X-Git-Url: http://git.hungrycats.org/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=hacks%2Fwebcollage;h=4b41ac16534580c5e9036ffaa01407e9c5edecce;hb=4cecfc89e5e889c7232693897c06168fb378bd5c;hp=9489f22f6d1d3b58390c986c3358eec97f4a100d;hpb=3f9592851ce4ed76a9979bfdd6ec7dc5c457e183;p=xscreensaver diff --git a/hacks/webcollage b/hacks/webcollage index 9489f22f..4b41ac16 100755 --- a/hacks/webcollage +++ b/hacks/webcollage @@ -12,12 +12,26 @@ # 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\ +# 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: +# +# default-n: webcollage -root -driftnet \n\ +# +# Driftnet is the Unix implementation of the MacOS "EtherPEG" program. + + require 5; use strict; @@ -33,31 +47,44 @@ 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: 1.90 $ }; $version =~ s/^[^0-9]+([0-9.]+).*$/$1/; +my $version = q{ $Revision: 1.102 $ }; $version =~ s/^[^0-9]+([0-9.]+).*$/$1/; my $copyright = "WebCollage $version, Copyright (c) 1999-2002" . " Jamie Zawinski \n" . " http://www.jwz.org/xscreensaver/\n"; -my @search_methods = ( 40, "imagevista", \&pick_from_alta_vista_images, - 30, "altavista", \&pick_from_alta_vista_text, - 19, "yahoorand", \&pick_from_yahoo_random_link, +my @search_methods = ( 77, "altavista", \&pick_from_alta_vista_random_link, + 12, "yahoorand", \&pick_from_yahoo_random_link, 9, "lycos", \&pick_from_lycos_text, 2, "yahoonews", \&pick_from_yahoo_news_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, + # 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, ); # programs we can use to write to the root window (tried in ascending order.) @@ -110,19 +137,20 @@ 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. ); @@ -141,6 +169,8 @@ 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... ); @@ -182,6 +212,12 @@ my $min_gif_area = (120 * 120); my $no_output_p = 0; my $urls_only_p = 0; +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; my %rejected_urls; @@ -274,6 +310,7 @@ sub get_document_1 { my $cookie = $cookies{$them}; my $user_agent = "$progname/$version"; + if ($url =~ m@^http://www\.altavista\.com/@) { # block this, you turkeys. $user_agent = "Mozilla/4.76 [en] (X11; U; Linux 2.2.16-22 i686; Nav)"; @@ -352,6 +389,10 @@ 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; @@ -569,8 +610,7 @@ sub pick_image_from_body { } elsif ( ! m@^[^/:?]+:@ ) { $_ = "$base$link"; s@/\./@/@g; - while (s@/\.\./@/@g) { - } + 1 while (s@/[^/]+/\.\./@/@g); } # skip non-http @@ -793,9 +833,9 @@ sub pick_from_search_engine { $search_count = $1; } elsif ($body =~ m@found about ((\d{1,3})(,\d{3})*|\d+) results@) { $search_count = $1; - } elsif ($body =~ m@\b\d+ - \d+ of (\d+)\b@i) { # imagevista + } elsif ($body =~ m@\b\d+ - \d+ of (\d+)\b@i) { # avimages $search_count = $1; - } elsif ($body =~ m@About ((\d{1,3})(,\d{3})*) images@i) { # imagevista + } elsif ($body =~ m@About ((\d{1,3})(,\d{3})*) images@i) { # avimages $search_count = $1; } elsif ($body =~ m@We found ((\d{1,3})(,\d{3})*|\d+) results@i) { # *vista $search_count = $1; @@ -805,9 +845,9 @@ sub pick_from_search_engine { $search_count = $1; # lycos } elsif ($body =~ m@WEB.*?RESULTS.*?\b((\d{1,3})(,\d{3})*)\b.*?Matches@i) { $search_count = $1; # hotbot - } elsif ($body =~ m@no photos were found containing@i) { # imagevista + } elsif ($body =~ m@no photos were found containing@i) { # avimages $search_count = "0"; - } elsif ($body =~ m@found no document matching@i) { # altavista + } elsif ($body =~ m@found no document matching@i) { # avtext $search_count = "0"; } 1 while ($search_count =~ s/^(\d+)(\d{3})/$1,$2/); @@ -974,6 +1014,50 @@ sub pick_from_yahoo_random_link { } } + +############################################################################ +# +# Pick images from random pages returned by the Alta Vista Random Link +# +############################################################################ + +# altavista +my $alta_vista_random_link = "http://www.altavista.com/image/randomlink"; + + +# Picks a random page; picks a random image on that page; +# returns two URLs: the page containing the image, and the image. +# Returns () if nothing found this time. +# +sub pick_from_alta_vista_random_link { + my ( $timeout ) = @_; + + print STDERR "\n\n" if ($verbose_load); + LOG ($verbose_load, "URL: $alta_vista_random_link"); + + $last_search = $alta_vista_random_link; # for warnings + + $suppress_audit = 1; + + my ( $base, $body ) = get_document ($alta_vista_random_link, + undef, $timeout); + if (!$base || !$body) { + $body = undef; + return; + } + + LOG ($verbose_load, "redirected to: $base"); + + my $img = pick_image_from_body ($base, $body); + $body = undef; + + if ($img) { + return ($base, $img); + } else { + return (); + } +} + ############################################################################ # @@ -982,22 +1066,20 @@ sub pick_from_yahoo_random_link { ############################################################################ -my $alta_vista_images_url = "http://www.altavista.com/cgi-bin/query" . +my $alta_vista_images_url = "http://www.altavista.com/image/results" . "?ipht=1" . # photos "&igrph=1" . # graphics "&iclr=1" . # color "&ibw=1" . # b&w "&micat=1" . # no partner sites - "&imgset=1" . # no partner sites - "&stype=simage" . # do image search - "&mmW=1" . # unknown, but required + "&sc=on" . # "site collapse" "&q="; -# imagevista +# avimages sub pick_from_alta_vista_images { my ( $timeout ) = @_; - my $words = random_words(1); + my $words = random_words(0); my $page = (int(rand(9)) + 1); my $search_url = $alta_vista_images_url . $words; @@ -1012,10 +1094,9 @@ sub pick_from_alta_vista_images { my @candidates = (); foreach my $u (@subpages) { - # altavista is encoding their URLs now. - next unless ($u =~ - m@^/r\?ck_sm=[a-zA-Z0-9]+\&ref=[a-zA-Z0-9]+(\&uid=[a-zA-Z0-9]+)?