Zanie Blue

> @zanieb Do we want to block this on _"run the tests without all dependencies installed"_ or not? (My take would probably be... we could consider it but as a...

Maybe it'd be best to be able to specify dumps/loads separately both for user control and to avoid doing `getattr` to get the `dumps/loads` methods? Perhaps: ``` httpx.Client(json_encoder=orjson.dumps, json_decoder=orjson.loads) ```

@xbeastx I think it's quite clearly articulated at https://github.com/encode/httpx/pull/1352#issuecomment-845817581 why that pull request went stale. There is some additional helpful context at https://github.com/encode/httpx/pull/1730#issuecomment-874011751 and discussion at https://github.com/encode/httpx/discussions/1740

I strongly agree that global state is not a good path forward for the library. I like the `request_class` and `response_class` approach — that would also help with some other...

Looks similar to the previously resolved https://github.com/encode/httpcore/pull/679 Thanks for all the detailed commentary!

See https://github.com/encode/httpcore/pull/643 to resolve type errors (outside of tests) by adding `retries` to SOCKS proxy classes.

> I don't think we need test cases where what we're doing is passing through an argument, and then inspecting the private implementation details to see that it's been set....

My need for this is not pressing as I've implemented wrapper around the client but I'd love to keep discussing the best way to handle this when you have time!

This is annoying... sorry. The solution is probably to create a temporary [git config](https://git-scm.com/docs/git-config#SCOPES) then configure or disable the [git credential helper](https://git-scm.com/docs/gitcredentials) — I don't think we should need one...

@konstin is this fixed.... or did we just turn the test off?