Fixes #216, via @harrytodorov.

Fixes #214; issue noted by @anooprthomas.

(And reenable the unit test that confirms this.)

It's wasteful to expand the provided Float to an Interval since then
redundant work is done in the full (Interval op interval) implementation.

This gives a ~5% speedup for scenes with many quadric shapes, where
Interval is used heavily in the intersection routines.

Silences a compiler warning with nvcc 11.5.

(The MSVC build fails when these are inconsistent.)

* Add rendercoordsys as a scene Option

* Remove quote protection as it is not needed

There should be no reason to bother, but there are frequent and
inexplicable timeouts on github's Windows runners when it tries to check
out this repository; these often lead to bogus build failures. The hope is
that this works around that...

This works around a NaN case in SampleSphericalTriangle(). (Ideally that
function would be more numerically robust, but that's a bigger issue...)

Issue #211.

Via @shadeops. Issue #210.

A missing ResolveFilename() in the function that associated loaded normal
map images with materials caused normal maps to be dropped when pbrt was
started in a different directory than the scene file was in.

Fixes issue #210.

Due to round-off, BasicSceneBuilder::CTMIsAnimated() may return true even
though the corresponding AnimatedTransform ends up storing the same two
transforms. This was triggering an assert in the AnimatedPrimitive
constructor...

Fixes #204.

Accessing parameters of the std::moved-from entity was not effective...

Null-valued member variables were leading to trying to initialize std::strings
with nullptr rather than printing "nullptr".

Update filesystem/ and stb/ to latest versions.
Start using git submodules for lodepng (and update to latest.)

Reduce the epsilon to better resolve fine shadows (especially in the Moana Island scene.)

Also add a hack to fix a rare infinite loop when finding all intersections
along a ray for subsurface scattering with quadric shapes (root cause bug
not yet chased down.)

The Future class is now gone, and its functionality ("is the result
ready?", "give me the result value") is now folded into AsyncJob.  This
allows retrieving the result from multiple threads, with synchronization
handled in AsyncJob.  (This is more flexible than std::future allows, since
we don't need either move semantics or the ability to pass along exceptions
when the future is harvested.)

This fixes a number of race conditions that helgrind was flagging during
startup with complex scenes.  (Our prior use of std::future wasn't correct
in that we had unprotected access attempts from multiple threads.)

This allows users to not need to worry about thread safety in their
callback implementations. In particular, this fixes a race in the
BVHAggregate's use of ThreadLocal that rarely led to crashes or complaints
about malloc heap corruption during BVH construction.

Given C++'s rules for initialization of statics, I don't think this is
strictly necessary, but it makes helgrind happy and causes no harm.

There were two signed integer multiplications that readily overflowed. They
are now both unsigned, which is well-defined under overflow.

Issue #177.

(Rather than just printing a warning.) Issue #152.