htmlscrubber: Do not scrub url anchors that contain colons.

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

diff --git a/IkiWiki/Plugin/aggregate.pm b/IkiWiki/Plugin/aggregate.pm
index 9b531b6128c1de09d81e9c1b60da25ab736e2732..7789c4c2a9d38b7e616d9d3b4df8e4d96fe74ec6 100644 (file)
--- a/IkiWiki/Plugin/aggregate.pm
+++ b/IkiWiki/Plugin/aggregate.pm
@@ -4,7 +4,7 @@ package IkiWiki::Plugin::aggregate;
 
 use warnings;
 use strict;
-use IkiWiki 2.00;
+use IkiWiki 3.00;
 use HTML::Parser;
 use HTML::Tagset;
 use HTML::Entities;
@@ -46,7 +46,7 @@ sub getsetup () {
                },
                aggregateinternal => {
                        type => "boolean",
-                       example => 0,
+                       example => 1,
                        description => "enable aggregation to internal pages?",
                        safe => 0, # enabling needs manual transition
                        rebuild => 0,
@@ -61,6 +61,10 @@ sub getsetup () {
 }
 
 sub checkconfig () {
+       if (! defined $config{aggregateinternal}) {
+               $config{aggregateinternal}=1;
+       }
+
        if ($config{aggregate} && ! ($config{post_commit} && 
                                     IkiWiki::commit_hook_enabled())) {
                launchaggregation();
@@ -294,7 +298,7 @@ sub loadstate () {
        return if $state_loaded;
        $state_loaded=1;
        if (-e "$config{wikistatedir}/aggregate") {
-               open(IN, "$config{wikistatedir}/aggregate") ||
+               open(IN, "<", "$config{wikistatedir}/aggregate") ||
                        die "$config{wikistatedir}/aggregate: $!";
                while (<IN>) {
                        $_=IkiWiki::possibly_foolish_untaint($_);
@@ -331,7 +335,7 @@ sub savestate () {
        garbage_collect();
        my $newfile="$config{wikistatedir}/aggregate.new";
        my $cleanup = sub { unlink($newfile) };
-       open (OUT, ">$newfile") || error("open $newfile: $!", $cleanup);
+       open (OUT, ">", $newfile) || error("open $newfile: $!", $cleanup);
        foreach my $data (values %feeds, values %guids) {
                my @line;
                foreach my $field (keys %$data) {
@@ -343,7 +347,8 @@ sub savestate () {
                                push @line, "tag=$_" foreach @{$data->{tags}};
                        }
                        else {
-                               push @line, "$field=".$data->{$field};
+                               push @line, "$field=".$data->{$field}
+                                       if defined $data->{$field};
                        }
                }
                print OUT join(" ", @line)."\n" || error("write $newfile: $!", $cleanup);
@@ -351,6 +356,20 @@ sub savestate () {
        close OUT || error("save $newfile: $!", $cleanup);
        rename($newfile, "$config{wikistatedir}/aggregate") ||
                error("rename $newfile: $!", $cleanup);
+
+       my $timestamp=undef;
+       foreach my $feed (keys %feeds) {
+               my $t=$feeds{$feed}->{lastupdate}+$feeds{$feed}->{updateinterval};
+               if (! defined $timestamp || $timestamp > $t) {
+                       $timestamp=$t;
+               }
+       }
+       $newfile=~s/\.new$/time/;
+       open (OUT, ">", $newfile) || error("open $newfile: $!", $cleanup);
+       if (defined $timestamp) {
+               print OUT $timestamp."\n";
+       }
+       close OUT || error("save $newfile: $!", $cleanup);
 }
 
 sub garbage_collect () {
@@ -399,6 +418,7 @@ sub mergestate () {
        }
 
        # New guids can be created during aggregation.
+       # Guids have a few fields that may be updated during aggregation.
        # It's also possible that guids were removed from the on-disk state
        # while the aggregation was in process. That would only happen if
        # their feed was also removed, so any removed guids added back here
@@ -407,6 +427,11 @@ sub mergestate () {
                if (! exists $guids{$guid}) {
                        $guids{$guid}=$myguids{$guid};
                }
+               else {
+                       foreach my $field (qw{md5}) {
+                               $guids{$guid}->{$field}=$myguids{$guid}->{$field};
+                       }
+               }
        }
 }
 
@@ -529,6 +554,11 @@ sub aggregate (@) {
                }
 
                foreach my $entry ($f->entries) {
+                       # XML::Feed doesn't work around XML::Atom's bizarre
+                       # API, so we will. Real unicode strings? Yes please.
+                       # See [[bugs/Aggregated_Atom_feeds_are_double-encoded]]
+                       local $XML::Atom::ForceUnicode = 1;
+
                        my $c=$entry->content;
                        # atom feeds may have no content, only a summary
                        if (! defined $c && ref $entry->summary) {
@@ -634,11 +664,13 @@ sub add_page (@) {
                # creation time on record for the new page.
                utime $mtime, $mtime, "$config{srcdir}/".htmlfn($guid->{page});
                # Store it in pagectime for expiry code to use also.
-               $IkiWiki::pagectime{$guid->{page}}=$mtime;
+               $IkiWiki::pagectime{$guid->{page}}=$mtime
+                       unless exists $IkiWiki::pagectime{$guid->{page}};
        }
        else {
                # Dummy value for expiry code.
-               $IkiWiki::pagectime{$guid->{page}}=time;
+               $IkiWiki::pagectime{$guid->{page}}=time
+                       unless exists $IkiWiki::pagectime{$guid->{page}};
        }
 }