Allow let(:send)

[soulcutter]

Not sure if there was a reason that treat_symbols_as_metadata_keys_with_true_values was hanging around in here, so let me know if I should blow away that commit.

This fixes rspec/rspec-core#1163

[soulcutter]

@xaviershay Looks like the build is hitting something similar to rspec/rspec-mocks#458 ?

[xaviershay]

probably because this is a string not a symbol. If you make it :WHAT it should pass, or rebase off my PR to fix that other issue (pretty sure that's ready to merge now).

[soulcutter]

Ahh I think I see. Thanks for taking a look. (although this is not the spot where the error occurs)

[myronmarston]

I'm confused...why does it matter that it's a symbol vs a string? And what does this have to do with @xaviershay's PR? It had to do with declaring a test double, but I don't see that being done here...

[soulcutter]

@myronmarston It wasn't here, it was from

rspec-rails/spec/rspec/rails/mocks/mock_model_spec.rb

Lines 199 to 203 in 2325288

	it "does not stub column accessor if already stubbed in declaration (with string)" do
	model = mock_model(MockableModel, "column_a" => "a")
	model.respond_to?("column_a")
	model.column_a.should eq("a")
	end

[myronmarston]

Ah, got it. Thanks!

[soulcutter]

The build was fixed when @xaviershay 's rspec-mocks changes got merged in.

/cc @alindeman

[soulcutter]

Should we backport this as well?

[myronmarston]

Should we backport this as well?

Given that it's as bug, it would be good to backport it to 2.99 and 2.14.

[myronmarston]

This LGTM, @soulcutter. Will leave to you to merge so you can take care of backporting it at the same time (I worry about merging and us forgetting to back port since it's not still open).