Now, that I am trying to document this, I am hesitating with the resulting interface:
1.per default, Session keeps the previous payload around.
2.user has to explicitly clear the internal state (if he knows the library well enough)

Do you know if people are actually using session to send the same payload multiple times? Because if they are not, I´d rather call RemoveContent() after every request that carries a body.

MultiperformPerformTests.MultiperformMixedMethodsPerformTest seems to be flaky - did you see it have issues before?

Now, that I am trying to document this, I am hesitating with the resulting interface: 1.per default, Session keeps the previous payload around. 2.user has to explicitly clear the internal state (if he knows the library well enough)

Do you know if people are actually using session to send the same payload multiple times? Because if they are not, I´d rather call RemoveContent() after every request that carries a body.

Ah, yes :D
Yes, people are actually using this behaviour (and I'm as well for some of my API requests). Because of this and since it would be a really drastic change in how cpr::Session works I'd keep it as it is.

This always has been a bit of a discussion point but we decided to go this way. Even if it's the more intuitive way that in reality might almost always require more code to clear the content before reusing.

MultiperformPerformTests.MultiperformMixedMethodsPerformTest seems to be flaky - did you see it have issues before?

Yes, especially on macOS. This is a limitation of the mock server we are using for our tests.

Approved, merged and is now part of the 1.11.x branch. I will try to make a new release over the next weekend.

thank you very much for the good collaboration!