[Minuit2] Refactor to initialize class members in the class declaration

This makes the code less error prone, because initialization can't be
forgotten or be inconsistent between constructors.

Author: guitargeek

math/minuit2/inc/Minuit2/MnUserParameterState.h

@@ -38,11 +38,7 @@ class MnUserParameterState {
    /// default constructor (invalid state)
    MnUserParameterState()
       : fValid(false),
-        fCovarianceValid(false),
         fCovStatus(-1),
-        fFVal(0),
-        fEDM(0),
-        fNFcn(0),
         fParameters(MnUserParameters()),
         fCovariance(MnUserCovariance()),
         fIntParameters(std::vector<double>()),
@@ -157,11 +153,11 @@ class MnUserParameterState {
 
 private:
    bool fValid;
-   bool fCovarianceValid;
+   bool fCovarianceValid = false;
    int fCovStatus; // covariance matrix status
-   double fFVal;
-   double fEDM;
-   unsigned int fNFcn;
+   double fFVal = 0.;
+   double fEDM = 0.;
+   unsigned int fNFcn = 0;
 
    MnUserParameters fParameters;
    MnUserCovariance fCovariance;

math/minuit2/src/MnUserParameterState.cxx

@@ -20,19 +20,11 @@ namespace Minuit2 {
 // construct from user parameters (before minimization)
 //
 MnUserParameterState::MnUserParameterState(std::span<const double> par, std::span<const double> err)
-   : fValid(true),
-     fCovarianceValid(false),
-     fCovStatus(-1),
-     fFVal(0.),
-     fEDM(0.),
-     fNFcn(0),
-     fParameters(MnUserParameters(par, err)),
-     fIntParameters(par.begin(), par.end())
+   : fValid(true), fCovStatus(-1), fParameters(MnUserParameters(par, err)), fIntParameters(par.begin(), par.end())
 {
 }
 
-MnUserParameterState::MnUserParameterState(const MnUserParameters &par)
-   : fValid(true), fCovarianceValid(false), fCovStatus(-1), fFVal(0.), fEDM(0.), fNFcn(0), fParameters(par)
+MnUserParameterState::MnUserParameterState(const MnUserParameters &par) : fValid(true), fCovStatus(-1), fParameters(par)
 {
    // construct from user parameters (before minimization)
 
@@ -50,7 +42,7 @@ MnUserParameterState::MnUserParameterState(const MnUserParameters &par)
 // construct from user parameters + errors (before minimization)
 //
 MnUserParameterState::MnUserParameterState(std::span<const double> par, std::span<const double> cov, unsigned int nrow)
-   : fValid(true), fCovStatus(-1), fFVal(0.), fEDM(0.), fNFcn(0), fIntParameters(par.begin(), par.end())
+   : fValid(true), fCovStatus(-1), fIntParameters(par.begin(), par.end())
 {
    // construct from user parameters + errors (before minimization) using std::vector for parameter error and    // an
    // std::vector of size n*(n+1)/2 for the covariance matrix  and n (rank of cov matrix)
@@ -66,7 +58,7 @@ MnUserParameterState::MnUserParameterState(std::span<const double> par, std::spa
 }
 
 MnUserParameterState::MnUserParameterState(std::span<const double> par, const MnUserCovariance &cov)
-   : fValid(true), fCovStatus(-1), fFVal(0.), fEDM(0.), fNFcn(0), fIntParameters(par.begin(), par.end())
+   : fValid(true), fCovStatus(-1), fIntParameters(par.begin(), par.end())
 {
    // construct from user parameters + errors (before minimization) using std::vector (params) and MnUserCovariance
    // class
@@ -82,7 +74,7 @@ MnUserParameterState::MnUserParameterState(std::span<const double> par, const Mn
 }
 
 MnUserParameterState::MnUserParameterState(const MnUserParameters &par, const MnUserCovariance &cov)
-   : fValid(true), fCovStatus(-1), fFVal(0.), fEDM(0.), fNFcn(0), fParameters(par)
+   : fValid(true), fCovStatus(-1), fParameters(par)
 {
    // construct from user parameters + errors (before minimization) using
    // MnUserParameters and MnUserCovariance objects
@@ -102,7 +94,7 @@ MnUserParameterState::MnUserParameterState(const MnUserParameters &par, const Mn
 //
 //
 MnUserParameterState::MnUserParameterState(const MinimumState &st, double up, const MnUserTransformation &trafo)
-   : fValid(st.IsValid()), fCovarianceValid(false), fCovStatus(-1), fFVal(st.Fval()), fEDM(st.Edm()), fNFcn(st.NFcn())
+   : fValid(st.IsValid()), fCovStatus(-1), fFVal(st.Fval()), fEDM(st.Edm()), fNFcn(st.NFcn())
 {
    //
    // construct from internal parameters (after minimization)