- automatic architecture detection (first architecture argument
    only needed if non-default compiler is required)
  - better detection and avoidance of duplicate libraries for the
    authentication modules (afs, krb5 and srp modules can be build
    at the same time)
Mods by Christian Holm and me.


git-svn-id: http://root.cern.ch/svn/root/trunk@7197 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@7196 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@7195 27541ba8-7e3a-0410-8455-c3a389f83636

from the header file to the implementation file.
Add a protection in case of NaN to avoid an infinite loop.
(thanks Gero Flucke for the suggestion)


git-svn-id: http://root.cern.ch/svn/root/trunk@7194 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@7193 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@7192 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@7191 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@7190 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@7189 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@7188 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@7187 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@7186 27541ba8-7e3a-0410-8455-c3a389f83636

G__REDIRECTIO was not set for the windows build (using MS compiler).
Defining it solves the problem.


git-svn-id: http://root.cern.ch/svn/root/trunk@7185 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@7184 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@7183 27541ba8-7e3a-0410-8455-c3a389f83636

not needed (anymore) to compile and link graf/src/G* with -g. The "redo"
script had as side effect to relink also libCore with -g which was the
real solution.


git-svn-id: http://root.cern.ch/svn/root/trunk@7182 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@7181 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@7180 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@7179 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@7178 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@7177 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@7176 27541ba8-7e3a-0410-8455-c3a389f83636

new authentication scheme supports, the following methods: user/passwd
(but save not clear passwd over wire), ssh, krb5, srp, globus and gui/uid
(like RFIO uses). Everything is documents here:
http://root.cern.ch/root/Auth.html.


git-svn-id: http://root.cern.ch/svn/root/trunk@7175 27541ba8-7e3a-0410-8455-c3a389f83636

new authentication scheme supports, the following methods: user/passwd
(but save not clear passwd over wire), ssh, krb5, srp, globus and gui/uid
(like RFIO uses). Everything is documents here:
http://root.cern.ch/root/Auth.html.


git-svn-id: http://root.cern.ch/svn/root/trunk@7174 27541ba8-7e3a-0410-8455-c3a389f83636

authentication patch that is following.


git-svn-id: http://root.cern.ch/svn/root/trunk@7173 27541ba8-7e3a-0410-8455-c3a389f83636

Alice can be drawn in 14.5s instead of 1400s.

-Add bit kVisOnScreen in TGeoAtt.
-Replace calls to TGeoPainter::IsOnScreen by calls to TGeoAtt::TestBit(TGeoAtt::kVisOnScreen)
-delete function TVirtualGeoPainter::IsOnScreen and TGeoPainter::IsOnScreen
-In TGeoNode replace call to TGeoPainter::IsOnScreen by TGeoAtt::TestBit(TGeoAtt::kVisOnScreen)


git-svn-id: http://root.cern.ch/svn/root/trunk@7172 27541ba8-7e3a-0410-8455-c3a389f83636

std::multimap in gcc 3 and above takes an allocator<pair<key, value> > (as does the std::map).


git-svn-id: http://root.cern.ch/svn/root/trunk@7171 27541ba8-7e3a-0410-8455-c3a389f83636

TFormula::TFormula(TFormula const &) was indirecly accessing uninitialized memory.
This is fixed by initializing the value in it.

TF1::Copy was leaking memory (not deleting existing arrays), once this was fixed,
this required to add the proper initialization in TF1::TF1(TF1 const &).


git-svn-id: http://root.cern.ch/svn/root/trunk@7170 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@7169 27541ba8-7e3a-0410-8455-c3a389f83636

loop on all array elements. This gives a factor 10 when drawing Alice in ray-tracing mode.
Yet a better solution would be to build a sorted table of pointers to visible volumes
followed by a binary search.


git-svn-id: http://root.cern.ch/svn/root/trunk@7168 27541ba8-7e3a-0410-8455-c3a389f83636

- a couple of fixes in TGeoTrd2::DistToIn() and TGeoTorus::DistToIn()
- computation of normals for composite shapes - these can now be
visualized with raytracing.


git-svn-id: http://root.cern.ch/svn/root/trunk@7167 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@7166 27541ba8-7e3a-0410-8455-c3a389f83636

  o RooAbsReal

    - Fix bug in asymmetry plotting of slice of a RooSimultaneous


git-svn-id: http://root.cern.ch/svn/root/trunk@7165 27541ba8-7e3a-0410-8455-c3a389f83636

"running Root under supervision of valgrind I've noticed a possible
problem with strcpy(G__ifile.name,shlfile) call in
int G__shl_load(shlfile).

G__shl_load is called by G__shl_load(G__ifile.name) in loadfile.c:1928
resulting in undefined operation strcpy(G__ifile.name,G__ifile.name)"


git-svn-id: http://root.cern.ch/svn/root/trunk@7164 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@7163 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@7162 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@7161 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@7160 27541ba8-7e3a-0410-8455-c3a389f83636

The package has 3 classes TMultiLayerPerceptron, TNeuron, TSynapse

// TMultiLayerPerceptron
//
// This class describes a neural network.
// There are facilities to train the network and use the output.
//
// The input layer is made of inactive neurons (returning the
// normalized input), hidden layers are made of sigmoids and output
// neurons are linear.
//
// The basic input is a TTree and two (training and test) TEventLists.
// For classification jobs, a branch (maybe in a TFriend) must contain
// the expected output.
// 6 learning methods are available: kStochastic, kBatch,
// kSteepestDescent, kRibierePolak, kFletcherReeves and kBFGS.
//
// This implementation is *inspired* from the mlpfit package from
// J.Schwindling et al.


git-svn-id: http://root.cern.ch/svn/root/trunk@7159 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@7158 27541ba8-7e3a-0410-8455-c3a389f83636