Mailing list - Entries of 2005

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

Re: [qftestJUI] Problem finding widget indexes

  • Subject: Re: [qftestJUI] Problem finding widget indexes
  • From: Gregor Schmid <Gregor.Schmid@?.de>
  • Date: 06 Dec 2005 16:52:05 +0100

Hi Paul,

have you set hierarchical resolution for both recording and replay?

If qftestJUI records just one component for left and right side,
either hierarchical mode is not set for recording or the parent
components don't have distinguishing names.

If still in doubt, please temporarily set name override mode to "Plain
attribute" for recording, create a new test-suite, select
Record->Receive components. In the SUT, press F11, right click
anywhere and select "Whole window". That should record all components
in the window, no matter what their name is in the new suite. Please
save the suite and send it to me.

As for why qftest selects the same component during replay even though
you have two different representations of these: I cannot tell without
a run-log. Please change the option "Run-log->Content->Log level for
the SUT" to plain message, replay a sequence with the two clicks and
send me the resulting log.

Best regards,

"Paul Bindels" <paul.bindels@?.com> writes:

> Thanks for the feedback.  I am however still having problems here.
> I was already running in "hierarchical name resolution" mode.   I see in the
> "Structure" of the component that there is an index count of 2.  This makes
> sense because there is 2 check boxes associated with it.  One check box is
> on the left and one check box is on the right of the parent class component.
> When I record the checking of a checkbox on the left, it puts the index as
> zero (as expected).  When I record the checking of a checkbox on the right,
> it puts the index as one (again as expected).  However when I do a replay,
> the test doesn't check the correct check box.  When I manipulate the index
> of the component manually, it doesn't make any difference during replay.  It
> still checks the same box as it did originally.  Am I missing something on
> how the Structure index is used?  Why doesn't qftest pick the correct index
> during the replay?  Why can't I change the index value of the component
> Another similar problem, is that when I record the checking of each box,
> qftest only creates one component because the component has the same name.
> This makes it impossible to check both the box on the left and the right
> (even if I could get the indexing part to work).
> I have attached a small qft file to show you what we are trying to do and
> picture of the checkboxes.
> I will try the name resolver next but that looks pretty complicated so I
> would like to try the other options first.
> -----Original Message-----
> From: Gregor Schmid [mailto:Gregor.Schmid@?.de]
> Sent: Friday, November 18, 2005 8:37 AM
> To: qftestJUI-list@?.de
> Cc: Paul Bindels
> Subject: Re: [qftestJUI] Problem finding widget indexes
> Hi Paul,
> please see the documentation for the option
> "Record->Components->Name override mode"
> manual section 14.2.3 (and the similar option under replay).
> If you setting is name overrides everything, qftest will not be able (or
> rather not try) to tell components with identical names apart (though you
> should get a warning about non-unique names).
> Your options are to switch to "hierarchical name resolution" mode or even
> "plain attribute", but if possible you should try to avoid that.
> It's better to try to use per-window unique names and only if that can't be
> achieved, use hierarchical resoution or maybe write a name-resolver to make
> a few remaining name collisions unique (which would be a feasible solution
> for this special case).
> Best regards,
>     Greg
> "Paul Bindels" <paul.bindels@?.com> writes:
> > Hi,
> > We are having a problem when playing back a sequence that uses widgets
> > of the same name.
> > We have a widget which contains two Jtextfields (one editable, one non
> > editable) which both have the same name. We cannot consistently select
> > one or the other to edit the text or to do verification.  When we play
> > back sometimes the non-editable text window gets selected. Our
> > developers say the parent (the widget) has an index for each of the
> > text boxes so QfTest should be able to uniquely identify the one we
> > want if we have access to it. We don't see this capability in QfTest.
> > One solution is to force a unique name for each internal component but
> > this requires redesign. Is there a work around? Is there a planned
> enhancement to provide this capability?
> >
> > Design
> > Widget
> >   - Jtextfield AttrAddress (non-eidtable)
> >   - Jtextfield AttrAddress (editable)
> >
> > QfTest only shows one
> > Widget
> >   - Jtextfield AttrAddress## (no way to know which one)
> >
> > Thank you,
> > Paul

Gregor Schmid                                Gregor.Schmid@?.de
Quality First Software GmbH           
Tulpenstr. 41                                Tel: +49 8171 919870
DE-82538 Geretsried                          Fax: +49 8171 919876