One more layout fix.


git-svn-id: http://root.cern.ch/svn/root/trunk@24505 27541ba8-7e3a-0410-8455-c3a389f83636

fix HTML issues reported by w3c validator.


git-svn-id: http://root.cern.ch/svn/root/trunk@24504 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@24497 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@24496 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@24492 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@24482 27541ba8-7e3a-0410-8455-c3a389f83636

- fix a coding convention violation
- improve and update release notes 
- improve doc, in particular doxygen groups


git-svn-id: http://root.cern.ch/svn/root/trunk@24477 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@24463 27541ba8-7e3a-0410-8455-c3a389f83636

- from Bertrand: disable warning 4250 on Windows 9.0. This is a known deficit of 9.0 in that particular case, 
  see for example http://connect.microsoft.com/VisualStudio/feedback/ViewFeedback.aspx?FeedbackID=101259


git-svn-id: http://root.cern.ch/svn/root/trunk@24424 27541ba8-7e3a-0410-8455-c3a389f83636

  Unfortunatly explicit constructors cannot be added since arenot supported by CINT. One should then use in 
  Python or CINT dummy constructor + operator=.


git-svn-id: http://root.cern.ch/svn/root/trunk@24404 27541ba8-7e3a-0410-8455-c3a389f83636

Main changes are: 

- improve integrator interface adding template functions taking as imput any C++ callable object for evaluation 
  of the integrals
- use ROOT convention for enumeration names (start with k, like kADAPTIVE for the integration type). Change for 
  Integration types, Interpolation types (in MathMore) and Minimizer1D types. Change all classes and tests using this 
  enumeration name  
- improve RichardsonDerivator class (used by TF1::Derivative). Remove an un-needed function call
  Fix also a bug in calculation of the error returned by TF1::DerivativeError
- change interface for the function interfaced used for fitting (IParamFunction function and IParamGrad function)
  Requires now that the parameter values are passed when evaluating the function and the gradient and make the method
  operator()(double *x, double *p) const. Derived class must implement DoEvalPar (and not DoEval as before) 
  for implementing this interface. This change makes the caching of parameter not hidden and the interface is now 
  thread-safe.  In the IParamGradFunction interface remove the funcitonality for gradient of the coordinates (x) 
  since it is not needed. Define only the parameter gradient. 
- improve also the ParamFunction class used by the concrete function like Polynomial. Adapt Polynomial class to new 
  parameteric function interface
- improve Fit Result class with function to get parameter correlations after the fit and to fill any generic matrix 
with covariance or correlation result. 
- add method also in the minimizer interface to get correlation values 
- add strategy in the minimizer control parameters
- increase timing of stressMathCore and stressMathMore


git-svn-id: http://root.cern.ch/svn/root/trunk@24403 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@24401 27541ba8-7e3a-0410-8455-c3a389f83636

- use std::string for parameter name and remove un-necessary limitation on the length of parameter name
   maintan the Name() method returning a char * (for backward compatibility) and add a new method GetName() returning 
   a reference to a std::string
- add also a method in MInuit2Minimizer to get correlation coefficient between parameter i and j  and a method to get 
  the global Correlation coefficient of parameter i
- add new complex tests of fitting used for parallelization studies 
- compile in ROOT also the tests in MnSim and MnTutorial


git-svn-id: http://root.cern.ch/svn/root/trunk@24400 27541ba8-7e3a-0410-8455-c3a389f83636

strangely enough Wndows does not support %p, use good old 0x%lx.


git-svn-id: http://root.cern.ch/svn/root/trunk@24317 27541ba8-7e3a-0410-8455-c3a389f83636

using %p as format flag instead of 0x%x allows pointers to be passed
on 64bits, too.


git-svn-id: http://root.cern.ch/svn/root/trunk@24300 27541ba8-7e3a-0410-8455-c3a389f83636

a fix for TFFTRealComplex::GetPointComplex(Int_t *,...) function, for the case of transforms of more than 2 dimensions. The real index in the internal 1-d flat array is now computed correctly, taking into account that only ~half of entries are stored because of Hermitian symmetry. And thanks to Christophe Yeche for pointing it out.


