#!/bin/sh
#
# smistrip --
#
#	Extract MIB and PIB modules from text files, like RFCs or I-Ds.
#
# Copyright (c) 1999 Frank Strauss, Technical University of Braunschweig.
# Copyright (c) Niels Baggesen, Jochen Friedrich
#
# Modified by Niels Baggesen to be somewhat more aggressive in suppressing
# blank lines, and support the -x option.
#
# Modified by Jochen Friedrich to merge the changes of libsmi back in and
# make the aggressive suppressing of blank lines optional.
#
# See the file "COPYING" for information on usage and redistribution
# of this file, and for a DISCLAIMER OF ALL WARRANTIES.
#
# $Id: smistrip.in 3434 2006-04-07 07:02:51Z strauss $
#
# NOTE, that this script relies on awk (tested with GNU awk) and getopts
# (shell builtin like in bash or standalone).
#

AWK=/usr/bin/awk
GETOPTS=getopts
VERSION=0.4.8


do_version () {
    echo "smistrip $VERSION"
}



do_usage () {
    echo "Usage: smistrip [-Vvhna] [-i dir] [-d dir] [-s suffix] [-m modules] file1 [file2 [...]]"
    echo "-V         show version and license information"
    echo "-v         verbose"
    echo "-h         show usage information"
    echo "-n         do not write module files"
    echo "-a         strip blank lines more aggressively"
    echo "-i dir     try to read files from directory dir"
    echo "-d dir     write module to directory dir"
    echo "-x suffix  append suffix to the module file name"
    echo "-m modules strip only the specified modules. For a list of modules"
    echo "           use : as a separator"
    echo "file1 ...  input files to parse (RFCs, I-Ds, ...)"
}



do_strip () {
    if [ "$indir" ] ; then
	FILE="$indir/$1"
    else
	FILE="$1"
    fi
    if [ ! -f "$FILE" -a -f "$FILE.gz" ] ; then
	FILE="$FILE.gz"
    fi
    echo "$FILE" | grep -q '\.gz$'
    if [ $? -ne 0 ] ; then
    	CMD=cat
    else
    	CMD=zcat
    fi

    $CMD "$FILE" | \
    tr -d '\015' | \
    $AWK -vtest="$test" -vdir="$dir" -vsingle="$single" -vsuffix="$suffix" -vverbose="$verbose" -vaggressive="$aggressive" '

    BEGIN {
	if (length(single) != 0) {
	    single = ":"single":"
	}
	else {
	    single = ""
	}
    }

    END {
	if (single != "" && single != ":") {
	    gsub(":", " ", single)
	    print "WARNING: Module(s) not found:" single
	}
    }

    # start of module
    /^[ \t]*[A-Za-z0-9-]* *(PIB-)?DEFINITIONS *(::=)? *(BEGIN)? *$/ {
	module = $1
	collect = 1
	macro = 0
	skip = 0
	n = 0
    }

    # at the end of a page we go back one line (which is expected to
    # be a separator line), and start the counter skipped to skip the
    # next few lines.
    /\[[pP]age [iv0-9]*\] */ {
	# some drafts do not use that separator line. so keep it if
	# there are non-blank characters.
	if (!aggressive && n && collect) {
	    if (!(line[n-1] == ""))
		print "WARNING: the line ::"line[n-1]":: should be a separator before a page break. It was kept. ";
	    else n--;
	    skip = 3
	}
	collect = 0
	next
    }

    /^[ \t]*(::=|DESCRIPTION|SYNTAX|MAX-ACCESS|MIN-ACCESS|ACCESS|STATUS|REFERENCE|INDEX|AUGMENTS|DEFVAL|UNITS|DISPLAY|")/ {
	skip = 0
	if (collect && aggressive)
	    if (line[n-1] == "") n--
    }

    # a blank line - suppress multiple
    /^[ \t\r]*$/ {
        if (collect && (skip == 0)) {
	    if (aggressive && n) {
                if (line[n-1] != "" && line[n-1] !~ /,[ \t\r]*$/) line[n++] = ""
	    }
	    else line[n++] = ""
	}
	if (skip > 0) skip--;
        next
    }

    # collect non-blank line when inside mib module
    /[^ \f\t]/ {
	if (length(module) > 0) {
	    if (!collect)
		collect = 1     # page header, stop skipping
	    else if (skip == 0)
		line[n++] = $0
	}
	if (skip > 0) skip--;
    }

    # remember when we enter a macro definition
    /^[ \t]*[A-Za-z0-9-]* *MACRO *::=/ {
	macro = 1
	skip = 0
    }

    # end of module
    /^[ \t]*END[ \t]*$/ {
	skip = 0
	if (macro == 0) {
	    if (single == "" || match(single, ":"module":")) {
	        sub(":"module, "", single)
		strip = 99
		for (i=0 ; i < n ; i++) {
		    # find the minimum column that contains non-blank characters
		    # in order to cut a blank prefix off. Ignore lines that only
                    # contain white spaces.
                    if (!(line[i] ~ /^[ \t]*$/)) {
        	        p = match(line[i], "[^ ]")
		        if ((p < strip) && (length(line[i]) > p)) { strip = p }
                    }
		}
    
		if (test != "1") {
		    if (dir) {
		       f = dir"/"module suffix
		    } else {
		       f = module suffix
		    }
		    for (i=0 ; i < n ; i++) {
			print substr(line[i], strip) >f
		    }
		}
		if (verbose) {
		    print module ": " n " lines."
		}
    
	    }
	    else
		print "NOTE: " module ": ignored."
	    module = ""
	} else {
	    macro = 0
	}
    }
    '
}



while $GETOPTS Vvhnam:i:d:x: c ; do
    case $c in
	v)	verbose=1
		;;
	n)	test=1
		;;
	a)	aggressive=1
		;;
	m)	single=$OPTARG
		;;
	i)	indir=$OPTARG
		;;
	d)	dir=$OPTARG
		;;
	x)	suffix=$OPTARG
		;;
	h)	do_usage
		exit 0
		;;
	V)	do_version
		exit 0
		;;
	*)	do_usage
		exit 1
		;;
    esac
done

shift `expr $OPTIND - 1`



if [ $# -eq 0 ] ; then
    do_strip -
else 
    for f in "$@" ; do
	do_strip "$f"
    done
fi

exit 0