X-Git-Url: http://git.vanrenterghem.biz/git.ikiwiki.info.git/blobdiff_plain/b31e8c08260b0da95014caea078f3155d4707e28..e9e75f5f29d78b0631f98d9b0fa93184fcb181fe:/IkiWiki/Plugin/meta.pm diff --git a/IkiWiki/Plugin/meta.pm b/IkiWiki/Plugin/meta.pm index 88c942fa4..15bb29b3f 100644 --- a/IkiWiki/Plugin/meta.pm +++ b/IkiWiki/Plugin/meta.pm @@ -4,27 +4,31 @@ 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, + }, +} -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} } @$needsbuild) { # remove state, it will be re-added # if the preprocessor directive is still # there during the rebuild @@ -34,33 +38,37 @@ sub needsbuild (@) { #{{{ } } -sub scrub ($) { #{{{ +sub scrub ($$) { if (IkiWiki::Plugin::htmlscrubber->can("sanitize")) { - return IkiWiki::Plugin::htmlscrubber::sanitize(content => shift); + return IkiWiki::Plugin::htmlscrubber::sanitize( + content => shift, destpage => shift); } else { return shift; } -} #}}} +} + +sub safeurl ($) { + my $url=shift; + if (exists $IkiWiki::Plugin::htmlscrubber::{safe_url_regexp} && + defined $IkiWiki::Plugin::htmlscrubber::safe_url_regexp) { + return $url=~/$IkiWiki::Plugin::htmlscrubber::safe_url_regexp/; + } + else { + return 1; + } +} -sub htmlize ($$$) { #{{{ +sub htmlize ($$$) { my $page = shift; my $destpage = shift; - my $text = shift; - $text=IkiWiki::htmlize($page, pagetype($pagesources{$page}), + return IkiWiki::htmlize($page, $destpage, pagetype($pagesources{$page}), IkiWiki::linkify($page, $destpage, - IkiWiki::preprocess($page, $destpage, $text))); - - # hack to get rid of enclosing junk added by markdown - $text=~s!^

!!; - $text=~s!

$!!; - chomp $text; - - return $text; + IkiWiki::preprocess($page, $destpage, shift))); } -sub preprocess (@) { #{{{ +sub preprocess (@) { return "" unless @_; my %params=@_; my $key=shift; @@ -77,12 +85,41 @@ 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}=HTML::Entities::encode_numeric($value); + # fallthrough + } + elsif ($key eq 'description') { + $pagestate{$page}{meta}{description}=HTML::Entities::encode_numeric($value); + # fallthrough + } + elsif ($key eq 'guid') { + $pagestate{$page}{meta}{guid}=HTML::Entities::encode_numeric($value); + # fallthrough + } + elsif ($key eq 'license') { + push @{$metaheaders{$page}}, ''; + $pagestate{$page}{meta}{license}=$value; + return ""; + } + elsif ($key eq 'copyright') { + push @{$metaheaders{$page}}, ''; + $pagestate{$page}{meta}{copyright}=$value; + return ""; + } + elsif ($key eq 'link' && ! %params) { + # hidden WikiLink + push @{$links{$page}}, $value; + return ""; } - elsif ($key eq 'permalink') { - $permalink{$page}=$value; - push @{$meta{$page}}, scrub(''); + elsif ($key eq 'author') { + $pagestate{$page}{meta}{author}=$value; + # fallthorough + } + elsif ($key eq 'authorurl') { + $pagestate{$page}{meta}{authorurl}=$value if safeurl($value); + # fallthrough } elsif ($key eq 'date') { eval q{use Date::Parse}; @@ -91,6 +128,19 @@ sub preprocess (@) { #{{{ $IkiWiki::pagectime{$page}=$time if defined $time; } } + + if (! defined wantarray) { + # avoid collecting duplicate data during scan pass + return; + } + + # Metadata collection that happens only during preprocessing pass. + if ($key eq 'permalink') { + if (safeurl($value)) { + $pagestate{$page}{meta}{permalink}=$value; + push @{$metaheaders{$page}}, scrub('', $destpage); + } + } elsif ($key eq 'stylesheet') { my $rel=exists $params{rel} ? $params{rel} : "alternate stylesheet"; my $title=exists $params{title} ? $params{title} : $value; @@ -98,28 +148,30 @@ 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{$page}}, '"; } elsif ($key eq 'openid') { - if (exists $params{server}) { - push @{$meta{$page}}, ''; + push @{$metaheaders{$page}}, ''; + } + if (safeurl($value)) { + push @{$metaheaders{$page}}, ''; + push @{$metaheaders{$page}}, ''; + } + if (exists $params{"xrds-location"} && safeurl($params{"xrds-location"})) { + push @{$metaheaders{$page}}, ''; } - push @{$meta{$page}}, ''; - } - elsif ($key eq 'license') { - push @{$meta{$page}}, ''; - $license{$page}=$value; - } - elsif ($key eq 'copyright') { - push @{$meta{$page}}, ''; - $copyright{$page}=$value; } elsif ($key eq 'redir') { return "" if $page ne $destpage; @@ -130,7 +182,7 @@ sub preprocess (@) { #{{{ add_depends($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") } $value=urlto($link, $page); @@ -143,7 +195,7 @@ 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}; @@ -153,70 +205,111 @@ sub preprocess (@) { #{{{ $value=encode_entities($value); } my $delay=int(exists $params{delay} ? $params{delay} : 0); - my $redir=""; + my $redir=""; if (! $safe) { - $redir=scrub($redir); + $redir=scrub($redir, $destpage); } - push @{$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("\n", $destpage); + } + } + elsif ($key eq 'robots') { + push @{$metaheaders{$page}}, ''; } else { - push @{$meta{$page}}, scrub(''); - if ($key eq 'author') { - $author{$page}=$value; - } - elsif ($key eq 'authorurl') { - $authorurl{$page}=$value; - } + push @{$metaheaders{$page}}, scrub('', $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")) { + $template->param(title => $pagestate{$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})) { - $template->param(license => htmlize($page, $destpage, $license{$page})); - } - if (exists $copyright{$page} && $template->query(name => "copyright") && - ($page eq $destpage || ! exists $copyright{$destpage} || - $copyright{$page} ne $copyright{$destpage})) { - $template->param(copyright => htmlize($page, $destpage, $copyright{$page})); - } -} # }}} + + foreach my $field (qw{author authorurl permalink}) { + $template->param($field => $pagestate{$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 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=~/^$re$/i) { + return IkiWiki::SuccessReason->new("$re matches $field of $page"); + } + else { + return IkiWiki::FailReason->new("$re does not match $field of $page"); + } + } + else { + return IkiWiki::FailReason->new("$page does not have a $field"); + } +} + +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", @_); +} 1