This way, we don't end up with breakage if rustls is built with more
than one provider enabled. This can (and does) happen if a transitive
dependency (such as reqwest *hinthint*) depends on a specific,
non-default rustls backend.
This allows to detect and handle dying streams without getting stuck
forever.
Timeouts are always wrong, though, so we put the burden of choosing the
right values (mostly) on the creator of a stream.
