X-Git-Url: http://matita.cs.unibo.it/gitweb/?a=blobdiff_plain;f=helm%2Fhttp_getter%2Fhttp_getter.pl.in;h=e9d1f4a3e04609137ad68b8db5658759be797f81;hb=5d7d6bd5090f3f82279bef0b93b4b361a5b1d751;hp=483b71b1dba642540fe86d0296ecbab2ad56c98f;hpb=3370bb5149deef15ba35b913c5641cd3bf9aea01;p=helm.git

diff --git a/helm/http_getter/http_getter.pl.in b/helm/http_getter/http_getter.pl.in
index 483b71b1d..e9d1f4a3e 100755
--- a/helm/http_getter/http_getter.pl.in
+++ b/helm/http_getter/http_getter.pl.in
@@ -29,7 +29,8 @@ my $VERSION = "@VERSION@";
 
 # various variables
 my ($HELM_LIB_PATH);
-my (%map);
+my $cgi_dir = "@HELM_CGI_DIR@";
+my (%xml_map, %rdf_map, %xslt_map);
 
 # First of all, let's load HELM configuration
 use Env;
@@ -42,17 +43,29 @@ if (defined ($HELM_LIB_DIR)) {
    $HELM_LIB_PATH = $DEFAULT_HELM_LIB_DIR."/configuration.pl";
 }
 
-# <ZACK>: TODO temporary, move this setting to configuration file
-# set the cache mode, may be "gzipped" or "normal"
-my $cachemode = $ENV{'HTTP_GETTER_CACHE_MODE'} || 'gzipped';
+# next require defines: $helm_dir, $html_link, $dtd_dir, $uris_dbm, $indexname
+require $HELM_LIB_PATH;
+
+# TEMP: TODO put these vars in configuration file configuration.xml
+# <move_to_conf_file>
+my $cachemode = $ENV{'HTTP_GETTER_CACHE_MODE'} ||
+    'gzipped';
 if (($cachemode ne 'gzipped') and ($cachemode ne 'normal')) {
  die "Invalid HTTP_GETTER_CACHE_MODE environment variable, must be".
  "'normal' or 'gzipped'\n";
 }
-# </ZACK>
 
-# next require defines: $helm_dir, $html_link, $dtd_dir, $uris_dbm
-require $HELM_LIB_PATH;
+$helm_rdf_dir = $ENV{'HTTP_GETTER_RDF_DIR'} || $helm_rdf_dir;
+$rdf_dbm = $ENV{'HTTP_GETTER_RDF_DBM'} || $rdf_dbm;
+$xslt_dbm = $ENV{'HTTP_GETTER_XSLT_DBM'} || $xslt_dbm;
+$rdf_indexname = $ENV{'HTTP_GETTER_RDF_INDEXNAME'} || $rdf_indexname;
+$xslt_indexname = $ENV{'HTTP_GETTER_XSLT_INDEXNAME'} || $xslt_indexname;
+$servers_file = $ENV{'HTTP_GETTER_SERVERS_FILE'} || $servers_file;
+$uris_dbm = $ENV{'HTTP_GETTER_URIS_DBM'} || $uris_dbm;
+$dtdbaseurl = $ENV{'HTTP_GETTER_DTD_BASE_URL'} || $dtdbaseurl;
+$getterport = $ENV{'HTTP_GETTER_PORT'} || $getterport;
+  
+# </move_to_conf_file>
 
 # Let's override the configuration file
 $style_dir = $ENV{"HELM_STYLE_DIR"} if (defined ($ENV{"HELM_STYLE_DIR"}));
@@ -65,13 +78,15 @@ use LWP::UserAgent;
 use DB_File;
 use Compress::Zlib;
 use CGI;
+use URI;
 use URI::Escape;
 
 #CSC: mancano i controlli sulle condizioni di errore di molte funzioni
 #CSC: ==> non e' robusto
 #CSC: altra roba da sistemare segnata con CSC
 
-my $d = new HTTP::Daemon LocalPort => 8081;
+my $d = new HTTP::Daemon LocalPort => $getterport
+ or die "Error: port $getterport not available.";
 my $myownurl = $d->url;
 
 # Let's patch the returned URL
@@ -82,16 +97,33 @@ $myownurl  =~ s/http:\/\/(.*):(.*)/$1/;
 ($myownurl) = gethostbyname($myownurl);
 $myownurl = "http://".$myownurl.":".$myownport;
 
-tie(%map, 'DB_File', $uris_dbm.".db", O_RDONLY, 0664);
+tie(%xml_map, 'DB_File', $uris_dbm, O_RDWR, 0664); # open dbs
+tie(%rdf_map, 'DB_File', $rdf_dbm, O_RDWR, 0664);
+tie(%xslt_map, 'DB_File', $xslt_dbm, O_RDWR, 0664);
+
+print "\n";
+print "HTTP Getter $VERSION\n"; # print hello information
 print "Please contact me at: <URL:", $myownurl."/", ">\n";
+print "\n";
+print "servers_file: $servers_file\n";
 print "helm_dir: $helm_dir\n";
+print "helm_rdf_dir: $helm_rdf_dir\n";
 print "style_dir: $style_dir\n";
 print "dtd_dir: $dtd_dir\n";
-print "urls_of_uris.db: $uris_dbm.db\n";
+print "urls_of_uris db: $uris_dbm\n";
+print "rdf db: $rdf_dbm\n";
+print "xslt db: $xslt_dbm\n";
 print "cache mode: $cachemode\n";
+print "indexname: $indexname\n";
+print "rdf_indexname: $rdf_indexname\n";
+print "xslt_indexname: $xslt_indexname\n";
+print "dtdbaseurl: $dtdbaseurl\n";
+print "getterport: $getterport\n";
+print "\n";
 
