3 # Copyright (C) 2000, HELM Team.
5 # This file is part of HELM, an Hypertextual, Electronic
6 # Library of Mathematics, developed at the Computer Science
7 # Department, University of Bologna, Italy.
9 # HELM is free software; you can redistribute it and/or
10 # modify it under the terms of the GNU General Public License
11 # as published by the Free Software Foundation; either version 2
12 # of the License, or (at your option) any later version.
14 # HELM is distributed in the hope that it will be useful,
15 # but WITHOUT ANY WARRANTY; without even the implied warranty of
16 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
17 # GNU General Public License for more details.
19 # You should have received a copy of the GNU General Public License
20 # along with HELM; if not, write to the Free Software
21 # Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
23 # For details, see the HELM World-Wide-Web page,
24 # http://cs.unibo.it/helm/.
28 my $VERSION = "@VERSION@";
32 my $cgi_dir = "@HELM_CGI_DIR@";
33 my (%xml_map, %rdf_map, %xslt_map);
35 # First of all, let's load HELM configuration
37 my $HELM_LIB_DIR = $ENV{"HELM_LIB_DIR"};
38 # this should be the only fixed constant
39 my $DEFAULT_HELM_LIB_DIR = "@HELM_LIB_DIR@";
40 if (defined ($HELM_LIB_DIR)) {
41 $HELM_LIB_PATH = $HELM_LIB_DIR."/configuration.pl";
43 $HELM_LIB_PATH = $DEFAULT_HELM_LIB_DIR."/configuration.pl";
46 # next require defines: $helm_dir, $html_link, $dtd_dir, $uris_dbm, $indexname
47 require $HELM_LIB_PATH;
49 # TEMP: TODO put these vars in configuration file configuration.xml
51 my $cachemode = $ENV{'HTTP_GETTER_CACHE_MODE'} ||
53 if (($cachemode ne 'gzipped') and ($cachemode ne 'normal')) {
54 die "Invalid HTTP_GETTER_CACHE_MODE environment variable, must be".
55 "'normal' or 'gzipped'\n";
58 my $helm_rdf_dir = $ENV{'HTTP_GETTER_RDF_DIR'} ||
59 "/usr/local/helm/rdf_library";
60 my $rdf_dbm = $ENV{'HTTP_GETTER_RDF_DBM'} ||
61 "/usr/local/helm/rdf_urls_of_uris.db";
62 my $xslt_dbm = $ENV{'HTTP_GETTER_XSLT_DBM'} ||
63 "/usr/local/helm/xslt_urls_of_uris.db";
64 my $rdf_indexname = $ENV{'HTTP_GETTER_RDF_INDEXNAME'} ||
66 my $xslt_indexname = $ENV{'HTTP_GETTER_XSLT_INDEXNAME'} ||
68 $servers_file = $ENV{'HTTP_GETTER_SERVERS_FILE'} || $servers_file;
70 # </move_to_conf_file>
72 # Let's override the configuration file
73 $style_dir = $ENV{"HELM_STYLE_DIR"} if (defined ($ENV{"HELM_STYLE_DIR"}));
74 $dtd_dir = $ENV{"HELM_DTD_DIR"} if (defined ($ENV{"HELM_DTD_DIR"}));
86 #CSC: mancano i controlli sulle condizioni di errore di molte funzioni
87 #CSC: ==> non e' robusto
88 #CSC: altra roba da sistemare segnata con CSC
90 my $d = new HTTP::Daemon LocalPort => 8081
91 or die "Error: port 8081 not available.";
92 my $myownurl = $d->url;
94 # Let's patch the returned URL
95 $myownurl =~ s/\/$//; # chop the final slash
96 my $myownport = $myownurl;
97 $myownport =~ s/http:\/\/(.*):(.*)/$2/;
98 $myownurl =~ s/http:\/\/(.*):(.*)/$1/;
99 ($myownurl) = gethostbyname($myownurl);
