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 $http_method = $r->method;
101 my $http_path = $r->url->path;
102 my $http_query = uri_unescape($r->url->query);
103 my $cgi = new CGI("$http_query");
104 my $inputuri = $cgi->param('uri');
105 print "\nRequest: ".$r->url."\n\n";
107 print "\nUnescaped query: ".$http_query."\n";
109 if ($http_method eq 'GET' and $http_path eq "/getciconly") {
110 # finds the uri, url and filename
111 my $cicuri = $inputuri;
112 my $answerformat = $cgi->param('format');
113 $answerformat = "" if (not defined($answerformat));
115 if (($answerformat ne "gz") and ($answerformat ne "normal")
116 and ($answerformat ne "")) {
117 die "Wrong output format: $answerformat, must be 'normal' ".
121 my $cicfilename = $cicuri;
122 $cicfilename =~ s/cic:(.*)/$1/;
123 $cicfilename =~ s/theory:(.*)/$1/;
125 my $cicurl = $map{$cicuri};
127 if ($cicurl =~ /\.xml$/) { # non gzipped file
129 } elsif ($cicurl =~ /\.xml\.gz$/) { # gzipped file
130 $extension = ".xml.gz";
131 } else { # error: unknown extension
132 die "unexpected extension in url: $cicurl, might be '.xml'".
135 $cicfilename = $helm_dir.$cicfilename.$extension;
137 if (!defined($cicurl)) {
138 print "\nNOT FOUND!!!!!\n";
139 $c->send_error(RC_NOT_FOUND)
141 print_request("cic",$cicuri,$cicurl,$cicfilename);
144 my $ciccontent = download(0,"cic",$cicurl,$cicfilename,$answerformat);
146 # Answering the client
147 answer($c,$ciccontent);
149 } elsif ($http_method eq 'GET' and $http_path eq "/getdtd") {
150 my $filename = $inputuri;
151 $filename = $dtd_dir."/".$filename;
152 print "DTD: $inputuri ==> ($filename)\n";
153 if (stat($filename)) {
154 print "Using local copy\n";
155 open(FD, $filename) or die "Cannot open $filename\n";
158 # Vary bad heuristic for substituion of absolute URLs
160 s/ENTITY (.*) SYSTEM\s+"/ENTITY $1 SYSTEM "$myownurl\/getdtd?uri=/g;
166 die "Could not find DTD!";
168 } elsif ($http_method eq 'GET' and $http_path eq "/getstyleconf") {
169 my $filename = $inputuri;
170 $filename = $style_dir."/config/".$filename;
171 if (stat($filename)) {
172 open(FD, $filename) or die "Cannot open $filename\n";
175 s/DOCTYPE (.*) SYSTEM\s+"/DOCTYPE $1 SYSTEM "$myownurl\/getstyleconf?uri=/g;
181 die "Could not find Style Configuration File!";
183 } elsif ($http_method eq 'GET' and $http_path eq "/getxslt") {
184 my $filename = $inputuri;
185 $filename = $style_dir."/".$filename;
186 print "XSLT: $inputuri ==> ($filename)\n";
187 if (stat($filename)) {
188 print "Using local copy\n";
189 open(FD, $filename) or die "Cannot open $filename\n";
192 # Vary bad heuristic for substituion of absolute URLs
194 s/xsl:import\s+href="/xsl:import href="$myownurl\/getxslt?uri=/g ;
195 s/xsl:include\s+href="/xsl:include href="$myownurl\/getxslt?uri=/g ;
201 die "Could not find XSLT!";
203 } elsif ($http_method eq 'GET' and $http_path eq "/conf") {
204 my $quoted_html_link = $html_link;
205 $quoted_html_link =~ s/&/&/g;
206 $quoted_html_link =~ s/</</g;
207 $quoted_html_link =~ s/>/>/g;
208 $quoted_html_link =~ s/'/'/g;
209 $quoted_html_link =~ s/"/"/g;
210 print "\nConfiguration requested, returned #$quoted_html_link#\n";
211 $cont = "<?xml version=\"1.0\"?><html_link>$quoted_html_link</html_link>";
213 } elsif ($http_method eq 'GET' and $http_path eq "/update") {
214 # rebuild urls_of_uris.db
215 print "Update requested...\n";
217 kill(USR1,getppid()); # signal changes to parent
219 answer($c,"<html><body><h1>Update done</h1></body></html>");
220 } elsif ($http_method eq 'GET' and $http_path eq "/ls") {
221 # send back keys that begin with a given uri
222 my ($uritype,$uripattern,$uriprefix);
223 my $baseuri = $cgi->param('baseuri');
224 chop $baseuri if ($baseuri =~ /.*\/$/); # remove trailing "/"
225 my $outype = $cgi->param('format'); # output type, might be 'txt' or 'xml'
226 $uripattern = $baseuri;
227 $uripattern =~ s/^.*:(.*)/$1/;
228 if ($baseuri =~ /^cic:/) {
230 } elsif ($baseuri =~ /^theory:/) {
233 $uritype = "invalid";
235 if ($uritype ne "invalid") { # uri is valid
236 if (($outype ne 'txt') and ($outype ne 'xml')) { # invalid out type
237 print "Invalid output type specified: $outype\n";
238 answer($c,"<html><body><h1>Invalid output type, may be ".
