Fix anonymous controller route helpers

[thomasfedb]

Continuing on from #881.

Changed orig_routes to @orig_routes as it is referred to in method_missing on line 138. Bug introduced in 86aea05. Added a test that fails when fix is reverted.

[thomasfedb]

Travis is failing with:

/home/travis/build/rspec/rspec-rails/lib/rspec/rails/matchers/relation_match_array.rb:2:
in `<top (required)>': uninitialized constant RSpec::Matchers::BuiltIn::MatchArray (NameError)

Which leaves me baffled, as I didn't touch that file or anything I can think would affect it.

[thomasfedb]

Travis is failing with:

/home/travis/build/rspec/rspec-rails/lib/rspec/rails/matchers/relation_match_array.rb:2:
in `<top (required)>': uninitialized constant RSpec::Matchers::BuiltIn::MatchArray (NameError)

Which leaves me baffled, as I didn't touch that file or anything I can think would affect it.

[myronmarston]

In rspec-expectations I recently renamed that class. See rspec/rspec-expectations@6bde36e and rspec/rspec-expectations#398.

Use RSpec::Matchers::BuiltIn::ContainExactly instead.

[myronmarston]

In rspec-expectations I recently renamed that class. See rspec/rspec-expectations@6bde36e and rspec/rspec-expectations#398.

Use RSpec::Matchers::BuiltIn::ContainExactly instead.

[thomasfedb]

@myronmarston Okay, thanks.

I guess that should be a separate PR however.
Would you be prepared to do so? Or shall I?

[thomasfedb]

@myronmarston Okay, thanks.

I guess that should be a separate PR however.
Would you be prepared to do so? Or shall I?

[myronmarston]

Would you be prepared to do so? Or shall I?

Have at it!

[myronmarston]

Would you be prepared to do so? Or shall I?

Have at it!

[thomasfedb]

@myronmarston The new ContainExactly doesn't recognise ActiveRecord::Relation objects as being enumerable, for the obvious reason that ActiveRecord::Relation.new.ancestors doesn't include Enumerable.

I could either monkeypatch RSpec::Matchers::Composable and change the enumerable? method in the rspec-rails gem, or perhaps you could check for #to_a in the rspec-expectations gem.

This was introduced in rspec/rspec-expectations@ec93540.

[thomasfedb]

@myronmarston The new ContainExactly doesn't recognise ActiveRecord::Relation objects as being enumerable, for the obvious reason that ActiveRecord::Relation.new.ancestors doesn't include Enumerable.

I could either monkeypatch RSpec::Matchers::Composable and change the enumerable? method in the rspec-rails gem, or perhaps you could check for #to_a in the rspec-expectations gem.

This was introduced in rspec/rspec-expectations@ec93540.

[myronmarston]

I've got a fix in rspec/rspec-expectations#404.

[myronmarston]

I've got a fix in rspec/rspec-expectations#404.

[myronmarston]

I just merged that PR so you should be good to go now :). Sorry about the regression!

[myronmarston]

I just merged that PR so you should be good to go now :). Sorry about the regression!

[thomasfedb]

Cheers @myronmarston.

[thomasfedb]

Cheers @myronmarston.

[thomasfedb]

Just squashed my tests into the same commit for neatness, and to trigger Travis.

[thomasfedb]

Just squashed my tests into the same commit for neatness, and to trigger Travis.

[thomasfedb]

Travis is failing on one spec that appears unrelated, as mentioned in #906 by @alindeman.

[thomasfedb]

Travis is failing on one spec that appears unrelated, as mentioned in #906 by @alindeman.

[alindeman]

For now, I decided simply to partially revert the commit. I'd love to have it covered by a spec, but I think we need a common way to run examples (extracted from -core?) first.

[alindeman]

For now, I decided simply to partially revert the commit. I'd love to have it covered by a spec, but I think we need a common way to run examples (extracted from -core?) first.

[thomasfedb]

Consider using the spec attached to this issue once you're ready.

[thomasfedb]

Consider using the spec attached to this issue once you're ready.

[myronmarston]

I think we need a common way to run examples (extracted from -core?) first.

rspec-support could be a good home for the supporting code. I'm working on improving rspec-mocks so that no monkey patching is needed to get it to work in that fashion.

[myronmarston]

I think we need a common way to run examples (extracted from -core?) first.

rspec-support could be a good home for the supporting code. I'm working on improving rspec-mocks so that no monkey patching is needed to get it to work in that fashion.

[thomasfedb]

@myronmarston I did notice that the spec was a little tricky to write.

[thomasfedb]

@myronmarston I did notice that the spec was a little tricky to write.