this is done until the present ABI issues between gcc and clang are
solved, most notably the ones concerning shared_ptrs.

The CMake build system is broken with genvector=OFF (does not really
disable genvector, and tests fail if that is fixed). Therefore, it
was decided that it's better to just make this always ON and remove
the option for disabling it altogether.

Addresses ROOT-9311

Fix the configuration to find Blas and decide either to build CPU or GPU architecture. Define this in RCondifure.h

This reverts commit dc71c77684f6475ff508593a23044890ff02a82f.

Main application for such commands - creation of the image files.
ROOT code waits from the client to store image in the file.
By default 100 s will be used. Normally image creation is much faster

One should be aware that certificate should be provided as well

...and use it in RIntegerSequence.hxx

Changing option name according better option semantic/documentation cases and extending it to be more generic (supporting zlib, lz4 and lzma)

Only when required, full version of TRootSniffer with all extra
libraries will be loaded. Which is actually happens by default.

Now only full-functional sniffer can produce or parse XML files

Here will be placed the code, which depends from gpad, tree, image and
other libraries

By default OFF, should be enabled only when appropriate software is
installed on the system

At configure time one should add information into RConfigure.h file, if
these components are exists. Only at such case one should try to load
these libraries.

After installing the new VMware Fusion 10 the PATH on macOS looks like:
/usr/local/bin:/usr/bin:/bin:/usr/sbin:/sbin:/Applications/VMware Fusion.app/Contents/Public:
this choked up setxrd.sh where PATH was not quoted. Added quotes around
all path instances for protection in both scripts. Also unset drop_from_pathi
at end of script to avoid bash namespace polution.

Some MesaGL implementations (since 2016) have trouble (segv) when one tries
to store vertex array operations into display lists. The practice is indeed
somewhat questionable :) The following options will allow complete or partial
disablement of display list usage.
OpenGL.UseDisplayLists:                     1
OpenGL.UseDisplayListsForVertexArrays:      1

OpenGL.UseDisplayLists already works.

Handling of OpenGL.UseDisplayListsForVertexArrays still needs to be
implemented as it requires review of all GL rendering classes.

It seems my last commit doesn't properly implement the way we
add a define to the dictionary. It seems this is the proper way
to do it instead of just calling add_definitions.