 $SIG{CHLD} = "IGNORE"; # do not accumulate defunct processes
 $SIG{USR1} = \&update; # sent by the child to make the parent update
+
 while (my $c = $d->accept) {
  if (fork() == 0) {
     while (my $r = $c->get_request) {
@@ -104,15 +136,16 @@ while (my $c = $d->accept) {
         my $inputuri = $cgi->param('uri');
         print "\nRequest: ".$r->url."\n\n";
 
-	print "\nUnescaped query: ".$http_query."\n";
+        print "\nUnescaped query: ".$http_query."\n";
 
-        if ($http_method eq 'GET' and $http_path eq "/getciconly") {
+         # "getxml" works with rdf uris
+        if ($http_method eq 'GET' and $http_path eq "/getxml") {
             # finds the uri, url and filename
-            my $cicuri = $inputuri;
+
             my $answerformat = $cgi->param('format');
             my $patch_dtd = $cgi->param('patch_dtd');
             $answerformat = "" if (not defined($answerformat));
-            $patch_dtd = "" if (not defined($patch_dtd));
+            $patch_dtd = "yes" if (not defined($patch_dtd));
             if (($answerformat ne "gz") and ($answerformat ne "normal")
                and ($answerformat ne "")) {
              die "Wrong output format: $answerformat, must be 'normal' ".
@@ -123,35 +156,91 @@ while (my $c = $d->accept) {
              die "Wrong param, patch_dtd must be 'yes' or 'no'\n";
             }
 
-            my $cicfilename = $cicuri;
-            $cicfilename =~ s/cic:(.*)/$1/;
-            $cicfilename =~ s/theory:(.*)/$1/;
-
-            my $cicurl   = $map{$cicuri};
-            my $extension;
-            if ($cicurl =~ /\.xml$/) { # non gzipped file
-              $extension = ".xml";
-            } elsif ($cicurl =~ /\.xml\.gz$/) { # gzipped file
-              $extension = ".xml.gz";
-            } else { # error: unknown extension
-             die "unexpected extension in url: $cicurl, might be '.xml'".
-              "or '.xml.gz'";
+            my $filename = $inputuri;
+            my $prefix;
+            if (not isRdfUri($inputuri)) { # standad cic: or theory: uri
+             $filename =~ s/^cic:(.*)/$1/;
+             $filename =~ s/^theory:(.*)/$1/;
+            } else { # rdf uri
+             # The "helm:rdf/<<...>>//cic:" prefix is kept, but quoted
+             # (so that a "helm:rdf/<<...>>//cic:" dir is created
+             $prefix = $filename;
+             $filename =~ s/^(.*)\/\/cic:(.*)/$2/;
+             $filename =~ s/^(.*)\/\/theory:(.*)/$2/;
+             $prefix =~ s/\Q$filename\E//;
+             $prefix =~ s/_/__/g;
+             $prefix =~ s/\//_/g;
+             $filename = $prefix.$filename;
             }
-            $cicfilename = $helm_dir.$cicfilename.$extension;
 
