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@";
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";
57 my $helm_rdf_dir = $ENV{'HTTP_GETTER_RDF_DIR'} ||
58 "/usr/local/helm/rdf_library";
59 my $rdf_dbm = $ENV{'HTTP_GETTER_RDF_DBM'} ||
60 "/usr/local/helm/rdf_urls_of_uris";
61 my $rdf_indexname = $ENV{'HTTP_GETTER_RDF_INDEXNAME'} ||
63 # </move_to_conf_file>
65 # Let's override the configuration file
66 $style_dir = $ENV{"HELM_STYLE_DIR"} if (defined ($ENV{"HELM_STYLE_DIR"}));
67 $dtd_dir = $ENV{"HELM_DTD_DIR"} if (defined ($ENV{"HELM_DTD_DIR"}));
78 #CSC: mancano i controlli sulle condizioni di errore di molte funzioni
79 #CSC: ==> non e' robusto
80 #CSC: altra roba da sistemare segnata con CSC
82 my $d = new HTTP::Daemon LocalPort => 8081;
83 my $myownurl = $d->url;
85 # Let's patch the returned URL
86 $myownurl =~ s/\/$//; # chop the final slash
87 my $myownport = $myownurl;
88 $myownport =~ s/http:\/\/(.*):(.*)/$2/;
89 $myownurl =~ s/http:\/\/(.*):(.*)/$1/;
90 ($myownurl) = gethostbyname($myownurl);
91 $myownurl = "http://".$myownurl.":".$myownport;
93 tie(%map, 'DB_File', $uris_dbm.".db", O_RDWR, 0664); # open dbs
94 tie(%rdf_map, 'DB_File', $rdf_dbm.".db", O_RDWR, 0664);
97 print "HTTP Getter $VERSION\n"; # print hello information
98 print "Please contact me at: <URL:", $myownurl."/", ">\n";
100 print "helm_dir: $helm_dir\n";
101 print "helm_rdf_dir: $helm_rdf_dir\n";
102 print "style_dir: $style_dir\n";
103 print "dtd_dir: $dtd_dir\n";
104 print "urls_of_uris db: $uris_dbm.db\n";
105 print "rdf db: $rdf_dbm.db\n";
106 print "cache mode: $cachemode\n";
107 print "indexname: $indexname\n";
108 print "rdf_indexname: $rdf_indexname\n";
111 $SIG{CHLD} = "IGNORE"; # do not accumulate defunct processes
112 $SIG{USR1} = \&update; # sent by the child to make the parent update
114 while (my $c = $d->accept) {
116 while (my $r = $c->get_request) {
117 #CSC: mancano i controlli di sicurezza
119 my $http_method = $r->method;
120 my $http_path = $r->url->path;
121 my $http_query = uri_unescape($r->url->query);
122 my $cgi = new CGI("$http_query");
123 my $inputuri = $cgi->param('uri');
124 print "\nRequest: ".$r->url."\n\n";
126 print "\nUnescaped query: ".$http_query."\n";
128 # "getxml" works with rdf uris
129 if ($http_method eq 'GET' and $http_path eq "/getxml") {
130 # finds the uri, url and filename
132 my $answerformat = $cgi->param('format');
133 my $patch_dtd = $cgi->param('patch_dtd');
134 $answerformat = "" if (not defined($answerformat));
135 $patch_dtd = "yes" if (not defined($patch_dtd));
136 if (($answerformat ne "gz") and ($answerformat ne "normal")
137 and ($answerformat ne "")) {
138 die "Wrong output format: $answerformat, must be 'normal' ".
