Annotation of /trunk/webthumb

#!/bin/sh
#
# This script is based on work of mendel from
# http://www.lafferty.ca/software/webshots/webshot
# webshot: hackish thing to take a picture of a whole webpage.
#
# idea from sippey (http://www.sippey.com/2004/blog-snaps/)
# inspiration from bradfitz
#
# usage: webshot url filename
#
# JPEG seems to be the best format.
#
# Modified by Dobrica Pavlinusic 2004-07-25
# - include signal grabbing
# - dynamic vncserver display 
# - correct WindowID parsing
# - wait for browser to load full page (useful on slow links)
#
# It will need follwing external utilities:
# vncserver
# mozilla-firefox
# HEAD (part or LWP perl library, it can be commented out)
# nc (netcat, to simulate web server)
# xwininfo (standard part of X-clients)
# import (part of ImageMagick package)
# xdotool (much more useful than xwit for this purpose)
#
# It will create snapshot of following size
W=1024
H=768

# and then resize it to (comment out to disable)
#RESIZE=200x

# turn locally visible server to watch progress
DEBUG=0
test ! -z "$3" && DEBUG=1

# wait for $WAIT seconds for page to load
WAIT=30

# some configurable paths
BROWSER=mozilla-firefox
XSERVER=Xvfb
URL=$1
FILE=${2:-screenshot.jpg}

FRAMESET="`mktemp`.html"
PROFILE_DIR=`mktemp -d`
PORT=8888

LOCAL_DISPLAY=$DISPLAY

if [ -z "$URL" ] ; then
        echo "usage: $0 http://url.to.capture [screenshot.jpg]"
        exit 1
fi

if [ -z "`which xdotool`" ] ; then
        echo "$0 need xdotool from http://www.semicomplete.com/blog/projects/xdotool/"
        exit 1
fi

if [ -z "`which $XSERVER`" ] ; then
        echo "$0 really need $XSERVER to operate. please install it."
        exit 1
fi

while netstat -ln | grep ":$PORT " >/dev/null ; do
        PORT=`expr $PORT + 1`
done

echo "### using port $PORT"

echo -n "testing URL $URL "
if HEAD $URL >/dev/null ; then
        echo "ok"
else
        echo "FAILED"
        exit 1
fi

DISPLAY_PORT=6042

while netstat -ln | grep ":$DISPLAY_PORT " >/dev/null ; do
        DISPLAY_PORT=`expr $DISPLAY_PORT + 1`
done

DISPLAY=:`expr $DISPLAY_PORT - 6000`

echo "using DISPLAY=$DISPLAY"

if [ "$DEBUG" == 1 ] ; then
        echo "using locally visible debug server on $LOCAL_DISPLAY"
        DISPLAY=$LOCAL_DISPLAY Xephyr -ac -screen ${W}x${H}x24 $DISPLAY 2>/dev/null &
else
        echo "starting $XSERVER"
        $XSERVER -ac -screen 0 ${W}x${H}x24 $DISPLAY 2>/dev/null &
fi

if [ -z "$!" ] ; then
        echo "ABORT: can't start X server!"
        exit
fi
XSERVER_PID=$!

echo "X server pid $XSERVER_PID"

function kill_x_server() {
        echo "Killing server $XSERVER_PID"
        kill $XSERVER_PID
        rm -f $FRAMESET
        rm -fr $PROFILE_DIR
        trap '' EXIT
        exit 1
}
trap 'kill_x_server' INT QUIT TERM SEGV EXIT

# create frameset to load site and after site is loaded trigger this script
cat > $FRAMESET <<END_OF_FRAMESET
<html>
<head>
<frameset rows="0,*" border=0 frameborder=no framespacing=0 onload="window.frames['trigger_frame'].window.location.replace('http://localhost:$PORT');">
  <frame src="about:blank" name="trigger_frame" scrolling=no marginwidth=0 marginheight=0>
  <frame src="$URL" name="page_frame" scrolling=no>
</frameset>
</head>
</html>
END_OF_FRAMESET

echo "Using frameset html $FRAMESET"

echo "preview available with: xwd -display $DISPLAY -root | xwud"

echo "making 'Screenshot' profile in $PROFILE_DIR"
$BROWSER -CreateProfile "Screenshot $PROFILE_DIR" 2>/dev/null | grep Success

echo "launching browser $BROWSER with $URL"
$BROWSER -P Screenshot -width $W -height $H -safemode $FRAMESET 2>/dev/null &
BROWSER_PID=$!