-            if (!defined($cicurl)) {
-             print "\nNOT FOUND!!!!!\n";
+            my $url = resolve ($inputuri); # resolve uri in url
+            if (not defined($url)) { # uri not found in uri2url map
+             die "NOT FOUND: uri \"$inputuri\" can't be resolved\n";
              $c->send_error(RC_NOT_FOUND)
-            } else {
-               print_request("cic",$cicuri,$cicurl,$cicfilename);
-
-               # Retrieves the file
-               my $ciccontent = download($patch_dtd,"cic",$cicurl,$cicfilename,$answerformat);
-
-               # Answering the client
-                answer($c,$ciccontent,$answerformat);
+            } else { # uri found and mapped to url
+              my $extension; # file extension
+              if ($url =~ /\.xml$/) { # non gzipped file
+                $extension = ".xml";
+              } elsif ($url =~ /\.xml\.gz$/) { # gzipped file
+                $extension = ".xml.gz";
+              } else { # error: unknown extension
+               die "unexpected extension in url: $url, might be '.xml'".
+                "or '.xml.gz'";
+              }
+
+              if (not isRdfUri ($inputuri)) { # save in uri std cache dir
+               $filename = $helm_dir.$filename.$extension;
+              } else { # save in rdf metadata cache dir
+               $filename = $helm_rdf_dir."/".$filename.$extension;
+              }
+
+              print_request($inputuri,$url,$filename);
+
+              # Retrieves the file
+              my $ciccontent = download($patch_dtd,$url,$filename,$answerformat);
+
+              if ($answerformat eq "normal") { # answer the client in text/xml
+               answer($c,$ciccontent,"text/xml","");
+              } else { # answer the client in text/xml, gzip encoding
+               answer($c,$ciccontent,"text/xml","x-gzip");
+              }
             }
+            
+                           # "/register" does not work with rdf uris
+        } elsif ($http_method eq 'GET' and $http_path eq "/register") {
+          my $inputurl = $cgi->param('url');
+          print "Register requested...\n";
+          $xml_map{$inputuri}=$inputurl;
+
+          # Now let's clean the cache
+          my $cicfilename = $inputuri;
+          $cicfilename =~ s/cic:(.*)/$1/;
+          $cicfilename =~ s/theory:(.*)/$1/;
+
+          print "Unlinking ".$helm_dir.$cicfilename.".xml[.gz]\n";
+          unlink ($helm_dir.$cicfilename.".xml");
+          unlink ($helm_dir.$cicfilename.".xml.gz");
+
+          kill(USR1,getppid()); # signal changes to parent
+          untie %xml_map;
+          print "done.\n";
+          html_nice_answer($c,"Register done");
+
+                                  # "/resolve" works with rdf uri
+        } elsif ($http_method eq 'GET' and $http_path eq "/resolve") {
+          #my $outputurl = $xml_map{$inputuri};
+          my $outputurl = resolve($inputuri);
+          $outputurl = "" if (not defined($outputurl));
+          $cont = "<?xml version=\"1.0\" ?>\n\n";
+          if ($outputurl eq "") {
+           $cont .= "<unresolved />\n";
+          } else {
+           $cont .= "<url value=\"$outputurl\" />\n";
+          }
+          answer($c,$cont,"text/xml","");
+
         } elsif ($http_method eq 'GET' and $http_path eq "/getdtd") {
+
             my $filename = $inputuri;
             $filename = $dtd_dir."/".$filename;
             print "DTD: $inputuri ==> ($filename)\n";
@@ -166,58 +255,85 @@ while (my $c = $d->accept) {
 	        $cont .= $_;
 	       }
                close(FD);
-               answer($c,$cont,"normal");
+               answer($c,$cont,"text/xml","");
             } else {
                die "Could not find DTD!";
             }
-        } elsif ($http_method eq 'GET' and $http_path eq "/getstyleconf") {
-            my $filename = $inputuri;
-            $filename = $style_dir."/config/".$filename;
-            if (stat($filename)) {
-               open(FD, $filename) or die "Cannot open $filename\n";
-               $cont = "";
-               while(<FD>) {
-	        s/DOCTYPE (.*) SYSTEM\s+"/DOCTYPE $1 SYSTEM "$myownurl\/getstyleconf?uri=/g;
-	        $cont .= $_;
-	       }
-               close(FD);
-               answer($c,$cont,"normal");
-            } else {
-               die "Could not find Style Configuration File!";
-            }
+
         } elsif ($http_method eq 'GET' and $http_path eq "/getxslt") {
-            my $filename = $inputuri;
-            $filename = $style_dir."/".$filename;
-            print "XSLT: $inputuri ==> ($filename)\n";
-            if (stat($filename)) {
-               print "Using local copy\n";
-               open(FD, $filename) or die "Cannot open $filename\n";
-               $cont = "";
-               while(<FD>) {
-                # Vary bad heuristic for substituion of absolute URLs
-                # for relative ones
-                s/xsl:import\s+href="/xsl:import href="$myownurl\/getxslt?uri=/g ;
-                s/xsl:include\s+href="/xsl:include href="$myownurl\/getxslt?uri=/g ;
-                $cont .= $_;
-               }
-               close(FD);
-               answer($c,$cont,"normal");
-            } else {
-               die "Could not find XSLT!";
-            }
+
+          $url = $xslt_map{$inputuri};
+          print "Downloading XSLT ...\n";
+          helm_wget($tmp_dir, $url);
+          print "XSLT downloaded!\n";
+          $inputuri =~ s/.*\///;
+          my $filename = $tmp_dir . "/" . $inputuri;
+          my $cont;
+          die "Could not find XSLT!\n" unless stat($filename);
+          open(FD, $filename) or die "Can't open xslt local copy: '$filename'\n";
+          while(<FD>) {
+            s/xsl:import\s+href="/xsl:import href="$myownurl\/getxslt?uri=/g;
+            s/xsl:include\s+href="/xsl:include href="$myownurl\/getxslt?uri=/g;
+            $cont .= $_;
+          }
+          close(FD);
+          answer($c, $cont, "text/xml", "");
+          unlink $filename or
+            print "Can't unlink XSLT temp file '$filename'";
+
+# <old_version>
+#             my $filename = $inputuri;
+#             $filename = $style_dir."/".$filename;
