Consider merge to nock, and nock to flatiron.

[heapwolf]

Let's consider bringing @pgte's nock (https://github.com/pgte/nock) into flatiron in favor of this project and merging any features from this project into nock.

[indexzero]

Not opposed to this, although there doesn't seem to be anything in this library worth porting. @pgte I have some reservations about overwriting http.request. Not in-so-far that I disagree with it, but that it's not 100% mocked.

All it takes is one smart-ass library author (or ancient library) to use http.createClient or instantiate http.ClientRequest directly. This circumvents nock.

Would you be open to overriding http.ClientRequest instead?

[pgte]

@indexzero Yup, no problem, fortunately http.ClientRequest is exposed. I opened this issue: nock/nock#26

[xla]

Only to add some thoughts from the nodejs-dev google group where it's stated that bare constructors will likely go away in the future and which would leave us with factories and the intended interfaces.

[indexzero]

@goldjunge I don't take away that point from the discussion at all: felixge, mikeal, polotek, and piscisaureus were all for it. That's a pretty solid list of +1s

[pgte]

@indexzero @hij1nx I recently finished a big refactoring of nock and http.ClientRequest overriding is included. It now contains 91 client tests (awesome 8 contributors).

[heapwolf]

w00t! Let's drop it into flatiron! Just transfer it from the admin page whenever you're ready. And then we can just delete this repo.

[indexzero]

@pgte @hij1nx Sounds good. We shouldn't straight up delete this repository for a couple of months though. A deprecation warning and redirect to nock will help transition anyone using this.

[heapwolf]

Agreed, closing issue.

[pgte]

@hij1nx I don't have write access to the flatiron org. How do you want to do this?

[bithavoc]

How about adding a disclaimer in the readme? "We recommend you to use Nock instead" or something.

[indexzero]

Added deprecation notice. Closing.