function kill_browser() {
        echo "Killing browser $BROWSER_PID"
        kill $BROWSER_PID
        echo "Killing server $XSERVER_PID"
        kill $XSERVER_PID
        rm -f $FRAMESET
        rm -fr $PROFILE_DIR
        trap '' EXIT
        exit 1
}
trap 'kill_browser' INT QUIT TERM SEGV EXIT

function ping_browser() {
        echo -n "ping browser"
        while ! ( $BROWSER -remote "ping();" 2>&1 ) >/dev/null ; do
                RID=`xdotool search Restore 2>/dev/null`
                if [ ! -z "$RID" ] ; then
                        echo -n "Esc"
                        xdotool focus $RID
                        xdotool key Escape
                        sleep 1
                fi
                echo -n "."
                sleep 1
        done
        echo " OK"
}

ping_browser

$BROWSER -remote "openURL($FRAMESET)" 2>/dev/null

echo "waiting for on_load event from browser $BROWSER_PID for ${WAIT}s"

# there is hard-coded limit here:
# we will wait $WAIT sec for page to load and render

echo -e "HTTP/1.0 304 Not modified\r\n\r\n" | nc -l -w $WAIT -p $PORT >/dev/null || echo "Timeout after $WAIT sec!"

# get Mozilla Firefox window id (for resize)
WINDOW_ID=`xwininfo -display $DISPLAY -root -tree | grep gecko | cut -d\" -f1 | sort -n | head -1`

if [ -z "$WINDOW_ID" ] ; then
        echo "can't find window with name 'Mozilla Firefox'"
        exit 1
else
        echo "working on firefox window $WINDOW_ID"
fi

xdotool search firefox

xdotool focus $WINDOW_ID
xdotool key F11

ping_browser

# try to deduce inside area of window

DUMP_ID=`xwininfo -display $DISPLAY -id $WINDOW_ID -tree | grep 0x | grep -v ${W}x${H} | grep ${W}x | head -1 | sed 's/^ *//' | cut -d' ' -f1`

if [ -z "$DUMP_ID" ] ; then
        echo "can't find inside area of window. Using whole browser!"
        DUMP_ID=$WINDOW_ID
fi

echo "saving window $DUMP_ID to $FILE"
if [ ! -z "$RESIZE" ] ; then
        RESIZE="-geometry $RESIZE"
fi

