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;
69 $uris_dbm = $ENV{'HTTP_GETTER_URIS_DBM'} || $uris_dbm;
71 # </move_to_conf_file>
73 # Let's override the configuration file
74 $style_dir = $ENV{"HELM_STYLE_DIR"} if (defined ($ENV{"HELM_STYLE_DIR"}));
75 $dtd_dir = $ENV{"HELM_DTD_DIR"} if (defined ($ENV{"HELM_DTD_DIR"}));
87 #CSC: mancano i controlli sulle condizioni di errore di molte funzioni
88 #CSC: ==> non e' robusto
89 #CSC: altra roba da sistemare segnata con CSC
91 my $d = new HTTP::Daemon LocalPort => 8081
92 or die "Error: port 8081 not available.";
93 my $myownurl = $d->url;
95 # Let's patch the returned URL
96 $myownurl =~ s/\/$//; # chop the final slash
97 my $myownport = $myownurl;
98 $myownport =~ s/http:\/\/(.*):(.*)/$2/;
99 $myownurl =~ s/http:\/\/(.*):(.*)/$1/;
100 ($myownurl) = gethostbyname($myownurl);
101 $myownurl = "http://".$myownurl.":".$myownport;
103 tie(%xml_map, 'DB_File', $uris_dbm, O_RDWR, 0664); # open dbs
104 tie(%rdf_map, 'DB_File', $rdf_dbm, O_RDWR, 0664);
105 tie(%xslt_map, 'DB_File', $xslt_dbm, O_RDWR, 0664);
108 print "HTTP Getter $VERSION\n"; # print hello information
109 print "Please contact me at: <URL:", $myownurl."/", ">\n";
111 print "servers_file: $servers_file\n";
112 print "helm_dir: $helm_dir\n";
113 print "helm_rdf_dir: $helm_rdf_dir\n";
114 print "style_dir: $style_dir\n";
115 print "dtd_dir: $dtd_dir\n";
116 print "urls_of_uris db: $uris_dbm\n";
117 print "rdf db: $rdf_dbm\n";
118 print "xslt db: $xslt_dbm\n";
119 print "cache mode: $cachemode\n";
120 print "indexname: $indexname\n";
121 print "rdf_indexname: $rdf_indexname\n";
122 print "xslt_indexname: $xslt_indexname\n";
125 $SIG{CHLD} = "IGNORE"; # do not accumulate defunct processes
126 $SIG{USR1} = \&update; # sent by the child to make the parent update
128 while (my $c = $d->accept) {
130 while (my $r = $c->get_request) {
131 #CSC: mancano i controlli di sicurezza
133 my $http_method = $r->method;
134 my $http_path = $r->url->path;
135 my $http_query = uri_unescape($r->url->query);
136 my $cgi = new CGI("$http_query");
137 my $inputuri = $cgi->param('uri');
138 print "\nRequest: ".$r->url."\n\n";
140 print "\nUnescaped query: ".$http_query."\n";
142 # "getxml" works with rdf uris
143 if ($http_method eq 'GET' and $http_path eq "/getxml") {
144 # finds the uri, url and filename
146 my $answerformat = $cgi->param('format');
147 my $patch_dtd = $cgi->param('patch_dtd');
148 $answerformat = "" if (not defined($answerformat));
149 $patch_dtd = "yes" if (not defined($patch_dtd));
150 if (($answerformat ne "gz") and ($answerformat ne "normal")
151 and ($answerformat ne "")) {
152 die "Wrong output format: $answerformat, must be 'normal' ".
