Merge commit 'upstream/master' into prv/po

Conflicts:

	IkiWiki/Plugin/editpage.pm
	IkiWiki/Plugin/skeleton.pm.example
	doc/plugins/write.mdwn
	t/syntax.t

IkiWiki/Plugin/blogspam.pm

@@ -0,0 +1,111 @@
+#!/usr/bin/perl
+package IkiWiki::Plugin::blogspam;
+
+use warnings;
+use strict;
+use IkiWiki 3.00;
+require RPC::XML;
+require RPC::XML::Client;
+
+my $defaulturl='http://test.blogspam.net:8888/';
+
+sub import {
+	hook(type => "getsetup", id => "blogspam",  call => \&getsetup);
+	hook(type => "checkcontent", id => "blogspam", call => \&checkcontent);
+}
+
+sub getsetup () {
+	return
+		plugin => {
+			safe => 1,
+			rebuild => 0,
+		},
+		blogspam_pagespec => {
+			type => 'pagespec',
+			example => 'postcomment(*)',
+			description => 'PageSpec of pages to check for spam',
+			link => 'ikiwiki/PageSpec',
+			safe => 1,
+			rebuild => 0,
+		},
+		blogspam_options => {
+			type => "string",
+			example => "blacklist=1.2.3.4,blacklist=8.7.6.5,max-links=10",
+			description => "options to send to blogspam server",
+			link => "http://blogspam.net/api/testComment.html#options",
+			safe => 1,
+			rebuild => 0,
+		},
+		blogspam_server => {
+			type => "string",
+			default => $defaulturl,
+			description => "blogspam server XML-RPC url",
+			safe => 1,
+			rebuild => 0,
+		},
+}
+
+sub checkcontent (@) {
+	my %params=@_;
+	
+ 	if (exists $config{blogspam_pagespec}) {
+		return undef
+			if ! pagespec_match($params{page}, $config{blogspam_pagespec},
+	                	location => $params{page});
+	}
+
+	my $url=$defaulturl;
+	$url = $params{blogspam_server} if exists $params{blogspam_server};
+	my $client = RPC::XML::Client->new($url);
+
+	my @options = split(",", $params{blogspam_options})
+		if exists $params{blogspam_options};
+
+	# Allow short comments and whitespace-only edits, unless the user
+	# has overridden min-words themselves.
+	push @options, "min-words=0"
+		unless grep /^min-words=/i, @options;
+	# Wiki pages can have a lot of urls, unless the user specifically
+	# wants to limit them.
+	push @options, "exclude=lotsaurls"
+		unless grep /^max-links/i, @options;
+	# Unless the user specified a size check, disable such checking.
+	push @options, "exclude=size"
+		unless grep /^(?:max|min)-size/i, @options;
+	# This test has absurd false positives on words like "alpha"
+	# and "buy".
+	push @options, "exclude=stopwords";
+
+	# blogspam API does not have a field for author url, so put it in
+	# the content to be checked.
+	if (exists $params{url}) {
+		$params{content}.="\n".$params{url};
+	}
+
+	my $res = $client->send_request('testComment', {
+		ip => $ENV{REMOTE_ADDR},
+		comment => $params{content},
+		subject => defined $params{subject} ? $params{subject} : "",
+		name => defined $params{author} ? $params{author} : "",
+		options => join(",", @options),
+		site => $config{url},
+		version => "ikiwiki ".$IkiWiki::version,
+	});
+
+	if (! ref $res || ! defined $res->value) {
+		debug("failed to get response from blogspam server ($url)");
+		return undef;
+	}
+	elsif ($res->value =~ /^SPAM:(.*)/) {
+		return gettext("Sorry, but that looks like spam to <a href=\"http://blogspam.net/\">blogspam</a>: ").$1;
+	}
+	elsif ($res->value ne 'OK') {
+		debug(gettext("blogspam server failure: ").$res->value);
+		return undef;
+	}
+	else {
+		return undef;
+	}
+}
+
+1

IkiWiki/Plugin/comments.pm

@@ -343,8 +343,6 @@ sub sessioncgi ($$) {
 		error(gettext("bad page name"));
 	}
 
