X-Git-Url: http://git.hungrycats.org/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=hacks%2Fwebcollage;h=e606655ec8e6b9db09162499ef1bafff151ca45b;hb=9c9d475ff889ed8be02e8ce8c17da28b93278fca;hp=bb3c66ef6ded01d0434310ae4c87528fb52b21b9;hpb=a94197e76a5dea5cb60542840809d6c20d0abbf3;p=xscreensaver diff --git a/hacks/webcollage b/hacks/webcollage index bb3c66ef..e606655e 100755 --- a/hacks/webcollage +++ b/hacks/webcollage @@ -1,6 +1,6 @@ #!/usr/bin/perl -w # -# webcollage, Copyright (c) 1999-2002 by Jamie Zawinski +# webcollage, Copyright (c) 1999-2004 by Jamie Zawinski # This program decorates the screen with random images from the web. # One satisfied customer described it as "a nonstop pop culture brainbath." # @@ -12,10 +12,31 @@ # software for any purpose. It is provided "as is" without express or # implied warranty. + # To run this as a display mode with xscreensaver, add this to `programs': # -# default-n: webcollage -root \n\ -# default-n: webcollage -root -filter 'vidwhacker -stdin -stdout' \n\ +# webcollage -root +# webcollage -root -filter 'vidwhacker -stdin -stdout' +# +# +# You can see this in action at http://www.jwz.org/webcollage/ -- +# it auto-reloads about once a minute. To make a page similar to +# that on your own system, do this: +# +# webcollage -size '800x600' -imagemap $HOME/www/webcollage/index +# +# +# If you have the "driftnet" program installed, webcollage can display a +# collage of images sniffed off your local ethernet, instead of pulled out +# of search engines: in that way, your screensaver can display the images +# that your co-workers are downloading! +# +# Driftnet is available here: http://www.ex-parrot.com/~chris/driftnet/ +# Use it like so: +# +# webcollage -root -driftnet +# +# Driftnet is the Unix implementation of the MacOS "EtherPEG" program. require 5; @@ -33,34 +54,55 @@ require POSIX; use Fcntl ':flock'; # import LOCK_* constants use POSIX qw(strftime); +use bytes; # Larry can take Unicode and shove it up his ass sideways. + # Perl 5.8.0 causes us to start getting incomprehensible + # errors about UTF-8 all over the place without this. + my $progname = $0; $progname =~ s@.*/@@g; -my $version = q{ $Revision: 1.87 $ }; $version =~ s/^[^0-9]+([0-9.]+).*$/$1/; -my $copyright = "WebCollage $version, Copyright (c) 1999-2002" . +my $version = q{ $Revision: 1.114 $ }; $version =~ s/^[^0-9]+([0-9.]+).*$/$1/; +my $copyright = "WebCollage $version, Copyright (c) 1999-2004" . " Jamie Zawinski \n" . - " http://www.jwz.org/xscreensaver/\n"; - - - -my @search_methods = ( 30, "imagevista", \&pick_from_alta_vista_images, - 28, "altavista", \&pick_from_alta_vista_text, - 18, "yahoorand", \&pick_from_yahoo_random_link, - 14, "googleimgs", \&pick_from_google_images, - 2, "yahoonews", \&pick_from_yahoo_news_text, - 8, "lycos", \&pick_from_lycos_text, - - # Hotbot gives me "no matches" just about every time. - # Then I try the same URL again, and it works. I guess - # it caches searches, and webcollage always busts its - # cache and time out? Or it just sucks. - # 0, "hotbot", \&pick_from_hotbot_text, + " http://www.jwz.org/webcollage/\n"; + + + +my @search_methods = ( 72, "altavista", \&pick_from_alta_vista_random_link, + 10, "livejournal", \&pick_from_livejournal_images, + 10, "yahoorand", \&pick_from_yahoo_random_link, + 8, "yahoonews", \&pick_from_yahoo_news_text, + + # Alta Vista has a new "random link" URL now. + # They added it specifically to better support webcollage! + # That was super cool of them. This is how we used to do + # it, before: + # + # 0, "avimages", \&pick_from_alta_vista_images, + # 0, "avtext", \&pick_from_alta_vista_text, + + # Google asked (nicely) for me to stop searching them. + # I asked them to add a "random link" url. They said + # "that would be easy, we'll think about it" and then + # never wrote back. Booo Google! Booooo! + # + # 0, "googlenums", \&pick_from_google_image_numbers, + # 