$HELM_LIB_PATH = $DEFAULT_HELM_LIB_DIR."/configuration.pl";
}
-# Let's override the configuration file
-$style_dir = $ENV{"HELM_STYLE_DIR"} if (defined ($ENV{"HELM_STYLE_DIR"}));
-$dtd_dir = $ENV{"HELM_DTD_DIR"} if (defined ($ENV{"HELM_DTD_DIR"}));
-
# <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
require $HELM_LIB_PATH;
+# Let's override the configuration file
+$style_dir = $ENV{"HELM_STYLE_DIR"} if (defined ($ENV{"HELM_STYLE_DIR"}));
+$dtd_dir = $ENV{"HELM_DTD_DIR"} if (defined ($ENV{"HELM_DTD_DIR"}));
+
use HTTP::Daemon;
use HTTP::Status;
use HTTP::Request;
use LWP::UserAgent;
use DB_File;
use Compress::Zlib;
+use CGI;
+use URI::Escape;
#CSC: mancano i controlli sulle condizioni di errore di molte funzioni
#CSC: ==> non e' robusto
while (my $r = $c->get_request) {
#CSC: mancano i controlli di sicurezza
- my $inputuri = $r->url;
+ 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");
+
+ print "\nUnescaped query: ".$http_query."\n";
if ($http_method eq 'GET' and $http_path eq "/getciconly") {
# finds the uri, url and filename
$cont = "<?xml version=\"1.0\"?><html_link>$quoted_html_link</html_link>";
answer($c,$cont);
} elsif ($http_method eq 'GET' and $http_path eq "/update") {
- print "Update requested...";
- update();
- kill(USR1,getppid());
+ # rebuild urls_of_uris.db
+ print "Update requested...\n";
+ mk_urls_of_uris();
+ kill(USR1,getppid()); # signal changes to parent
print " done\n";
answer($c,"<html><body><h1>Update done</h1></body></html>");
- } elsif ($http_method eq 'GET' and $http_path eq "/version") {
- print "Version requested!";
- answer($c,"<html><body><h1>HTTP Getter Version ".$VERSION."</h1></body></html>");
+ } elsif ($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'
+ $uripattern = $baseuri;
+ $uripattern =~ s/^.*:(.*)/$1/;
+ if ($baseuri =~ /^cic:/) {
+ $uritype = "cic";
+ } elsif ($baseuri =~ /^theory:/) {
+ $uritype = "theory";
+ } else {
+ $uritype = "invalid";
+ }
+ if ($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>");
+ } else { # valid output
+ print "BASEURI $baseuri, FORMAT $outype\n";
+ $cont = finduris($uritype,$uripattern,$outype);
+ answer($c,$cont);
+ }
+ } 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>");
+ }
+ } elsif ($http_method eq 'GET' and $http_path eq "/help") {
+ print "Help requested!";
+ answer($c,"<html><body><h1>HTTP Getter Version ".
+ $VERSION."</h1></body></html>");
} else {
- print "\nINVALID REQUEST!!!!!\n";
+ print "\n";
+ print "INVALID REQUEST!!!!!\n";
+ print "(PATH: ",$http_path,", ";
+ print "QUERY: ",$http_query,")\n";
$c->send_error(RC_FORBIDDEN)
}
print "\nRequest solved: ".$r->url."\n\n";
#================================
+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
+
+ print "FINDURIS, uritype: $uritype, uripattern: $uripattern, ".
