Verifying Students have Posted to the Newsgroup

A utility that will download posts from a newsgroup, newsarchiver.py, is attached to this document (the Wiki insists on appending .txt; just rename it). It will be wrapped by a program similar to the following, which should then be run frequently (every minute) from a cron job:

#!/bin/bash

export PATH=`/bin/showpath gnu standard`
cd /u/cs135/newsgroup
./newsarchiver.py uw.cs.cs135 >/dev/null
tmpfile=a0_done.$$
realfile=a0_done
grep '^From: ' `grep -l '^References: $' *`\
        | perl -lne 'print $1 if /\<(.*?)@.*uwaterloo\.ca\>/' > $tmpfile
mv $tmpfile $realfile

The portion after the string "^References:" in the grep line must be replaced by the message-ID of the initial newsgroup post, and the course name should be replaced on each line it occurs with the proper course.

Changes that can/should be made and tested include:

• Making the use of a command like whoami to avoid hard-coding in a particular course name in several places so this script can be copied with fewer modifications.
• Requiring this script is in the same directory as newsarchiver.py (which it was anyway), so that the cd line should be able to be changed to the less fragile cd $(dirname $0).

The test.exe program can then take an action similar to the following:

#!/usr/bin/env bash
readonly POSTS="/u/$course/newsgroup/a0_done"
if [ `grep '^'$student'$' $POSTS |  wc -w` -gt 0 ]; then 
        echo '100' >&3
        echo "The appropriate email address was used for a newsgroup post." >&4
else
        echo '0' >&3
        echo "An e-mail from $student's uwaterloo address was not found on the newsgroup." >&4
fi

Please note the importance of wrapping the student userid in ^ and $. This is done to ensure that the userid is actually listed in full, and is not just a substring of another userid, which could lead to false positives.