Add the contains_all() and contains_any() functions

Here are some examples which presume the default value of IFS.

contains_all " cat mat " cat dog # returns 1
contains_all " cat mat " mat cat # returns 0
contains_any " cat mat " cat dog # returns 0
contains_any " cat mat " dog     # returns 1

Here are some examples showing that IFS is taken into account.

IFS=, contains_all "cat,mat" cat dog # returns 1
IFS=, contains_all "cat,mat" mat cat # returns 0
IFS=, contains_any "cat,mat" cat dog # returns 0
IFS=, contains_any "cat,mat" dog     # returns 1

Signed-off-by: Kerin Millar <kfm@plushkava.net>

2 files changed, 172 insertions, 25 deletions

diff --git a/functions.sh b/functions.sh
index 1926c40..a4fa946 100644
--- a/functions.sh
+++ b/functions.sh
@@ -17,7 +17,7 @@
 # COLUMNS          : may be used by _update_columns() to get the column count
 # EPOCHREALTIME    : potentially used by _update_time() to get the time
 # GENFUN_MODULES   : which of the optional function collections must be sourced
-# IFS              : multiple warn() operands are joined by its first character
+# IFS              : affects contains_all(), contains_any() and warn()
 # INVOCATION_ID    : used by from_unit()
 # PORTAGE_BIN_PATH : used by from_portage()
 # RC_OPENRC_PID    : used by from_runscript()
@@ -49,6 +49,96 @@ chdir()
 }
 
 #
+# Takes the first parameter as a string comprising zero or more words, composes
+# a set consisting of the intersection of those words, then determines whether
+# the intersection of the remaining parameters forms a subset thereof. The
+# words shall be collected by splitting the string into individual fields, in
+# accordance with section 2.6.5 of the Shell Command Language specification.
+# Therefore, the value of IFS shall be taken into account. If fewer than two
+# parameters are provided, or if the first parameter yields no fields, or if the
+# second set is disjoint from - or a superset of - the first, the return value
+# shall be greater than 0.
+#
+contains_all()
+{
+	[ "$#" -ge 2 ] && IFS=${IFS} awk -f - -- "$@" <<-'EOF'
+	BEGIN {
+		ifs = ENVIRON["IFS"]
+		haystack = ARGV[1]
+		argc = ARGC
+		ARGC = 1
+		if (length(ifs) == 0) {
+			FS = "^"
+		} else if (length(ifs) != 3 || ifs ~ /[^ \t\n]/) {
+			# Split by the first character of IFS.
+			FS = "[" substr(ifs, 1, 1) "]"
+		} else {
+			# Mimic default field splitting behaviour, per section 2.6.5.
+			FS = "[ \t\n]+"
+			sub("^" FS, "", haystack)
+		}
+		# In sh, fields are terminated, not separated.
+		sub(FS "$", "", haystack)
+		len = split(haystack, words)
+		for (i = 1; i <= len; i++) {
+			set2[words[i]]
+		}
+		for (i = 2; i < argc; i++) {
+			set1[ARGV[i]]
+		}
+		for (word in set2) {
+			delete set1[word]
+		}
+		for (word in set1) {
+			exit 1
+		}
+	}
+	EOF
+}
+
+#
+# Takes the first parameter as a string comprising zero or more words then
+# determines whether at least one of the remaining parameters can be matched
+# against any of those words. The words shall be collected by splitting the
+# string into individual fields, in accordance with section 2.6.5 of the Shell
+# Command Language specification. Therefore, the value of IFS shall be taken
+# into account. If fewer than two parameters are provided, or if the first
+# parameter yields no fields, or if none of the following parameters can be
+# matched, the return value shall be greater than 0.
+#
+contains_any()
+{
+	local had_noglob haystack i item needle retval
+
+	[ "$#" -ge 2 ] || return
+	haystack=$1
+	shift
+	i=0
+	case $- in
+		*f*)
+			had_noglob=1
+			;;
+		*)
+			had_noglob=0
+	esac
+	set -f
+	for needle; do
+		if [ "$(( i += 1 ))" -eq 1 ]; then
+			# shellcheck disable=2086
+			set -- ${haystack}
+		fi
+		for item; do
+			[ "${item}" = "${needle}" ] && break 2
+		done
+	done
+	retval=$?
+	if [ "${had_noglob}" -eq 0 ]; then
+		set +f
+	fi
+	return "${retval}"
+}
+
+#
 # Considers the first parameter as an URL then attempts to fetch it with either
 # curl(1) or wget(1). If the URL does not contain a scheme then the https://
 # scheme shall be presumed. Both utilities shall be invoked in a manner that
@@ -586,29 +676,6 @@ whenceforth()
 #------------------------------------------------------------------------------#
 
