This reverts commit 19159269.
The minicern library is used by ATLAS.

These are temporary static libraries that need to be linked to a
final shared library, so they must be compiled with -fPIC.

  There are some codes compiled at the start up time. For example,
    - #include \"cling/Interpreter/RuntimeUniverse.h\"
    - #include \"cling/Interpreter/DynamicLookupRuntimeUniverse.h\"
    - namespace cling { class Interpreter; namespace runtime { Interpreter* gCling }}}
    - PrintValue
  These are passed to Cling as string and initialized at the start up time.  So I think it makes sense to reduce top-level global variables, #includes and virtual functions.

  1. Global variables
    If we break at emitModule, we can get a list of global variables and functions which are actually deserialized. These include functions, variables, STL classes and all the functions derives from them.
 I tried to change them to for example constexpr, so that it's processed at compile time.
  2. Eagerly Deserialized decls
    Thanks to @Axel 's hint and tip, we could minimize eagerly deserialized decls deserialized in ASTReader::PassInterestingDeclsToConsumer. We already removed most of eagerly deserialized decls (Some are remaining to be removed, some are hard to remove and some don't cost a lot).

  So far, we got 9.2% of cpu time improvement and 8.8% of memory improvement at start up time in release build.

  - root.exe -q -l
    - master
      cpu time = 0.09186914285714286 sec (average of 7 times)
      res  memory = 142.008 Mbytes

    - HEAD
      cpu time = 0.08337842857142856 sec
      res  memory = 129.508 Mbytes

  - hsimple.C
  Improved by 13% of cpu time and 8.5% of memory
    - master
      cpu time = 0.0954708 sec (average)
      res  memory = 142.891 Mbytes

    - HEAD
      cpu time = 0.0833258 sec
      res  memory = 130.73 Mbytes

  With modules
    - Improvement by 17.7% in cputime and 2% in memory on root.exe -q -l
    (For memory, small improvement is because most of the memory is taken by LoadModules)
    - With this patch, modules is 11.2% slower in cpu time and 6% better in residential memory.

We will not fix them, some are even "fine".
See https://sft.its.cern.ch/jira/browse/ROOT-9179 for more info.

Removed in C++17. No functionality change intended.

[net/monalisa/src/TMonaLisaWriter.cxx:956] -> [net/monalisa/src/TMonaLisaWriter.cxx:957]: (error) Iterator 'iter' used after element has been erased.
[proof/proofd/src/XrdProofdProofServMgr.cxx:4648] -> [proof/proofd/src/XrdProofdProofServMgr.cxx:4646]: (error) Iterator 'iter' used after element has been erased.

[math/minuit2/src/Minuit2Minimizer.cxx:669] -> [math/minuit2/src/Minuit2Minimizer.cxx:669]: (style) Same expression on both sides of '||'.
[math/minuit2/src/Minuit2Minimizer.cxx:736] -> [math/minuit2/src/Minuit2Minimizer.cxx:736]: (style) Same expression on both sides of '||'.
[math/minuit2/src/Minuit2Minimizer.cxx:753] -> [math/minuit2/src/Minuit2Minimizer.cxx:753]: (style) Same expression on both sides of '||'.

[tree/tree/src/TBasket.cxx:852]: (style) Redundant condition: If 'flag >= 80', the comparison 'flag' is always true.
[misc/table/src/TFileIter.cxx:467] -> [misc/table/src/TFileIter.cxx:468]: (warning) Identical condition 'thisRunNumber<runNumber', second condition is always false

[math/mathcore/src/TMath.cxx:826]: (style) Array index 'ia' is used before limits check.
[math/mathcore/src/TMath.cxx:830]: (style) Array index 'ib' is used before limits check.
[math/mathcore/src/triangle.c:15434]: (style) Array index 'aspectindex' is used before limits check.
[roofit/roofit/src/RooIntegralMorph.cxx:375]: (style) Array index 'igapHigh' is used before limits check.

[core/winnt/src/TWinNTSystem.cxx:5035]: (style) Statements following return, break, continue, goto or throw will never be executed.
[tree/treeplayer/src/TSelectorDraw.cxx:380]: (style) Statements following return, break, continue, goto or throw will never be executed.

[net/auth/src/TAuthenticate.cxx:4205]: (error) Resource leak: fd
[roofit/roofitcore/src/BidirMMapPipe.cxx:1880]: (error) Memory leak: s
[tree/tree/src/TLeaf.cxx:222]: (error) Memory leak: countname
[tree/tree/src/TLeaf.cxx:226]: (error) Memory leak: countname
[tmva/tmva/src/MethodBase.cxx:2783]: (error) Memory leak: pdfS
[tmva/tmva/src/MethodBase.cxx:2783]: (error) Memory leak: pdfB

[roofit/roofitcore/src/RooAbsArg.cxx:280]: (style) Redundant checking of STL container element existence before removing it.