239 "\"txt\" or \"xml\"</h1></body></html>");
240 } else { # valid output
241 print "BASEURI $baseuri, FORMAT $outype\n";
242 $cont = finduris($uritype,$uripattern,$outype);
245 } else { # invalid uri
246 print "Invalid uri: $baseuri, may begin with 'cic:', ".
247 "'theory:' or '*:'\n";
248 answer($c,"<html><body><h1>Invalid uri , may begin with ".
249 "\"cic:\", \"theory:\" or \"*:\"</h1></body></html>");
251 } elsif ($http_method eq 'GET' and $http_path eq "/help") {
252 print "Help requested!";
253 answer($c,"<html><body><h1>HTTP Getter Version ".
254 $VERSION."</h1></body></html>");
257 print "INVALID REQUEST!!!!!\n";
258 print "(PATH: ",$http_path,", ";
259 print "QUERY: ",$http_query,")\n";
260 $c->send_error(RC_FORBIDDEN)
262 print "\nRequest solved: ".$r->url."\n\n";
266 print "\nCONNECTION CLOSED\n\n";
271 #================================
273 sub finduris { # find uris for cic and theory trees generation
274 my ($uritype,$uripattern,$format) = @_;
276 my ($uri,$localpart,$basepart,$dirname,$suffix,$flags,$key);
277 my (@itemz,@already_pushed_dir);
278 my (%objects,%dirs); # map uris to suffixes' flags
280 print "FINDURIS, uritype: $uritype, uripattern: $uripattern, ".
281 "format: $format\n\n";
283 if (($uritype eq "cic") or ($uritype eq "theory")) {
284 # get info only of one type: cic or theory
285 foreach (keys(%map)) { # select matching uris
287 if ($uri =~ /^$uritype:$uripattern\//) {
289 $localpart =~ s/^$uritype:$uripattern\/(.*)/$1/;
291 if ($localpart =~ /^[^\/]*$/) { # no slash, an OBJECT
292 $basepart = $localpart;
293 $basepart =~ s/^([^.]*\.[^.]*)(\.types)?(\.ann)?/$1/; # remove exts .types or
295 $flags = $objects{$basepart}; # get old flags
296 if ($localpart =~ /\.ann$/) {
297 $flags = add_flag("ann","YES",$flags);
299 $flags = add_flag("ann","NO",$flags);
301 if ($localpart =~ /\.types$/) {
302 $flags = add_flag("types","YES",$flags);
303 } elsif ($localpart =~ /\.types\.ann$/) {
304 $flags = add_flag("types","ANN",$flags);
306 $flags = add_flag("types","NO",$flags);
308 $objects{$basepart} = $flags; # save new flags
309 } else { # exists at least one slash, a DIR
310 ($dirname) = split (/\//, $localpart);
311 $dirs{$dirname} = ""; # no flags requirement for dir
316 die "getter internal error: unsupported uritype: \"$uritype\"";
318 # now generate OUTPUT:
319 # output will be saved in $content
320 if ($format eq "txt") { # TXT output
321 foreach $key (sort(keys %dirs)) {
322 $content .= "dir, " . $key . "\n";
324 foreach $key (sort(keys %objects)) {
325 $content .= "object, $key, " . $objects{$key} . "\n";
327 } elsif ($format eq "xml") { # XML output
328 $content .= '<?xml version="1.0" encoding="ISO-8859-1"?>' . "\n";
329 $content .= '<!DOCTYPE ls SYSTEM ';
330 $content .= '"http://localhost:8081/getdtd?uri=ls.dtd">' . "\n\n";
331 $content .= "<ls>\n";
332 foreach $key (sort(keys %dirs)) {
333 $content .= "\t<section>$key</section>\n";
335 foreach $key (sort(keys %objects)) {
336 $content .= "\t<object name=\"$key\">\n";
337 $flags = $objects{$key};
338 $flags =~ s/^<(.*)>$/$1/;
339 my ($annflag,$typesflag) = split /,/,$flags;
340 $content .= "\t\t<ann value=\"$annflag\" />\n";
341 $content .= "\t\t<types value=\"$typesflag\" />\n";
342 $content .= "\t</object>\n";
344 $content .= "</ls>\n";
345 } else { # may not enter this branch
346 die "Getter internal error: invalid format received by finduris sub";
352 # manage string like: "<ann_flag,type_flag>"
353 # "ann_flag" may be one of "ann_YES", "ann_NO"
354 # "type_flag" may be one of "types_NO", "types_YES", "types_ANN"
355 # when adding a flag the max between the current flag and the new flag
356 # is taken, the orders are ann_NO < ann_YES and types_NO < types_YES <
358 my ($flagtype,$newflag,$str) = @_;
359 $str = "<,>" if ($str eq "");
360 ($str =~ s/^<(.*,.*)>$/$1/) or die "Internal error: ".
