hello all,
I am beginner to this list, since i have just started hacking
mediawiki. I am trying to apply the LDAP authentication plugin to
mediawiki-1.5.2 in a FC-4. I'm using openldap-2.2.23-5.
I have read the documentation at :
http://meta.wikimedia.org/wiki/LDAP_Authentication and downloaded the
plugin from :http://bugzilla.wikipedia.org/attachment.cgi?id=1042&action=view
I have dropped the above file in /includes and changed
Localsettings.php accordingly. Luckily everything seems to be working
fine.
But everything crashed when I tried to Group based authentication in
mediawiki. I have added group entry to the ldap server, restarted it
and then inlcuded the configuration parameters for group based
authentication. Now my Localsettings.php looks like this:
<code>
require_once( 'LdapAuthentication.php' );
$wgAuth = new LdapAuthenticationPlugin();
$wgLDAPDomainNames = array( "libregeek" );
$wgLDAPServerNames = array( "libregeek"=>"localhost" );
$wgLDAPSearchStrings = array(
"libregeek"=>"uid=USER-NAME,ou=People,dc=libregeek,dc=net" );
$wgLDAPUseSSL = false;
$wgLDAPUseLocal = false;
$wgLDAPAddLDAPUsers = false;
$wgLDAPUpdateLDAP = false;
$wgLDAPMailPassword = false;
$wgLDAPRetrievePrefs = false;
$wgMinimalPasswordLength = 1;
$wgLDAPGroupDN = "cn=itpeople,ou=Groups,dc=libregeek,dc=net";
$wgLDAPProxyAgent = "cn=root,dc=libregeek,dc=net";
$wgLDAPProxyAgentPassword = "secret"; //this should work with hashes btw
$wgLDAPBaseDNs = array("libregeek"=>"dc=libregeek,dc=net");
$wgLDAPSearchAttributes =
array("libregeek"=>"(uid=USER-NAME)(wikiAccess=TRUE)");
</code>
With this configuration I can't login . It simply says "The password
you entered is incorrect (or missing). Please try again.".
is there any missing configuration items ??
I just tried to print the global variables corresponding to Proxies in
LdapAuthentication.php and it gives nothing.
One more thing I didn't got the correct idea of Proxy. is it like the
"Manager" account who has read permission to all the directory
entries.
what may have gone wrong?
please help
regards
Manilal
Some of Magnus's recent work inspired me to dig into Article.php and rework how
redirects are handled.
My current in-progress code is attached on this bug:
http://bugzilla.wikimedia.org/show_bug.cgi?id=1103
Formerly, the following of redirects was handled inside Article's
getContent/fetchContent functions, and which would get called inside the view()
action handler. This is problematic for several reasons:
* When loading text for other purposes, you could trip a global redirect by
accident. Ugly parameters got hacked in to try to work around that.
* Redirecting from one type of page to another required all kinds of extra hacks
or just plain didn't work; regular articles, image pages, and category pages are
handled by different variants of the Article class while Special pages do
something else entirely.
* It's daaaaaamn ugly; lots of checks of globals and URL parameters in the
middle of text-loading code, and the global title might change while you're
halfway through a function.
In the new code, the setup code in Wiki.php creates an Article object, asks if
it's a redirect, and if so follows it to get another Article object. It feels
like a cleaner interface to me, and minimizes the amount of weird global changes
appearing as side effects of low-level content-loading methods.
With some further work, it should even be possible to handle redirects to
Special pages properly; currently those are done with a URL redirect (or not at
all, if interwiki redirects are disabled). Instead, we could load up the special
page and shove the 'redirected from' link+subheading at it like we do for
regular pages.
Since this changes some of the guts of the wiki, I haven't committed it quite
yet but plan to soon. Please test, and let me know about anything that breaks or
post a fix to the patch. (Remember, CVS HEAD is our production code and code
should always be carefully tested before committing there.)
-- brion vibber (brion @ pobox.com)
Jan:
Thanks a lot for your answer.I have another maybe a little foolish question. How can I get into the webserver-commandline?I' ve tried to look for the help in meta pages but can't find it.
ffaarr
>Hello ffaarr,
>the category-links are included in the articles? But the articles are not
>assigned to the categorys?
