]> git.vanrenterghem.biz Git - git.ikiwiki.info.git/blobdiff - IkiWiki/Plugin/aggregate.pm
* Encode link and guid urls in rss feeds to avoid illegal utf8 slipping in.
[git.ikiwiki.info.git] / IkiWiki / Plugin / aggregate.pm
index 99b26a546e91adb9ed35a7fea7ecb860884d8868..7c4ab3d106b87bf5cc4ce834bad15e7aaba7c7e1 100644 (file)
@@ -5,6 +5,10 @@ package IkiWiki::Plugin::aggregate;
 use warnings;
 use strict;
 use IkiWiki;
+use HTML::Entities;
+use HTML::Parser;
+use HTML::Tagset;
+use URI;
 
 my %feeds;
 my %guids;
@@ -70,11 +74,11 @@ sub preprocess (@) { #{{{
        $feed->{name}=$name;
        $feed->{sourcepage}=$params{page};
        $feed->{url}=$params{url};
-       my $dir=exists $params{dir} ? $params{dir} : IkiWiki::titlepage($params{name});
+       my $dir=exists $params{dir} ? $params{dir} : $params{page}."/".IkiWiki::titlepage($params{name});
        $dir=~s/^\/+//;
        ($dir)=$dir=~/$IkiWiki::config{wiki_file_regexp}/;
        $feed->{dir}=$dir;
-       $feed->{feedurl}=defined $params{feedurl} ? $params{feedurl} : $params{url};
+       $feed->{feedurl}=defined $params{feedurl} ? $params{feedurl} : "";
        $feed->{updateinterval}=defined $params{updateinterval} ? $params{updateinterval} * 60 : 15 * 60;
        $feed->{expireage}=defined $params{expireage} ? $params{expireage} : 0;
        $feed->{expirecount}=defined $params{expirecount} ? $params{expirecount} : 0;
@@ -83,6 +87,7 @@ sub preprocess (@) { #{{{
        $feed->{numposts}=0 unless defined $feed->{numposts};
        $feed->{newposts}=0 unless defined $feed->{newposts};
        $feed->{message}="new feed" unless defined $feed->{message};
+       $feed->{tags}=[];
        while (@_) {
                my $key=shift;
                my $value=shift;
@@ -93,7 +98,7 @@ sub preprocess (@) { #{{{
 
        return "<a href=\"".$feed->{url}."\">".$feed->{name}."</a>: ".
               "<i>".$feed->{message}."</i> (".$feed->{numposts}.
-              " stored posts; ".$feed->{newposts}." new)";
+              " stored posts; ".$feed->{newposts}." new)<br />";
 } # }}}
 
 sub delete (@) { #{{{
@@ -116,8 +121,9 @@ sub loadstate () { #{{{
                        my $data={};
                        foreach my $i (split(/ /, $_)) {
                                my ($field, $val)=split(/=/, $i, 2);
-                               if ($field eq "name" || $field eq "message") {
-                                       $data->{$field}=IkiWiki::pagetitle($val);
+                               if ($field eq "name" || $field eq "feed" ||
+                                   $field eq "guid" || $field eq "message") {
+                                       $data->{$field}=decode_entities($val, " \t\n");
                                }
                                elsif ($field eq "tag") {
                                        push @{$data->{tags}}, $val;
@@ -140,6 +146,8 @@ sub loadstate () { #{{{
 } #}}}
 
 sub savestate () { #{{{
+       eval q{use HTML::Entities};
+       die $@ if $@;
        open (OUT, ">$IkiWiki::config{wikistatedir}/aggregate" ||
                die "$IkiWiki::config{wikistatedir}/aggregate: $!");
        foreach my $data (values %feeds, values %guids) {
@@ -159,8 +167,9 @@ sub savestate () { #{{{
 
                my @line;
                foreach my $field (keys %$data) {
-                       if ($field eq "name" || $field eq "message") {
-                               push @line, "$field=".IkiWiki::titlepage($data->{$field});
+                       if ($field eq "name" || $field eq "feed" ||
+                           $field eq "guid" || $field eq "message") {
+                               push @line, "$field=".encode_entities($data->{$field}, " \t\n");
                        }
                        elsif ($field eq "tags") {
                                push @line, "tag=$_" foreach @{$data->{tags}};
@@ -180,38 +189,47 @@ sub aggregate () { #{{{
        eval q{use HTML::Entities};
        die $@ if $@;
 
