Bleh, had some old initialization code in place that meant I was using
32 connections. (Always verify with 'netstat'...) Logic is now to
use 1/4 of MeshMaxConncurrentRequests to head in the direction of 8
at a time. Full count is used to implement a high-water level keeping
llcorehttp in work.
Initial work completed on linux, moving over to windows to do debug
and refinement. This includes 5/6 handlers based on existing responders
and use of llcorehttp for the mesh header fetch.
llcorehttp implements a nice trace facility but it hasn't been attached to
an external control to date. This hands over control to a *non-persistent*
setting that can be used for QA or field diagnostics.
Isolate llcorehttp initialization into a utility class (LLAppCoreHttp)
that provides glue between app and library (sets up policies, handles
notifications). Introduce 'TextureFetchConcurrency' debug setting to
provide some field control when absolutely necessary.