141 if (($patch_dtd ne "yes") and ($patch_dtd ne "no")
142 and ($patch_dtd ne "")) {
143 die "Wrong param, patch_dtd must be 'yes' or 'no'\n";
146 my $filename = $inputuri;
147 if (not isRdfUri($inputuri)) { # standad cic: or theory: uri
148 $filename =~ s/^cic:(.*)/$1/;
149 $filename =~ s/^theory:(.*)/$1/;
151 $filename =~ s/^(.*)\/\/cic:(.*)/$2/;
152 $filename =~ s/^(.*)\/\/theory:(.*)/$2/;
155 my $url = resolve ($inputuri); # resolve uri in url
156 if (not defined($url)) { # uri not found in uri2url map
157 die "NOT FOUND: uri \"$inputuri\" can't be resolved\n";
158 $c->send_error(RC_NOT_FOUND)
159 } else { # uri found and mapped to url
160 my $extension; # file extension
161 if ($url =~ /\.xml$/) { # non gzipped file
163 } elsif ($url =~ /\.xml\.gz$/) { # gzipped file
164 $extension = ".xml.gz";
165 } else { # error: unknown extension
166 die "unexpected extension in url: $url, might be '.xml'".
170 if (not isRdfUri ($inputuri)) { # save in uri std cache dir
171 $filename = $helm_dir.$filename.$extension;
172 } else { # save in rdf metadata cache dir
173 $filename = $helm_rdf_dir.$filename.$extension;
176 print_request($inputuri,$url,$filename);
179 my $ciccontent = download($patch_dtd,$url,$filename,$answerformat);
181 if ($answerformat eq "normal") { # answer the client in text/xml
182 answer($c,$ciccontent,"text/xml","");
183 } else { # answer the client in text/xml, gzip encoding
184 answer($c,$ciccontent,"text/xml","x-gzip");
188 # "/register" does not work with rdf uris
189 } elsif ($http_method eq 'GET' and $http_path eq "/register") {
190 my $inputurl = $cgi->param('url');
191 print "Register requested...\n";
192 $map{$inputuri}=$inputurl;
194 # Now let's clean the cache
195 my $cicfilename = $inputuri;
196 $cicfilename =~ s/cic:(.*)/$1/;
197 $cicfilename =~ s/theory:(.*)/$1/;
199 print "Unlinking ".$helm_dir.$cicfilename.".xml[.gz]\n";
200 unlink ($helm_dir.$cicfilename.".xml");
201 unlink ($helm_dir.$cicfilename.".xml.gz");
203 kill(USR1,getppid()); # signal changes to parent
206 html_nice_answer($c,"Register done");
208 # "/resolve" works with rdf uri
209 } elsif ($http_method eq 'GET' and $http_path eq "/resolve") {
210 #my $outputurl = $map{$inputuri};
211 my $outputurl = resolve($inputuri);
212 $outputurl = "" if (not defined($outputurl));
213 $cont = "<?xml version=\"1.0\" ?>\n\n";
214 if ($outputurl eq "") {
215 $cont .= "<unresolved />\n";
217 $cont .= "<url value=\"$outputurl\" />\n";
219 answer($c,$cont,"text/xml","");
221 } elsif ($http_method eq 'GET' and $http_path eq "/getdtd") {
222 my $filename = $inputuri;
223 $filename = $dtd_dir."/".$filename;
224 print "DTD: $inputuri ==> ($filename)\n";
225 if (stat($filename)) {
226 print "Using local copy\n";
227 open(FD, $filename) or die "Cannot open $filename\n";
230 # Vary bad heuristic for substituion of absolute URLs
232 s/ENTITY (.*) SYSTEM\s+"/ENTITY $1 SYSTEM "$myownurl\/getdtd?uri=/g;
236 answer($c,$cont,"text/xml","");
238 die "Could not find DTD!";
241 } elsif ($http_method eq 'GET' and $http_path eq "/getstyleconf") {
242 my $filename = $inputuri;
243 $filename = $style_dir."/config/".$filename;
