Extracts XML from application log to a file.

getxml.sh

#!/bin/sh
[ $1 ] || { echo "Usage: getxml <redacted-app-logfile>" ; exit 1 ; }
[ -f $1 ] || { ls $1 ; exit 1 ; }

eventlog=$1

# determine input type and output file name
string0="LoggerMessageProcessor - Outbound message before validation: "
string1="LoggerMessageProcessor - CustomerProfileEvent Message: "
baseline_events=`zgrep -m 1 "$string0" $eventlog|wc -l|awk '{print $1}'`
parallel_events=`zgrep -m 1 "$string1" $eventlog|wc -l|awk '{print $1}'`
tag=""
[ $baseline_events -gt 0 ] && { matcher="$string0" ; tag="base" ; }
[ $parallel_events -gt 0 ] && { matcher="$string1" ; tag="test" ; }
[ "$tag" == "" ] && { echo "nothing to extract -- exiting" ; exit 1 ; }
outputfile="$eventlog.$tag.$$.xml"

# namespace string to replace if the test application munges it
# nsa=baseline, nsb=backwards version
nsa='xmlns:ns2="http://types.mycompany.com/BusinessEventMessage/Customer/CustomerProfileEvent/1/0" xmlns:ns3="http://types.mycompany.com/BusinessEventMessage/Common/BEMMetaTypes/1/0"'
nsb='xmlns:ns3="http://types.mycompany.com/BusinessEventMessage/Common/BEMMetaTypes/1/0" xmlns:ns2="http://types.mycompany.com/BusinessEventMessage/Customer/CustomerProfileEvent/1/0"'

# Now extract the XML
# Tweak output slightly to improve matching between Baseline and Test XML
zgrep -e "$matcher" $eventlog | grep "\[INFO\]" \
  | awk -F"$matcher" 'NF>1 {print $2}' \
  | sed 's?'"${nsb}"'?'"${nsa}"'?' \
  | awk -F'ns3:BusinessEventId' '{print $1 FS substr($2,1,20) "....</" FS $3 }' \
  | awk -F'ns3:MessageTime' '{print $1 FS substr($2,1,11) "....</" FS $3 }' \
  | sort > $outputfile

# EXPLANATION of the above
#   grep: only want INFO messages from the log 
#   awk:  get the XML part, after $matcher string
#   sed:  swaps $nsb back to $nsa, with quoting to handle (/") chars
#   awk:  truncate BusinessEventId 
#   awk:  truncate MessageTime 
#   sort: helps XML matching

echo "XML extracted to $outputfile"

The resulting XML files are used to compare baseline vs. test versions of the application, using a simple diff comparison.