tools/fix_date.sh

   1 #!/bin/sh
   2
   3 #  * Copyright 2004 Tristan Chabredier <wwp@claws-mail.org>
   4 #  *
   5 #  * This file is free software; you can redistribute it and/or modify it
   6 #  * under the terms of the GNU General Public License as published by
   7 #  * the Free Software Foundation; either version 3 of the License, or
   8 #  * (at your option) any later version.
   9 #  *
  10 #  * This program is distributed in the hope that it will be useful, but
  11 #  * WITHOUT ANY WARRANTY; without even the implied warranty of
  12 #  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  13 #  * General Public License for more details.
  14 #  *
  15 #  * You should have received a copy of the GNU General Public License
  16 #  * along with this program; if not, write to the Free Software
  17 #  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18 #
  19 # fix_date.sh           helper script to fix non-standard date or add missing
  20 #                                       date header to emails
  21
  22 # usage: fix_date.sh <filename> [<filename> ..]
  23 # It will replace the Date: value w/ the one picked up from more recent
  24 # Fetchinfo time header, Received: field.. Otherwise, it will take the file
  25 #  modification time (using a RFC 2822-compliant form).
  26 # Any already existing X-Original-Date is kept, if missing we're adding it
  27 # if the Date: was set (even if set w/ non conform value)
  28
  29 # TODO: fallback to X-OriginalArrivalTime: ?
  30
  31 VERSION="0.1.4"
  32
  33
  34 version()
  35 {
  36         echo "$VERSION"
  37         exit 0
  38 }
  39
  40 usage()
  41 {
  42         echo "usage:"
  43         echo "  ${0##*/} [<switches>] <filename> [<filename> ..]"
  44         echo "switches:"
  45         echo "  -h --help     display this help then exit"
  46         echo "  -v --version  display version information then exit"
  47         echo "  -d --debug    turn on debug information (be more verbose)"
  48         echo "  -f --force    always force (re-)writing of Date: header"
  49         echo "  -r --rfc      force re-writing of Date: header when it's not RFC-compliant"
  50         echo "  -s --strict   use RFC-strict matching patterns for dates"
  51         echo "  --            end of switches (in case a filename starts with a -)"
  52     echo "this script requires coreutils (cat, cut, head, tr), dos2unix, grep and set"
  53     echo "in PATH to work"
  54         exit $1
  55 }
  56
  57 date_valid()
  58 {
  59         test $STRICT -eq 1 && \
  60                 REGEXP="$DATE_REGEXP_STRICT" || \
  61                 REGEXP="$DATE_REGEXP"
  62
  63         echo "$1" | grep -qEim 1 "$REGEXP"
  64         DATE_VALID=$?
  65 }
  66
  67 dump_date_fields()
  68 {
  69         test -z "$X_ORIGINAL_DATE" -a -n "$DATE" && \
  70                 echo "X-Original-Date:$DATE" >> "$TMP"
  71         echo "Date:$REPLACEMENT_DATE" >> "$TMP"
  72 }
  73
  74
  75 # use --force to always (re-)write the Date header
  76 # otherwise, the Date header will be written if only it doesn't exist
  77 FORCE=0
  78 # use --rfc to (re-)write the Date header when it's not RFC-compliant
  79 # otherwise, the Date header will be written if only it doesn't exist
  80 RFC=0
  81 # use --debug to display more information about what's performed
  82 DEBUG=0
  83 # use --strict to use strict matching patterns for date validation
  84 STRICT=0
  85 # 0 = valid, always valid until --strict is used, then date_valid overrides this value
  86 DATE_VALID=0
  87 # max header lines (300 is a reasonable minimum value but 600 has already been encountered, set to 1000 by security)
  88 MAX_HEADER_LINES=1000
