I’ve also had a look at your fork and patches. Unfortunately you haven’t
added any tests, so it’s hard to simply cherry-pick things. I’ll go
ahead and add tests at least for the bugs you’ve reported (not sure
about the additional features like load_path, yet).
I currently don’t think this is a bug but actually the correct
behaviour. Could you please provide example scenarios where you want
fallback locale lookups take precedence over defaults? (See the example
given in the ticket comment above.)
One thing though. Do you really think that claiming authorship for this
one is correct?
It’s quite obviously a slightly modified version of the Globalize2
LoadPath.
We’ve decided against moving this functionality to I18n a while ago. You
might want to make it a plugin instead. But if you open source it
somewhere, please be sure to maintain the actual authorship.
Thanks a lot. You found much better and consistent way to fix them
(all together )
I currently don’t think this is a bug but actually the correct behaviour. Could you please provide example scenarios where you want fallback locale lookups take precedence over defaults? (See the example given in the ticket comment above.)
As I wrote in ticket comment, the example is
ActiveRecord::Base#human_attribute_name method which always passes
default which is a attribute_key_name.to_s.humanize string.
So e.g. when I set locale to pl-PL and have:
pl:
activerecord:
attributes:
user:
name: “name in pl”
I will get “name” for User.human_attribute_name(:name), what’s not
correct.
One thing though. Do you really think that claiming authorship for this one is correct?