Fix query_yn for some Unicode languages

[VlasovVitaly]

Fix issue #11877
pot file updated too.

[BevapDin]

@VlasovVitaly Translating the four letters (y/n/Y/N) separately and combining the translations into the displayed string would theoretically allow any characters (single Unicode code point) to be used.

That means adding translations for "letter for yes (uppercase)", "letter for yes (lowercase)", "letter for no (uppercase)" and "letter for no (lowercase)" (using a translation context).

The current code can only uppercase ASCII characters, that would be avoid as well. Of course it requires more work from the translators, so would this be an acceptable solution?

[VlasovVitaly]

I thinking about this method and found couple of major problems:

1. As you noticed, working with Uppercase/Lowercase is confusing. This can be resolved by adding good comments for translators.
2. I can't image what happens when only one letter will be translated and other is not.
3. User need to change keyboard layout to press y/n directly in all non-latin languages. This can be possible resolved by adding keyboard mapping list. Russian example '' => 'д' => 'y', 'н' => 'n'. Most awesome thing is we can use this mapping in any input context. So anyone can use own language key bindings. But i guess this will be broken for oriental languages.