X-Git-Url: http://git.vanrenterghem.biz/git.ikiwiki.info.git/blobdiff_plain/deb39a1f14a840761f32011f28fe698015d8b298..ea2ab6f885b97081c37f9034d8b7e12a0e34ad3e:/IkiWiki/Plugin/meta.pm diff --git a/IkiWiki/Plugin/meta.pm b/IkiWiki/Plugin/meta.pm index 586dbcb81..e7b96bdf1 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, 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 @@ -32,35 +37,40 @@ sub needsbuild (@) { #{{{ } } } + return $needsbuild; } -sub scrub ($) { #{{{ +sub scrub ($$$) { if (IkiWiki::Plugin::htmlscrubber->can("sanitize")) { - return IkiWiki::Plugin::htmlscrubber::sanitize(content => shift); + return IkiWiki::Plugin::htmlscrubber::sanitize( + content => shift, page => shift, destpage => shift); } else { return shift; } -} #}}} +} + +sub safeurl ($) { + my $url=shift; + if (exists $IkiWiki::Plugin::htmlscrubber::{safe_url_regexp} && + defined $IkiWiki::Plugin::htmlscrubber::safe_url_regexp) { + return $url=~/$IkiWiki::Plugin::htmlscrubber::safe_url_regexp/; + } + else { + return 1; + } +} -sub htmlize ($$$) { #{{{ +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 $key=shift; @@ -79,48 +89,93 @@ sub preprocess (@) { #{{{ # Metadata collection that needs to happen during the scan pass. if ($key eq 'title') { - $title{$page}=HTML::Entities::encode_numeric($value); - } - elsif ($key eq 'date') { - eval q{use Date::Parse}; - if (! $@) { - my $time = str2time($value); - $IkiWiki::pagectime{$page}=$time if defined $time; + $pagestate{$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 @{$meta{$page}}, ''; - $license{$page}=$value; + push @{$metaheaders{$page}}, ''; + $pagestate{$page}{meta}{license}=$value; return ""; } elsif ($key eq 'copyright') { - push @{$meta{$page}}, ''; - $copyright{$page}=$value; + push @{$metaheaders{$page}}, ''; + $pagestate{$page}{meta}{copyright}=$value; return ""; } elsif ($key eq 'link' && ! %params) { # hidden WikiLink - push @{$links{$page}}, $value; + add_link($page, $value); return ""; } + elsif ($key eq 'enclosure') { + my $link=bestlink($page, $value); + if (! length $link) { + error gettext("enclosure not found") + } + add_depends($page, $link, deptype("presence")); + + $value=urlto($link, $page, 1); + $pagestate{$page}{meta}{enclosure}=$value; + $pagestate{$page}{meta}{enclosurefile}=$link; + # fallthrough + } elsif ($key eq 'author') { - $author{$page}=$value; + $pagestate{$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') { - $authorurl{$page}=$value; + $pagestate{$page}{meta}{authorurl}=$value if safeurl($value); # fallthrough } + elsif ($key eq 'permalink') { + $pagestate{$page}{meta}{permalink}=$value if safeurl($value); + # fallthrough + } + elsif ($key eq 'date') { + eval q{use Date::Parse}; + if (! $@) { + my $time = str2time($value); + $IkiWiki::pagectime{$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 $time; + } + } if (! defined wantarray) { # avoid collecting duplicate data during scan pass return; } - # Metadata collection that happens only during preprocessing pass. + # Metadata handling that happens only during preprocessing pass. if ($key eq 'permalink') { - $permalink{$page}=$value; - push @{$meta{$page}}, scrub(''); + if (safeurl($value)) { + push @{$metaheaders{$page}}, scrub('', $page, $destpage); + } } elsif ($key eq 'stylesheet') { my $rel=exists $params{rel} ? $params{rel} : "alternate stylesheet"; @@ -129,20 +184,57 @@ 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}}, '"; + "\" type=\"text/css\" />", $page, $destpage); + } + elsif ($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 $js) { + error gettext("script not found"); + } + push @{$metaheaders{$page}}, scrub('', + $page, $destpage); } elsif ($key eq 'openid') { - if (exists $params{server}) { - push @{$meta{$page}}, ''; + 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}}, '' if $delegate ne 2; + push @{$metaheaders{$page}}, '' if $delegate ne 1; + } + if (safeurl($value)) { + push @{$metaheaders{$page}}, '' if $delegate ne 2; + push @{$metaheaders{$page}}, '' 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{$page}}, ''; + } + } + elsif ($key eq 'foaf') { + if (safeurl($value)) { + push @{$metaheaders{$page}}, ''; } - push @{$meta{$page}}, ''; } elsif ($key eq 'redir') { return "" if $page ne $destpage; @@ -150,11 +242,11 @@ sub preprocess (@) { #{{{ if ($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; @@ -166,7 +258,7 @@ 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}; @@ -176,61 +268,202 @@ 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, $page, $destpage); } - push @{$meta{$page}}, $redir; + push @{$metaheaders{$page}}, $redir; } elsif ($key eq 'link') { if (%params) { - $meta{$page}.=scrub("\n"); + " />\n", $page, $destpage); } } + elsif ($key eq 'robots') { + push @{$metaheaders{$page}}, ''; + } + elsif ($key eq 'description' || $key eq 'author') { + push @{$metaheaders{$page}}, ''; + } + elsif ($key eq 'name') { + push @{$metaheaders{$page}}, scrub('', $page, $destpage); + } + elsif ($key eq 'keywords') { + # Make sure the keyword string is safe: only allow alphanumeric + # characters, space and comma and strip the rest. + $value =~ s/[^[:alnum:], ]+//g; + push @{$metaheaders{$page}}, ''; + } else { - push @{$meta{$page}}, scrub(''); + push @{$metaheaders{$page}}, scrub('', $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 (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 $pagestate{$page}{meta}{enclosure}) { + $template->param(enclosure => HTML::Entities::encode_entities(IkiWiki::urlabs($pagestate{$page}{meta}{enclosure}, $config{url}))); + } + + foreach my $field (qw{authorurl}) { + eval q{use HTML::Entities}; + $template->param($field => HTML::Entities::encode_entities($pagestate{$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 author}) { + 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); } - 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})); +} + +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