wordcount: fix test

Don't just limit to full URL: accept any non-whitespace characters as
a link.

modules/calendar.pl

@@ -1,4 +1,4 @@
-# Copyright (C) 2004, 2005, 2006  Alex Schroeder <alex@emacswiki.org>
+# Copyright (C) 2004–2023  Alex Schroeder <alex@gnu.org>
 # Copyright (C) 2006  Ingo Belka
 #
 # This program is free software; you can redistribute it and/or modify
@@ -112,9 +112,7 @@ sub DoCollect {
   my $search = GetParam('search', '');
   ReportError(T('The match parameter is missing.')) unless $match or $search;
   print GetHeader('', Ts('Page Collection for %s', $match||$search), '');
-  my @pages = (grep(/$match/, $search
-		    ? SearchTitleAndBody($search)
-		    : AllPagesList()));
+  my @pages = Matched($match, $search ? SearchTitleAndBody($search) : AllPagesList());
   if (!$CollectingJournal) {
     $CollectingJournal = 1;
     # Now save information required for saving the cache of the current page.

modules/index.pl

@@ -1,4 +1,4 @@
-# Copyright (C) 2004, 2007  Alex Schroeder <alex@emacswiki.org>
+# Copyright (C) 2004–2023  Alex Schroeder <alex@gnu.org>
 #
 # This program is free software; you can redistribute it and/or modify
 # it under the terms of the GNU General Public License as published by
@@ -47,8 +47,5 @@ sub PrintableIndexPages {
   push(@pages, AllPagesList()) if GetParam('pages', 1);
   push(@pages, keys %PermanentAnchors) if GetParam('permanentanchors', 1);
   push(@pages, keys %NearSource) if GetParam('near', 0);
-  my $match = GetParam('match', '');
-  @pages = grep /$match/i, @pages if $match;
-  @pages = sort @pages;
-  return @pages;
+  return sort Matched(GetParam('match'), @pages);
 }

modules/journal-rss.pl

@@ -1,4 +1,4 @@
-# Copyright (C) 2004–2021  Alex Schroeder <alex@gnu.org>
+# Copyright (C) 2004–2023  Alex Schroeder <alex@gnu.org>
 #
 # This program is free software; you can redistribute it and/or modify
 # it under the terms of the GNU General Public License as published by
@@ -86,7 +86,7 @@ sub JournalRssGetRcLines {
   my $reverse = GetParam('reverse', 0);
   my $monthly = GetParam('monthly', 0);
   my $offset = GetParam('offset', 0);
-  my @pages = sort JournalSort (grep(/$match/, $search ? SearchTitleAndBody($search) : AllPagesList()));
+  my @pages = sort JournalSort (Matched($match, $search ? SearchTitleAndBody($search) : AllPagesList()));
   if ($monthly and not $match) {
     my ($sec, $min, $hour, $mday, $mon, $year, $wday, $yday) = gmtime();
     $match = '^' . sprintf("%04d-%02d", $year+1900, $mon+1) . '-\d\d';

modules/mail.pl

@@ -1,4 +1,4 @@
-# Copyright (C) 2009–2020  Alex Schroeder <alex@gnu.org>
+# Copyright (C) 2009–2022  Alex Schroeder <alex@gnu.org>
 # Copyright (C) 2015 Aleks-Daniel Jakimenko <alex.jakimenko@gmail.com>
 #
 # This program is free software; you can redistribute it and/or modify it under

modules/markdown-rule.pl

@@ -1,5 +1,5 @@
 #! /usr/bin/perl
-# Copyright (C) 2014–2019  Alex Schroeder <alex@gnu.org>
+# Copyright (C) 2014–2022  Alex Schroeder <alex@gnu.org>
 
