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@";
34 # First of all, let's load HELM configuration
36 my $HELM_LIB_DIR = $ENV{"HELM_LIB_DIR"};
37 # this should be the only fixed constant
38 my $DEFAULT_HELM_LIB_DIR = "@HELM_LIB_DIR@";
39 if (defined ($HELM_LIB_DIR)) {
40 $HELM_LIB_PATH = $HELM_LIB_DIR."/configuration.pl";
42 $HELM_LIB_PATH = $DEFAULT_HELM_LIB_DIR."/configuration.pl";
45 # <ZACK>: TODO temporary, move this setting to configuration file
46 # set the cache mode, may be "gzipped" or "normal"
47 my $cachemode = $ENV{'HTTP_GETTER_CACHE_MODE'} || 'gzipped';
48 if (($cachemode ne 'gzipped') and ($cachemode ne 'normal')) {
49 die "Invalid HTTP_GETTER_CACHE_MODE environment variable, must be".
50 "'normal' or 'gzipped'\n";
54 # next require defines: $helm_dir, $html_link, $dtd_dir, $uris_dbm
55 require $HELM_LIB_PATH;
57 # Let's override the configuration file
58 $style_dir = $ENV{"HELM_STYLE_DIR"} if (defined ($ENV{"HELM_STYLE_DIR"}));
59 $dtd_dir = $ENV{"HELM_DTD_DIR"} if (defined ($ENV{"HELM_DTD_DIR"}));
70 #CSC: mancano i controlli sulle condizioni di errore di molte funzioni
71 #CSC: ==> non e' robusto
72 #CSC: altra roba da sistemare segnata con CSC
74 my $d = new HTTP::Daemon LocalPort => 8081;
75 my $myownurl = $d->url;
77 # Let's patch the returned URL
78 $myownurl =~ s/\/$//; # chop the final slash
79 my $myownport = $myownurl;
80 $myownport =~ s/http:\/\/(.*):(.*)/$2/;
81 $myownurl =~ s/http:\/\/(.*):(.*)/$1/;
82 ($myownurl) = gethostbyname($myownurl);
83 $myownurl = "http://".$myownurl.":".$myownport;
85 tie(%map, 'DB_File', $uris_dbm.".db", O_RDONLY, 0664);
86 print "Please contact me at: <URL:", $myownurl."/", ">\n";
87 print "helm_dir: $helm_dir\n";
88 print "style_dir: $style_dir\n";
89 print "dtd_dir: $dtd_dir\n";
90 print "urls_of_uris.db: $uris_dbm.db\n";
91 print "cache mode: $cachemode\n";
93 $SIG{CHLD} = "IGNORE"; # do not accumulate defunct processes
94 $SIG{USR1} = \&update; # sent by the child to make the parent update
95 while (my $c = $d->accept) {
97 while (my $r = $c->get_request) {
98 #CSC: mancano i controlli di sicurezza
100 my $inputuri = uri_unescape($r->url);
101 $inputuri =~ s/^[^?]*\?uri=(.*)/$1/;
102 print "\nRequest: ".$r->url."\n\n";
103 my $http_method = $r->method;
104 my $http_path = $r->url->path;
105 my $http_query = uri_unescape($r->url->query);
106 my $cgi = new CGI("$http_query");
108 print "\nUnescaped query: ".$http_query."\n";
110 if ($http_method eq 'GET' and $http_path eq "/getciconly") {
111 # finds the uri, url and filename
112 my $cicuri = $inputuri;
114 my $cicfilename = $cicuri;
115 $cicfilename =~ s/cic:(.*)/$1/;
116 $cicfilename =~ s/theory:(.*)/$1/;
118 my $cicurl = $map{$cicuri};
120 if ($cicurl =~ /\.xml$/) { # non gzipped file
122 } elsif ($cicurl =~ /\.xml\.gz$/) { # gzipped file
123 $extension = ".xml.gz";
124 } else { # error: unknown extension
125 die "unexpected extension in url: $cicurl, might be '.xml'".