\&r=(.*)@); - $u = url_unquote($2); + # avtext is encoding their URLs now. + next unless ($u =~ m@^/r.*\&r=([^&]+).*@); + $u = url_unquote($1); next unless ($u =~ m@^http://@i); # skip non-HTTP or relative URLs next if ($u =~ m@[/.]altavista\.com\b@i); # skip altavista builtins @@ -1151,18 +1232,19 @@ sub pick_from_google_image_numbers { ############################################################################ -my $alta_vista_url_1 = "http://www.altavista.com/cgi-bin/query?pg=q" . - "&text=yes&kl=XX&stype=stext&q="; -my $alta_vista_url_2 = "http://www.altavista.com/sites/search/web?pg=q" . - "&kl=XX&search=Search&q="; - -my $alta_vista_url = $alta_vista_url_2; +my $alta_vista_url = "http://www.altavista.com/web/results" . + "?pg=aq" . + "&aqmode=s" . + "&filetype=html" . + "&sc=on" . # "site collapse" + "&nbq=50" . + "&aqo="; -# altavista +# 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 . $words; @@ -1182,8 +1264,7 @@ sub pick_from_alta_vista_text { # onMouseOver to make it look like they're not! Well, it makes it # easier for us to identify search results... # - next unless ($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"); @@ -1202,23 +1283,28 @@ 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; + # 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); @@ -1227,7 +1313,7 @@ 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"); @@ -1246,17 +1332,22 @@ 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); + # 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); @@ -1264,10 +1355,12 @@ 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"); @@ -1286,14 +1379,21 @@ 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); + my $words = random_words(0); my $search_url = $yahoo_news_url . $words; my ($search_hit_count, @subpages) = @@ -1315,6 +1415,113 @@ sub pick_from_yahoo_news_text { + +############################################################################ +# +# Pick images by waiting for driftnet to populate a temp dir with files. +# Requires driftnet version 0.1.5 or later. +# (Driftnet is a program by Chris Lightfoot that sniffs your local ethernet +# for images being downloaded by others.) +# Driftnet/webcollage integration by jwz. +# +############################################################################ + +# driftnet +sub pick_from_driftnet { + my ( $timeout ) = @_; + + my $id = $driftnet_magic; + my $dir = $driftnet_dir; + my $start = time; + my $now; + + error ("\$driftnet_dir unset?") unless ($dir); + $dir =~ s@/+$@@; + + error ("$dir unreadable") unless (-d "$dir/."); + + $timeout = $http_timeout unless ($timeout); + $last_search = $id; + + while ($now = time, $now < $start + $timeout) { + local *DIR; + opendir (DIR, $dir) || error ("$dir: $!"); + while (my $file = readdir(DIR)) { + next if ($file =~ m/^\./); + $file = "$dir/$file"; + closedir DIR; + LOG ($verbose_load, "picked file $file ($id)"); + return ($id, $file); + } + closedir DIR; + } + LOG (($verbose_net || $verbose_load), "timed out for $id"); + return (); +} + + +sub get_driftnet_file { + my ($file) = @_; + + error ("\$driftnet_dir unset?") unless ($driftnet_dir); + + my $id = $driftnet_magic; + my $re = qr/$driftnet_dir/; + error ("$id: $file not in $driftnet_dir?") + unless ($file =~ m@^$re@o); + + local *IN; + open (IN, $file) || error ("$id: $file: $!"); + my $body = ''; + while () { $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)); +} + ############################################################################ # @@ -1326,7 +1533,6 @@ 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 { @@ -1550,6 +1756,12 @@ 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 $done = 0; foreach (@recent_images) { @@ -1703,6 +1915,24 @@ sub bellrand { } +sub signal_cleanup { + my ($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 @@ -1739,12 +1969,8 @@ my $img_height; 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; } @@ -1923,17 +2149,6 @@ 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."; @@ -2342,12 +2557,29 @@ sub paste_image { } +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"; @@ -2399,6 +2631,13 @@ 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 $which = shift @ARGV; my @rest = @search_methods; @@ -2418,11 +2657,13 @@ sub main { LOG (1, "DEBUG: using only \"$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: $progname " . + "[-root] [-display dpy] [-verbose] [-debug which]\n" . + "\t\t [-timeout secs] [-delay secs] [-filter cmd] [-filter2 cmd]\n" . + "\t\t [-no-output] [-urls-only] [-background color] [-size WxH]\n" . + "\t\t [-dictionary dictionary-file] [-http-proxy host[:port]]\n" . + "\t\t [-driftnet [driftnet-program-and-args]]\n" . + "\n"; exit 1; } } @@ -2502,6 +2743,10 @@ sub main { pick_dictionary(); } + init_signals(); + + spawn_driftnet ($driftnet_cmd) if ($driftnet_cmd); + if ($urls_only_p) { url_only_output; } else {