ticket:13415 Refactoring shadowed virtual method in IFixedDecimal.

ticket:13415 Refactoring shadowed virtual method in IFixedDecimal.

[sffc, 2017-10-20 22:25:34]

On 2017/10/20 22:25:10, sffc wrote:

This is for b/68045724

[markus.icu, 2017-10-21 05:15:24]

This is not really wrong, but normally the fix for this problem is to add "using
IFixedDecimal::getPluralOperand;" in DecimalQuantity, next to the declaration of
the real function.

See https://stackoverflow.com/questions/9995421/gcc-woverloaded-virtual-warnings

See also class DecimalFormat which has using NumberFormat::format; because it
does not override every format() function.

[sffc, 2017-10-21 10:21:21]

In Java this is a pure interface, so it's confusing that this class had a
non-pure virtual method in the first place. I took the opportunity to
refactor the code and replace the method with a helper function.

On Oct 20, 2017 10:15 PM, <markus.icu@gmail.com> wrote:

> This is not really wrong, but normally the fix for this problem is to
> add "using IFixedDecimal::getPluralOperand;" in DecimalQuantity, next to
> the declaration of the real function.
>
> See
> https://stackoverflow.com/questions/9995421/gcc-woverloaded-
> virtual-warnings
>
> See also class DecimalFormat which has using NumberFormat::format;
> because it does not override every format() function.
>
> https://codereview.appspot.com/333920043/
>

[markus.icu, 2017-10-21 16:20:25]

LGTM

Java interfaces used to not allow function implementations.

But, either way is ok with me, and more parallel code C vs. J is good.

https://codereview.appspot.com/333920043/diff/1/i18n/plurrule.cpp
File i18n/plurrule.cpp (right):

https://codereview.appspot.com/333920043/diff/1/i18n/plurrule.cpp#newcode793
i18n/plurrule.cpp:793: PluralOperand operand =
tokenTypeToPluralOperand(digitsType);
optional/nit: I find such functions more readable as

    PluralOperand operand = pluralOperandFromTokenType(digitsType);

[sffc, 2017-10-23 07:13:13]

Committed revision 40626.