244 if (stat($filename)) {
245 open(FD, $filename) or die "Cannot open $filename\n";
248 s/DOCTYPE (.*) SYSTEM\s+"/DOCTYPE $1 SYSTEM "$myownurl\/getstyleconf?uri=/g;
252 answer($c,$cont,"text/plain","");
254 die "Could not find Style Configuration File!";
257 } elsif ($http_method eq 'GET' and $http_path eq "/getxslt") {
258 my $filename = $inputuri;
259 $filename = $style_dir."/".$filename;
260 print "XSLT: $inputuri ==> ($filename)\n";
261 if (stat($filename)) {
262 print "Using local copy\n";
263 open(FD, $filename) or die "Cannot open $filename\n";
266 # Vary bad heuristic for substituion of absolute URLs
268 s/xsl:import\s+href="/xsl:import href="$myownurl\/getxslt?uri=/g ;
269 s/xsl:include\s+href="/xsl:include href="$myownurl\/getxslt?uri=/g ;
273 answer($c,$cont,"text/xml","");
275 die "Could not find XSLT!";
278 # "/update" works with rdf uri
279 } elsif ($http_method eq 'GET' and $http_path eq "/update") {
280 # rebuild urls_of_uris db and rdf uris db
281 print "Update requested...\n";
283 kill(USR1,getppid()); # signal changes to parent
285 html_nice_answer($c,"Update done");
287 # "/ls" does not work with rdf uris
288 } elsif ($http_method eq 'GET' and $http_path eq "/ls") {
289 # send back keys that begin with a given uri
290 my ($uritype,$uripattern,$uriprefix);
291 my $baseuri = $cgi->param('baseuri');
292 chop $baseuri if ($baseuri =~ /.*\/$/); # remove trailing "/"
293 # output type, might be 'txt' or 'xml':
294 my $outype = $cgi->param('format');
295 $uripattern = $baseuri;
296 $uripattern =~ s/^.*:(.*)/$1/;
297 if ($baseuri =~ /^cic:/) {
299 } elsif ($baseuri =~ /^theory:/) {
302 $uritype = "invalid";
304 if ($uritype ne "invalid") { # uri is valid
305 if (($outype ne 'txt') and ($outype ne 'xml')) { # invalid out type
306 print "Invalid output type specified: $outype\n";
307 html_nice_answer($c,"Invalid output type, must be ".
309 } else { # valid output
310 print "BASEURI $baseuri, FORMAT $outype\n";
311 $cont = finduris($uritype,$uripattern,$outype);
312 if ($outype eq 'txt') {
313 answer($c,$cont,"text/plain","");
314 } elsif ($outype eq 'xml') {
315 answer($c,$cont,"text/xml","");
317 die "Internal error, exiting!";
320 } else { # invalid uri
321 print "Invalid uri: $baseuri, may begin with 'cic:', ".
322 "'theory:' or '*:'\n";
323 html_nice_answer($c,"Invalid uri , must begin with ".
324 "'cic:' or 'theory:'");
327 } elsif ($http_method eq 'GET' and $http_path eq "/help") {
329 print "Help requested!\n";
330 html_nice_answer($c,"HTTP Getter Version: $VERSION");
332 } elsif ($http_method eq 'GET' and $http_path =~ /\.cgi$/) {
334 print "CGI requested!\n";
335 if ($http_path !~ /^\/[^\/]*\.cgi$/) {
336 html_nice_answer($c,"Invalid CGI name: $http_path, ".
337 "you can't request CGI that contain a slash in the name\n");
338 } elsif (stat "$cgi_dir"."$http_path") {
339 if (not -x "$cgi_dir/$http_path") {
340 html_nice_answer($c,"CGI $http_path found but not ".
342 } else { # exec CGI and anser back its output
345 $ENV{'QUERY_STRING'} = $http_query;
346 my $cgi_output = `$cgi_dir/$http_path`;
347 answer($c,$cgi_output,"","");
351 html_nice_answer($c,"CGI '$http_path' not found ".
