GNOME Bugzilla – Bug 455276
AccessibleName for HTML text labels is the raw HTML
Last modified: 2007-11-14 08:43:14 UTC
See the Orca Bug 455259. The AccessibleName is the raw HTML for Java label containing HTML. For example, the SwingSet2 TabbedPane demo has a page tab with the label "Bouncing Babies!" in blue. The AccessibleName is "<html><font color=blue><bold><center>Bouncing Babies!</center></bold></font></html>".
Hi Jeff: any ideas on this one?
I don't think this belongs to a bug of jab. Jab just forwards the message to AT and it hardly changes anything. So the message should be explained by AT or stripped the html marks by java applications.
(In reply to comment #2) > I don't think this belongs to a bug of jab. Jab just forwards the message to AT > and it hardly changes anything. So the message should be explained by AT or > stripped the html marks by java applications. I don't believe the accessible name of an object should contain HTML markup. Instead, I believe the intent of the name is that it is a human consumable string. As such, rather than requiring all assistive technologies to strip out the HTML markup from the name, it might be better for the JABG to do the stripping.
Because JABG doesn't know the syntax and semantic of the name returned from the accessible object, it of course has no way to explain it. Futhermore, if there's processing string in JABG, this will have a negative effect on preformance. I think the best way is to let java applications return the real name without html markups in their implementation of getAccessibleName for components.
(In reply to comment #4) > Because JABG doesn't know the syntax and semantic of the name returned from the > accessible object, it of course has no way to explain it. Futhermore, if > there's processing string in JABG, this will have a negative effect on > preformance. > > I think the best way is to let java applications return the real name without > html markups in their implementation of getAccessibleName for components. > To make sure I understand: your recommendation is that application developers that use HTML markup should explicitly set the accessible name? If so, this seems like something we may continually struggle with. I wonder if we might want to put an RFE directly into Java Swing to do some parsing of the string first and do the automatic setting of the name if it contains markup.
Yes. The application developers should call setAccessibleName to provide the name for AT. This is also the way that all GNOME C applications do. Hence, I'll close it as "not a defect".