[PULL] 1.8 fixes, unreviewed

[Keith Packard]

On Tue, 2 Feb 2010 17:29:04 +1000, Peter Hutterer <peter.hutterer at who-t.net> wrote:

> Peter Hutterer (5):
>       xkb: convert XkbUseCorePtr and XkbUseCoreKbd into PickPointer/Keyboard.

This looks like it duplicates functionality in _XkbLookupAnyDevice which
has precisely these same checks.

>       Xi: reset the sli pointers after copying device
>       classes. (#25640)

As far as I can tell (reading XkbAllocSrvLedInfo) this looks
correct. I don't understand xkb any better though.

Reviewed-by: Keith Packard <keithp at keithp.com>

>       dix: When attaching a device, set CoreEvents to TRUE.

This patch does two things -- makes coreEvents track dev->u.master !=
NULL and also has miPointerSetPosition explicitly have the pointer track
slave device positions only when they have no master device. That second
change seems redundant to me.

>       dix: if owner-events is true for passive grabs, add the window mask (#25400)

Wow. This patch is not correct, but it does identify a bug in the code.
As the grab is activated before the event is delivered, I'd say that the
right fix would be to use DeliverGrabbedEvent instead of open-coding
similar stuff in CheckPassiveGrabsOnWindow.

>       Xi: assume BadMode for non-BadMatch errors returned from
>       SetDeviceMode.

Yeah, this seems like a fine hack except that all protocol requests are
allowed to return BadAlloc, so you should probably check for this one as
well.

-- 
keith.packard at intel.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://lists.x.org/archives/xorg-devel/attachments/20100212/dc0ec628/attachment.pgp