352 "in CGI dir '$cgi_dir'");
355 } else { # unsupported request
357 print "INVALID REQUEST!!!!!\n";
358 print "(PATH: ",$http_path,", ";
359 print "QUERY: ",$http_query,")\n";
360 $c->send_error(RC_FORBIDDEN)
363 print "\nRequest solved: ".$r->url."\n\n";
368 print "\nCONNECTION CLOSED\n\n";
373 #================================
375 sub isRdfUri { # return true if the uri is an rdf uri, false otherwise
377 # helm:rdf/cic:www.cs.unibo.it/helm/rdf/foo_schema//cic:\
378 # /Coq/Init/Logic/True_rec.con.types
380 # the format is "helm:rdf/<metadata_tree>:<metadata_scheme>//<xml_file_uri>"
383 if ($uri =~ /^helm:rdf(.*):(.*)\/\/(.*)/) {
390 sub resolve { # resolve an uri in a url, work both with standard cic: or theory:
392 print "RESOLVE subroutine\n";
394 print "GIVEN URI: \"$uri\"\n";
395 if (isRdfUri ($uri)) { # rdf uri, resolve using rdf db
396 print "IS A RDF URI\n";
397 print "I WILL RETURN '$rdf_map{$uri}'\n";
398 return ($rdf_map{$uri});
399 } else { # standard cic: or theory: uri, resolve using std uri db
400 print "IS NOT A RDF URI\n";
401 print "I WILL RETURN '$map{$uri}'\n";
406 sub finduris { # find uris for cic and theory trees generation
407 my ($uritype,$uripattern,$format) = @_;
409 my ($uri,$localpart,$basepart,$dirname,$suffix,$flags,$key);
410 my (@itemz,@already_pushed_dir);
411 my (%objects,%dirs); # map uris to suffixes' flags
412 my $debug=1; # for debug
414 print "FINDURIS, uritype: $uritype, uripattern: $uripattern, ".
415 "format: $format\n\n" if defined($debug);
417 if (($uritype eq "cic") or ($uritype eq "theory")) {
418 # get info only of one type: cic or theory
419 foreach (keys(%map)) { # select matching uris
421 if ($uri =~ /^$uritype:$uripattern(\/|$|\.)/) {
422 if ($uri =~ /^$uritype:$uripattern\//) { # directory match
424 $localpart =~ s/^$uritype:$uripattern\/(.*)/$1/;
425 } elsif ($uri =~ /^$uritype:$uripattern($|\.)/) { # file match
427 $localpart =~ s/^.*\/([^\/]*)/$1/;
429 die "Internal error, seems that requested match is none of ".
430 "directory match or file match";
432 print "LOCALPART: $localpart\n" if defined($debug);
434 if ($localpart =~ /^[^\/]*$/) { # no slash, an OBJECT
435 $basepart = $localpart;
436 $basepart =~ s/^([^.]*\.[^.]*)(\.types)?(\.ann)?/$1/;
437 # remove exts .types or
439 $flags = $objects{$basepart}; # get old flags
440 if ($localpart =~ /\.ann$/) {
441 $flags = add_flag("ann","YES",$flags);
443 $flags = add_flag("ann","NO",$flags);
445 if ($localpart =~ /\.types$/) {
446 $flags = add_flag("types","YES",$flags);
447 } elsif ($localpart =~ /\.types\.ann$/) {
448 $flags = add_flag("types","ANN",$flags);
450 $flags = add_flag("types","NO",$flags);
452 $objects{$basepart} = $flags; # save new flags
453 } else { # exists at least one slash, a DIR
454 ($dirname) = split (/\//, $localpart);
455 $dirs{$dirname} = ""; # no flags requirement for dir
460 die "getter internal error: unsupported uritype: \"$uritype\"";
462 # now generate OUTPUT:
463 # output will be saved in $content
464 if ($format eq "txt") { # TXT output
465 foreach $key (sort(keys %dirs)) {
466 $content .= "dir, " . $key . "\n";
468 foreach $key (sort(keys %objects)) {
469 $content .= "object, $key, " . $objects{$key} . "\n";
471 } elsif ($format eq "xml") { # XML output
472 $content .= '<?xml version="1.0" encoding="ISO-8859-1"?>' . "\n";
473 $content .= "<!DOCTYPE ls SYSTEM ";
474 $content .= "\"$myownurl/getdtd?uri=ls.dtd\">" . "\n\n";
475 $content .= "<ls>\n";
476 foreach $key (sort(keys %dirs)) {
477 $content .= "\t<section>$key</section>\n";
479 foreach $key (sort(keys %objects)) {
480 $content .= "\t<object name=\"$key\">\n";
481 $flags = $objects{$key};
482 $flags =~ s/^<(.*)>$/$1/;
483 my ($annflag,$typesflag) = split /,/,$flags;
484 $content .= "\t\t<ann value=\"$annflag\" />\n";
485 $content .= "\t\t<types value=\"$typesflag\" />\n";
486 $content .= "\t</object>\n";
488 $content .= "</ls>\n";
489 } else { # may not enter this branch
490 die "Getter internal error: invalid format received by finduris sub";
496 # manage string like: "<ann_flag,type_flag>"
497 # "ann_flag" may be one of "ann_YES", "ann_NO"
498 # "type_flag" may be one of "types_NO", "types_YES", "types_ANN"
499 # when adding a flag the max between the current flag and the new flag
500 # is taken, the orders are ann_NO < ann_YES and types_NO < types_YES <
502 my ($flagtype,$newflag,$str) = @_;
503 $str = "<,>" if ($str eq "");
504 ($str =~ s/^<(.*,.*)>$/$1/) or die "Internal error: ".