import -window $DUMP_ID $RESIZE $FILE

if [ "$DEBUG" == 1 ] ; then
        echo -n "press enter to exit! "
        read
fi
1	dpavlin	1	#!/bin/sh
2			#
3			# This script is based on work of mendel from
4			# http://www.lafferty.ca/software/webshots/webshot
5			# webshot: hackish thing to take a picture of a whole webpage.
6			#
7			# idea from sippey (http://www.sippey.com/2004/blog-snaps/)
8			# inspiration from bradfitz
9			#
10			# usage: webshot url filename
11			#
12			# JPEG seems to be the best format.
13			#
14			# Modified by Dobrica Pavlinusic 2004-07-25
15			# - include signal grabbing
16			# - dynamic vncserver display
17			# - correct WindowID parsing
18			# - wait for browser to load full page (useful on slow links)
19			#
20			# It will need follwing external utilities:
21			# vncserver
22			# mozilla-firefox
23			# HEAD (part or LWP perl library, it can be commented out)
24			# nc (netcat, to simulate web server)
25			# xwininfo (standard part of X-clients)
26			# import (part of ImageMagick package)
27	dpavlin	62	# xdotool (much more useful than xwit for this purpose)
28	dpavlin	1	#
29			# It will create snapshot of following size
30	dpavlin	2	W=1024
31			H=768
32	dpavlin	1
33	dpavlin	2	# and then resize it to (comment out to disable)
34			#RESIZE=200x
35	dpavlin	1
36	dpavlin	64	# turn locally visible server to watch progress
37	dpavlin	61	DEBUG=0
38			test ! -z "$3" && DEBUG=1
39
40			# wait for $WAIT seconds for page to load
41			WAIT=30
42
43	dpavlin	1	# some configurable paths
44			BROWSER=mozilla-firefox
45	dpavlin	60	XSERVER=Xvfb
46	dpavlin	1	URL=$1
47			FILE=${2:-screenshot.jpg}
48
49			FRAMESET="`mktemp`.html"
50	dpavlin	60	PROFILE_DIR=`mktemp -d`
51	dpavlin	1	PORT=8888
52
53	dpavlin	61	LOCAL_DISPLAY=$DISPLAY
54
55	dpavlin	2	if [ -z "$URL" ] ; then
56			echo "usage: $0 http://url.to.capture [screenshot.jpg]"
57			exit 1
58			fi
59
60	dpavlin	62	if [ -z "`which xdotool`" ] ; then
61			echo "$0 need xdotool from http://www.semicomplete.com/blog/projects/xdotool/"
62	dpavlin	2	exit 1
63			fi
64
65	dpavlin	60	if [ -z "`which $XSERVER`" ] ; then
66			echo "$0 really need $XSERVER to operate. please install it."
67			exit 1
68			fi
69
70	dpavlin	61	while netstat -ln \| grep ":$PORT " >/dev/null ; do
71			PORT=`expr $PORT + 1`
72			done
73
74			echo "### using port $PORT"
75
76	dpavlin	1	echo -n "testing URL $URL "
77			if HEAD $URL >/dev/null ; then
78			echo "ok"
79			else
80			echo "FAILED"
81			exit 1
82			fi
83
84	dpavlin	64	DISPLAY_PORT=6042
85	dpavlin	1
86	dpavlin	64	while netstat -ln \| grep ":$DISPLAY_PORT " >/dev/null ; do
87			DISPLAY_PORT=`expr $DISPLAY_PORT + 1`
88			done
89	dpavlin	1
90	dpavlin	64	DISPLAY=:`expr $DISPLAY_PORT - 6000`
91	dpavlin	60
92	dpavlin	64	echo "using DISPLAY=$DISPLAY"
93
94	dpavlin	61	if [ "$DEBUG" == 1 ] ; then
95	dpavlin	62	echo "using locally visible debug server on $LOCAL_DISPLAY"
96	dpavlin	64	DISPLAY=$LOCAL_DISPLAY Xephyr -ac -screen ${W}x${H}x24 $DISPLAY 2>/dev/null &
97	dpavlin	61	else
98	dpavlin	62	echo "starting $XSERVER"
99	dpavlin	61	$XSERVER -ac -screen 0 ${W}x${H}x24 $DISPLAY 2>/dev/null &
100			fi
101
102	dpavlin	64	if [ -z "$!" ] ; then
103			echo "ABORT: can't start X server!"
104			exit
105			fi
106	dpavlin	60	XSERVER_PID=$!
107	dpavlin	64
108	dpavlin	62	echo "X server pid $XSERVER_PID"
109	dpavlin	60
110			function kill_x_server() {
111	dpavlin	61	echo "Killing server $XSERVER_PID"
112	dpavlin	60	kill $XSERVER_PID
113	dpavlin	1	rm -f $FRAMESET
114	dpavlin	60	rm -fr $PROFILE_DIR
115	dpavlin	62	trap '' EXIT