  89
  90 while [ -n "$1" ]
  91 do
  92         case "$1" in
  93         -h|--help)              usage 0;;
  94         -v|--version)   version;;
  95         -f|--force)             FORCE=1;;
  96         -d|--debug)             DEBUG=1;;
  97         -r|--rfc)               RFC=1;;
  98         -s|--strict)    STRICT=1;;
  99         --)                             shift
 100                                         break;;
 101         -*)                             echo "error: unrecognized switch '$1'"
 102                                         usage 1;;
 103         *)                              break;;
 104         esac
 105         shift
 106 done
 107
 108 if [ $FORCE -eq 1 -a $RFC -eq 1 ]
 109 then
 110         echo "error: use either --force or --rfc, but not both at the same time"
 111         usage 1
 112 fi
 113
 114 test $# -lt 1 && \
 115         usage 1
 116
 117 for PROG in dos2unix grep sed
 118 do
 119     type "$PROG" >/dev/null 2>&1 || \
 120         { echo "error: $PROG not found in PATH"; exit 1; }
 121 done
 122
 123 TMPDIR="/tmp"
 124 TMP="$TMPDIR/${0##*/}.$$.tmp"
 125 echo > "$TMP" >/dev/null 2>&1
 126 if [ $? -eq 0 ]
 127 then
 128     rm -f "$TMP" >/dev/null 2>&1
 129 else
 130     TMPDIR="$HOME"
 131     TMP="$TMPDIR/${0##*/}.$$.tmp"
 132 fi
 133 HEADERS="$TMPDIR/${0##*/}.$$.headers.tmp"
 134 BODY="$TMPDIR/${0##*/}.$$.body.tmp"
 135
 136 DATE_REGEXP='( (Mon|Tue|Wed|Thu|Fri|Sat|Sun),)? [0-9]+ (Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec) [0-9]+ [0-9]+:[0-9]+:[0-9]+ [+-]?[0-9]+'
 137 DATE_REGEXP_STRICT='(Mon|Tue|Wed|Thu|Fri|Sat|Sun), [0-9]+ (Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec) [0-9]+ [0-9]+:[0-9]+:[0-9]+ [+-]?[0-9]+'
 138
 139 while [ -n "$1" ]
 140 do
 141         # skip if file is empty or doesn't exist
 142         if [ ! -s "$1" ]
 143         then
 144                 test $DEBUG -eq 1 && \
 145                         echo "$1: no found or empty, skipping"
 146                 shift
 147                 continue
 148         fi
 149         SKIP=0
 150
 151         # split headers and body
 152         # find the empty line that separates body (if any) from headers,
 153         # work on a temporary dos2unix'ed copy because body might
 154         # contain DOS CRLF and grep '^$' won't work
 155         head -$MAX_HEADER_LINES "$1" | dos2unix > "$TMP"
 156         SEP=`grep -nEm1 "^$" "$TMP" 2>/dev/null | cut -d ':' -f 1`
 157         rm -f "$TMP"
 158         if [ -z "$SEP" -o "$SEP" = "0" -o $? -ne 0 ]
 159         then
 160                 cp -f "$1" "$HEADERS"
 161                 :> "$BODY"
 162                 test $DEBUG -eq 1 && \
 163                         echo "$1: no body part could be found before line $MAX_HEADER_LINES"
 164         else
 165                 sed -n '1,'`expr $SEP - 1`'p' "$1" > "$HEADERS"
 166                 sed '1,'`expr $SEP - 1`'d' "$1" > "$BODY"
 167         fi
 168
 169         # work on headers only
 170
 171         # get the Date and X-Original-Date
 172         X_ORIGINAL_DATE=`sed -n '/^X-Original-Date:/,/^[^\t]/p' "$HEADERS" | head -n -1 | cut -d ':' -f 2-`
 173         DATE=`sed -n '/^Date:/,/^[^\t]/p' "$HEADERS" | head -n -1 | cut -d ':' -f 2-`
 174
 175         # work on headers, minus Date and X-Original-Date
 176         test -n "$X_ORIGINAL_DATE" && \
 177                 sed -i '/^X-Original-Date:/,/^[^\t]/d' "$HEADERS"
 178         test -n "$DATE" && \
 179                 sed -i '/^Date:/,/^[^\t]/d' "$HEADERS"
 180
 181         # find a replacement date in Fetchinfo: header
 182         FETCH_DATE=`grep -im1 'X-FETCH-TIME: ' "$HEADERS" | cut -d ' ' -f 2-`
 183
 184         # or in Received: headers ..