505 "wrong string format for flag adding in $str";
506 my ($annflag,$typeflag) = split /,/,$str;
507 if ($flagtype eq "ann") { # ANN flag handling
508 if ($newflag eq "YES") {
510 } elsif ($newflag eq "NO") {
511 $annflag = "NO" unless ($annflag eq "YES");
513 die "Internal error: annflag must be \"YES\" or \"NO\"";
515 } elsif ($flagtype eq "types") { # TYPES flag handling
516 if ($newflag eq "ANN") {
518 } elsif ($newflag eq "YES") {
519 $typeflag = "YES" unless ($typeflag eq "ANN");
520 } elsif ($newflag eq "NO") {
522 unless (($typeflag eq "ANN") or ($typeflag eq "YES"));
524 die "Internal error: typeflag must be \"YES\", \"NO\" or \"ANN\"";
527 die "Internal error: unsupported flagtype \"$flagtype\"";
529 $str = "<$annflag,$typeflag>";
532 #CSC: Too much powerful: creates even /home, /home/users/, ...
533 #CSC: Does not raise errors if could not create dirs/files
537 my @dirs = split /\//,$pathname;
539 foreach $dir (@dirs) {
540 $tmp = ((defined($tmp)) ? $tmp."\/".$dir : "");
548 my ($uri,$url,$filename) = @_;
549 print "<download_request>\n";
552 print "filename: $filename\n\n";
553 print "</download_request>\n";
562 sub gunzip { # gunzip a file and return the deflated content
565 my ($gz, $buffer, $cont);
567 print "deflating $filename ...\n";
568 $gz = gzopen($filename, "r")
569 or die "Cannot open gzip'ed file $filename: $gzerrno";
571 while ( $gz->gzread($buffer) > 0 ) {
574 die "Error while reading : $gzerrno\n" if $gzerrno != Z_STREAM_END ;
580 sub gzip { # gzip the content argument and save it to filename argument
581 my ($cont, $filename) = @_;
582 #my $debug=1; # debug only
584 print "gzopening $filename ...\n" if (defined($debug));
585 my $gz = gzopen($filename, "w")
586 or die "Cannot gzopen for writing file $filename: $gzerrno";
587 print "gzwriting ...\n" if (defined($debug));
588 $gz->gzwrite($cont) or die "error writing: $gzerrno, exiting!\n";
589 print "gzclosing ...\n" if (defined($debug));
594 my ($patch_dtd,$url,$filename,$format) = @_;
597 # print "DOWNLOAD subs receives url: \"$url\"\n";
598 # print "DOWNLOAD subs receives filename: \"$filename\"\n";
600 my $resourcetype; # retrieve mode: "normal" (.xml) or "gzipped" (.xml.gz)
601 if ($filename =~ /\.xml$/) { # set retrieve mode
602 $resourcetype = "normal";
603 } elsif ($filename =~ /\.xml\.gz$/) {
604 $resourcetype = "gzipped";
606 die "Unsupported download extension, might be '.gz' or '.xml'\n";
608 my $basefname = $filename;
609 $basefname =~ s/\.gz$//; # get base resource name removing trailing .gz
610 $cont = ""; # modified by side-effect by the callback function
613 if (stat($basefname)) {
614 $localfname=$basefname;
615 } elsif (stat($basefname.".gz")) {
616 $localfname=$basefname.".gz";
618 if ($localfname ne "") { # we already have local copy of requested file
619 # check both possible cache entry: gzipped or normal
620 print "Using local copy.\n";
621 if ($localfname =~ /\.xml\.gz$/) { # deflating cached file and return it
622 $cont = gunzip($localfname);
623 } elsif ($localfname =~ /\.xml$/) { # just return cached file
624 open(FD, $localfname) or die "Cannot open $localfname";
625 while(<FD>) { $cont .= $_; }
628 die "Internal error: unexpected file name $localfname,"
629 ."must end with '.gz' or '.xml.gz'\n";
631 } else { # download file from net
632 print "Downloading file ...\n"; # download file
633 my $ua = LWP::UserAgent->new;
634 my $request = HTTP::Request->new(GET => "$url");
