Provided by: libpcp4-dev_7.0.2-1_amd64 bug

NAME

       pmtimespecNow,    pmtimespecInc,    pmtimespecDec,    pmtimespecAdd,   pmtimespecSub,   pmtimespecToReal,
       pmtimespecFromReal, pmtimespecTotimeval, pmtimespecPrint, pmtimespecPrintInterval - helper  routines  for
       time stored as a struct timespec

C SYNOPSIS

       #include <pcp/pmapi.h>

       void pmtimespecNow(struct timespec *tp);
       void pmtimespecInc(struct timespec *ap, const struct timespec *bp);
       void pmtimespecDec(struct timespec *ap, const struct timespec *bp);
       double pmtimespecAdd(const struct timespec *ap, const struct timespec *bp);
       double pmtimespecSub(const struct timespec *ap, const struct timespec *bp);
       double pmtimespecToReal(const struct timespec *tp);
       void pmtimespecFromReal(double secs, struct timespec *tp);
       void pmtimespecTotimeval(struct timespec *tsp, struct timeval *tvp);
       void pmtimespecPrint(FILE *f, const struct timespec *tp);
       void pmtimespecPrintInterval(FILE *f, const struct timespec *tp);

       cc ... -lpcp

DESCRIPTION

       pmtimespecNow  is  a  platform-independent method that returns the current system time since the Epoch in
       tp.

       pmtimespecInc adds the time in ap to the time in bp and stores the result in ap.  Similarly pmtimespecDec
       subtracts the time in bp from the time in ap and stores the result in ap.

       pmtimespecAdd (and pmtimespecSub) add (and subtract) times and return the result as a double value.   The
       time in the ap argument is not changed.

       pmtimespecToReal  converts the time in tp to an equivalent double value.  pmtimespecFromReal provides the
       reverse conversion, taking the time in secs and returning the equivalent time in tp.

       pmtimespecTotimeval converts the struct timespec tsp  into  an  equivalent  struct  timeval  tvp  (modulo
       nanosecond to microsecond truncation).

       For  pmtimespecAdd, pmtimespecSub, pmtimespecToReal and pmtimespecFromReal there is an issue of numerical
       precision.  If the double result or operand has more than 6 decimal digits of precision  in  the  seconds
       part of the value, then the fractional value is converted to microseconds before arithmetic computations,
       so the nanoseconds field is only accurate to 6 decimal digits in this case.

       pmtimespecPrint  prints  the timestamp from tp on the stream f in the local time (as returned by pmLocal‐
       time(3)) in the format HH:MM:SS.XXXXXXXXX.

       pmtimespecPrintInterval  prints  the  interval  from  tp  on  the  stream  f  in  a   minimalist   format
       [[HHh]MMm]SS[.XXXXXXXXX]s  where the hours, minutes and nanoseconds are only reported if they are non-ze‐
       ro, and any trailing zeroes in the nanoseconds are  omitted.   For  example  if  tp->tv_sec  was  75  and
       tp->tv_nsec was 51000000 then the output would be 1h15.51s.

NOTES

       For  variants  of  these  interfaces  using microsecond precision and the timeval structure, refer to pm‐
       timevalNow(3).

COMPATIBILITY

       Prior to PCP 7.0 the routine pmtimespecPrint was called pmPrintHighResStamp.  To  support  PMAPI  transi‐
       tion, the old routine can still be used if applications are recompiled with -DPMAPI_VERSION=2.

       pmPrintHighResStamp is now deprecated and compile-time support for this routine name will be removed in a
       future release.

SEE ALSO

       PMAPI(3), pmLocaltime(3) and pmtimevalNow(3).

Performance Co-Pilot                                   PCP                                      PMTIMESPECNOW(3)