X-Git-Url: http://matita.cs.unibo.it/gitweb/?a=blobdiff_plain;f=helm%2Fhttp_getter%2Fhttp_getter.pl.in;h=40fd3118d1b1b156635587bbb49e6eaf8e4a03bf;hb=0e5feffa1e6379b84125357ea356907a56d65c3c;hp=0f9c0a1094756a40a7aee2bb0b4e10fc5a926cf5;hpb=fdbdf1853e5bb351f340194e7cecd6210f832e1e;p=helm.git

diff --git a/helm/http_getter/http_getter.pl.in b/helm/http_getter/http_getter.pl.in
index 0f9c0a109..40fd3118d 100755
--- a/helm/http_getter/http_getter.pl.in
+++ b/helm/http_getter/http_getter.pl.in
@@ -23,8 +23,15 @@
 # For details, see the HELM World-Wide-Web page,
 # http://cs.unibo.it/helm/.
 
+#use strict;
+
 my $VERSION = "@VERSION@";
 
+# various variables
+my ($HELM_LIB_PATH);
+my $cgi_dir = "@HELM_CGI_DIR@";
+my (%map, %rdf_map);
+
 # First of all, let's load HELM configuration
 use Env;
 my $HELM_LIB_DIR = $ENV{"HELM_LIB_DIR"};
@@ -45,9 +52,16 @@ if (($cachemode ne 'gzipped') and ($cachemode ne 'normal')) {
 }
 # </ZACK>
 
-# next require defines: $helm_dir, $html_link, $dtd_dir, $uris_dbm
+# 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>
+$helm_rdf_dir = "/usr/local/helm/rdf";
+$rdf_dbm = "/usr/local/helm/rdf_urls";
+$rdf_indexname = "rdf_index.txt";
+# </move_to_conf_file>
+
 # Let's override the configuration file
 $style_dir = $ENV{"HELM_STYLE_DIR"} if (defined ($ENV{"HELM_STYLE_DIR"}));
 $dtd_dir = $ENV{"HELM_DTD_DIR"} if (defined ($ENV{"HELM_DTD_DIR"}));
@@ -76,13 +90,22 @@ $myownurl  =~ s/http:\/\/(.*):(.*)/$1/;
 ($myownurl) = gethostbyname($myownurl);
 $myownurl = "http://".$myownurl.":".$myownport;
 
-tie(%map, 'DB_File', $uris_dbm.".db", O_RDONLY, 0664);
+tie(%map, 'DB_File', $uris_dbm.".db", O_RDWR, 0664); # open dbs
+tie(%rdf_map, 'DB_File', $rdf_dbm.".db", O_RDWR, 0664);
+
+print "\n";
+print "HTTP Getter $VERSION\n"; # print hello information
 print "Please contact me at: <URL:", $myownurl."/", ">\n";
+print "\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 "cache mode: $cachemode\n";
+print "indexname: $indexname\n";
+print "rdf_indexname: $rdf_indexname\n";
+print "\n";
 
 $SIG{CHLD} = "IGNORE"; # do not accumulate defunct processes
 $SIG{USR1} = \&update; # sent by the child to make the parent update
@@ -91,147 +114,111 @@ while (my $c = $d->accept) {
     while (my $r = $c->get_request) {
         #CSC: mancano i controlli di sicurezza
         