+#             print "XSLT: $inputuri ==> ($filename)\n";
+#             if (stat($filename)) {
+#                print "Using local copy\n";
+#                open(FD, $filename) or die "Cannot open $filename\n";
+#                $cont = "";
+#                while(<FD>) {
+#                 # Vary bad heuristic for substituion of absolute URLs
+#                 # for relative ones
+#                 s/xsl:import\s+href="/xsl:import href="$myownurl\/getxslt?uri=/g ;
+#                 s/xsl:include\s+href="/xsl:include href="$myownurl\/getxslt?uri=/g ;
+#                 $cont .= $_;
+#                }
+#                close(FD);
+#                answer($c,$cont,"text/xml","");
+#             } else {
+#                die "Could not find XSLT!";
+#             }
+# </old_version>
+
+                                    # "/update" works with rdf uri
         } elsif ($http_method eq 'GET' and $http_path eq "/update") {
-            # rebuild urls_of_uris.db
+            # rebuild urls_of_uris db and rdf uris db
            print "Update requested...\n";
-           mk_urls_of_uris();
+           update_dbs();
            kill(USR1,getppid()); # signal changes to parent
            print " done\n";
-           answer($c,"<html><body><h1>Update done</h1></body></html>","normal");
+           html_nice_answer($c,"Update done");
+
+                                 # "/getalluris" does not work with rdf uris
+        } elsif ($http_method eq 'GET' and $http_path eq "/getalluris") {
+            # send back all the keys in xml
+            print "BASEURI $baseuri, FORMAT $outype\n";
+            $cont = getalluris();
+            answer($c,$cont,"text/xml","");
+
+                                 # "/getallrdfuris"
+        } elsif ($http_method eq 'GET' and $http_path eq "/getallrdfuris") {
+            # send back all the keys in xml
+            my $class = $cgi->param('class');
+            print "BASEURI $baseuri, FORMAT $outype, CLASS $class\n";
+            $cont = getallrdfuris($class);
+            answer($c,$cont,"text/xml","");
+
         } elsif ($http_method eq 'GET' and $http_path eq "/ls") {
             # send back keys that begin with a given uri
            my ($uritype,$uripattern,$uriprefix);
            my $baseuri = $cgi->param('baseuri');
            chop $baseuri if ($baseuri =~ /.*\/$/); # remove trailing "/"
-           my $outype = $cgi->param('format'); # output type, might be 'txt' or 'xml'
+                         # output type, might be 'txt' or 'xml':
+           my $outype = $cgi->param('format');
            $uripattern = $baseuri;
            $uripattern =~ s/^.*:(.*)/$1/;
            if ($baseuri =~ /^cic:/) {
@@ -230,32 +346,76 @@ while (my $c = $d->accept) {
            if ($uritype ne "invalid") { # uri is valid
             if (($outype ne 'txt') and ($outype ne 'xml')) { # invalid out type
              print "Invalid output type specified: $outype\n";
-             answer($c,"<html><body><h1>Invalid output type, may be ".
-              "\"txt\" or \"xml\"</h1></body></html>","normal");
+             html_nice_answer($c,"Invalid output type, must be ".
+              "'txt' or 'xml'");
             } else { # valid output
              print "BASEURI $baseuri, FORMAT $outype\n";
              $cont = finduris($uritype,$uripattern,$outype);
-             answer($c,$cont,"normal");
+             if ($outype eq 'txt') {
+              answer($c,$cont,"text/plain","");
+             } elsif ($outype eq 'xml') {
+              answer($c,$cont,"text/xml","");
+             } else {
+              die "Internal error, exiting!";
+             }
             }
            } else { # invalid uri
             print "Invalid uri: $baseuri, may begin with 'cic:', ".
              "'theory:' or '*:'\n";
-            answer($c,"<html><body><h1>Invalid uri , may begin with ".
-             "\"cic:\", \"theory:\" or \"*:\"</h1></body></html>","normal");
+            html_nice_answer($c,"Invalid uri , must begin with ".
+             "'cic:' or 'theory:'");
            }
+
         } elsif ($http_method eq 'GET' and $http_path eq "/help") {
-           print "Help requested!";
-           answer($c,"<html><body><h1>HTTP Getter Version ".
-            $VERSION."</h1></body></html>","normal");
-        } else {
+          # help request
+           print "Help requested!\n";
+           html_nice_answer($c,"HTTP Getter Version: $VERSION");
+
+        } elsif ($http_method eq 'GET' and $http_path eq "/getempty") {
+          # return an empty, but valid, xml document
+          my $emptyxml = <<EOXML;
+<?xml version="1.0"?>
+<!DOCTYPE empty [
+  <!ELEMENT empty EMPTY>
+]>
+<empty />
+EOXML
+          answer($c, $emptyxml, "", "");
+
+        } elsif ($http_method eq 'GET' and $http_path =~ /\.cgi$/) {
+          # cgi handling
+           print "CGI requested!\n";
+           if ($http_path !~ /^\/[^\/]*\.cgi$/) {
+            html_nice_answer($c,"Invalid CGI name: $http_path, ".
+            "you can't request CGI that contain a slash in the name\n");
+           } elsif (stat "$cgi_dir"."$http_path") {
+            if (not -x "$cgi_dir/$http_path") {
+             html_nice_answer($c,"CGI $http_path found but not ".
+              "executable\n");
+            } else { # exec CGI and anser back its output
+             my %old_ENV = %ENV;
+             %ENV = ();
+             $ENV{'QUERY_STRING'} = $http_query;
+             my $cgi_output = `$cgi_dir/$http_path`;
+             answer($c,$cgi_output,"","");
+             %ENV = %old_ENV;
+            }
+           } else {
+            html_nice_answer($c,"CGI '$http_path' not found ".
+             "in CGI dir '$cgi_dir'");
+           }
+
+        } else { # unsupported request
             print "\n";
             print "INVALID REQUEST!!!!!\n";
             print "(PATH: ",$http_path,", ";
             print "QUERY: ",$http_query,")\n";
             $c->send_error(RC_FORBIDDEN)
         }
+
         print "\nRequest solved: ".$r->url."\n\n";
     }
+
     $c->close;
     undef($c);
     print "\nCONNECTION CLOSED\n\n";
@@ -265,28 +425,113 @@ while (my $c = $d->accept) {
 
