TSchemaRule.cxx

Go to the documentation of this file.

// @(#)root/core:$Id$
// author: Lukasz Janyst <ljanyst@cern.ch>
 
#include "TSchemaRule.h"
 
#include "RConversionRuleParser.h"
#include "TClassEdit.h"
#include "TSchemaRuleProcessor.h"
#include "TSchemaType.h"
#include "TObjArray.h"
#include "TObjString.h"
#include "TROOT.h"
 
#include <utility>
#include <iostream>
#include <vector>
#include <list>
#include <string>
#include <cstdlib>
#include <sstream>
 
ClassImp(TSchemaRule);
 
using namespace ROOT;
 
namespace {
   static Bool_t IsIrrelevantCharacter(char c)
   {
      return (c == ' ' || c == '\n' || c == '\t');
   }
 
   static void AdvanceOverIrrelevantCharacter(const char*& str)
   {
      while (IsIrrelevantCharacter(*str)) {
         ++str;
      }
   }
 
   // check if lhs matches rhs if we ignore irelevant white space differences
 
   static Bool_t IsCodeEquivalent(const TString& lhs, const TString& rhs)
   {
      Bool_t result = kTRUE;
 
      if (lhs != rhs) {
         const char null = '\0';
         const char* left = lhs.Data();
         const char* right = rhs.Data();
         Bool_t literal = false;
         // skip initial white space
         AdvanceOverIrrelevantCharacter(left);
         AdvanceOverIrrelevantCharacter(right);
 
         while (*left != null || *right  != null) {
            // If both chars are white space, skip consecutive white space
            if (!literal && IsIrrelevantCharacter(*left) && IsIrrelevantCharacter(*right)) {
               AdvanceOverIrrelevantCharacter(left);
               AdvanceOverIrrelevantCharacter(right);
               continue;
            }
            // Check if one string has trailing white space, and the other doesn't.
            if (*left == null || *right == null) {
               AdvanceOverIrrelevantCharacter(left);
               AdvanceOverIrrelevantCharacter(right);
               result = (*left == null && *right == null);
               break;
            }
 
            if (*left != *right) {
               result = kFALSE;
               break;
            }
 
            if (*left == '"') {
               literal = !literal;
            }
            ++left;
            ++right;
         }
      }
      return result;
   }
 
}
 
////////////////////////////////////////////////////////////////////////////////
/// Default Constructor.
 
TSchemaRule::TSchemaRule(): fVersionVect( nullptr ), fChecksumVect( nullptr ),
                            fTargetVect( nullptr ), fSourceVect( nullptr ),
                            fIncludeVect( nullptr ), fEmbed( kTRUE ),
                            fReadFuncPtr( nullptr ), fReadRawFuncPtr( nullptr ),
                            fRuleType( kNone )
{
}
 
////////////////////////////////////////////////////////////////////////////////
/// Destructor.
 
TSchemaRule::~TSchemaRule()
{
   delete fVersionVect;
   delete fChecksumVect;
   delete fTargetVect;
   delete fSourceVect;
   delete fIncludeVect;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Copy Constructor.
 
TSchemaRule::TSchemaRule( const TSchemaRule& rhs ): TObject( rhs ),
                            fVersionVect( nullptr ), fChecksumVect( nullptr ),
                            fTargetVect( nullptr ), fSourceVect( nullptr ),
                            fIncludeVect( nullptr ), fEmbed( kTRUE ),
                            fReadFuncPtr( nullptr ), fReadRawFuncPtr( nullptr ),
                            fRuleType( kNone )
{
   *this = rhs;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Copy operator.
 