0, "googleimgs", \&pick_from_google_images, + + # I suspect Hotbot is actually the same search engine + # data as Lycos. + # + # 0, "hotbot", \&pick_from_hotbot_text, + + # Eh, Lycos sucks anyway. + # 0, "lycos", \&pick_from_lycos_text, ); # programs we can use to write to the root window (tried in ascending order.) # my @root_displayers = ( + "xscreensaver-getimage -root -file", "chbg -once -xscreensaver -max_size 100", - "xv -root -quit -viewonly -noresetroot -quick24 -rmode 5" . + "xv -root -quit -viewonly +noresetroot -quick24 -rmode 5" . " -rfg black -rbg black", "xli -quiet -onroot -center -border black", "xloadimage -quiet -onroot -center -border black", @@ -106,19 +148,20 @@ my $opacity = 0.85; # my %poisoners = ( "die.net" => 1, # 'l33t h4ck3r d00dz. - "genforum.genealogy.com" => 1, # Cluttering altavista with human names. - "rootsweb.com" => 1, # Cluttering altavista with human names. + "genforum.genealogy.com" => 1, # Cluttering avtext with human names. + "rootsweb.com" => 1, # Cluttering avtext with human names. "akamai.net" => 1, # Lots of sites have their images on Akamai. - # But those are pretty much all banners. + "akamaitech.net" => 1, # But those are pretty much all banners. # Since Akamai is super-expensive, let's # go out on a limb and assume that all of # their customers are rich-and-boring. - "bartleby.com" => 1, # Dictionary, cluttering altavista. - "encyclopedia.com" => 1, # Dictionary, cluttering altavista. - "onlinedictionary.datasegment.com" => 1, # Dictionary, cluttering altavista. - "hotlinkpics.com" => 1, # Porn site that has poisoned imagevista + "bartleby.com" => 1, # Dictionary, cluttering avtext. + "encyclopedia.com" => 1, # Dictionary, cluttering avtext. + "onlinedictionary.datasegment.com" => 1, # Dictionary, cluttering avtext. + "hotlinkpics.com" => 1, # Porn site that has poisoned avimages # (I don't see how they did it, though!) "alwayshotels.com" => 1, # Poisoned Lycos pretty heavily. + "nextag.com" => 1, # Poisoned Alta Vista real good. ); @@ -137,6 +180,8 @@ my %warningless_sites = ( "yimg.com" => 1, # This is where dailynews.yahoo.com stores "eimg.com" => 1, # its images, so pick_from_yahoo_news_text() # hits this every time. + + "driftnet" => 1, # builtin... ); @@ -177,6 +222,13 @@ my $min_gif_area = (120 * 120); my $no_output_p = 0; my $urls_only_p = 0; +my $imagemap_base = undef; + +my @pids_to_kill = (); # forked pids we should kill when we exit, if any. + +my $driftnet_magic = 'driftnet'; +my $driftnet_dir = undef; +my $default_driftnet_cmd = "driftnet -a -m 100"; my $wordlist; @@ -218,6 +270,11 @@ sub get_document_1 { my ($url_proto, $dummy, $serverstring, $path) = split(/\//, $url, 4); $path = "" unless $path; + if (!$url_proto || !$serverstring) { + LOG (($verbose_net || $verbose_load), "unparsable URL: $url"); + return (); + } + my ($them,$port) = split(/:/, $serverstring); $port = 80 unless $port; @@ -225,6 +282,7 @@ sub get_document_1 { my $port2 = $port; if ($http_proxy) { $serverstring = $http_proxy if $http_proxy; + $serverstring =~ s@^[a-z]+://@@; ($them2,$port2) = split(/:/, $serverstring); $port2 = 80 unless $port2; } @@ -270,7 +328,9 @@ sub get_document_1 { my $cookie = $cookies{$them}; my $user_agent = "$progname/$version"; - if ($url =~ m@^http://www\.altavista\.com/@) { + + if ($url =~ m@^http://www\.altavista\.com/@ || + $url =~ m@^http://random\.yahoo\.com/@) { # block this, you turkeys. $user_agent = "Mozilla/4.76 [en] (X11; U; Linux 2.2.16-22 i686; Nav)"; } @@ -293,6 +353,11 @@ sub get_document_1 { print S $hdrs; my $http = || ""; + # Kludge: the Yahoo Random Link is now returning as its first + # line "Status: 301" instead of "HTTP/1.0 301 Found". Fix it... + # + $http =~ s@^Status:\s+(\d+)\b@HTTP/1.0 $1@i; + $_ = $http; s/[\r\n]+$//s; LOG ($verbose_http, " <== $_"); @@ -348,6 +413,10 @@ sub get_document { my ( $url, $referer, $timeout ) = @_; my $start = time; + if (defined($referer) && $referer eq $driftnet_magic) { + return get_driftnet_file ($url); + } + my $orig_url = $url; my $loop_count = 0; my $max_loop_count = 4; @@ -374,6 +443,7 @@ sub get_document { if ( $http =~ m@^HTTP/[0-9.]