class RooMinimizer: public TObject

 Cleanup method called by atexit handler installed by RooSentinel
 to delete all global heap objects when the program is terminated

 Construct MINUIT interface to given function. Function can be anything,
 but is typically a -log(likelihood) implemented by RooNLLVar or a chi^2
 (implemented by RooChi2Var). Other frequent use cases are a RooAddition
 of a RooNLLVar plus a penalty or constraint term. This class propagates
 all RooFit information (floating parameters, their values and errors)
 to MINUIT before each MINUIT call and propagates all MINUIT information
 back to the RooFit object at the end of each call (updated parameter
 values, their (asymmetric errors) etc. The default MINUIT error level
 for HESSE and MINOS error analysis is taken from the defaultErrorLevel()
 value of the input function.

 Destructor

 Change MINUIT strategy to istrat. Accepted codes
 are 0,1,2 and represent MINUIT strategies for dealing
 most efficiently with fast FCNs (0), expensive FCNs (2)
 and 'intermediate' FCNs (1)

 Set the level for MINUIT error analysis to the given
 value. This function overrides the default value
 that is taken in the RooMinimizer constructor from
 the defaultErrorLevel() method of the input function

 Change MINUIT epsilon

 Choose the minimzer algorithm.

 Parse traditional RooAbsPdf::fitTo driver options

  m - Run Migrad only
  h - Run Hesse to estimate errors
  v - Verbose mode
  l - Log parameters after each Minuit steps to file
  t - Activate profile timer
  r - Save fit result
  0 - Run Migrad with strategy 0

 Execute MIGRAD. Changes in parameter values
 and calculated errors are automatically
 propagated back the RooRealVars representing
 the floating parameters in the MINUIT operation

 Execute HESSE. Changes in parameter values
 and calculated errors are automatically
 propagated back the RooRealVars representing
 the floating parameters in the MINUIT operation

 Execute MINOS. Changes in parameter values
 and calculated errors are automatically
 propagated back the RooRealVars representing
 the floating parameters in the MINUIT operation

 Execute MINOS for given list of parameters. Changes in parameter values
 and calculated errors are automatically
 propagated back the RooRealVars representing
 the floating parameters in the MINUIT operation

 Execute SEEK. Changes in parameter values
 and calculated errors are automatically
 propagated back the RooRealVars representing
 the floating parameters in the MINUIT operation

 Execute SIMPLEX. Changes in parameter values
 and calculated errors are automatically
 propagated back the RooRealVars representing
 the floating parameters in the MINUIT operation

 Execute IMPROVE. Changes in parameter values
 and calculated errors are automatically
 propagated back the RooRealVars representing
 the floating parameters in the MINUIT operation

 Change the MINUIT internal printing level

 If flag is true, perform constant term optimization on
 function being minimized.

 Save and return a RooFitResult snaphot of current minimizer status.
 This snapshot contains the values of all constant parameters,
 the value of all floating parameters at RooMinimizer construction and
 after the last MINUIT operation, the MINUIT status, variance quality,
 EDM setting, number of calls with evaluation problems, the minimized
 function value and the full correlation matrix

 Create and draw a TH2 with the error contours in parameters var1 and v2 at up to 6 'sigma' settings
 where 'sigma' is calculated as n*n*errorLevel

 Start profiling timer

 Stop profiling timer and report results of last session

 Apply results of given external covariance matrix. i.e. propagate its errors
 to all RRV parameter representations and give this matrix instead of the
 HESSE matrix at the next save() call

 Import the results of the last fit performed, interpreting
 the fit parameters as the given varList of parameters.