# Project:Support desk

(Redirected from Support Desk)
vde   Welcome to MediaWiki.org's Support desk, where you can ask MediaWiki questions!

There are also other places where to ask: IRC, mailing lists, Q&A etc.

## Post a new question

• MediaWiki
• PHP
• Database
2. Please include the URL of your wiki unless you absolutely can't. It's often a lot easier for us to identify the source of the problem if we can look for ourselves.
3. To start a new thread, click "Start a new discussion".
Start a new discussion
 First page Previous page Next page Last page

## deleting user accounts from user list

running MediaWiki 1.24.1

Three questions:

1.Is it possible to delete user accounts after creating an account, especially if you have sent a random password and cant acquire the password?

2.Is it possible to change the password after creating the account if you have sent a randomly generated password to their email address?

3.Also, if you've selected to send a user a random password to their email when creating the account is it possible to resend this account verification email to them if they accidentally deleted it.

18:39, 27 July 2015

"1.Is it possible to delete user accounts after creating an account, especially if you have sent a random password and cant acquire the password?"

For help with Q1, you can try this Extension:UserMerge

21:06, 28 July 2015

## Starting conversion of LiquidThreads to Flow at mediawiki.org

LiquidThreads (LQT) has not been well-supported in a long time. Flow is in active development, and more real-world use-cases will help focus attention on the higher-priority features that are needed. To that end, LQT pages at mediawiki.org will start being converted to Flow in the next couple of weeks. This page, as the most active on mediawiki.org will be the last to be converted.

Please see details, and an emphatic request for feedback, at Topic:Sdoatsbslsafx6lw. Much thanks.

07:28, 17 March 2015
Edited by 2 users.
Last edit: 12:53, 5 June 2015

11:43, 18 March 2015

Update: All pages except for this one, and VisualEditor/Feedback (which will be converted tonight), have been converted to Flow. This final page will be converted during the Wednesday workday.

Thanks again to everyone who submitted bug reports, suggestions, constructive criticism, and feature requests. You might be interested to see the slides from the recent Wikimania presentation, File:User(s) Talk(ing) - Wikimania 2015.pdf, or to signup as a user research volunteer via http://wikimedia.org/research (they're currently doing some Flow research, but they cover many other aspects of the site in other months). As always, your comments and suggestions are welcome, encouraged, and needed, at Talk:Flow.

19:08, 28 July 2015

## ElasticSearch extension not supported for CirrusSearch Extension downloaded from latest WIKI version 1.25.1

Hi,

I have upgraded WIKI for latest stable version 1.25.1. As MWSearch is obsolete extension so I have tried install CirrusSearch with ElasticSearch extention. But unfortunately I found below error message :

"

Warning: Invalid argument supplied for foreach() in /data/www/html/wikinew/extensions/Elastica/ElasticaConnection.php on line 78 Warning: Search backend error during fetching elasticsearch version after 2. Param host does not exist [Called from CirrusSearch\ElasticsearchIntermediary::failure in /data/www/html/wikinew/extensions/CirrusSearch/includes/ElasticsearchIntermediary.php at line 98] in /data/www/html/wikinew/includes/debug/MWDebug.php on line 300


content index...

       Fetching Elasticsearch version...unable to determine, aborting.PHP Notice:  Undefined index: version in /data/www/html/wikinew/extensions/CirrusSearch/includes/Maintenance/ConfigUtils.php on line 53


Notice: Undefined index: version in /data/www/html/wikinew/extensions/CirrusSearch/includes/Maintenance/ConfigUtils.php on line 53 ...Not supported! Only Elasticsearch 1.x is supported. Your version: . "

15:07, 15 July 2015

It seems you haven't a running Elasticsearch server, or you haven't configured it in the configuration variable $wgCirrusSearchServers. 11:10, 16 July 2015 Hi, I configured ElasticSearch Server name in variable$wgCirrusSearchServers. But when I run the script updateSearchIndexConfig.php under /extensions/CirrusSearch/maintenance and I got below Error/Warning message.

content index...