-FEED:  foreach my $feed (values %feeds) {
-               next unless time - $feed->{lastupdate} >= $feed->{updateinterval};
+       foreach my $feed (values %feeds) {
+               next unless $IkiWiki::config{rebuild} || 
+                       time - $feed->{lastupdate} >= $feed->{updateinterval};
                $feed->{lastupdate}=time;
                $feed->{newposts}=0;
                $IkiWiki::forcerebuild{$feed->{sourcepage}}=1;
 
                IkiWiki::debug("checking feed ".$feed->{name}." ...");
 
-               my @urls=XML::Feed->find_feeds($feed->{feedurl});
-               if (! @urls) {
-                       $feed->{message}="could not find feed at ".$feed->{feedurl};
-                       IkiWiki::debug($feed->{message});
-               }
-               foreach my $url (@urls) {
-                       my $f=XML::Feed->parse(URI->new($url));
-                       if (! $f) {
-                               $feed->{message}=XML::Feed->errstr;
+               if (! length $feed->{feedurl}) {
+                       my @urls=XML::Feed->find_feeds($feed->{url});
+                       if (! @urls) {
+                               $feed->{message}="could not find feed at ".$feed->{feedurl};
                                IkiWiki::debug($feed->{message});
-                               next FEED;
+                               next;
                        }
+                       $feed->{feedurl}=pop @urls;
+               }
+               my $f=eval{XML::Feed->parse(URI->new($feed->{feedurl}))};
+               if ($@) {
+                       $feed->{message}="feed crashed XML::Feed! $@";
+                       IkiWiki::debug($feed->{message});
+                       next;
+               }
+               if (! $f) {
+                       $feed->{message}=XML::Feed->errstr;
+                       IkiWiki::debug($feed->{message});
+                       next;
+               }
 
-                       foreach my $entry ($f->entries) {
-                               add_page(
-                                       feed => $feed,
-                                       title => defined $entry->title ? decode_entities($entry->title) : "untitled",
-                                       link => $entry->link,
-                                       content => $entry->content->body,
-                                       guid => defined $entry->id ? $entry->id : time."_".$feed->name,
-                                       ctime => $entry->issued ? ($entry->issued->epoch || time) : time,
-                               );
-                       }
+               foreach my $entry ($f->entries) {
+                       add_page(
+                               feed => $feed,
+                               title => defined $entry->title ? decode_entities($entry->title) : "untitled",
+                               link => $entry->link,
+                               content => $entry->content->body,
+                               guid => defined $entry->id ? $entry->id : time."_".$feed->name,
+                               ctime => $entry->issued ? ($entry->issued->epoch || time) : time,
+                       );
                }
+
                $feed->{message}="processed ok";
        }
 
@@ -237,12 +255,16 @@ sub add_page (@) { #{{{
                $feed->{newposts}++;
 