[core/dictgen/src/rootcling_impl.cxx:457]: (performance) Possible inefficient checking for 'fieldSelRules' emptiness.
[core/dictgen/src/rootcling_impl.cxx:4573]: (performance) Possible inefficient checking for 'filesIncludedByLinkdef' emptiness.
[gui/canvaspainter/v7/src/TCanvasPainter.cxx:322]: (performance) Possible inefficient checking for 'fWebConn' emptiness.
[gui/canvaspainter/v7/src/TCanvasPainter.cxx:374]: (performance) Possible inefficient checking for 'fCmds' emptiness.
[gui/canvaspainter/v7/src/TCanvasPainter.cxx:419]: (performance) Possible inefficient checking for 'fWebConn' emptiness.
[gui/canvaspainter/v7/src/TCanvasPainter.cxx:475]: (performance) Possible inefficient checking for 'fWebConn' emptiness.
[gui/canvaspainter/v7/src/TCanvasPainter.cxx:587]: (performance) Possible inefficient checking for 'fCmds' emptiness.
[gui/canvaspainter/v7/src/TCanvasPainter.cxx:787]: (performance) Possible inefficient checking for 'fCmds' emptiness.
[html/src/TDocParser.cxx:684]: (performance) Possible inefficient checking for 'currentType' emptiness.
[proof/proofd/src/XrdProofdManager.cxx:804]: (performance) Possible inefficient checking for 'uwrks' emptiness.
[proof/proofd/src/XrdProofdSandbox.cxx:467]: (performance) Possible inefficient checking for 'staglst' emptiness.
[proof/proofd/src/XrdProofdSandbox.cxx:571]: (performance) Possible inefficient checking for 'actln' emptiness.
[roofit/roofitcore/src/RooSimWSTool.cxx:574]: (performance) Possible inefficient checking for 'slist' emptiness.

all the reports Prefer prefix ++/-- operators for non-primitive types

This refactors the CMake build files to no longer manually call
the CMake functions for generating and linking the dicionairies.
One reason is to reduce boilerplate, the other is that it is now
no longer possible to have naming mismatches between the dictionary
generation arguments and linking function arguments which was causing
the race conditions we had in our build system.

This also now follows more strictly the separation between dependencies
like Math, Tree and Core and pure linking flags to external libraries.

The naming of the library in the two function calls ROOT_GENERATE_DICTIONARY
and ROOT_LINKER_LIBRARY was often not matching. This causes that
ROOT_LINKER_LIBRARY couldn't correctly set its dependencies to the output
file of ROOT_GENERATE_DICTIONARY and therefore causes race conditions in
the build system.

Note: We're attached all dependencies to a custom target, and
NOT to the output files generated by rootcling. This should mitigate
the race conditions we experience when multiple targets in different
Make jobs request the same output file as a dependency and then
suffer from this Make problem that is described here:
https://cmake.org/Bug/view.php?id=10082

This patch also adds a (commented out) piece of code that can print
warnings if we don't have a fitting G__*.cxx file for a
ROOT_LINKER_LIBRARY call. This is sometimes intended, so I didn't
enable this warning by default. We should enable this code
by default in the future once we have a way to express if we
intentionally don't provide the G__*.cxx file when calling
ROOT_LINKER_LIBRARY.

ROOT_GENERATE_DICTIONARY allows to specify dependencies, but we currently
don't actually specify those. As we need this for getting the module
dependencies right (as we can't build missing modules on demand), we should
add those dependencies here. Then those dependencies propagate to the
rootcling invocation which will in the future also generate the C++ module
for the selected dictionary.

ClassImp is a macro defined in Rtypes.h. It can be written without trailing
semicolon. However, clang-format is based on lexing and it doesn't know at
all about preprocessors and how to expand macros. When visiting ClassImp(X)
it thinks this is not a completed entity (as it doesn't expand macros) and
considers the next line as continuation of the previous.

Inserting a semicolon fixes the issue and makes formatting with clang-format
in the ROOT codebase less tricky.

Many headers contains redundant header guards around #include directives:

    #ifndef ROOT_TTree
    #include "TTree.h"
    #endif

This patch removes the #ifndef's around these includes as they
don't serve any pratical purpose anymore and are no longer part
of the current ROOT coding convention.

This patch also fixes the 153 typos that are contained in the
symbols of the #ifndef directives.

Signed-off-by: Vassil Vassilev <vvasilev@cern.ch>

Make cannot deal with 'export *', they have to be escaped and fixed later on.

This is the recommended way of using the modules feature and should pave our way
of enabling -fmodules-local-submodules-visibility mode.

Use the standard #include <typeinfo> and #include <iosfwd> instead.

Our two build systems can generate almost entirely a modulemap file laying out
one module per library. Due to some non-modular implementation in ROOT's core we
still need to treat some header files specially.

This patch appends modular header files to the modulemap while skipping the
predefined ones. Configure-make relates all headers that are installed in
$ROOTSYS/include to their library whereas relates only headers which are part of
library's dictionary.

- Spell check

- Spell check

gcc complains that passing any other value than 0 is unsafe and show be used only in debug situations

Makes modules builds with -fno-implicit-module-maps go further.

Either include the missing header files or add forward declarations,
where the complete types are not required.