#!/bin/sh # cvt2iso9660 -- pervert filenames to conform to ISO9660 format. # # Copyright (C) 1992 Free Software Foundation, Inc. # # This program is free software; you can redistribute it and/or modify # it under the terms of the GNU General Public License as published by # the Free Software Foundation; either version 2, or (at your option) # any later version. # # This program is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # GNU General Public License for more details. # # You should have received a copy of the GNU General Public License along # with this program; if not, you can either send email to this program's # author (see below) or write to: The Free Software Foundation, Inc; 675 # Mass Ave; Cambridge, MA 02139, USA. # # Please send bug reports, etc. to friedman@prep.ai.mit.edu # # This script traverses down directories specified on the command line # renaming things so they conform to the ISO9660 character set with no more # than 8 characters plus a three-character extension separated by a `.' to # a file. # A translation table of munged file name to original name is also created # in each directory. # # This script is known not to work with /bin/sh on some versions of Ultrix # (though it works with sh5 on those systems). The problem stems from # being unable to redirect io with the `read' builtin. # # Using this script is likely to prevert your precious bodily fluids. # # Created 27-Oct-92, last modified 30-Oct-92 # Version 1.0 # umask 000 # Name by which this script was invoked. progname=`basename $0` usage="Usage: ${progname} {-D} {-h} {-v} [dir1] {dir2} {...} {--debug} {--help} {--verbose} " # To prevent hairy quoting and escaping later. bq="\`" eq="'" # Unset any variables set by options below, if you don't want to take the # chance that they may be exported from the environment # Don't actually use `unset' since old bourne shells don't have this # command. Instead, assign them an empty value. debug="" verbose="" # Parse command line arguments. # If you add new options be sure to change the wildcards below to make # sure they are unambiguous (i.e. only match one possible long option) # Be sure to show at least one instance of the full long option name to # document what the long option is canonically called. # Long options which take arguments will need a `*' appended to the # canonical name to match the value appended after the `=' character. while [ "$#" != "0" ]; do case z${1} in z-D | z--debug | z--d* ) debug=t shift ;; z-h | z--help | z--h* ) echo "${usage}" 1>&2 exit 1 ;; z-v | z--verbose | z--v* ) verbose="t" shift ;; z-- ) # Stop option processing shift break ;; z-* ) echo "${progname}: unknown option \`$1'" 1>&2 echo "${usage}" 1>&2 exit 1 ;; * ) break ;; esac done test "z${debug}" != "z" && set -x if [ $# = 0 ]; then echo "${progname}: At least one directory argument is required." 1>&2 echo "${usage}" 1>&2 exit 1 fi file_list="/tmp/${progname}$$" newfile_list="/tmp/${progname}n$$" # Name of table of mutated-filename to original-filename list. transtab_basename="transtab." orig_pwd="`pwd`" # Remaining arguments are path trees to convert. for directory in "$@" ; do cd "${orig_pwd}" # Restore from any previous loop cd "${directory}" # May be relative # First build a list of all the files and directories. Hopefully `find' # will print directories before the files in them, so that renames will # happen in the correct order. if [ -n "${verbose}" ]; then echo "*** ${progname}: Building list of files in ${directory}..." 1>&2 fi find . -print | sed 's/^\.\/*//;/^$/d' > "${file_list}" if [ $? != 0 ]; then echo "*** ${progname}: Couldn't operate on ${directory}. Skipping." 