1. PHP Warning: Invalid argument supplied for foreach() in /data/www/html/wikinew/extensions/Elastica/ElasticaConnection.php on line 78
2. Warning: Invalid argument supplied for foreach() in /data/www/html/wikinew/extensions/Elastica/ElasticaConnection.php on line 78 Fetching Elasticsearch version...unable to determine, aborting.PHP Notice: Undefined index: version in /data/www/html/wikinew/extensions/CirrusSearch/includes/Maintenance/ConfigUtils.php on line 53
3. Notice: Undefined index: version in /data/www/html/wikinew/extensions/CirrusSearch/includes/Maintenance/ConfigUtils.php on line 53
4. ...Not supported!
5. Only Elasticsearch 1.x is supported. Your version: .

Please suggest for the solution since it says Only Elasticsearch 1.x is supported.

Info: I have used 'CirrusSearch-REL1_25-ace4f80.tar.gz' & 'Elastica-REL1_25-eb56a6d.tar.gz'

Thank You !

12:22, 16 July 2015

After I configured in the configuration variable $wgCirrusSearchServers = array( 'elasticsearch0', 'elasticsearch1' ); Will you please let me know what a ElasticSearch Server is? Is it the same server where I hosted MediaWIKI 1.25.1 with the extensions 'CirrusSearch-REL1_25-ace4f80.tar.gz' & 'Elastica-REL1_25-eb56a6d.tar.gz' ? Still I get the below Err Notice after running script 'updateSearchIndexConfig.php' content index... 1. Fetching Elasticsearch version...unable to determine, aborting.PHP Notice: Undefined index: version in /data/www/html/wikinew/extensions/CirrusSearch/includes/Maintenance/ConfigUtils.php on line 53 2. Notice: Undefined index: version in /data/www/html/wikinew/extensions/CirrusSearch/includes/Maintenance/ConfigUtils.php on line 53 3. ...Not supported! 4. Only Elasticsearch 1.x is supported. Your version: . Please suggest. Thank you ! 13:49, 16 July 2015 You need to install elasticsearch on a server, it's a search service. You need to install it manually using the shell, it's normally not pre-installed. If it isn't installed, you can't use CirrusSearch (CirrusSearch itself doesn't work without elasticsearch). 06:02, 17 July 2015 I have successfully installed elasticsearch as a service using shell and I can able to access the service through the default port 9200 or by using curl. Below is the set of configuration on my 'LocalSettings.php' 1. require_once( "$IP/extensions/Elastica/Elastica.php" );
2. require_once( "$IP/extensions/CirrusSearch/CirrusSearch.php" ); 3. #$wgDisableSearchUpdate = true;
4. $wgCirrusSearchServers = array( 'example.domain.com' ); 5.$wgSearchType = 'CirrusSearch';

Still I have the below error while searching full text "

### An error has occurred while searching: We could not complete your search due to a temporary problem. Please try again later."

Kindly provide a fix to the above ERROR !

14:28, 28 July 2015

## Can't enable skin Vector.php

Hello,

I've just installed MW 1.24.2 and tried to display the main page. Unfortunately there is an error saying: \mediawiki\skins\Vector/SkinVector.php): failed to open stream: No such file or directory in ...\mediawiki\includes\AutoLoader.php on line 1282

