Problem with int cast

Hi there,

I’m not sure if someone already reported this but after some searching I couldn’t find anything like it.

I noticed something strange with the CINT interpreter (this doesn’t happen if I compile the code).

I wanted to truncate a result of a calculation of doubles into an int. Consider the snippet as an example:

Double_t LogTanMin = -4.;
Double_t LogTanW = 0.01; 
Double_t LogTan = -3.900;

Int_t x = (Int_t) ((LogTan-LogTanMin)/LogTanW)

The output of the previous code is:

which is wrong (should be 1).
Moreover, if I assign the result to a Double_t and then cast I keep getting a wrong result, but if instead I cast it to Float_t and then cast it to Int_t then it comes right.

ouble_t LogTanMin = -4.;
Double_t LogTanW = 0.01; 
Double_t LogTan = -3.900;

Float_t xtempf =  ((LogTan-LogTanMin)/LogTanW)
Int_t x = (Int_t) xtempf

the result comes out correct:

It only happens in the interpreter (ROOT 5.22/00) and probably it is a known feature. I could have missed it if it wasn’t for the fact that I started getting the wrong samples from an histogram.

Any ideas?

Thanks in advance,
Nuno

I cannot reproduce your problem. Note that the result should be 10 and not 1 !!
In general one must be very careful when casting the result of a floating point operation to an integer. for example adding 0.5 to the right expression is in general a good idea.

Rene