 #
-# Considers the first parameter as containing zero or more blank-separated words
-# then determines whether any of the remaining parameters can be matched in
-# their capacity as discrete words.
-#
-_contains_word()
-{
-	local word wordlist
-
-	wordlist=$1 word=$2
-	case ${word} in
-		''|*[[:blank:]]*)
-			;;
-		*)
-			case " ${wordlist} " in
-				*[[:blank:]]"${word}"[[:blank:]]*)
-					return
-					;;
-			esac
-	esac
-	false
-}
-
-#
 # Determines whether the terminal is a dumb one.
 #
 _has_dumb_terminal()
@@ -761,7 +828,7 @@ _want_module()
 	local basename
 
 	basename=${1##*/}
-	_contains_word "${GENFUN_MODULES}" "${basename%.sh}"
+	contains_any "${GENFUN_MODULES}" "${basename%.sh}"
 }
 
 #
diff --git a/test-functions b/test-functions
index 34ff54a..59c0b29 100755
--- a/test-functions
+++ b/test-functions
@@ -719,6 +719,84 @@ test_substr() {
 	iterate_tests 6 "$@"
 }
 
+test_contains_all() {
+	set -- \
+		ge  1  N/A         N/A         N/A         N/A  \
+		ge  1  'foo  bar'  ''          N/A         N/A  \
+		ge  1  'foo  bar'  ''          ' '         N/A  \
+		ge  1  'foo  bar'  ''          ' bar'      N/A  \
+		ge  1  'foo  bar'  ''          ' bar'      N/A  \
+		ge  1  'foo  bar'  ''          'foo '      N/A  \
+		ge  1  'foo  bar'  ''          'foo  bar'  N/A  \
+		ge  1  'foo  bar'  ' '         ''          N/A  \
+		ge  1  'foo  bar'  ' '         ' '         N/A  \
+		ge  1  'foo  bar'  ' '         N/A         N/A  \
+		ge  1  'foo  bar'  ' bar'      ''          N/A  \
+		ge  1  'foo  bar'  ' bar'      N/A         N/A  \
+		ge  1  'foo  bar'  'foo '      ''          N/A  \
+		ge  1  'foo  bar'  'foo '      ' bar'      N/A  \
+		ge  1  'foo  bar'  'foo '      N/A         N/A  \
+		ge  1  'foo  bar'  'foo  bar'  ''          N/A  \
+		ge  1  'foo  bar'  'foo  bar'  N/A         N/A  \
+		ge  1  'foo  bar'  N/A         N/A         N/A  \
+		ge  1  'foo  bar'  bar         foo         ''   \
+		ge  1  'foo  bar'  bar         foo         ' '  \
+		ge  1  'foo  bar'  baz         bar         foo  \
+		ge  1  'foo  bar'  fo          ba          N/A  \
+		ge  1  'foo  bar'  foo         bar         ''   \
+		ge  1  'foo  bar'  foo         bar         ' '  \
+		ge  1  'foo  bar'  foo         bar         baz  \
+		ge  1  'foo  bar'  o           a           N/A  \
+		ge  1  'foo  bar'  oo          ar          N/A  \
+		eq  0  'foo  bar'  foo         bar         N/A  \
+		eq  0  'foo  bar'  bar         foo         N/A
+
+	callback() {
+		shift
+		test_description="contains_all $(quote_args "$@")"
+		contains_all "$@"
+	}
+
+	iterate_tests 6 "$@"
+}
+
+test_contains_any() {
+	set -- \
+		ge  1  N/A        N/A        N/A        \
+		ge  1  'foo bar'  N/A        N/A        \
+		ge  1  'foo bar'  fo         ba         \
+		ge  1  'foo bar'  oo         ar         \
+		ge  1  'foo bar'  o          a          \
+		ge  1  'foo bar'  'foo bar'  'foo bar'  \
+		ge  1  'foo bar'  'foo bar'  _          \
+		ge  1  'foo bar'  _          'foo bar'  \
+		ge  1  'foo bar'  'foo '     ' bar'     \
+		ge  1  'foo bar'  'foo '     _          \
+		ge  1  'foo bar'  _          ' bar'     \
+		ge  1  'foo bar'  ' bar'     _          \
+		ge  1  'foo bar'  _          'foo '     \
+		ge  1  'foo bar'  ''         ''         \
+		ge  1  'foo bar'  ''         _          \
+		ge  1  'foo bar'  _          ''         \
+		ge  1  'foo bar'  ' '        ' '        \
+		ge  1  'foo bar'  ' '        _          \
+		ge  1  'foo bar'  _          ' '        \
+		eq  0  'foo bar'  foo        bar        \
+		eq  0  'foo bar'  bar        foo        \
+		eq  0  'foo bar'  foo        _          \
+		eq  0  'foo bar'  _          bar        \
+		eq  0  'foo bar'  bar        _          \
+		eq  0  'foo bar'  _          foo
+
+	callback() {
+		shift
+		test_description="contains_any $(quote_args "$@")"
+		contains_any "$@"
+	}
+
+	iterate_tests 5 "$@"
+}
+
 iterate_tests() {
 	slice_width=$1
 	shift
@@ -794,6 +872,8 @@ test_is_subset || rc=1
 test_trueof_all || rc=1
 test_trueof_any || rc=1
 #test_substr || rc=1
+test_contains_all || rc=1
+test_contains_any || rc=1
 
 cleanup_tmpdir