Fix rendering explicit template in 1.1

[Burgestrand]

See #121 for reported issue.

To get CI to pass I first needed to fix that, which is why 1799dca is part of this PR. See #122 for that in isolation.

Original issue report as follows:

Hi! This report is still early, because I haven't figured out the why yet, but I have figured out the what.

Ruby 3.2.1, Rails 7.1.2.

The change in b0cc8bc broke this code:

def template 
  render(
    template: "orders/layouts/single_room",
    locals: { view: self, dato: }
  )
end

orders/layouts/single_room is an ERB file, but I'm not sure that matters because it never gets that far in the backtrace anyway.

With this backtrace:

     ActionView::Template::Error:
       'nil' is not an ActiveModel-compatible object. It must implement :to_partial_path.
     # /Users/kim/.asdf/installs/ruby/3.2.1/lib/ruby/gems/3.2.0/gems/actionview-7.1.2/lib/action_view/renderer/abstract_renderer.rb:82:in `partial_path'
     # /Users/kim/.asdf/installs/ruby/3.2.1/lib/ruby/gems/3.2.0/gems/actionview-7.1.2/lib/action_view/renderer/object_renderer.rb:20:in `render_object_derive_partial'
     # /Users/kim/.asdf/installs/ruby/3.2.1/lib/ruby/gems/3.2.0/gems/actionview-7.1.2/lib/action_view/renderer/renderer.rb:96:in `render_partial_to_object'
     # /Users/kim/.asdf/installs/ruby/3.2.1/lib/ruby/gems/3.2.0/gems/actionview-7.1.2/lib/action_view/renderer/renderer.rb:55:in `render_partial'
     # /Users/kim/.asdf/installs/ruby/3.2.1/lib/ruby/gems/3.2.0/gems/actionview-7.1.2/lib/action_view/helpers/rendering_helper.rb:35:in `block in render'
     # /Users/kim/.asdf/installs/ruby/3.2.1/lib/ruby/gems/3.2.0/gems/actionview-7.1.2/lib/action_view/base.rb:291:in `in_rendering_context'
     # /Users/kim/.asdf/installs/ruby/3.2.1/lib/ruby/gems/3.2.0/gems/actionview-7.1.2/lib/action_view/helpers/rendering_helper.rb:33:in `render'
     # /Users/kim/.asdf/installs/ruby/3.2.1/lib/ruby/gems/3.2.0/gems/phlex-rails-1.1.0/lib/phlex/rails/sgml/overrides.rb:26:in `render'

So far I've only verified that reverting that change only is sufficient for the code to not raise. I tested this by modifying the single line in the installed phlex-rails gem on my machine and re-running my tests.

[Burgestrand]

I believe I found a suitable fix, and so this is ready for review.

The PR that introduced this problem is #113. I do not know if the change I made just now brings that problem back.

[Burgestrand]

I found a more focused test approach, and added a regression test for #113 too. Turns out my initial fix did indeed break that, but now with my recent change (force-pushed, I didn't think that piece of history was worth preserving) this passes both use-cases.

With that I'm done with this until you've had a look.

[joeldrapper]

Thanks for this. Your fix looks good.

[joeldrapper]

I’m merging this in #130 with a minor fix for older versions of Ruby.

[Burgestrand]

Great, thanks! Oh, I forgot to update the changelog.

[joeldrapper]

No problem. I’ll sort the change log now. I’m going to do a patch release with this fix.