+ 30[123]@ ) { $_ = $head; + my ( $location ) = m@^location:[ \t]*(.*)$@im; if ( $location ) { $location =~ s/[\r\n]$//; @@ -492,9 +562,9 @@ sub pick_image_from_body { # randomly from the set of images on the web. All the logic here for # rejecting some images is really a set of heuristics for rejecting # images that are not really images: for rejecting *text* that is in - # GIF/JPEG form. I don't want text, I want pictures, and I want the - # content of the pictures to be randomly selected from among all the - # available content. + # GIF/JPEG/PNG form. I don't want text, I want pictures, and I want + # the content of the pictures to be randomly selected from among all + # the available content. # # So, filtering out "dirty" pictures by looking for "dirty" keywords # would be wrong: dirty pictures exist, like it or not, so webcollage @@ -552,7 +622,7 @@ sub pick_image_from_body { } elsif ( m/^(img|a) .*(src|href) ?= ?\"? ?(.*?)[ >\"]/io ) { - my $was_inline = ( "$1" eq "a" || "$1" eq "A" ); + my $was_inline = (! ( "$1" eq "a" || "$1" eq "A" )); my $link = $3; my ( $width ) = m/width ?=[ \"]*(\d+)/oi; my ( $height ) = m/height ?=[ \"]*(\d+)/oi; @@ -565,8 +635,7 @@ sub pick_image_from_body { } elsif ( ! m@^[^/:?]+:@ ) { $_ = "$base$link"; s@/\./@/@g; - while (s@/\.\./@/@g) { - } + 1 while (s@/[^/]+/\.\./@/@g); } # skip non-http @@ -575,7 +644,7 @@ sub pick_image_from_body { } # skip non-image - if ( ! m@[.](gif|jpg|jpeg|pjpg|pjpeg)$@io ) { + if ( ! m@[.](gif|jpg|jpeg|pjpg|pjpeg|png)$@io ) { next; } @@ -624,9 +693,9 @@ sub pick_image_from_body { $urls[++$#urls] = $url; $unique_urls{$url}++; - # jpegs are preferable to gifs. + # JPEGs are preferable to GIFs and PNGs. $_ = $url; - if ( ! m@[.]gif$@io ) { + if ( ! m@[.](gif|png)$@io ) { $urls[++$#urls] = $url; } @@ -742,6 +811,24 @@ sub url_unquote { return $s; } +sub html_quote { + my ($s) = @_; + $s =~ s/&/&/gi; + $s =~ s//>/gi; + $s =~ s/\"/"/gi; + return $s; +} + +sub html_unquote { + my ($s) = @_; + $s =~ s/</ 1) { + $search_url .= "&start=" . $page*$num; # page number + $search_url .= "&num=" . $num; #images per page + } + + my ($search_hit_count, @subpages) = + pick_from_search_engine ($timeout, $search_url, $words); + + my @candidates = (); + my %referers; + foreach my $u (@subpages) { + next unless ($u =~ m@imgres\?imgurl@i); # All pics start with this + next if ($u =~ m@[/.]google\.com\b@i); # skip google builtins + + if ($u =~ m@^/imgres\?imgurl=(.*?)\&imgrefurl=(.*?)\&@) { + my $ref = $2; + my $img = "http://$1"; + + LOG ($verbose_filter, " candidate: $ref"); + push @candidates, $img; + $referers{$img} = $ref; + } + } + + @candidates = depoison (@candidates); + return () if ($#candidates < 0); + my $i = int(rand($#candidates+1)); + my $img = $candidates[$i]; + my $ref = $referers{$img}; + + LOG ($verbose_load, "picked image " . ($i+1) . ": $img (on $ref)"); + return ($ref, $img); +} + + + +############################################################################ +# +# Pick images by feeding random words into Alta Vista Text Search +# +############################################################################ + + +my $alta_vista_url = "http://www.altavista.com/web/results" . + "?pg=aq" . + "&aqmode=s" . + "&filetype=html" . + "&sc=on" . # "site collapse" + "&nbq=50" . + "&aqo="; + +# avtext sub pick_from_alta_vista_text { my ( $timeout ) = @_; - my $words = random_words(1); + my $words = random_words(0); my $page = (int(rand(9)) + 1); my $search_url = $alta_vista_url . $words; @@ -1119,8 +1307,7 @@ sub pick_from_alta_vista_text { # onMouseOver to make it look like they're not! Well, it makes it # easier for us to identify search results... # - next unless ($u =~ - m@^/r\?ck_sm=[a-zA-Z0-9]+\&ref=[a-zA-Z0-9]+\&uid=[a-zA-Z0-9]+\&r=(.