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)
# xwit (not so standard command, but very useful)
# import (part of ImageMagick package)
#
# It will create snapshot of following size
W=1024
H=768

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

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

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

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

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

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

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

echo "starting X server $XSERVER"

export DISPLAY=:42

$XSERVER -ac -screen 0 ${W}x${H}x24 $DISPLAY &
# debug server
#Xnest -display :0 -ac -geometry ${W}x${H} -depth 24 $DISPLAY &

XSERVER_PID=$!
echo "using pid $xserver_pid for X server"

function kill_x_server() {
        kill $BROWSER_PID
        kill $XSERVER_PID
        rm -f $FRAMESET
        rm -fr $PROFILE_DIR
}
trap 'echo "QUIT!" >&2; kill_x_server; exit 1' INT QUIT TERM SEGV

# 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" | grep Success

echo "launching browser $BROWSER with $URL"
$BROWSER -P Screenshot -width $W -height $H $FRAMESET &
BROWSER_PID=$!

echo "staring web server (wait for page to load)"

# there is hard-coded limit here:
# we will wait 60sec for page to load and render

echo -e "HTTP/1.0 304 Not modified\r\n\r\n" | nc -l -w 60 -p $PORT >/dev/null

# 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
fi

# move window to foreground
$XWIT -display $DISPLAY -id $WINDOW_ID -pop

echo "resizing window $WINDOW_ID to maximum size"
$XWIT -display $DISPLAY -id $WINDOW_ID -move 0 0
$XWIT -display $DISPLAY -id $WINDOW_ID -resize $W $H


echo -n "wating for browser ping..."
while ! ( $BROWSER -remote "ping();" 2>&1 ) >/dev/null ; do
        echo -n "."
        sleep 1
done
echo

# 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

kill_x_server

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			# xwit (not so standard command, but very useful)
27			# import (part of ImageMagick package)
28			#
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			# some configurable paths
37			BROWSER=mozilla-firefox
38	dpavlin	60	XSERVER=Xvfb
39	dpavlin	2	XWIT=xwit
40	dpavlin	1	URL=$1
41			FILE=${2:-screenshot.jpg}
42
43			FRAMESET="`mktemp`.html"
44	dpavlin	60	PROFILE_DIR=`mktemp -d`
45	dpavlin	1	PORT=8888
46
47	dpavlin	2	if [ -z "$URL" ] ; then
48			echo "usage: $0 http://url.to.capture [screenshot.jpg]"
49			exit 1
50			fi
51
52			if [ -z "`which $XWIT`" ] ; then
53	dpavlin	60	echo "$0 really need $XWIT to operate. please install it."
54	dpavlin	2	exit 1
55			fi
56
57	dpavlin	60	if [ -z "`which $XSERVER`" ] ; then
58			echo "$0 really need $XSERVER to operate. please install it."
59			exit 1
60			fi
61
62	dpavlin	1	echo -n "testing URL $URL "
63			if HEAD $URL >/dev/null ; then
64			echo "ok"
65			else
66			echo "FAILED"
67			exit 1
68			fi
69
70	dpavlin	60	echo "starting X server $XSERVER"
71	dpavlin	1
72	dpavlin	60	export DISPLAY=:42
73	dpavlin	1
74	dpavlin	60	$XSERVER -ac -screen 0 ${W}x${H}x24 $DISPLAY &
75			# debug server
76			#Xnest -display :0 -ac -geometry ${W}x${H} -depth 24 $DISPLAY &
77
78			XSERVER_PID=$!
79			echo "using pid $xserver_pid for X server"
80
81			function kill_x_server() {
82			kill $BROWSER_PID
83			kill $XSERVER_PID
84	dpavlin	1	rm -f $FRAMESET
85	dpavlin	60	rm -fr $PROFILE_DIR
86	dpavlin	1	}
87	dpavlin	60	trap 'echo "QUIT!" >&2; kill_x_server; exit 1' INT QUIT TERM SEGV
88	dpavlin	1
89			# create frameset to load site and after site is loaded trigger this script
90			cat > $FRAMESET <<END_OF_FRAMESET
91			<html>
92			<head>
93			<frameset rows="0,*" border=0 frameborder=no framespacing=0 onload="window.frames['trigger_frame'].window.location.replace('http://localhost:$PORT');">
94			<frame src="about:blank" name="trigger_frame" scrolling=no marginwidth=0 marginheight=0>
95			<frame src="$URL" name="page_frame" scrolling=no>
96			</frameset>
97			</head>
98			</html>
99			END_OF_FRAMESET
100
101			echo "Using frameset html $FRAMESET"
102
103	dpavlin	60	echo "preview available with: xwd -display $DISPLAY -root \| xwud"
104	dpavlin	1
105	dpavlin	60	echo "making 'Screenshot' profile in $PROFILE_DIR"
106			$BROWSER -CreateProfile "Screenshot $PROFILE_DIR" \| grep Success
107	dpavlin	1
108			echo "launching browser $BROWSER with $URL"
109	dpavlin	60	$BROWSER -P Screenshot -width $W -height $H $FRAMESET &
110			BROWSER_PID=$!
111	dpavlin	1
112			echo "staring web server (wait for page to load)"
113
114			# there is hard-coded limit here:
115			# we will wait 60sec for page to load and render
116
117			echo -e "HTTP/1.0 304 Not modified\r\n\r\n" \| nc -l -w 60 -p $PORT >/dev/null
118
119			# get Mozilla Firefox window id (for resize)
120	dpavlin	60	WINDOW_ID=`xwininfo -display $DISPLAY -root -tree \| grep gecko \| cut -d\" -f1 \| sort -n \| head -1`
121	dpavlin	1
122			if [ -z "$WINDOW_ID" ] ; then
123			echo "can't find window with name 'Mozilla Firefox'"
124			exit 1
125			fi
126
127	dpavlin	2	# move window to foreground
128			$XWIT -display $DISPLAY -id $WINDOW_ID -pop
129	dpavlin	1
130			echo "resizing window $WINDOW_ID to maximum size"
131	dpavlin	2	$XWIT -display $DISPLAY -id $WINDOW_ID -move 0 0
132			$XWIT -display $DISPLAY -id $WINDOW_ID -resize $W $H
133	dpavlin	1
134
135			echo -n "wating for browser ping..."
136			while ! ( $BROWSER -remote "ping();" 2>&1 ) >/dev/null ; do
137			echo -n "."
138			sleep 1
139			done
140			echo
141
142			# try to deduce inside area of window
143
144			DUMP_ID=`xwininfo -display $DISPLAY -id $WINDOW_ID -tree \| grep 0x \| grep -v ${W}x${H} \| grep ${W}x \| head -1 \| sed 's/^ *//' \| cut -d' ' -f1`
145
146			if [ -z "$DUMP_ID" ] ; then
147			echo "can't find inside area of window. Using whole browser!"
148			DUMP_ID=$WINDOW_ID
149			fi
150
151			echo "saving window $DUMP_ID to $FILE"
152	dpavlin	2	if [ ! -z "$RESIZE" ] ; then
153			RESIZE="-geometry $RESIZE"
154			fi
155	dpavlin	60
156	dpavlin	2	import -window $DUMP_ID $RESIZE $FILE
157	dpavlin	1
158	dpavlin	60	kill_x_server
159	dpavlin	1