 #================================
 
+sub isRdfUri { # return true if the uri is an rdf uri, false otherwise
+# typycal rdf uri:
+# helm:rdf/cic:www.cs.unibo.it/helm/rdf/foo_schema//cic:\
+#  /Coq/Init/Logic/True_rec.con.types
+#
+# the format is "helm:rdf/<metadata_tree>:<metadata_scheme>//<xml_file_uri>"
+#
+ my ($uri) = @_;
+ if ($uri =~ /^helm:rdf(.*):(.*)\/\/(.*)/) {
+  return 1;
+ } else {
+  return 0;
+ }
+}
+
+sub resolve { # resolve an uri in a url, work both with standard cic: or theory:
+              # uris and rdf uris
+ print "RESOLVE subroutine\n";
+ my ($uri) = @_;
+ print "GIVEN URI: \"$uri\"\n";
+ if (isRdfUri ($uri)) { # rdf uri, resolve using rdf db
+  print "IS A RDF URI\n";
+  print "I WILL RETURN '$rdf_map{$uri}'\n";
+  return ($rdf_map{$uri});
+ } else { # standard cic: or theory: uri, resolve using std uri db
+  print "IS NOT A RDF URI\n";
+  print "I WILL RETURN '$xml_map{$uri}'\n";
+  return ($xml_map{$uri});
+ }
+}
+
+sub getalluris { # get all the keys whose prefix is cic
+ my $content = "";
+ my ($uri);
+ my $debug=1; # for debug
+
+  $content .= '<?xml version="1.0" encoding="ISO-8859-1"?>' . "\n";
+  $content .= "<!DOCTYPE alluris SYSTEM ";
+  $content .= "\"$myownurl/getdtd?uri=alluris.dtd\">" . "\n\n";
+  $content .= "<alluris>\n";
+  foreach $uri (sort (keys(%xml_map))) { # select matching uris
+   if ($uri =~ /^cic:/ && not $uri =~ /.types$/) {
+    print "GETALLURI: $uri\n" if defined($debug);
+    $content .= "\t<uri value=\"$uri\"/>\n";
+   }
+  }
+  $content .= "</alluris>\n";
+ return $content;
+}
+
+sub getallrdfuris {
+ my $class = $_[0];
+ my $content = "";
+ my ($uri);
+ my $debug=1; # for debug
+
+  $content .= '<?xml version="1.0" encoding="ISO-8859-1"?>' . "\n";
+  $content .= "<!DOCTYPE allrdfuris SYSTEM ";
+  $content .= "\"$myownurl/getdtd?uri=alluris.dtd\">" . "\n\n";
+  $content .= "<allrdfuris>\n";
+  foreach $uri (sort (keys(%rdf_map))) {
+   if ($class eq "forward" &&
+        $uri =~ /^helm:rdf:www.cs.unibo.it\/helm\/rdf\/forward/
+       ||
+        $class eq "backward" &&
+         $uri =~ /^helm:rdf:www.cs.unibo.it\/helm\/rdf\/backward/) {
+      print "GETALLRDFURI: $uri\n" if defined($debug);
+      $content .= "\t<uri value=\"$uri\"/>\n";
+   }
+  }
+  $content .= "</allrdfuris>\n";
+ return $content;
+}
+
 sub finduris { # find uris for cic and theory trees generation
  my ($uritype,$uripattern,$format) = @_;
  my $content = "";
  my ($uri,$localpart,$basepart,$dirname,$suffix,$flags,$key);
  my (@itemz,@already_pushed_dir);
  my (%objects,%dirs); # map uris to suffixes' flags
+ my $debug=1; # for debug
 
  print "FINDURIS, uritype: $uritype, uripattern: $uripattern, ".
-  "format: $format\n\n";
+  "format: $format\n\n" if defined($debug);
  
  if (($uritype eq "cic") or ($uritype eq "theory")) {
    # get info only of one type: cic or theory
-  foreach (keys(%map)) { # select matching uris
+  foreach (keys(%xml_map)) { # select matching uris
    $uri = $_;
-   if ($uri =~ /^$uritype:$uripattern\//) {
-    $localpart = $uri;
-    $localpart =~ s/^$uritype:$uripattern\/(.*)/$1/;
+   if ($uri =~ /^$uritype:$uripattern(\/|$|\.)/) {
+    if ($uri =~ /^$uritype:$uripattern\//) { # directory match
+     $localpart = $uri;
+     $localpart =~ s/^$uritype:$uripattern\/(.*)/$1/;
+    } elsif ($uri =~ /^$uritype:$uripattern($|\.)/) { # file match
+     $localpart = $uri;
+     $localpart =~ s/^.*\/([^\/]*)/$1/;
+    } else {
+     die "Internal error, seems that requested match is none of ".
+      "directory match or file match";
+    }
+    print "LOCALPART: $localpart\n" if defined($debug);
 
