Generates a regular expression (regex) that matches the provided version number or higher. Useful for Jamf Pro's "matches regex" operator in searches and smart groups where the results need to be the current version of an app or higher.

Match Version Number or Higher.bash

#!/bin/bash

<<ABOUT_THIS_SCRIPT
-------------------------------------------------------------------------------

	Written by:William Smith
	Professional Services Engineer
	Jamf
	bill@talkingmoose.net
	https://gist.github.com/2cf20236e665fcd7ec41311d50c89c0e
	
	Originally posted: April 12, 2020
	
	Modified: May 6, 2020
	Changes:
	Adding support to break long regex strings for Jamf Pro.
	
	Modified: May 24, 2020
	Changes:
	Displaying sequence characters in verbose reporting instead of number.
	Now accounting for version strings with non-numeric characters.
	Added warning if sequence begins with "0".
	Added warning if sequence contains non-standard characters.
	Accounting for multple Jamf Pro regex strings.

	Purpose: Generate a regular expression (regex) string that matches
	the provided version number or higher.
	
	Instructions: Run the script in Terminal, supplying a version number
	string as the first argument:
	
	e.g. '/path/to/Match Version Number or Higher.bash' 16.17
	
	Or run the script in Terminal without any argument to use the example
	version number string within the script.
	
	Optionally, set verbose to "On" or "Off".

	Except where otherwise noted, this work is licensed under
	http://creativecommons.org/licenses/by/4.0/

	"Perhaps it is the forgetting not the remembering that is the essence
	of what makes us human. To make sense of the world, we must filter it."
	
-------------------------------------------------------------------------------
ABOUT_THIS_SCRIPT

# ----- set verbosity and provide a version number string ---------------------

# turn on for step-by-step explanation while building the regex or off to provide only the regex
verbose="On" # "On" or "Off"
usingJamf="Yes" # "Yes" or "No"

# from supplied argument in Terminal
versionString=$1

# confirm version string only contains numbers and periods or is blank
if [[ $versionString =~ [^[:digit:].] ]]; then
	warning="Yes"
fi

# sample version strings
if [[ "$versionString" = "" ]]; then
	# versionString="79.0.3945.117" # e.g. Google Chrome
	# versionString="16.17" # Microsoft Office 2019
	# versionString="74.0.1" # Mozilla Firefox
	# versionString="19.10.2.41" # Citrix Workspace
	# versionString="20.006.20034" # Adobe Acrobat Reader DC
	# versionString="19.021.20058" # Adobe Acrobat Pro DC
	# versionString="21.0.3" # Adobe Photoshop 2020
	# versionString="100.86.91" # Microsoft Defender
	# versionString="5.0.3 (24978.0517)" # Zoom.us
	versionString="5.0.3-24978.0517 (4323)" # just a long and complicated test string
fi

# ----- functions -------------------------------------------------------------
	
# enables or disables verbose mode
function logcomment() {

	if [[ "$verbose" = "On" ]]; then
		echo "$1"
	fi
}

# processes a digit within a sequence
function evaluateSequence()	{

	# ----- process the first digit in a sequence -----------------------------

	# prepend exact characters leading up to the current character under evaluation
	if [[ "$regex" != "" ]]; then
		regex="$regex|"
	fi

	# get the sequence ( e.g. "74" of "74.0.1" )
	sequence=$( /usr/bin/awk -F "." -v i=$aSequence '{ print $i }' <<< "$adjustedVersionString" )
	logcomment "Sequence $aSequence is \"$sequence\""
	
	# show warning if sequence begins with "0"
	if [[ "$sequence" =~ ^0.+ ]]; then
		warning="Yes"
	fi
	
	# get count of digits in the sequence ( e.g. 2 digits in "74" )
	digitCount=$( /usr/bin/tr -d '\r\n' <<< "$sequence" | /usr/bin/wc -c | /usr/bin/xargs ) # e.g. 2
	logcomment "Count of digits in sequence \"$sequence\" is $digitCount"
	logcomment
	
	# generate regex for the first number of the sequence rolling over to add another digit ( e.g. 99 > 100 )
	logcomment "Count of digits in sequence \"$sequence\" may roll over to $((digitCount + 1)) or more digits"
	buildRegex="$regexPrefix\d{$((digitCount + 1)),}"
	logcomment "Regex for $((digitCount + 1)) or more digits is \"$buildRegex\""
	
	# add a wildcard to end of string to match everything else
	logcomment "Wildcard everything else"
	buildRegex="$buildRegex.*"
	
	# show complete regex for this digit
	logcomment "Complete regex is \"$buildRegex\""
	regex="$regex$buildRegex"
	
	# show the entire regex as the script progresses through each digit
	logcomment "Progressive regex: $regex"
	logcomment
	
	# ----- process the remaining digits in a sequence ------------------------
	
	# create array of digits in sequence ( e.g. "7, 4" )
	digits=()
	for ((i = 0; i < ${#sequence}; i++)); do
		digits+=(${sequence:$i:1})
	done

	# iterate over each digit of the sequence
	# for aDigit in ${digits[*]}
	for indexNumber in "${!digits[@]}"
	do
		# ----- the number 8 can only roll up to 9 ----------------------------
		
		if [[ "${digits[$indexNumber]}" -eq 8 ]]; then
			logcomment "Because digit $((indexNumber + 1 )) in sequence \"$sequence\" is \"8\", roll it to \"9\""
			buildRegex="9"
			
			if [[ $((digitCount - indexNumber - 1 )) -ne 0 ]]; then			
				logcomment "Because remaining count of digits in sequence \"$sequence\" is $((digitCount - indexNumber - 1 )), pad the sequence with $((digitCount - indexNumber - 1 )) more digit(s)"
				buildRegex="$buildRegex\d{$((digitCount - indexNumber - 1 )),}"
				logcomment "Regex for $((digitCount - indexNumber - 1 )) more digit(s) is \d{$((digitCount - indexNumber - 1 )),}"
			fi
			
			logcomment "Wildcard everything else"
			buildRegex="$regexPrefix$buildRegex.*"
			logcomment "Complete regex is \"$buildRegex\""
			
			logcomment "Progressive regex: $regex|$buildRegex"
			regex="$regex|$buildRegex"
			logcomment
			