git-svn-id: http://root.cern.ch/svn/root/trunk@24198 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@24183 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@24161 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@24127 27541ba8-7e3a-0410-8455-c3a389f83636

The concrete implementation of the class is in TCint.
The new functions are essentially interfaces to the existing CINT C++ classes
like G__ClassInfo, G__CallFunc, G__DataMemberInfo, G__MethodInfo, G__TypeInfo
and G__TypedefInfo. Most of these functions will call the Reflex interface
once the CINT/Reflex interface will be available later this year.
All direct calls to CINT are replaced by calls like gCint->ClassInfo_xxx where gCint
points to the TCint implementation.
The existing calls to TCint via gInterpreter are kept because on Windows
gInterpreter does not call directly TCint, but instead call TWin32InterpreterProxy
that in turns call TCint.
On Linux and Mac, gInterpreter and gCint are equivalent.

Once the transformation using gCint will be completed, it will be possible
to specify at startup time which version of CINT (CINT or CINT7) by loading
dynamically TCint or TCint7.--This line, and those below, will be ignored--

M    core/meta/src/TStreamerElement.cxx
M    core/meta/src/TCint.cxx
M    core/meta/src/TGlobal.cxx
M    core/meta/src/TBaseClass.cxx
M    core/meta/src/TDataMember.cxx
M    core/meta/src/TInterpreter.cxx
M    core/meta/src/TClass.cxx
M    core/meta/src/TMethodArg.cxx
M    core/meta/src/TDataType.cxx
M    core/meta/src/TMethodCall.cxx
M    core/meta/src/TIsAProxy.cxx
M    core/meta/src/TMethod.cxx
M    core/meta/src/TFunction.cxx
M    core/meta/inc/TClass.h
M    core/meta/inc/TDictionary.h
M    core/meta/inc/TMethodArg.h
M    core/meta/inc/TDataType.h
M    core/meta/inc/TMethodCall.h
M    core/meta/inc/TMethod.h
M    core/meta/inc/TFunction.h
M    core/meta/inc/TCint.h
M    core/meta/inc/TGlobal.h
M    core/meta/inc/TBaseClass.h
M    core/meta/inc/TDataMember.h
M    core/meta/inc/TInterpreter.h
M    core/thread/src/TThread.cxx
M    core/base/src/TVirtualFitter.cxx
M    core/base/src/TStorage.cxx
M    core/base/src/TQConnection.cxx
M    core/base/src/TROOT.cxx
M    core/base/src/TSystem.cxx
M    core/base/src/TQObject.cxx
M    core/rint/src/TRint.cxx
M    core/rint/src/TTabCom.cxx
M    math/foam/src/TFoam.cxx
M    math/minuit2/src/TFitterMinuit.cxx
M    math/mathcore/src/FunctorCint.cxx
M    math/minuit/src/TMinuit.cxx
M    tree/tree/src/TTree.cxx
M    tree/tree/src/TSelectorCint.cxx
M    tree/tree/src/TSelector.cxx
M    tree/tree/inc/TSelectorCint.h
M    roofit/roofitcore/src/RooGenCategory.cxx
M    roofit/roofitcore/src/RooWorkspace.cxx
M    hist/hist/src/TF1.cxx
M    gui/guibuilder/src/TGuiBldDragManager.cxx
M    net/alien/inc/TAlienSystem.h


git-svn-id: http://root.cern.ch/svn/root/trunk@24077 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@24027 27541ba8-7e3a-0410-8455-c3a389f83636

- fix a potential problem found by Timur (see http://root.cern.ch/phpBB2/viewtopic.php?p=27949#27949)


git-svn-id: http://root.cern.ch/svn/root/trunk@23970 27541ba8-7e3a-0410-8455-c3a389f83636

Use as the name of the template argument of the templated member function a symbol that is _not_ also the name of a function in order to avoid a fatal confusion in order compilers (gcc 3.3)


git-svn-id: http://root.cern.ch/svn/root/trunk@23879 27541ba8-7e3a-0410-8455-c3a389f83636

