* A big step towards solving #910
1) Rank atoms by inverse atomic number (ensures Hs go last)
2) Use heavy-atom degree when doing layout (ensures Hs do not change gross geometry)
* basic testing
* take degree into account with ranks too
* further improvements of relative ranking.
definitely need some cleanup of this now
* some cleanup
* add an explicit test
- added #ifdef M_PI (...) #endif in all relevant places
- made length() and sqLength() method consistent
with respect to usage of pow(x, 2) vs x*x in
Code/Geometry/point.h
- removed gzip-related boost.iostreams dependency and
replaced with portable "cmake -E tar xzf" command
in Code/ForceField/MMFF/CMakeLists.txt
This does not address the non-canonical rendering piece -- that will
be considerably more work -- but it does take care of the
epoxides being rendered inside the ring.
instead of using the property map interface.
A nice side-effect is that the wart of having to use property maps to loop over bonds or atom neighbors
is now gone.
This potentially breaks lots of client C++ code.