-        my $inputuri = uri_unescape($r->url); 
-        $inputuri =~ s/^[^?]*\?uri=(.*)/$1/;
-        print "\nRequest: ".$r->url."\n\n";
         my $http_method = $r->method;
         my $http_path = $r->url->path;
         my $http_query = uri_unescape($r->url->query);
         my $cgi = new CGI("$http_query");
+        my $inputuri = $cgi->param('uri');
+        print "\nRequest: ".$r->url."\n\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 $cicfilename = $cicuri;
-            $cicfilename =~ s/cic:(.*)/$1/;
-            $cicfilename =~ s/theory:(.*)/$1/;
+            my $answerformat = $cgi->param('format');
+            my $patch_dtd = $cgi->param('patch_dtd');
+            $answerformat = "" if (not defined($answerformat));
+            $patch_dtd = "yes" if (not defined($patch_dtd));
+            if (($answerformat ne "gz") and ($answerformat ne "normal")
+               and ($answerformat ne "")) {
+             die "Wrong output format: $answerformat, must be 'normal' ".
+                 "or 'gz'\n";
+            }
+            if (($patch_dtd ne "yes") and ($patch_dtd ne "no")
+               and ($patch_dtd ne "")) {
+             die "Wrong param, patch_dtd must be 'yes' or 'no'\n";
+            }
+
+            my $filename = $inputuri;
+            if (not isRdfUri($inputuri)) { # standad cic: or theory: uri
+             $filename =~ s/^cic:(.*)/$1/;
+             $filename =~ s/^theory:(.*)/$1/;
+            } else { # rdf uri
+             $filename =~ s/^(.*)\/\/cic:(.*)/$2/;
+             $filename =~ s/^(.*)\/\/theory:(.*)/$2/;
+            }
 
-            my $cicurl   = $map{$cicuri};
-            my $extension;
-            if ($cicurl =~ /\.xml$/) { # non gzipped file
+            my $url = resolve ($inputuri); # resolve uri in url
+            if (not defined($url)) {
+             die "uri \"$inputuri\" can't be resolved\n";
+            }
+
+            my $extension; # file extension
+            if ($url =~ /\.xml$/) { # non gzipped file
               $extension = ".xml";
-            } elsif ($cicurl =~ /\.xml\.gz$/) { # gzipped file
+            } elsif ($url =~ /\.xml\.gz$/) { # gzipped file
               $extension = ".xml.gz";
             } else { # error: unknown extension
-             die "unexpected extension in url: $cicurl, might be '.xml'".
+             die "unexpected extension in url: $url, might be '.xml'".
               "or '.xml.gz'";
             }
-            $cicfilename = $helm_dir.$cicfilename.$extension;
-
-            if (!defined($cicurl)) {
-             print "\nNOT FOUND!!!!!\n";
-             $c->send_error(RC_NOT_FOUND)
-            } else {
-               print_request("cic",$cicuri,$cicurl,$cicfilename);
-
-               # Retrieves the file
-               my $ciccontent = download(0,"cic",$cicurl,$cicfilename);
 
-               # Answering the client
-               answer($c,$ciccontent);
-            }
-        } elsif ($http_method eq 'GET' and $http_path eq "/get") {
-            # finds the uris, urls and filenames
-            my $cicuri = $inputuri,
-               $typesuri = $inputuri,
-               $annuri = $inputuri;
-            my $annsuffix;
-            if ($inputuri =~ /\.types$/) {
-               $cicuri    =~ s/(.*)\.types$/$1/;
-               undef($annuri);
-            } elsif ($inputuri =~ /\.types\.ann$/) {
-               $cicuri    =~ s/(.*)\.types\.ann$/$1/;
-               $typesuri  =~ s/(.*)\.ann$/$1/;
-               $annsuffix = ".types.ann";
-            } elsif ($inputuri =~ /\.ann$/) {
-               $cicuri  =~ s/(.*)\.ann$/$1/;
-               undef($typesuri);
-               $annsuffix = ".ann";
+            if (not isRdfUri ($inputuri)) {
+             $filename = $helm_dir.$filename.$extension;
             } else {
-               undef($typesuri);
-               undef($annuri);
+             $filename = $helm_rdf_dir.$filename.$extension;
             }
 