-	# FIXME: is this right? Or should we be using the candidate subpage
-	# (whatever that might mean) as the base URL?
 	my $baseurl = urlto($page, undef, 1);
 
 	$form->title(sprintf(gettext("commenting on %s"),
@@ -469,10 +467,22 @@ sub sessioncgi ($$) {
 	}
 
 	if ($form->submitted eq POST_COMMENT && $form->validate) {
-		my $file = "$location._comment";
-
 		IkiWiki::checksessionexpiry($cgi, $session);
 		
+		$postcomment=1;
+		IkiWiki::check_content(content => $form->field('editcontent'),
+			subject => $form->field('subject'),
+			$config{comments_allowauthor} ? (
+				author => $form->field('author'),
+				url => $form->field('url'),
+			) : (),
+			page => $location,
+			cgi => $cgi, session => $session
+		);
+		$postcomment=0;
+		
+		my $file = "$location._comment";
+
 		# FIXME: could probably do some sort of graceful retry
 		# on error? Would require significant unwinding though
 		writefile($file, $config{srcdir}, $content);

IkiWiki/Plugin/editpage.pm

@@ -78,7 +78,43 @@ sub check_canedit ($$$;$) {
 			}
 		}
 	});
-	return $canedit;
+	return defined $canedit ? $canedit : 1;
+}
+
+sub check_content (@) {
+	my %params=@_;
+	
+	return 1 if ! exists $hooks{checkcontent}; # optimisation
+
+	if (exists $pagesources{$params{page}}) {
+		my @diff;
+		my %old=map { $_ => 1 }
+		        split("\n", readfile(srcfile($pagesources{$params{page}})));
+		foreach my $line (split("\n", $params{content})) {
+			push @diff, $line if ! exists $old{$_};
+		}
+		$params{content}=join("\n", @diff);
+	}
+
+	my $ok;
+	run_hooks(checkcontent => sub {
+		return if defined $ok;
+		my $ret=shift->(%params);
+		if (defined $ret) {
+			if ($ret eq "") {
+				$ok=1;
+			}
+			elsif (ref $ret eq 'CODE') {
+				$ret->();
+				$ok=0;
+			}
+			elsif (defined $ret) {
+				error($ret);
+			}
+		}
+
+	});
+	return defined $ok ? $ok : 1;
 }
 
 sub check_cansave ($$$$) {
@@ -396,8 +432,17 @@ sub cgi_editpage ($$) {
 			exit;
 		}
 			
+		my $message="";
+		if (defined $form->field('comments') &&
+		    length $form->field('comments')) {
+			$message=$form->field('comments');
+		}
+		
 		my $content=$form->field('editcontent');
 		check_cansave($page, $content, $q, $session);
+		check_content(content => $content, page => $page,
+			cgi => $q, session => $session,
+			subject => $message);
 		run_hooks(editcontent => sub {
 			$content=shift->(
 				content => $content,
@@ -431,12 +476,6 @@ sub cgi_editpage ($$) {
 		
 		my $conflict;
 		if ($config{rcs}) {
-			my $message="";
-			if (defined $form->field('comments') &&
-			    length $form->field('comments')) {
-				$message=$form->field('comments');
-			}
-			
 			if (! $exists) {
 				rcs_add($file);
 			}

IkiWiki/Plugin/skeleton.pm.example

@@ -31,6 +31,7 @@ sub import {
 	hook(type => "sessioncgi", id => "skeleton", call => \&sessioncgi);
 	hook(type => "canedit", id => "skeleton", call => \&canedit);
 	hook(type => "cansave", id => "skeleton", call => \&cansave);
+	hook(type => "checkcontent", id => "skeleton", call => \&checkcontent);
 	hook(type => "editcontent", id => "skeleton", call => \&editcontent);
 	hook(type => "formbuilder_setup", id => "skeleton", call => \&formbuilder_setup);
 	hook(type => "formbuilder", id => "skeleton", call => \&formbuilder);
@@ -181,6 +182,7 @@ sub canedit ($$$) {
 	debug("skeleton plugin running in canedit");
 }
 
