1.1 --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
     1.2 +++ b/javadoc-search.pl	Mon Sep 27 16:57:26 2010 -0400
     1.3 @@ -0,0 +1,465 @@
     1.4 +#!/usr/bin/perl
     1.5 +
     1.6 +use strict;
     1.7 +use vars qw / $javadocRoots $indexFile $resultFrame $linkFrame $gzip /;
     1.8 +#####################################################################
     1.9 +#
    1.10 +# javadoc-search
    1.11 +#
    1.12 +# http://sourceforge.net/projects/javadoc-search/
    1.13 +#
    1.14 +# Copyright (C) 2002-2003, Nicholas Sushkin.
    1.15 +#
    1.16 +#     This program is free software; you can redistribute it and/or modify
    1.17 +#     it under the terms of the GNU General Public License as published by
    1.18 +#     the Free Software Foundation; either version 2 of the License, or
    1.19 +#     (at your option) any later version.
    1.20 +#
    1.21 +#     This program is distributed in the hope that it will be useful,
    1.22 +#     but WITHOUT ANY WARRANTY; without even the implied warranty of
    1.23 +#     MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
    1.24 +#     GNU General Public License for more details.
    1.25 +#
    1.26 +#     You should have received a copy of the GNU General Public License
    1.27 +#     along with this program; if not, write to the Free Software
    1.28 +#     Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
    1.29 +# 
    1.30 +#
    1.31 +# Instructions are at http://javadoc-search.sourceforge.net/
    1.32 +#
    1.33 +#
    1.34 +#####################################################################
    1.35 +#
    1.36 +# $Id: javadoc-search.pl,v 1.8 2003/08/19 00:11:09 nsushkin Exp $
    1.37 +#
    1.38 +#####################################################################
    1.39 +
    1.40 +### Indexing setup
    1.41 +
    1.42 +# list of subdirectories of docroot to search for javadoc index files
    1.43 +# Windows and Apache - prefix with drive letter and use only lowercase letters.
    1.44 +#    $javadocRoots = [ "c:/program\ files/apache/htdocs/docs" ];
    1.45 +# Unix 
    1.46 +#    $javadocRoots = [ "/usr/local/apache/htdocs/docs" ];
    1.47 +$javadocRoots = [ "/usr/local/apache/htdocs/docs" ];
    1.48 +
    1.49 +# index file, if ends with .gz will use gzip
    1.50 +# Windows
    1.51 +#   $indexFile = "c:/bin/apache\ group/apache/javadoc-index.txt";
    1.52 +# Unix
    1.53 +#   $indexFile = "/usr/local/apache/htdocs/docs/javadoc-index.txt.gz";
    1.54 +$indexFile = "/usr/local/apache/htdocs/docs/javadoc-index.txt.gz";
    1.55 +
    1.56 +### Display style setup
    1.57 +
    1.58 +# if $resultFrame is not empty, display search results in this frame
    1.59 +$resultFrame = "";
    1.60 +
    1.61 +# if $linkFrame is not empty, open javadoc links in this frame
    1.62 +$linkFrame = "";
    1.63 +
    1.64 +### Auxillary programs setup
    1.65 +
    1.66 +# which gzip
    1.67 +$gzip = "/usr/bin/gzip";
    1.68 +
    1.69 +# nothing below is customizable
    1.70 +#####################################################################
    1.71 +use Getopt::Std;
    1.72 +use File::Find;
    1.73 +
    1.74 +my @indexFiles;
    1.75 +my %links;
    1.76 +
    1.77 +(my $rev = '$Revision: 1.8 $') =~ s/^\$\w+: ([\.\d]+) \$$/$1/;
    1.78 +
    1.79 +sub usage
    1.80 +{
    1.81 +    print STDERR <<EOUSAGE;
    1.82 +Javadoc Search $rev
    1.83 +
    1.84 +Usage: $0 [-h] [-r]
    1.85 +
    1.86 +   -h     prints this help
    1.87 +   -r     builds index
    1.88 +
    1.89 +   This is a cgi script to search javadoc for classes, members, and methods
    1.90 +   using an index.
