Add support for KeyLogWriter

[kegsay]

Fixes #343 - style is done in the same way as tls.Config.

Manually tested both as a client and as a server, then reading the resulting logs in Wireshark and confirming that data can be decrypted.

[codecov]

Codecov Report

Merging #344 (8d125fe) into master (eee772b) will decrease coverage by 0.04%.
The diff coverage is 80.00%.

@@            Coverage Diff             @@
##           master     #344      +/-   ##
==========================================
- Coverage   75.84%   75.80%   -0.05%     
==========================================
  Files          86       86              
  Lines        3660     3670      +10     
==========================================
+ Hits         2776     2782       +6     
- Misses        600      603       +3     
- Partials      284      285       +1     

Flag	Coverage Δ
go	75.82% <80.00%> (-0.05%)	⬇️
wasm	65.90% <80.00%> (+0.04%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
config.go	100.00% <ø> (ø)
handshaker.go	74.30% <71.42%> (-0.15%)	⬇️
conn.go	82.60% <100.00%> (+0.03%)	⬆️
flight4handler.go	83.76% <100.00%> (+0.06%)	⬆️
flight5handler.go	81.73% <100.00%> (+0.08%)	⬆️
internal/net/dpipe/dpipe.go	93.24% <0.00%> (-2.71%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update eee772b...2c19c14. Read the comment docs.

[Sean-Der]

This looks fantastic! Would be happy to take right away. Would you mind adding a test? Then should be good to go :)

I can fix all the other little things (linters and so on)

[daenney]

I'd like to see some actual tests for this. Since the KeyLogWriter is an io.Writer, you can put in a bytes.Buffer during test and check what ends up in there.

[kegsay]

It's hard to see at what level you wish to add tests for:

• Verify something is written to the io.Writer which looks roughly right according to https://developer.mozilla.org/en-US/docs/Mozilla/Projects/NSS/Key_Log_Format - effectively a unit test for writeKeyLog.
• Verify that something is written at the right stage of the handshake process. - tests at the handshakeFSM level.
• Verify that the data written to io.Writer can actually decrypt traffic. - E2E test.

The latter being the most useful but also the most effort to write tests for.

[Sean-Der]

I am happy with 1/2!

I think we can just remove randomness and then just assert a certain value is generated everytime. Just enough to make sure someone doesn't regress your work in the future :)

[kegsay]

Test was added a few weeks ago.

[daenney]

The tests are fine, but as you can see from the PR status, our commit linter disagrees: https://github.com/pion/dtls/pull/344/checks?check_run_id=1838025192#step:3:15. It's a small thing, but we like to keep this consistent on this repo and everything else in the Pion org. Might be easiest to fold the test commit into the original one, that should make everything happy.

[Sean-Der]

Thank you so much for contributing this @kegsay! I squashed your commits into 4879d34 also released v2.0.8