TSchemaRule& TSchemaRule::operator = ( const TSchemaRule& rhs )
{
   if( this != &rhs ) {
      fVersion        = rhs.fVersion;
      fChecksum       = rhs.fChecksum;
      fSourceClass    = rhs.fSourceClass;
      fTarget         = rhs.fTarget;
      fSource         = rhs.fSource;
      fInclude        = rhs.fInclude;
      fCode           = rhs.fCode;
      fEmbed          = rhs.fEmbed;
      fReadFuncPtr    = rhs.fReadFuncPtr;
      fReadRawFuncPtr = rhs.fReadRawFuncPtr;
      fRuleType       = rhs.fRuleType;
      fAttributes     = rhs.fAttributes;
   }
   return *this;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Return true if the rule have the same effects.
 
Bool_t TSchemaRule::operator == ( const TSchemaRule& rhs ) const
{
   if( this != &rhs ) {
      Bool_t result = ( fVersion == rhs.fVersion
                       && fChecksum == rhs.fChecksum
                       && fSourceClass == rhs.fSourceClass
                       && fTargetClass == rhs.fTargetClass
                       && fSource == rhs.fSource
                       && fTarget == rhs.fTarget
                       && fInclude == rhs.fInclude
                       && IsCodeEquivalent(fCode, rhs.fCode)
                       && fEmbed == rhs.fEmbed
                       && fRuleType == rhs.fRuleType
                       && fAttributes == rhs.fAttributes );
      if (result &&
          ( (fReadRawFuncPtr != rhs.fReadRawFuncPtr && fReadRawFuncPtr != nullptr && rhs.fReadRawFuncPtr != nullptr)
           ||  (fReadFuncPtr != rhs.fReadFuncPtr && fReadFuncPtr != nullptr && rhs.fReadFuncPtr != nullptr) ) )
      {
         result = kFALSE;
      }
 
      return result;
   }
   return kTRUE;
}
 
////////////////////////////////////////////////////////////////////////////////
/// The ls function lists the contents of a class on stdout. Ls output
/// is typically much less verbose then Dump().
 
void TSchemaRule::ls(Option_t *targetname) const
{
   TROOT::IndentLevel();
   std::cout << "Schema Evolution Rule: ";
   if (fRuleType==kReadRule) std::cout <<  "read ";
   else if (fRuleType==kReadRawRule) std::cout << "readraw ";
   std::cout << "\n";
   TROOT::IncreaseDirLevel();
   TROOT::IndentLevel();
   std::cout << "sourceClass=\"" << fSourceClass << "\" ";
   if (fVersion.Length())  std::cout << "version=\"" << fVersion << "\" ";
   if (fChecksum.Length()) std::cout << "checksum=\"" << fChecksum << "\" ";
   if (targetname && targetname[0]) std::cout << "targetClass=\"" << targetname << "\" ";
   else std::cout << "targetClass\"" << fTargetClass << "\" ";
   std::cout << "\n";
   TROOT::IndentLevel();
   std::cout << "source=\"" << fSource << "\" ";
   std::cout << "target=\"" << fTarget << "\" ";
   std::cout << "\n";
   if (fInclude.Length()) {
      TROOT::IndentLevel();
      std::cout << "include=\"" << fInclude << "\" " << "\n";
   }
   if (fAttributes.Length()) {
      TROOT::IndentLevel();
      std::cout << "attributes=\"" << fAttributes << "\"" << "\n";
   }
   if (fCode.Length()) {
      TROOT::IndentLevel();
      std::cout << "code=\"{" << fCode << "}\" "
      << "\n";
   }
   TROOT::DecreaseDirLevel();
}
 
////////////////////////////////////////////////////////////////////////////////
/// Add to the string 'out' the string representation of the rule.
/// if options contains:
///  's' : add the short form of the rule is possible
///  'x' : add the xml form of the rule
 
void TSchemaRule::AsString(TString &out, const char *options) const
{
   TString opt(options);
   opt.ToLower();
   Bool_t shortform = opt.Contains('s');
   Bool_t xmlform = opt.Contains('x');
 
