qmailanalog wants seconds and fractional seconds (since start of epoch, I
suppose, although I don't think it actually matters since I believe all of
its calculations are relative). Dan, were you planning on releasing a new
version of qmailanalog sometime at or after qmail 2.0? Is the logging
format going to change sufficiently that I shouldn't put a lot of time
into approving log analysis for 1.03?
Attached is a small program based on tai64nlocal that will convert from
tai64n format to what qmailanalog expects.
--
Russ Allbery ([EMAIL PROTECTED]) <URL:http://www.eyrie.org/~eagle/>
/* $Id$
Convert external TAI64N timestamps to fractional seconds since epoch.
Written by Russ Allbery <[EMAIL PROTECTED]>
This work is in the public domain.
Usage:
tai64nfrac < input > output
Expects the input stream to be a sequence of lines beginning with @, a
timestamp in external TAI64N format, and a space. Replaces the @ and the
timestamp with fractional seconds since epoch (1970-01-01 00:00:00 UTC).
The input time format is the format written by tai64n and multilog. The
output time format is expected by qmailanalog. */
#include <stdio.h>
/* Read a TAI64N external format timestamp from stdin and write fractional
seconds since epoch (TAI, not UTC) to stdout. Return the character after
the timestamp. */
int
decode(void)
{
int c;
unsigned long u;
unsigned long seconds = 0;
unsigned long nanoseconds = 0;
while ((c = getchar()) != EOF) {
u = c - '0';
if (u >= 10) {
u = c - 'a';
if (u >= 6) break;
u += 10;
}
seconds <<= 4;
seconds += nanoseconds >> 28;
nanoseconds &= 0xfffffff;
nanoseconds <<= 4;
nanoseconds += u;
}
seconds -= 4611686018427387914ULL;
printf("%lu.%lu ", seconds, nanoseconds);
return c;
}
int
main(void)
{
int c;
unsigned long seconds;
unsigned long nanoseconds;
while ((c = getchar()) != EOF) {
if (c == '@') c = decode();
while (c != EOF) {
putchar(c);
if (c == '\n') break;
c = getchar();
}
}
}
