Less fancy token pasting, replacing them with inline functions where
possible, entirely removing unnecessary ones, and fixing smaller
issues here and there.
Part of P0186, funded by [Anonymous] and Blue Bolt.
The bullet update function really needs this one declared in C land to
link without a fixup overflow? Guess we'll have another technical debt
push before then…
Part of P0148, funded by [Anonymous].
The main point of the previous strings/ subdirectory was to bundle all
hardcoded strings for translators. And sure, *technically*, gaiji
strings are *both* strings *and* something you might want to translate.
But mainly, they're sprites with an attached enum, and their own
directory. Changes to the enum quickly tend to involve changes to the
strings that use these values, so it makes sense to keep both in the
same directory.
Especially since 82% of the previous strings/ directories consisted of
such gaiji strings.
That leaves the strings/ directory rather empty and nondescript though.
Recently though, I've been wanting to generally move all Shift-JIS text
to this directory. While that wouldn't *solve* the typical "text editor
accidentally a file upon save, due to wrongly detected encoding" issue,
it's at least a mitigation: If all Shift-JIS strings are in files that
contain nothing *but* Shift-JIS strings, a wrongly detected encoding
becomes immediately noticeable.
For that job, strings/ can have a more descriptive name though. Hence,
shiftjis/.
Part of P0141, funded by [Anonymous] and rosenrose.
Funny how the actual scores are stored as little-endian gaiji strings
in the bold font, yet never actually used as such.
Part of P0063, funded by -Tom-.
nmlgc