Silverfrost Forums

Welcome to our forums

Version 8.40 loss of precision

4 Dec 2018 4:05 #22924

In version 8.40 I get a compilation warning message that I was not getting in earlier versions for the following code:

Module m
  Integer, Parameter, Public :: dp = Kind(1.0d0) ! - double precision -
  Integer, Parameter, Public :: sp = Kind(1.0e0) ! - single precision -
#if DP==1
  Integer, Parameter, Public :: rp = dp          ! - default precision -
#else
  Integer, Parameter, Public :: rp = sp          ! - default precision -
#endif
!
  Real(Kind=rp), Parameter, Public :: million = 1000000.0_rp ! - one million -
End Module m

I compile with /CFPP /VPARAM DP 1

The error message is:

A REAL constant has been trucated with possible loss of precision

4 Dec 2018 8:09 #22926

Sorry about this. We are still getting the occasional false error report with this new test for loss of precision. I have fixed this particular failure for the next release after v8.40.

4 Dec 2018 12:13 #22930

Thanks, Paul. Does that mean that the error report is spurious and I can simply ignore it?

4 Dec 2018 1:09 #22931

Yes it does.

Please login to reply.