Fix the issues found in the review of 1.0.0

[DAB-LABS]

I had 1.0.0 reviewed by an outside party before asking Home Assistant to switch to it. The review confirmed the library talks to devices like the original does.

They also mentioned that the new version keeps one network socket open per device, a slow reply to an old request could get mistaken for the reply to a new one. Poor connectivity could show up as a wrong sensor reading or state.

We think we fixed it after confirming on bench hardware. Tested on Python 3.13 and 3.14, 252 tests, plus a live run against an RM4 Pro.

Coding assistant's notes on this change:

Replies are now matched to their request by the packet counter the device echoes back at offset 0x28 (verified on an RM4 Pro, six for six). A reply carrying the counter of a request that already timed out is dropped. A reply whose counter matches nothing we sent is still accepted, so firmware that does not echo the counter keeps working as before.

Also in this change: concurrent callers hitting an expired session key share one re-authentication, and the "logged out" code (-2) now triggers re-auth; changing device.host reopens the socket; aclose() during a request fails it immediately instead of waiting out the timeout; capture() treats ReadError (-10) as "nothing yet" alongside StorageError (-5), which the CLI inherits; a capture() generator abandoned without closing no longer blocks the next one, and there is a new capture_active property; the README explains that broadlink and python-broadlink cannot share an environment; and the changelog no longer describes the carried-over device commits as intact (they were squash-merged with Co-authored-by credit).