155 if (($patch_dtd ne "yes") and ($patch_dtd ne "no")
156 and ($patch_dtd ne "")) {
157 die "Wrong param, patch_dtd must be 'yes' or 'no'\n";
160 my $filename = $inputuri;
162 if (not isRdfUri($inputuri)) { # standad cic: or theory: uri
163 $filename =~ s/^cic:(.*)/$1/;
164 $filename =~ s/^theory:(.*)/$1/;
166 # The "helm:rdf/<<...>>//cic:" prefix is kept, but quoted
167 # (so that a "helm:rdf/<<...>>//cic:" dir is created
169 $filename =~ s/^(.*)\/\/cic:(.*)/$2/;
170 $filename =~ s/^(.*)\/\/theory:(.*)/$2/;
171 $prefix =~ s/\Q$filename\E//;
174 $filename = $prefix.$filename;
177 my $url = resolve ($inputuri); # resolve uri in url
178 if (not defined($url)) { # uri not found in uri2url map
179 die "NOT FOUND: uri \"$inputuri\" can't be resolved\n";
180 $c->send_error(RC_NOT_FOUND)
181 } else { # uri found and mapped to url
182 my $extension; # file extension
183 if ($url =~ /\.xml$/) { # non gzipped file
185 } elsif ($url =~ /\.xml\.gz$/) { # gzipped file
186 $extension = ".xml.gz";
187 } else { # error: unknown extension
188 die "unexpected extension in url: $url, might be '.xml'".
192 if (not isRdfUri ($inputuri)) { # save in uri std cache dir
193 $filename = $helm_dir.$filename.$extension;
194 } else { # save in rdf metadata cache dir
195 $filename = $helm_rdf_dir."/".$filename.$extension;
198 print_request($inputuri,$url,$filename);
201 my $ciccontent = download($patch_dtd,$url,$filename,$answerformat);
203 if ($answerformat eq "normal") { # answer the client in text/xml
204 answer($c,$ciccontent,"text/xml","");
205 } else { # answer the client in text/xml, gzip encoding
206 answer($c,$ciccontent,"text/xml","x-gzip");
210 # "/register" does not work with rdf uris
211 } elsif ($http_method eq 'GET' and $http_path eq "/register") {
212 my $inputurl = $cgi->param('url');
213 print "Register requested...\n";
214 $xml_map{$inputuri}=$inputurl;
216 # Now let's clean the cache
217 my $cicfilename = $inputuri;
218 $cicfilename =~ s/cic:(.*)/$1/;
219 $cicfilename =~ s/theory:(.*)/$1/;
221 print "Unlinking ".$helm_dir.$cicfilename.".xml[.gz]\n";
222 unlink ($helm_dir.$cicfilename.".xml");
223 unlink ($helm_dir.$cicfilename.".xml.gz");
225 kill(USR1,getppid()); # signal changes to parent
228 html_nice_answer($c,"Register done");
230 # "/resolve" works with rdf uri
231 } elsif ($http_method eq 'GET' and $http_path eq "/resolve") {
232 #my $outputurl = $xml_map{$inputuri};
233 my $outputurl = resolve($inputuri);
234 $outputurl = "" if (not defined($outputurl));
235 $cont = "<?xml version=\"1.0\" ?>\n\n";
236 if ($outputurl eq "") {
237 $cont .= "<unresolved />\n";
239 $cont .= "<url value=\"$outputurl\" />\n";
241 answer($c,$cont,"text/xml","");
243 } elsif ($http_method eq 'GET' and $http_path eq "/getdtd") {
245 my $filename = $inputuri;
246 $filename = $dtd_dir."/".$filename;
247 print "DTD: $inputuri ==> ($filename)\n";
248 if (stat($filename)) {
249 print "Using local copy\n";
250 open(FD, $filename) or die "Cannot open $filename\n";
253 # Vary bad heuristic for substituion of absolute URLs
