| || |
The recognition of third party objects and the use of enablers in coded programs.
User Jas Singh (Jas@mercia.com.nospam) posted:
Please can you help us ?
We need more information about this subject area.
We are still not certain about how specifically we would deal with the
incorporation of enablers in the code and how this would work ? Are we
correct in assuming that we would need to incorporate code into the
functions, calls and procedures where the programs are being called ?
The problem that we face is whether we need to write additional code in
order that the Rational Testing software can recognise the third party
objects and DLL's as well as those objects that are embedded into the
application but originate from a third party organisation other than the
language producing company.
Also we need to know to have access to the components required in order to
recognise Java and ActiveX components in accordance to the customisable
Proxy mechanism. If we are required to add ActiveX enablers at various
levels of the application (Both visually in the C++ form designer and
programmatically in a CAX Window) as would happen in VB, Java, HTML and
Power Builder as well. And writing proxies for our controls. This would
not help as some of our proxies may never call and recorded script is
always Generic Object. (Which we can not use because some of them use
coded co-ordinates and we need to able to test on several versions of
windows and user configurations).
For example, I know that that all of the current "standard" Java Swing and
AWT objects are supported, but I don't know who decides which of those
objects are "standard". If that makes any sense at all. You need to get
information in order to see what is generally supported, and what add-ons
are available from rational for additional areas that they support.
Does any one in your organisation know where they have achieved this
aready ? We may be doing this differently or more effectively than the
way I have described above ? Does any one have experience of doing
something similar ? Thanks in advance.
Kindest and Best Regards,
Tel : 0121 359 2703
e-mail : firstname.lastname@example.org
Re: The recognition of third party objects and the use of enablers in coded programs.
User DAVID MARSH (DAVID.MARSH@thales-cs.com.nospam) posted:
This depends on your level of paranoia (mine is variable, but not too high
The enabler reveals information to Robot, same principle in other test
tools. I have read that this is done by inserting a dummy object and
navigating from there, but I really don't know.
If you were to ship with the enabler enabled then if I hooked up Robot I
could find out quite a lot about your application that you might not want me
to know. When I have looked at some objects with robot I have seen that
only a subset of the data that Robot sees is actually displayed. The
additional data might give insights, any thing from missing columns on grids
to what appeared to be a simple control having a lot of data attached to it.
Then there is the identification of objects, good programming practices say
you should have meaningful names to help maintenance. So just looking at
the names of all your objects would give some clues on your application.
Finally with the enabler in you might find reverse engineering easier by
running some data driven tests to find boundary points.
I've never heard of it being done, but if you start talking to project
managers about this some of them get rather nervous.