 # This program is free software: you can redistribute it and/or modify it under
 # the terms of the GNU General Public License as published by the Free Software
@@ -194,12 +194,11 @@ sub MarkdownRule {
     return OpenHtmlEnvironment('pre',1) . $str; # always level 1
   }
   # link: [an example](http://example.com/ "Title")
-  elsif (m/\G\[((?:[^]\n]+\n?)+)\]\($FullUrlPattern(\s+"(.+?)")?\)/cg) {
+  elsif (m/\G\[((?:[^]\n]+\n?)+)\]\((\S+)(\s+"(.+?)")?\)/cg) {
     my ($text, $url, $title) = ($1, $2, $4);
-    $url =~ /^($UrlProtocols)/;
     my %params;
     $params{-href} = $url;
-    $params{-class} = "url $1";
+    $params{-class} = "url";
     $params{-title} = $title if $title;
     return $q->a(\%params, $text);
   }

modules/namespaces.pl

@@ -1,4 +1,4 @@
-# Copyright (C) 2004, 2005, 2006, 2007, 2008, 2009, 2012  Alex Schroeder <alex@gnu.org>
+# Copyright (C) 2004–2022  Alex Schroeder <alex@gnu.org>
 #
 # This program is free software; you can redistribute it and/or modify it under
 # the terms of the GNU General Public License as published by the Free Software

