Mercurial > notdcc

comparison misc/fetch-testmsg-whitelist.in @ 0:c7f6b056b673

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
First import of vendor version
author Peter Gervai <grin@grin.hu>
date Tue, 10 Mar 2009 13:49:58 +0100
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:c7f6b056b673
1 #!/bin/sh
2
3 # Fetch a list of "empty" mail messages for whitelisting. Many free mail
4 # service providers add HTML or other text to mail. That causes empty
5 # and nearly empty mail messages to have valid DCC checksums and not be
6 # ignored by DCC clients.
7
8 # The fetched file can be included in whiteclnt files. For example, the
9 # following line in @prefix@whiteclnt would whitelist many common
10 # empty messages
11 # include @prefix@/testmsg-whitelist
12
13 # By default the script fetches http://www.iecc.com/dcc-testmsg-whitelist.txt
14 # to @prefix@/testmsg-whitelist
15
16 # The script should be run at most once a day.
17
18
19
20 # Rhyolite Software DCC 1.3.103-1.31 $Revision$
21 # @configure_input@
22
23 exec </dev/null
24
25 HTTP_REFERER=DCC-1.3.103-script; export HTTP_REFERER
26
27 DCC_HOMEDIR=@prefix@
28 URL=http://www.iecc.com/dcc-testmsg-whitelist.txt
29 TGT=testmsg-whitelist
30 LOG=$TGT.log
31 PGM=@FETCH_CMD@
32 FORCE=
33
34 USAGE="`basename $0`: [-xf] [-h homedir] [-p fetch-pgm] [-s src-URL]"
35 while getopts "xfh:p:s:" c; do
36 case $c in
37 x) set -x;;
38 f) FORCE=yes;;
39 h) DCC_HOMEDIR="$OPTARG";;
40 p) PGM="$OPTARG";;
41 s) URL="$OPTARG";;
42 *) echo 1>&2 "$USAGE"; exit 1;;
43 esac
44 done
45 shift `expr $OPTIND - 1 || true`
46 if test "$#" -ne 0; then
47 echo 1>&2 "$USAGE"
48 exit 1
49 fi
50
51
52 FNAME=`expr "$URL" : '.*/\([^/][^/]*\)'`
53 if test -z "$FNAME"; then
54 FNAME="$URL"
55 if test -z "$FNAME"; then
56 echo 1>&2 "source file not specified"
57 exit 1
58 fi
59 fi
60
61 cd $DCC_HOMEDIR
62
63 # don't bother if the file exists and is no more than 2 weeks old
64 if test -f "$TGT" -a -z "$FORCE"; then
65 if test "`find $TGT -mtime -7 -type f`"; then
66 date "+%n%x %X: $TGT is too recent to fetch again" >>$LOG
67 exit 0
68 fi
69 fi
70
71 # Delay for an arbitrary, somewhat random number of seconds to try to spread
72 # the load on the HTTP server for the list. Some versions of cksum yield
73 # 10 digit numbers that some versions of expr think are negative.
74 RND=`ps | cksum | sed -e 's/ */ + /' -e 's/\([0-9]\{6\}\)\([0-9]\)/\1 + \2/g'`
75 RND=`expr \( $RND \) % 123`
76 sleep $RND
77
78 # use fetch, wget, curl, or ftp that understands URLs
79 rm -f $FNAME
80 PGM_B=`basename $PGM`
81 if test "$PGM_B" = wget; then
82 PGM_B=
83 # Do not use --mirror because -r results in a 0 exit status
84 # even on failure.
85 # Do not use --no-remove-listing, -nr, or --dont-remove-listing
86 # because none of them are supported by all versions of wget.
87 # At least some versions of wget exit with 0 after having done
88 # nothing but emitting a usage message.
89 if $PGM -nd --retr-symlinks -N --no-host-directories \
90 --passive-ftp @FETCH_WGET_OPTS@ $URL >$LOG 2>&1; then
91 if test -s $FNAME; then
92 if test -n "`sed -n -e 2q \
93 -e 's/.*DOCTYPE.*/HTML/p' \
94 -e 's/<HEAD>/HTML/p' -e 's/<head>/HTML/p' $FNAME`"; then
95 rm $FNAME
96 fi
97 fi
98 fi
99 fi
100
101 if test "$PGM_B" = fetch; then
102 PGM_B=
103 $PGM -p -q -m $URL >$LOG 2>&1
104 fi
105
106 if test "$PGM_B" = curl; then
107 PGM_B=
108 $PGM -s -S --connect-timeout 30 --max-time 600 \
109 @FETCH_CURL_OPTS@ $URL -o $FNAME >$LOG 2>&1
110 # --fail does not work on at least some versions of curl
111 if test -s $FNAME; then
112 if test -n "`sed -n -e 2q \
113 -e 's/.*DOCTYPE.*/HTML/p' \
114 -e 's/<HEAD>/HTML/p' -e 's/<head>/HTML/p' $FNAME`"; then
115 rm $FNAME
116 fi
117 fi
118 fi
119
120 if test "$PGM_B" = ftp; then
121 PGM_B=
122 $PGM -p $URL >$LOG 2>&1
123 # if that did not work, try ancient anonymous FTP
124 if test ! -s $FNAME; then
125 HOST=`expr "$URL" : "ftp://\([^/]*\)/"`
126 RFILE=`expr "$URL" : "ftp://[^/]*/\(.*\)"`
127 echo "try old anonymous FTP"
128 (echo "user anonymous `hostname`"; echo "get $RFILE $FNAME") \
129 | ftp -n $HOST
130 fi
131 # some versions of ftp like to leave empty files
132 if test ! -s $FNAME; then
133 rm -f $FNAME
134 fi
135 fi
136 # if some other program was specified, use it and hope it is simple enough
137 if test -n "$PGM_B"; then
138 $PGM $URL
139 fi
140
141 if test ! -s "$FNAME"; then
142 echo 1>&2 "failed to fetch $FNAME with $PGM"
143 exit 1
144 fi
145
146 # work around wget timestamping
147 # We use the mtime of the file to reduce fetching of the file.
148 # We use `wget -N` to prevent .1 files
149 touch $FNAME
150
151 if test "$FNAME" != "$TGT"; then
152 mv -f "$FNAME" "$TGT"
153 fi
154
155 date "+%n%x %X: fetched $TGT" >>$LOG