		# ----- anything 0 through 7 will roll up to the next number ----------
		
		elif [[ "${digits[$indexNumber]}" -lt 8 ]]; then
			logcomment "Because digit $((indexNumber + 1 )) in sequence \"$sequence\" is \"${digits[$indexNumber]}\", roll it to \"$((${digits[$indexNumber]} + 1))\" or higher"
			buildRegex="[$((${digits[$indexNumber]} + 1))-9]"
			logcomment "Regex for $((${digits[$indexNumber]} + 1)) or higher is \"$buildRegex\""
			
			if [[ $((digitCount - indexNumber - 1 )) -ne 0 ]]; then			
				logcomment "Because remaining count of digits in sequence \"$sequence\" is $((digitCount - indexNumber - 1 )), pad the sequence with $((digitCount - indexNumber - 1 )) more digit(s)"
				buildRegex="$buildRegex\d{$((digitCount - indexNumber - 1 )),}"
				logcomment "Regex for $((digitCount - indexNumber - 1 )) more digit(s) is \d{$((digitCount - indexNumber - 1 )),}"
			fi
			
			logcomment "Wildcard everything else"
			buildRegex="$regexPrefix$buildRegex.*"
			logcomment "Complete regex is \"$buildRegex\""
			
			logcomment "Progressive regex: $regex|$buildRegex"
			regex="$regex|$buildRegex"
			logcomment
		
		# ----- nothing to do if the digit is 9 -------------------------------
		# ----- (the preceding digit is already rolled up) --------------------
		
		else
			logcomment "Because \"Digit $((indexNumber + 1 ))\" in sequence \"$sequence\" is 9, do nothing"
			logcomment
		fi
		
		regexPrefix="$regexPrefix${digits[$indexNumber]}"
	done
}

# ----- run the script --------------------------------------------------------

# verify the version string to the user
logcomment "Version string is $versionString"

# replace non-numeric sequences of characters with periods
adjustedVersionString=$( /usr/bin/sed -E 's/[^0-9]+/./g' <<< "$versionString" | /usr/bin/sed -E 's/[^0-9]$//g' )
logcomment "Adjusted version string for parsing is \"$adjustedVersionString\""

# number of "sequences" separated by a divider
sequenceCount=$( /usr/bin/awk -F "." '{ print NF }' <<< "$adjustedVersionString" ) # e.g. 4
logcomment "Number of sequences is $sequenceCount"

# create a list of sequence dividers in the version string separated by "###"
sequenceDividers=$( /usr/bin/sed -E 's/[0-9]+/###/g' <<< "$versionString" )
logcomment "Replacing digits in sequences to get the sequence dividers \"$sequenceDividers\""
logcomment

# 14 special regex characters that may appear as sequence dividers that will need escaping
regexSpecialCharacters="\&$.|?*+()[]{}"

# used to track unchanged digits to the left of the current digit being evaluated
regexPrefix=""

# evaluate the version string
for ((aSequence=1;aSequence<=$sequenceCount;aSequence++))
do
	logcomment "Evaluating sequence $aSequence of $sequenceCount"
	evaluateSequence
	
	# resetting variable
	dividers=""
	
	# add sequence divider to end of the sequence
	divider=$( /usr/bin/awk -F "###" -v divider=$(( aSequence + 1 )) '{ print $divider }' <<< "$sequenceDividers" )
	
	for (( aCharacter=0; aCharacter<${#divider}; aCharacter++ ))
	do
		logcomment "Next character is \"${divider:$aCharacter:1}\""
		
		if [[ "$regexSpecialCharacters" = *"${divider:$aCharacter:1}"* ]]; then
			dividers="$dividers\\${divider:$aCharacter:1}"
			logcomment "Escaping \"${divider:$aCharacter:1}\" to create \"\\${divider:$aCharacter:1}\""
			
		else
			dividers="$dividers${divider:$aCharacter:1}"
			logcomment "This character does not need escaping"
		fi
	done
	regexPrefix="$regexPrefix$dividers"
	logcomment "Progressive regex: $regex|$regexPrefix"
	logcomment
done

# include original version string at end of regex, escaping special regex characters
escapedVersionString=""
for (( aCharacter=0; aCharacter<${#versionString}; aCharacter++ ))
do
	
	if [[ "$regexSpecialCharacters" = *"${versionString:$aCharacter:1}"* ]]; then
		escapedVersionString="$escapedVersionString\\${versionString:$aCharacter:1}"
	else
		escapedVersionString="$escapedVersionString${versionString:$aCharacter:1}"
	fi
done

regex="$regex|$escapedVersionString"
logcomment "Adding original version string to end of regex as a potential match."
logcomment

if [[ "$warning" = "Yes" ]]; then
	echo
	echo "==============================================="
	echo "                                               "
	echo "                    WARNING                    "
	echo "                                               "
	echo "   This version string contains non-standard   "
	echo "   characters or number sequences that begin   "
	echo "   with a zero (i.e. \"0123\", which is the    "
	echo "   same as \"123\").                           "
	echo "                                               "
	echo "   Use regexes with caution.                   "
	echo "                                               "
	echo "==============================================="
	echo
fi
	
# return full regex including start and end of string characters (e.g. ^ and $ )
regex="^($regex.*)$"

# get characterCount of regex
regexCharacterCount=$( /usr/bin/wc -c <<< "$regex" | /usr/bin/xargs )

# display the regex for the version string and its character count
echo
echo "Regex for \"$versionString\" or higher ($regexCharacterCount characters):

$regex"
echo

if [[ "$usingJamf" = "Yes" ]] && [[ "$regexCharacterCount" -gt 255 ]]; then

	# get count of characters in generated regex string
	regexCharacters=${#regex}
	
	# determine number of regex strings needed, accounting for beginning ^ and ending $ characters
	jamfStringCount="$((regexCharacters / 254 + 1))"
	
	# get number of sequences separated by | in regex
	sequenceCount=$( /usr/bin/awk -F "|" '{ print NF }' <<< "$regex" )
	
	# divide the count of sequences in half
	breakDelimiterPosition=$((sequenceCount / jamfStringCount))
	
	# replace middle | operator(s) with the letter "b"
	dividedRegex="$regex"
	
	for (( aBreak=0; aBreak<$jamfStringCount; aBreak++ ))
	do
		breakDelimiterPosition=$((breakDelimiterPosition * aBreak + breakDelimiterPosition))	
		dividedRegex=$( /usr/bin/sed "s/|/b/$breakDelimiterPosition" <<< "$dividedRegex" )
	done

