Giorgos Keramidas wrote: >On 2006-08-11 20:03, Julian Elischer <julian_at_elischer.org> wrote: > > >>At various times I've wanted to add timestamps to logfiles as they are >>generated.. >> >>usually this has involved perl or something to do it. >> >> > >Indeed, I use a tiny Perl script to do this: > > 1 #!/usr/bin/perl -wT > 2 # > 3 # A utility to timestamp logs. Useful for piping the output of long > 4 # running commands through it, before saving to a file. > 5 # > 6 # $Id: tslog,v 1.4 2006/03/30 15:00:49 giorgos Exp $ > 7 > 8 use strict; > 9 use POSIX qw(strftime); > 10 use Time::HiRes qw( gettimeofday ); > 11 > 12 my ($line, $microseconds, $now, $seconds, $tag); > 13 > 14 if ($#ARGV == 0) { > 15 $tag = $ARGV[0] . " --- "; > 16 } else { > 17 $tag = ""; > 18 } > 19 > 20 $| = 1; > 21 while (defined($line = <STDIN>)) { > 22 chomp $line; > 23 ($seconds, $microseconds) = gettimeofday; > 24 $now = strftime "%Y-%m-%d %H:%M:%S", gmtime($seconds); > 25 printf "%s%s.%06d - %s\n", $tag, $now, $microseconds, $line; > 26 } > >The $tag part is there so you can do stuff like: > >$ ls -l | tslog FOO | head >FOO --- 2006-08-12 03:12:17.954283 - total 386 >FOO --- 2006-08-12 03:12:17.954784 - -rw-rw-r-- 1 giorgos giorgos - 1 Aug 9 20:10 #kot# >FOO --- 2006-08-12 03:12:17.955014 - drwxrwxr-x 2 giorgos giorgos - 512 Aug 8 13:27 CVS >FOO --- 2006-08-12 03:12:17.955378 - drwxrwxr-x 5 giorgos giorgos - 512 Oct 23 2004 GNUstep >FOO --- 2006-08-12 03:12:17.957108 - -rw-rw-r-- 1 giorgos giorgos - 6945 Jul 10 15:59 GOTHMOG >FOO --- 2006-08-12 03:12:17.957475 - drwx------ 4 giorgos giorgos - 1536 Aug 12 06:10 Mail >FOO --- 2006-08-12 03:12:17.959169 - -rw------- 1 giorgos giorgos - 226927 Aug 12 05:29 Mailbox >FOO --- 2006-08-12 03:12:17.959528 - -rw------- 1 giorgos giorgos - 4553 May 23 19:22 NOTES >FOO --- 2006-08-12 03:12:17.961219 - drwxr-xr-x 5 giorgos giorgos - 512 Jun 22 16:13 News >FOO --- 2006-08-12 03:12:17.961576 - drwxrwxr-x 2 giorgos giorgos - 512 May 30 14:30 RCS >$ > >which includes the "FOO --- " prefix at the beginning of default output: > >$ ls -l | tslog | head >2006-08-12 03:12:26.670881 - total 386 >2006-08-12 03:12:26.671382 - -rw-rw-r-- 1 giorgos giorgos - 1 Aug 9 20:10 #kot# >2006-08-12 03:12:26.671608 - drwxrwxr-x 2 giorgos giorgos - 512 Aug 8 13:27 CVS >2006-08-12 03:12:26.673455 - drwxrwxr-x 5 giorgos giorgos - 512 Oct 23 2004 GNUstep >2006-08-12 03:12:26.673926 - -rw-rw-r-- 1 giorgos giorgos - 6945 Jul 10 15:59 GOTHMOG >2006-08-12 03:12:26.675527 - drwx------ 4 giorgos giorgos - 1536 Aug 12 06:10 Mail >2006-08-12 03:12:26.675963 - -rw------- 1 giorgos giorgos - 226927 Aug 12 05:29 Mailbox >2006-08-12 03:12:26.677567 - -rw------- 1 giorgos giorgos - 4553 May 23 19:22 NOTES >2006-08-12 03:12:26.678002 - drwxr-xr-x 5 giorgos giorgos - 512 Jun 22 16:13 News >2006-08-12 03:12:26.679604 - drwxrwxr-x 2 giorgos giorgos - 512 May 30 14:30 RCS >$ > > > >>finally I broke down and just added a small bit to date(1) >> >>the -s option tells date to add a timestamp on the front of every line >>read in through stdin >> >> > >I'm not suggesting that this is not a worthwhile addition, but it does >have some minor glitches that a Perl version doesn't need to work >around. The following part, for example, is a bit buggy unless I'm not >reading it correctly: > > > >>+ if (sflag) { >>+ char linebuf[2048]; >>+ time_t otval = 0; >>+ >>+ while (fgets(linebuf, 2048, stdin)) { >>+ if (!rflag && time(&tval) == -1) >>+ err(1, "time"); >>+ >>+ if (tval != otval) { >>+ lt = *localtime(&tval); >>+ badv = vary_apply(v, <); >>+ if (badv) { >>+ fprintf(stderr, "%s: Cannot apply date adjustment\n", >>+ badv->arg); >>+ vary_destroy(v); >>+ usage(); >>+ } >>+ (void)strftime(buf, sizeof(buf), format, <); >>+ otval = tval; >>+ } >>+ (void)printf("%s", buf); >>+ fputs(linebuf, stdout); >>+ if (fflush(stdout)) { >>+ err(1, "stdout"); >>+ } >>+ } >> >> > >If you receive lines with a length greater than 2048 - strlen(length of >timestamp) there is a case that this will print output like: > > Fri Aug 11 19:53:34 PDT 2006: ...lots of stuff here...Fri Aug 11 19:53:34 PDT 2006: ... > > > Yes I said I hacked it in :-) In my app you will never have such long lines.. basically you need something that reads lines and tells you how much it read.. (I have no idea WHY fgets need sto return the START.. you already KNOW that!) it'd be nice if you didn't have to to a strlen() on each line. >To make sure that lines are not concatenated too early and still check >that linebuf[] is not overflow, we may have to use something like the >following (untested) patch: > > ... > > if (sflag) { > char linebuf[sizeof(buf)]; > time_t otval = 0; > int complete = 0; > size_t len; > > while (fgets(linebuf, 2048, stdin) != NULL) { > if (rlfag == 0 && time(&tval) == -1) > err(1, "time"); > > if (complete != 0 && tval != otval) { > lt = *localtime(&tval); > badv = vary_apply(v, <); > if (badv) { > fprintf(stderr, "%s: Cannot apply date adjustment\n", > badv->arg); > vary_destroy(v); > usage(); > } > (void)strftime(buf, sizeof(buf), format, <); > otval = tval; > } > if (complete != 0) { > (void)printf("%s", buf); > complete = 0; > } > fputs(linebuf, stdout); > if (fflush(stdout)) > err(1, "stdout"); > > /* > * If we just printed a newline, make sure we > * print a new timestamp when we loop back for > * the next batch of input characters. > */ > len = strlen(linebuf); > if (linebuf[len - 1] == '\n') > complete = 1; > } > } else ... > >Received on Sat Aug 12 2006 - 01:54:46 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:38:58 UTC