     if ($localpart =~ /^[^\/]*$/) { # no slash, an OBJECT
      $basepart = $localpart;
-     $basepart =~ s/^([^.]*\.[^.]*)(\.types)?(\.ann)?/$1/; # remove exts .types or
-                                              # .types.ann
+     $basepart =~ s/^([^.]*\.[^.]*)((\.body)|(\.types))?(\.ann)?/$1/;
+                                              # remove exts .types, .body,
+                                              # .types.ann or .body.ann
      $flags = $objects{$basepart}; # get old flags
      if ($localpart =~ /\.ann$/) {
       $flags = add_flag("ann","YES",$flags);
@@ -300,6 +545,13 @@ sub finduris { # find uris for cic and theory trees generation
      } else {
       $flags = add_flag("types","NO",$flags);
      }
+     if ($localpart =~ /\.body$/) {
+      $flags = add_flag("body","YES",$flags);
+     } elsif ($localpart =~ /\.body\.ann$/) {
+      $flags = add_flag("body","ANN",$flags);
+     } else {
+      $flags = add_flag("body","NO",$flags);
+     }
      $objects{$basepart} = $flags; # save new flags
     } else { # exists at least one slash, a DIR
      ($dirname) = split (/\//, $localpart);
@@ -321,8 +573,8 @@ sub finduris { # find uris for cic and theory trees generation
   }
  } elsif ($format eq "xml") { # XML output
   $content .= '<?xml version="1.0" encoding="ISO-8859-1"?>' . "\n";
-  $content .= '<!DOCTYPE ls SYSTEM ';
-  $content .= "$myownurl/getdtd?uri=ls.dtd>" . "\n\n";
+  $content .= "<!DOCTYPE ls SYSTEM ";
+  $content .= "\"$myownurl/getdtd?uri=ls.dtd\">" . "\n\n";
   $content .= "<ls>\n";
   foreach $key (sort(keys %dirs)) {
    $content .= "\t<section>$key</section>\n";
@@ -331,9 +583,10 @@ sub finduris { # find uris for cic and theory trees generation
    $content .= "\t<object name=\"$key\">\n";
    $flags = $objects{$key};
    $flags =~ s/^<(.*)>$/$1/;
-   my ($annflag,$typesflag) = split /,/,$flags;
+   my ($annflag,$typesflag,$bodyflag) = split /,/,$flags;
    $content .= "\t\t<ann value=\"$annflag\" />\n";
    $content .= "\t\t<types value=\"$typesflag\" />\n";
+   $content .= "\t\t<body value=\"$bodyflag\" />\n";
    $content .= "\t</object>\n";
   }
   $content .= "</ls>\n";
@@ -344,17 +597,18 @@ sub finduris { # find uris for cic and theory trees generation
 }
 
 sub add_flag {
-# manage string like: "<ann_flag,type_flag>"
+# manage string like: "<ann_flag,type_flag,body_flag>"
 # "ann_flag" may be one of "ann_YES", "ann_NO"
 # "type_flag" may be one of "types_NO", "types_YES", "types_ANN"
+# "body_flag" may be one of "body_NO", "body_YES", "body_ANN"
 # when adding a flag the max between the current flag and the new flag
-# is taken, the orders are ann_NO < ann_YES and types_NO < types_YES <
-# types_ANN
+# is taken, the orders are ann_NO < ann_YES, types_NO < types_YES <
+# types_ANN and body_NO < body_YES < body_ANN
  my ($flagtype,$newflag,$str) = @_;
- $str = "<,>" if ($str eq "");
- ($str =~ s/^<(.*,.*)>$/$1/) or die "Internal error: ".
+ $str = "<,,>" if ($str eq "");
+ ($str =~ s/^<(.*,.*,.*)>$/$1/) or die "Internal error: ".
    "wrong string format for flag adding in $str";
- my ($annflag,$typeflag) = split /,/,$str;
+ my ($annflag,$typeflag,$bodyflag) = split /,/,$str;
  if ($flagtype eq "ann") { # ANN flag handling
   if ($newflag eq "YES") {
    $annflag = "YES";
@@ -374,10 +628,21 @@ sub add_flag {
   } else {
    die "Internal error: typeflag must be \"YES\", \"NO\" or \"ANN\"";
   }
+ } elsif ($flagtype eq "body") { # BODY flag handling
+  if ($newflag eq "ANN") {
+   $bodyflag = "ANN";
+  } elsif ($newflag eq "YES") {
+   $bodyflag = "YES" unless ($bodyflag eq "ANN");
+  } elsif ($newflag eq "NO") {
+   $bodyflag = "NO"
+     unless (($bodyflag eq "ANN") or ($bodyflag eq "YES"));
+  } else {
+   die "Internal error: typeflag must be \"YES\", \"NO\" or \"ANN\"";
+  }
  } else {
   die "Internal error: unsupported flagtype \"$flagtype\"";
  }
- $str = "<$annflag,$typeflag>";
+ $str = "<$annflag,$typeflag,$bodyflag>";
 }
 
 #CSC: Too much powerful: creates even /home, /home/users/, ...
@@ -396,16 +661,12 @@ sub mkdirs
 
 sub print_request
 {
- my ($str,$uri,$url,$filename) = @_;
- print $str."uri: $uri\n";
- print $str."url: $url\n";
- print $str."filename: $filename\n\n";
-}
-
-sub callback
-{
- my ($data) = @_;
- $cont .= $data;
+ my ($uri,$url,$filename) = @_;
+ print "<download_request>\n";
+ print "uri: $uri\n";
+ print "url: $url\n";
+ print "filename: $filename\n\n";
+ print "</download_request>\n";
 }
 
 sub gunzip { # gunzip a file and return the deflated content
@@ -440,10 +701,11 @@ sub gzip {	# gzip the content argument and save it to filename argument
 }
 
 sub download {
- my ($patch_dtd,$str,$url,$filename,$format) = @_;
+ my ($patch_dtd,$url,$filename,$format) = @_;
  my ($gz, $buffer);
 
