[mythtv] Re: [mythtv-commits] mythtv commit: r7155 - in trunk/mythtv/libs by danielk

[Marcelo Toledo]

Daniel Kristjansson <danielk at cuymedia.net> writes:

> It's used by GetLanguageList(). If no languages have been specified
> for the EPG/subtitles it grabs the result of GetLanguage() which is
> our UI language. But GetLanguage returns a two character code and we
> need three character codes in all the contexts where we use a 
> language list so I used iso639.h to translate these.

I imagine that these functions are related to the language that we'll
use to show translated text (I might be wrong though). Even if it's not
it, but are related to identifying a language, this will not work for
everyone.

I think we shouldn't use iso639 because of this:

"But the two-character codes of ISO 639 are neither sufficiently
mnemonic nor complete for the world's languages: whereas ISO 639
supplies codes for only about 136 languages, the Ethnologue published by
the Summer Institute of Linguistics identifies over 6100 languages (see
Ethnologue: Languages of the World, ed. Barbara Grimes. 11th
edition. Dallas, TX: Summer Institute of Linguistics, 1988)."

And also because it will not work for every languages today and we have
a solution that already work and that's what every application uses. I
think the UNIX spec. of internationalization would fit in the system and
would work for everyone.

It uses the syntax: language[_territory][.codeset]

that's why GetLanguages is not returning only 2 charters as you said, it
already have two languages that returns in this form, they are
Traditional Chinese (ZH_TW) and Brazilian Portuguese (PT_BR), it happens
because portuguese (pt) is not the same as (brazilian portuguese), as
might also differ, I don't know from en_us, en_uk and so on.

-- 
Marcelo Toledo
marcelo at marcelotoledo.org
http://www.marcelotoledo.org
Mobile: 55 71 9957-6560