Use as the name of the template argument of the templated member function a symbol that is _not_ also the name of a function in order to avoid a fatal confusion in order compilers (gcc 3.3)


git-svn-id: http://root.cern.ch/svn/root/trunk@23871 27541ba8-7e3a-0410-8455-c3a389f83636

- improve chi2test to consider the case of weighted histograms (options UW or WW) and the histograms 
  have zero bin errors. 
 (For example in case an histogram is compared to a function represented as an histogram with zero errors)

- suppress in unuran tests info message produced by  chi2 test


git-svn-id: http://root.cern.ch/svn/root/trunk@23859 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@23858 27541ba8-7e3a-0410-8455-c3a389f83636

call TMath::LnGamma instead, otherwise we can explode memory
when fitting histograms with big bin contents with the Likelihood option.


git-svn-id: http://root.cern.ch/svn/root/trunk@23843 27541ba8-7e3a-0410-8455-c3a389f83636

this should make the W3C validator happy.


git-svn-id: http://root.cern.ch/svn/root/trunk@23827 27541ba8-7e3a-0410-8455-c3a389f83636

- improve chi2 test for the treatment of empty bins. Fix cases with option WW or UW and both histograms had empty bins.
  For UU and only one histogram has an empty bin do not decrease ndf as suggested by I. Reid. 

- fix bug in evaluation of chi2 pdf when x = 0 and ndf <=2.


git-svn-id: http://root.cern.ch/svn/root/trunk@23823 27541ba8-7e3a-0410-8455-c3a389f83636

- fixed a bug in TH1::Divide with binomial errors (see https://savannah.cern.ch/bugs/?36372)
- remove some commented code in TF1, update rel notes and improve graphics output of a tutorial


git-svn-id: http://root.cern.ch/svn/root/trunk@23766 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@23747 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@23745 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@23735 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@23704 27541ba8-7e3a-0410-8455-c3a389f83636

add  #pragma link C++ global ROOT_TMath as suggested by Philippe. This avoids via the header guard to re-include the definition of TMath.h when running in CINT. 
This fixes a crash observed running stressFit.cxx in interpreted mode


git-svn-id: http://root.cern.ch/svn/root/trunk@23686 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@23675 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@23673 27541ba8-7e3a-0410-8455-c3a389f83636

git-svn-id: http://root.cern.ch/svn/root/trunk@23660 27541ba8-7e3a-0410-8455-c3a389f83636

- add comments in TMinuitMinimizer class fixing the  coding convention violation


git-svn-id: http://root.cern.ch/svn/root/trunk@23657 27541ba8-7e3a-0410-8455-c3a389f83636

Main changes are: 

  TMath: 

        from David: The following methods now have additional interface working with Iterators as well:  
		MinElement
		MaxElement
		LocMin
		LocMax
		Mean
		GeomMean
		RMS
		BinarySearch
		Sort -> (changed to SortItr)
       	- use now std in binary search algorithm. Old algorithm failed some tests and it is found to be a 
           little bit slower 
       	- do not use std instead in LocMin and LocMax since generic std algorithm is slower on 64 bit arch. 
       	- TMath::PoissonI 
            fix a problem for large values, by using directly the TMath::Poisson implementation
       	- TMath::Landau and TMath::Landau
  		remove duplications by using directly ROOT::Math::landau_pdf and ROOT::Math::landau_cdf
	- TMath::Prob
   		use ROOT::Math::chisquared_cdf_c  (gives also less rounding errors  for small values)
	- TMath::GammaDist
   		use ROOT::Math::gamma_pdf , implementation now returns correct value, 0,  for large values of 
		shape parameter instead of nan
	- TMath::LogNormal:   
  		use ROOT::Math::lognormal_pdf 

	- compile TMath dictionary with -O2 
	- fix a warning in -Wshadow      

        - TMath.h is now  reordered by functionality and alphabetical order

  test: 
	various improvement from David, see http://root.cern.ch/viewcvs?rev=22880&root=root&view=rev	

  - add dictionary for the enumeration used in the Integration class 
  
  - from David: some clean up of Minimizer class (remove typedef definitions)


git-svn-id: http://root.cern.ch/svn/root/trunk@23655 27541ba8-7e3a-0410-8455-c3a389f83636