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 "/getxml") {
110 # finds the uri, url and filename
111 my $cicuri = $inputuri;
112 my $answerformat = $cgi->param('format');
113 my $patch_dtd = $cgi->param('patch_dtd');
114 $answerformat = "" if (not defined($answerformat));
115 $patch_dtd = "" if (not defined($patch_dtd));
116 if (($answerformat ne "gz") and ($answerformat ne "normal")
117 and ($answerformat ne "")) {
118 die "Wrong output format: $answerformat, must be 'normal' ".
121 if (($patch_dtd ne "yes") and ($patch_dtd ne "no")
122 and ($patch_dtd ne "")) {
123 die "Wrong param, patch_dtd must be 'yes' or 'no'\n";
126 my $cicfilename = $cicuri;
127 $cicfilename =~ s/cic:(.*)/$1/;
128 $cicfilename =~ s/theory:(.*)/$1/;
130 my $cicurl = $map{$cicuri};
131 if (not defined($cicurl)) {
132 die "uri \"$cicuri\" can't be resolved\n";
135 if ($cicurl =~ /\.xml$/) { # non gzipped file
137 } elsif ($cicurl =~ /\.xml\.gz$/) { # gzipped file
138 $extension = ".xml.gz";
139 } else { # error: unknown extension
140 die "unexpected extension in url: $cicurl, might be '.xml'".
143 $cicfilename = $helm_dir.$cicfilename.$extension;
145 if (!defined($cicurl)) {
146 print "\nNOT FOUND!!!!!\n";
147 $c->send_error(RC_NOT_FOUND)
149 print_request("cic",$cicuri,$cicurl,$cicfilename);
152 my $ciccontent = download($patch_dtd,"cic",$cicurl,$cicfilename,$answerformat);
154 # Answering the client
155 answer($c,$ciccontent,"text/xml","x-gzip");
157 } elsif ($http_method eq 'GET' and $http_path eq "/resolve") {
158 my $outputurl = $map{$inputuri};
159 $outputurl = "" if (not defined($outputurl));
160 $cont = "<?xml version=\"1.0\" ?>\n\n";
161 if ($outputurl eq "") {
162 $cont .= "<unresolved />\n";
164 $cont .= "<url value=\"$outputurl\" />\n";
166 answer($c,$cont,"text/xml","");
167 } elsif ($http_method eq 'GET' and $http_path eq "/getdtd") {
168 my $filename = $inputuri;
169 $filename = $dtd_dir."/".$filename;
170 print "DTD: $inputuri ==> ($filename)\n";
171 if (stat($filename)) {
172 print "Using local copy\n";
173 open(FD, $filename) or die "Cannot open $filename\n";
176 # Vary bad heuristic for substituion of absolute URLs
178 s/ENTITY (.*) SYSTEM\s+"/ENTITY $1 SYSTEM "$myownurl\/getdtd?uri=/g;
182 answer($c,$cont,"text/xml","");
184 die "Could not find DTD!";
186 } elsif ($http_method eq 'GET' and $http_path eq "/getstyleconf") {
187 my $filename = $inputuri;
188 $filename = $style_dir."/config/".$filename;
189 if (stat($filename)) {
190 open(FD, $filename) or die "Cannot open $filename\n";
193 s/DOCTYPE (.*) SYSTEM\s+"/DOCTYPE $1 SYSTEM "$myownurl\/getstyleconf?uri=/g;
197 answer($c,$cont,"text/plain","");
199 die "Could not find Style Configuration File!";
201 } elsif ($http_method eq 'GET' and $http_path eq "/getxslt") {
202 my $filename = $inputuri;
203 $filename = $style_dir."/".$filename;
