X-Git-Url: https://sipb.mit.edu/gitweb.cgi/ikiwiki.git/blobdiff_plain/c0f88082323b74d934c8c2d50666418bcd56b3c9..86781fc43ec263967350dd6f12af9f1d05914500:/IkiWiki/Plugin/aggregate.pm

diff --git a/IkiWiki/Plugin/aggregate.pm b/IkiWiki/Plugin/aggregate.pm
index 6f9c78075..a2c10ce5c 100644
--- a/IkiWiki/Plugin/aggregate.pm
+++ b/IkiWiki/Plugin/aggregate.pm
@@ -17,7 +17,7 @@ my %guids;
 sub import { #{{{
 	hook(type => "getopt", id => "aggregate", call => \&getopt);
 	hook(type => "checkconfig", id => "aggregate", call => \&checkconfig);
-	hook(type => "filter", id => "aggregate", call => \&filter);
+	hook(type => "needsbuild", id => "aggregate", call => \&needsbuild);
 	hook(type => "preprocess", id => "aggregate", call => \&preprocess);
         hook(type => "delete", id => "aggregate", call => \&delete);
 	hook(type => "savestate", id => "aggregate", call => \&savestate);
@@ -49,16 +49,18 @@ sub checkconfig () { #{{{
 	}
 } #}}}
 
-sub filter (@) { #{{{
-	my %params=@_;
-	my $page=$params{page};
-
+sub needsbuild (@) { #{{{
+	my $needsbuild=shift;
+	
 	loadstate(); # if not already loaded
-	# Mark all feeds originating on this page as removable;
-	# preprocess will unmark those that still exist.
-	remove_feeds($page);
 
-	return $params{content};
+	foreach my $feed (values %feeds) {
+		if (grep { $_ eq $pagesources{$feed->{sourcepage}} } @$needsbuild) {
+			# Mark all feeds originating on this page as removable;
+			# preprocess will unmark those that still exist.
+			remove_feeds($feed->{sourcepage});
+		}
+	}
 } # }}}
 
 sub preprocess (@) { #{{{
@@ -127,6 +129,7 @@ sub delete (@) { #{{{
 my $state_loaded=0;
 sub loadstate () { #{{{
 	return if $state_loaded;
+	$state_loaded=1;
 	if (-e "$config{wikistatedir}/aggregate") {
 		open(IN, "$config{wikistatedir}/aggregate") ||
 			die "$config{wikistatedir}/aggregate: $!";
@@ -157,17 +160,14 @@ sub loadstate () { #{{{
 		}
 
 		close IN;
-		
-		$state_loaded=1;
 	}
 } #}}}
 
 sub savestate () { #{{{
+	return unless $state_loaded;
 	eval q{use HTML::Entities};
 	error($@) if $@;
 	my $newfile="$config{wikistatedir}/aggregate.new";
-	# TODO: This cleanup function could use improvement. Any newly
-	# aggregated files are left behind unrecorded, and should be deleted.
 	my $cleanup = sub { unlink($newfile) };
 	open (OUT, ">$newfile") || error("open $newfile: $!", $cleanup);
 	foreach my $data (values %feeds, values %guids) {
@@ -180,7 +180,8 @@ sub savestate () { #{{{
 				}
 			}
 			else {
-				unlink pagefile($data->{page});
+				unlink pagefile($data->{page})
+					if exists $data->{page};
 			}
 			next;
 		}
@@ -220,6 +221,7 @@ sub expire () { #{{{
 	foreach my $feed (values %feeds) {
 		next unless $feed->{expireage} || $feed->{expirecount};
 		my $count=0;
+		my %seen;
 		foreach my $item (sort { $IkiWiki::pagectime{$b->{page}} <=> $IkiWiki::pagectime{$a->{page}} }
 		                  grep { exists $_->{page} && $_->{feed} eq $feed->{name} && $IkiWiki::pagectime{$_->{page}} }
 		                  values %guids) {
@@ -227,7 +229,7 @@ sub expire () { #{{{
 				my $days_old = (time - $IkiWiki::pagectime{$item->{page}}) / 60 / 60 / 24;
 				if ($days_old > $feed->{expireage}) {
 					debug(sprintf(gettext("expiring %s (%s days old)"),
-						$item->{page}, $days_old));
+						$item->{page}, int($days_old)));
 					$item->{expired}=1;
 				}
 			}
@@ -237,7 +239,10 @@ sub expire () { #{{{
 				$item->{expired}=1;
 			}
 			else {
-				$count++;
+				if (! $seen{$item->{page}}) {
+					$seen{$item->{page}}=1;
+					$count++;
+				}
 			}
 		}
 	}
@@ -296,6 +301,13 @@ sub aggregate () { #{{{
 			$content=Encode::decode_utf8($content);
 			$f=eval{XML::Feed->parse(\$content)};
 		}
+		if ($@) {
+			# Another possibility is badly escaped entities.
+			$feed->{message}.=" ".sprintf(gettext("(feed entities escaped)"));
+			$content=~s/\&(?!amp)(\w+);/&amp;$1;/g;
+			$content=Encode::decode_utf8($content);
+			$f=eval{XML::Feed->parse(\$content)};
+		}
 		if ($@) {
 			$feed->{message}=gettext("feed crashed XML::Feed!")." ($@)";
 			$feed->{error}=1;
@@ -312,6 +324,7 @@ sub aggregate () { #{{{
 		foreach my $entry ($f->entries) {
 			add_page(
 				feed => $feed,
+				copyright => $f->copyright,
 				title => defined $entry->title ? decode_entities($entry->title) : "untitled",
 				link => $entry->link,
 				content => defined $entry->content->body ? $entry->content->body : "",
@@ -391,6 +404,8 @@ sub add_page (@) { #{{{
 	$template->param(content => htmlescape(htmlabs($params{content}, $feed->{feedurl})));
 	$template->param(name => $feed->{name});
 	$template->param(url => $feed->{url});
+	$template->param(copyright => $params{copyright})
+		if defined $params{copyright} && length $params{copyright};
 	$template->param(permalink => urlabs($params{link}, $feed->{feedurl}))
 		if defined $params{link};
 	if (ref $feed->{tags}) {