255 s/ENTITY (.*) SYSTEM\s+"/ENTITY $1 SYSTEM "$myownurl\/getdtd?uri=/g;
259 answer($c,$cont,"text/xml","");
261 die "Could not find DTD!";
264 } elsif ($http_method eq 'GET' and $http_path eq "/getxslt") {
266 $url = $xslt_map{$inputuri};
267 print "Downloading XSLT ...\n";
268 helm_wget($tmp_dir, $url);
269 print "XSLT downloaded!\n";
270 $inputuri =~ s/.*\///;
271 my $filename = $tmp_dir . "/" . $inputuri;
273 die "Could not find XSLT!\n" unless stat($filename);
274 open(FD, $filename) or die "Can't open xslt local copy: '$filename'\n";
276 s/xsl:import\s+href="/xsl:import href="$myownurl\/getxslt?uri=/g;
277 s/xsl:include\s+href="/xsl:include href="$myownurl\/getxslt?uri=/g;
281 answer($c, $cont, "text/xml", "");
283 print "Can't unlink XSLT temp file '$filename'";
286 # my $filename = $inputuri;
287 # $filename = $style_dir."/".$filename;
288 # print "XSLT: $inputuri ==> ($filename)\n";
289 # if (stat($filename)) {
290 # print "Using local copy\n";
291 # open(FD, $filename) or die "Cannot open $filename\n";
294 # # Vary bad heuristic for substituion of absolute URLs
295 # # for relative ones
296 # s/xsl:import\s+href="/xsl:import href="$myownurl\/getxslt?uri=/g ;
297 # s/xsl:include\s+href="/xsl:include href="$myownurl\/getxslt?uri=/g ;
301 # answer($c,$cont,"text/xml","");
303 # die "Could not find XSLT!";
307 # "/update" works with rdf uri
308 } elsif ($http_method eq 'GET' and $http_path eq "/update") {
309 # rebuild urls_of_uris db and rdf uris db
310 print "Update requested...\n";
312 kill(USR1,getppid()); # signal changes to parent
314 html_nice_answer($c,"Update done");
316 # "/getalluris" does not work with rdf uris
317 } elsif ($http_method eq 'GET' and $http_path eq "/getalluris") {
318 # send back all the keys in xml
319 print "BASEURI $baseuri, FORMAT $outype\n";
320 $cont = getalluris();
321 answer($c,$cont,"text/xml","");
323 } elsif ($http_method eq 'GET' and $http_path eq "/ls") {
324 # send back keys that begin with a given uri
325 my ($uritype,$uripattern,$uriprefix);
326 my $baseuri = $cgi->param('baseuri');
327 chop $baseuri if ($baseuri =~ /.*\/$/); # remove trailing "/"
328 # output type, might be 'txt' or 'xml':
329 my $outype = $cgi->param('format');
330 $uripattern = $baseuri;
331 $uripattern =~ s/^.*:(.*)/$1/;
332 if ($baseuri =~ /^cic:/) {
334 } elsif ($baseuri =~ /^theory:/) {
337 $uritype = "invalid";
339 if ($uritype ne "invalid") { # uri is valid
340 if (($outype ne 'txt') and ($outype ne 'xml')) { # invalid out type
341 print "Invalid output type specified: $outype\n";
342 html_nice_answer($c,"Invalid output type, must be ".
344 } else { # valid output
345 print "BASEURI $baseuri, FORMAT $outype\n";
346 $cont = finduris($uritype,$uripattern,$outype);
347 if ($outype eq 'txt') {
348 answer($c,$cont,"text/plain","");
349 } elsif ($outype eq 'xml') {
350 answer($c,$cont,"text/xml","");
352 die "Internal error, exiting!";
355 } else { # invalid uri
356 print "Invalid uri: $baseuri, may begin with 'cic:', ".
357 "'theory:' or '*:'\n";
358 html_nice_answer($c,"Invalid uri , must begin with ".
