X-Git-Url: http://git.vanrenterghem.biz/git.ikiwiki.info.git/blobdiff_plain/1ac0fd03bb9e840b03aabe516557c24f43a9c450..414ccfb1e298a722d2afa8b8540ca84dc1ccc6f2:/IkiWiki/Plugin/meta.pm

diff --git a/IkiWiki/Plugin/meta.pm b/IkiWiki/Plugin/meta.pm
index 987262a48..220fff9dc 100644
--- a/IkiWiki/Plugin/meta.pm
+++ b/IkiWiki/Plugin/meta.pm
@@ -4,27 +4,32 @@ package IkiWiki::Plugin::meta;
 
 use warnings;
 use strict;
-use IkiWiki 2.00;
+use IkiWiki 3.00;
 
-my %meta;
-my %title;
-my %permalink;
-my %author;
-my %authorurl;
-my %license;
-my %copyright;
+my %metaheaders;
 
-sub import { #{{{
+sub import {
+	hook(type => "getsetup", id => "meta", call => \&getsetup);
 	hook(type => "needsbuild", id => "meta", call => \&needsbuild);
-	hook(type => "preprocess", id => "meta", call => \&preprocess);
+	hook(type => "preprocess", id => "meta", call => \&preprocess, scan => 1);
 	hook(type => "pagetemplate", id => "meta", call => \&pagetemplate);
-} # }}}
+}
+
+sub getsetup () {
+	return
+		plugin => {
+			safe => 1,
+			rebuild => undef,
+			section => "core",
+		},
+}
 
-sub needsbuild (@) { #{{{
+sub needsbuild (@) {
 	my $needsbuild=shift;
 	foreach my $page (keys %pagestate) {
 		if (exists $pagestate{$page}{meta}) {
-			if (grep { $_ eq $pagesources{$page} } @$needsbuild) {
+			if (exists $pagesources{$page} &&
+			    grep { $_ eq $pagesources{$page} } @$needsbuild) {
 				# remove state, it will be re-added
 				# if the preprocessor directive is still
 				# there during the rebuild
@@ -32,18 +37,40 @@ sub needsbuild (@) { #{{{
 			}
 		}
 	}
+	return $needsbuild;
 }
 
-sub scrub ($) { #{{{
+sub scrub ($$$) {
 	if (IkiWiki::Plugin::htmlscrubber->can("sanitize")) {
-		return IkiWiki::Plugin::htmlscrubber::sanitize(content => shift);
+		return IkiWiki::Plugin::htmlscrubber::sanitize(
+			content => shift, page => shift, destpage => shift);
 	}
 	else {
 		return shift;
 	}
-} #}}}
+}
+
+sub safeurl ($) {
+	my $url=shift;
+	if (exists $IkiWiki::Plugin::htmlscrubber::{safe_url_regexp} &&
+	    defined $IkiWiki::Plugin::htmlscrubber::safe_url_regexp) {
+		return $url=~/$IkiWiki::Plugin::htmlscrubber::safe_url_regexp/;
+	}
+	else {
+		return 1;
+	}
+}
 
-sub preprocess (@) { #{{{
+sub htmlize ($$$) {
+	my $page = shift;
+	my $destpage = shift;
+
+	return IkiWiki::htmlize($page, $destpage, pagetype($pagesources{$page}),
+		IkiWiki::linkify($page, $destpage,
+		IkiWiki::preprocess($page, $destpage, shift)));
+}
+
+sub preprocess (@) {
 	return "" unless @_;
 	my %params=@_;
 	my $key=shift;
@@ -60,12 +87,57 @@ sub preprocess (@) { #{{{
 	# fully encoded.
 	$value=decode_entities($value);
 