204 print "XSLT: $inputuri ==> ($filename)\n";
205 if (stat($filename)) {
206 print "Using local copy\n";
207 open(FD, $filename) or die "Cannot open $filename\n";
210 # Vary bad heuristic for substituion of absolute URLs
212 s/xsl:import\s+href="/xsl:import href="$myownurl\/getxslt?uri=/g ;
213 s/xsl:include\s+href="/xsl:include href="$myownurl\/getxslt?uri=/g ;
217 answer($c,$cont,"text/xml","");
219 die "Could not find XSLT!";
221 } elsif ($http_method eq 'GET' and $http_path eq "/update") {
222 # rebuild urls_of_uris.db
223 print "Update requested...\n";
225 kill(USR1,getppid()); # signal changes to parent
227 answer($c,"<html><body><h1>Update done</h1></body></html>","text/html","");
228 } elsif ($http_method eq 'GET' and $http_path eq "/ls") {
229 # send back keys that begin with a given uri
230 my ($uritype,$uripattern,$uriprefix);
231 my $baseuri = $cgi->param('baseuri');
232 chop $baseuri if ($baseuri =~ /.*\/$/); # remove trailing "/"
233 my $outype = $cgi->param('format'); # output type, might be 'txt' or 'xml'
234 $uripattern = $baseuri;
235 $uripattern =~ s/^.*:(.*)/$1/;
236 if ($baseuri =~ /^cic:/) {
238 } elsif ($baseuri =~ /^theory:/) {
241 $uritype = "invalid";
243 if ($uritype ne "invalid") { # uri is valid
244 if (($outype ne 'txt') and ($outype ne 'xml')) { # invalid out type
245 print "Invalid output type specified: $outype\n";
246 answer($c,"<html><body><h1>Invalid output type, may be ".
247 "\"txt\" or \"xml\"</h1></body></html>","text/html","");
248 } else { # valid output
249 print "BASEURI $baseuri, FORMAT $outype\n";
250 $cont = finduris($uritype,$uripattern,$outype);
251 if ($outype eq 'txt') {
252 answer($c,$cont,"text/plain","");
253 } elsif ($outype eq 'xml') {
254 answer($c,$cont,"text/xml","");
256 die "Internal error, exiting!";
259 } else { # invalid uri
260 print "Invalid uri: $baseuri, may begin with 'cic:', ".
261 "'theory:' or '*:'\n";
262 answer($c,"<html><body><h1>Invalid uri , may begin with ".
263 "\"cic:\", \"theory:\" or \"*:\"</h1></body></html>","text/html","");
265 } elsif ($http_method eq 'GET' and $http_path eq "/help") {
266 print "Help requested!";
267 answer($c,"<html><body><h1>HTTP Getter Version ".
268 $VERSION."</h1></body></html>","text/html","");
271 print "INVALID REQUEST!!!!!\n";
272 print "(PATH: ",$http_path,", ";
273 print "QUERY: ",$http_query,")\n";
274 $c->send_error(RC_FORBIDDEN)
276 print "\nRequest solved: ".$r->url."\n\n";
280 print "\nCONNECTION CLOSED\n\n";
285 #================================
287 sub finduris { # find uris for cic and theory trees generation
288 my ($uritype,$uripattern,$format) = @_;
290 my ($uri,$localpart,$basepart,$dirname,$suffix,$flags,$key);
291 my (@itemz,@already_pushed_dir);
292 my (%objects,%dirs); # map uris to suffixes' flags
294 print "FINDURIS, uritype: $uritype, uripattern: $uripattern, ".
