From: "Jeremy L. Buchmann" <jlbuchma AT pema DOT scs DOT unr DOT edu>
Newsgroups: comp.os.msdos.djgpp
Subject: Re: time and milliseconds
Date: 4 Jan 1999 06:48:49 GMT
Organization: University & Community College System of Nevada
Lines: 46
Message-ID: <76po8h$mo$1@pema.scs.unr.edu>
References: <76kg9d$3tl$1 AT nnrp1 DOT dejanews DOT com>
NNTP-Posting-Host: empire.scs.unr.edu
NNTP-Posting-User: unauthenticated_user
User-Agent: tin/pre-1.4-981114 ("The Watchman") (UNIX) (OSF1/V4.0 (alpha))
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp
Reply-To: djgpp AT delorie DOT com

leonj AT geocities DOT com wrote:
: Is there a way to retreive milliseconds from the time function?
: Is there a source clip to show this...?

The ftime() function is probably the most portable way of doing this.  I
have used it in Unix and it is listed in the DJGPP libc.a reference, so I
assume it works there, too.  It looks like this:

#include <sys/timeb.h>

int ftime(struct timeb *buf)

It stores the current time in the timeb structure buf.  
The format of struct timeb is:

struct timeb {
  time_t time; //This is your normal time() structure
  unsigned short millitm //milliseconds
  ... //some other stuff
}


Here is how I used it in a Unix program (compiled on Solaris if it
matters):

int SystemTime::click()
{
  timeb *tp = new timeb;
  if (ftime(tp) == 0) //if ftime() returns okay...
  {
    milliseconds = tp->millitm; //extract milliseconds here
    time_t *timecalptr = &tp->time;
    tm* tmptr = localtime(timecalptr);
    seconds = tmptr->tm_sec;
    minutes = tmptr->tm_min;
  }
}     

Like another poster said, the "granularity"  of system clocks garantees
that millisecond values won't be very accurate, but it's better than
nothing.

-------------------------------------------------------------------
Jeremy Buchmann       "Those who trade freedom for safety deserve
jlbuchma AT scs DOT unr DOT edu   neither freedom nor safety." -- Ben Franklin
-------------------------------------------------------------------