introduce estrlist for sets operations, use it partially

bin/estrlist

+#!/bin/bash
+# 2009-2010, 2012 Etersoft www.etersoft.ru
+# Author: Vitaly Lipatov <lav@etersoft.ru>
+# Public domain
+
+# TODO: rewrite with shell commands, perl or C
+# Python - http://www.linuxtopia.org/online_books/programming_books/python_programming/python_ch16s03.html
+# Shell  - http://linux.byexamples.com/archives/127/uniq-and-basic-set-theory/
+#        - http://maiaco.com/articles/shellSetOperations.php
+# Perl   - http://docstore.mik.ua/orelly/perl/cookbook/ch04_09.htm
+#        - http://blogs.perl.org/users/polettix/2012/03/sets-operations.html
+# http://rosettacode.org/wiki/Symmetric_difference
+# TODO: add unit tests
+# http://ru.wikipedia.org/wiki/Операции_над_множествами
+
+# Base set operations:
+# * union
+#   "1 2 3" "3 4 5" -> "1 2 3 4 5"
+# * intersection
+#   "1 2 3" "3 4 5" -> "3"
+# * relative complement (substracted, difference) ( A ? B – members in A but not in B )
+# http://en.wikipedia.org/wiki/Complement_%28set_theory%29
+#   "1 3" "1 2 3 4" -> "2 4"
+# * symmetric difference (симметричная разность) ( A ^ B – members in A or B but not both )
+# http://en.wikipedia.org/wiki/Symmetric_difference
+#   "1 2 3" "3 4 5" -> "1 2 4 5"
+
+filter_strip_spaces()
+{
+        # possible use just
+        #xargs echo
+        sed -e "s| \+| |g" | \
+                sed -e "s|^ ||" | sed -e "s| \$||"
+}
+
+strip_spaces()
+{
+        echo "$*" | filter_strip_spaces
+}
+
+union()
+{
+        local i
+        for i in $@ ; do
+                echo "$i"
+        done | sort -u
+}
+
+
+# remove_from_list "1." "11 12 21 22" -> "21 22"
+reg_remove()
+{
+        local i
+        local RES=
+        for i in $2 ; do
+                echo "$i" | grep -q -w "$1" || RES="$RES $i"
+        done
+        strip_spaces "$RES"
+}
+
+# Args: LIST1 LIST2
+# do_exclude_list print LIST2 exclude fields contains also in LIST1
+# Example: exclude "1 3" "1 2 3 4" -> "2 4"
+exclude()
+{
+        local i
+        local RES=
+        for i in $2 ; do
+                echo "$1" | grep -q -w "$i" || RES="$RES $i"
+        done
+        strip_spaces "$RES"
+}
+
+# regexclude_list "22 1." "11 12 21 22" -> "21"
+reg_exclude()
+{
+        local i
+        local RES="$2"
+        for i in $1 ; do
+                RES=$(reg_remove "$i" "$RES")
+        done
+        strip_spaces "$RES"
+}
+
+# FIXME:
+# reg_include "1." "11 12 21 22" -> "11 12"
+reg_include()
+{
+        local i
+        local RES=
+        for i in $2 ; do
+                echo "$i" | grep -q -w "$1" && RES="$RES $i"
+        done
+        strip_spaces "$RES"
+}
+
+
+--help()
+{
+        echo "estrlist developed for string list operations. See also cut, join, paste..."
+        echo "Usage: $0 <command> [args]"
+        echo "Commands:"
+        echo "strip_spaces [args]               - remove spaces between words"
+        echo "filter_stip_spaces                - remove spaces from words from standart input"
+        echo "reg_remove  <PATTERN> [word list] - remove words containing a match to the given PATTERN (grep notation)"
+        echo "exclude <list1> [list2]           - print list2 words contains also in list1"
+        echo "reg_exclude <PATTERN> [word list]  - print only words not matched with PATTERN"
+        echo
+        echo "Examples:"
+        echo "\$ $0 reg_remove \"1.\" \"11 12 21 22\""
+        $0 reg_remove "1." "11 12 21 22"
+        echo "\$ $0 exclude \"1 3\" \"1 2 3 4\""
+        $0 exclude "1 3" "1 2 3 4"
+        echo "\$ $0 reg_exclude \"22 1.\" \"11 12 21 22\""
+        $0 reg_exclude "22 1." "11 12 21 22"
+}
+
+COMMAND="$1"
+if [ -z "$COMMAND" ] ; then
+        echo "Run with --help for get command description."
+        exit 1
+fi
+
+shift
+"$COMMAND" "$@"

share/eterbuild/functions/strings

@@ -32,23 +32,18 @@ remove_from_list()
 # Example: do_exclude_list "1 3" "1 2 3 4" -> "2 4"
 do_exclude_list()
 {
-        local i
-        local RES=
-        for i in $2 ; do
-                echo "$1" | grep -q -w "$i" || RES="$RES $i"
-        done
-        strip_spaces "$RES"
+        $ETERBUILDBIN/estrlist exclude "$@"
 }
 
 # regexp_exclude_list "22 1." "11 12 21 22" -> "21"
 regexp_exclude_list()
 {
-        local i
-        local RES="$2"
-        for i in $1 ; do
-                RES=$(remove_from_list "$i" "$RES")
-        done
-        strip_spaces "$RES"
+        $ETERBUILDBIN/estrlist reg_exclude "$@"
+}
+
+estrlist()
+{
+        $ETERBUILDBIN/estrlist "$@"
 }