100 $myownurl = "http://".$myownurl.":".$myownport;
102 tie(%xml_map, 'DB_File', $uris_dbm, O_RDWR, 0664); # open dbs
103 tie(%rdf_map, 'DB_File', $rdf_dbm, O_RDWR, 0664);
104 tie(%xslt_map, 'DB_File', $xslt_dbm, O_RDWR, 0664);
107 print "HTTP Getter $VERSION\n"; # print hello information
108 print "Please contact me at: <URL:", $myownurl."/", ">\n";
110 print "servers_file: $servers_file\n";
111 print "helm_dir: $helm_dir\n";
112 print "helm_rdf_dir: $helm_rdf_dir\n";
113 print "style_dir: $style_dir\n";
114 print "dtd_dir: $dtd_dir\n";
115 print "urls_of_uris db: $uris_dbm\n";
116 print "rdf db: $rdf_dbm\n";
117 print "xslt db: $xslt_dbm\n";
118 print "cache mode: $cachemode\n";
119 print "indexname: $indexname\n";
120 print "rdf_indexname: $rdf_indexname\n";
121 print "xslt_indexname: $xslt_indexname\n";
124 $SIG{CHLD} = "IGNORE"; # do not accumulate defunct processes
125 $SIG{USR1} = \&update; # sent by the child to make the parent update
127 while (my $c = $d->accept) {
129 while (my $r = $c->get_request) {
130 #CSC: mancano i controlli di sicurezza
132 my $http_method = $r->method;
133 my $http_path = $r->url->path;
134 my $http_query = uri_unescape($r->url->query);
135 my $cgi = new CGI("$http_query");
136 my $inputuri = $cgi->param('uri');
137 print "\nRequest: ".$r->url."\n\n";
139 print "\nUnescaped query: ".$http_query."\n";
141 # "getxml" works with rdf uris
142 if ($http_method eq 'GET' and $http_path eq "/getxml") {
143 # finds the uri, url and filename
145 my $answerformat = $cgi->param('format');
146 my $patch_dtd = $cgi->param('patch_dtd');
147 $answerformat = "" if (not defined($answerformat));
148 $patch_dtd = "yes" if (not defined($patch_dtd));
149 if (($answerformat ne "gz") and ($answerformat ne "normal")
150 and ($answerformat ne "")) {
151 die "Wrong output format: $answerformat, must be 'normal' ".
154 if (($patch_dtd ne "yes") and ($patch_dtd ne "no")
155 and ($patch_dtd ne "")) {
156 die "Wrong param, patch_dtd must be 'yes' or 'no'\n";
159 my $filename = $inputuri;
161 if (not isRdfUri($inputuri)) { # standad cic: or theory: uri
162 $filename =~ s/^cic:(.*)/$1/;
163 $filename =~ s/^theory:(.*)/$1/;
165 # The "helm:rdf/<<...>>//cic:" prefix is kept, but quoted
166 # (so that a "helm:rdf/<<...>>//cic:" dir is created
168 $filename =~ s/^(.*)\/\/cic:(.*)/$2/;
169 $filename =~ s/^(.*)\/\/theory:(.*)/$2/;
170 $prefix =~ s/\Q$filename\E//;
173 $filename = $prefix.$filename;
176 my $url = resolve ($inputuri); # resolve uri in url
177 if (not defined($url)) { # uri not found in uri2url map
178 die "NOT FOUND: uri \"$inputuri\" can't be resolved\n";
179 $c->send_error(RC_NOT_FOUND)
180 } else { # uri found and mapped to url
181 my $extension; # file extension
182 if ($url =~ /\.xml$/) { # non gzipped file
184 } elsif ($url =~ /\.xml\.gz$/) { # gzipped file
185 $extension = ".xml.gz";
186 } else { # error: unknown extension
187 die "unexpected extension in url: $url, might be '.xml'".