+ "format: $format\n\n";
+
+ if (($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\/(.*)/$1/;
+
+ if ($localpart =~ /^[^\/]*$/) { # no slash, an OBJECT
+ $basepart = $localpart;
+ $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);
+ } else {
+ $flags = add_flag("ann","NO",$flags);
+ }
+ if ($localpart =~ /\.types$/) {
+ $flags = add_flag("types","YES",$flags);
+ } elsif ($localpart =~ /\.types\.ann$/) {
+ $flags = add_flag("types","ANN",$flags);
+ } else {
+ $flags = add_flag("types","NO",$flags);
+ }
+ $objects{$basepart} = $flags; # save new flags
+ } else { # exists at least one slash, a DIR
+ ($dirname) = split (/\//, $localpart);
+ $dirs{$dirname} = ""; # no flags requirement for dir
+ }
+ }
+ }
+ } else {
+ die "getter internal error: unsupported uritype: \"$uritype\"";
+ }
+ # now generate OUTPUT:
+ # output will be saved in $content
+ if ($format eq "txt") {
+ foreach $key (sort(keys %dirs)) {
+ $content .= "dir, " . $key . "\n";
+ }
+ foreach $key (sort(keys %objects)) {
+ $content .= "object, $key, " . $objects{$key} . "\n";
+ }
+ } elsif ($format eq "xml") {
+ $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";
+ $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";
+ } else { # may not enter this branch
+ die "Getter internal error: invalid format received by finduris sub";
+ }
+ return $content;
+}
+
+sub add_flag {
+# manage string like: "<ann_flag,type_flag>"
+# "ann_flag" may be one of "ann_YES", "ann_NO"
+# "type_flag" may be one of "types_NO", "types_YES", "types_ANN"
+# when adding a flag the max between the current flag and the new flag
+# is taken, the orders are ann_NO < ann_YES and types_NO < types_YES <
+# types_ANN
+ my ($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 ($typeflag eq "YES"));
+ } else {
+ die "Internal error: typeflag must be \"YES\", \"NO\" or \"ANN\"";
+ }
+ } else {
+ die "Internal error: unsupported flagtype \"$flagtype\"";
+ }
+ $str = "<$annflag,$typeflag>";
+}
#CSC: Too much powerful: creates even /home, /home/users/, ...
#CSC: Does not raise errors if could not create dirs/files
}
} else { # download file from net
print "Downloading the $str file\n"; # download file
- $ua = LWP::UserAgent->new;
- $request = HTTP::Request->new(GET => "$url");
- $response = $ua->request($request, \&callback);
+ 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
if ($remove_headers) {
$cont =~ s/<\?xml [^?]*\?>//sg;
$cont =~ s/<!DOCTYPE [^>]*>//sg;
+ } else {
+ $cont =~ s/DOCTYPE (.*) SYSTEM\s+"http:\/\/www.cs.unibo.it\/helm\/dtd\//DOCTYPE $1 SYSTEM "$myownurl\/getdtd?uri=/g;
}
return $cont;
}
$c->send_response($res);
}
+sub helm_wget {
+#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;
+ my $request = HTTP::Request->new(GET => "$URL");
+ my $response = $ua->request($request, \&callback);
+ my ($filename) = reverse (split "/", $URL); # get filename part of the URL
+ open (TEMP, "> $prefix/$filename")
+ || die "Cannot open temporary file: $prefix/$filename\n";
+ print TEMP $cont;
+ close TEMP;
+}
+
sub update {
untie %map;
tie(%map, 'DB_File', $uris_dbm.".db", O_RDONLY, 0664);
}
+
+sub mk_urls_of_uris {
+#rebuild $uris_dbm.db fetching resource indexes from servers
+ my (
+ $server, $idxfile, $uri, $url, $comp, $line,
+ @servers,
+ %urls_of_uris
+ );
+
+ 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";
+ }
+ tie(%urls_of_uris, 'DB_File', $uris_dbm.".db", O_RDWR|O_CREAT, 0664);
+
+ open (SRVS, "< $servers_file") or
+ die "cannot open servers file: $servers_file\n";
+ @servers = <SRVS>;
+ close (SRVS);
+ while ($server = pop @servers) { #cicle on servers in reverse order
+ print "processing server: $server ...\n";
+ chomp $server;
+ helm_wget($tmp_dir, $server."/".$indexname); #get index
+ $idxfile = $tmp_dir."/".$indexname;
+ open (INDEX, "< $idxfile") or
+ die "cannot open temporary index file: $idxfile\n";
+ while ($line = <INDEX>) { #parse index and add entry to urls_of_uris
+ chomp $line;
+ ($uri,$comp) = split /[ \t]+/, $line;
+ # build url:
+ if ($comp =~ /gz/) {
+ $url = $uri . ".xml" . ".gz";
+ } else {
+ $url = $uri . ".xml";
+ }
+ $url =~ s/cic:/$server/;
+ $url =~ s/theory:/$server/;
+ $urls_of_uris{$uri} = $url;
+ }
+ close INDEX;
+ die "cannot unlink temporary file: $idxfile\n"
+ if (unlink $idxfile) != 1;
+ }
+
+ untie(%urls_of_uris);
+ tie(%map, 'DB_File', $uris_dbm.".db", O_RDONLY, 0664);
+}
+
projects / helm.git / blobdiff