Author: mavcunha
Description:
I've just installed the 1.4rc1 to manage some developer documentation but the
Wiki was elected to manage all documents on my site. The first trouble I've got
was the read access rights.
The Whitelist(Read|Edit|Account) works fine but how can I isolate part of the
Wiki from anonymous readers? What I need was an option like WhitelistRead but
doing the exactly the reverse thing, this way I could protect some restricted
information without worry with the growth of the Wiki.
Hacking the code I've came with some rough solution, here is it: First you will have another option in LocalSettings.php called
$wgProtectedPages witch is an
array.
So my LocalSettings.php looks like:
--snip--
# Disabling anonymous edits.
$wgWhitelistEdit = true;
# Specify who may create new accounts: 0 means no, 1 means yes
$wgWhitelistAccount = array ( 'user' => 0, 'sysop' => 1, 'developer' => 0 );
$wgWhitelistRead = array('Main Page');
$wgProtectedPages = array (
'Protected',
'Systems.*LDAP$' #you can use RegExp here.
);
--snip--
For this work you should define $wgWhitelistRead, so putting the 'Main Page'is a good idea since you probably won't block this page.
Now, for the $wgProtectedPages you can define in each element an (Perl)
Regular Expression (I will use preg_match() to check the title) or the name of
the page you want to restrict.
And now the hack itself, in Title.php (witch asks for the isAllowed('read')in User.php), in function userCanRead() /Sorry but I have to paste some code here/:
--snip--
} else {
global $wgWhitelistRead, $wgProtectedPages; # our new option
/** If anon users can create an account,
they need to reach the login page first! */
if( $wgUser->isAllowed( 'createaccount' )
&& $this->getNamespace() == NS_SPECIAL
&& $this->getText() == 'Userlogin' ) {
return true;
}
$name = $this->getPrefixedText();
# the new code, witch tries with preg_match() find
# if this page is protected so anon can't read this.
if (is_array($wgProtectedPages) ) {
foreach($wgProtectedPages as $pageName) {
if( preg_match("/$pageName/",$name) ) {
return false;
}
}
return true;
}
/** some pages are explicitly allowed */
if( in_array( $name, $wgWhitelistRead ) ) {
return true;
}
--snip--With this kind of check I can protect a complete subtree, generate a
auto-protected for new pages using some keyword in the title, or protect just
one single page. As you can see $wgProtectedPages has precedence above
$wgWhitelistRead.
Hope I didn't broke something, any thought?
Version: unspecified
Severity: enhancement