trpfpe(3F)                                             trpfpe(3F)



NAME
     trpfpe, fpecnt - trap and repair floating point faults

SYNOPSIS
     subroutine trpfpe (numesg, rtnval)
     double precision rtnval

     integer function fpecnt ()

     common /fpeflt/ fperr
     logical fperr

DESCRIPTION
     trpfpe sets up a signal handler to trap arithmetic excep-
     tions. If the exception is due to a floating point arith-
     metic fault, the result of the operation is replaced with
     the rtnval specified.  rtnval must be a double precision
     value. For example, 0.0d0 or the return value of dflmax().

     The first numesg occurrences of a floating point arithmetic
     error will cause a message to be written to the standard
     error file.  Any exception that cannot be repaired will
     result in the default action, typically an abort with core
     image.

     fpecnt returns the number of faults since the last call to
     trpfpe.

     The logical value in the common block labeled fpeflt will be
     set to .TRUE. each time a fault occurs.

FILES
     /usr/lib/libF77.a

SEE ALSO
     signal(3F)

NOTES
     When compiling source with the -Qocs option, an explicit
     underscore needs to be added to the end of the routine
     names.

     When compiling source with the -Qocs option, two explicit
     underscores need to be added to the end of the name of the
     common block fpeflt.

BUGS
     This handler just prints a message.  There are no arithmetic
     faults (only traps) around, so that no instruction is inter-
     rupted before it completes, and the program counter points
     to the next floating point instruction (not necessarily next
     executable instruction after the one that got the



Page 1                        CX/UX Hf77 Fortran Reference Manual

trpfpe(3F)                                             trpfpe(3F)



     exception).

     If the operation involves changing the stack pointer, it
     cannot be repaired.  This seldom should be a problem with
     the f77 compiler, but such an operation might be produced by
     the optimizer.

















































Page 2                        CX/UX Hf77 Fortran Reference Manual