   TString end;
   if (xmlform) {
      /*
       <read sourceClass="ClassA" version="[2]" targetClass="ClassA" source="int m_unit;" target="m_unit" >
       <![CDATA[ { m_unit = 10*onfile.m_unit; } ]]>
       </read>
       */
      shortform = kFALSE;
      out += "<";
      if (fRuleType==kReadRule) { out += "read "; end = "</read>"; }
      else if (fRuleType==kReadRawRule) { out += "readraw "; end = "</readraw>"; }
      else { out += "-- "; end = "-->"; }
 
   } else {
      if (!shortform || fRuleType!=kReadRule) {
         out += "type=";
         if (fRuleType==kReadRule) out += "read ";
         else if (fRuleType==kReadRawRule) out += "readraw ";
         else out += " ";
      }
   }
   if (!shortform || (fSourceClass != fTargetClass) ) {
      out += "sourceClass=\"" + fSourceClass + "\" ";
      out += "targetClass=\"" + fTargetClass + "\" ";
   } else {
      out += fSourceClass + " ";
   }
   if (shortform && fTarget == fSource) {
      out += fSource + " ";
   }
   if (!shortform || (fVersion != "[1-]")) {
      if (fVersion.Length())  out += "version=\""     + fVersion + "\" ";
   }
   if (fChecksum.Length()) out += "checksum=\""    + fChecksum + "\" ";
   if (!shortform || fTarget != fSource) {
      out += "source=\""      + fSource + "\" ";
      out += "target=\""      + fTarget + "\" ";
   }
   if (fInclude.Length())  out += "include=\""     + fInclude + "\" ";
   if (fAttributes.Length()) out += "attributes=\"" + fAttributes + "\" ";
   if (xmlform) {
      out += "> ";
   }
   if (xmlform) {
      if (fCode.Length()) {
         out += "\n<![CDATA[ { " + fCode + " ]]>\n ";
      } else if (fReadFuncPtr) {
         // Can we guess?
         // out += "code=\" + nameof(fReadFuncPtr) + "\" ";
      } else if (fReadRawFuncPtr) {
         // Can we guess?
         // out += "code=\" + nameof(fReadRawFuncPtr) + "\" ";
      }
   } else {
      if (fCode.Length()) {
         out += "code=\"{" + fCode + "}\" ";
      } else if (fReadFuncPtr) {
         // Can we guess?
         // out += "code=\" + nameof(fReadFuncPtr) + "\" ";
      } else if (fReadRawFuncPtr) {
         // Can we guess?
         // out += "code=\" + nameof(fReadRawFuncPtr) + "\" ";
      }
   }
   if (xmlform) {
      out += end;
   }
}
 
////////////////////////////////////////////////////////////////////////////////
/// Zero out this rule object.
 
void TSchemaRule::Clear( const char * /* option */)
{
   fVersion.Clear();
   fChecksum.Clear();
   fSourceClass.Clear();
   fTarget.Clear();
   fSource.Clear();
   fInclude.Clear();
   fCode.Clear();
   fAttributes.Clear();
   fReadRawFuncPtr = nullptr;
   fReadFuncPtr = nullptr;
   fRuleType = kNone;
   delete fVersionVect;   fVersionVect = nullptr;
   delete fChecksumVect;  fChecksumVect = nullptr;
   delete fTargetVect;    fTargetVect = nullptr;
   delete fSourceVect;    fSourceVect = nullptr;
   delete fIncludeVect;   fIncludeVect = nullptr;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Set the content fot this object from the rule
/// See TClass::AddRule for details on the syntax.
 