I checked in the LocalSettings.php if there is a problem with the path, but the default skin is set for "vector" and require_once "$IP/skins/Vector/Vector.php"; seems to be alright as well. The path of SkinVector.php seems to be correct and I couldn't find any typos or similar in the directory. I tried to download the skin separately and paste it into the skins directory, but then it says its only compatible with MW1.25+. I realized the error shows up with every skin I try to enable. Any hints what might be the problem? I'm actually still thinking of a path problem, but I just don't know where to look after it. MediaWiki: 1.24.2 PHP: 5.6.3 Database: MySQL 5.6.17 14:10, 27 July 2015 When you install MediaWiki, have you then taken a checkout from the Git versioning system? Doing so would be a way I could imagine you get this error. In contrast, for normal end users, the recommended installation method is to use a tarball from here: http://releases.wikimedia.org/mediawiki/1.25/mediawiki-1.25.1.tar.gz. 1.25 btw. is the newest version - but a 1.24.2 tarball is available as well, if you want that. These tarballs contain everything you need, also the Vector skin in the correct version. :-) 88.130.109.19515:24, 27 July 2015 Hello and thank you for the reply! The reason why I don't use the 1.25 version is that I simply couldn't get the installation running - instead of getting the index.php to start from I always got a blank page, even tough I meet the requirements. In the meantime I could fix the problem with the vector skin. In Vector.php on line 36 I replaced "$GLOBALS['wgAutoloadClasses']['SkinVector'] = __DIR__ . '/SkinVector.php';" with "$GLOBALS['wgAutoloadClasses']['SkinVector'] = 'skins/Vector/SkinVector.php';" and it seems to work fine now. :) 07:03, 28 July 2015 Regarding blank pages: Have you checked phab:T44205 (and maybe Manual:Errors_and_symptoms#You_see_a_Blank_Page)? 08:42, 28 July 2015 Yes I have checked the errors and symptoms of a blank page, but none of them seems to work for me. Because I have a blank page before the install I don't have a LocalSettings to modify, nor I have access to the php.ini since I'm on a hosted server. Regarding the DOM-XML problem I have checked with the phpInfo, but DOM-XML support is enabled as well as the other requirements. 09:56, 28 July 2015 Your change in Vector.php looks strange: __DIR__ is defined since PHP 5.3 at least. If you have PHP 5.6, this change should not be needed and if you have an old PHP version, you should get a different error ("Use of undefined constant __DIR__..."). For the blank page, you should check the error log or have your host tell you, what after such error is inside. 88.130.74.14710:28, 28 July 2015 ## Boot Hello. I want to use some boot. Where should I start? I want to use boot f.e. to change categories on pages. 17:25, 27 July 2015 Which boot? 08:38, 28 July 2015 ## Create users by the admin did not work Hi, When creating a rather restricted Wiki (MediaWiki 1.25.1), I set: $wgGroupPermissions['*']['createaccount'] = false;
$wgGroupPermissions['*']['edit'] = false;$wgGroupPermissions['*']['read'] = false;


Then I created the users (Create Acccount) as admin and I did set "Use a temporary random password and send it to the specified email address". The users all got a mail with their username and a temporary passwort. But login with the temporary password did not work.

Do I have to do additional steps to 'enable' the users?

In the database there is no (hashed) password for the new users: select user_password from user where user_name='USERNAME'; returns an empty field. Shouldn't that contain the hashed (temporary) password?

Best regards

91.229.57.24015:17, 14 July 2015

Maybe another instance of task T102927?

21:18, 14 July 2015

I thought that as well, but I didn't write it, as -if I got it right- we are not speaking about confirmation emails here.

88.130.70.22622:29, 14 July 2015

Have you checked your server logs? Are there any related error/warning/notice messages?

23:08, 18 July 2015

I think that I have the same problem with version 1.25.1 Did you find a proper solution?

I found a workaround by telling the new users to use the lost password button, which would send them a new random password, which works OK.

79.86.101.24008:17, 19 July 2015

## HTML markup in search results

Hi,

In older versions of MW to removed HTML markup from search results you had to add $this->mText = strip_tags($this->mText ); to the SearchEngine.php but in 1.25.1 this file no longer has a function called getTextSnippet(). Any ideas where this function is or how to make it work now?

22:47, 27 July 2015

Hello,

MediaWiki:1.25.1 PHP:5.4.43 (cgi-fcgi) MySQL:5.5.42-cll Lua:5.1.5

http://forgewareinc.com/wiki

Everytime i upload an image i get a database error but no further info, if i look at the file list it states error code 127

File Page: http://forgewareinc.com/wiki/index.php?title=File:BluTech_logo.png

Any idea how i might fix this?