359 "'cic:' or 'theory:'");
362 } elsif ($http_method eq 'GET' and $http_path eq "/help") {
364 print "Help requested!\n";
365 html_nice_answer($c,"HTTP Getter Version: $VERSION");
367 } elsif ($http_method eq 'GET' and $http_path eq "/getempty") {
368 # return an empty, but valid, xml document
369 my $emptyxml = <<EOXML;
370 <?xml version="1.0"?>
372 <!ELEMENT empty EMPTY>
376 answer($c, $emptyxml, "", "");
378 } elsif ($http_method eq 'GET' and $http_path =~ /\.cgi$/) {
380 print "CGI requested!\n";
381 if ($http_path !~ /^\/[^\/]*\.cgi$/) {
382 html_nice_answer($c,"Invalid CGI name: $http_path, ".
383 "you can't request CGI that contain a slash in the name\n");
384 } elsif (stat "$cgi_dir"."$http_path") {
385 if (not -x "$cgi_dir/$http_path") {
386 html_nice_answer($c,"CGI $http_path found but not ".
388 } else { # exec CGI and anser back its output
391 $ENV{'QUERY_STRING'} = $http_query;
392 my $cgi_output = `$cgi_dir/$http_path`;
393 answer($c,$cgi_output,"","");
397 html_nice_answer($c,"CGI '$http_path' not found ".
398 "in CGI dir '$cgi_dir'");
401 } else { # unsupported request
403 print "INVALID REQUEST!!!!!\n";
404 print "(PATH: ",$http_path,", ";
405 print "QUERY: ",$http_query,")\n";
406 $c->send_error(RC_FORBIDDEN)
409 print "\nRequest solved: ".$r->url."\n\n";
414 print "\nCONNECTION CLOSED\n\n";
419 #================================
421 sub isRdfUri { # return true if the uri is an rdf uri, false otherwise
423 # helm:rdf/cic:www.cs.unibo.it/helm/rdf/foo_schema//cic:\
424 # /Coq/Init/Logic/True_rec.con.types
426 # the format is "helm:rdf/<metadata_tree>:<metadata_scheme>//<xml_file_uri>"
429 if ($uri =~ /^helm:rdf(.*):(.*)\/\/(.*)/) {
436 sub resolve { # resolve an uri in a url, work both with standard cic: or theory:
438 print "RESOLVE subroutine\n";
440 print "GIVEN URI: \"$uri\"\n";
441 if (isRdfUri ($uri)) { # rdf uri, resolve using rdf db
442 print "IS A RDF URI\n";
443 print "I WILL RETURN '$rdf_map{$uri}'\n";
444 return ($rdf_map{$uri});
445 } else { # standard cic: or theory: uri, resolve using std uri db
446 print "IS NOT A RDF URI\n";
447 print "I WILL RETURN '$xml_map{$uri}'\n";
448 return ($xml_map{$uri});
452 sub getalluris { # get all the keys whose prefix is cic
455 my $debug=1; # for debug
457 $content .= '<?xml version="1.0" encoding="ISO-8859-1"?>' . "\n";
458 $content .= "<!DOCTYPE alluris SYSTEM ";
459 $content .= "\"$myownurl/getdtd?uri=alluris.dtd\">" . "\n\n";
460 $content .= "<alluris>\n";
461 foreach $uri (sort (keys(%xml_map))) { # select matching uris
462 if ($uri =~ /^cic:/ && not $uri =~ /.types$/) {
463 print "GETALLURI: $uri\n" if defined($debug);
464 $content .= "\t<uri value=\"$uri\"/>\n";
467 $content .= "</alluris>\n";
471 sub finduris { # find uris for cic and theory trees generation
472 my ($uritype,$uripattern,$format) = @_;
474 my ($uri,$localpart,$basepart,$dirname,$suffix,$flags,$key);
475 my (@itemz,@already_pushed_dir);
476 my (%objects,%dirs); # map uris to suffixes' flags
477 my $debug=1; # for debug
479 print "FINDURIS, uritype: $uritype, uripattern: $uripattern, ".