modules/network-blocker.pl

@@ -0,0 +1,198 @@
+# -*- mode: perl -*-
+# Copyright (C) 2023  Alex Schroeder <alex@gnu.org>
+
+# This program is free software: you can redistribute it and/or modify it under
+# the terms of the GNU Affero General Public License as published by the Free
+# Software Foundation, either version 3 of the License, or (at your option) any
+# later version.
+#
+# This program is distributed in the hope that it will be useful, but WITHOUT
+# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
+# FOR A PARTICULAR PURPOSE. See the GNU Affero General Public License for more
+# details.
+#
+# You should have received a copy of the GNU Affero General Public License along
+# with this program. If not, see <https://www.gnu.org/licenses/>.
+
+=encoding utf8
+
+=head1 NAME
+
+Oddmuse Network Blocker
+
+=head1 DESCRIPTION
+
+This module hooks into regular Oddmuse Surge Protection. It adds the following
+features:
+
+Repeated offenders are blocked for increasingly longer times.
+
+For every offender, we record the CIDR their IP number belongs to. Everytime an
+IP number is blocked, all the CIDRs of the other blocked IPs are checked: if
+there are three or more blocked IP numbers sharing the same CIDRs, the CIDR
+itself is blocked.
+
+CIDR blocking works the same way: Repeated offenders are blocked for
+increasingly longer times.
+
+=head2 Behind a reverse proxy
+
+Make sure your config file copies the IP number to the correct environment
+variable:
+
+    $ENV{REMOTE_ADDR} = $ENV{HTTP_X_FORWARDED_FOR};
+
+=head1 SEE ALSO
+
+<Oddmuse Surge Protection|https://oddmuse.org/wiki/Surge_Protection>
+
+=cut
+
+use strict;
+use v5.10;
+use Net::IP qw(:PROC);
+use Net::DNS qw(rr);
+
+AddModuleDescription('network-blocker.pl', 'Network Blocker Extension');
+
+our ($Now, $DataDir, $SurgeProtectionViews, $SurgeProtectionTime);
+
+{
+  no warnings 'redefine';
+  *OldNetworkBlockerDelayRequired = \&DelayRequired;
+  *DelayRequired = \&NewNetworkBlockerDelayRequired;
+}
+
+# Block for at least this many seconds.
+my $NetworkBlockerMinimumPeriod = 30;
+
+# Every violation doubles the current period until this maximum is reached (four weeks).
+my $NetworkBlockerMaximumPeriod = 60 * 60 * 24 * 7 * 4;
+
+# All the blocked networks. Maps CIDR to an array [expiry timestamp, expiry
+# period].
+my %NetworkBlockerList;
+
+# Candidates are remembered for this many seconds.
+my $NetworkBlockerCachePeriod = 600;
+
+# All the candidate networks for a block. Maps IP to an array [ts, cidr, ...].
+# Candidates are removed after $NetworkBlockerCachePeriod.
+my %NetworkBlockerCandidates;
+
+sub NetworkBlockerRead {
+  my ($status, $data) = ReadFile("$DataDir/network-blocks");
+  return unless $status;
+  my @lines = split(/\n/, $data);
+  while ($_ = shift(@lines)) {
+    my @items = split(/,/);
+    $NetworkBlockerList{shift(@items)} = \@items;
+  }
+  # an empty line separates the two sections
+  while ($_ = shift(@lines)) {
+    my @items = split(/,/);
+    $NetworkBlockerCandidates{shift(@items)} = \@items;
+  }
+  return 1;
+}
+
+sub NetworkBlockerWrite {
+  RequestLockDir('network-blocks') or return '';
+  WriteStringToFile(
+    "$DataDir/network-blocks",
+    join("\n\n",
+         join("\n", map {
+           join(",", $_, @{$NetworkBlockerList{$_}})
+              } keys %NetworkBlockerList),
+         join("\n", map {
+           join(",", $_, @{$NetworkBlockerCandidates{$_}})
+              } keys %NetworkBlockerCandidates)));
+  ReleaseLockDir('network-blocks');
+}
+
+sub NewNetworkBlockerDelayRequired {
+  my $ip = shift;
+  # If $ip is a name and not an IP number, parsing fails. In this case, run the
+  # regular code.
+  my $ob = new Net::IP($ip);
+  return OldNetworkBlockerDelayRequired($ip) unless $ob;
+  # Read the file. If the file does not exist, no problem.
+  NetworkBlockerRead();
+  # See if the current IP number is one of the blocked CIDR ranges.
+  for my $cidr (keys %NetworkBlockerList) {
+    # Perhaps this CIDR block can be expired.
+    if ($NetworkBlockerList{$cidr}->[0] < $Now) {
+      delete $NetworkBlockerList{$cidr};
+      next;
+    }
+    # Forget the CIDR if it cannot be turned into a range.
+    my $range = new Net::IP($cidr);
+    if (not $range) {
+      warn "CIDR $cidr is blocked but has no range: " . Net::IP::Error();
+      delete $NetworkBlockerList{$cidr};
+      next;
+    }
+    # If the CIDR overlaps with the remote IP number, it's a block.
+    warn "Checking whether $ip is in $cidr\n";
+    my $overlap = $range->overlaps($ob);
+    # $IP_PARTIAL_OVERLAP (ranges overlap) $IP_NO_OVERLAP (no overlap)
+    # $IP_A_IN_B_OVERLAP (range2 contains range1) $IP_B_IN_A_OVERLAP (range1
+    # contains range2) $IP_IDENTICAL (ranges are identical) undef (problem)
+    if (defined $overlap and $overlap != $IP_NO_OVERLAP) {
+      # Double the block period unless it has reached $NetworkBlockerMaximumPeriod.
+      if ($NetworkBlockerList{$cidr}->[1] < $NetworkBlockerMaximumPeriod / 2) {
+        $NetworkBlockerList{$cidr}->[1] *= 2;
+      } else {
+        $NetworkBlockerList{$cidr}->[1] = $NetworkBlockerMaximumPeriod;
+      }
+      $NetworkBlockerList{$cidr}->[0] = $Now + $NetworkBlockerList{$cidr}->[1];
+      # And we're done!
+      NetworkBlockerWrite();
+      ReportError(Ts('Too many connections by %s', $cidr)
+		  . ': ' . Tss('Please do not fetch more than %1 pages in %2 seconds.',
+			       $SurgeProtectionViews, $SurgeProtectionTime),
+		  '503 SERVICE UNAVAILABLE');
+    }
+  }
+  # If the CIDR isn't blocked, let's see if Surge Protection wants to block it.
+  my $result = OldNetworkBlockerDelayRequired($ip);
+  warn "$ip was blocked\n" if $result;
+  # If the IP is to be blocked, determine its CIDRs and put them on a list. Sadly,
+  # routeviews does not support IPv6 at the moment!
+  if ($result and not ip_is_ipv6($ip) and not $NetworkBlockerCandidates{$ip}) {
+    my $reverse = $ob->reverse_ip();
+    $reverse =~ s/in-addr\.arpa\.$/asn.routeviews.org/;
+    my @candidates;
+    for my $rr (rr($reverse, "TXT")) {
+      next unless $rr->type eq "TXT";
+      my @data = $rr->txtdata;
+      push(@candidates, join("/", @data[1..2]));
+    }
+    warn "$ip is in @candidates\n";
+    $NetworkBlockerCandidates{$ip} = [$Now, @candidates];
+    # Expire any of the other candidates
+    for my $other_ip (keys %NetworkBlockerCandidates) {
+      if ($NetworkBlockerCandidates{$other_ip}->[0] < $Now - $NetworkBlockerCachePeriod) {
+        delete $NetworkBlockerCandidates{$other_ip};
+      }
+    }
+    # Determine if any of the CIDRs is to be blocked.
+    my $save;
+    for my $cidr (@candidates) {
+      # Count how often the candidate CIDRs show up for other IP numbers.
+      my $count = 0;
+      for my $other_ip (keys %NetworkBlockerCandidates) {
+        my @data = $NetworkBlockerCandidates{$other_ip};
+        for my $other_cidr (@data[1 .. $#data]) {
+          $count++ if $cidr eq $other_cidr;
+        }
+      }
+      if ($count >= 3) {
+        $NetworkBlockerList{$cidr} = [$Now + $NetworkBlockerMinimumPeriod, $NetworkBlockerMinimumPeriod];
+        $save = 1;
+      }
+    }
+    NetworkBlockerWrite() if $save;
+  }
+  return $result;
+}