                # assign it an unused page
-               my $page=$feed->{dir}."/".IkiWiki::titlepage($params{title});
+               my $page=IkiWiki::titlepage($params{title});
+               # escape slashes and periods in title so it doesn't specify
+               # directory name or trigger ".." disallowing code.
+               $page=~s!([/.])!"__".ord($1)."__"!eg;
+               $page=$feed->{dir}."/".$page;
+               $page=lc($page);
                ($page)=$page=~/$IkiWiki::config{wiki_file_regexp}/;
                if (! defined $page || ! length $page) {
                        $page=$feed->{dir}."/item";
                }
-               $page=~s/\.\.//g; # avoid ".." directory tricks
                my $c="";
                while (exists $IkiWiki::pagesources{$page.$c} ||
                       -e pagefile($page.$c)) {
@@ -259,18 +281,18 @@ sub add_page (@) { #{{{
        eval q{use Digest::MD5 'md5_hex'};
        require Encode;
        my $digest=md5_hex(Encode::encode_utf8($params{content}));
-       return unless ! exists $guid->{md5} || $guid->{md5} ne $digest;
+       return unless ! exists $guid->{md5} || $guid->{md5} ne $digest || $IkiWiki::config{rebuild};
        $guid->{md5}=$digest;
 
        # Create the page.
        my $template=IkiWiki::template("aggregatepost.tmpl", blind_cache => 1);
-       my $content=$params{content};
-       $params{content}=~s/(?<!\\)\[\[/\\\[\[/g; # escape accidental wikilinks
-                                                 # and preprocessor stuff
-       $template->param(content => $params{content});
-       $template->param(url => $feed->{url});
+       $template->param(title => $params{title})
+               if defined $params{title} && length($params{title});
+       $template->param(content => htmlescape(htmlabs($params{content}, $feed->{feedurl})));
        $template->param(name => $feed->{name});
-       $template->param(link => $params{link}) if defined $params{link};
+       $template->param(url => $feed->{url});
+       $template->param(permalink => urlabs($params{link}, $feed->{feedurl}))
+               if defined $params{link};
        if (ref $feed->{tags}) {
                $template->param(tags => [map { tag => $_ }, @{$feed->{tags}}]);
        }
@@ -282,6 +304,58 @@ sub add_page (@) { #{{{
        utime $mtime, $mtime, pagefile($guid->{page}) if defined $mtime;
 } #}}}
 
+sub htmlescape ($) { #{{{
+       # escape accidental wikilinks and preprocessor stuff
+       my $html=shift;
+       $html=~s/(?<!\\)\[\[/\\\[\[/g;
+       return $html;
+} #}}}
+
+sub urlabs ($$) { #{{{
+       my $url=shift;
+       my $urlbase=shift;
+
+       URI->new_abs($url, $urlbase)->as_string;
+} #}}}
+
+sub htmlabs ($$) { #{{{
+       # Convert links in html from relative to absolute.
+       # Note that this is a heuristic, which is not specified by the rss
+       # spec and may not be right for all feeds. Also, see Debian
+       # bug #XXXX TODO: get bug.
+       my $html=shift;
+       my $urlbase=shift;
+
+       my $ret="";
+       my $p = HTML::Parser->new(api_version => 3);
+       $p->handler(default => sub { $ret.=join("", @_) }, "text");
+       $p->handler(start => sub {
+               my ($tagname, $pos, $text) = @_;
+               if (ref $HTML::Tagset::linkElements{$tagname}) {
+                       while (4 <= @$pos) {
+                               # use attribute sets from right to left
+                               # to avoid invalidating the offsets
+                               # when replacing the values
+                               my($k_offset, $k_len, $v_offset, $v_len) =
+                                       splice(@$pos, -4);
+                               my $attrname = lc(substr($text, $k_offset, $k_len));
+                               next unless grep { $_ eq $attrname } @{$HTML::Tagset::linkElements{$tagname}};
+                               next unless $v_offset; # 0 v_offset means no value
+                               my $v = substr($text, $v_offset, $v_len);
+                               $v =~ s/^([\'\"])(.*)\1$/$2/;
+                               my $new_v=urlabs($v, $urlbase);
+                               $new_v =~ s/\"/&quot;/g; # since we quote with ""
+                               substr($text, $v_offset, $v_len) = qq("$new_v");
+                       }
+               }
+               $ret.=$text;
+       }, "tagname, tokenpos, text");
+       $p->parse($html);
+       $p->eof;
+
+       return $ret;
+} #}}}
+
 sub remove_feeds () { #{{{
        my $page=shift;