Missing accents get generated for some TTF and OTF fonts.

[Ken Moffat]

I am attempting to revise my coverage of OTF and TTF fonts, using
XeLaTeX and the following packages: fontspec, fancyhdr, multicol,
polyglossia, bidi.

I want to show whatever output a font provides for various unicode
codepoints (some less-common letters or symbols may be mismapped)
and usually a missing codepoint provides some sort of indication
(inverse ?, or X in box, or empty box, or else a space) so that at
my first stage I can decide which alphabets are NOT supported by the
font.

I have this week discovered that in some fonts (so far, not very
many) the missing codepoint is reported in the output but the PDF
nevertheless contains it. So far I have seen this with grave accents
on Latin and Cyrillic vowels, and on (very ugly) carons and acute
accents added to Noto CJK fonts.

This is just a heads up for anyone doing similar things - with  my
XeLaTeX stack the results might appear usable, but for my main
purpose (letting people choose libre screen fonts which might be
useful to them) this obviously is double plus unuseful.

I have not found any explanation for this online, maybe I use the
wrong search terms, or maybe it is such an uncommon issue that
nobody has documented it.

ĸen
-- 
Voluntary positions are like garden bindweed: untended, they will
fill any void you might have.           -- James Max