modules/rename-pages.pl

@@ -1,4 +1,4 @@
-# Copyright (C) 2019  Alex Schroeder <alex@gnu.org>
+# Copyright (C) 2019–2023  Alex Schroeder <alex@gnu.org>
 #
 # This program is free software; you can redistribute it and/or modify
 # it under the terms of the GNU General Public License as published by
@@ -65,6 +65,7 @@ sub RenamePageMenu {
 	 . GetHiddenValue('id', $id)
 	 . $q->textfield(-name=>'to', -size=>20)
 	 . ' '
-	 . $q->submit('Do it'));
+	 . $q->submit('Do it')
+         . $q->end_form());
   }
 }

modules/search-list.pl

@@ -1,4 +1,4 @@
-# Copyright (C) 2006, 2007, 2008  Alex Schroeder <alex@emacswiki.org>
+# Copyright (C) 2006–2023  Alex Schroeder <alex@gnu.org>
 #
 # This program is free software; you can redistribute it and/or modify
 # it under the terms of the GNU General Public License as published by
@@ -36,18 +36,12 @@ sub SearchListRule {
       $term = GetId();
     }
     local ($OpenPageName, %Page);
-    my %hash = ();
+    my @found;
     if ($variation eq 'list') {
-      foreach my $id (SearchTitleAndBody($term)) {
-        $hash{$id} = 1 unless $id eq $original; # skip the page with the query
-      }
+      @found = grep { $_ ne $original } SearchTitleAndBody($term);
+    } elsif ($variation eq 'titlelist') {
+      @found = grep { $_ ne $original } Matched($term, AllPagesList());
     }
-    if ($variation eq 'titlelist') {
-      foreach my $id (grep(/$term/, AllPagesList())) {
-        $hash{$id} = 1 unless $id eq $original; # skip the page with the query
-      }
-    }
-    my @found = keys %hash;
     if (defined &PageSort) {
       @found = sort PageSort @found;
     } else {
@@ -63,32 +57,24 @@ sub SearchListRule {
   return;
 }
 
