Determine plural cases based on the actual language rules

[dimach1977]

Currently language-specific rules are ignored and all localizations are assumed to have, for example, a binary case ("two") even when it isn't needed. This breaks compatibility with localization management tools that do know which languages should have which plural cases and don't provide a way to add unneeded cases.

[dimach1977]

To expand a bit. For example, in Russian and other Slavic languages, there's no need for a special "two" form (and the language specific rule here never returns PluralCase.TWO.)

Localization management tools, such as poeditor, will prompt to create only the language-specific plural cases. Thus it would be impossible to create a proper localization for Slavic languages: easy_localization would look for the "two" entry and if that isn't found would default to the "many" form instead of the "few", like it should. This result in strings like "2 дней".

[dimach1977]

Added tests. Note that there's existing test that encode incorrect behavior or that test for behavior irrelevant to the test (e.g. widget tests checking for specific plurals.) Update those too.

[CarlosVimos]

Hi,

I'm encountering an issue with the pluralization rules for the Polish language in the _pl_rule() function. According to the function definition:

PluralCase

_pl_rule() {
// ...
if (_v == 0 && _i % 10 >= 2 && _i % 10 <= 4 && (_i % 100 < 12 || _i % 100 > 14)) {
return FEW;
// ...
}

This should return the plural form FEW for numbers like 2, 3, 4, 22, 23, 24, and so on (excluding 12, 13, and 14). However, in my testing, it appears that number "2" is instead being categorized under OTHER.

Could this be a potential issue in the pluralization logic, or am I possibly missing something in my setup? I'm using version 3.0.3 of easy_localization.

Would this PR also solve this problem?

[bw-flagship]

@dimach1977 Could you remove the not-referenced function and update the branch? Then we can merge :)

[dimach1977]

@bw-flagship Done (+cleaned up some copy-pasta in other tests)

[bw-flagship]

There is a problem with the CI-Pipeline, we need to merge #631 first

[bw-flagship]

@dimach1977 there are some analyzer issues:

warning • The '!' will have no effect because the receiver can't be null • lib/src/localization.dart:141:55 • unnecessary_non_null_assertion
   info • Unused import: 'dart:async' • test/easy_localization_language_specific_test.dart:1:8 • unused_import
   info • Unused import: 'package:easy_localization/easy_localization.dart' • test/easy_localization_language_specific_test.dart:4:8 • unused_import
   info • Unused import: 'package:easy_logger/easy_logger.dart' • test/easy_localization_language_specific_test.dart:7:8 • unused_import
   info • Unused import: 'package:shared_preferences/shared_preferences.dart' • test/easy_localization_language_specific_test.dart:10:8 • unused_import
   info • Unused import: 'dart:async' • test/easy_localization_logger_test.dart:1:8 • unused_import
   info • Unused import: 'easy_localization_test.dart' • test/easy_localization_logger_test.dart:7:8 • unused_import
   info • Unused import: 'dart:async' • test/easy_localization_test.dart:1:8 • unused_import

[dimach1977]

@bw-flagship ✅