1>&2 continue fi # Build list of files with their properly preverted names # Comments for sed commands here because not all versions of sed have # comments. Each label here corresponds to label in sed commands. # :l1 Convert any non d-chars to underscore (see ISO9660 7.6.1) # :l2 Remove all but first 8 chars from each pathname component. tr '[A-Z]' '[a-z]' < "${file_list}" | sed ' : l1 s/[^\/A-Za-z0-9_]/_/g; : l2 s/\([^\/][^\/][^\/][^\/][^\/][^\/][^\/][^\/]\)[^\/]*/\1/g; ' > "${newfile_list}" if [ $? != 0 ]; then echo "*** ${progname}: Couldn't operate on ${directory}. Skipping." 1>&2 continue fi if [ -n "${verbose}" ]; then echo "*** ${progname}: Renaming files..." 1>&2 fi exec 3< "${file_list}" 4< "${newfile_list}" # Unique file magic counters. Hopefully no more than a 1000 conflicts # will ever occur in a single directory. d_i=0 f_i=0 while read orig_file 0<&3 ; do orig_file_basename=`basename "${orig_file}"` read newfile 0<&4 previous_newfile_dirname="${newfile_dirname}" # Roughly equivalent to `dirname ...`, but more portable (dirname is # known to be missing on many machines) # Note that this is done *after* the sed perversions above so that # the entire pathname is munged right. newfile_dirname="`echo ${newfile} | sed 's/\/[^\/]*$//'`" if [ "z${newfile_dirname}" = "z${newfile}" ]; then newfile_dirname="." fi # If newfile isn't a directory, fix basename so it complies with # ISO9660 7.5.1 (except no version number is appended). # Section 7.5.2 is adhered to because the name+extension is kept # under 13 characters, rather than the 30 which 7.5.2 allows. # # Comments for sed commands here because not all versions of sed have # comments. Each label here corresponds to label in sed commands. # :l1 Convert any non d-chars to underscore (see ISO9660 7.4.1) # except for `.', which is kept as SEPARATOR 1. # :l2 Move any characters after the first 8 past the `.' # :l3 Get rid of any `.' chars in filename after first one. # :l4 Strip all but first 3 chars after extension (if any) # :l5 If name has no `.' in it already, append one at end. if [ ! -d "${newfile_dirname}/${orig_file_basename}" ]; then newfile_basename="`echo \"${orig_file_basename}\" \ | tr '[A-Z]' '[a-z]' \ | sed ': l1 s/[^A-Za-z0-9_.]/_/g; : l2 s/\([^.][^.][^.][^.][^.][^.][^.][^.]\)/\1./g; : l3 /\.[^.]*\./{ s/\.\([^.]*\)\./.\1/g; t l3 } : l4 s/\.\(...\).*/.\1/g; : l5 s/^\([^.]*\)$/\1./; '`" else newfile_basename="`basename ${newfile}`" fi if [ "z${newfile_dirname}" = "z." ]; then newfile="${newfile_basename}" else newfile="${newfile_dirname}/${newfile_basename}" fi # Make unique filename, if necessary. (This is only necessary if the # new file name is different from the old one. If it isn't, the file # is already unique.) # By the way, how do you catch a unique lizard? Unique up on it. if [ "z${newfile_basename}" != "z${orig_file_basename}" ]; then # reset unique file number counters if directory name has changed # since last loop. if [ "z${newfile_dirname}" != "z${previous_newfile_dirname}" ]; then d_i=0 f_i=0 fi if [ -f "${newfile}" ]; then newfile_sans_extension="`echo ${newfile} | sed 's/\.[^.\/]*$//'`" while [ -f "${newfile}" ]; do newfile="${newfile_sans_extension}.${f_i}" f_i=`expr ${f_i} + 1` done elif [ -d "${newfile}" ]; then # Technique for unique-ifying directories is different. newfile_trunc="`echo ${newfile} \ | sed 's/\([^\/][^\/][^\/][^\/][^\/]\)[^\/]*$/\1/'`" while [ -d "${newfile}" ]; do newfile="${newfile_trunc}${d_i}" d_i=`expr ${d_i} + 1` done fi newfile_basename="`basename ${newfile}`" transtab="${newfile_dirname}/${transtab_basename}" : >> "${transtab}" # Create transtab if not there already. echo "mv ${newfile_basename} ${orig_file_basename}" >> "${transtab}" mv "${newfile_dirname}/${orig_file_basename}" "${newfile}" if [ $? = 0 -a "z${verbose}" != "z" ]; then if [ "${newfile_dirname}" = "." ]; then echo "${directory}/${orig_file_basename} -> ${directory}/${newfile}" else echo "${directory}/${newfile_dirname}/${orig_file_basename} -> ${directory}/${newfile}" fi fi else if [ "z${verbose}" != "z" ]; then echo "${directory}/${newfile} not renamed" fi fi done done if [ -n "${verbose}" ]; then echo "*** ${progname}: Done." 1>&2 fi rm -f "${file_list}" "${newfile_list}" exit 0 # eof