Bool_t TSchemaRule::SetFromRule( const char *rule )
{
   //-----------------------------------------------------------------------
   // Parse the rule and check it's validity
   /////////////////////////////////////////////////////////////////////////////
 
   ROOT::Internal::MembersMap_t rule_values;
 
   std::string error_string;
   if( !ROOT::ParseRule(rule, rule_values, error_string) ) {
      Error("SetFromRule","The rule (%s) is invalid: %s",rule,error_string.c_str());
      return kFALSE;
   }
   ROOT::Internal::MembersMap_t::const_iterator it1;
 
   it1 = rule_values.find( "type" );
   if( it1 != rule_values.end() ) {
      if (it1->second == "read" || it1->second == "Read") {
         SetRuleType( TSchemaRule::kReadRule );
      } else if (it1->second == "readraw" || it1->second == "ReadRaw") {
         SetRuleType( TSchemaRule::kReadRawRule );
      } else {
         SetRuleType( TSchemaRule::kNone );
      }
   } else {
      // Default to read.
      SetRuleType( TSchemaRule::kReadRule );
   }
   it1 = rule_values.find( "targetClass" );
   if( it1 != rule_values.end() ) SetTargetClass( it1->second );
   it1 = rule_values.find( "sourceClass" );
   if( it1 != rule_values.end() ) SetSourceClass( it1->second );
   it1 = rule_values.find( "target" );
   if( it1 != rule_values.end() ) SetTarget( it1->second );
   it1 = rule_values.find( "source" );
   if( it1 != rule_values.end() ) SetSource( it1->second );
   it1 = rule_values.find( "version" );
   if( it1 != rule_values.end() ) SetVersion( it1->second );
   it1 = rule_values.find( "checksum" );
   if( it1 != rule_values.end() ) SetChecksum( it1->second );
   it1 = rule_values.find( "embed" );
   if( it1 != rule_values.end() ) SetEmbed( it1->second == "false" ? false : true );
   it1 = rule_values.find( "include" );
   if( it1 != rule_values.end() ) SetInclude( it1->second );
   it1 = rule_values.find( "attributes" );
   if( it1 != rule_values.end() ) SetAttributes( it1->second );
   it1 = rule_values.find( "code" );
   if( it1 != rule_values.end() ) SetCode( it1->second );
   // if (code is functioname) {
   // switch (ruleobj->GetRuleType() ) {
   // case kRead: SetReadFunctionPointer(  )
   // case kReadRewa: SetReadRawFunctionPointer( )
   // }
 
   return kTRUE;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Set the version string - returns kFALSE if the format is incorrect
 
Bool_t TSchemaRule::SetVersion( const TString& version )
{
   fVersion = "";
   Bool_t ret = ProcessVersion( version );
   if( ret )
      fVersion = version;
   return ret;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Get the version string.
 
const char *TSchemaRule::GetVersion() const
{
   return fVersion;
}
 
 
////////////////////////////////////////////////////////////////////////////////
/// Check if given version number is defined in this rule
 
Bool_t TSchemaRule::TestVersion( Int_t version ) const
{
   if( fVersion.IsNull() )
      return kFALSE;
 
   if( !fVersionVect )
      ProcessVersion( fVersion ); // At this point the version string should always be correct
 
   if (version == -1) {
      version = 1;
   }
 
   if (fVersionVect)
      for (auto &it : *fVersionVect)
         if( version >= it.first && version <= it.second )
            return kTRUE;
 
   return kFALSE;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Set the checksum string - returns kFALSE if the format is incorrect
 
Bool_t TSchemaRule::SetChecksum( const TString& checksum )
{
   fChecksum = "";
   Bool_t ret = ProcessChecksum( checksum );
   if( ret )
      fChecksum = checksum;
   return ret;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Check if given checksum is defined in this rule
 
Bool_t TSchemaRule::TestChecksum( UInt_t checksum ) const
{
   if( fChecksum.IsNull() )
      return kFALSE;
 
   if( !fChecksumVect )
      ProcessChecksum( fChecksum ); // At this point the checksum string should always be correct
 
   if (fChecksumVect)
      for (auto &it : *fChecksumVect)
         if( checksum == it )
            return kTRUE;
 
   return kFALSE;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Set the source class of this rule (i.e. the onfile class).
 
