Wrong menu item role for a11y

[Luca Ferretti]

Binary package hint: indicator-session

The role used by Shutdown, Hibernate, Logout and Lock Screen menu items (Sleep/Suspend not tested) is wrong. Using Orca, it says those menu entries are check items. They are, instead, simple menu items.

This is no fair for impaired people (we are lying them about actual usage of focused items) and is a really ugly too see this mis-usage of GTK+ widgets.

ProblemType: Bug
DistroRelease: Ubuntu 10.04
Package: indicator-session 0.2.7-0ubuntu3
ProcVersionSignature: Ubuntu 2.6.32-19.28-generic 2.6.32.10+drm33.1
Uname: Linux 2.6.32-19-generic i686
Architecture: i386
Date: Fri Apr 2 10:24:48 2010
InstallationMedia: Ubuntu 10.04 "Lucid Lynx" - Beta i386 (20100318)
ProcEnviron:
 LANG=en_US.utf8
 SHELL=/bin/bash
SourcePackage: indicator-session

[Ted Gould (ted)]

On Fri, 2010-04-02 at 08:32 +0000, Luca Ferretti wrote:
> The role used by Shutdown, Hibernate, Logout and Lock Screen menu items
> (Sleep/Suspend not tested) is wrong. Using Orca, it says those menu
> entries are check items. They are, instead, simple menu items.

We use the check item to provide more configurability at runtime instead
of reallocating and rebuilding them. I'd be happy to update how Orca
sees the items, but I'm unsure of how to do that. Do you know where
that information is gotten by Orca?

> This is no fair for impaired people (we are lying them about actual
> usage of focused items) and is a really ugly too see this mis-usage of
> GTK+ widgets.

I think that's in inappropriate characterization of how it works, and
insulting people's work isn't very nice.

[Luca Ferretti (elle.uca)]

Ted, in the past days I tried to ask about this issue on #a11y on GIMPnet. Unfortunately nobody had interest in this bug :(
I'll try contacting Willy Walker directly by email (he wasn't on IRC those days) but honestly I ignore his current status after Sun->Oracle.

Orca reads the info using the AT-SPI/ATK framework. Basically ATK defines some "interfaces" (example a button "implements" actions `press`, `release`, `clic` for AtkAction interface and statuses `invalid`, `active`, `pressed` for AtkState interface) while AT-SPI is a "bus" between client applications (Orca, GOK..) and provider applications/libraries (GTK+/GNOME, Firefox, OpenOffice.org).

So, I suspect you should re-define some Atk interface, but honestly I suppose is not trivial or "clean" to do :|

PS It wasn't in my intention insulting anyone, but simply report the current situation. Maybe rude, but IMHO true.