Can anyone tell me if it is possible to get the AIX "date" command to output
the
current time down to milli-seconds. A review of the man pages shows how to
output time only as accurately as seconds.

I need this to log the start time of a regularly called script. This script
functions
properly only when ran as single instance. I had problems with a remote host
calling this twice within a second of each other. Some kind of scheduler
hiccup
on the calling side.

I have implemented a lock file locking mechanism to stop this happening in
the
future, but I would like to be able to log the start time of my script down
to
milli-seconds for logging purposes, and to aid in finding out why the remote
host (an Oracle ERP host) experiences a double call sometimes.

If the "date" command is not this accurate, can anyone suggest alternatives?
Thanks in advance!

Tom.

Hello, TOM,

have you tried the cycle in the script
once you have started it it will continue run after the cycle perioud


with the best
BRIEDIS

"Tom Brehony" <tombrehony.nospam@eircom.nospam.net> wrote in message news:<2m4a8lFihmcrU1@uni-berlin.de>...
> Can anyone tell me if it is possible to get the AIX "date" command to output
> the
> current time down to milli-seconds. A review of the man pages shows how to
> output time only as accurately as seconds.
>
> I need this to log the start time of a regularly called script. This script
> functions
> properly only when ran as single instance. I had problems with a remote host
> calling this twice within a second of each other. Some kind of scheduler
> hiccup
> on the calling side.
>
> I have implemented a lock file locking mechanism to stop this happening in
> the
> future, but I would like to be able to log the start time of my script down
> to
> milli-seconds for logging purposes, and to aid in finding out why the remote
> host (an Oracle ERP host) experiences a double call sometimes.
>
> If the "date" command is not this accurate, can anyone suggest alternatives?
> Thanks in advance!
>
> Tom.

Janis,

I am sorry but I do not understand what you are saying here.
What do you mean "the cycle in the script"?

Tom.

"janis.briedis" <janis.briedis@tl.lv> wrote in message
news:b06898f5.0407200640.14d623a7@posting.google.c om...
> Hello, TOM,
>
> have you tried the cycle in the script
> once you have started it it will continue run after the cycle perioud
>
>
> with the best
> BRIEDIS
>
> "Tom Brehony" <tombrehony.nospam@eircom.nospam.net> wrote in message
news:<2m4a8lFihmcrU1@uni-berlin.de>...
> > Can anyone tell me if it is possible to get the AIX "date" command to
output
> > the
> > current time down to milli-seconds. A review of the man pages shows how
to
> > output time only as accurately as seconds.
> >
> > I need this to log the start time of a regularly called script. This
script
> > functions
> > properly only when ran as single instance. I had problems with a remote
host
> > calling this twice within a second of each other. Some kind of scheduler
> > hiccup
> > on the calling side.
> >
> > I have implemented a lock file locking mechanism to stop this happening
in
> > the
> > future, but I would like to be able to log the start time of my script
down
> > to
> > milli-seconds for logging purposes, and to aid in finding out why the
remote
> > host (an Oracle ERP host) experiences a double call sometimes.
> >
> > If the "date" command is not this accurate, can anyone suggest
alternatives?
> > Thanks in advance!
> >
> > Tom.

"Tom Brehony" <tombrehony.nospam@eircom.nospam.net> writes:

>Can anyone tell me if it is possible to get the AIX "date" command to output
>the current time down to milli-seconds. A review of the man pages shows
>how to output time only as accurately as seconds.

I believe the answer is "no". However, here is a quick and dirty C program
you can use for that purpose:

#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>
#include <string.h>
#include <time.h>
#include <sys/time.h>

int main(int argc, char **argv)
{
struct timeval tv;
struct tm *tm;
char *ofmt;
char *nfmt;
char *osp, *nsp;
char buf[1024];

if (argc != 2) {
fprintf(stderr, "Usage: %s format\n", argv[0]);
exit(1);
}
ofmt = argv[1];

