Relative translation keys inside blocks resolve to partials scope instead of callers scope

[AckermannTM]

What are you trying to accomplish?

Relative translations in a block use the path of the block, instead of the caller. See #2413

What approach did you choose and why?

Created an issue to describe the problem and was asked to create a failing test PR.
It took me some time to put the PR together 😅

Anything you want to highlight for special attention from reviewers?

I just hope I did everything correctly

[myabc]

Hi @AckermannTM thanks for creating the bug reproduction! If I try running your test and inserting a binding.irb within the render block, I see that calling t/translate calls the ActionView implementation. Outside of the render block, it calls the ViewComponent implementation. This is actually what I would I would expect to see.

bundle exec appraisal rails-8.0 rake test TEST=test/sandbox/test/components/relative_translation_partial_component_test.rb

<%= render "shared/partial" do %>
  <%= t(".title") %>
  <% binding.irb %>
<% end %>

> method(:translate).source_location
=> ["/Users/alexbcoles/.gem/ruby/3.4.5/gems/actionview-8.0.2/lib/action_view/helpers/translation_helper.rb", 73]

vs

<%= render "shared/partial" do %>
  <%= t(".title") %>
<% end %>
<% binding.irb %>

> method(:translate).source_location
=> ["/Users/alexbcoles/git-repos/view_component/lib/view_component/translatable.rb", 93]

[AckermannTM]

Hey @myabc, thanks for looking into this! That explanation makes sense. Inside a render block it is ActionView handling translations, and outside it falls back to the ViewComponent implementation. So this is not a ViewComponent issue but the expected ActionView behavior when rendering a partial inside a component.

I thought it might be a ViewComponent issue since it worked in version 3.23.2, but now it is clear that this is expected behavior.

Looks like it’s time to replace all partials with components in my project 😄

[Spone]

Thanks @AckermannTM for taking the time to create this PR and @myabc for chiming in!

Since it's not a ViewComponent issue, do you think it would be useful to modify the failing tests in order to document the expected behavior?

[AckermannTM]

Hey @Spone , thanks for the suggestion! I'm sorry for replying this late.

I'm going to adjust the test to document this behavior.