+	# Metadata collection that needs to happen during the scan pass.
 	if ($key eq 'title') {
-		$title{$page}=HTML::Entities::encode_numeric($value);
+		$pagestate{$page}{meta}{title}=$value;
+		if (exists $params{sortas}) {
+			$pagestate{$page}{meta}{titlesort}=$params{sortas};
+		}
+		else {
+			delete $pagestate{$page}{meta}{titlesort};
+		}
+		return "";
+	}
+	elsif ($key eq 'description') {
+		$pagestate{$page}{meta}{description}=$value;
+		# fallthrough
+	}
+	elsif ($key eq 'guid') {
+		$pagestate{$page}{meta}{guid}=$value;
+		# fallthrough
+	}
+	elsif ($key eq 'license') {
+		push @{$metaheaders{$page}}, '<link rel="license" href="#page_license" />';
+		$pagestate{$page}{meta}{license}=$value;
+		return "";
+	}
+	elsif ($key eq 'copyright') {
+		push @{$metaheaders{$page}}, '<link rel="copyright" href="#page_copyright" />';
+		$pagestate{$page}{meta}{copyright}=$value;
+		return "";
+	}
+	elsif ($key eq 'link' && ! %params) {
+		# hidden WikiLink
+		add_link($page, $value);
+		return "";
+	}
+	elsif ($key eq 'author') {
+		$pagestate{$page}{meta}{author}=$value;
+		if (exists $params{sortas}) {
+			$pagestate{$page}{meta}{authorsort}=$params{sortas};
+		}
+		else {
+			delete $pagestate{$page}{meta}{authorsort};
+		}
+		# fallthorough
+	}
+	elsif ($key eq 'authorurl') {
+		$pagestate{$page}{meta}{authorurl}=$value if safeurl($value);
+		# fallthrough
 	}
 	elsif ($key eq 'permalink') {
-		$permalink{$page}=$value;
-		push @{$meta{$page}}, scrub('<link rel="bookmark" href="'.encode_entities($value).'" />');
+		$pagestate{$page}{meta}{permalink}=$value if safeurl($value);
+		# fallthrough
 	}
 	elsif ($key eq 'date') {
 		eval q{use Date::Parse};
@@ -74,6 +146,25 @@ sub preprocess (@) { #{{{
 			$IkiWiki::pagectime{$page}=$time if defined $time;
 		}
 	}
+	elsif ($key eq 'updated') {
+		eval q{use Date::Parse};
+		if (! $@) {
+			my $time = str2time($value);
+			$pagestate{$page}{meta}{updated}=$time if defined $time;
+		}
+	}
+
+	if (! defined wantarray) {
+		# avoid collecting duplicate data during scan pass
+		return;
+	}
+
+	# Metadata handling that happens only during preprocessing pass.
+	if ($key eq 'permalink') {
+		if (safeurl($value)) {
+			push @{$metaheaders{$page}}, scrub('<link rel="bookmark" href="'.encode_entities($value).'" />', $page, $destpage);
+		}
+	}
 	elsif ($key eq 'stylesheet') {
 		my $rel=exists $params{rel} ? $params{rel} : "alternate stylesheet";
 		my $title=exists $params{title} ? $params{title} : $value;
@@ -81,28 +172,57 @@ sub preprocess (@) { #{{{
 		# editable page as a stylesheet
 		my $stylesheet=bestlink($page, $value.".css");
 		if (! length $stylesheet) {
-			return "[[meta ".gettext("stylesheet not found")."]]";
+			error gettext("stylesheet not found")
 		}
-		push @{$meta{$page}}, '<link href="'.urlto($stylesheet, $page).
+		push @{$metaheaders{$page}}, scrub('<link href="'.urlto($stylesheet, $page).
 			'" rel="'.encode_entities($rel).
 			'" title="'.encode_entities($title).
-			"\" type=\"text/css\" />";
+			"\" type=\"text/css\" />", $page, $destpage);
 	}
-	elsif ($key eq 'openid') {
-		if (exists $params{server}) {
-			push @{$meta{$page}}, '<link href="'.encode_entities($params{server}).
-				'" rel="openid.server" />';
+	elsif ($key eq 'script') {
+		my $defer=exists $params{defer} ? ' defer="defer"' : '';
+		my $async=exists $params{async} ? ' async="async"' : '';
+		my $js=bestlink($page, $value.".js");
+		if (! length $js) {
+			error gettext("script not found");
 		}
-		push @{$meta{$page}}, '<link href="'.encode_entities($value).
-			'" rel="openid.delegate" />';
+		push @{$metaheaders{$page}}, scrub('<script src="'.urlto($js, $page).
+			'"' . $defer . $async . ' type="text/javascript"></script>',
+			$page, $destpage);
 	}
-	elsif ($key eq 'license') {
-		push @{$meta{$page}}, '<link rel="license" href="#page_license" />';
-		$license{$page}=$value;
+	elsif ($key eq 'openid') {
+		my $delegate=0; # both by default
+		if (exists $params{delegate}) {
+			$delegate = 1 if lc $params{delegate} eq 'openid';
+			$delegate = 2 if lc $params{delegate} eq 'openid2';
+		}
+		if (exists $params{server} && safeurl($params{server})) {
+			push @{$metaheaders{$page}}, '<link href="'.encode_entities($params{server}).
+				'" rel="openid.server" />' if $delegate ne 2;
+			push @{$metaheaders{$page}}, '<link href="'.encode_entities($params{server}).
+				'" rel="openid2.provider" />' if $delegate ne 1;
+		}
+		if (safeurl($value)) {
+			push @{$metaheaders{$page}}, '<link href="'.encode_entities($value).
+				'" rel="openid.delegate" />' if $delegate ne 2;
+			push @{$metaheaders{$page}}, '<link href="'.encode_entities($value).
+				'" rel="openid2.local_id" />' if $delegate ne 1;
+		}
+		if (exists $params{"xrds-location"} && safeurl($params{"xrds-location"})) {
+			# force url absolute
+			eval q{use URI};
+			error($@) if $@;
+			my $url=URI->new_abs($params{"xrds-location"}, $config{url});
+			push @{$metaheaders{$page}}, '<meta http-equiv="X-XRDS-Location" '.
+				'content="'.encode_entities($url).'" />';
+		}
 	}
-	elsif ($key eq 'copyright') {
-		push @{$meta{$page}}, '<link rel="copyright" href="#page_copyright" />';
-		$copyright{$page}=$value;
+	elsif ($key eq 'foaf') {
+		if (safeurl($value)) {
+			push @{$metaheaders{$page}}, '<link rel="meta" '.
+				'type="application/rdf+xml" title="FOAF" '.
+				'href="'.encode_entities($value).'" />';
+		}
 	}
 	elsif ($key eq 'redir') {
 		return "" if $page ne $destpage;
@@ -110,11 +230,11 @@ sub preprocess (@) { #{{{
 		if ($value !~ /^\w+:\/\//) {
 			my ($redir_page, $redir_anchor) = split /\#/, $value;
 
