Keep only the last event index in ExecutionContextSnapshot

[bilboquet]

• document all added functions
• try in sandbox /simulation/labnet
  • if part of node-launch, checked using the resync_check flag
• unit tests on the added/changed features
  • make tests compile
  • make tests pass
• add logs allowing easy debugging in case the changes caused problems
• if the API has changed, update the API specification

[bilboquet]

Slot execution time comparison

Code has been instrumented with many traces, hence the execution time for 5k transactions which used to be around 80ms on this computer raised to around 200ms on average.
Anyway the trace are the same in both cases and with the fix the code perform better.

test case	without fix	with fix
5k failing roll sells (each fails generate an event)	4707 ms	259 ms
Execution time of 5k roll buys	173 ms	161 ms
Execution time of 5k roll sells	198 ms	168 ms
Execution time of 5k transactions	203 ms	191 ms
2 Execution of 1k messages	executing 454 async msg took 270047 µs executing 454 async msg took 215032 µs executing 92 async msg took 30476 µs executing 454 async msg took 264684 µs executing 454 async msg took 219176 µs executing 92 async msg took 31194 µs	executing 454 async msg took 200977 µs executing 454 async msg took 169043 µs executing 92 async msg took 28021 µs executing 454 async msg took 201414 µs executing 454 async msg took 169810 µs executing 92 async msg took 27887 µs

[damip]

Nice study of execution times ! I think we are good thanks to this fix :)