    1.91 +EOUSAGE
    1.92 +        
    1.93 +}
    1.94 +
    1.95 +#
    1.96 +# Build index part
    1.97 +#
    1.98 +sub indexWrite
    1.99 +{
   1.100 +    print STDERR "Writing all ", indexGetLinkCount(), " links to $indexFile .. ";
   1.101 +
   1.102 +    if ($indexFile =~ /\.gz$/)
   1.103 +    {
   1.104 +        open(OUTPUT, "| $gzip -c > \"$indexFile\"") || die "unable to gzip to $indexFile: $!\n";
   1.105 +    } 
   1.106 +    else 
   1.107 +    {
   1.108 +        open(OUTPUT, ">" . $indexFile) || die "unable to write to $indexFile: $!\n";
   1.109 +    }
   1.110 +
   1.111 +    my $oldSearchKey = "";
   1.112 +    foreach (sort { $links{$a} cmp $links{$b} } keys %links)
   1.113 +    {
   1.114 +        my $searchKey = $links{$_};
   1.115 +        if ($searchKey ne $oldSearchKey) 
   1.116 +        { 
   1.117 +            if ($oldSearchKey ne "") { print OUTPUT "\n"; }
   1.118 +            print OUTPUT $searchKey; 
   1.119 +        }
   1.120 +        print OUTPUT "\t", $_;
   1.121 +
   1.122 +        $oldSearchKey = $searchKey;
   1.123 +    }
   1.124 +    print OUTPUT "\n";
   1.125 +
   1.126 +    close(OUTPUT);
   1.127 +
   1.128 +    print STDERR "done\n";
   1.129 +}
   1.130 +
   1.131 +sub indexGetLinkCount
   1.132 +{
   1.133 +    scalar(keys %links);
   1.134 +}
   1.135 +
   1.136 +sub indexRebuild
   1.137 +{
   1.138 +    find(\&indexFileFinder, @$javadocRoots);
   1.139 +
   1.140 +    foreach (@indexFiles)
   1.141 +    {
   1.142 +        parseLinks($$_[0], $$_[1]);
   1.143 +    }
   1.144 +}
   1.145 +
   1.146 +sub indexFileFinder
   1.147 +{
   1.148 +    return unless -f $File::Find::name;
   1.149 +    return unless /^index-.*html??/;
   1.150 +
   1.151 +    push @indexFiles, [$File::Find::name, $File::Find::dir];
   1.152 +}
   1.153 +
   1.154 +sub indexAddLink
   1.155 +{
   1.156 +    my $key = shift;
   1.157 +    my $link = shift;
   1.158 +
   1.159 +    $links{$link} = $key;
   1.160 +}
   1.161 +
   1.162 +sub parseLinks()
   1.163 +{
   1.164 +    my $fileName = shift;
   1.165 +    my $dirName = shift;
   1.166 +
   1.167 +    print STDERR "Parsing ", $fileName, ": ";
   1.168 +
   1.169 +    open(FILE, $fileName);
   1.170 +
   1.171 +    while(<FILE>)
   1.172 +    {
   1.173 +        if($_ =~ m/<A[^>]+?HREF\s*=\s*[\"\']?([^\'\" >]+?)[ \'\"]?(>| title=)/sig)
   1.174 +        {
   1.175 +            parseLink($1, $dirName);
   1.176 +        }
   1.177 +    }
   1.178 +
   1.179 +    close(FILE);
   1.180 +
   1.181 +    print STDERR indexGetLinkCount(), " links\n";
   1.182 +}
   1.183 +
   1.184 +sub parseLink()
   1.185 +{
   1.186 +    my $link = shift;
   1.187 +    my $dirName = shift;
   1.188 +
   1.189 +    return if $link =~ qr'^(http|ftp)://'i;
   1.190 +
   1.191 +    # print "# link=", $link, "\n";
   1.192 +
   1.193 +    my $abslink = ( $link =~ qr'^[\./]*([^\./].*)' ) ? $1 : $link;
   1.194 +    # print "# abslink= $abslink\n";
   1.195 +
   1.196 +    return unless $abslink =~ qr'/';
   1.197 +    
   1.198 +    my ($file,$anchor) = split '#', $abslink;
   1.199 +    # print "# file=$file, anchor=$anchor\n";
   1.200 +
   1.201 +    my $class = ($file =~ /(.*)\.html??/ ) ? $1 : return;
   1.202 +    $class =~ tr{\/}{\.};
   1.203 +
   1.204 +    # print "# class=", $class, "\n";
   1.205 +    # print "# dirname=", $dirName, "\n";
   1.206 +
   1.207 +    # cancel out ".."
