Fixes#707
Our scoring algorithm previously did not take into account symbolic
traits, so when `bold = false and italic = false`, regular, bold, italic
would all be equally likely to appear first.
This modifies our scoring algorithm to prioritize matching symbolic
traits. Further, we have a special case for no symbolic traits to
prioritize "Regular" named styles. We can expand this to other styles
too but we do not do this here.
We also modified the algorithm to always prefer fonts with more glyphs
over fonts with less, hopeful that we can load fewer fonts for other
glyphs later.
Fixes#757
The x/y is unsigned (as it should be, since top-left is 0,0). In the
case of a very small grid size and certain thickness settings, its
possible to overflow. In this scenario, the only reasonable thing to do
is just clamp to 0 because some sprites are going to look weird with
small enough grids anyways.
Fixes#668
We were previously only checking the first font result in the search.
This also fixes our CoreText scoring algorithm to prioritize faces that
have the codepoint we're searching for.
This is a weird one. By using intCast on the `idx` I am periodically
getting a panic on index out of bounds where the index is larger than
FontIndex can possibly be. Very strange!
I tried to just remove intCasts and believe it or not that worked.
Previously, `cat /dev/urandom` would trigger the issue in seconds and
now I've had it running 20+ minutes without the issue.
The additional `if` check is just a safety mechanism
Fixes#143
There were multiple issues with the previous calculation. First, we
expected dash width COULD be negative and protected against it, but our
dash width calculation type was unsigned! With the proper font metrics,
this led to an underflow safey panic.
Second, as part of the dash width calculation, we are tryign to downstep
the size of the gaps so we can try to fit dashes within a cell. But we
were not using those adjusted gap sizes. With the proper font metrics,
this could lead to an assertion failure seen in #143.
This fixes the calculations. They produce the same numbers, but do so in
a more Zig-idiomatic way while fixing the above two issues.
When font shaping grapheme clusters, we erroneously used the font index
of a font that only matches the first codepoint in the cell. This led to the
combining characters being [usually] unknown and rendering as boxes.
For a grapheme, we must find a font face that has a glyph for _all codepoints_
in the grapheme.
This also fixes an issue where we now properly render the unicode replacement
character if we can't find a font satisfying a codepoint.