+<<<<<<< HEAD:IkiWiki/Plugin/skeleton.pm.example
 sub cansave ($$$$) {
 	my $page=shift;
 	my $content=shift;
@@ -204,6 +206,12 @@ sub canrename ($$$) {
 	my $session=shift;
 
 	debug("skeleton plugin running in canrename");
+=======
+sub checkcontent (@) {
+	my %params=@_;
+
+	debug("skeleton plugin running in checkcontent");
+>>>>>>> upstream/master:IkiWiki/Plugin/skeleton.pm.example
 }
 
 sub editcontent ($$$) {

auto-blog.setup

@@ -35,7 +35,7 @@ IkiWiki::Setup::Automator->import(
 	cgiurl => "http://$domain/~$ENV{USER}/$wikiname/ikiwiki.cgi",
 	cgi_wrapper => "$ENV{HOME}/public_html/$wikiname/ikiwiki.cgi",
 	adminemail => "$ENV{USER}\@$domain",
-	add_plugins => [qw{goodstuff websetup comments}],
+	add_plugins => [qw{goodstuff websetup comments blogspam}],
 	disable_plugins => [qw{}],
 	libdir => "$ENV{HOME}/.ikiwiki",
 	rss => 1,
@@ -44,5 +44,6 @@ IkiWiki::Setup::Automator->import(
 	
 	example => "blog",
 	comments_pagespec => "posts/* and !*/Discussion",
+	blogspam_pagespec => "postcomment(*)",
 	discussion => 0,
 )

debian/changelog

@@ -1,5 +1,11 @@
 ikiwiki (3.02) UNRELEASED; urgency=low
 
+  * blogspam: New plugin, adding spam filtering for page editing / comment
+    posting using the BlogSpam.net API.
+  * Add auto-blog.setup, which will set up an ikiwiki instance tuned for use
+    in blogging.
+  * checkcontent: New hook, can be used to implement arbitrary content
+    filters, including spam filters.
   * table: Fix misparsed links in external files.
   * table: Find links in external files in scan pass.
   * rename: Show full names of affected pages.
@@ -16,8 +22,6 @@ ikiwiki (3.02) UNRELEASED; urgency=low
     the nostromo web server.
   * Recommend libterm-readline-gnu-perl since that makes auto.setup
     behave better.
-  * Add auto-blog.setup, which will set up an ikiwiki instance tuned for use
-    in blogging.
 
  -- Joey Hess <joeyh@debian.org>  Tue, 06 Jan 2009 15:02:52 -0500
 

doc/download.mdwn

@@ -17,7 +17,7 @@ ikiwiki, there have been lots of enhancements and bug fixes since those
 versions.
 """]]
 
-        apt-get install ikiwiki
+        aptitude install ikiwiki
 
 Or download the deb from <http://packages.debian.org/unstable/web/ikiwiki>.
 

doc/ikiwikiusers.mdwn

@@ -39,6 +39,7 @@ Projects
 * [monkeysphere](http://web.monkeysphere.info/)
 * [The Walden Effect](http://www.waldeneffect.org/)
 * The [Fortran Wiki](http://fortranwiki.org/)
+* [Monotone](http://monotone.ca/wiki/FrontPage/)
 
 Personal sites and blogs
 ========================

doc/plugins/blogspam.mdwn

@@ -0,0 +1,25 @@
+[[!template id=plugin name=blogspam author="[[Joey]]"]]
+[[!tag type/auth]]
+
+This plugin adds antispam support to ikiwiki, using the
+[blogspam.net](http://blogspam.net/) API. Both page edits and
+[[comment|comments]] postings can be checked for spam. Currently,
+detected spam is not saved for human review, it is just rejected.
+
+The plugin requires the [[!cpan RPC::XML]] perl module.
+
+You can control how content is tested via the `blogspam_options` setting.
+The list of options is [here](http://blogspam.net/api/testComment.html#options).
+By default, the options are configured in a way that is appropriate for
+wiki content. This includes turning off some of the more problimatic tests.
+
+The `blogspam_pagespec` setting is a [[ikiwiki/PageSpec]] that can be
+used to configure which pages are checked for spam. The default is to check
+all edits. If you only want to check [[comments]] (not wiki page edits),
+set it to "postcomment(*)".
+
+By default, the blogspam.net server is used to do the spam checking. To
+change this, the `blogspam_server` option can be set to the url for a
+different server implementing the same API. Note that content is sent
+unencrypted over the internet to the server, and the server sees
+the full text of the content.