635 my $response = $ua->request($request, \&callback);
637 # cache retrieved file to disk
638 # <ZACK/> TODO: inefficent, I haven't yet undestood how to deflate
639 # in memory gzipped file, without call "gzopen"
640 print "Storing file ...\n";
641 # print "Making dirs ...\n";
643 # print "Opening tmp file for writing ...\n";
644 open(FD, ">".$filename.".tmp") or die "Cannot open $filename.tmp\n";
645 # print "Writing on tmp file ...\n";
647 # print "Closing tmp file ...\n";
650 # handle cache conversion normal->gzipped or gzipped->normal as user choice
651 print "cachemode:$cachemode, resourcetype:$resourcetype\n"
652 if (defined($debug));
653 if (($cachemode eq 'normal') and ($resourcetype eq 'normal')) {
654 # cache the file as is
655 rename "$filename.tmp", $filename;
656 } elsif (($cachemode eq 'gzipped') and ($resourcetype eq 'gzipped')) {
657 # cache the file as is
658 # and update the $cont variabile with deflated content
659 rename "$filename.tmp", $filename;
660 $cont = gunzip($filename);
661 } elsif (($cachemode eq 'normal') and ($resourcetype eq 'gzipped')) {
662 # deflate cache entry
664 open(FD, "> $basefname") or die "cannot open $basefname\n";
665 $cont = gunzip($filename.".tmp");
668 unlink "$filename.tmp"; # delete old gzipped file
669 } elsif (($cachemode eq 'gzipped') and ($resourcetype eq 'normal')) {
670 # compress cache entry
671 print "gzipping ...\n" if (defined($debug));
672 gzip($cont, $basefname.".gz");
673 unlink "$filename.tmp"; # delete old uncompressed file
675 die "Internal error, unsopported cachemode, resourcetype couple\n";
677 # $cont now contained uncompressed data
679 if ($patch_dtd eq "yes") {
680 $cont =~ s/DOCTYPE (.*) SYSTEM\s+"http:\/\/www.cs.unibo.it\/helm\/dtd\//DOCTYPE $1 SYSTEM "$myownurl\/getdtd?uri=/g;
682 if ($format eq "gz") {
683 gzip($cont,"$basefname.tmp");
684 open (TMP, "< $basefname.tmp")
685 or die "Can't open tempfile: $filename.tmp, exiting!\n";
691 unlink ($basefname . ".tmp") or
692 die "cannot unlink temporary file: $basefname.tmp\n";
699 # answer the client setting content, Content-Type and Content-Enconding
701 my ($c,$cont,$contype,$contenc) = @_;
702 my $res = new HTTP::Response;
703 $res->content($cont);
704 $res->push_header("Content-Type" => $contype)
705 unless ($contype eq "");
706 $res->push_header("Content-Encoding" => $contenc)
707 unless ($contenc eq "");
708 $res->push_header("Cache-Control" => "no-cache");
709 $res->push_header("Pragma" => "no-cache");
710 $res->push_header("Expires" => "0");
711 $c->send_response($res);
714 sub html_nice_answer {
715 # answer the client whith a nice html document
716 my ($c,$content) = @_;
717 $content = "<html><body><h1>$content</h1></body></html>";
718 answer($c,$content,"text/html","");
722 #retrieve a file from an url and write it to a temp dir
723 #used for retrieve resource index from servers
725 my ($prefix, $URL) = @_;
726 my $ua = LWP::UserAgent->new;
727 my $request = HTTP::Request->new(GET => "$URL");
728 my $response = $ua->request($request, \&callback);
729 my ($filename) = reverse (split "/", $URL); # get filename part of the URL
730 open (TEMP, "> $prefix/$filename")
731 || die "Cannot open temporary file: $prefix/$filename\n";
738 tie(%map, 'DB_File', $uris_dbm.".db", O_RDWR, 0664);
740 tie(%rdf_map, 'DB_File', $rdf_dbm.".db", O_RDWR, 0664);
744 # rebuild dbs fetching resource indexes from servers.
