Thanks for the review.
http://gwt-code-reviews.appspot.com/1025801/diff/1/7 File tools/cldr-import/src/com/google/gwt/tools/cldr/LocalizedNamesProcessor.java (right): http://gwt-code-reviews.appspot.com/1025801/diff/1/7#newcode96 tools/cldr-import/src/com/google/gwt/tools/cldr/LocalizedNamesProcessor.java:96: if (!"ZZ".equals(regionCode) && regionCode.length() == 2) { On 2010/10/22 17:12:20, pdr wrote:
In the case of a 3-letter regionCode: can they always be ignored?
(because they
are aliases?)
3-letter region codes are like 001=World, 419=Latin America, etc, not country codes. ZZ is excluded because it means unknown or invalid region. http://gwt-code-reviews.appspot.com/1025801/diff/1/7#newcode129 tools/cldr-import/src/com/google/gwt/tools/cldr/LocalizedNamesProcessor.java:129: if (++count > 10 || region.getLiteratePopulation() < 3000000) { On 2010/10/22 17:12:20, pdr wrote:
Why stop at 3M?
When choosing a likely country for a given language, you don't want to include a country with only a small number of speakers of that language. Experimentation found 3M was a threshold that provided good results. http://gwt-code-reviews.appspot.com/1025801/diff/1/7#newcode367 tools/cldr-import/src/com/google/gwt/tools/cldr/LocalizedNamesProcessor.java:367: pw.println(" \"" + Processor.quote(code) + "\","); On 2010/10/22 17:12:20, pdr wrote:
This will print an extra comma
Yes, that is perfectly fine in Java. http://gwt-code-reviews.appspot.com/1025801/show -- http://groups.google.com/group/Google-Web-Toolkit-Contributors