feat: Create generic regex functions to support bash 3.x + perl and native bash 4+ better regexp

Author: marverix

.github/workflows/integration-tests.yaml

@@ -43,7 +43,7 @@ jobs:
             exit 1
           fi
 
-          required_commands=(tar unzip curl jq)
+          required_commands=(tar unzip curl jq openssl)
           for cmd in "${required_commands[@]}"; do
             if ! command -v "$cmd" &> /dev/null; then
               echo "::error title=Pre-requisite failed::$cmd is not installed"

README.md

@@ -29,12 +29,13 @@ Nowadays more and more command-line tools and applications are distributed via G
 
 ## Requirements
 
-- `bash`
+- `bash` (version 3 or higher)
+- `perl` (by default installed on most unix systems, required for bash < 4)
 - `jq`
 - `curl` or `wget`
-- `tar`
-- `unzip`
-- `openssl`
+- `tar` (by default installed on most unix systems)
+- `unzip` (by default installed on most unix systems)
+- `openssl` (by default installed on most unix systems)
 
 ## Installation
 

gah

@@ -71,6 +71,16 @@ function cleanup() {
 	fi
 }
 
+# Check if Bash supports non-greedy quantifiers (Bash 4.0+)
+function supports_bash_regex() {
+	local bash_major_version="${BASH_VERSINFO[0]}"
+	[[ "$bash_major_version" -ge 4 ]]
+}
+
+#endregion
+#--------------------------------------------------
+#region HTTP functions
+
 function http_get() {
 	local url="$1"
 	# if curl is available, use it, otherwise use wget.
@@ -196,6 +206,53 @@ function get_md_url_regexp() {
 	echo "\(https:\/\/[a-z0-9.\/]+\/$(get_filename_regexp)\)"
 }
 
+# Generic regex matching function
+# Usage: regex_match "string" "pattern"
+# Returns: 0 if matches, 1 if doesn't match
+function regex_match() {
+	local string="$1"
+	local pattern="$2"
+	
+	if supports_bash_regex; then
+		[[ "$string" =~ $pattern ]]
+	else
+		echo "$string" | perl -ne "exit(0) if /$pattern/; exit(1)" 2>/dev/null
+	fi
+}
+
+# Extract capture group from regex match
+# Usage: regex_capture "string" "pattern" capture_group_number
+# Returns: captured text or empty string
+function regex_capture() {
+	local string="$1"
+	local pattern="$2"
+	local capture_group="${3:-1}"
+	
+	if supports_bash_regex; then
+		if [[ "$string" =~ $pattern ]]; then
+			echo "${BASH_REMATCH[$capture_group]}"
+		fi
+	else
+		echo "$string" | perl -ne "print \$$capture_group if /$pattern/"
+	fi
+}
+
+# Extract full match from regex
+# Usage: regex_match_text "string" "pattern"
+# Returns: matched text or empty string
+function regex_match_text() {
+	local string="$1"
+	local pattern="$2"
+	
+	if supports_bash_regex; then
+		if [[ "$string" =~ $pattern ]]; then
+			echo "${BASH_REMATCH[0]}"
+		fi
+	else
+		echo "$string" | perl -ne "print \$& if /$pattern/"
+	fi
+}
+
 #endregion
 #--------------------------------------------------
 #region GitHub API functions
@@ -250,8 +307,9 @@ function fetch_release_info() {
 function get_repo_name() {
 	local release_json="$1"
 	local html_url=$(jq -r '.html_url' "$release_json")
-	if [[ "$html_url" =~ $REGEXP_NAME_FROM_HTML_URL ]]; then
-		echo "${BASH_REMATCH[1]}"
+	local match=$(regex_capture "$html_url" "$REGEXP_NAME_FROM_HTML_URL" 1)
+	if [[ -n "$match" ]]; then
+		echo "$match"
 	fi
 }
 
@@ -266,8 +324,8 @@ function find_download_url() {
 
 	for name in $(jq -r '.assets[].name' "$release_json"); do
 		local lower_name=$(echo "$name" | tr '[A-Z]' '[a-z]')
-		if [[ "$lower_name" =~ $regexp ]]; then
-			if [[ "$lower_name" =~ $REGEXP_IGNORE_PATTERN ]]; then
+		if regex_match "$lower_name" "$regexp"; then
+			if regex_match "$lower_name" "$REGEXP_IGNORE_PATTERN"; then
 				print_debug "  $name ... Ignored"
 			else
 				print_debug "  $name ... MATCH!"
@@ -292,13 +350,13 @@ function find_download_url() {
 
 	jq -r '.body' "$release_json" | while read -r line; do
 		lower_line=$(echo "$line" | tr '[A-Z]' '[a-z]')
-		if [[ "$lower_line" =~ $regexp ]]; then
+		local match=$(regex_match_text "$lower_line" "$regexp")
+		if [[ -n "$match" ]]; then
 			print_debug "  $line ... MATCH!"
 			found="true"
-			line="${BASH_REMATCH[0]}"
-			line=${line:1}
-			line=${line%?}
-			echo $line
+			match=${match:1}
+			match=${match%?}
+			echo "$match"
 		else
 			print_debug "  $line ... Doesn't match"
 		fi
@@ -526,16 +584,17 @@ function command_install() {
 		print_debug "OK, lower case file name: $lower_file_name"
 
 		print_debug "Checking if file should be skipped"
-		if [[ "$lower_file_name" =~ $REGEXP_SKIP_FILES ]]; then
+		if regex_match "$lower_file_name" "$REGEXP_SKIP_FILES"; then
 			print_debug "File name matches skip pattern - Skipping: $file_name"
 			continue
 		fi
 		print_debug "No skip pattern matched - File will be installed"
 
 		print_debug "Removing version/os/arch parts from file name if present"
 		local regexp=$(get_name_regexp)
-		if [[ "$file_name" =~ $regexp ]]; then
-			file_name="${BASH_REMATCH[1]}"
+		local match=$(regex_capture "$file_name" "$regexp" 1)
+		if [[ -n "$match" ]]; then
+			file_name="$match"
 			print_debug "OK, cleaned file name"
 		fi
 
