Circular buffering for optimal performance

[cobaltfusion]

Circular buffering for optimal performance

[janwilmans]

Not every LogSource might have a handle to wait for, how would we deal with that? create an event for the logsource are trigger it from a poll-loop ?

[janwilmans]

std::vector does not release any memory allocated unless shrink_to_fit() is called.
I have tested a circular buffer implementation (of my own creation) but the performance is not nearly as good as std::vector. This might be because std::vector can use move-symantics and the circular buffer requires a byte-by-byte copy.

also, because the circular buffer does not solve any other problems, I have decided to remove the implementation and close this issue.

Preventing the std::merge is accomplished but a WaitForMultipleObjects and this does not rely on the circular buffer for serialization.