*)@); + next unless ($u =~ m@^/r.*\&r=([^&]+).*@); $u = url_unquote($1); LOG ($verbose_filter, " candidate: $u"); @@ -1139,23 +1326,30 @@ sub pick_from_alta_vista_text { # ############################################################################ -my $hotbot_search_url = "http://hotbot.lycos.com/" . - "?SM=SC" . - "&DV=0" . - "&LG=any" . - "&FVI=1" . - "&DC=100" . - "&DE=0" . - "&SQ=1" . - "&TR=13" . - "&AM1=MC" . - "&MT="; +my $hotbot_search_url =("http://hotbot.lycos.com/default.asp" . + "?ca=w" . + "&descriptiontype=0" . + "&imagetoggle=1" . + "&matchmode=any" . + "&nummod=2" . + "&recordcount=50" . + "&sitegroup=1" . + "&stem=1" . + "&cobrand=undefined" . + "&query="); sub pick_from_hotbot_text { my ( $timeout ) = @_; - my $words = random_words(0); - my $search_url = $hotbot_search_url . $words; + $last_search = $hotbot_search_url; # for warnings + + # lycos seems to always give us back dictionaries and word lists if + # we search for more than one word... + # + my $words = random_word(); + + my $start = int(rand(8)) * 10 + 1; + my $search_url = $hotbot_search_url . $words . "&first=$start&page=more"; my ($search_hit_count, @subpages) = pick_from_search_engine ($timeout, $search_url, $words); @@ -1164,7 +1358,7 @@ sub pick_from_hotbot_text { foreach my $u (@subpages) { # Hotbot plays redirection games too - next unless ($u =~ m@^/director.asp\?target=([^&]+)@); + next unless ($u =~ m@/director.asp\?.*\btarget=([^&]+)@); $u = url_decode($1); LOG ($verbose_filter, " candidate: $u"); @@ -1183,17 +1377,24 @@ sub pick_from_hotbot_text { # ############################################################################ -my $lycos_search_url = "http://lycospro.lycos.com/srchpro/" . +my $lycos_search_url = "http://search.lycos.com/default.asp" . "?lpv=1" . - "&t=any" . + "&loc=searchhp" . + "&tab=web" . "&query="; sub pick_from_lycos_text { my ( $timeout ) = @_; - my $words = random_words(0); + $last_search = $lycos_search_url; # for warnings + + # lycos seems to always give us back dictionaries and word lists if + # we search for more than one word... + # + my $words = random_word(); + my $start = int(rand(8)) * 10 + 1; - my $search_url = $lycos_search_url . $words . "&start=$start"; + my $search_url = $lycos_search_url . $words . "&first=$start&page=more"; my ($search_hit_count, @subpages) = pick_from_search_engine ($timeout, $search_url, $words); @@ -1201,10 +1402,12 @@ sub pick_from_lycos_text { my @candidates = (); foreach my $u (@subpages) { - # Lycos plays exact the same redirection game as hotbot. - # Note that "id=0" is used for internal advertising links, - # and 1+ are used for search results. - next unless ($u =~ m@^http://click.hotbot.com/director.asp\?id=[1-9]\d*&target=([^&]+)@); + # Lycos plays redirection games. + next unless ($u =~ m@^http://click.lycos.com/director.asp + .* + \btarget=([^&]+) + .* + @x); $u = url_decode($1); LOG ($verbose_filter, " candidate: $u"); @@ -1223,14 +1426,23 @@ sub pick_from_lycos_text { # ############################################################################ -my $yahoo_news_url = "http://search.news.yahoo.com/search/news_photos?" . - "&z=&n=100&o=o&2=&3=&p="; +my $yahoo_news_url = "http://search.news.yahoo.com/search/news" . + "?a=1" . + "&c=news_photos" . + "&s=-%24s%2C-date" . + "&n=100" . + "&o=o" . + "&2=" . + "&3=" . + "&p="; # yahoonews sub pick_from_yahoo_news_text { my ( $timeout ) = @_; - my $words = random_words(1); + $last_search = $yahoo_news_url; # for warnings + + my $words = random_words(0); my $search_url = $yahoo_news_url . $words; my ($search_hit_count, @subpages) = @@ -1251,6 +1463,157 @@ sub pick_from_yahoo_news_text { } + +############################################################################ +# +# Pick images from LiveJournal's list of recently-posted images. +# +############################################################################ + +my $livejournal_img_url = "http://www.livejournal.com/stats/latest-img.bml"; + +# livejournal +sub pick_from_livejournal_images { + my ( $timeout ) = @_; + + $last_search = $livejournal_img_url; # for warnings + + my ( $base, $body ) = get_document ($livejournal_img_url, undef, $timeout); + return () unless $body; + + my @candidates = (); + + $body =~ s/\n/ /gs; + $body =~ s/() { $body .