295 "format: $format\n\n";
297 if (($uritype eq "cic") or ($uritype eq "theory")) {
298 # get info only of one type: cic or theory
299 foreach (keys(%map)) { # select matching uris
301 if ($uri =~ /^$uritype:$uripattern\//) {
303 $localpart =~ s/^$uritype:$uripattern\/(.*)/$1/;
305 if ($localpart =~ /^[^\/]*$/) { # no slash, an OBJECT
306 $basepart = $localpart;
307 $basepart =~ s/^([^.]*\.[^.]*)(\.types)?(\.ann)?/$1/; # remove exts .types or
309 $flags = $objects{$basepart}; # get old flags
310 if ($localpart =~ /\.ann$/) {
311 $flags = add_flag("ann","YES",$flags);
313 $flags = add_flag("ann","NO",$flags);
315 if ($localpart =~ /\.types$/) {
316 $flags = add_flag("types","YES",$flags);
317 } elsif ($localpart =~ /\.types\.ann$/) {
318 $flags = add_flag("types","ANN",$flags);
320 $flags = add_flag("types","NO",$flags);
322 $objects{$basepart} = $flags; # save new flags
323 } else { # exists at least one slash, a DIR
324 ($dirname) = split (/\//, $localpart);
325 $dirs{$dirname} = ""; # no flags requirement for dir
330 die "getter internal error: unsupported uritype: \"$uritype\"";
332 # now generate OUTPUT:
333 # output will be saved in $content
334 if ($format eq "txt") { # TXT output
335 foreach $key (sort(keys %dirs)) {
336 $content .= "dir, " . $key . "\n";
338 foreach $key (sort(keys %objects)) {
339 $content .= "object, $key, " . $objects{$key} . "\n";
341 } elsif ($format eq "xml") { # XML output
342 $content .= '<?xml version="1.0" encoding="ISO-8859-1"?>' . "\n";
343 $content .= "<!DOCTYPE ls SYSTEM ";
344 $content .= "\"$myownurl/getdtd?uri=ls.dtd\">" . "\n\n";
345 $content .= "<ls>\n";
346 foreach $key (sort(keys %dirs)) {
347 $content .= "\t<section>$key</section>\n";
349 foreach $key (sort(keys %objects)) {
350 $content .= "\t<object name=\"$key\">\n";
351 $flags = $objects{$key};
352 $flags =~ s/^<(.*)>$/$1/;
353 my ($annflag,$typesflag) = split /,/,$flags;
354 $content .= "\t\t<ann value=\"$annflag\" />\n";
355 $content .= "\t\t<types value=\"$typesflag\" />\n";
356 $content .= "\t</object>\n";
358 $content .= "</ls>\n";
359 } else { # may not enter this branch
360 die "Getter internal error: invalid format received by finduris sub";
366 # manage string like: "<ann_flag,type_flag>"
367 # "ann_flag" may be one of "ann_YES", "ann_NO"
368 # "type_flag" may be one of "types_NO", "types_YES", "types_ANN"
369 # when adding a flag the max between the current flag and the new flag
370 # is taken, the orders are ann_NO < ann_YES and types_NO < types_YES <
372 my ($flagtype,$newflag,$str) = @_;
373 $str = "<,>" if ($str eq "");
374 ($str =~ s/^<(.*,.*)>$/$1/) or die "Internal error: ".
375 "wrong string format for flag adding in $str";
376 my ($annflag,$typeflag) = split /,/,$str;
377 if ($flagtype eq "ann") { # ANN flag handling
378 if ($newflag eq "YES") {
380 } elsif ($newflag eq "NO") {
381 $annflag = "NO" unless ($annflag eq "YES");
383 die "Internal error: annflag must be \"YES\" or \"NO\"";
385 } elsif ($flagtype eq "types") { # TYPES flag handling
386 if ($newflag eq "ANN") {
388 } elsif ($newflag eq "YES") {
389 $typeflag = "YES" unless ($typeflag eq "ANN");
390 } elsif ($newflag eq "NO") {
392 unless (($typeflag eq "ANN") or ($typeflag eq "YES"));
394 die "Internal error: typeflag must be \"YES\", \"NO\" or \"ANN\"";
397 die "Internal error: unsupported flagtype \"$flagtype\"";
399 $str = "<$annflag,$typeflag>";
402 #CSC: Too much powerful: creates even /home, /home/users/, ...
