> Well, you'll probably end up with one weird-looking pattern "8́"
> (looking like "eight with acute" and in fact saying "do not hyphenate
> before the combining acute accent"), but such is life ...
Exactly. For Latin-script languages there shouldn’t be too many of
those anyway. In fact, I fail to see why it should be necessary at all
to have a digit in the middle of a combining sequence.
> We should probably nevertheless do some double-checking to make sure
> that XeTeX doesn't do any strange normalization at this point.
Maybe. It’s hard to say without a concrete use case. But the
sequence <00E6 LATIN SMALL LETTER AE, 0301 COMBINING ACUTE ACCENT> (ǽ)
we’ve been discussing recently won’t cause any problem.
> The codes are determined by the Unicode standard and you can find the
> definitions in unicode-letters.tex (luatex-unicode-letters.tex). We
> would put those definitions to loadhyph-<languagename>.tex. I’m not
> yet sure how to handle this properly for LuaTeX, but I believe that
> *we* (me and Arthur) have some homework to do.
This should be taken care of by the files produced by the LaTeX team,
that are also used by plain Tex.
Best,
Arthur
