use HTML::Parser;
use HTML::Tagset;
use HTML::Entities;
-use URI;
use open qw{:utf8 :std};
my %feeds;
sub import {
hook(type => "getopt", id => "aggregate", call => \&getopt);
hook(type => "getsetup", id => "aggregate", call => \&getsetup);
- hook(type => "checkconfig", id => "aggregate", call => \&checkconfig);
+ hook(type => "checkconfig", id => "aggregate", call => \&checkconfig,
+ last => 1);
hook(type => "needsbuild", id => "aggregate", call => \&needsbuild);
hook(type => "preprocess", id => "aggregate", call => \&preprocess);
hook(type => "delete", id => "aggregate", call => \&delete);
$config{aggregateinternal}=1;
}
+ # This is done here rather than in a refresh hook because it
+ # needs to run before the wiki is locked.
if ($config{aggregate} && ! ($config{post_commit} &&
IkiWiki::commit_hook_enabled())) {
launchaggregation();
my @feeds=needsaggregate();
return unless @feeds;
if (! lockaggregate()) {
- debug("an aggregation process is already running");
- return;
+ error("an aggregation process is already running");
}
# force a later rebuild of source pages
$IkiWiki::forcerebuild{$_->{sourcepage}}=1
$config{aggregateinternal} = 0;
my $oldname = "$config{srcdir}/".htmlfn($data->{page});
+ if (! -e $oldname) {
+ $oldname = $IkiWiki::Plugin::transient::transientdir."/".htmlfn($data->{page});
+ }
+
my $oldoutput = $config{destdir}."/".IkiWiki::htmlpage($data->{page});
$config{aggregateinternal} = 1;
- my $newname = "$config{srcdir}/".htmlfn($data->{page});
+ my $newname = $IkiWiki::Plugin::transient::transientdir."/".htmlfn($data->{page});
debug "moving $oldname -> $newname";
if (-e $newname) {
if (-e $oldoutput) {
require IkiWiki::Render;
debug("removing output file $oldoutput");
- IkiWiki::prune($oldoutput);
+ IkiWiki::prune($oldoutput, $config{destdir});
}
}
foreach my $guid (values %guids) {
# any guid whose feed is gone should be removed
if (! exists $feeds{$guid->{feed}}) {
- unlink "$config{srcdir}/".htmlfn($guid->{page})
- if exists $guid->{page};
+ if (exists $guid->{page}) {
+ unlink $IkiWiki::Plugin::transient::transientdir."/".htmlfn($guid->{page})
+ || unlink "$config{srcdir}/".htmlfn($guid->{page});
+ }
delete $guids{$guid->{guid}};
}
# handle expired guids
elsif ($guid->{expired} && exists $guid->{page}) {
unlink "$config{srcdir}/".htmlfn($guid->{page});
+ unlink $IkiWiki::Plugin::transient::transientdir."/".htmlfn($guid->{page});
delete $guid->{page};
delete $guid->{md5};
}
}
sub aggregate (@) {
+ eval q{use Net::INET6Glue::INET_is_INET6}; # may not be available
eval q{use XML::Feed};
error($@) if $@;
eval q{use URI::Fetch};
}
$feed->{feedurl}=pop @urls;
}
- my $res=URI::Fetch->fetch($feed->{feedurl});
+ my $ua=useragent();
+ my $res=URI::Fetch->fetch($feed->{feedurl}, UserAgent=>$ua);
if (! $res) {
$feed->{message}=URI::Fetch->errstr;
$feed->{error}=1;
};
}
if ($@) {
- $feed->{message}=gettext("feed crashed XML::Feed!")." ($@)";
+ # gettext can clobber $@
+ my $error = $@;
+ $feed->{message}=gettext("feed crashed XML::Feed!")." ($error)";
$feed->{error}=1;
debug($feed->{message});
next;
feed => $feed,
copyright => $f->copyright,
title => defined $entry->title ? decode_entities($entry->title) : "untitled",
+ author => defined $entry->author ? decode_entities($entry->author) : "",
link => $entry->link,
content => (defined $c && defined $c->body) ? $c->body : "",
guid => defined $entry->id ? $entry->id : time."_".$feed->{name},
# updating an existing post
$guid=$guids{$params{guid}};
return if $guid->{expired};
+ write_page($feed, $guid, $mtime, \%params);
}
else {
# new post
}
my $c="";
while (exists $IkiWiki::pagecase{lc $page.$c} ||
+ -e $IkiWiki::Plugin::transient::transientdir."/".htmlfn($page.$c) ||
-e "$config{srcdir}/".htmlfn($page.$c)) {
$c++
}
+ $page=$page.$c;
- # Make sure that the file name isn't too long.
