Thanks Andrew. Your suggestion is pretty much what I was looking for - and
you're the second person to hit me with this "apachectl restart" business.
Sigh! So many commands, so little time!
- fleet -
At 18:55 11/9/99 -0800, you wrote:
>--- [EMAIL PROTECTED] wrote:
>I looked at the file scripts at both rebol.org and rebol.com without
>finding any similar routine. I assume I "load" the log file; but I have no
>idea how to get rid of the lines I don't want.
>
> - fleet -
>--- end of quote ---
>You'd probably want to make your script in several pieces. The first
would use read/lines to read the apache logfile (which has 1 record/line)
and save back to the log file the data without previous days' entries.
Depending on what you're trying to do, you might just delete the orig log
file if you're doing this every day.
>Here's the hard part: after changing the log file (which presumably takes
so little time that at most only an access or two are lost) you need to do
an apachectl restart. This would be something for rebol/command. Or, if
you can stand to lose a minute of logging/day, you might make a cron entry
1 minute after calling your rebol script restarts apache.
>
>In this situation you would have 2 cron entries, maybe a minute apart and
a script like the following:
>
>REBOL []
>today: make block! []
>notoday: make block! []
>foreach entry read/lines %logfile [
> either find entry now/date [
> append today join to-string entry {^/}
> ][
> insert notoday entry
> ]
>]
>write %logfile today
>;this should take very little time. around here is when your other cron
entry should restart apache or rebol/command could do it.
>
>; here you can parse/sort/save/whatever the entries from previous days.
>
>
>Something of this form (though not this exact code. now/date may not be
in the right format, for example) should do it.
>
>--
>Andrew Grossman
>
>