128 $cicfilename = $helm_dir.$cicfilename.$extension;
130 if (!defined($cicurl)) {
131 print "\nNOT FOUND!!!!!\n";
132 $c->send_error(RC_NOT_FOUND)
134 print_request("cic",$cicuri,$cicurl,$cicfilename);
137 my $ciccontent = download(0,"cic",$cicurl,$cicfilename);
139 # Answering the client
140 answer($c,$ciccontent);
142 } elsif ($http_method eq 'GET' and $http_path eq "/get") {
143 # finds the uris, urls and filenames
144 my $cicuri = $inputuri,
145 $typesuri = $inputuri,
148 if ($inputuri =~ /\.types$/) {
149 $cicuri =~ s/(.*)\.types$/$1/;
151 } elsif ($inputuri =~ /\.types\.ann$/) {
152 $cicuri =~ s/(.*)\.types\.ann$/$1/;
153 $typesuri =~ s/(.*)\.ann$/$1/;
154 $annsuffix = ".types.ann";
155 } elsif ($inputuri =~ /\.ann$/) {
156 $cicuri =~ s/(.*)\.ann$/$1/;
164 my $cicfilename = $cicuri;
165 $cicfilename =~ s/cic:(.*)/$1/;
166 $cicfilename =~ s/theory:(.*)/$1/;
167 $cicfilename = $helm_dir.$cicfilename;
169 my $cicurl = $map{$cicuri};
170 my $typesurl = $map{$typesuri} if (defined($typesuri));
171 my $annurl = $map{$annuri} if (defined($annuri));
172 my ($cicext, $typesext, $annext);
173 if ($cicurl =~ /\.xml$/) { # normal file
175 } elsif ($cicurl =~ /\.xml\.gz$/) { # gzipped file
178 die "unexpected extension in url: $cicurl;".
179 "might be '.xml' or '.xml.gz'";
181 if (defined($typesuri)) { # extension selection for types file
182 if ($typesurl =~ /\.xml$/) { # normal file
183 $typesext = ".types.xml";
184 } elsif ($typesurl =~ /\.xml\.gz$/) { # gzipped file
185 $typesext = ".types.xml.gz";
187 die "unexpected extension in url: $typesurl;".
188 "might be '.xml' or '.xml.gz'";
191 if (defined($annuri)) { # extension selection for annotation file
192 if ($annurl =~ /\.xml$/) { # normal file
194 } elsif ($annurl =~ /\.xml\.gz$/) { # gzipped file
197 die "unexpected extension in url: $annurl".
198 "might be '.xml' or '.xml.gz'";
201 my $typesfilename = $cicfilename.$typesext if $typesuri;
202 my $annfilename = $cicfilename.$annsuffix.$annext if $annuri;
203 $cicfilename .= $cicext;
205 if (!defined($cicurl) ||
206 (!defined($typesurl) && $typesuri) ||
207 (!defined($annuri) && $annuri))
209 print "\nNOT FOUND!!!!!\n";
210 $c->send_error(RC_NOT_FOUND)
212 print_request("cic",$cicuri,$cicurl,$cicfilename);
213 print_request("types",$typesuri,$typesurl,$typesfilename)
215 print_request("ann",$annuri,$annurl,$annfilename)
218 # Retrieves the files
220 my $ciccontent = download(1,"cic",$cicurl,$cicfilename);
222 download(1,"types",$typesurl,$typesfilename) if ($typesuri);
224 download(1,"ann",$annurl,$annfilename) if ($annuri);
226 # Merging the files together
229 <?xml version="1.0" encoding="UTF-8"?>
230 <cicxml uri="$cicuri">
237 # Answering the client
240 } elsif ($http_method eq 'GET' and $http_path eq "/getdtd") {
241 my $filename = $inputuri;
242 $filename = $dtd_dir."/".$filename;
243 print "DTD: $inputuri ==> ($filename)\n";
244 if (stat($filename)) {
245 print "Using local copy\n";
246 open(FD, $filename) or die "Cannot open $filename\n";
249 # Vary bad heuristic for substituion of absolute URLs
251 s/ENTITY (.*) SYSTEM\s+"/ENTITY $1 SYSTEM "$myownurl\/getdtd?uri=/g;
257 die "Could not find DTD!";
259 } elsif ($http_method eq 'GET' and $http_path eq "/getstyleconf") {
260 my $filename = $inputuri;
261 $filename = $style_dir."/config/".$filename;
262 if (stat($filename)) {
