When rendering text, if none of the fonts in the cascade list contain a
glyph for a given code point, we now query Skia's font manager to find
a system font that can render it.
This fixes an issue where having an emoji font first in the font-family
cascade list would cause us to get the space width from the emoji font
as well.
We now look at adjacent text chunks when deciding which font to use for
space widths, instead of just blindly obeying the font-family value.
By keeping track of the enclosing range around all Unicode ranges of a
FontCascadeList entry, we can quickly reject any code point that's
outside all ranges.
This knocks font_for_code_point() from 7% to 3% in the profile when
scrolling on https://screenshotone.com/