- #my $debug=1; # for DEBUG only
+# print "DOWNLOAD subs receives url: \"$url\"\n";
+# print "DOWNLOAD subs receives filename: \"$filename\"\n";
 
  my $resourcetype;	# retrieve mode: "normal" (.xml) or "gzipped" (.xml.gz)
  if ($filename =~ /\.xml$/) {	# set retrieve mode
@@ -465,7 +727,7 @@ sub download {
  }
  if ($localfname ne "") { # we already have local copy of requested file
       # check both possible cache entry: gzipped or normal
-  print "Using local copy for the $str file\n";
+  print "Using local copy.\n";
   if ($localfname =~ /\.xml\.gz$/) { # deflating cached file and return it
    $cont = gunzip($localfname);
   } elsif ($localfname =~ /\.xml$/) { # just return cached file
@@ -477,7 +739,7 @@ sub download {
    ."must end with '.gz' or '.xml.gz'\n";
   }
  } else { # download file from net
-   print "Downloading the $str file\n"; # download file
+   print "Downloading file ...\n"; # download file
    my $ua = LWP::UserAgent->new;
    my $request = HTTP::Request->new(GET => "$url");
    my $response = $ua->request($request, \&callback);
@@ -485,18 +747,19 @@ sub download {
    # cache retrieved file to disk
    # <ZACK/> TODO: inefficent, I haven't yet undestood how to deflate
    #    in memory gzipped file, without call "gzopen"
-   print "Storing the $str file\n";
-   print "Making dirs ...\n" if (defined($debug));
+   print "Storing file ...\n";
+#   print "Making dirs ...\n";
    mkdirs($filename);
-   print "Opening tmp file for writing ...\n" if (defined($debug));
+#   print "Opening tmp file for writing ...\n";
    open(FD, ">".$filename.".tmp") or die "Cannot open $filename.tmp\n";
-   print "Writing on tmp file ...\n" if (defined($debug));
+#   print "Writing on tmp file ...\n";
    print FD $cont;
-   print "Closing tmp file ...\n" if (defined($debug));
+#   print "Closing tmp file ...\n";
    close(FD);
 
    # handle cache conversion normal->gzipped or gzipped->normal as user choice
-   print "cachemode:$cachemode, resourcetype:$resourcetype\n" if (defined($debug));
+   print "cachemode:$cachemode, resourcetype:$resourcetype\n"
+    if (defined($debug));
    if (($cachemode eq 'normal') and ($resourcetype eq 'normal')) {
      # cache the file as is
     rename "$filename.tmp", $filename; 
@@ -524,7 +787,7 @@ sub download {
    # $cont now contained uncompressed data
  }
  if ($patch_dtd eq "yes") {
-    $cont =~ s/DOCTYPE (.*) SYSTEM\s+"http:\/\/www.cs.unibo.it\/helm\/dtd\//DOCTYPE $1 SYSTEM "$myownurl\/getdtd?uri=/g;
+    $cont =~ s/DOCTYPE (.*) SYSTEM\s+"$dtdbaseurl\//DOCTYPE $1 SYSTEM "$myownurl\/getdtd?uri=/g;
  }
  if ($format eq "gz") {
   gzip($cont,"$basefname.tmp");
@@ -543,19 +806,37 @@ sub download {
 }
 
 sub answer {
- my ($c,$cont,$contype) = @_; # $contype must be "normal" or "gz"
+# answer the client setting content, Content-Type and Content-Enconding
+# of the answer
+ my ($c,$cont,$contype,$contenc) = @_;
  my $res = new HTTP::Response;
  $res->content($cont);
- $res->push_header("Content-Type" => "text/xml");
- if ($contype eq "gz") {
-  $res->push_header("Content-Encoding" => "x-gzip");
- }
+ $res->push_header("Content-Type" => $contype)
+   unless ($contype eq "");
+ $res->push_header("Content-Encoding" => $contenc)
+   unless ($contenc eq "");
+ $res->push_header("Cache-Control" => "no-cache");
+ $res->push_header("Pragma" => "no-cache");
+ $res->push_header("Expires" => "0");
  $c->send_response($res);
+ $c->force_last_request();
+}
+
+sub html_nice_answer {
+# answer the client whith a nice html document
+ my ($c,$content) = @_;
+ $content = "<html><body><h1>$content</h1></body></html>";
+ answer($c,$content,"text/html","");
+}
+
+sub callback {
+ my ($data) = @_;
+ $cont .= $data;
 }
 
 sub helm_wget {
-#retrieve a file from an url and write it to a temp dir
-#used for retrieve resource index from servers
+# retrieve a file from an url and write it to a temp dir
+# used for retrieve resource index from servers
  $cont = "";
  my ($prefix, $URL) = @_;
  my $ua = LWP::UserAgent->new;
@@ -569,37 +850,55 @@ sub helm_wget {
 }
 
 sub update {
- untie %map;
- tie(%map, 'DB_File', $uris_dbm.".db", O_RDONLY, 0664);
+# retie dbs
+ untie %xml_map;
+ tie(%xml_map, 'DB_File', $uris_dbm, O_RDWR, 0664);
+ untie %rdf_map;
+ tie(%rdf_map, 'DB_File', $rdf_dbm, O_RDWR, 0664);
+ untie %xslt_map;
+ tie(%xslt_map, 'DB_File', $xslt_dbm, O_RDWR, 0664);
 }
 