263 open(FD, $filename) or die "Cannot open $filename\n";
266 s/DOCTYPE (.*) SYSTEM\s+"/DOCTYPE $1 SYSTEM "$myownurl\/getstyleconf?uri=/g;
272 die "Could not find Style Configuration File!";
274 } elsif ($http_method eq 'GET' and $http_path eq "/getxslt") {
275 my $filename = $inputuri;
276 $filename = $style_dir."/".$filename;
277 print "XSLT: $inputuri ==> ($filename)\n";
278 if (stat($filename)) {
279 print "Using local copy\n";
280 open(FD, $filename) or die "Cannot open $filename\n";
283 # Vary bad heuristic for substituion of absolute URLs
285 s/xsl:import\s+href="/xsl:import href="$myownurl\/getxslt?uri=/g ;
286 s/xsl:include\s+href="/xsl:include href="$myownurl\/getxslt?uri=/g ;
292 die "Could not find XSLT!";
294 } elsif ($http_method eq 'GET' and $http_path eq "/conf") {
295 my $quoted_html_link = $html_link;
296 $quoted_html_link =~ s/&/&/g;
297 $quoted_html_link =~ s/</</g;
298 $quoted_html_link =~ s/>/>/g;
299 $quoted_html_link =~ s/'/'/g;
300 $quoted_html_link =~ s/"/"/g;
301 print "\nConfiguration requested, returned #$quoted_html_link#\n";
302 $cont = "<?xml version=\"1.0\"?><html_link>$quoted_html_link</html_link>";
304 } elsif ($http_method eq 'GET' and $http_path eq "/update") {
305 # rebuild urls_of_uris.db
306 print "Update requested...\n";
308 kill(USR1,getppid()); # signal changes to parent
310 answer($c,"<html><body><h1>Update done</h1></body></html>");
311 } elsif ($http_method eq 'GET' and $http_path eq "/ls") {
312 # send back keys that begin with a given uri
313 my ($uritype,$uripattern,$uriprefix);
314 my $baseuri = $cgi->param('baseuri');
315 chop $baseuri if ($baseuri =~ /.*\/$/); # remove trailing "/"
316 my $outype = $cgi->param('format'); # output type, might be 'txt' or 'xml'
317 $uripattern = $baseuri;
318 $uripattern =~ s/^.*:(.*)/$1/;
319 if ($baseuri =~ /^cic:/) {
321 } elsif ($baseuri =~ /^theory:/) {
324 $uritype = "invalid";
326 if ($uritype ne "invalid") { # uri is valid
327 if (($outype ne 'txt') and ($outype ne 'xml')) { # invalid out type
328 print "Invalid output type specified: $outype\n";
329 answer($c,"<html><body><h1>Invalid output type, may be ".
330 "\"txt\" or \"xml\"</h1></body></html>");
331 } else { # valid output
332 print "BASEURI $baseuri, FORMAT $outype\n";
333 $cont = finduris($uritype,$uripattern,$outype);
336 } else { # invalid uri
337 print "Invalid uri: $baseuri, may begin with 'cic:', ".
338 "'theory:' or '*:'\n";
339 answer($c,"<html><body><h1>Invalid uri , may begin with ".
340 "\"cic:\", \"theory:\" or \"*:\"</h1></body></html>");
342 } elsif ($http_method eq 'GET' and $http_path eq "/help") {
343 print "Help requested!";
344 answer($c,"<html><body><h1>HTTP Getter Version ".
345 $VERSION."</h1></body></html>");
348 print "INVALID REQUEST!!!!!\n";
349 print "(PATH: ",$http_path,", ";
350 print "QUERY: ",$http_query,")\n";
351 $c->send_error(RC_FORBIDDEN)
353 print "\nRequest solved: ".$r->url."\n\n";
357 print "\nCONNECTION CLOSED\n\n";
362 #================================
364 sub finduris { # find uris for cic and theory trees generation
365 my ($uritype,$uripattern,$format) = @_;
367 my ($uri,$localpart,$basepart,$dirname,$suffix,$flags,$key);
368 my (@itemz,@already_pushed_dir);
369 my (%objects,%dirs); # map uris to suffixes' flags
371 print "FINDURIS, uritype: $uritype, uripattern: $uripattern, ".