	# print Jamf Pro instructions and both regex strings
	echo
	echo "Jamf Pro has a field character limit of 255 characters."
	echo "This regex exceeds that field character limit."
	echo "Add additional \"Application Version\" criteria to your search"
	echo "and paste each regex string into the the additional fields."
	echo
	echo
	echo "For example:"
	echo
	echo "              Application Title       is                Google Chrome.app"
	echo "and     (     Application Version     matches regex     <Regex 1>"
	echo "or            Application Version     matches regex     <Regex 2>     )"
	echo
	echo
	
	# display each Jamf Pro string
	for (( aBreak=0; aBreak<$jamfStringCount; aBreak++ ))
	do
		regexString=$( /usr/bin/awk -F "b" -v divider=$(( aBreak + 1 )) '{ print $divider }' <<< "$dividedRegex" )
		
		# add beginning of line characters if needed
		if [[ "$regexString" != "^("* ]]; then
			regexString="^($regexString"
		fi
		
		# add end of line characters if needed
		if [[ "$regexString" != *")$" ]]; then
			regexString="$regexString)$"
		fi
		
		# display each regex string
		echo "Regex $((aBreak + 1)):"
		echo "$regexString"
		echo
	done
fi

exit 0

Sample Output 1

Regex for "79.0.3945.117" or higher (249 characters):

^(\d{3,}.*|[8-9]\d{1,}.*|79\.\d{2,}.*|79\.[1-9].*|79\.0\.\d{5,}.*|79\.0\.[4-9]\d{3,}.*|79\.0\.39[5-9]\d{1,}.*|79\.0\.394[6-9].*|79\.0\.3945\.\d{4,}.*|79\.0\.3945\.[2-9]\d{2,}.*|79\.0\.3945\.1[2-9]\d{1,}.*|79\.0\.3945\.11[8-9].*|79\.0\.3945\.117.*)$

Sample Output 2

Version string is 79.0.3945.117
Adjusted version string for parsing is "79.0.3945.117"
Number of sequences is 4
Replacing digits in sequences to get the sequence dividers "###.###.###.###"

Evaluating sequence 1 of 4
Sequence 1 is "79"
Count of digits in sequence "79" is 2

Count of digits in sequence "79" may roll over to 3 or more digits
Regex for 3 or more digits is "\d{3,}"
Wildcard everything else
Complete regex is "\d{3,}.*"
Progressive regex: \d{3,}.*

Because digit 1 in sequence "79" is "7", roll it to "8" or higher
Regex for 8 or higher is "[8-9]"
Because remaining count of digits in sequence "79" is 1, pad the sequence with 1 more digit(s)
Regex for 1 more digit(s) is \d{1,}
Wildcard everything else
Complete regex is "[8-9]\d{1,}.*"
Progressive regex: \d{3,}.*|[8-9]\d{1,}.*

Because "Digit 2" in sequence "79" is 9, do nothing

Next character is "."
Escaping "." to create "\."
Progressive regex: \d{3,}.*|[8-9]\d{1,}.*|79\.

Evaluating sequence 2 of 4
Sequence 2 is "0"
Count of digits in sequence "0" is 1

Count of digits in sequence "0" may roll over to 2 or more digits
Regex for 2 or more digits is "79\.\d{2,}"
Wildcard everything else
Complete regex is "79\.\d{2,}.*"
Progressive regex: \d{3,}.*|[8-9]\d{1,}.*|79\.\d{2,}.*

Because digit 1 in sequence "0" is "0", roll it to "1" or higher
Regex for 1 or higher is "[1-9]"
Wildcard everything else
Complete regex is "79\.[1-9].*"
Progressive regex: \d{3,}.*|[8-9]\d{1,}.*|79\.\d{2,}.*|79\.[1-9].*

Next character is "."
Escaping "." to create "\."
Progressive regex: \d{3,}.*|[8-9]\d{1,}.*|79\.\d{2,}.*|79\.[1-9].*|79\.0\.

Evaluating sequence 3 of 4
Sequence 3 is "3945"
Count of digits in sequence "3945" is 4

Count of digits in sequence "3945" may roll over to 5 or more digits
Regex for 5 or more digits is "79\.0\.\d{5,}"
Wildcard everything else
Complete regex is "79\.0\.\d{5,}.*"
Progressive regex: \d{3,}.*|[8-9]\d{1,}.*|79\.\d{2,}.*|79\.[1-9].*|79\.0\.\d{5,}.*

Because digit 1 in sequence "3945" is "3", roll it to "4" or higher
Regex for 4 or higher is "[4-9]"
Because remaining count of digits in sequence "3945" is 3, pad the sequence with 3 more digit(s)
Regex for 3 more digit(s) is \d{3,}
Wildcard everything else
Complete regex is "79\.0\.[4-9]\d{3,}.*"
Progressive regex: \d{3,}.*|[8-9]\d{1,}.*|79\.\d{2,}.*|79\.[1-9].*|79\.0\.\d{5,}.*|79\.0\.[4-9]\d{3,}.*

Because "Digit 2" in sequence "3945" is 9, do nothing

Because digit 3 in sequence "3945" is "4", roll it to "5" or higher
Regex for 5 or higher is "[5-9]"
Because remaining count of digits in sequence "3945" is 1, pad the sequence with 1 more digit(s)
Regex for 1 more digit(s) is \d{1,}
Wildcard everything else
Complete regex is "79\.0\.39[5-9]\d{1,}.*"
Progressive regex: \d{3,}.*|[8-9]\d{1,}.*|79\.\d{2,}.*|79\.[1-9].*|79\.0\.\d{5,}.*|79\.0\.[4-9]\d{3,}.*|79\.0\.39[5-9]\d{1,}.*

Because digit 4 in sequence "3945" is "5", roll it to "6" or higher
Regex for 6 or higher is "[6-9]"
Wildcard everything else
Complete regex is "79\.0\.394[6-9].*"
Progressive regex: \d{3,}.*|[8-9]\d{1,}.*|79\.\d{2,}.*|79\.[1-9].*|79\.0\.\d{5,}.*|79\.0\.[4-9]\d{3,}.*|79\.0\.39[5-9]\d{1,}.*|79\.0\.394[6-9].*

Next character is "."
Escaping "." to create "\."
Progressive regex: \d{3,}.*|[8-9]\d{1,}.*|79\.\d{2,}.*|79\.[1-9].*|79\.0\.\d{5,}.*|79\.0\.[4-9]\d{3,}.*|79\.0\.39[5-9]\d{1,}.*|79\.0\.394[6-9].*|79\.0\.3945\.