void TSchemaRule::SetSourceClass( const TString& classname )
{
   std::string normalizedName;
   TClassEdit::GetNormalizedName(normalizedName, classname);
   fSourceClass = normalizedName;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Get the source class of this rule (i.e. the onfile class).
 
const char *TSchemaRule::GetSourceClass() const
{
   return fSourceClass;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Set the target class of this rule (i.e. the in memory class).
 
void TSchemaRule::SetTargetClass( const TString& classname )
{
   std::string normalizedName;
   TClassEdit::GetNormalizedName(normalizedName, classname);
   fTargetClass = normalizedName;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Get the targte class of this rule (i.e. the in memory class).
 
const char *TSchemaRule::GetTargetClass() const
{
   return fTargetClass;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Set the target member of this rule (i.e. the in memory data member).
 
void TSchemaRule::SetTarget( const TString& target )
{
   fTarget = target;
 
   if( target == "" ) {
      delete fTargetVect;
      fTargetVect = nullptr;
      return;
   }
 
   if( !fTargetVect ) {
      fTargetVect = new TObjArray();
      fTargetVect->SetOwner();
   }
   ProcessList( fTargetVect, target );
}
 
////////////////////////////////////////////////////////////////////////////////
/// Get the target data members of this rule as a simple string (i.e. the in memory data member).
 
const char *TSchemaRule::GetTargetString() const
{
   return fTarget;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Get the target data members of this rule (i.e. the in memory data member).
 
const TObjArray*  TSchemaRule::GetTarget() const
{
   if( fTarget == "" )
      return nullptr;
 
   if( !fTargetVect ) {
      fTargetVect = new TObjArray();
      fTargetVect->SetOwner();
      ProcessList( fTargetVect, fTarget );
   }
 
   return fTargetVect;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Set the list of source members.  This should be in the form of a declaration:
///     Int_t fOldMember; TNamed fName;
 
void TSchemaRule::SetSource( const TString& source )
{
   fSource = source;
 
   if( source == "" ) {
      delete fSourceVect;
      fSourceVect = nullptr;
      return;
   }
 
   if( !fSourceVect ) {
      fSourceVect = new TObjArray();
      fSourceVect->SetOwner();
   }
 
   ProcessDeclaration( fSourceVect, source );
}
 
////////////////////////////////////////////////////////////////////////////////
/// Get the list of source members as a TObjArray of TNamed object,
/// with the name being the member name and the title being its type.
 
const TObjArray* TSchemaRule::GetSource() const
{
   if( fSource == "" )
      return nullptr;
 
   if( !fSourceVect ) {
      fSourceVect = new TObjArray();
      fSourceVect->SetOwner();
      ProcessDeclaration( fSourceVect, fSource );
   }
   return fSourceVect;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Set the comma separated list of header files to include to be able
/// to compile this rule.
 
void TSchemaRule::SetInclude( const TString& incl )
{
   fInclude = incl;
 
   if( incl == "" ) {
      delete fIncludeVect;
      fIncludeVect = nullptr;
      return;
   }
 
   if( !fIncludeVect ) {
      fIncludeVect = new TObjArray();
      fIncludeVect->SetOwner();
   }
 
   ProcessList( fIncludeVect, incl );
}
 
////////////////////////////////////////////////////////////////////////////////
/// Return the list of header files to include to be able to
/// compile this rule as a TObjArray of TObjString
 
const TObjArray* TSchemaRule::GetInclude() const
{
   if( fInclude == "" )
      return nullptr;
 
   if( !fIncludeVect ) {
      fIncludeVect = new TObjArray();
      fIncludeVect->SetOwner();
      ProcessList( fIncludeVect, fInclude );
   }
 
   return fIncludeVect;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Set whether this rule should be save in the ROOT file (if true)
 
void TSchemaRule::SetEmbed( Bool_t embed )
{
   fEmbed = embed;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Return true if this rule should be saved in the ROOT File.
 
