[PATCH v4 0/5] Convert input co-ordinates to fixed-point

[Daniel Stone]

Hi,
Hopefully this patchset solves all the concerns that came up
previously.  wl_signed_24_8_t has been renamed to wl_fixed_t, and is
still an int32_t under the hood, but all users have been audited for
implicit conversions.  I also changed the macros to be static inline
functions, but since all types that could kinda conceivably be an
integer if you looked them at sideways will get promoted anyway, it
sadly doesn't offer much more type-safety.

Weston also now uses GLfloat rather than double internally, to match the
output and transformation maths.  However, in fairness, there are quite
a few internal users of double, so this might be something we want to
revisit later (maybe even a config option? who knows).

Hopefully this should be good to apply now.

Cheers,
Daniel