Evaluating sequence 4 of 4
Sequence 4 is "117"
Count of digits in sequence "117" is 3

Count of digits in sequence "117" may roll over to 4 or more digits
Regex for 4 or more digits is "79\.0\.3945\.\d{4,}"
Wildcard everything else
Complete regex is "79\.0\.3945\.\d{4,}.*"
Progressive regex: \d{3,}.*|[8-9]\d{1,}.*|79\.\d{2,}.*|79\.[1-9].*|79\.0\.\d{5,}.*|79\.0\.[4-9]\d{3,}.*|79\.0\.39[5-9]\d{1,}.*|79\.0\.394[6-9].*|79\.0\.3945\.\d{4,}.*

Because digit 1 in sequence "117" is "1", roll it to "2" or higher
Regex for 2 or higher is "[2-9]"
Because remaining count of digits in sequence "117" is 2, pad the sequence with 2 more digit(s)
Regex for 2 more digit(s) is \d{2,}
Wildcard everything else
Complete regex is "79\.0\.3945\.[2-9]\d{2,}.*"
Progressive regex: \d{3,}.*|[8-9]\d{1,}.*|79\.\d{2,}.*|79\.[1-9].*|79\.0\.\d{5,}.*|79\.0\.[4-9]\d{3,}.*|79\.0\.39[5-9]\d{1,}.*|79\.0\.394[6-9].*|79\.0\.3945\.\d{4,}.*|79\.0\.3945\.[2-9]\d{2,}.*

Because digit 2 in sequence "117" is "1", roll it to "2" or higher
Regex for 2 or higher is "[2-9]"
Because remaining count of digits in sequence "117" is 1, pad the sequence with 1 more digit(s)
Regex for 1 more digit(s) is \d{1,}
Wildcard everything else
Complete regex is "79\.0\.3945\.1[2-9]\d{1,}.*"
Progressive regex: \d{3,}.*|[8-9]\d{1,}.*|79\.\d{2,}.*|79\.[1-9].*|79\.0\.\d{5,}.*|79\.0\.[4-9]\d{3,}.*|79\.0\.39[5-9]\d{1,}.*|79\.0\.394[6-9].*|79\.0\.3945\.\d{4,}.*|79\.0\.3945\.[2-9]\d{2,}.*|79\.0\.3945\.1[2-9]\d{1,}.*

Because digit 3 in sequence "117" is "7", roll it to "8" or higher
Regex for 8 or higher is "[8-9]"
Wildcard everything else
Complete regex is "79\.0\.3945\.11[8-9].*"
Progressive regex: \d{3,}.*|[8-9]\d{1,}.*|79\.\d{2,}.*|79\.[1-9].*|79\.0\.\d{5,}.*|79\.0\.[4-9]\d{3,}.*|79\.0\.39[5-9]\d{1,}.*|79\.0\.394[6-9].*|79\.0\.3945\.\d{4,}.*|79\.0\.3945\.[2-9]\d{2,}.*|79\.0\.3945\.1[2-9]\d{1,}.*|79\.0\.3945\.11[8-9].*

Progressive regex: \d{3,}.*|[8-9]\d{1,}.*|79\.\d{2,}.*|79\.[1-9].*|79\.0\.\d{5,}.*|79\.0\.[4-9]\d{3,}.*|79\.0\.39[5-9]\d{1,}.*|79\.0\.394[6-9].*|79\.0\.3945\.\d{4,}.*|79\.0\.3945\.[2-9]\d{2,}.*|79\.0\.3945\.1[2-9]\d{1,}.*|79\.0\.3945\.11[8-9].*|79\.0\.3945\.117

Adding original version string to end of regex as a potential match.


Regex for "79.0.3945.117" or higher (249 characters):

^(\d{3,}.*|[8-9]\d{1,}.*|79\.\d{2,}.*|79\.[1-9].*|79\.0\.\d{5,}.*|79\.0\.[4-9]\d{3,}.*|79\.0\.39[5-9]\d{1,}.*|79\.0\.394[6-9].*|79\.0\.3945\.\d{4,}.*|79\.0\.3945\.[2-9]\d{2,}.*|79\.0\.3945\.1[2-9]\d{1,}.*|79\.0\.3945\.11[8-9].*|79\.0\.3945\.117.*)$

Sample Output 3

Version string is 5.0.3-24978.0517 (4323)
Adjusted version string for parsing is "5.0.3.24978.0517.4323"
Number of sequences is 6
Replacing digits in sequences to get the sequence dividers "###.###.###-###.### (###)"

Evaluating sequence 1 of 6
Sequence 1 is "5"
Count of digits in sequence "5" is 1

Count of digits in sequence "5" may roll over to 2 or more digits
Regex for 2 or more digits is "\d{2,}"
Wildcard everything else
Complete regex is "\d{2,}.*"
Progressive regex: \d{2,}.*

Because digit 1 in sequence "5" is "5", roll it to "6" or higher
Regex for 6 or higher is "[6-9]"
Wildcard everything else
Complete regex is "[6-9].*"
Progressive regex: \d{2,}.*|[6-9].*

Next character is "."
Escaping "." to create "\."
Progressive regex: \d{2,}.*|[6-9].*|5\.

