my %feeds;
my %guids;
-sub import { #{{{
+sub import {
hook(type => "getopt", id => "aggregate", call => \&getopt);
hook(type => "getsetup", id => "aggregate", call => \&getsetup);
hook(type => "checkconfig", id => "aggregate", call => \&checkconfig);
if (exists $config{aggregate_webtrigger} && $config{aggregate_webtrigger}) {
hook(type => "cgi", id => "aggregate", call => \&cgi);
}
-} # }}}
+}
-sub getopt () { #{{{
+sub getopt () {
eval q{use Getopt::Long};
error($@) if $@;
Getopt::Long::Configure('pass_through');
"aggregate" => \$config{aggregate},
"aggregateinternal!" => \$config{aggregateinternal},
);
-} #}}}
+}
-sub getsetup () { #{{{
+sub getsetup () {
return
plugin => {
safe => 1,
},
aggregateinternal => {
type => "boolean",
- example => 0,
+ example => 1,
description => "enable aggregation to internal pages?",
safe => 0, # enabling needs manual transition
rebuild => 0,
safe => 1,
rebuild => 0,
},
-} #}}}
+}
+
+sub checkconfig () {
+ if (! defined $config{aggregateinternal}) {
+ $config{aggregateinternal}=1;
+ }
-sub checkconfig () { #{{{
if ($config{aggregate} && ! ($config{post_commit} &&
IkiWiki::commit_hook_enabled())) {
launchaggregation();
}
-} #}}}
+}
-sub cgi ($) { #{{{
+sub cgi ($) {
my $cgi=shift;
if (defined $cgi->param('do') &&
}
exit 0;
}
-} #}}}
+}
-sub launchaggregation () { #{{{
+sub launchaggregation () {
# See if any feeds need aggregation.
loadstate();
my @feeds=needsaggregate();
unlockaggregate();
return 1;
-} #}}}
+}
# Pages with extension _aggregated have plain html markup, pass through.
-sub htmlize (@) { #{{{
+sub htmlize (@) {
my %params=@_;
return $params{content};
-} #}}}
+}
# Used by ikiwiki-transition aggregateinternal.
-sub migrate_to_internal { #{{{
+sub migrate_to_internal {
if (! lockaggregate()) {
error("an aggregation process is currently running");
}
IkiWiki::unlockwiki;
unlockaggregate();
-} #}}}
+}
-sub needsbuild (@) { #{{{
+sub needsbuild (@) {
my $needsbuild=shift;
loadstate();
markunseen($feed->{sourcepage});
}
}
-} # }}}
+}
-sub preprocess (@) { #{{{
+sub preprocess (@) {
my %params=@_;
foreach my $required (qw{name url}) {
$feed->{template}=$params{template} . ".tmpl";
delete $feed->{unseen};
$feed->{lastupdate}=0 unless defined $feed->{lastupdate};
+ $feed->{lasttry}=$feed->{lastupdate} unless defined $feed->{lasttry};
$feed->{numposts}=0 unless defined $feed->{numposts};
$feed->{newposts}=0 unless defined $feed->{newposts};
$feed->{message}=gettext("new feed") unless defined $feed->{message};
($feed->{newposts} ? "; ".$feed->{newposts}.
" ".gettext("new") : "").
")";
-} # }}}
+}
-sub delete (@) { #{{{
+sub delete (@) {
my @files=@_;
# Remove feed data for removed pages.
my $page=pagename($file);
markunseen($page);
}
-} #}}}
+}
-sub markunseen ($) { #{{{
+sub markunseen ($) {
my $page=shift;
foreach my $id (keys %feeds) {
$feeds{$id}->{unseen}=1;
}
}
-} #}}}
+}
my $state_loaded=0;
-sub loadstate () { #{{{
+sub loadstate () {
return if $state_loaded;
$state_loaded=1;
if (-e "$config{wikistatedir}/aggregate") {
close IN;
}
-} #}}}
+}
-sub savestate () { #{{{
+sub savestate () {
return unless $state_loaded;
garbage_collect();
my $newfile="$config{wikistatedir}/aggregate.new";
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);
close OUT || error("save $newfile: $!", $cleanup);
rename($newfile, "$config{wikistatedir}/aggregate") ||
error("rename $newfile: $!", $cleanup);
-} #}}}
+}
-sub garbage_collect () { #{{{
+sub garbage_collect () {
foreach my $name (keys %feeds) {
# remove any feeds that were not seen while building the pages
# that used to contain them
delete $guid->{md5};