191 if (not isRdfUri ($inputuri)) { # save in uri std cache dir
192 $filename = $helm_dir.$filename.$extension;
193 } else { # save in rdf metadata cache dir
194 $filename = $helm_rdf_dir."/".$filename.$extension;
197 print_request($inputuri,$url,$filename);
200 my $ciccontent = download($patch_dtd,$url,$filename,$answerformat);
202 if ($answerformat eq "normal") { # answer the client in text/xml
203 answer($c,$ciccontent,"text/xml","");
204 } else { # answer the client in text/xml, gzip encoding
205 answer($c,$ciccontent,"text/xml","x-gzip");
209 # "/register" does not work with rdf uris
210 } elsif ($http_method eq 'GET' and $http_path eq "/register") {
211 my $inputurl = $cgi->param('url');
212 print "Register requested...\n";
213 $xml_map{$inputuri}=$inputurl;
215 # Now let's clean the cache
216 my $cicfilename = $inputuri;
217 $cicfilename =~ s/cic:(.*)/$1/;
218 $cicfilename =~ s/theory:(.*)/$1/;
220 print "Unlinking ".$helm_dir.$cicfilename.".xml[.gz]\n";
221 unlink ($helm_dir.$cicfilename.".xml");
222 unlink ($helm_dir.$cicfilename.".xml.gz");
224 kill(USR1,getppid()); # signal changes to parent
227 html_nice_answer($c,"Register done");
229 # "/resolve" works with rdf uri
230 } elsif ($http_method eq 'GET' and $http_path eq "/resolve") {
231 #my $outputurl = $xml_map{$inputuri};
232 my $outputurl = resolve($inputuri);
233 $outputurl = "" if (not defined($outputurl));
234 $cont = "<?xml version=\"1.0\" ?>\n\n";
235 if ($outputurl eq "") {
236 $cont .= "<unresolved />\n";
238 $cont .= "<url value=\"$outputurl\" />\n";
240 answer($c,$cont,"text/xml","");
242 } elsif ($http_method eq 'GET' and $http_path eq "/getdtd") {
244 my $filename = $inputuri;
245 $filename = $dtd_dir."/".$filename;
246 print "DTD: $inputuri ==> ($filename)\n";
247 if (stat($filename)) {
248 print "Using local copy\n";
249 open(FD, $filename) or die "Cannot open $filename\n";
252 # Vary bad heuristic for substituion of absolute URLs
254 s/ENTITY (.*) SYSTEM\s+"/ENTITY $1 SYSTEM "$myownurl\/getdtd?uri=/g;
258 answer($c,$cont,"text/xml","");
260 die "Could not find DTD!";
263 } elsif ($http_method eq 'GET' and $http_path eq "/getxslt") {
265 $url = $xslt_map{$inputuri};
266 print "Downloading XSLT ...\n";
267 helm_wget($tmp_dir, $url);
268 print "XSLT downloaded!\n";
269 $inputuri =~ s/.*\///;
270 my $filename = $tmp_dir . "/" . $inputuri;
272 die "Could not find XSLT!\n" unless stat($filename);
273 open(FD, $filename) or die "Can't open xslt local copy: '$filename'\n";
275 s/xsl:import\s+href="/xsl:import href="$myownurl\/getxslt?uri=/g;
276 s/xsl:include\s+href="/xsl:include href="$myownurl\/getxslt?uri=/g;
280 answer($c, $cont, "text/xml", "");
282 print "Can't unlink XSLT temp file '$filename'";
285 # my $filename = $inputuri;
286 # $filename = $style_dir."/".$filename;
287 # print "XSLT: $inputuri ==> ($filename)\n";
288 # if (stat($filename)) {
289 # print "Using local copy\n";
290 # open(FD, $filename) or die "Cannot open $filename\n";
293 # # Vary bad heuristic for substituion of absolute URLs
294 # # for relative ones
295 # s/xsl:import\s+href="/xsl:import href="$myownurl\/getxslt?uri=/g ;
296 # s/xsl:include\s+href="/xsl:include href="$myownurl\/getxslt?uri=/g ;
300 # answer($c,$cont,"text/xml","");
302 # die "Could not find XSLT!";
306 # "/update" works with rdf uri
307 } elsif ($http_method eq 'GET' and $http_path eq "/update") {
308 # rebuild urls_of_uris db and rdf uris db
309 print "Update requested...\n";
311 kill(USR1,getppid()); # signal changes to parent
313 html_nice_answer($c,"Update done");
315 # "/ls" does not work with rdf uris
316 } elsif ($http_method eq 'GET' and $http_path eq "/ls") {
317 # send back keys that begin with a given uri
318 my ($uritype,$uripattern,$uriprefix);
319 my $baseuri = $cgi->param('baseuri');
320 chop $baseuri if ($baseuri =~ /.*\/$/); # remove trailing "/"
321 # output type, might be 'txt' or 'xml':
322 my $outype = $cgi->param('format');
323 $uripattern = $baseuri;
324 $uripattern =~ s/^.*:(.*)/$1/;
325 if ($baseuri =~ /^cic:/) {
327 } elsif ($baseuri =~ /^theory:/) {
330 $uritype = "invalid";
332 if ($uritype ne "invalid") { # uri is valid
333 if (($outype ne 'txt') and ($outype ne 'xml')) { # invalid out type
334 print "Invalid output type specified: $outype\n";
335 html_nice_answer($c,"Invalid output type, must be ".
