Re: [g-a-devel]Image Interface: -1 image size

["Padraig O'Briain" <Padraig Obriain Sun COM>]

Mario,

Some buttons have images so we implement AtkImage interface on all buttons (or 
at least for accessible objects correcponding to a GtkButton). I agree that it 
is not very satisfactory .

The assitive technology needs to determine whether there is actually an image.
There is code in gnopernicus in get_specialization_from_acc_real in SRObject.c 
which does just that.

I cannot speak to your other question.

Padraig

> 
> Hello.
> 
> During experimenting with the AT-SPI I found something which
> is kind of odd to me.  Maybe someone on this list can enlighten me
> as to why this would happen, or if this is related to some yet
> undiscovered bug in AT-SPI/ATK:
> 
> I found out that many objects on the desktop which implement
> the image interface seem to return -1 -1 for their size,
> and 0 for their position, yet, their state is visible.
> 
> Here is some minimal code to demonstrate it:
> 
> gspi> (define (bogus-images)
>         (find (desktop 0)  ;; Recursively look for objects satisfying 
predicate
>               (lambda (obj)
>                 (and (image? obj)     ;; Object implements the Image Interface
>                      (visible? obj)   ;; Object has state VISIBLE
>                      (< (caddr (image-extents obj)) 0)))))
> gspi> (length (bogus-images))
> 9
> 
> ;; So we have 9 objects implementing the image interface, but
> ;; with < 0 size, they look like this:
> 
> gspi> (map image-extents (bogus-images))
> ((0 0 -1 -1) (0 0 -1 -1) (0 0 -1 -1)
>  (0 0 -1 -1) (0 0 -1 -1) (0 0 -1 -1)
>  (0 0 -1 -1) (0 0 -1 -1) (0 0 -1 -1))
> 
> ;; Yet, their component extents are valid:
> 
> gspi> (map get-extents (bogus-images))
> ((236 743 96 22) (141 743 95 22) (46 743 95 22)
>  (169 0 24 24) (193 0 24 24) (4 191 200 30)
>  (4 221 200 30) (4 281 200 30) (92 95 21 53))
> 
> ;; Those kind of objects seem to be common to many apps:
> 
> gspi> (map find-application-name (bogus-images))
> ("gnome-panel" "gnome-panel" "gnome-panel"
>  "gnome-panel" "gnome-panel" "gnopernicus"
>  "gnopernicus" "gnopernicus" "gedit")
> 
> 
> Does anyone know what is the reasoning behind those Image-Interface
> implementing objects?
> 
> P.S.: A somewhat related question I have is why the Image
> interface does not provide any functions for retrieving
> a pixel representation of the actual image.  Is it expected
> that image descriptions are always provided by the
> Application programmer, and never dynamically added by the
> assistive technologies?  At least under Windows, screen
> readers have a image labeling feature which allows for adding
> descriptions to certain icons on the fly.  I guess that
> such a functionality would require the assistive technology
> to actually retrieve the pixel content of an image.  Or is
> it expected to do this via direct X Windows System calls?
> 
> -- 
> CYa,
>   Mario | Debian Developer <URL:http://debian.org/>
>         | Get my public key via finger mlang db debian org
>         | 1024D/7FC1A0854909BCCDBE6C102DDFFC022A6B113E44
> _______________________________________________
> Gnome-accessibility-devel mailing list
> Gnome-accessibility-devel gnome org
> http://mail.gnome.org/mailman/listinfo/gnome-accessibility-devel