Timeout pairs after not receiving data for long enough

[mickel8]

This PR:

• timeouts pairs if they don't receive any data for long enough (5 seconds)
• sets end-of-candidates flag after 10 seconds (this is chosen arbitrarly) starting from calling set_remote_credentials (see RFC 8863)
• converts selected_pair into selected_pair_id
• renames handle_timeout to handle_ta_timeout

[codecov]

Codecov Report

Attention: Patch coverage is 71.30435% with 33 lines in your changes are missing coverage. Please review.

Project coverage is 81.06%. Comparing base (fcf31c9) to head (befbe87).

❗ Current head befbe87 differs from pull request most recent head 7ed7cac. Consider uploading reports for the commit 7ed7cac to get more accurate results

Additional details and impacted files

@@            Coverage Diff             @@
##           master      #40      +/-   ##
==========================================
+ Coverage   80.63%   81.06%   +0.43%     
==========================================
  Files          25       25              
  Lines        1301     1368      +67     
==========================================
+ Hits         1049     1109      +60     
- Misses        252      259       +7     

Files	Coverage Δ
lib/ex_ice/priv/candidate_pair.ex	63.63% <100.00%> (ø)
lib/ex_ice/priv/conn_check_handler/controlling.ex	78.26% <100.00%> (+10.07%)	⬆️
lib/ex_ice/ice_agent.ex	46.80% <33.33%> (-2.09%)	⬇️
lib/ex_ice/priv/ice_agent.ex	84.56% <80.28%> (+0.42%)	⬆️
lib/ex_ice/priv/conn_check_handler/controlled.ex	53.84% <48.48%> (+3.84%)	⬆️

... and 1 file with indirect coverage changes

---

Continue to review full report in Codecov by Sentry.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update fcf31c9...7ed7cac. Read the comment docs.

[LVala]

Maybe better to log the now - pair.last_seen instead of @pair_timeout?

In theory we could go almost twice as long as @pair_timeout without marking it as failed (which is fine, I think), so I can think of a case when this might be a bit misleading.

Edit: Now I see that your're checking every div(@pair_timeout, 2) but I think the point still stands, I guess?