337 } else { # valid output
338 print "BASEURI $baseuri, FORMAT $outype\n";
339 $cont = finduris($uritype,$uripattern,$outype);
340 if ($outype eq 'txt') {
341 answer($c,$cont,"text/plain","");
342 } elsif ($outype eq 'xml') {
343 answer($c,$cont,"text/xml","");
345 die "Internal error, exiting!";
348 } else { # invalid uri
349 print "Invalid uri: $baseuri, may begin with 'cic:', ".
350 "'theory:' or '*:'\n";
351 html_nice_answer($c,"Invalid uri , must begin with ".
352 "'cic:' or 'theory:'");
355 } elsif ($http_method eq 'GET' and $http_path eq "/help") {
357 print "Help requested!\n";
358 html_nice_answer($c,"HTTP Getter Version: $VERSION");
360 } elsif ($http_method eq 'GET' and $http_path eq "/getempty") {
361 # return an empty, but valid, xml document
362 my $emptyxml = <<EOXML;
363 <?xml version="1.0"?>
365 <!ELEMENT empty EMPTY>
369 answer($c, $emptyxml, "", "");
371 } elsif ($http_method eq 'GET' and $http_path =~ /\.cgi$/) {
373 print "CGI requested!\n";
374 if ($http_path !~ /^\/[^\/]*\.cgi$/) {
375 html_nice_answer($c,"Invalid CGI name: $http_path, ".
376 "you can't request CGI that contain a slash in the name\n");
377 } elsif (stat "$cgi_dir"."$http_path") {
378 if (not -x "$cgi_dir/$http_path") {
379 html_nice_answer($c,"CGI $http_path found but not ".
381 } else { # exec CGI and anser back its output
384 $ENV{'QUERY_STRING'} = $http_query;
385 my $cgi_output = `$cgi_dir/$http_path`;
386 answer($c,$cgi_output,"","");
390 html_nice_answer($c,"CGI '$http_path' not found ".
391 "in CGI dir '$cgi_dir'");
394 } else { # unsupported request
396 print "INVALID REQUEST!!!!!\n";
397 print "(PATH: ",$http_path,", ";
398 print "QUERY: ",$http_query,")\n";
399 $c->send_error(RC_FORBIDDEN)
402 print "\nRequest solved: ".$r->url."\n\n";
407 print "\nCONNECTION CLOSED\n\n";
412 #================================
414 sub isRdfUri { # return true if the uri is an rdf uri, false otherwise
416 # helm:rdf/cic:www.cs.unibo.it/helm/rdf/foo_schema//cic:\
417 # /Coq/Init/Logic/True_rec.con.types
419 # the format is "helm:rdf/<metadata_tree>:<metadata_scheme>//<xml_file_uri>"
422 if ($uri =~ /^helm:rdf(.*):(.*)\/\/(.*)/) {
429 sub resolve { # resolve an uri in a url, work both with standard cic: or theory:
431 print "RESOLVE subroutine\n";
433 print "GIVEN URI: \"$uri\"\n";
434 if (isRdfUri ($uri)) { # rdf uri, resolve using rdf db
435 print "IS A RDF URI\n";
436 print "I WILL RETURN '$rdf_map{$uri}'\n";
437 return ($rdf_map{$uri});
438 } else { # standard cic: or theory: uri, resolve using std uri db
439 print "IS NOT A RDF URI\n";
440 print "I WILL RETURN '$xml_map{$uri}'\n";
441 return ($xml_map{$uri});
445 sub finduris { # find uris for cic and theory trees generation
446 my ($uritype,$uripattern,$format) = @_;
448 my ($uri,$localpart,$basepart,$dirname,$suffix,$flags,$key);
449 my (@itemz,@already_pushed_dir);
450 my (%objects,%dirs); # map uris to suffixes' flags
451 my $debug=1; # for debug
453 print "FINDURIS, uritype: $uritype, uripattern: $uripattern, ".