361 "wrong string format for flag adding in $str";
362 my ($annflag,$typeflag) = split /,/,$str;
363 if ($flagtype eq "ann") { # ANN flag handling
364 if ($newflag eq "YES") {
366 } elsif ($newflag eq "NO") {
367 $annflag = "NO" unless ($annflag eq "YES");
369 die "Internal error: annflag must be \"YES\" or \"NO\"";
371 } elsif ($flagtype eq "types") { # TYPES flag handling
372 if ($newflag eq "ANN") {
374 } elsif ($newflag eq "YES") {
375 $typeflag = "YES" unless ($typeflag eq "ANN");
376 } elsif ($newflag eq "NO") {
378 unless (($typeflag eq "ANN") or ($typeflag eq "YES"));
380 die "Internal error: typeflag must be \"YES\", \"NO\" or \"ANN\"";
383 die "Internal error: unsupported flagtype \"$flagtype\"";
385 $str = "<$annflag,$typeflag>";
388 #CSC: Too much powerful: creates even /home, /home/users/, ...
389 #CSC: Does not raise errors if could not create dirs/files
393 my @dirs = split /\//,$pathname;
395 foreach $dir (@dirs) {
396 $tmp = ((defined($tmp)) ? $tmp."\/".$dir : "");
404 my ($str,$uri,$url,$filename) = @_;
405 print $str."uri: $uri\n";
406 print $str."url: $url\n";
407 print $str."filename: $filename\n\n";
416 sub gunzip { # gunzip a file and return the deflated content
419 my ($gz, $buffer, $cont);
421 print "deflating $filename ...\n";
422 $gz = gzopen($filename, "r")
423 or die "Cannot open gzip'ed file $filename: $gzerrno";
425 while ( $gz->gzread($buffer) > 0 ) {
428 die "Error while reading : $gzerrno\n" if $gzerrno != Z_STREAM_END ;
434 sub gzip { # gzip the content argument and save it to filename argument
435 my ($cont, $filename) = @_;
436 #my $debug=1; # debug only
438 print "gzopening $filename ...\n" if (defined($debug));
439 my $gz = gzopen($filename, "w")
440 or die "Cannot gzopen for writing file $filename: $gzerrno";
441 print "gzwriting ...\n" if (defined($debug));
442 $gz->gzwrite($cont) or die "error writing: $gzerrno, exiting!\n";
443 print "gzclosing ...\n" if (defined($debug));
449 my ($remove_headers,$str,$url,$filename,$format) = @_;
452 #my $debug=1; # for DEBUG only
454 my $resourcetype; # retrieve mode: "normal" (.xml) or "gzipped" (.xml.gz)
455 if ($filename =~ /\.xml$/) { # set retrieve mode
456 $resourcetype = "normal";
457 } elsif ($filename =~ /\.xml\.gz$/) {
458 $resourcetype = "gzipped";
460 die "Unsupported download extension, might be '.gz' or '.xml'\n";
462 my $basefname = $filename;
463 $basefname =~ s/\.gz$//; # get base resource name removing trailing .gz
464 $cont = ""; # modified by side-effect by the callback function
467 if (stat($basefname)) {
468 $localfname=$basefname;
469 } elsif (stat($basefname.".gz")) {
470 $localfname=$basefname.".gz";
472 if ($localfname ne "") { # we already have local copy of requested file
473 # check both possible cache entry: gzipped or normal
474 print "Using local copy for the $str file\n";
475 if ($localfname =~ /\.xml\.gz$/) { # deflating cached file and return it
476 $cont = gunzip($localfname);
477 } elsif ($localfname =~ /\.xml$/) { # just return cached file
478 open(FD, $localfname) or die "Cannot open $localfname";
479 while(<FD>) { $cont .= $_; }
482 die "Internal error: unexpected file name $localfname,"
483 ."must end with '.gz' or '.xml.gz'\n";
485 } else { # download file from net
486 print "Downloading the $str file\n"; # download file
487 my $ua = LWP::UserAgent->new;
488 my $request = HTTP::Request->new(GET => "$url");
489 my $response = $ua->request($request, \&callback);
491 # cache retrieved file to disk
492 # <ZACK/> TODO: inefficent, I haven't yet undestood how to deflate
493 # in memory gzipped file, without call "gzopen"
494 print "Storing the $str file\n";
495 print "Making dirs ...\n" if (defined($debug));
