Comment # 36 on bug 942896 from Stanislav Brabec

Mindaugas Baranauskas:

The parameter for Unicode map is a bit more complicated.

Very old fonts and fonts in ancient formats don't contain any Unicode map. They
contain just up to 256 characters, and it is expected, that the locale has the
same character map. If it is not true, Unicode map has to be passed.

Newer fonts still support only up to 256 characters, but they embed a map
between positions in font and Unicode characters. If this map is present, the
font can be loaded without specifying Unicode map. The map inside the font is
used.

New fonts support more thab 256 characters, and they embed a map between
positions in font and Unicode characters. If the font is loaded without
specifying Unicode map, the map inside the font is used.


Regarding your test: Are you aware of a race condition that could affect your
tests? See bug 904214.

As far as I understand, fonts are correct even in Canadian and Slovak, but the
keymap is bad. It could be the mentioned bug.


Regarding the patch:
https://github.com/yast/yast-country/pull/61/files

You included Serbian locale (sr_RS.UTF-8), which uses Cyrillic. Serbian should
definitely use font supporting Cyrillic. (In difference to sr@Latin locale,
which uses Latin alphabet.)

Regarding Japanese: It seems to be OK. It used lat9w before (=ISO-8859-15).