454 "format: $format\n\n" if defined($debug);
456 if (($uritype eq "cic") or ($uritype eq "theory")) {
457 # get info only of one type: cic or theory
458 foreach (keys(%xml_map)) { # select matching uris
460 if ($uri =~ /^$uritype:$uripattern(\/|$|\.)/) {
461 if ($uri =~ /^$uritype:$uripattern\//) { # directory match
463 $localpart =~ s/^$uritype:$uripattern\/(.*)/$1/;
464 } elsif ($uri =~ /^$uritype:$uripattern($|\.)/) { # file match
466 $localpart =~ s/^.*\/([^\/]*)/$1/;
468 die "Internal error, seems that requested match is none of ".
469 "directory match or file match";
471 print "LOCALPART: $localpart\n" if defined($debug);
473 if ($localpart =~ /^[^\/]*$/) { # no slash, an OBJECT
474 $basepart = $localpart;
475 $basepart =~ s/^([^.]*\.[^.]*)(\.types)?(\.ann)?/$1/;
476 # remove exts .types or
478 $flags = $objects{$basepart}; # get old flags
479 if ($localpart =~ /\.ann$/) {
480 $flags = add_flag("ann","YES",$flags);
482 $flags = add_flag("ann","NO",$flags);
484 if ($localpart =~ /\.types$/) {
485 $flags = add_flag("types","YES",$flags);
486 } elsif ($localpart =~ /\.types\.ann$/) {
487 $flags = add_flag("types","ANN",$flags);
489 $flags = add_flag("types","NO",$flags);
491 $objects{$basepart} = $flags; # save new flags
492 } else { # exists at least one slash, a DIR
493 ($dirname) = split (/\//, $localpart);
494 $dirs{$dirname} = ""; # no flags requirement for dir
499 die "getter internal error: unsupported uritype: \"$uritype\"";
501 # now generate OUTPUT:
502 # output will be saved in $content
503 if ($format eq "txt") { # TXT output
504 foreach $key (sort(keys %dirs)) {
505 $content .= "dir, " . $key . "\n";
507 foreach $key (sort(keys %objects)) {
508 $content .= "object, $key, " . $objects{$key} . "\n";
510 } elsif ($format eq "xml") { # XML output
511 $content .= '<?xml version="1.0" encoding="ISO-8859-1"?>' . "\n";
512 $content .= "<!DOCTYPE ls SYSTEM ";
513 $content .= "\"$myownurl/getdtd?uri=ls.dtd\">" . "\n\n";
514 $content .= "<ls>\n";
515 foreach $key (sort(keys %dirs)) {
516 $content .= "\t<section>$key</section>\n";
518 foreach $key (sort(keys %objects)) {
519 $content .= "\t<object name=\"$key\">\n";
520 $flags = $objects{$key};
521 $flags =~ s/^<(.*)>$/$1/;
522 my ($annflag,$typesflag) = split /,/,$flags;
523 $content .= "\t\t<ann value=\"$annflag\" />\n";
524 $content .= "\t\t<types value=\"$typesflag\" />\n";
525 $content .= "\t</object>\n";
527 $content .= "</ls>\n";
528 } else { # may not enter this branch
529 die "Getter internal error: invalid format received by finduris sub";
535 # manage string like: "<ann_flag,type_flag>"
536 # "ann_flag" may be one of "ann_YES", "ann_NO"
537 # "type_flag" may be one of "types_NO", "types_YES", "types_ANN"
538 # when adding a flag the max between the current flag and the new flag
539 # is taken, the orders are ann_NO < ann_YES and types_NO < types_YES <
541 my ($flagtype,$newflag,$str) = @_;
542 $str = "<,>" if ($str eq "");
543 ($str =~ s/^<(.*,.*)>$/$1/) or die "Internal error: ".
