Below is a snip of my responce on the issue from the Message duplication
problem thread.
I don't know which of the two scripts is the faster one.
I would be interested in finding out how the scripts compare for small,
medium, and large subscriber files on the time each take to be sorted.
<snip>
The only argument is the listname subscriber file.
#!/usr/bin/perl
$list = $ARGV[0];
system("mv $list $list.bak ; cat $list.bak | tr A-Z a-z | sort -t@ -k2 >
$list ; chmod 660 $list ; chown majordomo:majordomo $list\n");
exit(0);
Note the 4th line is a continuation of the system command on the 3rd line.
</snip>
At 09:36 AM 6/21/01 -0500, you wrote:
>The recent discussions on how to improve majordomo/sendmail performance by
>sorting the list files by domain prompted me to try to do this for my own
>lists. Finding no script to do this already written, I put together the
>following perl script to do the sorting. The directory the list files are in
>is defined in the chdir function in the script. The script assumes the list
>files are the only files in the directory _without_ a period in the name.
>
>#! /usr/bin/perl
>
>chdir "/var/lib/majordomo/lists/testdir";
>@listnames = `ls`;
>@lists = grep (!/\./, @listnames);
>unlink 'lists.out';
>
>for my $list (@lists)
> {
> chomp $list;
> open LISTFILE, $list;
> print $list."\n";
> @fileaddresses = ();
>
> while (<LISTFILE>)
> {
> chomp;
> ($account, $host) = split (/\@/, $_);
> push @fileaddresses, lc($host)."@".lc($account)."\n";
> }
>
> @sortedaddresses = sort @fileaddresses;
> open OUTFILE, ">>sortedlist.out";
>
> for my $sortedaddress (@sortedaddresses)
> {
> chomp $sortedaddress;
> ($sortedhost, $sorteddest) = split (/\@/, $sortedaddress);
> print OUTFILE $sorteddest."@".$sortedhost."\n";
> }
> close OUTFILE;
>
> `mv sortedlist.out $list`;
> unlink "sortedlist.out";
> }
>
>All comments/suggested improvements are welcome.
>
>Bill Anderson
>MIS Network Administrator
>MO Dept of Economic Development
>
>
>
Follow-Ups:
|
|