372 "format: $format\n\n";
374 if (($uritype eq "cic") or ($uritype eq "theory")) {
375 # get info only of one type: cic or theory
376 foreach (keys(%map)) { # select matching uris
378 if ($uri =~ /^$uritype:$uripattern\//) {
380 $localpart =~ s/^$uritype:$uripattern\/(.*)/$1/;
382 if ($localpart =~ /^[^\/]*$/) { # no slash, an OBJECT
383 $basepart = $localpart;
384 $basepart =~ s/^([^.]*\.[^.]*)(\.types)?(\.ann)?/$1/; # remove exts .types or
386 $flags = $objects{$basepart}; # get old flags
387 if ($localpart =~ /\.ann$/) {
388 $flags = add_flag("ann","YES",$flags);
390 $flags = add_flag("ann","NO",$flags);
392 if ($localpart =~ /\.types$/) {
393 $flags = add_flag("types","YES",$flags);
394 } elsif ($localpart =~ /\.types\.ann$/) {
395 $flags = add_flag("types","ANN",$flags);
397 $flags = add_flag("types","NO",$flags);
399 $objects{$basepart} = $flags; # save new flags
400 } else { # exists at least one slash, a DIR
401 ($dirname) = split (/\//, $localpart);
402 $dirs{$dirname} = ""; # no flags requirement for dir
407 die "getter internal error: unsupported uritype: \"$uritype\"";
409 # now generate OUTPUT:
410 # output will be saved in $content
411 if ($format eq "txt") { # TXT output
412 foreach $key (sort(keys %dirs)) {
413 $content .= "dir, " . $key . "\n";
415 foreach $key (sort(keys %objects)) {
416 $content .= "object, $key, " . $objects{$key} . "\n";
418 } elsif ($format eq "xml") { # XML output
419 $content .= '<?xml version="1.0" encoding="ISO-8859-1"?>' . "\n";
420 $content .= '<!DOCTYPE ls SYSTEM ';
421 $content .= '"http://localhost:8081/getdtd?uri=ls.dtd">' . "\n\n";
422 $content .= "<ls>\n";
423 foreach $key (sort(keys %dirs)) {
424 $content .= "\t<section>$key</section>\n";
426 foreach $key (sort(keys %objects)) {
427 $content .= "\t<object name=\"$key\">\n";
428 $flags = $objects{$key};
429 $flags =~ s/^<(.*)>$/$1/;
430 my ($annflag,$typesflag) = split /,/,$flags;
431 $content .= "\t\t<ann value=\"$annflag\" />\n";
432 $content .= "\t\t<types value=\"$typesflag\" />\n";
433 $content .= "\t</object>\n";
435 $content .= "</ls>\n";
436 } else { # may not enter this branch
437 die "Getter internal error: invalid format received by finduris sub";
443 # manage string like: "<ann_flag,type_flag>"
444 # "ann_flag" may be one of "ann_YES", "ann_NO"
445 # "type_flag" may be one of "types_NO", "types_YES", "types_ANN"
446 # when adding a flag the max between the current flag and the new flag
447 # is taken, the orders are ann_NO < ann_YES and types_NO < types_YES <
449 my ($flagtype,$newflag,$str) = @_;
450 $str = "<,>" if ($str eq "");
451 ($str =~ s/^<(.*,.*)>$/$1/) or die "Internal error: ".