544 "wrong string format for flag adding in $str";
545 my ($annflag,$typeflag) = split /,/,$str;
546 if ($flagtype eq "ann") { # ANN flag handling
547 if ($newflag eq "YES") {
549 } elsif ($newflag eq "NO") {
550 $annflag = "NO" unless ($annflag eq "YES");
552 die "Internal error: annflag must be \"YES\" or \"NO\"";
554 } elsif ($flagtype eq "types") { # TYPES flag handling
555 if ($newflag eq "ANN") {
557 } elsif ($newflag eq "YES") {
558 $typeflag = "YES" unless ($typeflag eq "ANN");
559 } elsif ($newflag eq "NO") {
561 unless (($typeflag eq "ANN") or ($typeflag eq "YES"));
563 die "Internal error: typeflag must be \"YES\", \"NO\" or \"ANN\"";
566 die "Internal error: unsupported flagtype \"$flagtype\"";
568 $str = "<$annflag,$typeflag>";
571 #CSC: Too much powerful: creates even /home, /home/users/, ...
572 #CSC: Does not raise errors if could not create dirs/files
576 my @dirs = split /\//,$pathname;
578 foreach $dir (@dirs) {
579 $tmp = ((defined($tmp)) ? $tmp."\/".$dir : "");
587 my ($uri,$url,$filename) = @_;
588 print "<download_request>\n";
591 print "filename: $filename\n\n";
592 print "</download_request>\n";
595 sub gunzip { # gunzip a file and return the deflated content
598 my ($gz, $buffer, $cont);
600 print "deflating $filename ...\n";
601 $gz = gzopen($filename, "r")
602 or die "Cannot open gzip'ed file $filename: $gzerrno";
604 while ( $gz->gzread($buffer) > 0 ) {
607 die "Error while reading : $gzerrno\n" if $gzerrno != Z_STREAM_END ;
613 sub gzip { # gzip the content argument and save it to filename argument
614 my ($cont, $filename) = @_;
615 #my $debug=1; # debug only
617 print "gzopening $filename ...\n" if (defined($debug));
618 my $gz = gzopen($filename, "w")
619 or die "Cannot gzopen for writing file $filename: $gzerrno";
620 print "gzwriting ...\n" if (defined($debug));
621 $gz->gzwrite($cont) or die "error writing: $gzerrno, exiting!\n";
622 print "gzclosing ...\n" if (defined($debug));
627 my ($patch_dtd,$url,$filename,$format) = @_;
630 # print "DOWNLOAD subs receives url: \"$url\"\n";
631 # print "DOWNLOAD subs receives filename: \"$filename\"\n";
633 my $resourcetype; # retrieve mode: "normal" (.xml) or "gzipped" (.xml.gz)
634 if ($filename =~ /\.xml$/) { # set retrieve mode
635 $resourcetype = "normal";
636 } elsif ($filename =~ /\.xml\.gz$/) {
637 $resourcetype = "gzipped";
639 die "Unsupported download extension, might be '.gz' or '.xml'\n";
641 my $basefname = $filename;
642 $basefname =~ s/\.gz$//; # get base resource name removing trailing .gz
643 $cont = ""; # modified by side-effect by the callback function
646 if (stat($basefname)) {
647 $localfname=$basefname;
648 } elsif (stat($basefname.".gz")) {
649 $localfname=$basefname.".gz";
651 if ($localfname ne "") { # we already have local copy of requested file
652 # check both possible cache entry: gzipped or normal
653 print "Using local copy.\n";
654 if ($localfname =~ /\.xml\.gz$/) { # deflating cached file and return it
655 $cont = gunzip($localfname);
656 } elsif ($localfname =~ /\.xml$/) { # just return cached file
657 open(FD, $localfname) or die "Cannot open $localfname";
658 while(<FD>) { $cont .= $_; }
661 die "Internal error: unexpected file name $localfname,"
662 ."must end with '.gz' or '.xml.gz'\n";
664 } else { # download file from net
665 print "Downloading file ...\n"; # download file
666 my $ua = LWP::UserAgent->new;
667 my $request = HTTP::Request->new(GET => "$url");