Evaluating sequence 2 of 6
Sequence 2 is "0"
Count of digits in sequence "0" is 1

Count of digits in sequence "0" may roll over to 2 or more digits
Regex for 2 or more digits is "5\.\d{2,}"
Wildcard everything else
Complete regex is "5\.\d{2,}.*"
Progressive regex: \d{2,}.*|[6-9].*|5\.\d{2,}.*

Because digit 1 in sequence "0" is "0", roll it to "1" or higher
Regex for 1 or higher is "[1-9]"
Wildcard everything else
Complete regex is "5\.[1-9].*"
Progressive regex: \d{2,}.*|[6-9].*|5\.\d{2,}.*|5\.[1-9].*

Next character is "."
Escaping "." to create "\."
Progressive regex: \d{2,}.*|[6-9].*|5\.\d{2,}.*|5\.[1-9].*|5\.0\.

Evaluating sequence 3 of 6
Sequence 3 is "3"
Count of digits in sequence "3" is 1

Count of digits in sequence "3" may roll over to 2 or more digits
Regex for 2 or more digits is "5\.0\.\d{2,}"
Wildcard everything else
Complete regex is "5\.0\.\d{2,}.*"
Progressive regex: \d{2,}.*|[6-9].*|5\.\d{2,}.*|5\.[1-9].*|5\.0\.\d{2,}.*

Because digit 1 in sequence "3" is "3", roll it to "4" or higher
Regex for 4 or higher is "[4-9]"
Wildcard everything else
Complete regex is "5\.0\.[4-9].*"
Progressive regex: \d{2,}.*|[6-9].*|5\.\d{2,}.*|5\.[1-9].*|5\.0\.\d{2,}.*|5\.0\.[4-9].*

Next character is "-"
This character does not need escaping
Progressive regex: \d{2,}.*|[6-9].*|5\.\d{2,}.*|5\.[1-9].*|5\.0\.\d{2,}.*|5\.0\.[4-9].*|5\.0\.3-

Evaluating sequence 4 of 6
Sequence 4 is "24978"
Count of digits in sequence "24978" is 5

Count of digits in sequence "24978" may roll over to 6 or more digits
Regex for 6 or more digits is "5\.0\.3-\d{6,}"
Wildcard everything else
Complete regex is "5\.0\.3-\d{6,}.*"
Progressive regex: \d{2,}.*|[6-9].*|5\.\d{2,}.*|5\.[1-9].*|5\.0\.\d{2,}.*|5\.0\.[4-9].*|5\.0\.3-\d{6,}.*

Because digit 1 in sequence "24978" is "2", roll it to "3" or higher
Regex for 3 or higher is "[3-9]"
Because remaining count of digits in sequence "24978" is 4, pad the sequence with 4 more digit(s)
Regex for 4 more digit(s) is \d{4,}
Wildcard everything else
Complete regex is "5\.0\.3-[3-9]\d{4,}.*"
Progressive regex: \d{2,}.*|[6-9].*|5\.\d{2,}.*|5\.[1-9].*|5\.0\.\d{2,}.*|5\.0\.[4-9].*|5\.0\.3-\d{6,}.*|5\.0\.3-[3-9]\d{4,}.*

Because digit 2 in sequence "24978" is "4", roll it to "5" or higher
Regex for 5 or higher is "[5-9]"
Because remaining count of digits in sequence "24978" is 3, pad the sequence with 3 more digit(s)
Regex for 3 more digit(s) is \d{3,}
Wildcard everything else
Complete regex is "5\.0\.3-2[5-9]\d{3,}.*"
Progressive regex: \d{2,}.*|[6-9].*|5\.\d{2,}.*|5\.[1-9].*|5\.0\.\d{2,}.*|5\.0\.[4-9].*|5\.0\.3-\d{6,}.*|5\.0\.3-[3-9]\d{4,}.*|5\.0\.3-2[5-9]\d{3,}.*

Because "Digit 3" in sequence "24978" is 9, do nothing

Because digit 4 in sequence "24978" is "7", roll it to "8" or higher
Regex for 8 or higher is "[8-9]"
Because remaining count of digits in sequence "24978" is 1, pad the sequence with 1 more digit(s)
Regex for 1 more digit(s) is \d{1,}
Wildcard everything else
Complete regex is "5\.0\.3-249[8-9]\d{1,}.*"
Progressive regex: \d{2,}.*|[6-9].*|5\.\d{2,}.*|5\.[1-9].*|5\.0\.\d{2,}.*|5\.0\.[4-9].*|5\.0\.3-\d{6,}.*|5\.0\.3-[3-9]\d{4,}.*|5\.0\.3-2[5-9]\d{3,}.*|5\.0\.3-249[8-9]\d{1,}.*

Because digit 5 in sequence "24978" is "8", roll it to "9"
Wildcard everything else
Complete regex is "5\.0\.3-24979.*"
Progressive regex: \d{2,}.*|[6-9].*|5\.\d{2,}.*|5\.[1-9].*|5\.0\.\d{2,}.*|5\.0\.[4-9].*|5\.0\.3-\d{6,}.*|5\.0\.3-[3-9]\d{4,}.*|5\.0\.3-2[5-9]\d{3,}.*|5\.0\.3-249[8-9]\d{1,}.*|5\.0\.3-24979.*

Next character is "."
Escaping "." to create "\."
Progressive regex: \d{2,}.*|[6-9].*|5\.\d{2,}.*|5\.[1-9].*|5\.0\.\d{2,}.*|5\.0\.[4-9].*|5\.0\.3-\d{6,}.*|5\.0\.3-[3-9]\d{4,}.*|5\.0\.3-2[5-9]\d{3,}.*|5\.0\.3-249[8-9]\d{1,}.*|5\.0\.3-24979.*|5\.0\.3-24978\.