-			add_depends($page, $redir_page);
 			my $link=bestlink($page, $redir_page);
 			if (! length $link) {
-				return "[[meta ".gettext("redir page not found")."]]";
+				error gettext("redir page not found")
 			}
+			add_depends($page, $link, deptype("presence"));
 
 			$value=urlto($link, $page);
 			$value.='#'.$redir_anchor if defined $redir_anchor;
@@ -126,7 +246,7 @@ sub preprocess (@) { #{{{
 			my %seen;
 			while (exists $pagestate{$at}{meta}{redir}) {
 				if ($seen{$at}) {
-					return "[[meta ".gettext("redir cycle is not allowed")."]]";
+					error gettext("redir cycle is not allowed")
 				}
 				$seen{$at}=1;
 				$at=$pagestate{$at}{meta}{redir};
@@ -136,70 +256,191 @@ sub preprocess (@) { #{{{
 			$value=encode_entities($value);
 		}
 		my $delay=int(exists $params{delay} ? $params{delay} : 0);
-		my $redir="<meta http-equiv=\"refresh\" content=\"$delay; URL=$value\">";
+		my $redir="<meta http-equiv=\"refresh\" content=\"$delay; URL=$value\" />";
 		if (! $safe) {
-			$redir=scrub($redir);
+			$redir=scrub($redir, $page, $destpage);
 		}
-		push @{$meta{$page}}, $redir;
+		push @{$metaheaders{$page}}, $redir;
 	}
 	elsif ($key eq 'link') {
-		return "[[meta ".gettext("link is no longer supported")."]]";
+		if (%params) {
+			push @{$metaheaders{$page}}, scrub("<link href=\"".encode_entities($value)."\" ".
+				join(" ", map {
+					encode_entities($_)."=\"".encode_entities(decode_entities($params{$_}))."\""
+				} keys %params).
+				" />\n", $page, $destpage);
+		}
+	}
+	elsif ($key eq 'robots') {
+		push @{$metaheaders{$page}}, '<meta name="robots"'.
+			' content="'.encode_entities($value).'" />';
+	}
+	elsif ($key eq 'description') {
+		push @{$metaheaders{$page}}, '<meta name="'.
+			encode_entities($key).
+			'" content="'.encode_entities($value).'" />';
+	}
+	elsif ($key eq 'name') {
+		push @{$metaheaders{$page}}, scrub('<meta '.$key.'="'.
+			encode_entities($value).
+			join(' ', map { "$_=\"$params{$_}\"" } keys %params).
+			' />', $page, $destpage);
 	}
 	else {
-		push @{$meta{$page}}, scrub('<meta name="'.encode_entities($key).
-			'" content="'.encode_entities($value).'" />');
-		if ($key eq 'author') {
-			$author{$page}=$value;
-		}
-		elsif ($key eq 'authorurl') {
-			$authorurl{$page}=$value;
-		}
+		push @{$metaheaders{$page}}, scrub('<meta name="'.
+			encode_entities($key).'" content="'.
+			encode_entities($value).'" />', $page, $destpage);
 	}
 
