19 Apr
2011
19 Apr
'11
1:38 p.m.
Ryou Ezoe wrote:
I don't care library internally convert between UTFs. At least, conversion between UTFs are trivial and more importantly consistent so I don't care.
The user want to develop their software using their native language. Localization should be done by mapping their language to other languages. So the library should offer mapping to any unicode string to other unicode string.
But you have to propose something specific. How should the library be changed? I asked a direct question: if the library accepts translate( L"日本語" ) encodes the wide literal as UTF-8 and uses it to look up a msgid in the .po file, will this be enough for you? Or do you want something else, and if so, what?