EDIT: After turning on detailed errors it now shows: Notice: Uncommitted DB writes (transaction from DatabaseBase::query (WikiPage::pageData)). in /home/forgewareinc/public_html/wiki/includes/db/Database.php on line 4362

20:41, 27 July 2015

## Search result without image as first.

Hello. I have problem with search result in a mediawiki. In the result of search I'm getting first image and then later pages. It is possibe to change search setting? I would like to have first pages and then images. If there is not possible to change setting. Could you please tell me how I can turn off the image search?

194.11.254.13210:50, 27 July 2015

## Log has not be created for today

using php 5.3.4

added the page curation extension, and when I try to nominate a page for deleteion and give a detail description..I get a popup that reads "log has not been created for today"

can someone help?

You did run "update.php" after adding? See Extension:PageTriage

07:04, 27 July 2015

Hi,

I just installed the latest version of mediawiki and the default footer links are not really what I want.

1st How do I make it not a string? I have been through some guides but they were all for some older versions and php files/sections they refer to either do not exist anymore or contain completely different code (I am not an IT guy, I have to idea how to code so I have to rely on examples/guides) :s

2nd Where do I change the actual text of this "This page was last modified on 23 July 2015, at 13:30", I think it's too long so I would like to make the text shorter (not to mention this "Content is available under GNU Free Documentation License 1.3 or later unless otherwise noted.").. my footer is only 700px wide, those two things get wrapped almost twice..

3rd How do I add a new link/text, for example, how many times a page has been viewed. In an older version I had something like "This page has been accessed 153,112 times." but now I do not get that at all, is it because of a different skin (currently Monobook)?

83.6.207.15913:11, 25 July 2015
07:00, 27 July 2015

## delete watchlisttoken

Hi, im using MediaWiki 1.24, PHP 5.3, MySQL 5.1.72 and VectorSkin on my Site (matura.marienberg.at )

How can I delete the watchlisttoken (the white/blue star next to the searchbox) for all users who are logged in. I've already hidden the watchlist-Link in the personal-url with css (MediaWiki:Common.css) but I can't find how to get rid of this token.

Thanks, Ray

10:36, 25 July 2015

#ca-watch { display:none!important; }


will hide the star symbol just as you hide the other tabs currently.

88.130.83.24612:51, 25 July 2015

Thanks a lot!

1. ca-watch { display:none!important; }
2. ca-unwatch { display:none!important; }

the token was gone.

91.114.194.14406:48, 27 July 2015

I don't know what the proper name for the #ca-watch star is, but I'm almost sure the real "watchlist token" is the one described in MediaWiki:Prefs-watchlist-token.

14:03, 25 July 2015

What my CSS hides is the star symbol.

What Ricordisamoa‎ speaks about is a text, which shows up in the user preferences (I believe?). If you in fact want to hide options in the user preferences, then you should not use CSS to do that, but you should use the configuration variable $wgHiddenPrefs in the LocalSettings.php file! $wgHiddenPrefs[] = 'watchlisttoken';


in LocalSettings.php should hide that entry (according to the docs, but untested).

88.130.83.24614:23, 25 July 2015

## Memory Leak? Memory running out.

Hi there, I noticed my media wiki is absorbing a lot of memory especially mysqld.bin.I am not sure why. I have 2gb ram and 1 cpu allocated to the VM for my mediawiki. At the same time, the nodejs has 4 workers. I am not sure why either. the only thing using the nodejs i though would be the parser.

The installed software i am using are: MediaWiki 1.25.1 PHP 5.5.25 (fpm-fcgi) MySQL 5.6.23

The only extensions i have are: Admin Links category Tree launchExternal ParserFunctions DiscussionThreading Universal LanguageSelector VisualEditor WikiEditor

I have a snapshot of the top: http://ctrlv.in/609700

Hope to hear from someone.

