don't process data metadata in scan, as it's expensive and the metadata should not...

[git.ikiwiki.info.git] / IkiWiki / Plugin / meta.pm

diff --git a/IkiWiki/Plugin/meta.pm b/IkiWiki/Plugin/meta.pm
index 968e6ccee789de718bac558f6dd30cf58b0acaab..42fddb0bb695855ecbedec0001f8d395e4d66957 100644 (file)
--- a/IkiWiki/Plugin/meta.pm
+++ b/IkiWiki/Plugin/meta.pm
@@ -15,19 +15,24 @@ my %license;
 my %copyright;
 
 sub import { #{{{
 my %copyright;
 
 sub import { #{{{

+       hook(type => "needsbuild", id => "meta", call => \&needsbuild);

        hook(type => "preprocess", id => "meta", call => \&preprocess, scan => 1);
        hook(type => "preprocess", id => "meta", call => \&preprocess, scan => 1);

-       hook(type => "filter", id => "meta", call => \&filter);

        hook(type => "pagetemplate", id => "meta", call => \&pagetemplate);
 } # }}}
 
        hook(type => "pagetemplate", id => "meta", call => \&pagetemplate);
 } # }}}
 

-sub filter (@) { #{{{
-       my %params=@_;
-       
-       $meta{$params{page}}='';
-       delete $pagestate{$params{page}}{meta}{redir};
-
-       return $params{content};
-} # }}}
+sub needsbuild (@) { #{{{
+       my $needsbuild=shift;
+       foreach my $page (keys %pagestate) {
+               if (exists $pagestate{$page}{meta}) {
+                       if (grep { $_ eq $pagesources{$page} } @$needsbuild) {
+                               # remove state, it will be re-added
+                               # if the preprocessor directive is still
+                               # there during the rebuild
+                               delete $pagestate{$page}{meta};
+                       }
+               }
+       }
+}

 
 sub scrub ($) { #{{{
        if (IkiWiki::Plugin::htmlscrubber->can("sanitize")) {
 
 sub scrub ($) { #{{{
        if (IkiWiki::Plugin::htmlscrubber->can("sanitize")) {


@@ -38,10 +43,25 @@ sub scrub ($) { #{{{
        }
 } #}}}
 
        }
 } #}}}
 

+sub htmlize ($$$) { #{{{
+       my $page = shift;
+       my $destpage = shift;
+       my $text = shift;
+
+       $text=IkiWiki::htmlize($page, pagetype($pagesources{$page}),
+               IkiWiki::linkify($page, $destpage,
+               IkiWiki::preprocess($page, $destpage, $text)));
+
+       # hack to get rid of enclosing junk added by markdown
+       $text=~s!^<p>!!;
+       $text=~s!</p>$!!;
+       chomp $text;
+
+       return $text;
+}
+

 sub preprocess (@) { #{{{
 sub preprocess (@) { #{{{

-       if (! @_) {
-               return "";
-       }
+       return "" unless @_;

        my %params=@_;
        my $key=shift;
        my $value=$params{$key};
        my %params=@_;
        my $key=shift;
        my $value=$params{$key};


@@ -53,63 +73,55 @@ sub preprocess (@) { #{{{
        delete $params{preview};
 
        eval q{use HTML::Entities};
        delete $params{preview};
 
        eval q{use HTML::Entities};

-       # Always dencode, even if encoding later, since it might not be
+       # Always decode, even if encoding later, since it might not be

        # fully encoded.
        $value=decode_entities($value);
 
        # fully encoded.
        $value=decode_entities($value);
 

-       if ($key eq 'link') {
-               if (%params) {
-                       $meta{$page}.=scrub("<link href=\"".encode_entities($value)."\" ".
-                               join(" ", map {
-                                       encode_entities($_)."=\"".encode_entities(decode_entities($params{$_}))."\""
-                               } keys %params).
-                               " />\n");
-               }
-               else {
-                       # hidden WikiLink
-                       push @{$links{$page}}, $value;
-               }
+       # Metadata collection that needs to happen during the scan pass.
+       if ($key eq 'title') {
+               $title{$page}=HTML::Entities::encode_numeric($value);