Evaluating sequence 5 of 6
Sequence 5 is "0517"
Count of digits in sequence "0517" is 4

Count of digits in sequence "0517" may roll over to 5 or more digits
Regex for 5 or more digits is "5\.0\.3-24978\.\d{5,}"
Wildcard everything else
Complete regex is "5\.0\.3-24978\.\d{5,}.*"
Progressive regex: \d{2,}.*|[6-9].*|5\.\d{2,}.*|5\.[1-9].*|5\.0\.\d{2,}.*|5\.0\.[4-9].*|5\.0\.3-\d{6,}.*|5\.0\.3-[3-9]\d{4,}.*|5\.0\.3-2[5-9]\d{3,}.*|5\.0\.3-249[8-9]\d{1,}.*|5\.0\.3-24979.*|5\.0\.3-24978\.\d{5,}.*

Because digit 1 in sequence "0517" is "0", roll it to "1" or higher
Regex for 1 or higher is "[1-9]"
Because remaining count of digits in sequence "0517" is 3, pad the sequence with 3 more digit(s)
Regex for 3 more digit(s) is \d{3,}
Wildcard everything else
Complete regex is "5\.0\.3-24978\.[1-9]\d{3,}.*"
Progressive regex: \d{2,}.*|[6-9].*|5\.\d{2,}.*|5\.[1-9].*|5\.0\.\d{2,}.*|5\.0\.[4-9].*|5\.0\.3-\d{6,}.*|5\.0\.3-[3-9]\d{4,}.*|5\.0\.3-2[5-9]\d{3,}.*|5\.0\.3-249[8-9]\d{1,}.*|5\.0\.3-24979.*|5\.0\.3-24978\.\d{5,}.*|5\.0\.3-24978\.[1-9]\d{3,}.*

Because digit 2 in sequence "0517" is "5", roll it to "6" or higher
Regex for 6 or higher is "[6-9]"
Because remaining count of digits in sequence "0517" is 2, pad the sequence with 2 more digit(s)
Regex for 2 more digit(s) is \d{2,}
Wildcard everything else
Complete regex is "5\.0\.3-24978\.0[6-9]\d{2,}.*"
Progressive regex: \d{2,}.*|[6-9].*|5\.\d{2,}.*|5\.[1-9].*|5\.0\.\d{2,}.*|5\.0\.[4-9].*|5\.0\.3-\d{6,}.*|5\.0\.3-[3-9]\d{4,}.*|5\.0\.3-2[5-9]\d{3,}.*|5\.0\.3-249[8-9]\d{1,}.*|5\.0\.3-24979.*|5\.0\.3-24978\.\d{5,}.*|5\.0\.3-24978\.[1-9]\d{3,}.*|5\.0\.3-24978\.0[6-9]\d{2,}.*

Because digit 3 in sequence "0517" is "1", roll it to "2" or higher
Regex for 2 or higher is "[2-9]"
Because remaining count of digits in sequence "0517" is 1, pad the sequence with 1 more digit(s)
Regex for 1 more digit(s) is \d{1,}
Wildcard everything else
Complete regex is "5\.0\.3-24978\.05[2-9]\d{1,}.*"
Progressive regex: \d{2,}.*|[6-9].*|5\.\d{2,}.*|5\.[1-9].*|5\.0\.\d{2,}.*|5\.0\.[4-9].*|5\.0\.3-\d{6,}.*|5\.0\.3-[3-9]\d{4,}.*|5\.0\.3-2[5-9]\d{3,}.*|5\.0\.3-249[8-9]\d{1,}.*|5\.0\.3-24979.*|5\.0\.3-24978\.\d{5,}.*|5\.0\.3-24978\.[1-9]\d{3,}.*|5\.0\.3-24978\.0[6-9]\d{2,}.*|5\.0\.3-24978\.05[2-9]\d{1,}.*

Because digit 4 in sequence "0517" is "7", roll it to "8" or higher
Regex for 8 or higher is "[8-9]"
Wildcard everything else
Complete regex is "5\.0\.3-24978\.051[8-9].*"
Progressive regex: \d{2,}.*|[6-9].*|5\.\d{2,}.*|5\.[1-9].*|5\.0\.\d{2,}.*|5\.0\.[4-9].*|5\.0\.3-\d{6,}.*|5\.0\.3-[3-9]\d{4,}.*|5\.0\.3-2[5-9]\d{3,}.*|5\.0\.3-249[8-9]\d{1,}.*|5\.0\.3-24979.*|5\.0\.3-24978\.\d{5,}.*|5\.0\.3-24978\.[1-9]\d{3,}.*|5\.0\.3-24978\.0[6-9]\d{2,}.*|5\.0\.3-24978\.05[2-9]\d{1,}.*|5\.0\.3-24978\.051[8-9].*

Next character is " "
This character does not need escaping
Next character is "("
Escaping "(" to create "\("
Progressive regex: \d{2,}.*|[6-9].*|5\.\d{2,}.*|5\.[1-9].*|5\.0\.\d{2,}.*|5\.0\.[4-9].*|5\.0\.3-\d{6,}.*|5\.0\.3-[3-9]\d{4,}.*|5\.0\.3-2[5-9]\d{3,}.*|5\.0\.3-249[8-9]\d{1,}.*|5\.0\.3-24979.*|5\.0\.3-24978\.\d{5,}.*|5\.0\.3-24978\.[1-9]\d{3,}.*|5\.0\.3-24978\.0[6-9]\d{2,}.*|5\.0\.3-24978\.05[2-9]\d{1,}.*|5\.0\.3-24978\.051[8-9].*|5\.0\.3-24978\.0517 \(

Evaluating sequence 6 of 6
Sequence 6 is "4323"
Count of digits in sequence "4323" is 4

Count of digits in sequence "4323" may roll over to 5 or more digits
Regex for 5 or more digits is "5\.0\.3-24978\.0517 \(\d{5,}"
Wildcard everything else
Complete regex is "5\.0\.3-24978\.0517 \(\d{5,}.*"
Progressive regex: \d{2,}.*|[6-9].*|5\.\d{2,}.*|5\.[1-9].*|5\.0\.\d{2,}.*|5\.0\.[4-9].*|5\.0\.3-\d{6,}.*|5\.0\.3-[3-9]\d{4,}.*|5\.0\.3-2[5-9]\d{3,}.*|5\.0\.3-249[8-9]\d{1,}.*|5\.0\.3-24979.*|5\.0\.3-24978\.\d{5,}.*|5\.0\.3-24978\.[1-9]\d{3,}.*|5\.0\.3-24978\.0[6-9]\d{2,}.*|5\.0\.3-24978\.05[2-9]\d{1,}.*|5\.0\.3-24978\.051[8-9].*|5\.0\.3-24978\.0517 \(\d{5,}.*

