Xlib

=============================

Improved support for newer versions of Mac OS X, a couple of new
extensions, and several bugfixes.

Composite extension
-------------------

Support for the composite extension, used to implement a composition
manager (added for plcm work in plwm).

By itself this extension is not very useful, it is intended to be used
together with the DAMAGE and XFIXES extensions. Typically you would also
need RENDER or glX or some similar method of creating fancy graphics.

XF86 special function keysyms
-----------------------------

Keysym definitions for special function keys found on modern keyboards,
e.g. raise and lower volume, start specific applications, etc. Have a
look in [Xlib/keysymdef/xf86.py](Xlib/keysymdef/xf86.py) to see what
there are and experiment with xev to see what your keyboard generates.
These definitions aren't brought in by default, so you must do this
after importing `Xlib.XK`:

python
Xlib.XK.load_keysym_group('xf86')


RANDR extension
---------------

The RANDR extension complements XINERAMA as a way of getting data about
the physical screens making up a virtual screen in X. An example of
usage can be found in [examples/xrandr.py](examples/xrandr.py).

---

=============================================================

A couple of new extensions, a Python 2.5 fix and a couple of aliases
(`Display.get_atom()` now uses the internal cache and added
`Window.raise_window()`). Tabs converted to spaces (SF id: 1559082).

RECORD extension (SF id: 1538663)
---------------------------------

Alex Badea contributed a RECORD extension module, allowing Python Xlib
programs to capture mouse and keyboard events (or all other core or
extension events) easily. A demo is in the examples directory. See
<http://refspecs.freestandards.org/X11/recordlib.pdf> for more
information.

XINERAMA extension
------------------

Mike Meyer contributed a Xinerama extension module, allowing Python Xlib
programs to interrogate the X server about positions and sizes of
multiple screens. Specifications are a bit tricky to find -
<http://sourceforge.net/projects/xinerama/> has some older specs and the
source code of the xorg project (libs & server code) has "definitive"
information.

Python 2.5 fix (SF id: 1623900)
-------------------------------

Bugfix to correct handling of XAuthority file parsing under Python 2.5
causing failed authentication.

---

==============================================================

A small release to incorporate a number of minor corrections and bug
fixes, including small changes to keysym handling, `.Xauthority`
parsing, several fixes to sending/receiving/flushing data, addition of
`WithdrawnState` to `WMHints`. petli completed documentation for
`Display` objects.

---

==========================

SHAPE extension
---------------

Jeffrey Boser contributed a SHAPE extension module, allowing Python Xlib
programs to use shaped windows. Take a look at examples/shapewin.py for
ideas on how to use it. For more information on shaped windows, see
<http://ftp.x.org/pub/R6.6/xc/doc/hardcopy/Xext/shape.PS.gz>

Python 2.2 fix
--------------

In Python 2.2 `FCNTL.FD_CLOEXEC` has disappeared and `FCNTL` on the whole
is deprecated, so that had to be dealt with to make the Xlib work with
that version.

---

==========================

Regression tests for the protocol definition
--------------------------------------------

Regressions tests have been created for all requests, replies and
events. The tests use PyUnit, and the old resource database test has
been updated to use it too.

A lot of protocol bugfixes
--------------------------

The bugs discovered by the regression tests have been fixed.
Additionally, a subtle bug in the core engine which could cause a
"can't happen"-error has also been found and fixed.

---

==========================

Event bugfix
------------

The xlib failed to parse the type code of events sent from other clients
using `SendEvent`. This has been fixed, adding the field `send_event' to
all event objects.

Event documentation
-------------------

The section "Event Types" in the manual has been written, detailing
all event types in the core protocol. The manual is now ten pages
thicker.

Basic support for GetImage/PutImage
-----------------------------------

The Drawable methods `put_image()` and `get_image()` have been
implemented, but handling image data is still up to the user. There is
however, thanks to Ilpo Nyyssönen, a trivial method `put_pil_image()`
that will work on some combinations of image and drawable depth. It's
not perfect, but it's a start.

---