- # NB: This doesn't check for path length limits.
- my $max=POSIX::pathconf($config{srcdir}, &POSIX::_PC_NAME_MAX);
- if (defined $max && length(htmlfn($page)) >= $max) {
+ $guid->{page}=$page;
+ eval { write_page($feed, $guid, $mtime, \%params) };
+ if ($@) {
+ # assume failure was due to a too long filename
$c="";
$page=$feed->{dir}."/item";
while (exists $IkiWiki::pagecase{lc $page.$c} ||
- -e "$config{srcdir}/".htmlfn($page.$c)) {
+ -e $IkiWiki::Plugin::transient::transientdir."/".htmlfn($page.$c) ||
+ -e "$config{srcdir}/".htmlfn($page.$c)) {
$c++
}
+ $page=$page.$c;
+
+ $guid->{page}=$page;
+ write_page($feed, $guid, $mtime, \%params);
}
- $guid->{page}=$page;
debug(sprintf(gettext("creating new page %s"), $page));
}
+}
+
+sub write_page ($$$$$) {
+ my $feed=shift;
+ my $guid=shift;
+ my $mtime=shift;
+ my %params=%{shift()};
+
$guid->{feed}=$feed->{name};
# To write or not to write? Need to avoid writing unchanged pages
$template=template($feed->{template}, blind_cache => 1);
};
if ($@) {
- print STDERR gettext("failed to process template:")." $@";
+ # gettext can clobber $@
+ my $error = $@;
+ print STDERR gettext("failed to process template:")." $error";
return;
}
$template->param(title => $params{title})
if defined $params{title} && length($params{title});
+ $template->param(author => $params{author})
+ if defined $params{author} && length($params{author}
+ && $params{author} ne $feed->{name});
$template->param(content => wikiescape(htmlabs($params{content},
defined $params{base} ? $params{base} : $feed->{feedurl})));
$template->param(name => $feed->{name});
$template->param(url => $feed->{url});
$template->param(copyright => $params{copyright})
if defined $params{copyright} && length $params{copyright};
- $template->param(permalink => urlabs($params{link}, $feed->{feedurl}))
+ $template->param(permalink => IkiWiki::urlabs($params{link}, $feed->{feedurl}))
if defined $params{link};
if (ref $feed->{tags}) {
$template->param(tags => [map { tag => $_ }, @{$feed->{tags}}]);
}
- writefile(htmlfn($guid->{page}), $config{srcdir},
- $template->output);
+ writefile(htmlfn($guid->{page}),
+ $IkiWiki::Plugin::transient::transientdir, $template->output);
if (defined $mtime && $mtime <= time) {
# Set the mtime, this lets the build process get the right
# creation time on record for the new page.
- utime $mtime, $mtime, "$config{srcdir}/".htmlfn($guid->{page});
+ utime $mtime, $mtime,
+ $IkiWiki::Plugin::transient::transientdir."/".htmlfn($guid->{page});
# Store it in pagectime for expiry code to use also.
$IkiWiki::pagectime{$guid->{page}}=$mtime
unless exists $IkiWiki::pagectime{$guid->{page}};
return encode_entities(shift, '\[\]');
}
-sub urlabs ($$) {
- my $url=shift;
- my $urlbase=shift;
-
- URI->new_abs($url, $urlbase)->as_string;
-}
-
sub htmlabs ($$) {
# Convert links in html from relative to absolute.
# Note that this is a heuristic, which is not specified by the rss
next unless $v_offset; # 0 v_offset means no value
my $v = substr($text, $v_offset, $v_len);
$v =~ s/^([\'\"])(.*)\1$/$2/;
- my $new_v=urlabs($v, $urlbase);
+ my $new_v=IkiWiki::urlabs($v, $urlbase);
$new_v =~ s/\"/"/g; # since we quote with ""
substr($text, $v_offset, $v_len) = qq("$new_v");
}