-
 # Add a new action list
 
 $Action{list} = \&DoList;
 
 sub DoList {
-my $id = shift;
-my $match = GetParam('match', '');
-my $search = GetParam('search', '');
+  my $id = shift;
+  my $match = GetParam('match', '');
+  my $search = GetParam('search', '');
   ReportError(T('The search parameter is missing.')) unless $match or $search;
   print GetHeader('', Ts('Page list for %s', $match||$search), '');
   local (%Page, $OpenPageName);
-    my %hash = ();
-    foreach my $id (grep(/$match/, $search
-                    ? SearchTitleAndBody($search)
-                    : AllPagesList()))  {
-      $hash{$id} = 1;
-    }
-    my @found = keys %hash;
-    if (defined &PageSort) {
-      @found = sort PageSort @found;
-    } else {
-      @found = sort(@found);
-    }
-    @found = map { $q->li(GetPageLink($_)) } @found;
-    print $q->start_div({-class=>'search list'}),
-      $q->ul(@found), $q->end_div;
+  my @found = Matched($match, $search ? SearchTitleAndBody($search) : AllPagesList());
+  if (defined &PageSort) {
+    @found = sort PageSort @found;
+  } else {
+    @found = sort(@found);
+  }
+  @found = map { $q->li(GetPageLink($_)) } @found;
+  print $q->start_div({-class=>'search list'}), $q->ul(@found), $q->end_div;
   PrintFooter();
 }

modules/sistersites.pl

@@ -1,4 +1,4 @@
-# Copyright (C) 2007–2014  Alex Schroeder <alex@gnu.org>
+# Copyright (C) 2007–2023  Alex Schroeder <alex@gnu.org>
 #
 # This program is free software: you can redistribute it and/or modify it under
 # the terms of the GNU General Public License as published by the Free Software
@@ -36,8 +36,7 @@ sub SisterPages {
   push(@pages, AllPagesList()) if GetParam('pages', 1);
   push(@pages, keys %PermanentAnchors) if GetParam('permanentanchors', 1);
   push(@pages, keys %NearSource) if GetParam('near', 0);
-  my $match = GetParam('match', '');
-  @pages = grep /$match/i, @pages if $match;
+  @pages = Matched(GetParam('match', ''), @pages);
   @pages = sort @pages;
   return @pages;
 }

modules/wordcount.pl

@@ -23,7 +23,7 @@
 use strict;
 use v5.10;
 
-our (@MyInitVariables, $HtmlHeaders);
+our (@MyInitVariables, $HtmlHeaders, $EditNote);
 
 AddModuleDescription('wordcount.pl', 'Word Count Extension');
 
@@ -57,3 +57,5 @@ sub WordcountAddScript {
 	}
 	</script>";
 }
+
+$EditNote = "Words: <span id='textWordCount'></span>" . $EditNote;

scripts/tarballs.pl

@@ -85,7 +85,9 @@ versions of Oddmuse.</p>
 <ul>
 % for my $tarball (@$tarballs) {
 <li>
+% if ($tarball ne 'latest') {
 <a href="https://oddmuse.org/releases/<%= $tarball %>.tar.gz"><%= $tarball %>.tar.gz</a>
+% }
 (files for <%= link_to release => {tarball => $tarball} => begin %>\
 <%= $tarball =%><%= end %>)
 </li>