if (gettimeofday(&tv, NULL) < 0 ) {
perror("gettimeofday");
exit(1);
}
tm = localtime(&tv.tv_sec);
nfmt = malloc(strlen(ofmt) * 6 / 2 + 100);
if (nfmt == NULL) {
perror("malloc new fmt");
exit(2);
}
/*
* Locate %s sequences and replace with microseconds
*/
nfmt[0] = '\0';
nsp = nfmt;
while ((osp = strstr(ofmt, "%s")) != NULL) {
size_t l;
l = osp - ofmt;
strncpy(nsp, ofmt, l);
sprintf(nsp+l, "%06d", tv.tv_usec );
ofmt = osp+2;
nsp += strlen(nsp);
}
strcpy(nsp, ofmt);
if (strftime(buf, sizeof(buf), nfmt, tm) == 0) {
perror("strftime");
exit(2);
}
puts(buf);

return 0;
}

Call it with a single argument, which is a format as accepted by
strftime(), augmented by the formst %s, which gives you 6 digits
of microseconds. Thus, to show the current date and time with
microseconds, use

./a.out %Y-%m-%d,%H:%M:%S.%s
2004-07-20,13:31:41.788645

>I have implemented a lock file locking mechanism to stop this happening in
>the future, but I would like to be able to log the start time of my script
>down to milli-seconds for logging purposes, and to aid in finding out why
>the remote host (an Oracle ERP host) experiences a double call sometimes.

Is it possible you are seeing the previous instance of the call getting
hung up, until the next instance somehow shakes it loose, and the two
then run at the same time?

--
Dale Talcott, IT Research Computing Services, Purdue University
aeh@quest.cc.purdue.edu http://quest.cc.purdue.edu/~aeh/

Excellent Dale, it's perfect!

Many thanks.

Tom.

"Dale Talcott" <aeh@quest.cc.purdue.edu> wrote in message
news:aeh.1090348515@quest.cc.purdue.edu...
> "Tom Brehony" <tombrehony.nospam@eircom.nospam.net> writes:
>
> >Can anyone tell me if it is possible to get the AIX "date" command to
output
> >the current time down to milli-seconds. A review of the man pages shows
> >how to output time only as accurately as seconds.
>
> I believe the answer is "no". However, here is a quick and dirty C
program
> you can use for that purpose:
>
> #include <stdio.h>
> #include <stdlib.h>
> #include <unistd.h>
> #include <string.h>
> #include <time.h>
> #include <sys/time.h>
>
> int main(int argc, char **argv)
> {
> struct timeval tv;
> struct tm *tm;
> char *ofmt;
> char *nfmt;
> char *osp, *nsp;
> char buf[1024];
>
> if (argc != 2) {
> fprintf(stderr, "Usage: %s format\n", argv[0]);
> exit(1);
> }
> ofmt = argv[1];
>
> if (gettimeofday(&tv, NULL) < 0 ) {
> perror("gettimeofday");
> exit(1);
> }
> tm = localtime(&tv.tv_sec);
> nfmt = malloc(strlen(ofmt) * 6 / 2 + 100);
> if (nfmt == NULL) {
> perror("malloc new fmt");
> exit(2);
> }
> /*
> * Locate %s sequences and replace with microseconds
> */
> nfmt[0] = '\0';
> nsp = nfmt;
> while ((osp = strstr(ofmt, "%s")) != NULL) {
> size_t l;
> l = osp - ofmt;
> strncpy(nsp, ofmt, l);
> sprintf(nsp+l, "%06d", tv.tv_usec );
> ofmt = osp+2;
> nsp += strlen(nsp);
> }
> strcpy(nsp, ofmt);
> if (strftime(buf, sizeof(buf), nfmt, tm) == 0) {
> perror("strftime");
> exit(2);
> }
> puts(buf);
>
> return 0;
> }
>
> Call it with a single argument, which is a format as accepted by
> strftime(), augmented by the formst %s, which gives you 6 digits
> of microseconds. Thus, to show the current date and time with
> microseconds, use
>
> ./a.out %Y-%m-%d,%H:%M:%S.%s
> 2004-07-20,13:31:41.788645
>
> >I have implemented a lock file locking mechanism to stop this happening
in
> >the future, but I would like to be able to log the start time of my
script
> >down to milli-seconds for logging purposes, and to aid in finding out why
> >the remote host (an Oracle ERP host) experiences a double call sometimes.
>
> Is it possible you are seeing the previous instance of the call getting
> hung up, until the next instance somehow shakes it loose, and the two
> then run at the same time?
>
> --
> Dale Talcott, IT Research Computing Services, Purdue University
> aeh@quest.cc.purdue.edu http://quest.cc.purdue.edu/~aeh/