   1.208 +    if ($link =~ /^\.\.\//)
   1.209 +    {
   1.210 +        my @linkSegs = split '/', $link;
   1.211 +        my @dirSegs = split '/', $dirName;
   1.212 +
   1.213 +        while (@linkSegs[0] eq '..' and $#dirSegs >= 0)
   1.214 +        {
   1.215 +            shift @linkSegs;
   1.216 +            pop @dirSegs;
   1.217 +        }
   1.218 +        $link = join '/', @linkSegs;
   1.219 +        $dirName = join '/', @dirSegs;
   1.220 +    }
   1.221 +    my $fullLink = ( $link =~ /^\// ) ? $link : $dirName . "/" . $link;
   1.222 +
   1.223 +    my $key = $anchor eq "" ? $class : $class . '#' . $anchor;
   1.224 +    indexAddLink($key, $fullLink);
   1.225 +}
   1.226 +
   1.227 +#
   1.228 +# CGI part
   1.229 +#
   1.230 +use CGI qw(:standard);
   1.231 +
   1.232 +my $searchTypes = {'c'=>'class','cm'=>'class#member','cma'=>'class#member(args)'};
   1.233 +
   1.234 +sub topFrameGenerate
   1.235 +{
   1.236 +    Delete('render');
   1.237 +    param('render', 'query');
   1.238 +    my $queryFrameUrl = self_url;
   1.239 +
   1.240 +    Delete('render');
   1.241 +    param('render', 'noResults');
   1.242 +    my $noResultsFrameUrl = self_url;
   1.243 +
   1.244 +    return
   1.245 +        header(),
   1.246 +        html(head(title('Javadoc search')),
   1.247 +             frameset({-rows => '150,*'},
   1.248 +                      frame({-name => 'query', -src => $queryFrameUrl}),
   1.249 +                      frame({-name => $resultFrame, -src => $noResultsFrameUrl})));
   1.250 +}
   1.251 +
   1.252 +sub noResultsGenerate
   1.253 +{
   1.254 +    return
   1.255 +        header,
   1.256 +        start_html('Javadoc search'),
   1.257 +        h2('No results yet'),
   1.258 +        end_html,
   1.259 +        "\n";
   1.260 +}
   1.261 +
   1.262 +sub searchFormGenerate
   1.263 +{
   1.264 +    Delete('render');
   1.265 +
   1.266 +    return
   1.267 +        h1('Javadoc search'),
   1.268 +        $resultFrame ? start_form(-target => $resultFrame, -action=>script_name) : start_form(-action=>script_name),
   1.269 +        "Search pattern ", 
   1.270 +        textfield(-name=>'pattern', -size=>15),
   1.271 +        " in ",
   1.272 +        scrolling_list(-name=>'type',
   1.273 +                       -values=>[sort(keys(%$searchTypes))],
   1.274 +                       -labels=>$searchTypes,
   1.275 +                       -defaults=>['c'],
   1.276 +                       -size=>1),
   1.277 +        " and display ",
   1.278 +        scrolling_list(-name=>'results',
   1.279 +                       -values=>['1', '10', '20', '50', '0'],
   1.280 +                       -labels=>{'1'=>'1', '10'=>'10', '20'=>'20', '50'=>'50', '0'=>'All'},
   1.281 +                       -defaults=>['10'],
   1.282 +                       -size=>1),
   1.283 +        " results ",
   1.284 +        submit(-name=>'go', -value=>'Go'), submit(-name=>'lucky', -value=>'Go!'),
   1.285 +        $resultFrame ? hidden(-name => 'render', -default => 'results') : "",
   1.286 +        end_form,
   1.287 +        $resultFrame ? "" : hr;
   1.288 +}
   1.289 +
   1.290 +sub doSearch
   1.291 +{
   1.292 +    my ($pattern, $type, $results) = @_;
   1.293 +
   1.294 +    if (! -r $indexFile)
   1.295 +    {
   1.296 +        print strong("Error: index file is not readable");
   1.297 +        return;
   1.298 +    }
   1.299 +
   1.300 +    if ($indexFile =~ /\.gz$/)
   1.301 +    {
   1.302 +        open(INPUT, "$gzip -dc \"$indexFile\" |") || die "unable to un-gzip $indexFile: $!\n";
   1.303 +    } 
   1.304 +    else 
   1.305 +    {
   1.306 +        open(INPUT, $indexFile) || die "unable to read $indexFile: $!\n";
   1.307 +    }
   1.308 +    
   1.309 +    my $docRoot = $ENV{'DOCUMENT_ROOT'};
   1.310 +    my @matches;
   1.311 +
   1.312 +  INPUTLINE: 
   1.313 +    while(<INPUT>)
   1.314 +    {
   1.315 +        chomp;
   1.316 +        my ($key, @urls) = split (/\t/);
   1.317 +        
   1.318 +        for ($key)
   1.319 +        {
   1.320 +            if (($type eq "c" and /\#/) ||
   1.321 +                ($type eq "cm" and !(/\#/ and !/\)/)) ||
