The previous implementation generated accents in NFKD -- e.g., i
followed by fn+e would generate í, which is actually an ordinary i
followed by U+0301 COMBINING ACUTE ACCENT. Unfortunately, it turns
out that a bunch of websites and apps (especially European ones
written in languages that use these a lot) were very poorly written,
and will misparse and/or crash if presented with Unicode NFKD. They
require and expect NFKC, with characters like í (U+00ED LATIN SMALL
I WITH ACUTE) that look visually identical -- and are in fact
normalization-equivalent -- but have to be encoded differently.
The new accent implementation handles this in a very flexible way.
Many new comments added as well, as it's also clear that this is going
to need a bit more expansion before it becomes a true polyglot keymap.
Co-authored-by: Yonatan Zunger <zunger@desiderata.lan>
554b937d21