[Webkit-unassigned] [Bug 99578] [EFL] Share WebKit-Gtk's Accessibility implementation with others WebKit ports.

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=99578





--- Comment #5 from Mario Sanchez Prada <mario at webkit.org>  2012-10-23 03:59:52 PST ---
(In reply to comment #4)
> [...]
> Hmm, good point. The problematic point would be the usage of gail_text_util 
> methods on the AtkText implementation. FWIW, GTK and clutter are not using 
> them anymore, but using a private fully pango-based library.

Yes, which reminds me of another issue with the a11y code here: pango is still being used in some places there and, as far as I know, its usage inside WebKit is discouraged and should be avoided.

I think the right replacement is TextBreakIterator, not 100% sure, though.

> So another task would be check what EFL is using for the text rendering.

Text rendering code is not an issue in WebCore/accessibility/gtk. Pango is being used only to get a PangoLayout to pass to gail_text_util_get_text() so I think getting rid of gail would be enough (as easy as it sounds, as hard as it might be :-))

And AFAIK, the EFL port is already avoiding pango and relying in TextBreakIterator instead, but Krysztof will know better, I suppose.

> Probably this bug will became a metabug of all the tasks needed to make the 
> move. Your paragraph detects two.

Yes, could be.

> > > (as I'm thinking on gtk-widget initialization related code)
> > 
> > Not sure what you mean. If you're talking about the initialization of the 
> > WebView (which is a GtkWidget) and its implementation of 
> > gtk_widget_get_accessible(), such code is not here in 
> > WebCore/accessibility/gtk, but in WebKit/gtk/webkit (for WK1) and 
> > WebKit2/UIProcess/API/gtk (for WK2).
> 
> Yes I was talking about that. So good to know that we can discard my concern
> as a new task.

Ok, great.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.