@@ -620,7 +679,7 @@ function main() {
 			throw_error 1 "Please provide either repo in format 'owner/repo_name' or known alias.\n$HELP_STRING"
 
 		elif [[ "$2" == *"/"* ]]; then
-			if [[ "$2" =~ ^[a-zA-Z0-9_-]+/[a-zA-Z0-9_-]+$ ]]; then
+			if regex_match "$2" "^[a-zA-Z0-9_-]+/[a-zA-Z0-9_-]+$"; then
 				repo="$2"
 			else
 				throw_error 2 "Given string '$2' is not in format 'owner/repo_name'.\n$HELP_STRING"

test/00_utils.bats

@@ -24,3 +24,15 @@ teardown() {
 
 	assert_output "amd64"
 }
+
+@test "supports_bash_regex should return true for Bash 4+" {
+	# This test will pass/fail based on actual Bash version
+	# If running Bash 4+, it should succeed
+	if [[ "${BASH_VERSINFO[0]}" -ge 4 ]]; then
+		run supports_bash_regex
+		assert_success
+	else
+		run supports_bash_regex
+		assert_failure
+	fi
+}

test/01_regexp_functions.bats

@@ -192,3 +192,63 @@ teardown() {
 	assert_success
 	assert_output '([a-z][a-z0-9_-]+?)([_-]v?[0-9.]+)?([._-](apple[._-])?(darwin|macos|osx)[._-](arm64|aarch64|universal)|[._-](arm64|aarch64|universal)[._-](apple[._-])?(darwin|macos|osx))([_-][a-z0-9_-]+)?(\.zip|\.tar\.gz|\.tgz|\.tar\.xz|\.txz|\.tar\.bz2|\.tbz)?'
 }
+
+@test "regex_match should match simple pattern" {
+	run regex_match "hello-world" "^hello-world$"
+	assert_success
+}
+
+@test "regex_match should not match wrong pattern" {
+	run regex_match "hello-world" "^goodbye$"
+	assert_failure
+}
+
+@test "regex_match should match complex pattern" {
+	run regex_match "app-v1.2.3-linux-amd64.tar.gz" "^[a-z]+-v[0-9.]+-linux-amd64"
+	assert_success
+}
+
+@test "regex_match should handle non-greedy quantifiers" {
+	run regex_match "app-v1.2.3-linux-amd64" "^([a-z]+?)-.+"
+	assert_success
+}
+
+@test "regex_capture should extract first capture group" {
+	result=$(regex_capture "https://github.com/owner/repo" '^https://github.com/([^/]+)/([^/]+)$' 1)
+	assert_equal "$result" "owner"
+}
+
+@test "regex_capture should extract second capture group" {
+	result=$(regex_capture "https://github.com/owner/repo" '^https://github.com/([^/]+)/([^/]+)$' 2)
+	assert_equal "$result" "repo"
+}
+
+@test "regex_capture should extract with non-greedy quantifier" {
+	result=$(regex_capture "app-v1.2.3-linux-amd64" '^([a-z]+?)-v' 1)
+	assert_equal "$result" "app"
+}
+
+@test "regex_capture should return empty string when no match" {
+	result=$(regex_capture "hello" '^goodbye(.+)$' 1)
+	assert_equal "$result" ""
+}
+
+@test "regex_capture should default to first capture group" {
+	result=$(regex_capture "app-v1.2.3" '^([a-z]+)-v([0-9.]+)$')
+	assert_equal "$result" "app"
+}
+
+@test "regex_match_text should extract full match" {
+	result=$(regex_match_text "Download from https://example.com/file.tar.gz here" 'https://[^ ]+')
+	assert_equal "$result" "https://example.com/file.tar.gz"
+}
+
+@test "regex_match_text should return empty string when no match" {
+	result=$(regex_match_text "hello world" 'https://[^ ]+')
+	assert_equal "$result" ""
+}
+
+@test "regex_match_text should handle complex pattern with non-greedy" {
+	result=$(regex_match_text "app-v1.2.3-linux-amd64.tar.gz" '([a-z]+?)-v[0-9.]+-linux-amd64\.tar\.gz')
+	assert_equal "$result" "app-v1.2.3-linux-amd64.tar.gz"
+}

tools/install.sh

@@ -65,6 +65,14 @@ fi
 print_green "OK"
 
 # Check if required commands are installed
+# Perl is only required for Bash 3.x (Bash 4+ has native regex support)
+if [ "${BASH_VERSINFO[0]}" -lt 4 ]; then
+	print_yellow "Bash 3.x detected - Perl is required for regex support"
+	require_command perl
+else
+	print_green "Bash 4+ detected - native regex support available"
+fi
+
 require_command tar
 require_command unzip
 require_command curl wget