NEWSPATHS

Section: User Commands (1)
Updated: 5/31/88
Index Return to Main Contents

NAME

newspaths - collect host connectivity information from Usenet articles

SYNOPSIS

newspaths [ prev_out ]

DESCRIPTION

Newspaths scans the history file, in /usr/lib/news/history, using it to locate Usenet articles. For each article, the Path: header is read, and split up into its constituent host pairs. An associative array keeps track of how many times each pair of hosts appears in a Path: header. Every article in the history file is scanned.

The output format consists of the last Message-ID in the history file, followed by a sorted list of host pairs and counts. Here's a sample:

Last-ID: <24524@oliveb.olivetti.com>
 42814 epimass!pyramid
 22610 pyramid!ames
 12824 pyramid!decwrl
  7854 ames!mailrus
  5426 oliveb!ames
  5391 ames!pasteur
  5212 decwrl!ucbvax
< lots more >

If an input file argument is given, it should be the name of a previous output of newspaths. The file and counts are read in, and the history file is scanned until the Last-ID is found. Normal processing begins with the next article given in the history file. The idea is to be able to continue to accumulate the counts over long periods of time (despite expire runs) and to count each article only once. It is an error if the given Message-ID is not found in the history file.

SUGGESTED USE

Every night, from cron, do

        newspaths old-output > new-output
        mv new-output old-output

After 1 month of operation at epimass , the resulting output is 135K bytes long.

AUTHOR

Joseph T. Buck, jbuck@epimass.epi.com

This document was created by man2html, using the manual pages.
Time: 21:54:10 GMT, February 02, 2023