* Revert "[cmake] Narrow the list of included directories."

This reverts commit 88c16588.

* Revert "[cxxmodules] Build cling runtime into module."

This reverts commit 76a9ce75.

therewith for tens of widely used classes such as variables, constants,
distributions and combinations thereof.

this was due to an improper assumption in the way in which the binning
was accessed.

XrdSecEntity::Reset() is only available in newer versions of XRootD.

ROOT-9595 #close

but add a comment to the jitted code in order to obtain a clear error message if
the user tries to use this type in a jitted action/transformation later in the chain.

There are a number of trivial operations that users often want to
perform on dataframes that are surprisingly hard to get right, for
example adding several Defines in a loop or conditionally adding a
Filter depending on a runtime boolean (both use-cases are challenging in
C++, trivial in python).

Difficulties boil down to the fact that different dataframe nodes have
different types (because their types incorporate e.g. the type of the
callable passed to a Filter and the type of their parent node in the
computation graph).

This commit adds a common base class ROOT::RDF::RNode to all
nodes of the graph (except leaves a.k.a results, which have a completely
different interface), so that users can, for example:

* take any dataframe node by reference in non-template functions as RNode&
* keep dataframe nodes in `std::vector<RInterface<RNode>>`
* keep non-const pointers to dataframe nodes

and so on.

Internal RDataFrame code is also simplified by the introduction of this
common base class.
The downside is that if this mechanism is abused users might end up with extra,
unnecessary virtual calls in their event loop -- on the other hand, this
mechanism should only be used in situations that before required either complex
template magic or dirty and slow tricks.

In order to this PR to work, RFilter and Datasources expose their names.

Before this commit only the root node, the RLoopManager, kept the list of the custom columns defined by the user. This meant that it was not possible to define two columns with the same names in two different branches.

After this commit, each node in the graph has a map with pointers to all the columns defined up to that node. Nodes in different branches may see different custom column. This allows operations like the following to work:

 ROOT::RDataFrame d(3);
   auto branch1 = d.Define("b2", []() { return 1; });
   auto branch2 = d.Define("b2", []() { return 2; });

- Improved overall code quality
- Fixed possible missing initialization of the custom columns. Now the loop manager has again an observer pointer to all the custom columns, only to initialize them.

This reverts commit 783cddc4.

This reverts commit e56b4a96.

There are two problems with this test and cxxmodules. First, it picks up a
header files from $ROOTSYS/include and builds a library. Naturally, we should
build pcm files, however, building module MathCore requires specific -I to
build the needed VecCore and Vc components. This test does not have access
to the include relevant include paths. Secondly, if we want to reuse the
modules from ROOT which makes most sense we can't because we get a hard
error such as: fatal error: malformed or corrupted AST file:
'SourceLocation remap refers to unknown module, cannot find include/pcms/1WYSNQV9VBZK7/stl-2OZGQN92C38MI.pcm

We can fix the first point by moving out all VecCore-related headers
such as Math/Types.h and all of its includers in a separate module. Thus,
the current test will not require the VecCore (as in the textual case).
Alternatively, we can trace the origin of the fatal error and try to remap
the source locations.

The test should reuse the prebuilt modules in the module cache. However,
due to a mismatch between the module-related flags exported by FindROOT.
There is a mismatch between the -D passed by cmake.

This patch make sure that the build arguments match more closely the build
setup of ROOT. Thus we can reuse the already built module files.

[RDF,Tutorial] Add tutorial processing NanoAOD-like file producing dimuon spectrum from Run2011A CMS OpenData