[InMemory] Fixed deadlock

[mhutchinson]

Interesting bug found by accident when trying to benchmark in memory:

1. Client calls Witness.Update, which calls Persistence.Update. The in-memory persistence acquires a write lock on its global mutex to safely update the checkpoint.
2. Persistence.Update calls the callback function provided to perform consistency checks.
3. To verify the signature on the previous checkpoint, the callback needs the log's public key. It calls Persistence.Log, which attempts to acquire a read lock on the same global mutex
4. Deadlock!

This introduces a sync.Map for the logs (which are largely intialized up-front), and uses a mutex for the checkpoint state.

[codecov-commenter]

Codecov Report

❌ Patch coverage is 54.54545% with 5 lines in your changes missing coverage. Please review.
✅ Project coverage is 24.63%. Comparing base (26841b7) to head (1cec140).

Files with missing lines	Patch %	Lines
persistence/inmemory/inmemory.go	54.54%	5 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #559      +/-   ##
==========================================
+ Coverage   24.35%   24.63%   +0.28%     
==========================================
  Files          27       27              
  Lines        2012     2005       -7     
==========================================
+ Hits          490      494       +4     
+ Misses       1417     1406      -11     
  Partials      105      105              

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.