]> git.vanrenterghem.biz Git - git.ikiwiki.info.git/blobdiff - IkiWiki/Plugin/inline.pm
Merge remote branch 'smcv/ready/glob-cache'
[git.ikiwiki.info.git] / IkiWiki / Plugin / inline.pm
index 5d99c9da230c7b9c370f0b1bdac0430ffe746cab..1fe40a5eab4235f82c46cbd0499032c84103e091 100644 (file)
@@ -104,7 +104,7 @@ sub checkconfig () {
 }
 
 sub format (@) {
-        my %params=@_;
+       my %params=@_;
 
        # Fill in the inline content generated earlier. This is actually an
        # optimisation.
@@ -329,12 +329,14 @@ sub preprocess_inline (@) {
        if (! $feedonly) {
                my $template;
                if (! $raw) {
+                       # cannot use wiki pages as templates; template not sanitized due to
+                       # format hook hack
                        eval {
                                $template=template_depends($params{template}.".tmpl", $params{page},
                                        blind_cache => 1);
                        };
-                       if ($@ || ! $template) {
-                               error sprintf(gettext("template %s not found"), $params{template}.".tmpl");
+                       if ($@) {
+                               error sprintf(gettext("failed to process template %s"), $params{template}.".tmpl").": $@";
                        }
                }
                my $needcontent=$raw || (!($archive && $quick) && $template->query(name => 'content'));
@@ -353,10 +355,11 @@ sub preprocess_inline (@) {
                                $template->param(pageurl => urlto($page, $params{destpage}));
                                $template->param(inlinepage => $page);
                                $template->param(title => pagetitle(basename($page)));
-                               $template->param(ctime => displaytime($pagectime{$page}, $params{timeformat}));
+                               $template->param(ctime => displaytime($pagectime{$page}, $params{timeformat}, 1));
                                $template->param(mtime => displaytime($pagemtime{$page}, $params{timeformat}));
                                $template->param(first => 1) if $page eq $list[0];
                                $template->param(last => 1) if $page eq $list[$#list];
+                               $template->param(html5 => $config{html5});
        
                                if ($actions) {
                                        my $file = $pagesources{$page};
@@ -470,6 +473,13 @@ sub get_inline_content ($$) {
                       filter($page, $destpage,
                       readfile(srcfile($file))))));
                $nested--;
+               if (isinternal($page)) {
+                       # make inlined text of internal pages searchable
+                       run_hooks(indexhtml => sub {
+                               shift->(page => $page, destpage => $page,
+                                       content => $ret);
+                       });
+               }
        }
        
        if ($cached_destpage ne $destpage) {
@@ -495,37 +505,60 @@ sub date_822 ($) {
        return $ret;
 }
 
-sub date_3339 ($) {
-       my $time=shift;
-
-       my $lc_time=POSIX::setlocale(&POSIX::LC_TIME);
-       POSIX::setlocale(&POSIX::LC_TIME, "C");
-       my $ret=POSIX::strftime("%Y-%m-%dT%H:%M:%SZ", gmtime($time));
-       POSIX::setlocale(&POSIX::LC_TIME, $lc_time);
-       return $ret;
-}
-
 sub absolute_urls ($$) {
-       # sucky sub because rss sucks
-       my $content=shift;
+       # needed because rss sucks
+       my $html=shift;
        my $baseurl=shift;
 
        my $url=$baseurl;
        $url=~s/[^\/]+$//;
+       my $urltop; # calculated if needed
 
-        # what is the non path part of the url?
-        my $top_uri = URI->new($url);
-        $top_uri->path_query(""); # reset the path
-        my $urltop = $top_uri->as_string;
-
-       $content=~s/(<a(?:\s+(?:class|id)\s*="?\w+"?)?)\s+href=\s*"(#[^"]+)"/$1 href="$baseurl$2"/mig;
-        # relative to another wiki page
-       $content=~s/(<a(?:\s+(?:class|id)\s*="?\w+"?)?)\s+href=\s*"(?!\w+:)([^\/][^"]*)"/$1 href="$url$2"/mig;
-       $content=~s/(<img(?:\s+(?:class|id|width|height)\s*="?\w+"?)*)\s+src=\s*"(?!\w+:)([^\/][^"]*)"/$1 src="$url$2"/mig;
-        # relative to the top of the site
-       $content=~s/(<a(?:\s+(?:class|id)\s*="?\w+"?)?)\s+href=\s*"(?!\w+:)(\/[^"]*)"/$1 href="$urltop$2"/mig;
-       $content=~s/(<img(?:\s+(?:class|id|width|height)\s*="?\w+"?)*)\s+src=\s*"(?!\w+:)(\/[^"]*)"/$1 src="$urltop$2"/mig;
-       return $content;
+       my $ret="";
+
+       eval q{use HTML::Parser; use HTML::Tagset};
+       die $@ if $@;
+       my $p = HTML::Parser->new(api_version => 3);
+       $p->handler(default => sub { $ret.=join("", @_) }, "text");
+       $p->handler(start => sub {
+               my ($tagname, $pos, $text) = @_;
+               if (ref $HTML::Tagset::linkElements{$tagname}) {
+                       while (4 <= @$pos) {
+                               # use attribute sets from right to left
+                               # to avoid invalidating the offsets
+                               # when replacing the values
+                               my ($k_offset, $k_len, $v_offset, $v_len) =
+                                       splice(@$pos, -4);
+                               my $attrname = lc(substr($text, $k_offset, $k_len));
+                               next unless grep { $_ eq $attrname } @{$HTML::Tagset::linkElements{$tagname}};
+                               next unless $v_offset; # 0 v_offset means no value
+                               my $v = substr($text, $v_offset, $v_len);
+                               $v =~ s/^([\'\"])(.*)\1$/$2/;
+                               if ($v=~/^#/) {
+                                       $v=$baseurl.$v; # anchor
+                               }
+                               elsif ($v=~/^(?!\w+:)[^\/]/) {
+                                       $v=$url.$v; # relative url
+                               }
+                               elsif ($v=~/^\//) {
+                                       if (! defined $urltop) {
+                                               # what is the non path part of the url?
+                                               my $top_uri = URI->new($url);
+                                               $top_uri->path_query(""); # reset the path
+                                               $urltop = $top_uri->as_string;
+                                       }
+                                       $v=$urltop.$v; # url relative to top of site
+                               }
+                               $v =~ s/\"/&quot;/g; # since we quote with ""
+                               substr($text, $v_offset, $v_len) = qq("$v");
+                       }
+               }
+               $ret.=$text;
+       }, "tagname, tokenpos, text");
+       $p->parse($html);
+       $p->eof;
+
+       return $ret;
 }
 
 sub genfeed ($$$$$@) {