Count how often we change programs during a frame.

Opportunistically use the coloring program for
drawing underlines instead of the color program.
This avoids program changes in the middle of
text.

For the Emoji text scrolling benchmark, this reduces
the program changes per frame from > 1000 to around 100.

This will be used for coloring from the texture.

All the rest of debug spew goes to stderr here.

Add underlines into the mix with the Emoji heavy
text, to see if we can optimize lines.

ngl: Use fp16 for colors

See merge request GNOME/gtk!3405

wayland: Improve font setting fallback more

Closes #3838

See merge request GNOME/gtk!3412

alias attributes don't work on Darwin, so
do without.

Use an IFUNC resolver to determine whether we can use
intrinsics for FP16 conversion. This requires the functions
to be no longer inline.

Sadly, it turns out that __builtin_cpu_supports ("f16c")
doesn't compile on the systems where we want it to prevent
us from getting a SIGILL at runtime.

And add a compile time check for the presence of this
CPU feature.