Float comparison false positive deffect

[Jaroslav Fojtik]

Please note that extent of float type is much more higher than 0xFFFFFFFFFFFFFFFF.
Your false positive is a complete nonsense. The condition is apparently not always true.

CWE: 571
Condition 'ff>0xFFFFFFFFFFFFFFFF' is always true

/** Write float valur to datapool  variant. The variant type is resolved internally. */
void DatapoolVariant::SetAsFloat(float ff)
{    
 if(size>1)
  {
    if(pVoid==NULL) return;
    switch(type)
    {
     case 'Q': if(ff<0)          {*ArrQ=0;break;}
                if(ff>0xFFFFFFFFFFFFFFFF) {*ArrQ=0xFFFFFFFFFFFFFFFF;break;}
                *ArrQ = (UINT64)ff; 
                break;
      case 'q': *Arrq = (INT64)ff; 
                 break;

[CHR]

I can reproduce with this snippet:

void SetAsFloat(float ff, char type)
{    
    switch(type)
    {
     case 'Q': if(ff<0)                  {*ArrQ=0;break;}
               if(ff>0xFFFFFFFFFFFFFFFF) {*ArrQ=0xFFFFFFFFFFFFFFFF;break;}
               break;
    }
}

This seems to be an instance of this problem: https://trac.cppcheck.net/ticket/10150