Because digit 1 in sequence "4323" is "4", roll it to "5" or higher
Regex for 5 or higher is "[5-9]"
Because remaining count of digits in sequence "4323" is 3, pad the sequence with 3 more digit(s)
Regex for 3 more digit(s) is \d{3,}
Wildcard everything else
Complete regex is "5\.0\.3-24978\.0517 \([5-9]\d{3,}.*"
Progressive regex: \d{2,}.*|[6-9].*|5\.\d{2,}.*|5\.[1-9].*|5\.0\.\d{2,}.*|5\.0\.[4-9].*|5\.0\.3-\d{6,}.*|5\.0\.3-[3-9]\d{4,}.*|5\.0\.3-2[5-9]\d{3,}.*|5\.0\.3-249[8-9]\d{1,}.*|5\.0\.3-24979.*|5\.0\.3-24978\.\d{5,}.*|5\.0\.3-24978\.[1-9]\d{3,}.*|5\.0\.3-24978\.0[6-9]\d{2,}.*|5\.0\.3-24978\.05[2-9]\d{1,}.*|5\.0\.3-24978\.051[8-9].*|5\.0\.3-24978\.0517 \(\d{5,}.*|5\.0\.3-24978\.0517 \([5-9]\d{3,}.*

Because digit 2 in sequence "4323" is "3", roll it to "4" or higher
Regex for 4 or higher is "[4-9]"
Because remaining count of digits in sequence "4323" is 2, pad the sequence with 2 more digit(s)
Regex for 2 more digit(s) is \d{2,}
Wildcard everything else
Complete regex is "5\.0\.3-24978\.0517 \(4[4-9]\d{2,}.*"
Progressive regex: \d{2,}.*|[6-9].*|5\.\d{2,}.*|5\.[1-9].*|5\.0\.\d{2,}.*|5\.0\.[4-9].*|5\.0\.3-\d{6,}.*|5\.0\.3-[3-9]\d{4,}.*|5\.0\.3-2[5-9]\d{3,}.*|5\.0\.3-249[8-9]\d{1,}.*|5\.0\.3-24979.*|5\.0\.3-24978\.\d{5,}.*|5\.0\.3-24978\.[1-9]\d{3,}.*|5\.0\.3-24978\.0[6-9]\d{2,}.*|5\.0\.3-24978\.05[2-9]\d{1,}.*|5\.0\.3-24978\.051[8-9].*|5\.0\.3-24978\.0517 \(\d{5,}.*|5\.0\.3-24978\.0517 \([5-9]\d{3,}.*|5\.0\.3-24978\.0517 \(4[4-9]\d{2,}.*

Because digit 3 in sequence "4323" is "2", roll it to "3" or higher
Regex for 3 or higher is "[3-9]"
Because remaining count of digits in sequence "4323" is 1, pad the sequence with 1 more digit(s)
Regex for 1 more digit(s) is \d{1,}
Wildcard everything else
Complete regex is "5\.0\.3-24978\.0517 \(43[3-9]\d{1,}.*"
Progressive regex: \d{2,}.*|[6-9].*|5\.\d{2,}.*|5\.[1-9].*|5\.0\.\d{2,}.*|5\.0\.[4-9].*|5\.0\.3-\d{6,}.*|5\.0\.3-[3-9]\d{4,}.*|5\.0\.3-2[5-9]\d{3,}.*|5\.0\.3-249[8-9]\d{1,}.*|5\.0\.3-24979.*|5\.0\.3-24978\.\d{5,}.*|5\.0\.3-24978\.[1-9]\d{3,}.*|5\.0\.3-24978\.0[6-9]\d{2,}.*|5\.0\.3-24978\.05[2-9]\d{1,}.*|5\.0\.3-24978\.051[8-9].*|5\.0\.3-24978\.0517 \(\d{5,}.*|5\.0\.3-24978\.0517 \([5-9]\d{3,}.*|5\.0\.3-24978\.0517 \(4[4-9]\d{2,}.*|5\.0\.3-24978\.0517 \(43[3-9]\d{1,}.*

Because digit 4 in sequence "4323" is "3", roll it to "4" or higher
Regex for 4 or higher is "[4-9]"
Wildcard everything else
Complete regex is "5\.0\.3-24978\.0517 \(432[4-9].*"
Progressive regex: \d{2,}.*|[6-9].*|5\.\d{2,}.*|5\.[1-9].*|5\.0\.\d{2,}.*|5\.0\.[4-9].*|5\.0\.3-\d{6,}.*|5\.0\.3-[3-9]\d{4,}.*|5\.0\.3-2[5-9]\d{3,}.*|5\.0\.3-249[8-9]\d{1,}.*|5\.0\.3-24979.*|5\.0\.3-24978\.\d{5,}.*|5\.0\.3-24978\.[1-9]\d{3,}.*|5\.0\.3-24978\.0[6-9]\d{2,}.*|5\.0\.3-24978\.05[2-9]\d{1,}.*|5\.0\.3-24978\.051[8-9].*|5\.0\.3-24978\.0517 \(\d{5,}.*|5\.0\.3-24978\.0517 \([5-9]\d{3,}.*|5\.0\.3-24978\.0517 \(4[4-9]\d{2,}.*|5\.0\.3-24978\.0517 \(43[3-9]\d{1,}.*|5\.0\.3-24978\.0517 \(432[4-9].*

Next character is ")"
Escaping ")" to create "\)"
Progressive regex: \d{2,}.*|[6-9].*|5\.\d{2,}.*|5\.[1-9].*|5\.0\.\d{2,}.*|5\.0\.[4-9].*|5\.0\.3-\d{6,}.*|5\.0\.3-[3-9]\d{4,}.*|5\.0\.3-2[5-9]\d{3,}.*|5\.0\.3-249[8-9]\d{1,}.*|5\.0\.3-24979.*|5\.0\.3-24978\.\d{5,}.*|5\.0\.3-24978\.[1-9]\d{3,}.*|5\.0\.3-24978\.0[6-9]\d{2,}.*|5\.0\.3-24978\.05[2-9]\d{1,}.*|5\.0\.3-24978\.051[8-9].*|5\.0\.3-24978\.0517 \(\d{5,}.*|5\.0\.3-24978\.0517 \([5-9]\d{3,}.*|5\.0\.3-24978\.0517 \(4[4-9]\d{2,}.*|5\.0\.3-24978\.0517 \(43[3-9]\d{1,}.*|5\.0\.3-24978\.0517 \(432[4-9].*|5\.0\.3-24978\.0517 \(4323\)

Adding original version string to end of regex as a potential match.