668 my $response = $ua->request($request, \&callback);
670 # cache retrieved file to disk
671 # <ZACK/> TODO: inefficent, I haven't yet undestood how to deflate
672 # in memory gzipped file, without call "gzopen"
673 print "Storing file ...\n";
674 # print "Making dirs ...\n";
676 # print "Opening tmp file for writing ...\n";
677 open(FD, ">".$filename.".tmp") or die "Cannot open $filename.tmp\n";
678 # print "Writing on tmp file ...\n";
680 # print "Closing tmp file ...\n";
683 # handle cache conversion normal->gzipped or gzipped->normal as user choice
684 print "cachemode:$cachemode, resourcetype:$resourcetype\n"
685 if (defined($debug));
686 if (($cachemode eq 'normal') and ($resourcetype eq 'normal')) {
687 # cache the file as is
688 rename "$filename.tmp", $filename;
689 } elsif (($cachemode eq 'gzipped') and ($resourcetype eq 'gzipped')) {
690 # cache the file as is
691 # and update the $cont variabile with deflated content
692 rename "$filename.tmp", $filename;
693 $cont = gunzip($filename);
694 } elsif (($cachemode eq 'normal') and ($resourcetype eq 'gzipped')) {
695 # deflate cache entry
697 open(FD, "> $basefname") or die "cannot open $basefname\n";
698 $cont = gunzip($filename.".tmp");
701 unlink "$filename.tmp"; # delete old gzipped file
702 } elsif (($cachemode eq 'gzipped') and ($resourcetype eq 'normal')) {
703 # compress cache entry
704 print "gzipping ...\n" if (defined($debug));
705 gzip($cont, $basefname.".gz");
706 unlink "$filename.tmp"; # delete old uncompressed file
708 die "Internal error, unsopported cachemode, resourcetype couple\n";
710 # $cont now contained uncompressed data
712 if ($patch_dtd eq "yes") {
713 $cont =~ s/DOCTYPE (.*) SYSTEM\s+"http:\/\/www.cs.unibo.it\/helm\/dtd\//DOCTYPE $1 SYSTEM "$myownurl\/getdtd?uri=/g;
715 if ($format eq "gz") {
716 gzip($cont,"$basefname.tmp");
717 open (TMP, "< $basefname.tmp")
718 or die "Can't open tempfile: $filename.tmp, exiting!\n";
724 unlink ($basefname . ".tmp") or
725 die "cannot unlink temporary file: $basefname.tmp\n";
732 # answer the client setting content, Content-Type and Content-Enconding
734 my ($c,$cont,$contype,$contenc) = @_;
735 my $res = new HTTP::Response;
736 $res->content($cont);
737 $res->push_header("Content-Type" => $contype)
738 unless ($contype eq "");
739 $res->push_header("Content-Encoding" => $contenc)
740 unless ($contenc eq "");
741 $res->push_header("Cache-Control" => "no-cache");
742 $res->push_header("Pragma" => "no-cache");
743 $res->push_header("Expires" => "0");
744 $c->send_response($res);
747 sub html_nice_answer {
748 # answer the client whith a nice html document
749 my ($c,$content) = @_;
750 $content = "<html><body><h1>$content</h1></body></html>";
751 answer($c,$content,"text/html","");
760 # retrieve a file from an url and write it to a temp dir
761 # used for retrieve resource index from servers
763 my ($prefix, $URL) = @_;
764 my $ua = LWP::UserAgent->new;
765 my $request = HTTP::Request->new(GET => "$URL");
766 my $response = $ua->request($request, \&callback);
767 my ($filename) = reverse (split "/", $URL); # get filename part of the URL
768 open (TEMP, "> $prefix/$filename")
769 || die "Cannot open temporary file: $prefix/$filename\n";
777 tie(%xml_map, 'DB_File', $uris_dbm, O_RDWR, 0664);
779 tie(%rdf_map, 'DB_File', $rdf_dbm, O_RDWR, 0664);
781 tie(%xslt_map, 'DB_File', $xslt_dbm, O_RDWR, 0664);
785 # rebuild dbs fetching resource indexes from servers.