 185         test $STRICT -eq 1 && \
 186                 REGEXP="$DATE_REGEXP" || \
 187                 REGEXP="$DATE_REGEXP_STRICT"
 188         RECEIVED_DATE=`sed -n '/^Received:/,/^[^\t]/p' "$HEADERS" | head -n -1 | grep -Eoim 1 "$REGEXP"`
 189
 190         # .. or from file properties
 191         FILE_DATE=`LC_ALL=POSIX LANG=POSIX ls -l --time-style="+%a, %d %b %Y %X %z" "$1" | tr -s ' ' | cut -d ' ' -f 6-11`
 192         # we could also use the system date as a possible replacement
 193         SYSTEM_DATE="`date -R`"
 194
 195         # determine which replacement date to use
 196         if [ -z "$FETCH_DATE" ]
 197         then
 198                 if [ -z "$RECEIVED_DATE" ]
 199                 then
 200                         # don't forget the leading whitespace here
 201                         REPLACEMENT_DATE=" $FILE_DATE"
 202                         REPLACEMENT="file date"
 203 #                       REPLACEMENT_DATE=" $SYSTEM_DATE"
 204 #                       REPLACEMENT="system date"
 205                 else
 206                         REPLACEMENT_DATE="$RECEIVED_DATE"
 207                         REPLACEMENT="received date"
 208                 fi
 209         else
 210                 # don't forget the leading whitespace here
 211                 REPLACEMENT_DATE=" $FETCH_DATE"
 212                 REPLACEMENT="Fetchinfo time header"
 213         fi
 214
 215         # ensure that the original X-Original-Date is kept
 216         :> "$TMP"
 217         if [ -n "$X_ORIGINAL_DATE" ]
 218         then
 219                 echo "X-Original-Date:$X_ORIGINAL_DATE" >> "$TMP"
 220         fi
 221
 222         # replace/set the date and write all lines
 223         test $RFC -eq 1 && \
 224                 date_valid "$DATE"
 225         if [ -z "$DATE" ]
 226         then
 227                 test $DEBUG -eq 1 && \
 228                         echo "$1: date not found, using $REPLACEMENT now"
 229                 dump_date_fields
 230         else
 231                 if [ $FORCE -eq 1 ]
 232                 then
 233                         test $DEBUG -eq 1 && \
 234                                 echo "$1: date already found, replacing with $REPLACEMENT"
 235                         dump_date_fields
 236                 else
 237                         if [ $RFC -eq 1 ]
 238                         then
 239                                 if [ $DATE_VALID -ne 0 ]
 240                                 then
 241                                         test $DEBUG -eq 1 && \
 242                                                 echo "$1: date already found but not RFC-compliant, replacing with $REPLACEMENT"
 243                                         dump_date_fields
 244                                 else
 245                                         test $DEBUG -eq 1 && \
 246                                                 echo "$1: date already found and RFC-compliant, skipping"
 247                                         SKIP=1
 248                                 fi
 249                         else
 250                                 test $DEBUG -eq 1 && \
 251                                         echo "$1: date already found, skipping"
 252                                 SKIP=1
 253                         fi
 254                 fi
 255         fi
 256
 257         if [ $SKIP -eq 0 ]
 258         then
 259                 # uncomment the following line to backup the original file
 260                 #mv -f "$1" "$1.bak"
 261
 262                 cat "$HEADERS" >> "$TMP"
 263                 cat "$BODY" >> "$TMP"
 264                 mv -f "$TMP" "$1"
 265                 if [ $? -ne 0 ]
 266                 then
 267                         echo "error while moving '$TMP' to '$1'"
 268                         exit 1
 269                 fi
 270         fi
 271         rm -f "$HEADERS" "$BODY" "$TMP" >/dev/null 2>&1
 272
 273         shift
 274 done
 275 exit 0