#!/bin/bash
###############################################################################
#  check_url 2002-jan-16 - gash - http://www.moosenugget.org                  #
#                                                                             #
#                                                                             #
#  simple bash2 script that incorporates curl / md5sum to notify the user     # 
#  when a url has changed                                                     #
#                                                                             #
#  curl   http://curl.haxx.se                                                 #
#  md5sum http://www.gnu.org/directory/textutils.html                         #
#                                                                             #
#  command line options:                                                      #
#    -s ... exit after one sweep                                              #
###############################################################################

clear

export PATH="/usr/bin:/bin:/usr/local/bin:/usr/sbin:/sbin:/usr/local/sbin"

### where to store the data files
DATADIR="/tmp"  # note: no trailing /

### this file is used to log the pages that have been updated, 
### in case the user is afk...
updatelog="$DATADIR/check_url.log"

debug="0"

curltimeout=30

pages="http://www.whitehouse.gov http://www.house.gov"

### how long to sleep between sweeps (seconds)
sleeptime=7200

### hey look, i'm theme-able!! :)
spinners=". o 0 O"
#spinners="| / - \\"

function say {
  date=`date '+%Y-%b-%d %T'`
  echo -e "[${date}]:\n  $@\n"
}

bash_ver=`echo $BASH_VERSION | awk -v FS='.' '{print $1}'`

if [ "$bash_ver" = "1" ] ; then
  say "This script requires BASH 2.0 or greater:\n  http://www.gnu.org/directory/bash.html"
  exit
fi

required_proggies="curl md5sum echo sed cat diff sleep rm"

for a in $required_proggies ; do
  which $a 1> /dev/null 2> /dev/null
  if [ $? = 1 ] ; then say "required program not found: $a." ; exit 1 ; fi
done

echo "Monitoring URLs:"
for page in $pages ; do
  echo "  o $page"
done

echo ""

while [ 1 ] ; do
  for page in $pages ; do
    pagefile=`echo $page | sed 's/\///g'`
    pagefile="$DATADIR/$pagefile"

    if [ "$debug" = "1" ] ; then echo "Converted: $pagefile" ; fi
 
    if [ -f "$pagefile" ] ; then
      if [ "$debug" = "1" ] ; then echo -n "Checking: $page..." ; fi
  
      curl --connect-timeout $curltimeout -o ${pagefile}.new $page 1> /dev/null 2> /dev/null
  
      if  [ $? = 0 ] ; then
        ### this keeps md5sum from creating a filename within the md5sum file
        cat ${pagefile}.new | md5sum > ${pagefile}.new.md5
    
        diff -u ${pagefile}.md5 ${pagefile}.new.md5 1> /dev/null 2> /dev/null
     
        if [ $? = 0 ] ; then
          rm ${pagefile}.new ${pagefile}.new.md5
          if [ "$debug" = "1" ] ; then echo "unchanged." ; fi
        else
          ### log the update
          date=`date '+%Y-%b-%d %T'`
          echo -e "[${date}]:\n  $page has changed\n" >> $updatelog
          diff -u $pagefile ${pagefile}.new >> $updatelog
          echo -e "--------------------------------------------------------------------------------\n" >> $updatelog
          ### notify the user
          say "\a\a\a$page has changed!\n  (see $updatelog for details)"
          mv ${pagefile}.new $pagefile
          mv ${pagefile}.new.md5 ${pagefile}.md5
        fi
      else
        say "Error ($?): $page retreival failed."
      fi
    else
      date=`date '+%Y-%b-%d %T'`
      echo -e -n "[${date}]:\n  $page doesn't exist locally, getting..."
    
      curl --connect-timeout $curltimeout -o $pagefile $page 1> /dev/null 2> /dev/null
    
      if  [ $? = 0 ] ; then
        echo -e "done.\n"
    
        ### this keeps md5sum from creating a filename within the md5sum file
        cat $pagefile | md5sum > ${pagefile}.md5
      else
        date=`date '+%Y-%b-%d %T'`
        echo -e "\n\n[${date}]:\n  Error ($?): $page failed, unable to create local copy.\n"
      fi
    fi
  done

  ### exit after one sweep
  if [ "$1" = "-s" -o "$1" = "-S" ] ; then
    echo -e "Exiting.\n"
    exit
  fi

  ### do we have a terminal?
  if [ -t 1 ]; then
    echo -n "Sleeping..."
    count=1
    while [ $count -lt $sleeptime ] ; do
      for spinner in $spinners ; do
        echo -n "$spinner"
        sleep 1s
        echo -e -n "\b"
      done
      # count += # spinners
      let count=$count+4
    done
      echo -e -n "\r            \r"
  else
    sleep ${sleeptime}s
  fi
done

# EOF