Osg viewer truncates update viewer update commands

[braineo]

How to reproduce

Draw more than 1000 axes in a loop

import numpy
viewer = env.GetViewer()

indices  = range(2000)

Ts = []
offset = 0
for i in indices:
    T = numpy.eye(4)
    T [:3, 3] = [offset, 0 , 0]
    #T[:3, 3] = numpy.random.random(3)
    Ts += [T]
    offset += 0.01

handlers = []
for i in indices:
    handlers.append(misc.DrawAxes(env, Ts[i], dist=0.025, linewidth=4))

Observation

Only 1000 Axes are drawn on the viewer, suspect the following piece truncating the number of adding child to scene graph

// plugins/qtosgrave/qtosgviewer.cpp
void QtOSGViewer::_PostToGUIThread(const boost::function<void()>& fn, bool block)
{
    if( _nQuitMainLoop != -1 ) {
        // viewer quit, so anything posted won't get processed
        return;
    }

    boost::mutex::scoped_lock lockmsg(_mutexGUIFunctions);
    if( _listGUIFunctions.size() > 1000 ) {
        // can happen if system is especially slow
        //RAVELOG_WARN_FORMAT("too many gui post commands, ignoring %d", _listGUIFunctions.size());
        return;
    }

@Hsifnus

[Hsifnus]

The trivial solution to this would be to expand or remove the limit on _listGUIFunctions.size(). That being said, would it be better if certain functions sent to _PostToGUIThread were able to be condensed down into a single function pushed onto _listGUIFunction that accumulates similar function calls and calls each one in sequence?