Code optimizations for MessageSorter and FieldMap classes
Brought to you by:
orenmnero
Hi, everyone. In the attachment you'll find some code optimizations which can yield in some scenarios, according to quickfix performance test suite, up to 20-50%. This was achieved by reducing the amount of data copying and unnecessary memory management code. This applied to the latest commit [2269]
In the second version of this patch I've added the reference counted field buffer usage in FIX::message_order struct - this should greately help in those scenarios where message has dozens of repeating groups. Boost::shared_array<> turned to be too slow for the task =/
Looks great, I matched your results.