497 print "Opening tmp file for writing ...\n" if (defined($debug));
498 open(FD, ">".$filename.".tmp") or die "Cannot open $filename.tmp\n";
499 print "Writing on tmp file ...\n" if (defined($debug));
501 print "Closing tmp file ...\n" if (defined($debug));
504 # handle cache conversion normal->gzipped or gzipped->normal as user choice
505 print "cachemode:$cachemode, resourcetype:$resourcetype\n" if (defined($debug));
506 if (($cachemode eq 'normal') and ($resourcetype eq 'normal')) {
507 # cache the file as is
508 rename "$filename.tmp", $filename;
509 } elsif (($cachemode eq 'gzipped') and ($resourcetype eq 'gzipped')) {
510 # cache the file as is
511 # and update the $cont variabile with deflated content
512 rename "$filename.tmp", $filename;
513 $cont = gunzip($filename);
514 } elsif (($cachemode eq 'normal') and ($resourcetype eq 'gzipped')) {
515 # deflate cache entry
517 open(FD, "> $basefname") or die "cannot open $basefname\n";
518 $cont = gunzip($filename.".tmp");
521 unlink "$filename.tmp"; # delete old gzipped file
522 } elsif (($cachemode eq 'gzipped') and ($resourcetype eq 'normal')) {
523 # compress cache entry
524 print "gzipping ...\n" if (defined($debug));
525 gzip($cont, $basefname.".gz");
526 unlink "$filename.tmp"; # delete old uncompressed file
528 die "Internal error, unsopported cachemode, resourcetype couple\n";
530 # $cont now contained uncompressed data
532 if ($remove_headers) {
533 $cont =~ s/<\?xml [^?]*\?>//sg;
534 $cont =~ s/<!DOCTYPE [^>]*>//sg;
536 $cont =~ s/DOCTYPE (.*) SYSTEM\s+"http:\/\/www.cs.unibo.it\/helm\/dtd\//DOCTYPE $1 SYSTEM "$myownurl\/getdtd?uri=/g;
538 if ($format eq "gz") {
539 gzip($cont,"$basefname.tmp");
540 open (TMP, "< $basefname.tmp")
541 or die "Can't open tempfile: $filename.tmp, exiting!\n";
547 unlink ($basefname . ".tmp") or
548 die "cannot unlink temporary file: $basefname.tmp\n";
557 my $res = new HTTP::Response;
558 $res->content($cont);
559 $c->send_response($res);
563 #retrieve a file from an url and write it to a temp dir
564 #used for retrieve resource index from servers
566 my ($prefix, $URL) = @_;
567 my $ua = LWP::UserAgent->new;
568 my $request = HTTP::Request->new(GET => "$URL");
569 my $response = $ua->request($request, \&callback);
570 my ($filename) = reverse (split "/", $URL); # get filename part of the URL
571 open (TEMP, "> $prefix/$filename")
572 || die "Cannot open temporary file: $prefix/$filename\n";
579 tie(%map, 'DB_File', $uris_dbm.".db", O_RDONLY, 0664);
582 sub mk_urls_of_uris {
583 #rebuild $uris_dbm.db fetching resource indexes from servers
585 $server, $idxfile, $uri, $url, $comp, $line,
591 if (stat $uris_dbm.".db") { # remove old db file
592 unlink($uris_dbm.".db") or
593 die "cannot unlink old db file: $uris_dbm.db\n";
595 tie(%urls_of_uris, 'DB_File', $uris_dbm.".db", O_RDWR|O_CREAT, 0664);
597 open (SRVS, "< $servers_file") or
598 die "cannot open servers file: $servers_file\n";
601 while ($server = pop @servers) { #cicle on servers in reverse order
602 print "processing server: $server ...\n";
604 helm_wget($tmp_dir, $server."/".$indexname); #get index
605 $idxfile = $tmp_dir."/".$indexname;
606 open (INDEX, "< $idxfile") or
607 die "cannot open temporary index file: $idxfile\n";
608 while ($line = <INDEX>) { #parse index and add entry to urls_of_uris
610 ($uri,$comp) = split /[ \t]+/, $line;
613 $url = $uri . ".xml" . ".gz";
615 $url = $uri . ".xml";
617 $url =~ s/cic:/$server/;
618 $url =~ s/theory:/$server/;
619 $urls_of_uris{$uri} = $url;
622 die "cannot unlink temporary file: $idxfile\n"
623 if (unlink $idxfile) != 1;
626 untie(%urls_of_uris);
627 tie(%map, 'DB_File', $uris_dbm.".db", O_RDONLY, 0664);