X-Git-Url: http://git.vanrenterghem.biz/git.ikiwiki.info.git/blobdiff_plain/bb93fccf0690344aa77f9538a508959a6de09847..db9f2b1b2b97fd92b09f170c3b567ec59030bf0a:/IkiWiki/Plugin/aggregate.pm

diff --git a/IkiWiki/Plugin/aggregate.pm b/IkiWiki/Plugin/aggregate.pm
index 29bc6d0ce..59185e97f 100644
--- a/IkiWiki/Plugin/aggregate.pm
+++ b/IkiWiki/Plugin/aggregate.pm
@@ -4,11 +4,10 @@ package IkiWiki::Plugin::aggregate;
 
 use warnings;
 use strict;
-use IkiWiki 2.00;
+use IkiWiki 3.00;
 use HTML::Parser;
 use HTML::Tagset;
 use HTML::Entities;
-use URI;
 use open qw{:utf8 :std};
 
 my %feeds;
@@ -46,7 +45,7 @@ sub getsetup () {
 		},
 		aggregateinternal => {
 			type => "boolean",
-			example => 0,
+			example => 1,
 			description => "enable aggregation to internal pages?",
 			safe => 0, # enabling needs manual transition
 			rebuild => 0,
@@ -61,6 +60,10 @@ sub getsetup () {
 }
 
 sub checkconfig () {
+	if (! defined $config{aggregateinternal}) {
+		$config{aggregateinternal}=1;
+	}
+
 	if ($config{aggregate} && ! ($config{post_commit} && 
 	                             IkiWiki::commit_hook_enabled())) {
 		launchaggregation();
@@ -206,6 +209,8 @@ sub needsbuild (@) {
 			markunseen($feed->{sourcepage});
 		}
 	}
+
+	return $needsbuild;
 }
 
 sub preprocess (@) {
@@ -245,6 +250,7 @@ sub preprocess (@) {
 	$feed->{template}=$params{template} . ".tmpl";
 	delete $feed->{unseen};
 	$feed->{lastupdate}=0 unless defined $feed->{lastupdate};
+	$feed->{lasttry}=$feed->{lastupdate} unless defined $feed->{lasttry};
 	$feed->{numposts}=0 unless defined $feed->{numposts};
 	$feed->{newposts}=0 unless defined $feed->{newposts};
 	$feed->{message}=gettext("new feed") unless defined $feed->{message};
@@ -293,7 +299,7 @@ sub loadstate () {
 	return if $state_loaded;
 	$state_loaded=1;
 	if (-e "$config{wikistatedir}/aggregate") {
-		open(IN, "$config{wikistatedir}/aggregate") ||
+		open(IN, "<", "$config{wikistatedir}/aggregate") ||
 			die "$config{wikistatedir}/aggregate: $!";
 		while (<IN>) {
 			$_=IkiWiki::possibly_foolish_untaint($_);
@@ -330,7 +336,7 @@ sub savestate () {
 	garbage_collect();
 	my $newfile="$config{wikistatedir}/aggregate.new";
 	my $cleanup = sub { unlink($newfile) };
-	open (OUT, ">$newfile") || error("open $newfile: $!", $cleanup);
+	open (OUT, ">", $newfile) || error("open $newfile: $!", $cleanup);
 	foreach my $data (values %feeds, values %guids) {
 		my @line;
 		foreach my $field (keys %$data) {
@@ -342,7 +348,8 @@ sub savestate () {
 				push @line, "tag=$_" foreach @{$data->{tags}};
 			}
 			else {
-				push @line, "$field=".$data->{$field};
+				push @line, "$field=".$data->{$field}
+					if defined $data->{$field};
 			}
 		}
 		print OUT join(" ", @line)."\n" || error("write $newfile: $!", $cleanup);
@@ -350,6 +357,20 @@ sub savestate () {
 	close OUT || error("save $newfile: $!", $cleanup);
 	rename($newfile, "$config{wikistatedir}/aggregate") ||
 		error("rename $newfile: $!", $cleanup);
+
+	my $timestamp=undef;
+	foreach my $feed (keys %feeds) {
+		my $t=$feeds{$feed}->{lastupdate}+$feeds{$feed}->{updateinterval};
+		if (! defined $timestamp || $timestamp > $t) {
+			$timestamp=$t;
+		}
+	}
+	$newfile=~s/\.new$/time/;
+	open (OUT, ">", $newfile) || error("open $newfile: $!", $cleanup);
+	if (defined $timestamp) {
+		print OUT $timestamp."\n";
+	}
+	close OUT || error("save $newfile: $!", $cleanup);
 }
 
 sub garbage_collect () {
@@ -390,14 +411,15 @@ sub mergestate () {
 	# fields.
 	foreach my $name (keys %myfeeds) {
 		if (exists $feeds{$name}) {
-			foreach my $field (qw{message lastupdate numposts
-			                      newposts error}) {
+			foreach my $field (qw{message lastupdate lasttry
+			                      numposts newposts error}) {
 				$feeds{$name}->{$field}=$myfeeds{$name}->{$field};
 			}
 		}
 	}
 