-            my $cicfilename = $cicuri;
-            $cicfilename =~ s/cic:(.*)/$1/;
-            $cicfilename =~ s/theory:(.*)/$1/;
-            $cicfilename = $helm_dir.$cicfilename;
-
-            my $cicurl   = $map{$cicuri};
-            my $typesurl = $map{$typesuri} if (defined($typesuri));
-            my $annurl   = $map{$annuri}  if (defined($annuri));
-            my ($cicext, $typesext, $annext);
-            if ($cicurl =~ /\.xml$/) { # normal file
-             $cicext = ".xml";
-            } elsif ($cicurl =~ /\.xml\.gz$/) { # gzipped file
-             $cicext = ".xml.gz";
-            } else {
-             die "unexpected extension in url: $cicurl;".
-             "might be '.xml' or '.xml.gz'";
-            }
-            if (defined($typesuri)) { # extension selection for types file
-             if ($typesurl =~ /\.xml$/) { # normal file
-              $typesext = ".types.xml";
-             } elsif ($typesurl =~ /\.xml\.gz$/) { # gzipped file
-              $typesext = ".types.xml.gz";
-             } else {
-              die "unexpected extension in url: $typesurl;".
-              "might be '.xml' or '.xml.gz'";
-             }
-            }
-            if (defined($annuri)) { # extension selection for annotation file
-             if ($annurl =~ /\.xml$/) { # normal file
-              $annext = ".xml";
-             } elsif ($annurl =~ /\.xml\.gz$/) { # gzipped file
-              $annext = ".xml.gz";
-             } else {
-              die "unexpected extension in url: $annurl".
-              "might be '.xml' or '.xml.gz'";
-             }
-            }
-            my $typesfilename = $cicfilename.$typesext if $typesuri;
-            my $annfilename  = $cicfilename.$annsuffix.$annext if $annuri;
-            $cicfilename .= $cicext;
-
-            if (!defined($cicurl) ||
-               (!defined($typesurl) && $typesuri) ||
-               (!defined($annuri) && $annuri))
-            {
+            if (!defined($url)) {
              print "\nNOT FOUND!!!!!\n";
              $c->send_error(RC_NOT_FOUND)
             } else {
-               print_request("cic",$cicuri,$cicurl,$cicfilename);
-               print_request("types",$typesuri,$typesurl,$typesfilename)
-                if ($typesuri);
-               print_request("ann",$annuri,$annurl,$annfilename)
-                if ($annuri);
- 
-               # Retrieves the files
+               print_request("cic",$inputuri,$url,$filename);
 
-               my $ciccontent = download(1,"cic",$cicurl,$cicfilename);
-               my $typescontent =
-                download(1,"types",$typesurl,$typesfilename) if ($typesuri);
-               my $anncontent =
-                download(1,"ann",$annurl,$annfilename) if ($annuri);
- 
-               # Merging the files together
- 
-               my $merged = <<EOT;
-<?xml version="1.0" encoding="UTF-8"?>
-<cicxml uri="$cicuri">
-$ciccontent
-$typescontent
-$anncontent
-</cicxml>
-EOT
+               # Retrieves the file
+               my $ciccontent = download($patch_dtd,"cic",$url,$filename,$answerformat);
 
