Mailingliste - Einträge 2006

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: [qftestJUI] Proper use of setname

  • Subject: RE: [qftestJUI] Proper use of setname
  • From: "Michaelis, Mark" <mark.michaelis@?.com>
  • Date: Mon, 26 Jun 2006 08:38:05 +0200

Hello Mark,
I had a different approach here: I started to write tests with a lot of pre-existing setNames in the code. At other places they were missing and sometimes there were names which make it even harder to write tests as there were different names for nearly the same component. Perhaps the last is the most interesting one: There is a way to specify names which may even make it harder to write tests. As workaround I wrote a ComponentNameResolver which maps the different names to one unique name.
I think best practice is to go the XP way, i. e. only add names if you need them. If you add a name to every component you might end up with:
  • a lot of unused names which still need to be maintained
  • and problems during development to find unique names (inside a given frame)
The one who writes the tests knows best what would be a proper name and would immediately see if there are double names as the test will fail.
I don't think that dynamic names are a good solution. As you already mentioned refactoring can easily break the tests. There is only one case where I used dynamic names: When setting names for the dividers in JSplitPanes I implemented a calculation of the name for these dividers based on the names of the panes which belong to the JSplitPane. This was in the "before-qftest-phase". Nowadays I would write a ComponentNameResolver for this.
I hope I could give you some ideas how we handle it over here. It would be interesting to know if there are other best-practices.
Kind Regards,
Mark Spielman wrote:

Our organization is beginning to use qftestJUI to test two of our GUI applications.  However I am finding it hard to understand when, where and how to appropriately use setname to make our lives easier?  We currently do not have extensive use of setname in either of our two products.  We have tried two approaches and both seem to work but are not necessarily ideal.  We are only in the early stages of test development.  What I was wondering is if there are general guidelines we can use to determine how to properly use setname to make our lives easier.  I’ve read the manual and am searching for slightly more detail, something along the lines of a best practices technical note or something?

The two approaches we’ve tried are:

  • Using java reflection to determine component names based on the coded variable names of the different components. In this case all components have names.  Here there’s a potential for a component to change names if ever a developer refactors the code.
  • Manually adding names to components as we discover difficulties in qftestJUI test implementation.  Components are named as needed but development of tests seems very slow.

What I’m wondering is if you were going to start a project from scratch, what is the ideal way to use setname to make testing and test development most efficient.




Mark Michaelis
Software Engineer Quality Assurance

CoreMedia AG
Ludwig-Erhard-Str. 18
20459 Hamburg, Germany

Any Content, Anywhere, in a Trusted Universe:
CoreMedia CMS and CoreMedia DRM

Attachment: smime.p7s
Description: S/MIME cryptographic signature