980480 - hit testing and boundaries fail on retina displays

Status: RESOLVED FIXED

(Core :: Disability Access APIs, defect)

Comment 1

[alexander :surkov (:asurkov)]

*2 helps for hit testing, /2 doesn't help for position.

Robert, do you know what's special in Retina displays?

Comment 2

[James Teh [:Jamie]]

Retina displays are high DPI. Is this bug referring to something internal to Firefox or did you test it using an AT?

In Windows, Firefox is DPI aware and exposes physical coordinates, so as long as the AT is DPI aware, it should just work. Note that NVDA 2013.3 and earlier aren't DPI aware, so this caused nasty problems until high DPI awareness was implemented.

Comment 3

[alexander :surkov (:asurkov)]

It's rather something Firefox internal because Chrome works with VoiceOver just fine

Comment 4

[alexander :surkov (:asurkov)]

Attachment: patch

Comment 5

[Trevor Saunders (:tbsaunde)]

Comment on attachment 8387669 [details] [diff] [review]
patch

>   nsCOMPtr<nsIAccessible> deepestFoundChild;
>-  mGeckoAccessible->GetDeepestChildAtPoint((int32_t)geckoPoint.x,
>-                                           (int32_t)geckoPoint.y,
>+  mGeckoAccessible->GetDeepestChildAtPoint(geckoPoint.x,
>+                                           geckoPoint.y,
>                                            getter_AddRefs(deepestFoundChild));

while your here you might as well change that to ChildAtPoint(x, y, eDeepestChild)

Comment 6

[Markus Stange [:mstange]]

Comment on attachment 8387669 [details] [diff] [review]
patch

>diff --git a/accessible/src/mac/mozAccessible.mm b/accessible/src/mac/mozAccessible.mm

>+  NSView* mainView = (NSView*)[[NSScreen screens] objectAtIndex:0];

These all need to be NSScreen instead of NSView.

Comment 7

[alexander :surkov (:asurkov)]

landed with comments addressed http://hg.mozilla.org/integration/mozilla-inbound/rev/9cf34c82f975

Comment 8

[Phil Ringnalda (:philor)]

https://hg.mozilla.org/mozilla-central/rev/9cf34c82f975