updating install.xml to use the new glod autobuilt library.

Updating cmake to ensure that linux uses the non-stub version of convex decomp.

Code reviewed by Falcon

pushing fix for glod to work on linux.

Patch taken from JIRA, committing locally to fix our project branch.
Code reviewed by Nyx / Brad

Another typo - was disabling the UI, whether the debug option was true or false
simple copy/paste error.

Fixing a typo in my previous patch - want to use setRenderTypeMask instead of
andRenderTypeMask.

Deminimis change, no review

minor cleanup and fixing the linux build.

Fix for crash when loading models (Don't hit UI from the main thread).  Added asserts to LLFastTimer to verify main thread use only.

tweaked some numbers that were incorrect when I originally wrote the algorithm

deminimis change, no review

First pass - uses the new algorithm to hopefully be more accurate of render load
on low-end machines. Also accounts for mesh complexity, including if a mesh is
weighted or non-weighted.

Code reviewed by davep

Probably a regression from the prevalidator rework - was enforcing non-negative int when precision=0 (i.e. non-float), should have been enforcing only full int range.

Fix double definition of create_jump_keys="true" introduced in the fix for VWR-20724 to pass coding policy check.