i18N demo: missing languages for many countries

[bill]

Barbados, for example, should have a child for "English", but doesn't.

It's because the CLDR main.zip file has no en_BB.xml file, so it didn't get registered.

Perhaps should be using another source like:

• ​http://www.alvestrand.no/pipermail/ietf-languages/2002-February/000081.html (except that only has languages that are used in multiple countries)
• ​http://en.wikipedia.org/wiki/List_of_countries_by_spoken_languages (may be hard to screen scrape and also seems to refer to languages/countries w/no ISO code)
• ​http://en.wikipedia.org/wiki/List_of_countries_and_capitals_in_native_languages (may be hard to screen scrape and also seems to refer to languages/countries w/no ISO code)

[bill]

(In [14211]) Use country data from GeoNames? instead of Wikipedia, and load data on demand. Note that the GeoNames? server provides localized data (ie, country names), so we intentionally pass the page's language setting as a parameter. The good thing about this change is that is shows off RPC. The bad thing is that the tree loads slowly (especially since the GeoNames? server can't filter countries by continent) Refs #4258, fixes #6808.

[Adam Peller]

hmm... looks like a trailing comma problem in IE with their data?

also, locale=ar in IE shows garbage for the data in the tree. an encoding problem?

does it ship down all languages regardless of the request?

[bill]

IE7 is working fine for me and I don't see any trailing commas in ​http://ws.geonames.org/countryInfoJSON?formatted=true.

As per the encoding problem, I see the garbage characters on IE but I think it's an IE problem or a dojo.xhrGet() problem.... I can download ​http://ws.geonames.org/countryInfoJSON?formatted=true&lang=ar (or ja) from firefox or safari and then successfully open the file (using TextEdit?) as UTF-8.

[bill]

(In [14271]) Assume that value returned from script IO is UTF-8. FF and Safari work this way already, but IE does not. Fixes #6808. !strict