22:15, 24 July 2015

mysqld is the Mysql database server, not MediaWiki (MediaWiki is handled by php and your webserver). You may need to configure mysql server properly according to your available memory.

Note that databse servers usually need to consume more memory to buffer query results and data that's accessed more frequently, so retrieving that data doesn't require to fetch it from the disk (which would be slow). Reducing the available memory for the database may make database access slower.

13:32, 25 July 2015
I guess what configuration should I setup for mysql? Or how can I limit it to 1gb? Rather than dynamically expanding?

14:51, 25 July 2015

You should probably document yourself about those settings, but there's http://www.mysqlcalculator.com/ (not sure how accurate is it).

16:37, 26 July 2015

## No links in the mobile view

I recently created Speedypedia [www.speedypedia.info], which is based on MediaWiki. I noticed that in the mobile view, links to articles which don't exist yet are removed, probably with JavaScript (for example, in the article about Haggai Matar there is no link to Anat Matar, and I think it changes after the page loads). Is it possible to remove this feature and display the links in the mobile view, like in the desktop view?

Thanks, Uri.

10:17, 23 July 2015

11:54, 23 July 2015

Add the following to your LocalSettings.php after the require_once line of MobileFrontend:

$wgMFShowRedLinks = true;$wgMFShowRedLinksAnon = true;


That's it.

15:53, 23 July 2015

Thank you, it worked! what are these options? @Fried, the red link only didn't appear in the mobile view, not in the desktop view.

16:16, 23 July 2015

Redlinks are deactivated in mobile for performance reason and because the mobile editor wasn't a really good function to welcome users (especially newcomers). So we deactivated them and added these two options to activate them, e.g. if third party wikis want to use them. The first one is for logged in users (if set to true, logged in users always see redlinks, but logged out users doesn't) and the second one is for logged out users.

Btw.: In the next release (1.26), redlinks will be enabled by default with a Click-to-action banner, if the user really wants to create a new page or not (redlinks are well known in the editing community, but a new user maybe doesn't know what it is, and the mobile editor doesn't have the nce warnings, like on desktop).

18:52, 23 July 2015

I have another problem - in the mobile view, if I search for an article which doesn't exist yet, I don't have a link to create this article (I only get this link in the desktop view). Is it also a feature we have to enable?

16:54, 23 July 2015

No, that is disabled by design. The mobile editor is a plain wikitext editor, which is a worse experience for newcomers. So, the eitor isn't really suitable for creating pages, too and we've disabled the create link for search. If you really want to enable it, you could do the following:

.mw-search-createlink {
display: inherit !important;
}

• Save and reload the search page (maybe you have to clear your browser cache and maybe any other existing cache layers, depends on your setup)
18:49, 23 July 2015

OK, thank you! I'll try.

05:20, 24 July 2015

Thank you, it worked!

12:49, 24 July 2015

12:55, 24 July 2015

Yeah, the problem is, that we never tested MobileFrontend with wikis, where you have to register to edit :) MF (=MobileFrontend) show this error message, whenever the page can not be edited by the current user, there is no configuration to handle this. Would you like to file a task in Phabricator for it? (Please note, that you're using REL1_25, so the WMF team(s) can decide what to do, normally we don't support other branches as master :)).

17:32, 24 July 2015

Task created, https://phabricator.wikimedia.org/T106943?workflow=create I don't understand what you mean REL1_25? If you want to know my version of MediaWiki, go to http://en.speedypedia.info/Special:Version (MediaWiki 1.25.1).

13:25, 25 July 2015

REL1_25 is the 1.25.X release branch in our git repository, it reflects the latest release of MediaWiki 1.25 :)

14:50, 26 July 2015

## Single Sign On - LDAP Extension

HELLO!

Need some help here. Appreciate the help from all subject matter experts. So right now, I am trying to implement auto login to my mediawiki site. Currently, users are able to login to mediawiki using their windows account and password.