                # Answering the client
-               answer($c,$merged);
+               if ($answerformat eq "normal") {
+                answer($c,$ciccontent,"text/xml","");
+               } else {
+                answer($c,$ciccontent,"text/xml","x-gzip");
+               }
             }
-         } elsif ($http_method eq 'GET' and $http_path eq "/getdtd") {
+          # "/register" does not work with rdf uris
+        } elsif ($http_method eq 'GET' and $http_path eq "/register") {
+          my $inputurl = $cgi->param('url');
+          print "Register requested...\n";
+          $map{$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.$cicfilename.".xml.gz");
+
+          kill(USR1,getppid()); # signal changes to parent
+          untie %map;
+          print "done\n";
+          html_nice_answer($c,"Register done");
+          # "/resolve" works with rdf uri
+        } elsif ($http_method eq 'GET' and $http_path eq "/resolve") {
+          #my $outputurl = $map{$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";
@@ -246,7 +233,7 @@ EOT
 	        $cont .= $_;
 	       }
                close(FD);
-               answer($c,$cont);
+               answer($c,$cont,"text/xml","");
             } else {
                die "Could not find DTD!";
             }
@@ -261,7 +248,7 @@ EOT
 	        $cont .= $_;
 	       }
                close(FD);
-               answer($c,$cont);
+               answer($c,$cont,"text/plain","");
             } else {
                die "Could not find Style Configuration File!";
             }
@@ -281,27 +268,18 @@ EOT
                 $cont .= $_;
                }
                close(FD);
-               answer($c,$cont);
+               answer($c,$cont,"text/xml","");
             } else {
                die "Could not find XSLT!";
             }
-        } elsif ($http_method eq 'GET' and $http_path eq "/conf") {
-            my $quoted_html_link = $html_link;
-            $quoted_html_link =~ s/&/&amp;/g;
-            $quoted_html_link =~ s/</&lt;/g;
-            $quoted_html_link =~ s/>/&gt;/g;
-            $quoted_html_link =~ s/'/&apos;/g;
-            $quoted_html_link =~ s/"/&quot;/g;
-            print "\nConfiguration requested, returned #$quoted_html_link#\n";
-	    $cont = "<?xml version=\"1.0\"?><html_link>$quoted_html_link</html_link>";
-            answer($c,$cont);
         } 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>");
+           html_nice_answer($c,"Update done");
+          # "/ls" does not work with rdf uris
         } elsif ($http_method eq 'GET' and $http_path eq "/ls") {
             # send back keys that begin with a given uri
            my ($uritype,$uripattern,$uriprefix);
@@ -320,24 +298,52 @@ EOT
            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>");
+             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);
+             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>");
+            html_nice_answer($c,"Invalid uri , must begin with ".
+             "'cic:' or 'theory:'");
            }
-        } elsif ($http_method eq 'GET' and $http_path eq "/version") {
-           print "Version requested!";
-           answer($c,"<html><body><h1>HTTP Getter Version ".
-            $VERSION."</h1></body></html>");
-        } else {
+        } elsif ($http_method eq 'GET' and $http_path eq "/help") {
+          # help request
+           print "Help requested!\n";
+           html_nice_answer($c,"HTTP Getter Version: $VERSION");
+        } elsif ($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,", ";
@@ -355,40 +361,76 @@ EOT
 
 #================================
 
+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
+ my ($uri) = @_;
+ if (isRdfUri ($uri)) { # rdf uri, resolve using rdf db
+  return ($rdf_map{$uri});
+ } else { # standard cic: or theory: uri, resolve using std uri db
+  return ($map{$uri});
+ }
+}
+
 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
    $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
+     $basepart =~ s/^([^.]*\.[^.]*)(\.types)?(\.ann)?/$1/;
+                                              # remove exts .types or
                                               # .types.ann
      $flags = $objects{$basepart}; # get old flags
      if ($localpart =~ /\.ann$/) {
-      $flags = add_flag("ann_YES",$flags);
+      $flags = add_flag("ann","YES",$flags);
      } else {
-      $flags = add_flag("ann_NO",$flags);
+      $flags = add_flag("ann","NO",$flags);
      }
      if ($localpart =~ /\.types$/) {
-      $flags = add_flag("types_YES",$flags);
+      $flags = add_flag("types","YES",$flags);
      } elsif ($localpart =~ /\.types\.ann$/) {
-      $flags = add_flag("types_ANN",$flags);
+      $flags = add_flag("types","ANN",$flags);
      } else {
-      $flags = add_flag("types_NO",$flags);
+      $flags = add_flag("types","NO",$flags);
      }
      $objects{$basepart} = $flags; # save new flags
     } else { # exists at least one slash, a DIR
@@ -402,37 +444,28 @@ sub finduris { # find uris for cic and theory trees generation
  }
                             # now generate OUTPUT:
                             # output will be saved in $content
- if ($format eq "txt") { 
+ if ($format eq "txt") { # TXT output
   foreach $key (sort(keys %dirs)) {
    $content .= "dir, " . $key . "\n";
   }
   foreach $key (sort(keys %objects)) {
-   $content .= "object, $key, (" . $objects{$key} . ")\n";
+   $content .= "object, $key, " . $objects{$key} . "\n";
   }
- } elsif ($format eq "xml") {
-
-# output sample:
-# <ls>
-# <section>pippo</section>
-# <section>pluto</section>
-# <object name="qui">
-#  <ann />
-#  <types />
-# </object>
-# <object name="quo">
-#  <conn />
-# </object>
-# </ls>
-
+ } elsif ($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 .= "<ls>\n";
   foreach $key (sort(keys %dirs)) {
    $content .= "\t<section>$key</section>\n";
   }
   foreach $key (sort(keys %objects)) {
    $content .= "\t<object name=\"$key\">\n";
-   foreach (split /,/, $objects{$key}) { # loop on flags
-    $content .= "\t\t<$_ />\n";
-   }
+   $flags = $objects{$key};
+   $flags =~ s/^<(.*)>$/$1/;
+   my ($annflag,$typesflag) = split /,/,$flags;
+   $content .= "\t\t<ann value=\"$annflag\" />\n";
+   $content .= "\t\t<types value=\"$typesflag\" />\n";
    $content .= "\t</object>\n";
   }
   $content .= "</ls>\n";
@@ -443,18 +476,40 @@ sub finduris { # find uris for cic and theory trees generation
 }
 
