that applications like
> QTP and WinRunner relay on the GUI map which ofcourse will be
> different with a RTL build (for eg: frames from the left would shift
> to the right and so on....). Would you have any more insote to offer
> to this and if so, what problems did you face and how did you metigate
> the same. Basically, i am looking for a best practice when it comes to
> automation harness :-).
Well, I am not as experienced in Winrunner but I remember it allows
referring to controls via IDs, at least SilkTest does. GUI maps can
contains ID tags same way as they do for associated texts. In that
case dealing about this problem with WR is not different from what you
described.
----
Best Wishes,
Vladimir
| |
|
|
|
|
|