WIP: test to show that mocking & resolving cannot be used concurrentl…

[sebastienbarre]

Per @helfer suggestion in apollographql/apollo-server#9, submit a PR with a failing test, showing a case where mocks and resolver can't be use concurrently.
Thank you.

[apollo-cla]

@sebastienbarre: Thank you for submitting a pull request! Before we can merge it, you'll need to sign the Meteor Contributor Agreement here: https://contribute.meteor.com/

[helfer]

@sebastienbarre Thanks, I think this is an excellent start. Now we just have to figure out how to make it do what we would expect 😬

According to the comments in the code it should already be doing that (mock.js), so the next step you could take would be to figure out why that test is failing by seeing which branch it takes for mocking the returnString field. The code isn't that straight forward (it's functions wrapping functions), so let me know if you have any questions!

[sebastienbarre]

@helfer thanks. So here is a stab at it. The if (preserveResolvers && field.resolve) was essentially bypassing the mockers, I moved it below so that I can merge the result of a resolver and the result of the mock, if they are both objects. This might look pretty clumsy, I trust you can make it look better through the magic of functional programming. Now I'm not sure what that big block below the we have to handle the root mutation and root query types differently comment does, but since it sets field.resolve, it might require checking preserveResolvers there too. The tests pass, so whatever that block is doing is not being exercised by the tests exercising preserveResolvers, right?
Thanks!

[helfer]

I think this should work! If I remember correctly how mocking is done, it shouldn't be an issue on the root value.

Can your rebase this on master so I can merge it?

[sebastienbarre]

@helfer done 👍

[helfer]

Perfect, thanks!

[sebastienbarre]

Unfortunately, this is now broken :( Remember when I mentioned "Now I'm not sure what that big block below the we have to handle the root mutation and root query types differently comment does"? Something had to be done there too. I'll have to come with a new test and fix...

[helfer]

Yes, I remember that. Unfortunately I think I forgot it and merged the PR before I took a look at that.