480 "format: $format\n\n" if defined($debug);
482 if (($uritype eq "cic") or ($uritype eq "theory")) {
483 # get info only of one type: cic or theory
484 foreach (keys(%xml_map)) { # select matching uris
486 if ($uri =~ /^$uritype:$uripattern(\/|$|\.)/) {
487 if ($uri =~ /^$uritype:$uripattern\//) { # directory match
489 $localpart =~ s/^$uritype:$uripattern\/(.*)/$1/;
490 } elsif ($uri =~ /^$uritype:$uripattern($|\.)/) { # file match
492 $localpart =~ s/^.*\/([^\/]*)/$1/;
494 die "Internal error, seems that requested match is none of ".
495 "directory match or file match";
497 print "LOCALPART: $localpart\n" if defined($debug);
499 if ($localpart =~ /^[^\/]*$/) { # no slash, an OBJECT
500 $basepart = $localpart;
501 $basepart =~ s/^([^.]*\.[^.]*)(\.types)?(\.ann)?/$1/;
502 # remove exts .types or
504 $flags = $objects{$basepart}; # get old flags
505 if ($localpart =~ /\.ann$/) {
506 $flags = add_flag("ann","YES",$flags);
508 $flags = add_flag("ann","NO",$flags);
510 if ($localpart =~ /\.types$/) {
511 $flags = add_flag("types","YES",$flags);
512 } elsif ($localpart =~ /\.types\.ann$/) {
513 $flags = add_flag("types","ANN",$flags);
515 $flags = add_flag("types","NO",$flags);
517 $objects{$basepart} = $flags; # save new flags
518 } else { # exists at least one slash, a DIR
519 ($dirname) = split (/\//, $localpart);
520 $dirs{$dirname} = ""; # no flags requirement for dir
525 die "getter internal error: unsupported uritype: \"$uritype\"";
527 # now generate OUTPUT:
528 # output will be saved in $content
529 if ($format eq "txt") { # TXT output
530 foreach $key (sort(keys %dirs)) {
531 $content .= "dir, " . $key . "\n";
533 foreach $key (sort(keys %objects)) {
534 $content .= "object, $key, " . $objects{$key} . "\n";
536 } elsif ($format eq "xml") { # XML output
537 $content .= '<?xml version="1.0" encoding="ISO-8859-1"?>' . "\n";
538 $content .= "<!DOCTYPE ls SYSTEM ";
539 $content .= "\"$myownurl/getdtd?uri=ls.dtd\">" . "\n\n";
540 $content .= "<ls>\n";
541 foreach $key (sort(keys %dirs)) {
542 $content .= "\t<section>$key</section>\n";
544 foreach $key (sort(keys %objects)) {
545 $content .= "\t<object name=\"$key\">\n";
546 $flags = $objects{$key};
547 $flags =~ s/^<(.*)>$/$1/;
548 my ($annflag,$typesflag) = split /,/,$flags;
549 $content .= "\t\t<ann value=\"$annflag\" />\n";
550 $content .= "\t\t<types value=\"$typesflag\" />\n";
551 $content .= "\t</object>\n";
553 $content .= "</ls>\n";
554 } else { # may not enter this branch
555 die "Getter internal error: invalid format received by finduris sub";
561 # manage string like: "<ann_flag,type_flag>"
562 # "ann_flag" may be one of "ann_YES", "ann_NO"
563 # "type_flag" may be one of "types_NO", "types_YES", "types_ANN"
564 # when adding a flag the max between the current flag and the new flag
565 # is taken, the orders are ann_NO < ann_YES and types_NO < types_YES <
567 my ($flagtype,$newflag,$str) = @_;
568 $str = "<,>" if ($str eq "");
569 ($str =~ s/^<(.*,.*)>$/$1/) or die "Internal error: ".