 sub add_flag {
-# for handle strings of flag, like "flag1,flag2,flag3"
-# add a flag to a string and return the new string
-# if the flag already exists, it is not added
- my ($flag,$str) = @_;
- if ($str =~ /^\w+(,\w+)*$/) { # strings of the form word,word,word,..
-  $str .= ",$flag" if ($str !~ /$flag/);
- } elsif ($str eq "") {
-  $str = $flag;
+# manage string like: "<ann_flag,type_flag>"
+# "ann_flag" may be one of "ann_YES", "ann_NO"
+# "type_flag" may be one of "types_NO", "types_YES", "types_ANN"
+# when adding a flag the max between the current flag and the new flag
+# is taken, the orders are ann_NO < ann_YES and types_NO < types_YES <
+# types_ANN
+ my ($flagtype,$newflag,$str) = @_;
+ $str = "<,>" if ($str eq "");
+ ($str =~ s/^<(.*,.*)>$/$1/) or die "Internal error: ".
+   "wrong string format for flag adding in $str";
+ my ($annflag,$typeflag) = split /,/,$str;
+ if ($flagtype eq "ann") { # ANN flag handling
+  if ($newflag eq "YES") {
+   $annflag = "YES";
+  } elsif ($newflag eq "NO") {
+   $annflag = "NO" unless ($annflag eq "YES");
+  } else {
+   die "Internal error: annflag must be \"YES\" or \"NO\"";
+  }
+ } elsif ($flagtype eq "types") { # TYPES flag handling
+  if ($newflag eq "ANN") {
+   $typeflag = "ANN";
+  } elsif ($newflag eq "YES") {
+   $typeflag = "YES" unless ($typeflag eq "ANN");
+  } elsif ($newflag eq "NO") {
+   $typeflag = "NO"
+     unless (($typeflag eq "ANN") or ($typeflag eq "YES"));
+  } else {
+   die "Internal error: typeflag must be \"YES\", \"NO\" or \"ANN\"";
+  }
  } else {
-  die "Internal error, malformed string: \"$str\" for flag adding";
+  die "Internal error: unsupported flagtype \"$flagtype\"";
  }
- return $str;
+ $str = "<$annflag,$typeflag>";
 }
 
 #CSC: Too much powerful: creates even /home, /home/users/, ...
@@ -505,20 +560,23 @@ sub gunzip { # gunzip a file and return the deflated content
 
 sub gzip {	# gzip the content argument and save it to filename argument
 	my ($cont, $filename) = @_;
+ #my $debug=1; # debug only
 
-	my ($gz, $cont);
-
-	$gz = gzopen($filename, "w")
+ print "gzopening $filename ...\n" if (defined($debug));
+	my $gz = gzopen($filename, "w")
 		or die "Cannot gzopen for writing file $filename: $gzerrno";
-	$gz->gzwrite($cont) or die "error writing: $gzerrno\n" ;
+ print "gzwriting ...\n" if (defined($debug));
+	$gz->gzwrite($cont) or die "error writing: $gzerrno, exiting!\n";
+ print "gzclosing ...\n" if (defined($debug));
 	$gz->gzclose();
 }
 
-sub download
-{
- my ($remove_headers,$str,$url,$filename) = @_;
+sub download {
+ my ($patch_dtd,$str,$url,$filename,$format) = @_;
  my ($gz, $buffer);
 