 	# New guids can be created during aggregation.
+	# Guids have a few fields that may be updated during aggregation.
 	# It's also possible that guids were removed from the on-disk state
 	# while the aggregation was in process. That would only happen if
 	# their feed was also removed, so any removed guids added back here
@@ -406,6 +428,11 @@ sub mergestate () {
 		if (! exists $guids{$guid}) {
 			$guids{$guid}=$myguids{$guid};
 		}
+		else {
+			foreach my $field (qw{md5}) {
+				$guids{$guid}->{$field}=$myguids{$guid}->{$field};
+			}
+		}
 	}
 }
 
@@ -458,10 +485,10 @@ sub aggregate (@) {
 	error($@) if $@;
 
 	foreach my $feed (@_) {
-		$feed->{lastupdate}=time;
+		$feed->{lasttry}=time;
 		$feed->{newposts}=0;
 		$feed->{message}=sprintf(gettext("last checked %s"),
-			displaytime($feed->{lastupdate}));
+			displaytime($feed->{lasttry}));
 		$feed->{error}=0;
 
 		debug(sprintf(gettext("checking feed %s ..."), $feed->{name}));
@@ -483,6 +510,10 @@ sub aggregate (@) {
 			debug($feed->{message});
 			next;
 		}
+
+		# lastupdate is only set if we were able to contact the server
+		$feed->{lastupdate}=$feed->{lasttry};
+
 		if ($res->status == URI::Fetch::URI_GONE()) {
 			$feed->{message}=gettext("feed not found");
 			$feed->{error}=1;
@@ -524,6 +555,11 @@ sub aggregate (@) {
 		}
 
 		foreach my $entry ($f->entries) {
+			# XML::Feed doesn't work around XML::Atom's bizarre
+			# API, so we will. Real unicode strings? Yes please.
+			# See [[bugs/Aggregated_Atom_feeds_are_double-encoded]]
+			local $XML::Atom::ForceUnicode = 1;
+
 			my $c=$entry->content;
 			# atom feeds may have no content, only a summary
 			if (! defined $c && ref $entry->summary) {
@@ -607,7 +643,14 @@ sub add_page (@) {
 	$guid->{md5}=$digest;
 
 	# Create the page.
-	my $template=template($feed->{template}, blind_cache => 1);
+	my $template;
+	eval {
+		$template=template($feed->{template}, blind_cache => 1);
+	};
+	if ($@) {
+		print STDERR gettext("failed to process template:")." $@";
+		return;
+	}
 	$template->param(title => $params{title})
 		if defined $params{title} && length($params{title});
 	$template->param(content => wikiescape(htmlabs($params{content},
@@ -616,7 +659,7 @@ sub add_page (@) {
 	$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}))
+	$template->param(permalink => IkiWiki::urlabs($params{link}, $feed->{feedurl}))
 		if defined $params{link};
 	if (ref $feed->{tags}) {
 		$template->param(tags => [map { tag => $_ }, @{$feed->{tags}}]);
@@ -629,11 +672,13 @@ sub add_page (@) {
 		# creation time on record for the new page.
 		utime $mtime, $mtime, "$config{srcdir}/".htmlfn($guid->{page});
 		# Store it in pagectime for expiry code to use also.
-		$IkiWiki::pagectime{$guid->{page}}=$mtime;
+		$IkiWiki::pagectime{$guid->{page}}=$mtime
+			unless exists $IkiWiki::pagectime{$guid->{page}};
 	}
 	else {
 		# Dummy value for expiry code.
-		$IkiWiki::pagectime{$guid->{page}}=time;
+		$IkiWiki::pagectime{$guid->{page}}=time
+			unless exists $IkiWiki::pagectime{$guid->{page}};
 	}
 }
 
@@ -642,13 +687,6 @@ sub wikiescape ($) {
 	return encode_entities(shift, '\[\]');
 }
 
-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
@@ -674,7 +712,7 @@ sub htmlabs ($$) {
 				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);
+				my $new_v=IkiWiki::urlabs($v, $urlbase);
 				$new_v =~ s/\"/&quot;/g; # since we quote with ""
 				substr($text, $v_offset, $v_len) = qq("$new_v");
 			}