= $_; } + close IN || error ("$id: $file: $!"); + unlink ($file) || error ("$id: $file: rm: $!"); + return ($id, $body); +} + + +sub spawn_driftnet { + my ($cmd) = @_; + + # make a directory to use. + while (1) { + my $tmp = $ENV{TEMPDIR} || "/tmp"; + $driftnet_dir = sprintf ("$tmp/driftcollage-%08x", rand(0xffffffff)); + LOG ($verbose_exec, "mkdir $driftnet_dir"); + last if mkdir ($driftnet_dir, 0700); + } + + if (! ($cmd =~ m/\s/)) { + # if the command didn't have any arguments in it, then it must be just + # a pointer to the executable. Append the default args to it. + my $dargs = $default_driftnet_cmd; + $dargs =~ s/^[^\s]+//; + $cmd .= $dargs; + } + + # point the driftnet command at our newly-minted private directory. + # + $cmd .= " -d $driftnet_dir"; + $cmd .= ">/dev/null" unless ($verbose_exec); + + my $pid = fork(); + if ($pid < 0) { error ("fork: $!\n"); } + if ($pid) { + # parent fork + push @pids_to_kill, $pid; + LOG ($verbose_exec, "forked for \"$cmd\""); + } else { + # child fork + nontrapping_system ($cmd) || error ("exec: $!"); + } + + # wait a bit, then make sure the process actually started up. + # + sleep (1); + error ("pid $pid failed to start \"$cmd\"") + unless (1 == kill (0, $pid)); +} ############################################################################ @@ -1263,7 +1626,6 @@ sub pick_from_yahoo_news_text { # Picks a random image on a random page, and returns two URLs: # the page containing the image, and the image. # Returns () if nothing found this time. -# Uses the url-randomizer 1 time in 5, else the image randomizer. # sub pick_image { @@ -1487,6 +1849,12 @@ sub save_recent_url { $_ = $url; my ($site) = m@^http://([^ \t\n\r/:]+)@; + return unless defined ($site); + + if ($base eq $driftnet_magic) { + $site = $driftnet_magic; + @recent_images = (); + } my $done = 0; foreach (@recent_images) { @@ -1571,10 +1939,12 @@ sub jpeg_size { while (ord($ch) != 0xDA && $i < $L) { # Find next marker, beginning with 0xFF. while (ord($ch) != 0xFF) { + return () if (length($body) <= $i); $ch = substr($body, $i, 1); $i++; } # markers can be padded with any number of 0xFF. while (ord($ch) == 0xFF) { + return () if (length($body) <= $i); $ch = substr($body, $i, 1); $i++; } @@ -1586,6 +1956,7 @@ sub jpeg_size { ($marker != 0xC4) && ($marker != 0xCC)) { # it's a SOFn marker $i += 3; + return () if (length($body) <= $i); my $s = substr($body, $i, 4); $i += 4; my ($a,$b,$c,$d) = unpack("C"x4, $s); return (($c<<8|$d), ($a<<8|$b)); @@ -1593,6 +1964,7 @@ sub jpeg_size { } else { # We must skip variables, since FFs in variable names aren't # valid JPEG markers. + return () if (length($body) <= $i); my $s = substr($body, $i, 2); $i += 2; my ($c1, $c2) = unpack ("C"x2, $s); my $length = ($c1 << 8) | $c2; @@ -1603,14 +1975,29 @@ sub jpeg_size { return (); } -# Given the raw body of a GIF or JPEG document, returns the dimensions of -# the image. +# Given the raw body of a PNG document, returns the dimensions of the image. +# +sub png_size { + my ($body) = @_; + return () unless ($body =~ m/^\211PNG\r/); + my ($bits) = ($body =~ m/^.{12}(.