Next, I would like to implement the auto-login function (without them entering their credentials). Below is my code.

require_once( "$IP/extensions/LdapAuthentication/LdapAuthentication.php" ); require_once( "$IP/extensions/LdapAuthentication/LdapAutoAuthentication.php" );
//$wgAuth = new LdapAuthenticationPlugin(); #Commented out for autologin$wgLDAPAutoAuthDomain = "MYDOMAIN"; //uncomment for auto login
$wgDiff3 = ""; # Authenticate against Active Directory$wgLDAPDomainNames = array( "MYDOMAIN" );
$wgLDAPServerNames = array( "MYDOMAIN" => "MYDOMAINSERVER" );$wgLDAPSearchStrings = array( "MYDOMAIN" => "USER-NAME@MYDOMAIN" );

$wgLDAPEncryptionType = array( "MYDOMAIN" => "clear" );$wgLDAPUseLocal = false;
$wgMinimalPasswordLength = 1;$wgLDAPBaseDNs = array(
'MYDOMAIN' => 'cn=Users,dc=domain,dc=com'
);

$wgLDAPSearchAttributes = array( 'MYDOMAIN' => 'sAMAccountName' );$wgLDAPGroupUseFullDN = array( "MYDOMAIN"=>true );
$wgLDAPBaseDNs = array( 'MYDOMAIN' => 'dc=MYDOMAIN,dc=example,dc=com' );$wgLDAPSearchAttributes = array( 'MYDOMAIN' => 'sAMAccountName' );
$wgLDAPGroupsUseMemberOf = array( "MYDOMAIN" => true ); //The objectclass of the groups we want to search for$wgLDAPGroupObjectclass = array( "MYDOMAIN"=>"group" );

//The attribute used for group members
$wgLDAPGroupAttribute = array( "MYDOMAIN"=>"member" ); //The naming attribute of the group$wgLDAPGroupNameAttribute = array( "MYDOMAIN"=>"cn" );
$userid =$_SERVER['REMOTE_USER'];
$wgLDAPAutoAuthUsername =$userid;
AutoAuthSetup();


Once I implemented the above, the domain name in my dropdownlist become empty and the username and password I entered denies my entry into the site although it is correct.

Why is this happening? Please assist! Greatly appreciated it!

THANKS ALL SO MUCH!

Just to add on ,the server that I'm running is Apache 2.2

175.156.223.20916:17, 16 July 2015

You will have to install and configure kerberos for apache. Once you do that, you can use the [Extension:Auth_remoteuser Auth_remoteuser] extension to handle authentication.

MarkAHershberger(talk)21:49, 18 July 2015

Hi Mark,

May I further clarify with you by installing and configuring kerbors for apache. Do you mean this website? (https://www.mediawiki.org/wiki/Extension:LDAP_Authentication/Kerberos_Configuration_Examples#Apache_configuration)? Or I have to install the mod_auth_kerb ?

Another issue is so this LDAP extension can be used together with https://www.mediawiki.org/wiki/Extension:Auth_remoteuser to perform single sign on?

mod_auth_kerb shouldn't be needed with IIS. That is what I would try first on Windows.

MarkAHershberger(talk)21:21, 25 July 2015

Or are you advising me to use Auth_remoteuser extension instead of LDAP extension? Thanks!

If you just need SSO authentication, then you should use Auth_remoteuser.

MarkAHershberger(talk)21:18, 25 July 2015

I've tried to install and configure mod_auth_kerb on my Apache Server. But I am using Windows and the information I googled tends to be more towards Linux. Any help please? Thanks!

203.120.42.15101:06, 20 July 2015

If you are using IIS, SSO should already work, right? And then you would need something like Auth_remoteuser to get the logged in user.