 	return "";
-} # }}}
+}
 
-sub pagetemplate (@) { #{{{
+sub pagetemplate (@) {
 	my %params=@_;
         my $page=$params{page};
         my $destpage=$params{destpage};
         my $template=$params{template};
 
-	if (exists $meta{$page} && $template->query(name => "meta")) {
+	if (exists $metaheaders{$page} && $template->query(name => "meta")) {
 		# avoid duplicate meta lines
 		my %seen;
-		$template->param(meta => join("\n", grep { (! $seen{$_}) && ($seen{$_}=1) } @{$meta{$page}}));
+		$template->param(meta => join("\n", grep { (! $seen{$_}) && ($seen{$_}=1) } @{$metaheaders{$page}}));
 	}
-	if (exists $title{$page} && $template->query(name => "title")) {
-		$template->param(title => $title{$page});
+	if (exists $pagestate{$page}{meta}{title} && $template->query(name => "title")) {
+		eval q{use HTML::Entities};
+		$template->param(title => HTML::Entities::encode_numeric($pagestate{$page}{meta}{title}));
 		$template->param(title_overridden => 1);
 	}
-	$template->param(permalink => $permalink{$page})
-		if exists $permalink{$page} && $template->query(name => "permalink");
-	$template->param(author => $author{$page})
-		if exists $author{$page} && $template->query(name => "author");
-	$template->param(authorurl => $authorurl{$page})
-		if exists $authorurl{$page} && $template->query(name => "authorurl");
-		
-	if ($page ne $destpage &&
-	    ((exists $license{$page}   && ! exists $license{$destpage}) ||
-	     (exists $copyright{$page} && ! exists $copyright{$destpage}))) {
-		# Force a scan of the destpage to get its copyright/license
-		# info. If the info is declared after an inline, it will
-		# otherwise not be available at this point.
-		IkiWiki::scan($pagesources{$destpage});
-	}
-
-	if (exists $license{$page} && $template->query(name => "license") &&
-	    ($page eq $destpage || ! exists $license{$destpage} ||
-	     $license{$page} ne $license{$destpage})) {
-		$template->param(license => IkiWiki::linkify($page, $destpage, $license{$page}));
-	}
-	if (exists $copyright{$page} && $template->query(name => "copyright") &&
-	    ($page eq $destpage || ! exists $copyright{$destpage} ||
-	     $copyright{$page} ne $copyright{$destpage})) {
-		$template->param(copyright => IkiWiki::linkify($page, $destpage, $copyright{$page}));
-	}
-} # }}}
+
+	foreach my $field (qw{author authorurl}) {
+		$template->param($field => $pagestate{$page}{meta}{$field})
+			if exists $pagestate{$page}{meta}{$field} && $template->query(name => $field);
+	}
+
+	foreach my $field (qw{permalink}) {
+		if (exists $pagestate{$page}{meta}{$field} && $template->query(name => $field)) {
+			eval q{use HTML::Entities};
+			$template->param($field => HTML::Entities::encode_entities(IkiWiki::urlabs($pagestate{$page}{meta}{$field}, $config{url})));
+		}
+	}
+
+	foreach my $field (qw{description}) {
+		eval q{use HTML::Entities};
