Lee Hinman

@fourtytoo would you mind adding a blurb to the README including this option? Otherwise I think it'd probably get lost without documentation.

Since the insecure option gets encoded into the connection manager, it'd be nice if an exception was thrown if someone tried to specify both at once, so that people know...

Hi @igrishaev, clj-http doesn't do anything with the stream in the event that is a non-200 exception. In the event of a non-200 exception a user may still want to...

> I think it'd make sense for clj-http to ship with something like this, > otherwise it's kinda hard to test a highly customized HTTP client. You can write your...

@quoll I think what I mentioned in #429 would be the best path forward, let me know what your thoughts are

Okay, do you know if there is a good way I can test NTLM auth? I don't have any services locally that I can stand up to try it and...

> Looks like it is setting additional header "Connection: close" when constructing a request (connection manager is not reusable). But for ntlm it is supposed to be "Connection: keep-alive". When...

@alex-k1 can you expand on what sort of server you are trying to authenticate against? IIS? I'd like to try and set up a reproduction with whatever OS/software you're using.

> The example the OP gave should work, even without the connection pool, right? Yes that's the intent, you should not have to use a connection pool to use NTLM...

Actually, what I previously wrote is incorrect, I've done more reading on NTLM auth and it turns out to require a persistent connection, rather than a once-per-request authentication. I've pushed...