570 "wrong string format for flag adding in $str";
571 my ($annflag,$typeflag) = split /,/,$str;
572 if ($flagtype eq "ann") { # ANN flag handling
573 if ($newflag eq "YES") {
575 } elsif ($newflag eq "NO") {
576 $annflag = "NO" unless ($annflag eq "YES");
578 die "Internal error: annflag must be \"YES\" or \"NO\"";
580 } elsif ($flagtype eq "types") { # TYPES flag handling
581 if ($newflag eq "ANN") {
583 } elsif ($newflag eq "YES") {
584 $typeflag = "YES" unless ($typeflag eq "ANN");
585 } elsif ($newflag eq "NO") {
587 unless (($typeflag eq "ANN") or ($typeflag eq "YES"));
589 die "Internal error: typeflag must be \"YES\", \"NO\" or \"ANN\"";
592 die "Internal error: unsupported flagtype \"$flagtype\"";
594 $str = "<$annflag,$typeflag>";
597 #CSC: Too much powerful: creates even /home, /home/users/, ...
598 #CSC: Does not raise errors if could not create dirs/files
602 my @dirs = split /\//,$pathname;
604 foreach $dir (@dirs) {
605 $tmp = ((defined($tmp)) ? $tmp."\/".$dir : "");
613 my ($uri,$url,$filename) = @_;
614 print "<download_request>\n";
617 print "filename: $filename\n\n";
618 print "</download_request>\n";
621 sub gunzip { # gunzip a file and return the deflated content
624 my ($gz, $buffer, $cont);
626 print "deflating $filename ...\n";
627 $gz = gzopen($filename, "r")
628 or die "Cannot open gzip'ed file $filename: $gzerrno";
630 while ( $gz->gzread($buffer) > 0 ) {
633 die "Error while reading : $gzerrno\n" if $gzerrno != Z_STREAM_END ;
639 sub gzip { # gzip the content argument and save it to filename argument
640 my ($cont, $filename) = @_;
641 #my $debug=1; # debug only
643 print "gzopening $filename ...\n" if (defined($debug));
644 my $gz = gzopen($filename, "w")
645 or die "Cannot gzopen for writing file $filename: $gzerrno";
646 print "gzwriting ...\n" if (defined($debug));
647 $gz->gzwrite($cont) or die "error writing: $gzerrno, exiting!\n";
648 print "gzclosing ...\n" if (defined($debug));
653 my ($patch_dtd,$url,$filename,$format) = @_;
656 # print "DOWNLOAD subs receives url: \"$url\"\n";
657 # print "DOWNLOAD subs receives filename: \"$filename\"\n";
659 my $resourcetype; # retrieve mode: "normal" (.xml) or "gzipped" (.xml.gz)
660 if ($filename =~ /\.xml$/) { # set retrieve mode
661 $resourcetype = "normal";
662 } elsif ($filename =~ /\.xml\.gz$/) {
663 $resourcetype = "gzipped";
665 die "Unsupported download extension, might be '.gz' or '.xml'\n";
667 my $basefname = $filename;
668 $basefname =~ s/\.gz$//; # get base resource name removing trailing .gz
669 $cont = ""; # modified by side-effect by the callback function
672 if (stat($basefname)) {
673 $localfname=$basefname;
674 } elsif (stat($basefname.".gz")) {
675 $localfname=$basefname.".gz";
677 if ($localfname ne "") { # we already have local copy of requested file
678 # check both possible cache entry: gzipped or normal
679 print "Using local copy.\n";
680 if ($localfname =~ /\.xml\.gz$/) { # deflating cached file and return it
681 $cont = gunzip($localfname);
682 } elsif ($localfname =~ /\.xml$/) { # just return cached file
683 open(FD, $localfname) or die "Cannot open $localfname";
684 while(<FD>) { $cont .= $_; }
687 die "Internal error: unexpected file name $localfname,"
688 ."must end with '.gz' or '.xml.gz'\n";
690 } else { # download file from net
691 print "Downloading file ...\n"; # download file
692 my $ua = LWP::UserAgent->new;
693 my $request = HTTP::Request->new(GET => "$url");
