Allow setting the timeout on every Call type

The code panics when changing the timeout of a call created using Call::unbounded_wait.
However, this doesn't make much sense, as both bounded and unbounded calls are represented by the same struct, with the only difference being an optional timeout.

In my use case, for instance, I have a service that receives a default Call instance and, only in specific cases, needs to set a timeout. Currently, this simple scenario is impossible to implement because there's no way to know if setting the timeout will cause a panic.

I’d propose removing the change_timeout() function and instead adding a simpler and more ergonomic alternative, such as:

    pub fn with_timeout(mut self, timeout_seconds: Option<u32>) -> Self {
        self.timeout_seconds = timeout_seconds;
        self
    }

Otherwise, please provide different structs for bounded and unbounded calls, so instead of a runtime panic, we can check at compile time whether a timeout can be set. But, at that point, please also add a way to convert from one type to the other.

Start by locating the Rust Call type and its unbounded_wait and change_timeout entry points. Review how bounded and unbounded calls store timeout state, then determine whether a unified timeout-setting API or separate convertible types fits the existing design. Done means changing a timeout no longer panics for an unbounded call and the chosen API is covered by relevant tests.