        }
        }

-       elsif ($key eq 'redir') {
-               my $safe=0;
-               if ($value !~ /^\w+:\/\//) {
-                       add_depends($page, $value);
-                       my $link=bestlink($page, $value);
-                       if (! length $link) {
-                               return "[[meta ".gettext("redir page not found")."]]";
-                       }
-                       $pagestate{$page}{meta}{redir}=$link;
-                       if ($pagestate{$link}{meta}{redir}) {
-                               # TODO: real cycle detection
-                               return "[[meta ".gettext("redir not allowed to point to a page that contains a redir")."]]";
-                       }
-                       $value=urlto($link, $destpage);
-                       $safe=1;
-               }
-               else {
-                       $value=encode_entities($value);
-               }
-               my $delay=int(exists $params{delay} ? $params{delay} : 0);
-               my $redir="<meta http-equiv=\"refresh\" content=\"$delay; URL=$value\">";
-               if (! $safe) {
-                       $redir=scrub($redir);
-               }
-               $meta{$page}.=$redir;
+       elsif ($key eq 'license') {
+               push @{$meta{$page}}, '<link rel="license" href="#page_license" />';
+               $license{$page}=$value;
+               return "";

        }
        }

-       elsif ($key eq 'title') {
-               $title{$page}=HTML::Entities::encode_numeric($value);
+       elsif ($key eq 'copyright') {
+               push @{$meta{$page}}, '<link rel="copyright" href="#page_copyright" />';
+               $copyright{$page}=$value;
+               return "";

        }
        }

-       elsif ($key eq 'permalink') {
-               $permalink{$page}=$value;
-               $meta{$page}.=scrub("<link rel=\"bookmark\" href=\"".encode_entities($value)."\" />\n");
+       elsif ($key eq 'link' && ! %params) {
+               # hidden WikiLink
+               push @{$links{$page}}, $value;
+               return "";
+       }
+       elsif ($key eq 'author') {
+               $author{$page}=$value;
+               # fallthorough
+       }
+       elsif ($key eq 'authorurl') {
+               $authorurl{$page}=$value;
+               # fallthrough
+       }
+
+       if (! defined wantarray) {
+               # avoid collecting duplicate data during scan pass
+               return;

        }
        }

-       elsif ($key eq 'date') {
+
+       # Metadata collection that happens only during preprocessing pass.
+       if ($key eq 'date') {

                eval q{use Date::Parse};
                if (! $@) {
                        my $time = str2time($value);
                        $IkiWiki::pagectime{$page}=$time if defined $time;
                }
        }
                eval q{use Date::Parse};
                if (! $@) {
                        my $time = str2time($value);
                        $IkiWiki::pagectime{$page}=$time if defined $time;
                }
        }

+       elsif ($key eq 'permalink') {
+               $permalink{$page}=$value;
+               push @{$meta{$page}}, scrub('<link rel="bookmark" href="'.encode_entities($value).'" />');
+       }

        elsif ($key eq 'stylesheet') {
                my $rel=exists $params{rel} ? $params{rel} : "alternate stylesheet";
                my $title=exists $params{title} ? $params{title} : $value;
        elsif ($key eq 'stylesheet') {
                my $rel=exists $params{rel} ? $params{rel} : "alternate stylesheet";
                my $title=exists $params{title} ? $params{title} : $value;


@@ -119,36 +131,69 @@ sub preprocess (@) { #{{{
                if (! length $stylesheet) {
                        return "[[meta ".gettext("stylesheet not found")."]]";
                }
                if (! length $stylesheet) {
                        return "[[meta ".gettext("stylesheet not found")."]]";
                }

-               $meta{$page}.='<link href="'.urlto($stylesheet, $page).
+               push @{$meta{$page}}, '<link href="'.urlto($stylesheet, $page).

                        '" rel="'.encode_entities($rel).
                        '" title="'.encode_entities($title).
                        '" rel="'.encode_entities($rel).
                        '" title="'.encode_entities($title).