>try to execute "maintenance/refreshlinks.php" on webserver-commandline.
>Regards,
>Jan
Anonymouse.org is a proxy popular with Chinese users to circumvent
censorship. Unfortunately, not all Wikipedias are accessible through
it. I just tried a large sample, and I found that the following do
*not* work: ja, ko, zh, mr, th. What happens is that anonymouse
redirects the main page of these Wikipedias back to the main page,
resulting in an infinite loop.
The brokenness of zh is of particular concern, for obvious reasons.
It might be a unicode issue, but ar, he, ru all work fine. I doubt that
is is a bug in anonymouse that affects only certain unicode characters.
I was also thinking that it might have to do with the Korean cluster,
but I belive ms is housed there, and it works fine through anonymouse.
I used wget -S --spider to look at the return values of our servers,
and I can't find a relevant difference between zh and he, say. However,
anonymouse probably uses a special user-agent in order to get a simple
page layout; I don't know which one that is so I can't compare what
exactly anonymouse sees. Could someone with access to the logs check
the user-agent of anonymouse?
Thanks,
Axel
The fact that zh doesn't work is particularly unsettling, for obvious
reasons.
Axel
__________________________________
Yahoo! for Good - Make a difference this year.
http://brand.yahoo.com/cybergivingweek2005/
I recently moved my MediaWiki 1.5.3 database from a MySQL 4.0 server to a MySQL
4.1 server. OLDPASSWORDS is enabled, and PHP 4.4.1 is using the newest MySQL
libraries. I am concurrently running a Simple Machines Forum (which works
flawlessly, showing that there is no problem with password hashing on MySQL 4.1)
After transfering to MySQL 4.1, all pages are returned blank
(<html><body></body></html) after an excessively long load time. I've tried
upgrading to 1.5.5, and even running a fresh install of 1.5.5, all to no avail.
During the install, the 'Initialising "MediaWiki" namespace' step takes
excessively long.
I have no access to my PHP error logs, because my host (servage.net) is about as
poor as they come when it deals with customer service.
Any help on this issue would be GREATLY appreciated.
Hello,
the installation of the ten new apaches currently suffers from some Fedora
Core 4 bugs.
srv80 hangs because the yum update of the diskdumptuils fails due to
incompatible changes in a patch.
The effect is described here:
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=170098
Srv79 is unreachable because a
yum update kernel kernel-smp kernel-smp-devel; reboot
didn't return. No ping response any more from srv79.
Regards,
jens
--
Telefonieren Sie schon oder sparen Sie noch?
NEU: GMX Phone_Flat http://www.gmx.net/de/go/telefonie
Hi! Since the switch to XML format, we've seen regular dumps every 1-2
weeks, which was much appreciated by people like me who run analysis
tools on them. Unfortunately, there hasn't been any new dump since the
last Dec 13-18 run. Does anybody know if/when they are scheduled to
resume?
--
Frédéric Brière <*> fbriere(a)fbriere.net
=> <fbriere(a)abacom.com> IS NO MORE: <http://www.abacomsucks.com> <=
good folk,
find at meta User:Nigelk/ConcatPages to have source code for
SpecialConcatPages.php, a special page to work with the NavMap extension
(at User:Nigelk/Nav). this is, you understand, a 0.1. it does this:
takes three POST or GET url parameters:
map (some extant map)
start (some page in the map)
end (some succeeding page in the map)
and tries to output the concatenation of that start, all the
pages in between (per the map), the end, and the children of the
end into the body of the special page.
no features at the moment other than that. the page points to a
demonstration.
no reason one could not apply this to a parameter that had arbitrary
pages listed in order. and i think i finally 'got' wfMsg and the
MessageCache, et al. only really tested it by me, with php 5.0.4 live
on the site.
comments, here or en Meta, as ever, welcome.
cheers,
nigel
I don't want to be an administrator, but I do want to vote on
a deletion review, or at least know enough to figure out whether or
not I should abstain. Would someone please grant me "deletedhistory"
permission only? Is that possible? Do I need to get nominated for
just that?
I am User:Nrcprm2026
Sincerely,
James Salsman