403 #CSC: Does not raise errors if could not create dirs/files
407 my @dirs = split /\//,$pathname;
409 foreach $dir (@dirs) {
410 $tmp = ((defined($tmp)) ? $tmp."\/".$dir : "");
418 my ($str,$uri,$url,$filename) = @_;
419 print $str."uri: $uri\n";
420 print $str."url: $url\n";
421 print $str."filename: $filename\n\n";
430 sub gunzip { # gunzip a file and return the deflated content
433 my ($gz, $buffer, $cont);
435 print "deflating $filename ...\n";
436 $gz = gzopen($filename, "r")
437 or die "Cannot open gzip'ed file $filename: $gzerrno";
439 while ( $gz->gzread($buffer) > 0 ) {
442 die "Error while reading : $gzerrno\n" if $gzerrno != Z_STREAM_END ;
448 sub gzip { # gzip the content argument and save it to filename argument
449 my ($cont, $filename) = @_;
450 #my $debug=1; # debug only
452 print "gzopening $filename ...\n" if (defined($debug));
453 my $gz = gzopen($filename, "w")
454 or die "Cannot gzopen for writing file $filename: $gzerrno";
455 print "gzwriting ...\n" if (defined($debug));
456 $gz->gzwrite($cont) or die "error writing: $gzerrno, exiting!\n";
457 print "gzclosing ...\n" if (defined($debug));
462 my ($patch_dtd,$str,$url,$filename,$format) = @_;
465 #my $debug=1; # for DEBUG only
467 my $resourcetype; # retrieve mode: "normal" (.xml) or "gzipped" (.xml.gz)
468 if ($filename =~ /\.xml$/) { # set retrieve mode
469 $resourcetype = "normal";
470 } elsif ($filename =~ /\.xml\.gz$/) {
471 $resourcetype = "gzipped";
473 die "Unsupported download extension, might be '.gz' or '.xml'\n";
475 my $basefname = $filename;
476 $basefname =~ s/\.gz$//; # get base resource name removing trailing .gz
477 $cont = ""; # modified by side-effect by the callback function
480 if (stat($basefname)) {
481 $localfname=$basefname;
482 } elsif (stat($basefname.".gz")) {
483 $localfname=$basefname.".gz";
485 if ($localfname ne "") { # we already have local copy of requested file
486 # check both possible cache entry: gzipped or normal
487 print "Using local copy for the $str file\n";
488 if ($localfname =~ /\.xml\.gz$/) { # deflating cached file and return it
489 $cont = gunzip($localfname);
490 } elsif ($localfname =~ /\.xml$/) { # just return cached file
491 open(FD, $localfname) or die "Cannot open $localfname";
492 while(<FD>) { $cont .= $_; }
495 die "Internal error: unexpected file name $localfname,"
496 ."must end with '.gz' or '.xml.gz'\n";
498 } else { # download file from net
499 print "Downloading the $str file\n"; # download file
500 my $ua = LWP::UserAgent->new;
501 my $request = HTTP::Request->new(GET => "$url");
502 my $response = $ua->request($request, \&callback);
504 # cache retrieved file to disk
505 # <ZACK/> TODO: inefficent, I haven't yet undestood how to deflate
506 # in memory gzipped file, without call "gzopen"
507 print "Storing the $str file\n";
508 print "Making dirs ...\n" if (defined($debug));
510 print "Opening tmp file for writing ...\n" if (defined($debug));
511 open(FD, ">".$filename.".tmp") or die "Cannot open $filename.tmp\n";
512 print "Writing on tmp file ...\n" if (defined($debug));
514 print "Closing tmp file ...\n" if (defined($debug));
517 # handle cache conversion normal->gzipped or gzipped->normal as user choice
518 print "cachemode:$cachemode, resourcetype:$resourcetype\n" if (defined($debug));