Bool_t TSchemaRule::GetEmbed() const
{
   return fEmbed;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Return kTRUE if this rule is valid.
 
Bool_t TSchemaRule::IsValid() const
{
   return (fVersionVect || fChecksumVect) && (fSourceClass.Length() != 0);
}
 
////////////////////////////////////////////////////////////////////////////////
/// Set the source code of this rule.
 
void TSchemaRule::SetCode( const TString& code )
{
   fCode = code;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Get the source code of this rule.
 
const char *TSchemaRule::GetCode() const
{
   return fCode;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Set the attributes code of this rule.
 
void TSchemaRule::SetAttributes( const TString& attributes )
{
   fAttributes = attributes;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Get the attributes code of this rule.
 
const char *TSchemaRule::GetAttributes() const
{
   return fAttributes;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Return true if one of the rule's data member target  is 'target'.
 
Bool_t TSchemaRule::HasTarget( const TString& target ) const
{
   if( !fTargetVect )
      return kFALSE;
 
   TObject*      obj;
   TObjArrayIter it( fTargetVect );
   while( (obj = it.Next()) ) {
      TObjString* str = (TObjString*)obj;
      if( str->GetString() == target )
         return kTRUE;
   }
   return kFALSE;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Return true if one of the rule's data member source is 'source'
 
Bool_t TSchemaRule::HasSource( const TString& source ) const
{
   if( !fSourceVect )
      return kFALSE;
 
   TObject*      obj;
   TObjArrayIter it( fSourceVect );
   while( (obj = it.Next()) ) {
      TSources* var = (TSources*)obj;
      if( var->GetName() == source )
         return kTRUE;
   }
   return kFALSE;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Set the pointer to the function to be run for the rule (if it is a read rule).
 
void TSchemaRule::SetReadFunctionPointer( TSchemaRule::ReadFuncPtr_t ptr )
{
   fReadFuncPtr = ptr;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Get the pointer to the function to be run for the rule (if it is a read rule).
 
TSchemaRule::ReadFuncPtr_t TSchemaRule::GetReadFunctionPointer() const
{
   return fReadFuncPtr;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Set the pointer to the function to be run for the rule (if it is a raw read rule).
 
void TSchemaRule::SetReadRawFunctionPointer( TSchemaRule::ReadRawFuncPtr_t ptr )
{
   fReadRawFuncPtr = ptr;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Get the pointer to the function to be run for the rule (if it is a raw read rule).
 
TSchemaRule::ReadRawFuncPtr_t TSchemaRule::GetReadRawFunctionPointer() const
{
   return fReadRawFuncPtr;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Set the type of the rule.
 
void TSchemaRule::SetRuleType( TSchemaRule::RuleType_t type )
{
   fRuleType = type;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Return kTRUE if the rule is a strict renaming of one of the data member of the class.
 
Bool_t TSchemaRule::IsAliasRule() const
{
   return fSourceClass != "" && (fVersion != "" || fChecksum != "") && fTarget == "" && fSource == "" && fInclude == "" && fCode == "" && fAttributes == "";
}
 
////////////////////////////////////////////////////////////////////////////////
/// Return kTRUE if the rule is a strict renaming of the class to a new name.
 
Bool_t TSchemaRule::IsRenameRule() const
{
   return fSourceClass != "" && (fVersion != "" || fChecksum != "") && fTarget != "" && fSource != "" && fInclude == "" && fCode == "" && fAttributes == "";
}
 
////////////////////////////////////////////////////////////////////////////////
/// Return the type of the rule.
 
TSchemaRule::RuleType_t TSchemaRule::GetRuleType() const
{
   return fRuleType;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Check if this rule conflicts with the given one.
 