MarkAHershberger(talk)21:20, 25 July 2015

## Open source exam and quiz banks: An idea for a new wiki

I am a college professor of physics who is concerned about the high cost of education. In the technical fields (math and science) there in more than enough information on the internet-at-large for a self-learner to acquire more than 50% of a four year college degree. Although the wikis (e.g. Wikipedia and Wikiversity) are still underdeveloped compared to the internet-at-large, we are approaching the point where this knowledge will also be available in a convenient form that is open source.

What is missing is a means by which a self-learner can easily document expertise in these subjects. Test banks are proprietary and carefully guarded by the publishers of textbooks.

Wikiversity:Quizbank represents my effort to fill this need. The idea is to post these questions publicly and use software to first store the questions on a personal computer, and then to create wikitext that can be uploaded to a wiki and printed out as pdf files. I have already done that with a Matlab code that randomly selects questions and renders them in wikitext with pagebreaks such that all the multiple choice options are on the same page as the question. I am not a computer programmer and my software is hard to use.

Here are some challenges that need to be addressed in the future, if this idea is to take off:

1. The Matlab software that retrieves and randomly selects questions from a copy of the quizbank that resides on a personal computer creates textfiles that must be uploaded to a wiki and then rendered as wikitext that can be printed. At the moment this is a public action. We need to arrange for these exams to be written and printed out in secret. I am able to do this because I am a Wikiversity custodian who can hide edits soon after they are made. And I make my exams many months in advance, and have only a small number of students. It is highly unlikely they would monitor my efforts and attempt to capture an exam as I write it. (Also, why cheat if the questions are already available?)
2. Teachers need to write test questions differently. One issue is the reluctance of students to ready anything. I tried to write tests based on Wikipedia articles, but found that all but the most dedicated students preferred to simply study the testbank.
1. One remedy is to select two or three questions from each exam and ask students to explain them in a short paragrah. ** Another remedy is to include written passages as part of the question. A large number of questions would be associated with that passage. As students prepare for the exams by reading the passage, they might find it easier to understand the passage than to attempt to memorize all the answers.
2. Another remedy is to require that students propose questions. The grading of such efforts is quite useful for any instructor who is a contributor to this bank. A related homework exercise for students is to verify that questions written for a specific Wikipedia article are still valid after the article has undergone a series of edits.
3. Another relatively trivial challenge is making the open source exam bank compatible with the conventional "secret" banks already available. The expectation is that colleges and universities would use the open source testbank as only a portion of the grade assessment and certification. There will always be room for those "secret" questions that only the instructor knows about.

There is no great rush to solve these problems. I am still struggling to get a small version of the quizbank prepared for the four courses that I teach. Serious work needs to be performed after I have established that this is a good way to teach. But if anybody has any suggestions, please let me know. I am only vaguely aware of what sort of extensions can be written because I am not really a programmer.

16:00, 22 July 2015

Hi Guy,

for that you write that you are not a programmer, you obviously have already made up big part of the software needed. You are writing much, but you do not have clear questions (yet) - so let me give you some thoughts:

You say that certain questions (and their answers obviously) need to be secret. Hiding a revision from view is one way to do that, and it might not even be the worst. If you set up an own wiki for that stuff, kind of your own one, then you could fully control what is going on there. E.g. you could take people's read access away from certain pages. However, MediaWiki is not made for that and you cannot regard that as "secure". There will most likely always be a way around such limitation. Making the complete wiki private would however be possible (and secure if you ask me).

If I understand correctly you or some co-workers will prepare tests. This should also continue that way - you do not want a computer algorithm to randomly pick questions for you. The Matlab software creates text files, which you are uploading to the wiki. Since I guess you know (or can control) the exact format, in which questions and answers are presented in these files, it should be rather easy to process them further. But if I understand you correctly, your software is already able to produce wikitext out of them.

88.130.105.13218:03, 22 July 2015

Thanks for responding. Let me give you two explanations and ask you a question:

The explanations:

1. I chose a program that randomly selects from the quizbank because that was easier to write. It works OK for the time being.

2. The "secret" question bank I refer to are those proprietary banks that professors keep on their hard drives. These questions will always play a role in grade assessment.