786 # Rebuild urls_of_uris db (1), rdf db (2) and xslt db (3)
788 $server, $idxfile, $rdf_idxfile, $uri, $url, $comp, $line,
790 %urls_of_uris, %rdf, %xslt # local proxy of dbs
796 foreach my $file ($uris_dbm, $rdf_dbm, $xslt_dbm) { # remove old db file
797 if (stat $file) { # remove if exists
798 unlink ($file) or die "can't unlink old db file: $file\n";
801 tie(%urls_of_uris, 'DB_File', $uris_dbm, O_RDWR|O_CREAT, 0664);
802 tie(%rdf, 'DB_File', $rdf_dbm, O_RDWR|O_CREAT, 0664);
803 tie(%xslt, 'DB_File', $xslt_dbm, O_RDWR|O_CREAT, 0664);
805 open (SRVS, "< $servers_file") or
806 die "cannot open servers file: $servers_file\n";
807 @servers = <SRVS>; # read all servers
810 while ($server = pop @servers) { # cicle on servers in _reverse_ order
811 print "processing server: $server ...\n";
813 helm_wget($tmp_dir, $server."/".$indexname); # get index
814 helm_wget($tmp_dir, $server."/".$rdf_indexname); # get rdf index
815 helm_wget($tmp_dir, $server."/".$xslt_indexname); # get xslt index
816 $idxfile = $tmp_dir."/".$indexname;
817 $rdf_idxfile = $tmp_dir."/".$rdf_indexname;
818 $xslt_idxfile = $tmp_dir . "/" . $xslt_indexname;
820 # (1) REBUILD URLS_OF_URIS DB
821 open (INDEX, "< $idxfile") or
822 die "cannot open temporary index file: $idxfile\n";
823 while (my $line = <INDEX>) { # parse index and add entry to urls_of_uris db
825 ($uri,$comp) = split /[ \t]+/, $line;
828 $url = $uri . ".xml" . ".gz";
830 $url = $uri . ".xml";
832 $url =~ s/cic:/$server/;
833 $url =~ s/theory:/$server/;
834 $urls_of_uris{$uri} = $url;
838 open (RDF_INDEX, "< $rdf_idxfile") or
839 die "cannot open temporary RDF index file: $rdf_idxfile\n";
840 while (my $line = <RDF_INDEX>) { # parse index and add entry to rdf db
842 ($uri,$comp) = split /[ \t]+/, $line; # comp is "gz" or nothing
845 $url = $uri . ".xml" . ".gz";
847 $url = $uri . ".xml";
849 $url =~ s/^helm:rdf(.*)\/\/cic:/$server/;
850 $url =~ s/^helm:rdf(.*)\/\/theory:/$server/;
854 # (3) REBUILD XSLT DB
855 open (XSLT_INDEX, "< $xslt_idxfile") or
856 die "cannot open temporary XSLT index file: $xslt_idxfile\n";
857 while (my $line = <XSLT_INDEX>) { # parse index and add entry to xslt db
859 my $stylesheet = $line;
860 my $url = $server . "/" . $stylesheet;
861 $xslt{$stylesheet} = $url;
864 close INDEX; # close indexes
869 foreach my $file ($idxfile, $rdf_idxfile, $xslt_idxfile) {
870 print "cannot unlink temporary file: $file\n" if (unlink $file) != 1;
875 untie(%urls_of_uris); # untie local proxies
878 tie(%xml_map, 'DB_File', $uris_dbm, O_RDWR, 0664); # retie global ones
879 tie(%rdf_map, 'DB_File', $rdf_dbm, O_RDWR, 0664);
880 tie(%xslt_map, 'DB_File', $xslt_dbm, O_RDWR, 0664);
884 # vim modline: do not remove!