Log message:
Update to 1.7.3:

Two important patches fixing deadlocks when using XIPassiveGrab requests.
You're recommended to updated.

Note that we now require libX11 1.6, that's a new dependency over 1.7.2 but
it's over a year old by now anyway.

Jasper St. Pierre (2):
      XIPassiveGrab: Fix display locking inside _XIPassiveGrabDevice for error paths
      XIPassiveGrab: Fix completely broken locking in XIGrabTouchBegin

Keith Packard (1):
      man: Update XIQueryVersion docs to match new version compatibility semantics

Michael Joost (1):
      Remove fallback for _XEatDataWords, require libX11 1.6 for it

Peter Hutterer (1):
      libXi 1.7.3

Log message:
Add builtin detection for libXi (mainly for the case X11_TYPE=native but builtin
version is older than required version).
Set API_DEPENDS for required package for the case using libXi from pkgsrc,
same as specified to build this package.

Log message:
Update to 1.7.2.

Changes in 1.7.2:
Only one minor change since the RC. Again, this release contains the fixes
for CVE-2013-1998, CVE-2013-1984 and CVE-2013-1995 so you're encouraged to
update.

Peter Hutterer (1):
      libXi 1.7.2

Thomas Klausner (1):
      Remove check that can never be true.

Changses in 1.7.1.901:
First and likely only RC for libXi 1.7.2. This one has a bunch of changes
for CVE-2013-1998, CVE-2013-1984 and CVE-2013-1995. These relate to various
integer overflows and other corruption that happens if we trust the server
a bit too much on the data we're being sent.

On top of those fixes, the sequence number in XI2 events is now set
propertly too (#64687).

Please test, if you find any issues let me know.

Alan Coopersmith (14):
      Expand comment on the memory vs. reply ordering in XIGetSelectedEvents()
      Use _XEatDataWords to avoid overflow of rep.length bit shifting
      Stack buffer overflow in XGetDeviceButtonMapping() [CVE-2013-1998 1/3]
      memory corruption in _XIPassiveGrabDevice() [CVE-2013-1998 2/3]
      unvalidated lengths in XQueryDeviceState() [CVE-2013-1998 3/3]
      integer overflow in XGetDeviceControl() [CVE-2013-1984 1/8]
      integer overflow in XGetFeedbackControl() [CVE-2013-1984 2/8]
      integer overflow in XGetDeviceDontPropagateList() [CVE-2013-1984 3/8]
      integer overflow in XGetDeviceMotionEvents() [CVE-2013-1984 4/8]
      integer overflow in XIGetProperty() [CVE-2013-1984 5/8]
      integer overflow in XIGetSelectedEvents() [CVE-2013-1984 6/8]
      Avoid integer overflow in XGetDeviceProperties() [CVE-2013-1984 7/8]
      Avoid integer overflow in XListInputDevices() [CVE-2013-1984 8/8]
      sign extension issue in XListInputDevices() [CVE-2013-1995]

Peter Hutterer (7):
      Copy the sequence number into the target event too (#64687)
      Don't overwrite the cookies serial number
      Fix potential corruption in mask_len handling
      Change size += to size = in XGetDeviceControl
      If the XGetDeviceDontPropagateList reply has an invalid length, return 0
      Include limits.h to prevent build error: missing INT_MAX
      libXi 1.7.1.901

Log message:
Needs libXfixes now (PR 47721)

Log message:
Update to 1.7.1:
Only a single fix: including the XFixes header to get the typedef for
PointerBarrier. Naturally, this adds a pkgconfig build-time dependency on
XFixes.

Why? The header shipped with 1.7 typedef'd PointerBarrier. If you #include
both XI and Xfixes headers, you will end up with a duplicate typedef.

This is not an issue on gcc >= 4.6 since the two typedef's are the same.
On earlier versions this will trigger an error. gcc 4.6 -pedantic-errors
will trigger the same error.

Peter Hutterer (2):
      Require XFixes for PointerBarrier, remove duplicate typedef                \ 
                                                                                 \ 
                   libXi 1.7.1

Log message:
adjust inputproto dependency

Log message:
redefinition of PointerBarrier.

gcc 4.6 won't complain about that, but earlier versions do:
http://gcc.gnu.org/git/?p=gcc.git;a=commitdiff;h=ce3765bf44e49ef0568a1ad4a0b7f807591d6412

gcc 4.6 with -pedantic-errors shows:
/opt/xorg/include/X11/extensions/XInput2.h:172:13: error: redefinition of
typedef ‘PointerBarrier’ [-pedantic]
In file included from test.c:1:0:
/opt/xorg/include/X11/extensions/Xfixes.h:255:13: note: previous declaration
of ‘PointerBarrier’ was here

PointerBarriers is defined in XFixes.h and here. So hook onto the only thing
we can in Xfixes.h and use that to figure out if we need to typedef
ourselves. XFIXES_MAJOR is defined in xfixeswire.h, so we can't hook onto it
directly.

Adding this ifdef here means we have include order dependency of XFixes.h
before XInput2.h unless we add a similar ifdef to the fixes headers.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
---
Nasty, but can't think of a better way atm.

Log message:
Update to 1.7:
The feature added in this revision is pointer barrier events and releases.
Pointer barriers themselves are an XFixes 5 feature, XI 2.3 adds two
disctinct features to barriers:
* the ability for clients to get notified if pointer movement is restricted
  by a barrier
* the ability for clients to allow a pointer to move through a barrier after
  it was constrained by the barrier

For a more verbose explanation of these features please refer to:
http://who-t.blogspot.com.au/2012/12/whats-new-in-xi-23-pointer-barrier.html

Log message:
Update to 1.6.2:
Two man page fixes and a fix to enable generic event copying for raw touch
events (using in e.g. XPeekEvent())

Benjamin Tissoires (1):
      Add missing XI_RawTouch* in XInputCopyCookie

Peter Hutterer (3):
      man: fix formatting issues in XGetDeviceControl(3)
      man: add generation of missing man pages for XIGrabTouchBegin
      libXi 1.6.2

Log message:
Drop superfluous PKG_DESTDIR_SUPPORT, "user-destdir" is default these days.