}
}
-} #}}}
+}
-sub mergestate () { #{{{
+sub mergestate () {
# Load the current state in from disk, and merge into it
# values from the state in memory that might have changed
# during aggregation.
# fields.
foreach my $name (keys %myfeeds) {
if (exists $feeds{$name}) {
- foreach my $field (qw{message lastupdate numposts
- newposts error}) {
+ foreach my $field (qw{message lastupdate lasttry
+ numposts newposts error}) {
$feeds{$name}->{$field}=$myfeeds{$name}->{$field};
}
}
$guids{$guid}=$myguids{$guid};
}
}
-} #}}}
+}
-sub clearstate () { #{{{
+sub clearstate () {
%feeds=();
%guids=();
$state_loaded=0;
-} #}}}
+}
-sub expire () { #{{{
+sub expire () {
foreach my $feed (values %feeds) {
next unless $feed->{expireage} || $feed->{expirecount};
my $count=0;
}
}
}
-} #}}}
+}
-sub needsaggregate () { #{{{
+sub needsaggregate () {
return values %feeds if $config{rebuild};
return grep { time - $_->{lastupdate} >= $_->{updateinterval} } values %feeds;
-} #}}}
+}
-sub aggregate (@) { #{{{
+sub aggregate (@) {
eval q{use XML::Feed};
error($@) if $@;
eval q{use URI::Fetch};
error($@) if $@;
foreach my $feed (@_) {
- $feed->{lastupdate}=time;
+ $feed->{lasttry}=time;
$feed->{newposts}=0;
$feed->{message}=sprintf(gettext("last checked %s"),
- displaytime($feed->{lastupdate}));
+ displaytime($feed->{lasttry}));
$feed->{error}=0;
debug(sprintf(gettext("checking feed %s ..."), $feed->{name}));
debug($feed->{message});
next;
}
+
+ # lastupdate is only set if we were able to contact the server
+ $feed->{lastupdate}=$feed->{lasttry};
+
if ($res->status == URI::Fetch::URI_GONE()) {
$feed->{message}=gettext("feed not found");
$feed->{error}=1;
# that contains invalid UTF-8 sequences. Convert
# feed to ascii to try to work around.
$feed->{message}.=" ".sprintf(gettext("(invalid UTF-8 stripped from feed)"));
- $content=Encode::decode_utf8($content, 0);
- $f=eval{XML::Feed->parse(\$content)};
+ $f=eval {
+ $content=Encode::decode_utf8($content, 0);
+ XML::Feed->parse(\$content)
+ };
}
if ($@) {
# Another possibility is badly escaped entities.
$feed->{message}.=" ".sprintf(gettext("(feed entities escaped)"));
$content=~s/\&(?!amp)(\w+);/&$1;/g;
- $content=Encode::decode_utf8($content, 0);
- $f=eval{XML::Feed->parse(\$content)};
+ $f=eval {
+ $content=Encode::decode_utf8($content, 0);
+ XML::Feed->parse(\$content)
+ };
}
if ($@) {
$feed->{message}=gettext("feed crashed XML::Feed!")." ($@)";
}
foreach my $entry ($f->entries) {
- my $content=$content=$entry->content->body;
+ my $c=$entry->content;
# atom feeds may have no content, only a summary
- if (! defined $content && ref $entry->summary) {
- $content=$entry->summary->body;
+ if (! defined $c && ref $entry->summary) {
+ $c=$entry->summary;
}
add_page(
copyright => $f->copyright,
title => defined $entry->title ? decode_entities($entry->title) : "untitled",
link => $entry->link,
- content => defined $content ? $content : "",
+ content => (defined $c && defined $c->body) ? $c->body : "",
guid => defined $entry->id ? $entry->id : time."_".$feed->{name},
ctime => $entry->issued ? ($entry->issued->epoch || time) : time,
+ base => (defined $c && $c->can("base")) ? $c->base : undef,
);
}
}
-} #}}}
+}
-sub add_page (@) { #{{{
+sub add_page (@) {
my %params=@_;
my $feed=$params{feed};
my $template=template($feed->{template}, blind_cache => 1);
$template->param(title => $params{title})
if defined $params{title} && length($params{title});
- $template->param(content => htmlescape(htmlabs($params{content}, $feed->{feedurl})));
+ $template->param(content => wikiescape(htmlabs($params{content},
+ defined $params{base} ? $params{base} : $feed->{feedurl})));
$template->param(name => $feed->{name});
$template->param(url => $feed->{url});
$template->param(copyright => $params{copyright})
# Dummy value for expiry code.
$IkiWiki::pagectime{$guid->{page}}=time;
}
-} #}}}
+}
-sub htmlescape ($) { #{{{
+sub wikiescape ($) {
# escape accidental wikilinks and preprocessor stuff
- my $html=shift;
- $html=~s/(?<!\\)\[\[/\\\[\[/g;
- return $html;
-} #}}}
+ return encode_entities(shift, '\[\]');
+}
-sub urlabs ($$) { #{{{
+sub urlabs ($$) {
my $url=shift;
my $urlbase=shift;
URI->new_abs($url, $urlbase)->as_string;
-} #}}}
+}
-sub htmlabs ($$) { #{{{
+sub htmlabs ($$) {
# Convert links in html from relative to absolute.
# Note that this is a heuristic, which is not specified by the rss
# spec and may not be right for all feeds. Also, see Debian
$p->eof;
return $ret;
-} #}}}
+}
-sub htmlfn ($) { #{{{
+sub htmlfn ($) {
return shift().".".($config{aggregateinternal} ? "_aggregated" : $config{htmlext});
-} #}}}
+}
my $aggregatelock;
-sub lockaggregate () { #{{{
+sub lockaggregate () {
# Take an exclusive lock to prevent multiple concurrent aggregators.
# Returns true if the lock was aquired.
if (! -d $config{wikistatedir}) {
return 0;
}
return 1;
-} #}}}
+}
-sub unlockaggregate () { #{{{
+sub unlockaggregate () {
return close($aggregatelock) if $aggregatelock;
return;
-} #}}}
+}
1