519 if (($cachemode eq 'normal') and ($resourcetype eq 'normal')) {
520 # cache the file as is
521 rename "$filename.tmp", $filename;
522 } elsif (($cachemode eq 'gzipped') and ($resourcetype eq 'gzipped')) {
523 # cache the file as is
524 # and update the $cont variabile with deflated content
525 rename "$filename.tmp", $filename;
526 $cont = gunzip($filename);
527 } elsif (($cachemode eq 'normal') and ($resourcetype eq 'gzipped')) {
528 # deflate cache entry
530 open(FD, "> $basefname") or die "cannot open $basefname\n";
531 $cont = gunzip($filename.".tmp");
534 unlink "$filename.tmp"; # delete old gzipped file
535 } elsif (($cachemode eq 'gzipped') and ($resourcetype eq 'normal')) {
536 # compress cache entry
537 print "gzipping ...\n" if (defined($debug));
538 gzip($cont, $basefname.".gz");
539 unlink "$filename.tmp"; # delete old uncompressed file
541 die "Internal error, unsopported cachemode, resourcetype couple\n";
543 # $cont now contained uncompressed data
545 if ($patch_dtd eq "yes") {
546 $cont =~ s/DOCTYPE (.*) SYSTEM\s+"http:\/\/www.cs.unibo.it\/helm\/dtd\//DOCTYPE $1 SYSTEM "$myownurl\/getdtd?uri=/g;
548 if ($format eq "gz") {
549 gzip($cont,"$basefname.tmp");
550 open (TMP, "< $basefname.tmp")
551 or die "Can't open tempfile: $filename.tmp, exiting!\n";
557 unlink ($basefname . ".tmp") or
558 die "cannot unlink temporary file: $basefname.tmp\n";
565 my ($c,$cont,$contype,$contenc) = @_;
566 my $res = new HTTP::Response;
567 $res->content($cont);
568 $res->push_header("Content-Type" => $contype);
569 $res->push_header("Content-Encoding" => $contenc)
570 unless ($contenc eq "");
571 $c->send_response($res);
575 #retrieve a file from an url and write it to a temp dir
576 #used for retrieve resource index from servers
578 my ($prefix, $URL) = @_;
579 my $ua = LWP::UserAgent->new;
580 my $request = HTTP::Request->new(GET => "$URL");
581 my $response = $ua->request($request, \&callback);
582 my ($filename) = reverse (split "/", $URL); # get filename part of the URL
583 open (TEMP, "> $prefix/$filename")
584 || die "Cannot open temporary file: $prefix/$filename\n";
591 tie(%map, 'DB_File', $uris_dbm.".db", O_RDONLY, 0664);
594 sub mk_urls_of_uris {
595 #rebuild $uris_dbm.db fetching resource indexes from servers
597 $server, $idxfile, $uri, $url, $comp, $line,
603 if (stat $uris_dbm.".db") { # remove old db file
604 unlink($uris_dbm.".db") or
605 die "cannot unlink old db file: $uris_dbm.db\n";
607 tie(%urls_of_uris, 'DB_File', $uris_dbm.".db", O_RDWR|O_CREAT, 0664);
609 open (SRVS, "< $servers_file") or
610 die "cannot open servers file: $servers_file\n";
613 while ($server = pop @servers) { #cicle on servers in reverse order
614 print "processing server: $server ...\n";
616 helm_wget($tmp_dir, $server."/".$indexname); #get index
617 $idxfile = $tmp_dir."/".$indexname;
618 open (INDEX, "< $idxfile") or
619 die "cannot open temporary index file: $idxfile\n";
620 while ($line = <INDEX>) { #parse index and add entry to urls_of_uris
622 ($uri,$comp) = split /[ \t]+/, $line;
625 $url = $uri . ".xml" . ".gz";
627 $url = $uri . ".xml";
629 $url =~ s/cic:/$server/;
630 $url =~ s/theory:/$server/;
631 $urls_of_uris{$uri} = $url;
634 die "cannot unlink temporary file: $idxfile\n"
635 if (unlink $idxfile) != 1;
638 untie(%urls_of_uris);
639 tie(%map, 'DB_File', $uris_dbm.".db", O_RDONLY, 0664);