[Webkit-unassigned] [Bug 43922] Support for multiple speech enabled elements in same page.

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=43922





--- Comment #2 from Jeremy Orlow <jorlow at chromium.org>  2010-08-13 04:54:06 PST ---
(From update of attachment 64231)
WebCore/page/SpeechInput.h:74
 +      HashMap<int, SpeechInputListener*> m_listeners;
Why do you need ids?  Why not just keep track of listeners via their pointers?


Will you only be able to cancel recording from WebKit, and not cancel processing?  If so, I think this might work since only one input box should ever be recording at once.  In fact, if you get a message from one while another is recording, I think you can take that as a signal to end recording in the other.  And that means only having one listener pointer at a time works fine.  Why does everything ened IDs though?  The multiple speech input boxes can all share the same backend since only one item will ever be listening at once...right?

If you can cancel processing from the controls, I guess it's more complicated because you could have multiple doing processing while another is recording and so on.  In that case, I feel like you could have a pointer to the recording object and a set of processing objects.  In that case, using IDs probably makes sense.  But can't the IDs just be memory addresses?  It seems a bit simpler than trying to dole out unique IDs.

Btw, can you try making a diagram (or even hand-drawing and scanning it in) of all the classes and how they interact and attach it to the top level bug?

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