t/markdown-rule.t

@@ -137,13 +137,13 @@ EOT
 
 xpath_run_tests(split(/\n/,<<'EOT'));
 [example](http://example.com/)
-//a[@class="url http"][@href="http://example.com/"][text()="example"]
+//a[@class="url"][@href="http://example.com/"][text()="example"]
 [an example](http://example.com/)
-//a[@class="url http"][@href="http://example.com/"][text()="an example"]
+//a[@class="url"][@href="http://example.com/"][text()="an example"]
 [an example](http://example.com/ "Title")
-//a[@class="url http"][@href="http://example.com/"][@title="Title"][text()="an example"]
+//a[@class="url"][@href="http://example.com/"][@title="Title"][text()="an example"]
 [an\nexample](http://example.com/)
-//a[@class="url http"][@href="http://example.com/"][text()="an\nexample"]
+//a[@class="url"][@href="http://example.com/"][text()="an\nexample"]
 \n[an\n\nexample](http://example.com/)
 //p[text()="[an"]/following-sibling::p//text()[contains(string(),"example](")]
 EOT

t/rollback-hang.t

@@ -0,0 +1,58 @@
+# Copyright (C) 2006–2023  Alex Schroeder <alex@gnu.org>
+#
+# This program is free software; you can redistribute it and/or modify it under
+# the terms of the GNU General Public License as published by the Free Software
+# Foundation; either version 3 of the License, or (at your option) any later
+# version.
+#
+# This program is distributed in the hope that it will be useful, but WITHOUT
+# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
+# FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License along with
+# this program. If not, see <http://www.gnu.org/licenses/>.
+
+require './t/test.pl';
+package OddMuse;
+use Test::More tests => 4;
+use utf8;
+
+# Reproduce a particular bug from alexschroeder.ch with the rc.log provided.
+WriteStringToFile($RcFile, <<'EOT');
+16853910992023-05-29_Net_newsHow to IRCAnonymousAlex2en
+16854004152023-05-29_Net_newsHow to IRCAnonymousAlex3en
+1685430599[[rollback]]1685400415Anonymous
+16855185032023-05-29_Net_newsAnonymousAlex4en
+EOT
+
+local $SIG{ALRM} = sub { fail "timeout!"; kill 'KILL', $$; };
+alarm 3;
+# this is recent changes from between the rollback and the page before it, so there are no pages to roll back
+my $page = get_page("action=rss full=1 short=0 from=1685413682");
+alarm 0;
+test_page($page, '2023-05-29 Net news');
+test_page_negative($page, 'rollback');
+
+# Reproduce a follow-up bug. First, rolling back just Test works as intended.
+WriteStringToFile($RcFile, <<'EOT');
+1691499987Testham127.0.0.1Berta1
+1691499988Mustuff127.0.0.1Chris1
+1691499989Testspam127.0.0.1Spammer2
+1691499990Test0Rollback to 2023-08-08 13:06 UTC127.0.0.1Alex3
+1691499990[[rollback]]1691499987Test
+EOT
+
+my $feed = get_page('action=rc raw=1');
+test_page($feed, 'title: Test');
+
+# Rolling back all of the wiki doesn't work.
+WriteStringToFile($RcFile, <<'EOT');
+1691499987Testham127.0.0.1Berta1
+1691499988Mustuff127.0.0.1Chris1
+1691499989Testspam127.0.0.1Spammer2
+1691499990Test0Rollback to 2023-08-08 13:06 UTC127.0.0.1Alex3
+1691499990[[rollback]]1691499987
+EOT
+
+$feed = get_page('action=rc raw=1');
+test_page($feed, 'title: Test');