452 "wrong string format for flag adding in $str";
453 my ($annflag,$typeflag) = split /,/,$str;
454 if ($flagtype eq "ann") { # ANN flag handling
455 if ($newflag eq "YES") {
457 } elsif ($newflag eq "NO") {
458 $annflag = "NO" unless ($annflag eq "YES");
460 die "Internal error: annflag must be \"YES\" or \"NO\"";
462 } elsif ($flagtype eq "types") { # TYPES flag handling
463 if ($newflag eq "ANN") {
465 } elsif ($newflag eq "YES") {
466 $typeflag = "YES" unless ($typeflag eq "ANN");
467 } elsif ($newflag eq "NO") {
469 unless (($typeflag eq "ANN") or ($typeflag eq "YES"));
471 die "Internal error: typeflag must be \"YES\", \"NO\" or \"ANN\"";
474 die "Internal error: unsupported flagtype \"$flagtype\"";
476 $str = "<$annflag,$typeflag>";
479 #CSC: Too much powerful: creates even /home, /home/users/, ...
480 #CSC: Does not raise errors if could not create dirs/files
484 my @dirs = split /\//,$pathname;
486 foreach $dir (@dirs) {
487 $tmp = ((defined($tmp)) ? $tmp."\/".$dir : "");
495 my ($str,$uri,$url,$filename) = @_;
496 print $str."uri: $uri\n";
497 print $str."url: $url\n";
498 print $str."filename: $filename\n\n";
507 sub gunzip { # gunzip a file and return the deflated content
510 my ($gz, $buffer, $cont);
512 print "deflating $filename ...\n";
513 $gz = gzopen($filename, "r")
514 or die "Cannot open gzip'ed file $filename: $gzerrno";
516 while ( $gz->gzread($buffer) > 0 ) {
519 die "Error while reading : $gzerrno\n" if $gzerrno != Z_STREAM_END ;
525 sub gzip { # gzip the content argument and save it to filename argument
526 my ($cont, $filename) = @_;
527 #my $debug=1; # debug only
529 print "gzopening $filename ...\n" if (defined($debug));
530 my $gz = gzopen($filename, "w")
531 or die "Cannot gzopen for writing file $filename: $gzerrno";
532 print "gzwriting ...\n" if (defined($debug));
533 $gz->gzwrite($cont) or die "error writing: $gzerrno, exiting!\n";
534 print "gzclosing ...\n" if (defined($debug));
540 my ($remove_headers,$str,$url,$filename) = @_;
543 #my $debug=1; # for DEBUG only
545 my $resourcetype; # retrieve mode: "normal" (.xml) or "gzipped" (.xml.gz)
546 if ($filename =~ /\.xml$/) { # set retrieve mode
547 $resourcetype = "normal";
548 } elsif ($filename =~ /\.xml\.gz$/) {
549 $resourcetype = "gzipped";
551 die "Unsupported download extension, might be '.gz' or '.xml'\n";
553 my $basefname = $filename;
554 $basefname =~ s/\.gz$//; # get base resource name removing trailing .gz
555 $cont = ""; # modified by side-effect by the callback function
558 if (stat($basefname)) {
559 $localfname=$basefname;
560 } elsif (stat($basefname.".gz")) {
561 $localfname=$basefname.".gz";
563 if ($localfname ne "") { # we already have local copy of requested file
564 # check both possible cache entry: gzipped or normal
565 print "Using local copy for the $str file\n";
566 if ($localfname =~ /\.xml\.gz$/) { # deflating cached file and return it
567 $cont = gunzip($localfname);
568 } elsif ($localfname =~ /\.xml$/) { # just return cached file
569 open(FD, $localfname) or die "Cannot open $localfname";
570 while(<FD>) { $cont .= $_; }
573 die "Internal error: unexpected file name $localfname,"
574 ."must end with '.gz' or '.xml.gz'\n";
576 } else { # download file from net
577 print "Downloading the $str file\n"; # download file
578 my $ua = LWP::UserAgent->new;
579 my $request = HTTP::Request->new(GET => "$url");
580 my $response = $ua->request($request, \&callback);
582 # cache retrieved file to disk
583 # <ZACK/> TODO: inefficent, I haven't yet undestood how to deflate