Bool_t TSchemaRule::Conflicts( const TSchemaRule* rule ) const
{
   //---------------------------------------------------------------------------
   // If the rules have different sources then the don't conflict
   /////////////////////////////////////////////////////////////////////////////
 
   if( fSourceClass != rule->fSourceClass )
      return kFALSE;
 
   //---------------------------------------------------------------------------
   // Check if the rules have common target
   /////////////////////////////////////////////////////////////////////////////
 
   if( !rule->GetTarget() )
      return kFALSE;
 
   Bool_t         haveCommonTargets = kFALSE;
   TObjArrayIter  titer( rule->GetTarget() );
   TObjString    *str;
   TObject       *obj;
 
   while( (obj = titer.Next() ) ) {
      str = (TObjString*)obj;
      if( HasTarget( str->String() ) )
         haveCommonTargets = kTRUE;
   }
 
   if( !haveCommonTargets )
      return kFALSE;
 
   //---------------------------------------------------------------------------
   // Check if there are conflicting checksums
   /////////////////////////////////////////////////////////////////////////////
 
   if( fChecksumVect ) {
      std::vector<UInt_t>::iterator it;
      for( it = fChecksumVect->begin(); it != fChecksumVect->end(); ++it )
         if( rule->TestChecksum( *it ) )
            return kTRUE;
   }
 
   //---------------------------------------------------------------------------
   // Check if there are conflicting versions
   /////////////////////////////////////////////////////////////////////////////
 
   if( fVersionVect && rule->fVersionVect )
   {
      std::vector<std::pair<Int_t, Int_t> >::iterator it1;
      std::vector<std::pair<Int_t, Int_t> >::iterator it2;
      for( it1 = fVersionVect->begin(); it1 != fVersionVect->end(); ++it1 ) {
         for( it2 = rule->fVersionVect->begin();
              it2 != rule->fVersionVect->end(); ++it2 ) {
            //------------------------------------------------------------------
            // the rules conflict it their version ranges intersect
            ////////////////////////////////////////////////////////////////////
 
            if( it1->first >= it2->first && it1->first <= it2->second )
               return kTRUE;
 
            if( it1->first < it2->first && it1->second >= it2->first )
               return kTRUE;
         }
      }
   }
   return kFALSE;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Check if specified version string is correct and build version vector.
 
Bool_t TSchemaRule::ProcessVersion( const TString& version ) const
{
   //---------------------------------------------------------------------------
   // Check if we have valid list
   /////////////////////////////////////////////////////////////////////////////
 
   if( version[0] != '[' || version[version.Length()-1] != ']' )
      return kFALSE;
   std::string ver = version.Data();
 
   std::list<std::string> versions;
   Internal::TSchemaRuleProcessor::SplitList( ver.substr( 1, ver.size()-2), versions );
 
   if( versions.empty() )
   {
      delete fVersionVect;
      fVersionVect = nullptr;
      return kFALSE;
   }
 
   if( !fVersionVect )
      fVersionVect = new std::vector<std::pair<Int_t, Int_t> >;
   fVersionVect->clear();
 
   //---------------------------------------------------------------------------
   // Check the validity of each list element
   /////////////////////////////////////////////////////////////////////////////
 
   std::list<std::string>::iterator it;
   for( it = versions.begin(); it != versions.end(); ++it ) {
      std::pair<Int_t, Int_t> verpair;
      if( !Internal::TSchemaRuleProcessor::ProcessVersion( *it, verpair ) )
      {
         delete fVersionVect;
         fVersionVect = nullptr;
         return kFALSE;
      }
      fVersionVect->push_back( verpair );
   }
   return kTRUE;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Check if specified checksum string is correct and build checksum vector.
 