===============================================
                                               
                    WARNING                    
                                               
   This version string contains non-standard   
   characters or number sequences that begin   
   with a zero (i.e. "0123", which is the    
   same as "123").                           
                                               
   Use regexes with caution.                   
                                               
===============================================


Regex for "5.0.3-24978.0517 (4323)" or higher (522 characters):

^(\d{2,}.*|[6-9].*|5\.\d{2,}.*|5\.[1-9].*|5\.0\.\d{2,}.*|5\.0\.[4-9].*|5\.0\.3-\d{6,}.*|5\.0\.3-[3-9]\d{4,}.*|5\.0\.3-2[5-9]\d{3,}.*|5\.0\.3-249[8-9]\d{1,}.*|5\.0\.3-24979.*|5\.0\.3-24978\.\d{5,}.*|5\.0\.3-24978\.[1-9]\d{3,}.*|5\.0\.3-24978\.0[6-9]\d{2,}.*|5\.0\.3-24978\.05[2-9]\d{1,}.*|5\.0\.3-24978\.051[8-9].*|5\.0\.3-24978\.0517 \(\d{5,}.*|5\.0\.3-24978\.0517 \([5-9]\d{3,}.*|5\.0\.3-24978\.0517 \(4[4-9]\d{2,}.*|5\.0\.3-24978\.0517 \(43[3-9]\d{1,}.*|5\.0\.3-24978\.0517 \(432[4-9].*|5\.0\.3-24978\.0517 \(4323\).*)$


Jamf Pro has a field character limit of 255 characters.
This regex exceeds that field character limit.
Add additional "Application Version" criteria to your search
and paste each regex string into the the additional fields.


For example:

              Application Title       is                Google Chrome.app
and     (     Application Version     matches regex     <Regex 1>
or            Application Version     matches regex     <Regex 2>     )


Regex 1:
^(\d{2,}.*|[6-9].*|5\.\d{2,}.*|5\.[1-9].*|5\.0\.\d{2,}.*|5\.0\.[4-9].*|5\.0\.3-\d{6,}.*)$

Regex 2:
^(5\.0\.3-[3-9]\d{4,}.*|5\.0\.3-2[5-9]\d{3,}.*|5\.0\.3-249[8-9]\d{1,}.*|5\.0\.3-24979.*|5\.0\.3-24978\.\d{5,}.*|5\.0\.3-24978\.[1-9]\d{3,}.*|5\.0\.3-24978\.0[6-9]\d{2,}.*|5\.0\.3-24978\.05[2-9]\d{1,}.*)$

Regex 3:
^(5\.0\.3-24978\.051[8-9].*|5\.0\.3-24978\.0517 \(\d{5,}.*|5\.0\.3-24978\.0517 \([5-9]\d{3,}.*|5\.0\.3-24978\.0517 \(4[4-9]\d{2,}.*|5\.0\.3-24978\.0517 \(43[3-9]\d{1,}.*|5\.0\.3-24978\.0517 \(432[4-9].*|5\.0\.3-24978\.0517 \(4323\).*)$

@grahampugh, I can certainly include something like this for munki.

Let's think about the right approach. Right now, I have a couple of variables that require toggling within the script to make it behave differently. I can add another toggle for "regex only". Would it be wise to convert all of these to options that can be called as part of the command? Something like -J | --jamf-output, and -V | --verbose-output?

I'd then make "regex only " the default behavior without any options.

Or does it satisfy a need to keep these as variables in the script itself? I'm leaning toward the first idea.

@talkingmoose something long those lines seems perfect. I don't mean to change the default behaviour that you envisaged (unless you want to!), but a flag such as -r | --regex-only or -q | quiet which outputted only the regex would make sense. I'm not sure if --jamf-output is an appropriate name, since the regex-only output could be used in any system.

I would personally either create a new AutoPkg processor to run the script and provide the regex as an output variable for subsequent processors (e.g. JSSImporter), or possibly incorporate the functionality directly into JSSImporter (I can do that, since I'm the maintainer) so that existing .jss recipes could be used. Only the SmartGroupTemplate.xml would need to be changed to incorporate the regex.

@grahampugh Currently Jamf Pro cannot handle a regex expression longer than 255 characters, and some version strings result in a regex longer than that, so the Jamf mode was created to split the output in multiple shorter regexen

@sdagley Ah, I see what you mean. That would be tricky (but essential) to somehow incorporate into the automation. Perhaps it could be done separately in the processor though. Or hopefully Jamf will lengthen the acceptable number of characters. @talkingmoose is that worth an FR?

@grahampugh I opened PI-008150 for the 255 character regex limit, but I just checked the Jamf Pro Known Issues page and it's not listed (nor does searching on "regex" have any hits). Hopefully @talkingmoose can determine the status of that PI.

@grahampugh I asked Jamf support for a status on PI-008150, and they replied that they're still waiting for a response from engineering. Since it was reported 2 months ago you might try adding your voice as being impacted to see if generates any extra traction.

Hi @talkingmoose, a colleague has come across an edge case where the following line (312) is being scuppered by the presence of an actual b in the version string:

		dividedRegex=$( /usr/bin/sed "s/|/b/$breakDelimiterPosition" <<< "$dividedRegex" )

I'm curious as to (1) why the | needs to be substituted at all, and (2) if there is a specific reason for choosing b as the replacement string that I'm not seeing? I'm thinking of replacing it with a little-used character such as § or ±.

Hi @grahampugh! It's been a while since I've revisited this gist, but I believe the sed substitution was a response to having to break long regex strings into smaller strings when used with Jamf Pro. Its fields had a limit of around 250 characters, and long version strings like Google's needed a solution. I probably used "b" for "break here at this 'or' statement".

You're right that most any character (or short string) here would work. You'll also need to adjust the awk field delimiter on line 334 to use the same character or short string.

In the interest of not having to deal with breaking a >255 character pattern into smaller versions I've adopted Cameron Moore's make_ge_version_regexp.sh fork of this script (https://github.com/moorereason/make_ge_version_regex) which generates much shorter patterns. While the README for it refers to the optimized patterns generated as being Jamf-unsafe I have not run into an instance where they haven't worked.