Todd Wolfson

Bump =/

It isn't out of the question but proxying a request is already a solved problem. I would be hesitant to introducing unnecessary/unreusable work.

Looks good for the most part, left mostly stylistic changes and a stack trace concern. Please get Travis CI green before landing. I will add @mlmorg as an npm publisher.

Reminder for when this is landed: - Update `CHANGELOG.md` - `git tag` version (`foundry` suggested)

I am confused what you mean by the Integration Contract test. Can you explain further?

I thought I wrote this down but I guess not. Consider using `object-hash` for our key serialization: https://www.npmjs.org/package/object-hash

I sadly don't have a lot of time to look into this =/ Feel free to rollback the problematic code, though I feel like the `n+1` performance was pretty egregious...

Can you provide more examples of when this comes into play? This is starting to highlight lines where we could have abstracted things better (e.g. make `fixed-server` a router middleware...

For external reference, I belive the current workaround is: ``` js response: [ myGlobalMiddleware(), function (req, res) {} ] ``` which runs the response handlers in order via a middleware...

I could still see this being practical if someone doesn't want `express.urlencoded()` for any POST-based mocks. Reopening the issue to prevent burial.