wiki.pl

@@ -1,5 +1,5 @@
 #! /usr/bin/env perl
-# Copyright (C) 2001-2020
+# Copyright (C) 2001-2023
 #     Alex Schroeder <alex@gnu.org>
 # Copyright (C) 2014-2015
 #     Alex Jakimenko <alex.jakimenko@gmail.com>
@@ -39,6 +39,7 @@ use B;
 use CGI qw/-utf8/;
 use CGI::Carp qw(fatalsToBrowser);
 use File::Glob ':glob';
+use List::Util qw(all max);
 use Encode qw(encode_utf8 decode_utf8);
 use sigtrap 'handler' => \&HandleSignals, 'normal-signals', 'error-signals';
 local $| = 1; # Do not buffer output (localized for mod_perl)
@@ -1543,28 +1544,34 @@ sub LatestChanges {
 sub StripRollbacks {
   my @result = @_;
   if (not (GetParam('all', $ShowAll) or GetParam('rollback', $ShowRollbacks))) { # strip rollbacks
-    my (%rollback);
+    my (%rollback); # used for single-page rollbacks
     for (my $i = $#result; $i >= 0; $i--) {
       # some fields have a different meaning if looking at rollbacks
       my ($ts, $id, $target_ts, $target_id) = @{$result[$i]};
+      # if this is a rollback marker
       if ($id eq '[[rollback]]') {
+        # if this is a single page rollback marker, strip it
 	if ($target_id) {
-	  $rollback{$target_id} = $target_ts; # single page rollback
-	  splice(@result, $i, 1);             # strip marker
+          # if this page is not already being rolled back, remember the target
+          # id and target ts so that those lines can be stripped below
+          if (not $rollback{$target_id} or $target_ts < $rollback{$target_id}) {
+            $rollback{$target_id} = $target_ts;
+          }
+          # the marker is always stripped
+	  splice(@result, $i, 1);
 	} else {
+          # if this is a global rollback, things are different: we're going to
+          # find the correct timestamp and strip all of those lines immediately
 	  my $end = $i;
-	  while ($ts > $target_ts and $i > 0) {
-	    $i--; # quickly skip all these lines
-	    $ts = $result[$i][0];
-	  }
-	  splice(@result, $i + 1, $end - $i);
-	  $i++; # compensate $i-- in for loop
+	  $i-- while $i > 0 and $target_ts < $result[$i-1][0];
+          # splice the lines found
+          splice(@result, $i, $end - $i + 1);
 	}
       } elsif ($rollback{$id} and $ts > $rollback{$id}) {
 	splice(@result, $i, 1); # strip rolled back single pages
       }
     }
-  } else {		  # just strip the marker left by DoRollback()
+  } else { # if rollbacks are not not shown, just strip the markers
     for (my $i = $#result; $i >= 0; $i--) {
       splice(@result, $i, 1) if $result[$i][1] eq '[[rollback]]'; # id
     }
@@ -3368,7 +3375,6 @@ sub SortIndex {
 
 sub DoIndex {
   my $raw = GetParam('raw', 0);
-  my $match = GetParam('match', '');
   my $limit = GetParam('n', '');
   my @pages = ();
   my @menu = ($q->label({-for=>'indexmatch'}, T('Filter:')) . ' '
@@ -3380,7 +3386,7 @@ sub DoIndex {
     push(@pages, $sub->()) if $value;
     push(@menu, $q->checkbox(-name=>$option, -checked=>$value, -label=>$text));
   }
-  @pages = grep /$match/i, @pages if $match;
+  @pages = Matched(GetParam('match', ''), @pages);
   @pages = sort SortIndex @pages;
   @pages = @pages[0 .. $limit - 1] if $limit;
   if ($raw) {
@@ -3553,11 +3559,23 @@ sub SearchTitleAndBody {
   return @found;
 }
 
-sub Filtered { # this is overwriten in extensions such as tags.pl
+# Filter the pages to be searched for $string. The default implementation
+# ignores $string and uses $match instead, just in case the user used both
+# search and match parameters. This is overwritten in extensions such as tags.pl
+# which extract tags from $string and use that to filter the pages.
+sub Filtered {
   my ($string, @pages) = @_;
-  my $match = GetParam('match', '');
-  @pages = grep /$match/i, @pages if $match;
-  return @pages;
+  return Matched(GetParam('match', ''), @pages);
+}
+
+sub Matched { # strictly for page titles
+  my ($string, @pages) = @_;
+  return @pages unless $string;
+  my @terms = grep { $_ } split(/[ _]+/, $string);
+  return grep {
+    my $id = $_;
+    all { $id =~ /$_/i } @terms;
+  } @pages;
 }
 
 sub SearchString {