{12})/s); + return () unless defined ($bits); + return () unless ($bits =~ /^IHDR/); + my ($ign, $w, $h) = unpack("a4N2", $bits); + return ($w, $h); +} + + +# Given the raw body of a GIF, JPEG, or PNG document, returns the dimensions +# of the image. # sub image_size { my ($body) = @_; my ($w, $h) = gif_size ($body); if ($w && $h) { return ($w, $h); } - return jpeg_size ($body); + ($w, $h) = jpeg_size ($body); + if ($w && $h) { return ($w, $h); } + return png_size ($body); } @@ -1636,6 +2023,24 @@ sub bellrand { } +sub signal_cleanup { + my ($sig) = @_; + print STDERR blurb() . (defined($sig) + ? "caught signal $sig." + : "exiting.") + . "\n" + if ($verbose_exec); + + x_cleanup(); + + if (@pids_to_kill) { + print STDERR blurb() . "killing: " . join(' ', @pids_to_kill) . "\n"; + kill ('TERM', @pids_to_kill); + } + + exit 1; +} + ############################################################################## # # Generating a list of urls only @@ -1659,25 +2064,34 @@ sub url_only_output { # ############################################################################## -my $image_ppm = ($ENV{TMPDIR} ? $ENV{TMPDIR} : "/tmp") . "/webcollage." . $$; -my $image_tmp1 = $image_ppm . "-1"; -my $image_tmp2 = $image_ppm . "-2"; +my $image_ppm = sprintf ("%s/webcollage-%08x", + ($ENV{TMPDIR} ? $ENV{TMPDIR} : "/tmp"), + rand(0xFFFFFFFF)); +my $image_tmp1 = sprintf ("%s/webcollage-1-%08x", + ($ENV{TMPDIR} ? $ENV{TMPDIR} : "/tmp"), + rand(0xFFFFFFFF)); +my $image_tmp2 = sprintf ("%s/webcollage-2-%08x", + ($ENV{TMPDIR} ? $ENV{TMPDIR} : "/tmp"), + rand(0xFFFFFFFF)); my $filter_cmd = undef; my $post_filter_cmd = undef; my $background = undef; +my @imagemap_areas = (); +my $imagemap_html_tmp = undef; +my $imagemap_jpg_tmp = undef; + + my $img_width; # size of the image being generated. my $img_height; -my $delay = 0; - +my $delay = 2; sub x_cleanup { - my ($sig) = @_; - print STDERR blurb() . "caught signal $sig.\n" if ($verbose_exec); unlink $image_ppm, $image_tmp1, $image_tmp2; - exit 1; + unlink $imagemap_html_tmp, $imagemap_jpg_tmp + if (defined ($imagemap_html_tmp)); } @@ -1713,9 +2127,9 @@ sub nontrapping_system { } -# Given the URL of a GIF or JPEG image, and the body of that image, writes a -# PPM to the given output file. Returns the width/height of the image if -# successful. +# Given the URL of a GIF, JPEG, or PNG image, and the body of that image, +# writes a PPM to the given output file. Returns the width/height of the +# image if successful. # sub image_to_pnm { my ($url, $body, $output) = @_; @@ -1727,9 +2141,12 @@ sub image_to_pnm { } elsif ((@_ = jpeg_size ($body))) { ($w, $h) = @_; $cmd = "djpeg"; + } elsif ((@_ = png_size ($body))) { + ($w, $h) = @_; + $cmd = "pngtopnm"; } else { LOG (($verbose_pbm || $verbose_load), - "not a GIF or JPG" . + "not a GIF, JPG, or PNG" . (($body =~ m@<(base|html|head|body|script|table|a href)>@i) ? " (looks like HTML)" : "") . ": $url"); @@ -1845,7 +2262,9 @@ sub x_or_pbm_output { if (!defined($webcollage_helper)) { # Only need these others if we don't have the helper. - @progs = (@progs, "giftopnm", "djpeg", "pnmpaste", "pnmscale", "pnmcut"); + @progs = (@progs, + "giftopnm", "pngtopnm", "djpeg", + "pnmpaste", "pnmscale", "pnmcut"); } foreach (@progs) { @@ -1856,17 +2275,6 @@ sub x_or_pbm_output { # $ppm_to_root_window_cmd = pick_root_displayer(); - - $SIG{HUP} = \&x_cleanup; - $SIG{INT} = \&x_cleanup; - $SIG{QUIT} = \&x_cleanup; - $SIG{ABRT} = \&x_cleanup; - $SIG{KILL} = \&x_cleanup; - $SIG{TERM} = \&x_cleanup; - - # Need this so that if giftopnm dies, we don't die. - $SIG{PIPE} = 'IGNORE'; - if (!$img_width || !$img_height) { $_ = "xdpyinfo"; which ($_) || error "$_ not found on \$PATH."; @@ -1922,13 +2330,17 @@ sub x_or_pbm_output { ($iw, $ih) = @_; $cmd = "djpeg |"; + } elsif ((@_ = png_size ($body))) { + ($iw, $ih) = @_; + $cmd = "pngtopnm |"; + } elsif ($body =~ m/^P\d\n(\d+) (\d+)\n/) { $iw = $1; $ih = $2; $cmd = ""; } else { - error "$bgimage is not a GIF, JPEG, or PPM."; + error "$bgimage is not a GIF, JPEG, PNG, or PPM."; } my $x = int (($img_width - $iw) / 2); @@ -1978,12 +2390,18 @@ sub paste_image { my ($iw, $ih); - if (defined ($webcollage_helper)) { + # If we are using the webcollage-helper, then we do not need to convert this + # image to a PPM. But, if we're using a filter command, we still must, since + # that's what the filters expect (webcollage-helper can read PPMs, so that's + # fine.) + # + if (defined ($webcollage_helper) && + !defined ($filter_cmd)) { ($iw, $ih) = image_size ($body); if (!