Bool_t TSchemaRule::ProcessChecksum( const TString& checksum ) const
{
   //---------------------------------------------------------------------------
   // Check if we have valid list
   /////////////////////////////////////////////////////////////////////////////
 
   if (!checksum[0])
      return kFALSE;
   std::string chk = (const char*)checksum;
   if( chk[0] != '[' || chk[chk.size()-1] != ']' )
      return kFALSE;
 
   std::list<std::string> checksums;
   Internal::TSchemaRuleProcessor::SplitList( chk.substr( 1, chk.size()-2), checksums );
 
   if( checksums.empty() ) {
      delete fChecksumVect;
      fChecksumVect = nullptr;
      return kFALSE;
   }
 
   if( !fChecksumVect )
      fChecksumVect = new std::vector<UInt_t>;
   fChecksumVect->clear();
 
   //---------------------------------------------------------------------------
   // Check the validity of each list element
   /////////////////////////////////////////////////////////////////////////////
 
   for( const auto& checksumStr : checksums ) {
      auto chksum = ParseChecksum( checksumStr.c_str() );
      if (chksum == 0u) {
         delete fChecksumVect;
         fChecksumVect = nullptr;
         return kFALSE;
      }
 
      fChecksumVect->push_back( chksum );
   }
   return kTRUE;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Parse the checksum in the given string. Returns either the checksum or zero
/// if the string is not a hex or decimal number.
 
UInt_t TSchemaRule::ParseChecksum(const char* checksum) const {
  std::istringstream converter(checksum);
  UInt_t chksum;
  converter >> std::hex >> chksum;
  if (converter.fail()) {
    converter.clear();
    converter.seekg(0);
    converter >> std::dec >> chksum;
  }
 
  if( converter.fail() ) {
    return 0u;
  }
 
  return chksum;
}
 
////////////////////////////////////////////////////////////////////////////////
/// Split the list as a comma separated list into a TObjArray of TObjString.
 
void TSchemaRule::ProcessList( TObjArray* array, const TString& list )
{
   std::list<std::string>           elems;
   std::list<std::string>::iterator it;
   Internal::TSchemaRuleProcessor::SplitList( (const char*)list, elems );
 
   array->Clear();
 
   if( elems.empty() )
      return;
 
   for( it = elems.begin(); it != elems.end(); ++it ) {
      TObjString *str = new TObjString;
      *str = it->c_str();
      array->Add( str );
   }
}
 
////////////////////////////////////////////////////////////////////////////////
/// Split the list as a declaration into as a TObjArray of TNamed(name,type).
 
void TSchemaRule::ProcessDeclaration( TObjArray* array, const TString& list )
{
   std::list<std::pair<ROOT::Internal::TSchemaType,std::string> >           elems;
   std::list<std::pair<ROOT::Internal::TSchemaType,std::string> >::iterator it;
   Internal::TSchemaRuleProcessor::SplitDeclaration( (const char*)list, elems );
 
   array->Clear();
 
   if( elems.empty() )
      return;
 
   for( it = elems.begin(); it != elems.end(); ++it ) {
      TSources *type = new TSources( it->second.c_str(), it->first.fType.c_str(), it->first.fDimensions.c_str() ) ;
      array->Add( type );
   }
}
 
#if 0
////////////////////////////////////////////////////////////////////////////////
/// Generate the actual function for the rule.
 
Bool_t TSchemaRule::GenerateFor( TStreamerInfo *info )
{
   String funcname = fSourceClass + "_to_" + fTargetClass;
   if (info) funcname += "_v" + info->GetClassVersion();
   TString names = fSource + "_" + fTarget;
   name.ReplaceAll(',','_');
   name.ReplaceAll(':','_');
   funcname += "_" + name;
 
   String filename = funcname + ".C";
   if (!false) {
      filename += '+';
   }
 
   std::ofstream fileout(filename);
 
 
      ROOT::WriteReadRawRuleFunc( *rIt, 0, mappedname, nameTypeMap, fileout );
      ROOT::WriteReadRuleFunc( *rIt, 0, mappedname, nameTypeMap, fileout );
 
   gROOT->LoadMacro(filename);
}
 
#endif