X-Git-Url: http://matita.cs.unibo.it/gitweb/?a=blobdiff_plain;f=helm%2Fhttp_getter%2Fhttp_getter.pl.in;h=e06a3a5e2fbd43e6d3f8cc93bfed0131164d3062;hb=faf311442f59ccbc652c7965d23ed2945fd40906;hp=67d295624f966a8faf967ca4aa5cedb6114e9942;hpb=a6d6fdf56d6d9271b86341f0c986f0a265f15e37;p=helm.git diff --git a/helm/http_getter/http_getter.pl.in b/helm/http_getter/http_getter.pl.in index 67d295624..e06a3a5e2 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); + # First of all, let's load HELM configuration use Env; my $HELM_LIB_DIR = $ENV{"HELM_LIB_DIR"}; @@ -76,7 +83,7 @@ $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); print "Please contact me at: \n"; print "helm_dir: $helm_dir\n"; print "style_dir: $style_dir\n"; @@ -91,25 +98,40 @@ 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 $http_path eq "/getciconly") { + if ($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)); + 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 $cicfilename = $cicuri; $cicfilename =~ s/cic:(.*)/$1/; $cicfilename =~ s/theory:(.*)/$1/; my $cicurl = $map{$cicuri}; + if (not defined($cicurl)) { + die "uri \"$cicuri\" can't be resolved\n"; + } my $extension; if ($cicurl =~ /\.xml$/) { # non gzipped file $extension = ".xml"; @@ -128,110 +150,44 @@ while (my $c = $d->accept) { print_request("cic",$cicuri,$cicurl,$cicfilename); # Retrieves the file - my $ciccontent = download(0,"cic",$cicurl,$cicfilename); + my $ciccontent = download($patch_dtd,"cic",$cicurl,$cicfilename,$answerformat); # Answering the client - answer($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); - $annsuffix = ".ann"; - } else { - undef($typesuri); - undef($annuri); - } - - 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)) - { - 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 - - 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 $merged = < - -$ciccontent -$typescontent -$anncontent - -EOT - - # 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 $http_path eq "/getdtd") { + } elsif ($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"); + } elsif ($http_method eq 'GET' and $http_path eq "/resolve") { + my $outputurl = $map{$inputuri}; + $outputurl = "" if (not defined($outputurl)); + $cont = "\n\n"; + if ($outputurl eq "") { + $cont .= "\n"; + } else { + $cont .= "\n"; + } + answer($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 +202,7 @@ EOT $cont .= $_; } close(FD); - answer($c,$cont); + answer($c,$cont,"text/xml",""); } else { die "Could not find DTD!"; } @@ -261,7 +217,7 @@ EOT $cont .= $_; } close(FD); - answer($c,$cont); + answer($c,$cont,"text/plain",""); } else { die "Could not find Style Configuration File!"; } @@ -281,27 +237,17 @@ 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/&/&/g; - $quoted_html_link =~ s//>/g; - $quoted_html_link =~ s/'/'/g; - $quoted_html_link =~ s/"/"/g; - print "\nConfiguration requested, returned #$quoted_html_link#\n"; - $cont = "$quoted_html_link"; - answer($c,$cont); } elsif ($http_method eq 'GET' and $http_path eq "/update") { # rebuild urls_of_uris.db print "Update requested...\n"; mk_urls_of_uris(); kill(USR1,getppid()); # signal changes to parent print " done\n"; - answer($c,"

Update done

"); + html_nice_answer($c,"Update done"); } elsif ($http_method eq 'GET' and $http_path eq "/ls") { # send back keys that begin with a given uri my ($uritype,$uripattern,$uriprefix); @@ -320,23 +266,49 @@ 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,"

Invalid output type, may be ". - "\"txt\" or \"xml\"

"); + 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,"

Invalid uri , may begin with ". - "\"cic:\", \"theory:\" or \"*:\"

"); + 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,"

HTTP Getter Version ". - $VERSION."

"); + print "Help requested!\n"; + html_nice_answer($c,"HTTP Getter Version: $VERSION"); + } elsif ($http_method eq 'GET' and $http_path =~ /\.cgi$/) { + 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 { print "\n"; print "INVALID REQUEST!!!!!\n"; @@ -361,21 +333,32 @@ sub finduris { # find uris for cic and theory trees generation 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 ($localpart =~ /\.ann$/) { @@ -411,8 +394,8 @@ sub finduris { # find uris for cic and theory trees generation } } elsif ($format eq "xml") { # XML output $content .= '' . "\n"; - $content .= '' . "\n\n"; + $content .= "" . "\n\n"; $content .= "\n"; foreach $key (sort(keys %dirs)) { $content .= "\t
$key
\n"; @@ -518,20 +501,23 @@ sub gunzip { # gunzip a file and return the deflated content sub gzip { # gzip the content argument and save it to filename argument my ($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"; @@ -564,64 +550,94 @@ 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 -# 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 + # 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; - $cont =~ s/]*>//sg; - } else { + if ($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() { + $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 = "

$content

"; + 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 @@ -639,7 +655,7 @@ 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); } sub mk_urls_of_uris { @@ -687,6 +703,6 @@ sub mk_urls_of_uris { } untie(%urls_of_uris); - tie(%map, 'DB_File', $uris_dbm.".db", O_RDONLY, 0664); + tie(%map, 'DB_File', $uris_dbm.".db", O_RDWR, 0664); }