584 # in memory gzipped file, without call "gzopen"
585 print "Storing the $str file\n";
586 print "Making dirs ...\n" if (defined($debug));
588 print "Opening tmp file for writing ...\n" if (defined($debug));
589 open(FD, ">".$filename.".tmp") or die "Cannot open $filename.tmp\n";
590 print "Writing on tmp file ...\n" if (defined($debug));
592 print "Closing tmp file ...\n" if (defined($debug));
595 # handle cache conversion normal->gzipped or gzipped->normal as user choice
596 print "cachemode:$cachemode, resourcetype:$resourcetype\n" if (defined($debug));
597 if (($cachemode eq 'normal') and ($resourcetype eq 'normal')) {
598 # cache the file as is
599 rename "$filename.tmp", $filename;
600 } elsif (($cachemode eq 'gzipped') and ($resourcetype eq 'gzipped')) {
601 # cache the file as is
602 # and update the $cont variabile with deflated content
603 rename "$filename.tmp", $filename;
604 $cont = gunzip($filename);
605 } elsif (($cachemode eq 'normal') and ($resourcetype eq 'gzipped')) {
606 # deflate cache entry
608 open(FD, "> $basefname") or die "cannot open $basefname\n";
609 $cont = gunzip($filename.".tmp");
612 unlink "$filename.tmp"; # delete old gzipped file
613 } elsif (($cachemode eq 'gzipped') and ($resourcetype eq 'normal')) {
614 # compress cache entry
615 print "gzipping ...\n" if (defined($debug));
616 gzip($cont, $basefname.".gz");
617 unlink "$filename.tmp"; # delete old uncompressed file
619 die "Internal error, unsopported cachemode, resourcetype couple\n";
621 # $cont now contained uncompressed data
624 if ($remove_headers) {
625 $cont =~ s/<\?xml [^?]*\?>//sg;
626 $cont =~ s/<!DOCTYPE [^>]*>//sg;
628 $cont =~ s/DOCTYPE (.*) SYSTEM\s+"http:\/\/www.cs.unibo.it\/helm\/dtd\//DOCTYPE $1 SYSTEM "$myownurl\/getdtd?uri=/g;
636 my $res = new HTTP::Response;
637 $res->content($cont);
638 $c->send_response($res);
642 #retrieve a file from an url and write it to a temp dir
643 #used for retrieve resource index from servers
645 my ($prefix, $URL) = @_;
646 my $ua = LWP::UserAgent->new;
647 my $request = HTTP::Request->new(GET => "$URL");
648 my $response = $ua->request($request, \&callback);
649 my ($filename) = reverse (split "/", $URL); # get filename part of the URL
650 open (TEMP, "> $prefix/$filename")
651 || die "Cannot open temporary file: $prefix/$filename\n";
658 tie(%map, 'DB_File', $uris_dbm.".db", O_RDONLY, 0664);
661 sub mk_urls_of_uris {
662 #rebuild $uris_dbm.db fetching resource indexes from servers
664 $server, $idxfile, $uri, $url, $comp, $line,
670 if (stat $uris_dbm.".db") { # remove old db file
671 unlink($uris_dbm.".db") or
672 die "cannot unlink old db file: $uris_dbm.db\n";
674 tie(%urls_of_uris, 'DB_File', $uris_dbm.".db", O_RDWR|O_CREAT, 0664);
676 open (SRVS, "< $servers_file") or
677 die "cannot open servers file: $servers_file\n";
680 while ($server = pop @servers) { #cicle on servers in reverse order
681 print "processing server: $server ...\n";
683 helm_wget($tmp_dir, $server."/".$indexname); #get index
684 $idxfile = $tmp_dir."/".$indexname;
685 open (INDEX, "< $idxfile") or
686 die "cannot open temporary index file: $idxfile\n";
687 while ($line = <INDEX>) { #parse index and add entry to urls_of_uris
689 ($uri,$comp) = split /[ \t]+/, $line;
692 $url = $uri . ".xml" . ".gz";
694 $url = $uri . ".xml";
696 $url =~ s/cic:/$server/;
697 $url =~ s/theory:/$server/;
698 $urls_of_uris{$uri} = $url;
701 die "cannot unlink temporary file: $idxfile\n"
702 if (unlink $idxfile) != 1;
705 untie(%urls_of_uris);
706 tie(%map, 'DB_File', $uris_dbm.".db", O_RDONLY, 0664);