+		$template->param($field => HTML::Entities::encode_numeric($pagestate{$page}{meta}{$field}))
+			if exists $pagestate{$page}{meta}{$field} && $template->query(name => $field);
+	}
+
+	foreach my $field (qw{license copyright}) {
+		if (exists $pagestate{$page}{meta}{$field} && $template->query(name => $field) &&
+		    ($page eq $destpage || ! exists $pagestate{$destpage}{meta}{$field} ||
+		     $pagestate{$page}{meta}{$field} ne $pagestate{$destpage}{meta}{$field})) {
+			$template->param($field => htmlize($page, $destpage, $pagestate{$page}{meta}{$field}));
+		}
+	}
+}
+
+sub get_sort_key {
+	my $page = shift;
+	my $meta = shift;
+
+	# e.g. titlesort (also makes sense for author)
+	my $key = $pagestate{$page}{meta}{$meta . "sort"};
+	return $key if defined $key;
+
+	# e.g. title
+	$key = $pagestate{$page}{meta}{$meta};
+	return $key if defined $key;
+
+	# fall back to closer-to-core things
+	if ($meta eq 'title') {
+		return pagetitle(IkiWiki::basename($page));
+	}
+	elsif ($meta eq 'date') {
+		return $IkiWiki::pagectime{$page};
+	}
+	elsif ($meta eq 'updated') {
+		return $IkiWiki::pagemtime{$page};
+	}
+	else {
+		return '';
+	}
+}
+
+sub match {
+	my $field=shift;
+	my $page=shift;
+	
+	# turn glob into a safe regexp
+	my $re=IkiWiki::glob2re(shift);
+
+	my $val;
+	if (exists $pagestate{$page}{meta}{$field}) {
+		$val=$pagestate{$page}{meta}{$field};
+	}
+	elsif ($field eq 'title') {
+		$val = pagetitle($page);
+	}
+
+	if (defined $val) {
+		if ($val=~$re) {
+			return IkiWiki::SuccessReason->new("$re matches $field of $page", $page => $IkiWiki::DEPEND_CONTENT, "" => 1);
+		}
+		else {
+			return IkiWiki::FailReason->new("$re does not match $field of $page", $page => $IkiWiki::DEPEND_CONTENT, "" => 1);
+		}
+	}
+	else {
+		return IkiWiki::FailReason->new("$page does not have a $field", $page => $IkiWiki::DEPEND_CONTENT);
+	}
+}
+
+package IkiWiki::PageSpec;
+
+sub match_title ($$;@) {
+	IkiWiki::Plugin::meta::match("title", @_);
+}
+
+sub match_author ($$;@) {
+	IkiWiki::Plugin::meta::match("author", @_);
+}
+
+sub match_authorurl ($$;@) {
+	IkiWiki::Plugin::meta::match("authorurl", @_);
+}
+
+sub match_license ($$;@) {
+	IkiWiki::Plugin::meta::match("license", @_);
+}
+
+sub match_copyright ($$;@) {
+	IkiWiki::Plugin::meta::match("copyright", @_);
+}
+
+sub match_guid ($$;@) {
+	IkiWiki::Plugin::meta::match("guid", @_);
+}
+
+package IkiWiki::SortSpec;
+
+sub cmp_meta {
+	my $meta = shift;
+	error(gettext("sort=meta requires a parameter")) unless defined $meta;
+
+	if ($meta eq 'updated' || $meta eq 'date') {
+		return IkiWiki::Plugin::meta::get_sort_key($a, $meta)
+			<=>
+			IkiWiki::Plugin::meta::get_sort_key($b, $meta);
+	}
+
+	return IkiWiki::Plugin::meta::get_sort_key($a, $meta)
+		cmp
+		IkiWiki::Plugin::meta::get_sort_key($b, $meta);
+}
+
+# A prototype of how sort=title could behave in 4.0 or something
+sub cmp_meta_title {
+	$_[0] = 'title';
+	return cmp_meta(@_);
+}
 
 1