Re: extended zvt widget

[Bill Haneman <bill haneman sun com>]

On Tue, 2002-04-30 at 14:41, Cristiano De Michele wrote:
> On Tue, 2002-04-30 at 16:29, Bill Haneman wrote:
> > On Mon, 2002-04-29 at 23:27, Cristiano De Michele wrote:
...>

> > However Cristiano, did you notice the accessibility
> > code in libzvt when you were working on it, and did you check to see if
> > it still worked after your changes?  That's something that would need to
> > be confirmed before applying any changes to libzvt.
> > 
> > best regards,
> > 
> > -Bill
> Hi Bill,
> I used profterm with my patched applied without any problem and 
> I also run vttest inside profterm having no problem, 
> should I do anything specific to test accessibility with my patches?

Hi Cristiano:

Yes, those two tests will tell you nothing at all about the
accessibility support.  If you test with ferret (from cvs module gail)
and simple-at (from at-spi) that will give you some information.  Ferret
is probably the easier of the two tests to set up and run, though its
coverage is rather limited.
in csh:

setenv GTK_MODULES gail:ferret

run the terminal window with the "Text" tab frontmost in the ferret
window, and check for inconsistencies.

Actually testing the libzvt accessibility coverage thoroughly is not a
trivial task, but ferret will give you some feedback and of course you
could patch ferret to give better coverage of the AtkText interfaces as
needed for libzvt.

You might glance at the code for accessibility support in libzvt (the
code that does things like implement atk_text_get_text_at_offset, etc.)
and also make sure you haven't changed the behavior of internal zvt
signal which the accessibility support relies on.  I think it's likely
that any substantial changes in libzvt (which it sounds as if you have
done) would break at least some of the accessibility support.  At this
late date no one should be changing code that accessibility uses without
trying to test it.  

I recognize that the available test tools for accessibility may make it
hard to do these regression tests at the moment, please feel free to
help in that regard :-)

best regards,

Bill

> ...I could give you some more infos on my patches of course
> but split them in separate bugs it's not feasible and even
> meaningless IMHO what's your opinio about? what should I do?
> 
> 						regards Cristiano
> 
> 
> 
> -- 
>   Cristiano De Michele,
>   Department of Physics,
>   University "Federico II" of Naples
>