$iw || !$ih) { LOG (($verbose_pbm || $verbose_load), - "not a GIF or JPG" . + "not a GIF, JPG, or PNG" . (($body =~ m@<(base|html|head|body|script|table|a href)>@i) ? " (looks like HTML)" : "") . ": $img"); @@ -2036,7 +2454,7 @@ sub paste_image { return 0 unless ($iw && $ih); } - my $target_w = $img_width; + my $target_w = $img_width; # max rectangle into which the image must fit my $target_h = $img_height; my $cmd = ""; @@ -2044,25 +2462,27 @@ sub paste_image { # Usually scale the image to fit on the screen -- but sometimes scale it - # to fit on half or a quarter of the screen. Note that we don't merely - # scale it to fit, we instead cut it in half until it fits -- that should - # give a wider distribution of sizes. + # to fit on half or a quarter of the screen. (We do this by reducing the + # size of the target rectangle.) Note that the image is not merely scaled + # to fit; we instead cut the image in half repeatedly until it fits in the + # target rectangle -- that gives a wider distribution of sizes. # - if (rand() < 0.3) { $target_w /= 2; $target_h /= 2; $scale /= 2; } - if (rand() < 0.3) { $target_w /= 2; $target_h /= 2; $scale /= 2; } + if (rand() < 0.3) { $target_w /= 2; $target_h /= 2; } # reduce target rect + if (rand() < 0.3) { $target_w /= 2; $target_h /= 2; } if ($iw > $target_w || $ih > $target_h) { while ($iw > $target_w || $ih > $target_h) { $iw = int($iw / 2); $ih = int($ih / 2); + $scale /= 2; } if ($iw <= 10 || $ih <= 10) { LOG ($verbose_pbm, "scaling to ${iw}x$ih would have been bogus."); return 0; } - LOG ($verbose_pbm, "scaling to ${iw}x$ih"); + LOG ($verbose_pbm, "scaling to ${iw}x$ih ($scale)"); $cmd .= " | pnmscale -xsize $iw -ysize $ih"; } @@ -2149,7 +2569,7 @@ sub paste_image { # If any cropping needs to happen, add pnmcut. # if ($crop_x != 0 || $crop_y != 0 || - $crop_w != $iw || $crop_h != $ih) { + $crop_w != $iw || $crop_h != $ih) { $iw = $crop_w; $ih = $crop_h; $cmd .= " | pnmcut $crop_x $crop_y $iw $ih"; @@ -2210,8 +2630,20 @@ sub paste_image { # cumulative. # if ($post_filter_cmd) { + + my $cmd; + $target = $image_tmp1; - $rc = nontrapping_system "($post_filter_cmd) < $image_ppm > $target"; + if (!defined ($webcollage_helper)) { + $cmd = "($post_filter_cmd) < $image_ppm > $target"; + } else { + # Blah, my scripts need the JPEG data, but some other folks need + # the PPM data -- what to do? Ignore the problem, that's what! +# $cmd = "djpeg < $image_ppm | ($post_filter_cmd) > $target"; + $cmd = "($post_filter_cmd) < $image_ppm > $target"; + } + + $rc = nontrapping_system ($cmd); if ($rc != 0) { LOG ($verbose_pbm, "filter failed: \"$post_filter_cmd\"\n"); return; @@ -2251,18 +2683,141 @@ sub paste_image { print STDOUT "image: ${iw}x${ih} @ $x,$y $base $source\n" if ($verbose_imgmap); + if ($imagemap_base) { + update_imagemap ($base, $x, $y, $iw, $ih, + $image_ppm, $img_width, $img_height); + } + clearlog(); return 1; } +sub update_imagemap { + my ($url, $x, $y, $w, $h, $image_ppm, $image_width, $image_height) = @_; + + $current_state = "imagemap"; + + my $max_areas = 200; + + $url = html_quote ($url); + my $x2 = $x + $w; + my $y2 = $y + $h; + my $area = ""; + unshift @imagemap_areas, $area; # put one on the front + if ($#imagemap_areas >= $max_areas) { + pop @imagemap_areas; # take one off the back. + } + + LOG ($verbose_pbm, "area: $x,$y,$x2,$y2 (${w}x$h)"); + + my $map_name = $imagemap_base; + $map_name =~ s@^.*/@@; + $map_name = 'collage' if ($map_name eq ''); + + my $imagemap_html = $imagemap_base . ".html"; + my $imagemap_jpg = $imagemap_base . ".jpg"; + + if (!defined ($imagemap_html_tmp)) { + $imagemap_html_tmp = $imagemap_html . sprintf (".%08x", rand(0xffffffff)); + $imagemap_jpg_tmp = $imagemap_jpg . sprintf (".