116	dpavlin	61	exit 1
117	dpavlin	1	}
118	dpavlin	61	trap 'kill_x_server' INT QUIT TERM SEGV EXIT
119	dpavlin	1
120			# create frameset to load site and after site is loaded trigger this script
121			cat > $FRAMESET <<END_OF_FRAMESET
122			<html>
123			<head>
124			<frameset rows="0,*" border=0 frameborder=no framespacing=0 onload="window.frames['trigger_frame'].window.location.replace('http://localhost:$PORT');">
125			<frame src="about:blank" name="trigger_frame" scrolling=no marginwidth=0 marginheight=0>
126			<frame src="$URL" name="page_frame" scrolling=no>
127			</frameset>
128			</head>
129			</html>
130			END_OF_FRAMESET
131
132			echo "Using frameset html $FRAMESET"
133
134	dpavlin	60	echo "preview available with: xwd -display $DISPLAY -root \| xwud"
135	dpavlin	1
136	dpavlin	60	echo "making 'Screenshot' profile in $PROFILE_DIR"
137	dpavlin	61	$BROWSER -CreateProfile "Screenshot $PROFILE_DIR" 2>/dev/null \| grep Success
138	dpavlin	1
139			echo "launching browser $BROWSER with $URL"
140	dpavlin	62	$BROWSER -P Screenshot -width $W -height $H -safemode $FRAMESET 2>/dev/null &
141	dpavlin	60	BROWSER_PID=$!
142	dpavlin	1
143	dpavlin	61	function kill_browser() {
144			echo "Killing browser $BROWSER_PID"
145			kill $BROWSER_PID
146			echo "Killing server $XSERVER_PID"
147			kill $XSERVER_PID
148			rm -f $FRAMESET
149			rm -fr $PROFILE_DIR
150	dpavlin	62	trap '' EXIT
151	dpavlin	61	exit 1
152			}
153			trap 'kill_browser' INT QUIT TERM SEGV EXIT
154	dpavlin	1
155	dpavlin	62	function ping_browser() {
156			echo -n "ping browser"
157			while ! ( $BROWSER -remote "ping();" 2>&1 ) >/dev/null ; do
158			RID=`xdotool search Restore 2>/dev/null`
159			if [ ! -z "$RID" ] ; then
160			echo -n "Esc"
161			xdotool focus $RID
162			xdotool key Escape
163			sleep 1
164			fi
165			echo -n "."
166			sleep 1
167			done
168			echo " OK"
169			}
170
171			ping_browser
172
173	dpavlin	63	$BROWSER -remote "openURL($FRAMESET)" 2>/dev/null
174	dpavlin	62
175	dpavlin	61	echo "waiting for on_load event from browser $BROWSER_PID for ${WAIT}s"
176
177	dpavlin	1	# there is hard-coded limit here:
178	dpavlin	61	# we will wait $WAIT sec for page to load and render
179	dpavlin	1
180	dpavlin	61	echo -e "HTTP/1.0 304 Not modified\r\n\r\n" \| nc -l -w $WAIT -p $PORT >/dev/null \|\| echo "Timeout after $WAIT sec!"
181	dpavlin	1
182			# get Mozilla Firefox window id (for resize)
183	dpavlin	60	WINDOW_ID=`xwininfo -display $DISPLAY -root -tree \| grep gecko \| cut -d\" -f1 \| sort -n \| head -1`
184	dpavlin	1
185			if [ -z "$WINDOW_ID" ] ; then
186			echo "can't find window with name 'Mozilla Firefox'"
187			exit 1
188	dpavlin	62	else
189			echo "working on firefox window $WINDOW_ID"
190	dpavlin	1	fi
191
192	dpavlin	62	xdotool search firefox
193	dpavlin	1
194	dpavlin	62	xdotool focus $WINDOW_ID
195			xdotool key F11
196	dpavlin	1
197	dpavlin	62	ping_browser
198	dpavlin	1
199			# try to deduce inside area of window
200
201			DUMP_ID=`xwininfo -display $DISPLAY -id $WINDOW_ID -tree \| grep 0x \| grep -v ${W}x${H} \| grep ${W}x \| head -1 \| sed 's/^ *//' \| cut -d' ' -f1`
202
203			if [ -z "$DUMP_ID" ] ; then
204			echo "can't find inside area of window. Using whole browser!"
205			DUMP_ID=$WINDOW_ID
206			fi
207
208			echo "saving window $DUMP_ID to $FILE"
209	dpavlin	2	if [ ! -z "$RESIZE" ] ; then
210			RESIZE="-geometry $RESIZE"
211			fi
212	dpavlin	60
213	dpavlin	2	import -window $DUMP_ID $RESIZE $FILE
214	dpavlin	1
215	dpavlin	62	if [ "$DEBUG" == 1 ] ; then
216			echo -n "press enter to exit! "
217			read
218			fi