Externals Update, 21.2 branch (2020.03.04.)

This MR is updating all projects in branch 21.2 to atlasexternals-1.0.61. While the jump from the currently used atlasexternals-1.0.58 may seem big, it's only the changes in the "last step" that are important to this branch.

The list of changes (atlasexternals@1.0.58...1.0.61) are the following:

• Added FindVBFNLO.cmake to AtlasLCG (needed by AthGeneration);
• Taught FastJet and FastJetExternals how to live without each other. I.e. to be able to build FastJetExternals against a version of FastJet coming from the system, or from an LCG release (needed by AthGeneration);
• Updated AthAnalysisExternals to use LCG_96b;
  • Removed the build of Eigen, FastJet and FastJetContrib from the project as now "new enough" versions are coming from LCG_96b;
• Updated all externals built by AnalysisBaseExternals to be in sync with the versions available in LCG_96b.

For a detailed update on the version changes of the externals introduced by the move to LCG_96b, see: atlasexternals!625 (merged)

The change in AthDerivation, to pick up FastJet and FastJetContrib from LCG, also brought a problem in the CMake configuration of JetRec to light. (It was hidden so far because those two externals were always coming from the same include and library paths before now.)

I was also trying to take this opportunity to make some fixes in AthDerivation, in how it builds on CentOS 7, which still has some issues. 😦

http://atlas-computing.web.cern.ch/atlas-computing/links/distDirectory/gitwww/GITWebArea/nightlies/21.2/2020-03-05T0018/AthDerivation/x86_64-centos7-gcc62-opt/AthDerivation/

But after looking at the issue for about an hour, I still don't understand it. 😦 So that will have to be handled separately after all...

Note that the AthDerivation release is not going to be fully shielded from this update. 🤔 So @egramsta and @boeriu should keep an eye on this MR as well.