-                       "\" type=\"text/css\" />\n";
+                       "\" type=\"text/css\" />";

        }
        elsif ($key eq 'openid') {
                if (exists $params{server}) {
        }
        elsif ($key eq 'openid') {
                if (exists $params{server}) {

-                       $meta{$page}.='<link href="'.encode_entities($params{server}).
-                               "\" rel=\"openid.server\" />\n";
+                       push @{$meta{$page}}, '<link href="'.encode_entities($params{server}).
+                               '" rel="openid.server" />';

                }
                }

-               $meta{$page}.='<link href="'.encode_entities($value).
-                       "\" rel=\"openid.delegate\" />\n";
+               push @{$meta{$page}}, '<link href="'.encode_entities($value).
+                       '" rel="openid.delegate" />';

        }
        }

-       elsif ($key eq 'license') {
-               $meta{$page}.="<link rel=\"license\" href=\"#page_license\" />\n";
-               $license{$page}=$value;
+       elsif ($key eq 'redir') {
+               return "" if $page ne $destpage;
+               my $safe=0;
+               if ($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")."]]";
+                       }
+
+                       $value=urlto($link, $page);
+                       $value.='#'.$redir_anchor if defined $redir_anchor;
+                       $safe=1;
+
+                       # redir cycle detection
+                       $pagestate{$page}{meta}{redir}=$link;
+                       my $at=$page;
+                       my %seen;
+                       while (exists $pagestate{$at}{meta}{redir}) {
+                               if ($seen{$at}) {
+                                       return "[[meta ".gettext("redir cycle is not allowed")."]]";
+                               }
+                               $seen{$at}=1;
+                               $at=$pagestate{$at}{meta}{redir};
+                       }
+               }
+               else {
+                       $value=encode_entities($value);
+               }
+               my $delay=int(exists $params{delay} ? $params{delay} : 0);
+               my $redir="<meta http-equiv=\"refresh\" content=\"$delay; URL=$value\">";
+               if (! $safe) {
+                       $redir=scrub($redir);
+               }
+               push @{$meta{$page}}, $redir;

        }
        }

-       elsif ($key eq 'copyright') {
-               $meta{$page}.="<link rel=\"copyright\" href=\"#page_copyright\" />\n";
-               $copyright{$page}=$value;
+       elsif ($key eq 'link') {
+               if (%params) {
+                       $meta{$page}.=scrub("<link href=\"".encode_entities($value)."\" ".
+                               join(" ", map {
+                                       encode_entities($_)."=\"".encode_entities(decode_entities($params{$_}))."\""
+                               } keys %params).
+                               " />\n");
+               }

        }
        else {
        }
        else {

-               $meta{$page}.=scrub("<meta name=\"".encode_entities($key).
-                       "\" content=\"".encode_entities($value)."\" />\n");
-               if ($key eq 'author') {
-                       $author{$page}=$value;
-               }
-               elsif ($key eq 'authorurl') {
-                       $authorurl{$page}=$value;
-               }
+               push @{$meta{$page}}, scrub('<meta name="'.encode_entities($key).
+                       '" content="'.encode_entities($value).'" />');

        }
 
        return "";
        }
 
        return "";


@@ -160,8 +205,11 @@ sub pagetemplate (@) { #{{{
         my $destpage=$params{destpage};
         my $template=$params{template};
 
         my $destpage=$params{destpage};
         my $template=$params{template};
 

-       $template->param(meta => $meta{$page})
-               if exists $meta{$page} && $template->query(name => "meta");
+       if (exists $meta{$page} && $template->query(name => "meta")) {
+               # avoid duplicate meta lines
+               my %seen;
+               $template->param(meta => join("\n", grep { (! $seen{$_}) && ($seen{$_}=1) } @{$meta{$page}}));
+       }

        if (exists $title{$page} && $template->query(name => "title")) {
                $template->param(title => $title{$page});
                $template->param(title_overridden => 1);
        if (exists $title{$page} && $template->query(name => "title")) {
                $template->param(title => $title{$page});
                $template->param(title_overridden => 1);


@@ -172,25 +220,16 @@ sub pagetemplate (@) { #{{{
                if exists $author{$page} && $template->query(name => "author");
        $template->param(authorurl => $authorurl{$page})
                if exists $authorurl{$page} && $template->query(name => "authorurl");
                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})) {
 
        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}));
+               $template->param(license => htmlize($page, $destpage, $license{$page}));

        }
        if (exists $copyright{$page} && $template->query(name => "copyright") &&
            ($page eq $destpage || ! exists $copyright{$destpage} ||
             $copyright{$page} ne $copyright{$destpage})) {
        }
        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}));
+               $template->param(copyright => htmlize($page, $destpage, $copyright{$page}));

        }
 } # }}}
 
        }
 } # }}}