745 # Rebuild urls_of_uris db (1) and rdf db (2)
747 $server, $idxfile, $rdf_idxfile, $uri, $url, $comp, $line,
749 %urls_of_uris, %rdf # local proxy of $map and $rdf_map
754 foreach my $file ($uris_dbm, $rdf_dbm) { # remove old db file
755 if (stat $file.".db") { # remove if exists
756 unlink ($file.".db") or die "can't unlink old db file: $file.db\n";
759 tie(%urls_of_uris, 'DB_File', $uris_dbm.".db", O_RDWR|O_CREAT, 0664);
760 tie(%rdf, 'DB_File', $rdf_dbm.".db", O_RDWR|O_CREAT, 0664);
762 open (SRVS, "< $servers_file") or
763 die "cannot open servers file: $servers_file\n";
764 @servers = <SRVS>; # read all servers
767 while ($server = pop @servers) { # cicle on servers in _reverse_ order
768 print "processing server: $server ...\n";
770 helm_wget($tmp_dir, $server."/".$indexname); # get index
771 helm_wget($tmp_dir, $server."/".$rdf_indexname); # get rdf index
772 $idxfile = $tmp_dir."/".$indexname;
773 $rdf_idxfile = $tmp_dir."/".$rdf_indexname;
775 open (INDEX, "< $idxfile") or # (1) REBUILD URLS_OF_URIS DB
776 die "cannot open temporary index file: $idxfile\n";
777 while ($line = <INDEX>) { # parse index and add entry to urls_of_uris db
779 ($uri,$comp) = split /[ \t]+/, $line;
782 $url = $uri . ".xml" . ".gz";
784 $url = $uri . ".xml";
786 $url =~ s/cic:/$server/;
787 $url =~ s/theory:/$server/;
788 $urls_of_uris{$uri} = $url;
791 open (RDF_INDEX, "< $rdf_idxfile") or # (2) REBUILD RDF DB
792 die "cannot open temporary rdf index file: $rdf_idxfile\n";
793 while ($line = <RDF_INDEX>) { # parse index and add entry to rdf db
795 ($uri,$comp) = split /[ \t]+/, $line; # comp is "gz" or nothing
798 $url = $uri . ".xml" . ".gz";
800 $url = $uri . ".xml";
802 $url =~ s/^helm:rdf(.*)\/\/cic:/$server/;
803 $url =~ s/^helm:rdf(.*)\/\/theory:/$server/;
807 close INDEX; # close indexes
809 die "cannot unlink temporary file: $idxfile\n" # remove temp files
810 if (unlink $idxfile) != 1;
811 die "cannot unlink temporary file: $rdf_idxfile\n"
812 if (unlink $rdf_idxfile) != 1;
816 untie(%urls_of_uris); # untie local proxies
818 tie(%map, 'DB_File', $uris_dbm.".db", O_RDWR, 0664); # retie global ones
819 tie(%rdf_map, 'DB_File', $rdf_dbm.".db", O_RDWR, 0664);