+ #my $debug=1; # for DEBUG only
+
  my $resourcetype;	# retrieve mode: "normal" (.xml) or "gzipped" (.xml.gz)
  if ($filename =~ /\.xml$/) {	# set retrieve mode
 	 $resourcetype = "normal";
@@ -551,64 +609,94 @@ sub download
    ."must end with '.gz' or '.xml.gz'\n";
   }
  } else { # download file from net
-    print "Downloading the $str file\n"; # download file
-    my $ua = LWP::UserAgent->new;
-    my $request = HTTP::Request->new(GET => "$url");
-    my $response = $ua->request($request, \&callback);
-               
-    # 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 $str file\n";
- mkdirs($filename);
- open(FD, ">".$filename.".tmp") or die "Cannot open $filename.tmp\n";
- print FD $cont;
- close(FD);
-
- # handle cache conversion normal->gzipped or gzipped->normal as user choice
- if (($cachemode eq 'normal') and ($resourcetype eq 'normal')) {
-   # cache the file as is
-  rename "$filename.tmp", $filename; 
- } elsif (($cachemode eq 'gzipped') and ($resourcetype eq 'gzipped')) {
-   # cache the file as is
-   # and update the $cont variabile with deflated content
-  rename "$filename.tmp", $filename; 
-  $cont = gunzip($filename);
- } elsif (($cachemode eq 'normal') and ($resourcetype eq 'gzipped')) {
-   # deflate cache entry
-   # and update $cont
-  open(FD, "> $basefname") or die "cannot open $basefname\n";
-  $cont = gunzip($filename.".tmp");
-  print FD $cont;
-  close(FD);
-  unlink "$filename.tmp"; # delete old gzipped file
- } elsif (($cachemode eq 'gzipped') and ($resourcetype eq 'normal')) {
-   # compress cache entry
-  gzip($cont, $basefname.".gz");
-  unlink "$filename.tmp"; # delete old uncompressed file
- } else {
-  die "Internal error, unsopported cachemode, resourcetype couple\n";
- }
- # $cont now contained uncompressed data
+   print "Downloading the $str file\n"; # download file
+   my $ua = LWP::UserAgent->new;
+   my $request = HTTP::Request->new(GET => "$url");
+   my $response = $ua->request($request, \&callback);
+   
+   # 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));
+   mkdirs($filename);
+   print "Opening tmp file for writing ...\n" if (defined($debug));
+   open(FD, ">".$filename.".tmp") or die "Cannot open $filename.tmp\n";
+   print "Writing on tmp file ...\n" if (defined($debug));
+   print FD $cont;
+   print "Closing tmp file ...\n" if (defined($debug));
+   close(FD);
 
+   # handle cache conversion normal->gzipped or gzipped->normal as user choice
+   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", $filename; 
+   } elsif (($cachemode eq 'gzipped') and ($resourcetype eq 'gzipped')) {
+     # cache the file as is
+     # and update the $cont variabile with deflated content
+    rename "$filename.tmp", $filename; 
+    $cont = gunzip($filename);
+   } elsif (($cachemode eq 'normal') and ($resourcetype eq 'gzipped')) {
+     # deflate cache entry
+     # and update $cont
+    open(FD, "> $basefname") or die "cannot open $basefname\n";
+    $cont = gunzip($filename.".tmp");
+    print FD $cont;
+    close(FD);
+    unlink "$filename.tmp"; # delete old gzipped file
+   } elsif (($cachemode eq 'gzipped') and ($resourcetype eq 'normal')) {
+     # compress cache entry
+    print "gzipping ...\n" if (defined($debug));
+    gzip($cont, $basefname.".gz");
+    unlink "$filename.tmp"; # delete old uncompressed file
+   } else {
+    die "Internal error, unsopported cachemode, resourcetype couple\n";
+   }
+   # $cont now contained uncompressed data
  }
- if ($remove_headers) {
-    $cont =~ s/<\?xml [^?]*\?>//sg;
-    $cont =~ s/<!DOCTYPE [^>]*>//sg;
- } else {
+ if ($patch_dtd eq "yes") {
     $cont =~ s/DOCTYPE (.*) SYSTEM\s+"http:\/\/www.cs.unibo.it\/helm\/dtd\//DOCTYPE $1 SYSTEM "$myownurl\/getdtd?uri=/g;
  }
+ if ($format eq "gz") {
+  gzip($cont,"$basefname.tmp");
+  open (TMP, "< $basefname.tmp")
+   or die "Can't open tempfile: $filename.tmp, exiting!\n";
+  $cont = "";
+  while(<TMP>) {
+   $cont .= $_;
+  }
+  close TMP;
+  unlink ($basefname . ".tmp") or
+   die "cannot unlink temporary file: $basefname.tmp\n";
+ }
+
  return $cont;
 }
 