%08x", rand(0xffffffff)); + } + + # Read the imagemap html file (if any) to get a template. + # + my $template_html = ''; + { + local *IN; + if (open (IN, "<$imagemap_html")) { + while () { $template_html .= $_; } + close IN; + LOG ($verbose_pbm, "read template $imagemap_html"); + } + + if ($template_html =~ m/^\s*$/s) { + $template_html = ("\n" . + "\n"); + LOG ($verbose_pbm, "created dummy template"); + } + } + + # Write the jpg to a tmp file + # + { + #my $cmd = "cjpeg < $image_ppm > $imagemap_jpg_tmp"; + my $cmd = "cp -p $image_ppm $imagemap_jpg_tmp"; + my $rc = nontrapping_system ($cmd); + if ($rc != 0) { + error ("imagemap jpeg failed: \"$cmd\"\n"); + } + } + + # Write the html to a tmp file + # + { + my $body = $template_html; + my $areas = join ("\n\t", @imagemap_areas); + my $map = ("\n\t$areas\n"); + my $img = (""); + $body =~ s@().*?()@$map@is; + $body =~ s@]*\bUSEMAP\b[^<>]*>@$img@is; + + # if there are magic webcollage spans in the html, update those too. + # + { + my @st = stat ($imagemap_jpg_tmp); + my $date = strftime("%d-%b-%Y %l:%M:%S %p %Z", localtime($st[9])); + my $size = int(($st[7] / 1024) + 0.5) . "K"; + $body =~ s@().*?()@$1$date$2@si; + $body =~ s@().*?()@$1$size$2@si; + } + + local *OUT; + open (OUT, ">$imagemap_html_tmp") || error ("$imagemap_html_tmp: $!"); + print OUT $body || error ("$imagemap_html_tmp: $!"); + close OUT || error ("$imagemap_html_tmp: $!"); + LOG ($verbose_pbm, "wrote $imagemap_html_tmp"); + } + + # Rename the two tmp files to the real files + # + rename ($imagemap_html_tmp, $imagemap_html) || + error "renaming $imagemap_html_tmp to $imagemap_html"; + LOG ($verbose_pbm, "wrote $imagemap_html"); + rename ($imagemap_jpg_tmp, $imagemap_jpg) || + error "renaming $imagemap_jpg_tmp to $imagemap_jpg"; + LOG ($verbose_pbm, "wrote $imagemap_jpg"); +} + + +sub init_signals { + + $SIG{HUP} = \&signal_cleanup; + $SIG{INT} = \&signal_cleanup; + $SIG{QUIT} = \&signal_cleanup; + $SIG{ABRT} = \&signal_cleanup; + $SIG{KILL} = \&signal_cleanup; + $SIG{TERM} = \&signal_cleanup; + + # Need this so that if giftopnm dies, we don't die. + $SIG{PIPE} = 'IGNORE'; +} + +END { signal_cleanup(); } + + sub main { $| = 1; srand(time ^ $$); my $verbose = 0; my $dict; + my $driftnet_cmd = 0; $current_state = "init"; $load_method = "none"; @@ -2288,6 +2843,9 @@ sub main { } elsif ($_ eq "-urls-only") { $urls_only_p = 1; $no_output_p = 1; + } elsif ($_ eq "-imagemap") { + $imagemap_base = shift @ARGV; + $no_output_p = 1; } elsif ($_ eq "-verbose") { $verbose++; } elsif (m/^-v+$/) { @@ -2314,6 +2872,13 @@ sub main { $http_proxy = shift @ARGV; } elsif ($_ eq "-dictionary" || $_ eq "-dict") { $dict = shift @ARGV; + } elsif ($_ eq "-driftnet" || $_ eq "--driftnet") { + @search_methods = ( 100, "driftnet", \&pick_from_driftnet ); + if (! ($ARGV[0] =~ m/^-/)) { + $driftnet_cmd = shift @ARGV; + } else { + $driftnet_cmd = $default_driftnet_cmd; + } } elsif ($_ eq "-debug" || $_ eq "--debug") { my $which = shift @ARGV; my @rest = @search_methods; @@ -2333,11 +2898,14 @@ sub main { LOG (1, "DEBUG: using only \"$which\""); } else { - print STDERR "$copyright\nusage: $progname [-root]" . - " [-display dpy] [-root] [-verbose] [-timeout secs]\n" . - "\t\t [-delay secs] [-filter cmd] [-filter2 cmd]\n" . - "\t\t [-dictionary dictionary-file]\n" . - "\t\t [-http-proxy host[:port]]\n"; + print STDERR "$copyright\nusage: $progname " . + "[-root] [-display dpy] [-verbose] [-debug which]\n" . + "\t\t [-timeout secs] [-delay secs] [-size WxH]\n" . + "\t\t [-no-output] [-urls-only] [-imagemap filename]\n" . + "\t\t [-filter cmd] [-filter2 cmd] [-background color]\n" . + "\t\t [-dictionary dictionary-file] [-http-proxy host[:port]]\n" . + "\t\t [-driftnet [driftnet-program-and-args]]\n" . + "\n"; exit 1; } } @@ -2417,6 +2985,10 @@ sub main { pick_dictionary(); } + init_signals(); + + spawn_driftnet ($driftnet_cmd) if ($driftnet_cmd); + if ($urls_only_p) { url_only_output; } else {