doc/plugins/write.mdwn

@@ -303,7 +303,7 @@ can check if the session object has a "name" parameter set.
 
 ### canedit
 
-	hook(type => "canedit", id => "foo", call => \&pagelocked);
+	hook(type => "canedit", id => "foo", call => \&canedit);
 
 This hook can be used to implement arbitrary access methods to control when
 a page can be edited using the web interface (commits from revision control
@@ -356,6 +356,26 @@ but is passed:
 * a session object
 * the named parameters `src`, `srcfile`, `dest` and `destfile`.
 
+### checkcontent
+	
+	hook(type => "checkcontent", id => "foo", call => \&checkcontent);
+
+This hook is called to check the content a user has entered on a page,
+before it is saved, and decide if it should be allowed.
+
+It is passed named parameters: `content`, `page`, `cgi`, and `session`. If
+the content the user has entered is a comment, it may also be passed some
+additional parameters: `author`, `url`, and `subject`. The `subject`
+parameter may also be filled with the user's comment about the change.
+
+Note: When the user edits an existing wiki page, the passed `content` will
+include only the lines that they added to the page, or modified.
+
+The hook should return `undef` on success. If the content is disallowed, it
+should return a message stating what the problem is, or a function
+that can be run to perform whatever action is necessary to allow the user
+to post the content.
+
 ### editcontent
 
 	hook(type => "editcontent", id => "foo", call => \&editcontent);

doc/sandbox.mdwn

@@ -1,4 +1,4 @@
-This is the SandBox, a page anyone can edit to try out ikiwiki.
+This is the [[SandBox]], a page anyone can edit to try out ikiwiki.
 
 hello
 
@@ -6,6 +6,10 @@ testing 1..2..3!!
 
 ----
 
+I am testing the edit box provided through ikiwiki.cgi.
+
+----
+
 Here's a paragraph.
 
 The following code block is pre-formatted:

doc/todo/anti-spam_protection.mdwn

@@ -17,3 +17,14 @@ Cheers,
 You might look at the Wikipedia page on "Spam\_in\_blogs" for more ideas.  In particular, would it be possible to force a subset of the pages (by regex, but you'd choose the regex to match those pages which are publicly writable) to use rel="nofollow" in all links.
 
 > I just wanted to leave a link here to the [[todo/require_CAPTCHA_to_edit]] plugin patch.  Unfortunately that plugin currently interacts badly with the openid plugin. -- [[Will]]
+
+
+---
+
+Ikiwiki now has a checkcontent hook that plugins can use to see content
+that is being entered and check it for spam/whatever.
+
+There is a blogspam plugin that uses the blogspam.org service
+to check for common spam signatures. --[[Joey]] 
+
+[[done]]

t/syntax.t

@@ -5,8 +5,8 @@ use Test::More;
 
 my @progs="ikiwiki.in";
 my @libs="IkiWiki.pm";
-# monotone, external, amazon_s3 skipped since they need perl modules
+# monotone, external, blogspam, amazon_s3, po skipped since they need perl modules
-push @libs, map { chomp; $_ } `find IkiWiki -type f -name \\*.pm | grep -v monotone.pm | grep -v external.pm | grep -v amazon_s3.pm | grep -v po.pm`;
+push @libs, map { chomp; $_ } `find IkiWiki -type f -name \\*.pm | grep -v monotone.pm | grep -v external.pm | grep -v blogspam.pm | grep -v amazon_s3.pm | grep -v po.pm`;
 push @libs, 'IkiWiki/Plugin/skeleton.pm.example';
 
 plan(tests => (@progs + @libs));