Reconciling RBNF rules and test data for zh-Hant

[Kip Cole via CLDR-Users]

As part of my build process I run json test data agains the rules defined in Cldr 31.   In some cases the data and rules don’t match.  For example:

Test data for zh-Hant (json format) - noting no negative rule for spellout-numbering-year:

      "SpelloutRules": {
        "%spellout-numbering-year": {
          "x.x": "=0.0=;",
          "0": "=%spellout-numbering=;",
          "1000": "=%spellout-numbering=;”,

RBNF for zh-Hant - note the negative numbers for the test

    "spellout-numbering-year": {
      "-1141": "負二負二負五負一",
      "-1142": "負二負二負五負二",
      "-1143": "負二負二負五負三",
      "-100": "負一百",
      "-75": "負七十五",
      "-50": "負五十",
      "-24": "負二十四",
      "0": "〇",
      "1": "一”,

If I understand correctly the test data may not be being actively maintained. If so I will consider the RBNF rules themselves as authoritative and adjust the test data according.

—Kip