twitter-cldr automatically sentence-casing elements (TimespanFormatter, patterns)

[pchew-change]

The values for TimespanFormatter:patterns are all sentence cases, with the letter of the first word capitalized, e.g.

• https://github.com/twitter/twitter-cldr-js/blob/master/lib/assets/javascripts/twitter_cldr/es-CO.js#L36
• https://www.evernote.com/l/AKvaqZp2bMdIiKl8oef6lRDy7j27aqyW1jAB/image.png
• "Hace {0} …"

The values in the CLDR code charts are all lower cased, e.g.:

• http://www.unicode.org/cldr/charts/29/summary/es.html#1426
• https://www.evernote.com/l/AKvOfiSJuOdGPrpwTuB2xeXnP_bjZ2l7rwgB/image.png
• "hace {0} …"

This causes problems when the value to be inserted is not in sentence-initial position, viz. in the middle of the sentence, e.g.:

• https://www.evernote.com/l/AKvMY_Qp-RhM8rTEeJ0bsaLUGsiEWTWyp9kB/image.png
• INCORRECT : "Desde el lanzamiento … _H_ace 3 días, …" (Since we started … 3 days ago, …)
• correct: "Desde el lanzamiento … hace 3 días, …" (Since we started … 3 days ago, …)

Please advise.

[camertron]

Hey @pchew-change, unfortunately the fields/units data built into this gem are from CLDR v23, not v29 as you linked. In v23, time spans were uppercased. The ruby version of twitter_cldr also suffers from this issue. As I mentioned in #80, I'm working on supporting the new units data introduced in CLDR v24, which should let us move to v29 without too many problems. Then I should be able to focus on improving the Javascript story.