694 my $response = $ua->request($request, \&callback);
696 # cache retrieved file to disk
697 # <ZACK/> TODO: inefficent, I haven't yet undestood how to deflate
698 # in memory gzipped file, without call "gzopen"
699 print "Storing file ...\n";
700 # print "Making dirs ...\n";
702 # print "Opening tmp file for writing ...\n";
703 open(FD, ">".$filename.".tmp") or die "Cannot open $filename.tmp\n";
704 # print "Writing on tmp file ...\n";
706 # print "Closing tmp file ...\n";
709 # handle cache conversion normal->gzipped or gzipped->normal as user choice
710 print "cachemode:$cachemode, resourcetype:$resourcetype\n"
711 if (defined($debug));
712 if (($cachemode eq 'normal') and ($resourcetype eq 'normal')) {
713 # cache the file as is
714 rename "$filename.tmp", $filename;
715 } elsif (($cachemode eq 'gzipped') and ($resourcetype eq 'gzipped')) {
716 # cache the file as is
717 # and update the $cont variabile with deflated content
718 rename "$filename.tmp", $filename;
719 $cont = gunzip($filename);
720 } elsif (($cachemode eq 'normal') and ($resourcetype eq 'gzipped')) {
721 # deflate cache entry
723 open(FD, "> $basefname") or die "cannot open $basefname\n";
724 $cont = gunzip($filename.".tmp");
727 unlink "$filename.tmp"; # delete old gzipped file
728 } elsif (($cachemode eq 'gzipped') and ($resourcetype eq 'normal')) {
729 # compress cache entry
730 print "gzipping ...\n" if (defined($debug));
731 gzip($cont, $basefname.".gz");
732 unlink "$filename.tmp"; # delete old uncompressed file
734 die "Internal error, unsopported cachemode, resourcetype couple\n";
736 # $cont now contained uncompressed data
738 if ($patch_dtd eq "yes") {
739 $cont =~ s/DOCTYPE (.*) SYSTEM\s+"http:\/\/www.cs.unibo.it\/helm\/dtd\//DOCTYPE $1 SYSTEM "$myownurl\/getdtd?uri=/g;
741 if ($format eq "gz") {
742 gzip($cont,"$basefname.tmp");
743 open (TMP, "< $basefname.tmp")
744 or die "Can't open tempfile: $filename.tmp, exiting!\n";
750 unlink ($basefname . ".tmp") or
751 die "cannot unlink temporary file: $basefname.tmp\n";
758 # answer the client setting content, Content-Type and Content-Enconding
760 my ($c,$cont,$contype,$contenc) = @_;
761 my $res = new HTTP::Response;
762 $res->content($cont);
763 $res->push_header("Content-Type" => $contype)
764 unless ($contype eq "");
765 $res->push_header("Content-Encoding" => $contenc)
766 unless ($contenc eq "");
767 $res->push_header("Cache-Control" => "no-cache");
768 $res->push_header("Pragma" => "no-cache");
769 $res->push_header("Expires" => "0");
770 $c->send_response($res);
773 sub html_nice_answer {
774 # answer the client whith a nice html document
775 my ($c,$content) = @_;
776 $content = "<html><body><h1>$content</h1></body></html>";
777 answer($c,$content,"text/html","");
786 # retrieve a file from an url and write it to a temp dir
787 # used for retrieve resource index from servers
789 my ($prefix, $URL) = @_;
790 my $ua = LWP::UserAgent->new;
791 my $request = HTTP::Request->new(GET => "$URL");
792 my $response = $ua->request($request, \&callback);
793 my ($filename) = reverse (split "/", $URL); # get filename part of the URL
794 open (TEMP, "> $prefix/$filename")
795 || die "Cannot open temporary file: $prefix/$filename\n";
803 tie(%xml_map, 'DB_File', $uris_dbm, O_RDWR, 0664);
805 tie(%rdf_map, 'DB_File', $rdf_dbm, O_RDWR, 0664);
807 tie(%xslt_map, 'DB_File', $xslt_dbm, O_RDWR, 0664);
811 # rebuild dbs fetching resource indexes from servers.
