[g-a-devel]Re: Something about AT-poke

[mindy liu <Mindy Liu sun com>]

Hi Michael,

I'm so glad to tell you that the "AT-poke" is more stable after following
your instruction.Thank you very much. :-) 

And during my using it, I find that there is no entry to set selection, there
are several methods in AT-SPI AccessibleText Interface,
I wonder how to test them?

In addition,  there are 4 buttons on AT-poke which labeled "copy""paste""cut""delete",
I tried them. I don't know whether my steps were right. First, I selected
a text string in the TextView window, and then click these buttons, the result
was correct. But I think AT-poke should have some input area to input the
selection start offset and end offset instead of just operating on text directly.

Thanks again.
 
glad to keep in touch 

-- Mindy

Michael Meeks wrote:
> Hi Mindy,
>
> On Tue, 2002-06-11 at 09:13, mindy liu wrote:
> > I'm a Software QA Engineer of Sun Beijing, and I'm working for netscape 
> > accessibility with Bolian, Silvia etc..
> >     
>
> 	Great - that sounds good. Firstly, you should really CC the Gnome a11y
> mailing list when you mail me, gnome-accessibility-devel gnome org, and
> you should be subscribed there - everything must be done in public and
> archived :-)
>
> > "At-poke" is good, so our team finally choose it as our test tool, but 
> > when I use it,  I find a problem, my operation steps list here:
> >       
>
> 	I'm glad you like at-poke.
>
> > 1. Start "AT-poke", at the same time "GTK+ Code Demos" is started
> >         
>
> 	the gtk-demo process is forked - I'm glad you have it installed, you're
> competant :-)
>
> > 2. In "At-poke" main window, double click the application "gtk-demo" ,
> > another window named "a tool for poking things" appears and shows the 
> > tree model
> >           
>
> 	Yes; good.
>
> > 3. In the second window, expand the tree
> >             
>
> 	Right so far.
>
> > 4. Click on items whose role are table-cell, the windows will diappear 
> > except the main window, and no application in that.
> >               
>
> 	Ok - what happens is this: the remote application seg faults and dies -
> this is nothing to do with at-poke, but most likely a buggy gail /
> at-spi implementation problem. Is it possible you have an old gail /
> at-spi lying around ? you need to use HEAD always.
>
> 	To fix the bug; you need to first do: export
> GTK_MODULES="gail:atk-bridge", then run 'gdb gtk-demo', then repeat your
> at-poking - and you should get a stack trace which the list would be
> most interested in.
>
> > Sorry to bother you,  looking forward to your reply, thanks a lot.
> > If you want more envionment infor, pls feel free to ask me.
> >                 
>
> 	No problem bother me as much as you like within reason. Another good
> thing you could do to make this much clearer for other people in future,
> is to get at-poke to pop up a dialog saying 'the XYZ application exited
> abnormally' or somesuch - although distinguishing this from a clean quit
> might be quite hard - presumably there is some de-registration event we
> can listen for (?).
>
> 	HTH, good to meet you,
>
> 	Regards,
>
> 		Michael.
>
>