   1.322 +                ($type eq "cma" and !(/\#/ and /\)/)))
   1.323 +            {
   1.324 +                next INPUTLINE;
   1.325 +            }
   1.326 +        }
   1.327 +
   1.328 +        if ($key =~ /$pattern/ and ($results == 0 or $#matches < $results) )
   1.329 +        {
   1.330 +            push @matches, [ ( $key, map { s/$docRoot//; $_ } @urls ) ];
   1.331 +        }
   1.332 +
   1.333 +        last if $results > 0 and $#matches >= $results;
   1.334 +
   1.335 +    }
   1.336 +    close(INPUT);
   1.337 +
   1.338 +    return \@matches;
   1.339 +}
   1.340 +
   1.341 +sub displayLuckyMatch
   1.342 +{
   1.343 +    my $matches = shift;
   1.344 +
   1.345 +    if ( $#$matches < 0 )
   1.346 +    {
   1.347 +        print 
   1.348 +            header,
   1.349 +            start_html('No matches'),
   1.350 +            p("No matches"),
   1.351 +            end_html;
   1.352 +    }
   1.353 +    else
   1.354 +    {
   1.355 +        print
   1.356 +            $linkFrame ? header(-target => $linkFrame) : "",
   1.357 +            redirect('-location'=> $$matches[0][1]);
   1.358 +    }
   1.359 +}
   1.360 +
   1.361 +
   1.362 +sub displayMatches
   1.363 +{
   1.364 +    my $matches = shift;
   1.365 +
   1.366 +    if ( $#$matches < 0 )
   1.367 +    {
   1.368 +        print p("No matches");
   1.369 +    }
   1.370 +    else
   1.371 +    {
   1.372 +        print 
   1.373 +            p(
   1.374 +              "Searching pattern ", em(escapeHTML(param('pattern'))),
   1.375 +              " in ", em(escapeHTML($$searchTypes{param('type')}))),
   1.376 +            h2("Search results"),
   1.377 +            "<ul>";
   1.378 +        
   1.379 +        foreach my $match (@$matches)
   1.380 +        {
   1.381 +            my $searchKey = escapeHTML(shift @$match);
   1.382 +            my $cnt = -1;
   1.383 +            print li( map {  
   1.384 +                a( $linkFrame ? {-href => $_, -target => $linkFrame } : { -href => $_}, ( $cnt++ >= 0 ? "alt".($cnt) : $searchKey) ) 
   1.385 +                } @$match );
   1.386 +            
   1.387 +        }
   1.388 +        
   1.389 +        print "</ul>";
   1.390 +    }
   1.391 +}
   1.392 +
   1.393 +#
   1.394 +# main
   1.395 +#
   1.396 +use vars qw / $opt_h $opt_r /;
   1.397 +
   1.398 +if (server_software() eq 'cmdline')
   1.399 +{
   1.400 +    getopts("rh");
   1.401 +
   1.402 +    if ($opt_h)
   1.403 +    {
   1.404 +        usage;
   1.405 +        exit;
   1.406 +    }
   1.407 +    
   1.408 +    if ($opt_r)
   1.409 +    {
   1.410 +        indexRebuild;
   1.411 +        indexWrite;
   1.412 +        exit;
   1.413 +    }
   1.414 +}
   1.415 +
   1.416 +
   1.417 +if (param('render'))
   1.418 +{
   1.419 +    $resultFrame = 'results';
   1.420 +    $linkFrame = '';
   1.421 +}
   1.422 +
   1.423 +my $thisFrame = param('render');
   1.424 +
   1.425 +if ($thisFrame eq 'topFrame')
   1.426 +{
   1.427 +    print topFrameGenerate();
   1.428 +    exit 0;
   1.429 +}
   1.430 +elsif ($thisFrame eq 'noResults')
   1.431 +{
   1.432 +    print noResultsGenerate();
   1.433 +    exit 0;
   1.434 +}
   1.435 +
   1.436 +unless (param('lucky') and param('pattern'))
   1.437 +{
   1.438 +    print 
   1.439 +        header,
   1.440 +        start_html('Javadoc search'),
   1.441 +        defined(param('pattern')) && $resultFrame ? "" : searchFormGenerate();
   1.442 +}
   1.443 +
   1.444 +if (param('pattern')) 
   1.445 +{
   1.446 +    my $matches = doSearch(param('pattern'), param('type'), param('results'));
   1.447 +
   1.448 +    if (param('lucky'))
   1.449 +    {
   1.450 +        displayLuckyMatch($matches);
   1.451 +    }
   1.452 +    else
   1.453 +    {
   1.454 +        displayMatches($matches);
   1.455 +    }
   1.456 +}
   1.457 +else
   1.458 +{
   1.459 +    if ($thisFrame ne 'query')
   1.460 +    {
   1.461 +        print p('Specify non-blank search pattern');
   1.462 +    }
   1.463 +}
   1.464 +
   1.465 +unless (param('lucky') and param('pattern'))
   1.466 +{
   1.467 +    print end_html;
   1.468 +}