I see that almost no other language has translated those (so Walloon will
be the first :) ).
If I understand correctly, qx/QX had to be translated to put some
letters that are necessary for that language, so I put
(there is also éèôû that are required for standard walloon orthography,
but a font having åçî will likely have the others too)
So, the Greek translation should put two greek letters there, the arabic
translation two arabic letters, etc.
The Ukrainian translation should put the "turned ghe" (which is not
always present on cyrillic fonts with only russian alphabet), etc.
And if input of latin letters is not possible in a given locale,
the qy/QY should be translated too (that would then put latin only
fonts out of the list of selectable fonts)
For a latin based language, should the qy/QY be translated also?
Doesn't keeping it untranslated would select ascii-only fonts,
that is, allowing the selection of a font which doesn't has the
needed glyphs for the language?
I would think that the comment in the po file is incorrect,
and that both qx/QX and qy/QY have to be "translated" for latin-based
languages, but only qx/QX for non latin languages that also allow ASCII
input (as ascii-only is good for typing in latin then, no need of
accentuated latin letters); and of course, both qx/QX and qy/QY translated
to non-latin scripts if tying ASCII has to be completly disabled.
the "oO" could be translated with non-latin letters (in addition to "oO")
for the few non-latin scripts that have case variations, like greek
and cyrillic, eg in Russian something like:
For common punctuation, Japanese could have:
(added the ideographic full stop and comma)
For digits, languages that use localized digits (like Farsi) should
put translate them, eg:
0175 in perso-arabic digits.
Is my interpretation of those special strings correct?
Ki ça vos våye bén,
http://chanae.walon.org/pablo/ PGP Key available, key ID: 0xD9B85466