Remove "overriding" translation keys

[MGPalmer]

Refs #431

All specs still pass, and this solves my overriding problem :)

[seejohnrun]

@MGPalmer thanks! do you mind adding a test or two to make sure formats can differ between locales? I'm unsure if they already exist

[MGPalmer]

Hmm I'm not 100% sure I understand what you'd like me to test...

However, looking over the code and specs, it seems to me like only the ice_cube....formats.default is being used, and that is pretty well tested.

In fact, setting a different to_s format seems broken:

mssngr(dev)> IceCube.to_s_time_format = "aaaa"
=> "aaaa"
mssngr(dev)> IceCube.to_s_time_format
=> "%B %-d, %Y"

Maybe because the instance variable set is not actually used in this getter:

  # Defines the format used by IceCube when printing out Schedule#to_s.
  # Defaults to '%B %e, %Y'
  def self.to_s_time_format
    IceCube::I18n.t("ice_cube.date.formats.default")
  end

  # Sets the format used by IceCube when printing out Schedule#to_s.
  def self.to_s_time_format=(format)
    @to_s_time_format = format
  end

but that's a different ticket ;)

[avit]

but that's a different ticket ;)

Yes, it looks like that setting was just left behind after translations were added. We should remove since it does nothing, but that could break code which actually sets it, so it should have a deprecation first.

[avit]

Thanks for this!

Unfortunately, I'm not sure if we can use this approach since those default values come from a rails gem, which would make us dependent on loading all of rails:

• https://github.com/svenfuchs/rails-i18n/blob/master/rails/locale/de.yml#L42
• https://github.com/svenfuchs/rails-i18n/blob/master/lib/rails_i18n/railtie.rb#L1

It would be nice to use those strings, but we would like to avoid this dependency if possible. (I suspect there's a way to reuse them without a runtime gem dependency.)

The real problem is that previously, the load order was causing other problems and now it sounds like we are overwriting these fields the wrong way. I will take another look at the c7b305c commit which broke this.

[avit]

👮 dropped line ending at EOF

[avit]

Also I'm surprised why the test suite is all green here with these removed... something to look into.

I'm still not 100% on this: do all users who use i18n use it with i18n-rails anyway?

[MGPalmer]

Also I'm surprised why the test suite is all green here with these removed... something to look into.

I might be mistaken, but my guess is that the format keys I've removed have not actually been used in ice_cube.

The real problem is that previously, the load order was causing other problems and now it sounds like we are overwriting these fields the wrong way. I will take another look at the c7b305c commit which broke this.

This would of course be my favorite solution!

[pacso]

This will be resolved by #489