-sub answer
-{
- my ($c,$cont) = @_;
+sub answer {
+# 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" => $contype)
+   unless ($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);
 }
 
+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 helm_wget {
 #retrieve a file from an url and write it to a temp dir
 #used for retrieve resource index from servers
@@ -626,36 +714,46 @@ sub helm_wget {
 
 sub update {
  untie %map;
- tie(%map, 'DB_File', $uris_dbm.".db", O_RDONLY, 0664);
+ tie(%map, 'DB_File', $uris_dbm.".db", O_RDWR, 0664);
+ untie %rdf_map;
+ tie(%rdf_map, 'DB_File', $uris_dbm.".db", 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) and rdf db (2)
  my (
-  $server, $idxfile, $uri, $url, $comp, $line,
+  $server, $idxfile, $rdf_idxfile, $uri, $url, $comp, $line,
   @servers,
-  %urls_of_uris
+  %urls_of_uris, %rdf # local proxy of $map and $rdf_map
  );
 
  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 %rdf_map;
+ foreach my $file ($uris_dbm, $rdf_dbm) { # remove old db file
+  if (stat $file.".db") { # remove if exists
+   unlink ($file.".db") or die "can't unlink old db file: $file.db\n";
+  }
  }
  tie(%urls_of_uris, 'DB_File', $uris_dbm.".db", O_RDWR|O_CREAT, 0664);
+ tie(%rdf, 'DB_File', $rdf_dbm.".db", 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
   $idxfile = $tmp_dir."/".$indexname;
-  open (INDEX, "< $idxfile") or
+  $rdf_idxfile = $tmp_dir."/".$rdf_indexname;
+
+  open (INDEX, "< $idxfile") or # (1) REBUILD URLS_OF_URIS DB
    die "cannot open temporary index file: $idxfile\n";
-  while ($line = <INDEX>) { #parse index and add entry to urls_of_uris
+  while ($line = <INDEX>) {  # parse index and add entry to urls_of_uris db
    chomp $line;
    ($uri,$comp) = split /[ \t]+/, $line;
              # build url:
@@ -668,12 +766,35 @@ sub mk_urls_of_uris {
    $url =~ s/theory:/$server/;
    $urls_of_uris{$uri} = $url;
   }
-  close INDEX;
-  die "cannot unlink temporary file: $idxfile\n"
+
+  open (RDF_INDEX, "< $rdf_idxfile") or # (2) REBUILD RDF DB
+   die "cannot open temporary rdf index file: $rdf_idxfile\n";
+  while ($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;
+  }
+
+  close INDEX;  # close indexes
+  close RDF_INDEX;
+  die "cannot unlink temporary file: $idxfile\n"  # remove temp files
    if (unlink $idxfile) != 1;
+  die "cannot unlink temporary file: $rdf_idxfile\n"
+   if (unlink $rdf_idxfile) != 1;
+
  }
 
- untie(%urls_of_uris);
- tie(%map, 'DB_File', $uris_dbm.".db", O_RDONLY, 0664);
-}
+ untie(%urls_of_uris); # untie local proxies
+ untie(%rdf);
+ tie(%map, 'DB_File', $uris_dbm.".db", O_RDWR, 0664); # retie global ones
+ tie(%rdf_map, 'DB_File', $rdf_dbm.".db", O_RDWR, 0664);
+} # update_dbs