-sub mk_urls_of_uris {
-#rebuild $uris_dbm.db fetching resource indexes from servers
+sub update_dbs {
+# rebuild dbs fetching resource indexes from servers.
+# Rebuild urls_of_uris db (1), rdf db (2) and xslt db (3)
  my (
-  $server, $idxfile, $uri, $url, $comp, $line,
+  $server, $idxfile, $rdf_idxfile, $uri, $url, $comp, $line,
   @servers,
-  %urls_of_uris
+  %urls_of_uris, %rdf, %xslt # local proxy of dbs
  );
 
- untie %map;
- if (stat $uris_dbm.".db") { # remove old db file
-  unlink($uris_dbm.".db") or
-   die "cannot unlink old db file: $uris_dbm.db\n";
+ untie %xml_map;
+ untie %rdf_map;
+ untie %xslt_map;
+ foreach my $file ($uris_dbm, $rdf_dbm, $xslt_dbm) { # remove old db file
+  if (stat $file) { # remove if exists
+   unlink ($file) or die "can't unlink old db file: $file\n";
+  }
  }
- tie(%urls_of_uris, 'DB_File', $uris_dbm.".db", O_RDWR|O_CREAT, 0664);
+ tie(%urls_of_uris, 'DB_File', $uris_dbm, O_RDWR|O_CREAT, 0664);
+ tie(%rdf, 'DB_File', $rdf_dbm, O_RDWR|O_CREAT, 0664);
+ tie(%xslt, 'DB_File', $xslt_dbm, O_RDWR|O_CREAT, 0664);
 
  open (SRVS, "< $servers_file") or
   die "cannot open servers file: $servers_file\n";
- @servers = <SRVS>;
+ @servers = <SRVS>; # read all servers
  close (SRVS);
- while ($server = pop @servers) { #cicle on servers in reverse order
+
+ while ($server = pop @servers) { # cicle on servers in _reverse_ order
   print "processing server: $server ...\n";
   chomp $server;
-  helm_wget($tmp_dir, $server."/".$indexname); #get index
+  helm_wget($tmp_dir, $server."/".$indexname); # get index
+  helm_wget($tmp_dir, $server."/".$rdf_indexname); # get rdf index
+  helm_wget($tmp_dir, $server."/".$xslt_indexname); # get xslt index
   $idxfile = $tmp_dir."/".$indexname;
+  $rdf_idxfile = $tmp_dir."/".$rdf_indexname;
+  $xslt_idxfile = $tmp_dir . "/" . $xslt_indexname;
+
+                  # (1) REBUILD URLS_OF_URIS DB
   open (INDEX, "< $idxfile") or
    die "cannot open temporary index file: $idxfile\n";
-  while ($line = <INDEX>) { #parse index and add entry to urls_of_uris
+  while (my $line = <INDEX>) {  # parse index and add entry to urls_of_uris db
    chomp $line;
    ($uri,$comp) = split /[ \t]+/, $line;
              # build url:
@@ -612,12 +911,54 @@ sub mk_urls_of_uris {
    $url =~ s/theory:/$server/;
    $urls_of_uris{$uri} = $url;
   }
-  close INDEX;
-  die "cannot unlink temporary file: $idxfile\n"
-   if (unlink $idxfile) != 1;
+
+                  # (2) REBUILD RDF DB
+  open (RDF_INDEX, "< $rdf_idxfile") or
+   die "cannot open temporary RDF index file: $rdf_idxfile\n";
+  while (my $line = <RDF_INDEX>) {  # parse index and add entry to rdf db
+   chomp $line;
+   ($uri,$comp) = split /[ \t]+/, $line; # comp is "gz" or nothing
+             # build url:
+   if ($comp =~ /gz/) { 
+    $url = $uri . ".xml" . ".gz";
+   } else {
+    $url = $uri . ".xml";
+   }
+   $url =~ s/^helm:rdf(.*)\/\/cic:/$server/;
+   $url =~ s/^helm:rdf(.*)\/\/theory:/$server/;
+   $rdf{$uri} = $url;
+  }
+
+                  # (3) REBUILD XSLT DB
+  open (XSLT_INDEX, "< $xslt_idxfile") or
+    die "cannot open temporary XSLT index file: $xslt_idxfile\n";
+  while (my $line = <XSLT_INDEX>) { # parse index and add entry to xslt db
+    chomp $line;
+    my $stylesheet = $line;
+    my $url = $server . "/" . $stylesheet;
+    $xslt{$stylesheet} = $url;
+  }
+
+  close INDEX;  # close indexes
+  close RDF_INDEX;
+  close XSLT_INDEX;
+
+                  # remove temp files
+  foreach my $file ($idxfile, $rdf_idxfile, $xslt_idxfile) {
+    print "cannot unlink temporary file: $file\n" if (unlink $file) != 1;
+  }
+
  }
 
- untie(%urls_of_uris);
- tie(%map, 'DB_File', $uris_dbm.".db", O_RDONLY, 0664);
-}
+ untie(%urls_of_uris); # untie local proxies
+ untie(%rdf);
+ untie(%xslt);
+ tie(%xml_map, 'DB_File', $uris_dbm, O_RDWR, 0664); # retie global ones
+ tie(%rdf_map, 'DB_File', $rdf_dbm, O_RDWR, 0664);
+ tie(%xslt_map, 'DB_File', $xslt_dbm, O_RDWR, 0664);
+
+} # update_dbs
+
+# vim modline: do not remove!
+# vim: set ft=perl: