of hardcoding the feature list in img2txt.

* Added a first, simple test program.

instance.

* Autodetect zzuf in the configure step. It might become handy.

* Updated NEWS and ChangeLog.

* Implemented CacaDisplay.Canvas which acts the same as caca_get_canvas().

Size and Rectangle versions.

library version and not rely on .so being present.

major/minor computation scheme.

* Use CuculFont.Render() in the C# test for a nice oldschool effect.

instead of (point, point). Yeah, I break the ABI again. But who uses these
    functions anyway?

allowing for much cleaner code (at the expense of an extra dependency).

* Slightly improved the C# test program.

fewer than Firefox, lol.

accordingly.
  * The C++ test example no longer needs "config.h".

we don’t end with those ugly identifiers in the documentation.

fixed the bindings so that they now work with the Microsoft CLR.

behaviour.

create a canvas when so.
  * Add caca_get_canvas() to retrieve the cucul canvas.

bindings.
  * Distribute the python/ subdirectory using automake.

cucul_attr_to_argb64() (previously _cucul_attr_to_rgb12bg,
    _cucul_attr_to_rgb12fg and _cucul_attr_to_argb4) in the official
    libcucul API.
  * Cleanup complete: libcaca no longer depends on "cucul_internals.h".

direct cv->chars and cv->attrs access by display drivers.

drivers no longer need to access the cv->refcount private member
    and __cucul_set_canvas_size() is no longer needed.