812 # Rebuild urls_of_uris db (1), rdf db (2) and xslt db (3)
814 $server, $idxfile, $rdf_idxfile, $uri, $url, $comp, $line,
816 %urls_of_uris, %rdf, %xslt # local proxy of dbs
822 foreach my $file ($uris_dbm, $rdf_dbm, $xslt_dbm) { # remove old db file
823 if (stat $file) { # remove if exists
824 unlink ($file) or die "can't unlink old db file: $file\n";
827 tie(%urls_of_uris, 'DB_File', $uris_dbm, O_RDWR|O_CREAT, 0664);
828 tie(%rdf, 'DB_File', $rdf_dbm, O_RDWR|O_CREAT, 0664);
829 tie(%xslt, 'DB_File', $xslt_dbm, O_RDWR|O_CREAT, 0664);
831 open (SRVS, "< $servers_file") or
832 die "cannot open servers file: $servers_file\n";
833 @servers = <SRVS>; # read all servers
836 while ($server = pop @servers) { # cicle on servers in _reverse_ order
837 print "processing server: $server ...\n";
839 helm_wget($tmp_dir, $server."/".$indexname); # get index
840 helm_wget($tmp_dir, $server."/".$rdf_indexname); # get rdf index
841 helm_wget($tmp_dir, $server."/".$xslt_indexname); # get xslt index
842 $idxfile = $tmp_dir."/".$indexname;
843 $rdf_idxfile = $tmp_dir."/".$rdf_indexname;
844 $xslt_idxfile = $tmp_dir . "/" . $xslt_indexname;
846 # (1) REBUILD URLS_OF_URIS DB
847 open (INDEX, "< $idxfile") or
848 die "cannot open temporary index file: $idxfile\n";
849 while (my $line = <INDEX>) { # parse index and add entry to urls_of_uris db
851 ($uri,$comp) = split /[ \t]+/, $line;
854 $url = $uri . ".xml" . ".gz";
856 $url = $uri . ".xml";
858 $url =~ s/cic:/$server/;
859 $url =~ s/theory:/$server/;
860 $urls_of_uris{$uri} = $url;
864 open (RDF_INDEX, "< $rdf_idxfile") or
865 die "cannot open temporary RDF index file: $rdf_idxfile\n";
866 while (my $line = <RDF_INDEX>) { # parse index and add entry to rdf db
868 ($uri,$comp) = split /[ \t]+/, $line; # comp is "gz" or nothing
871 $url = $uri . ".xml" . ".gz";
873 $url = $uri . ".xml";
875 $url =~ s/^helm:rdf(.*)\/\/cic:/$server/;
876 $url =~ s/^helm:rdf(.*)\/\/theory:/$server/;
880 # (3) REBUILD XSLT DB
881 open (XSLT_INDEX, "< $xslt_idxfile") or
882 die "cannot open temporary XSLT index file: $xslt_idxfile\n";
883 while (my $line = <XSLT_INDEX>) { # parse index and add entry to xslt db
885 my $stylesheet = $line;
886 my $url = $server . "/" . $stylesheet;
887 $xslt{$stylesheet} = $url;
890 close INDEX; # close indexes
895 foreach my $file ($idxfile, $rdf_idxfile, $xslt_idxfile) {
896 print "cannot unlink temporary file: $file\n" if (unlink $file) != 1;
901 untie(%urls_of_uris); # untie local proxies
904 tie(%xml_map, 'DB_File', $uris_dbm, O_RDWR, 0664); # retie global ones
905 tie(%rdf_map, 'DB_File', $rdf_dbm, O_RDWR, 0664);
906 tie(%xslt_map, 'DB_File', $xslt_dbm, O_RDWR, 0664);
910 # vim modline: do not remove!