My question is about "secret" exams from a "public" (open source) question bank that need to be rendered by first uploading to Wikiversity, and then printing out as a pdf file. At the moment this is a "public" act that anybody can see.

I don't need to do this now, but someday a larger open source organization (such as openStax College) might want to create tests "off-wiki", meaning that it is not necessary to go through Wikiversity or Wikipedia to render the wikitext into a pdf file. The solution would be to maintain a wikitext server with access that is by invitation only. I need to know if this feasible.

I have noticed that two by-invitation-only wikis are at UC Davis with Chemwiki and the OpenStax College textbooks (affiliated with Rice University). These environments are much different than wikimedia "wikitext" syntax, using mathjax for equations, for example.

Someday (maybe in a decade or so) I would like to see it possible for professors to render exams in wikitext in a way that nobody can see it until the day of the test. The wikitext would be created on personal computers, and ultimately rendered as pdf for paper-copy printout. If such "private" rendering of wikitext is feasible, I will continue to write exam questions on Wikiversity with the understanding that the technology will eventually catch up.

On the other hand, if Wikimedia is adamantly against supporting any wiki that permits "secret" or "private" rendering of wikitext, and if a private server cannot be created that renders equivalent Wikipedia-style wikitext, then I need to think about working with the syntax used by OpenStax College or Chemwiki. I have contacted both organizations, and they are both giving me moral support. But neither organization seems to have the resources to create the software I hope will happen sometime in the future.

Let me put the question simply with a hypothetical scenario: Suppose I were not an individual, but a consortium of dozens of professors who need a way to render wikitext in a fashion that nobody can see it until the day of the test. Would Wikimedia (or somebody else) be able to provide that support?

I believe the answer is yes.

01:10, 23 July 2015
Suppose I were not an individual, but a consortium of dozens of professors who need a way to render wikitext in a fashion that nobody can see it until the day of the test. Would Wikimedia (or somebody else) be able to provide that support?

It would be a SMOP, yes. You can set up your own copy of Wikiversity and and individual like myself or someone on the list of people on professional development and consulting would be able to help you.

MarkAHershberger(talk)21:15, 25 July 2015

## Cannot edit Common.css (or any MediaWiki:* page)

I have a wiki up and running (www.makehumancommunity.org/wiki if you're interested). As far as I know everything works as expected, and as the admin I'm able to use all special pages and edit all common pages. Except the ones in the MediaWiki namespace. Specifically, I can't edit MediaWiki:Common.css.

As soon as I navigate to that page, the interface looks as if I'm not logged in: The "log out" link changes to "log in", there is no edit tab etc.

Now, I'm pretty sure I caused this, either with the phpbb authentication plugin or with the plain url settings in apache, the latter which look like this:

 RewriteEngine On
RewriteCond %{REQUEST_FILENAME} !-f
RewriteCond %{REQUEST_FILENAME} !-d
RewriteRule ^/?wiki(/[^/]*)?$/home/joepal/mediawiki/index.php [L] RewriteRule ^/?wiki(/Special.*)?$ /home/joepal/mediawiki/index.php [L]
RewriteRule ^/?wiki(/MediaWiki.*)?\$ /home/joepal/mediawiki/index.php [L]
Alias /wiki /home/joepal/mediawiki


... but, can anyone suggest where/how to start debug this?

14:28, 25 July 2015

Your RewriteRules look strange; they will do funny things and they are not standard. The alias directive will only work if used inside a server config or virtual host file - it will not work in .htaccess.

88.130.83.24615:02, 25 July 2015

## Problem calling a special search page from the standard page search box

I previously implemented a special search page for a MediaWiki 1.17 install which works fine.

I now tried the same with the latest MediaWiki 1.25.1. The page by itself also works fine, except that it cannot be called by the standard page search box. I believe the problem lies in the handling of "search' in the parameters to index.php, i.e. from the docs:

"search - if non-blank (and if searches are enabled), override all other parameters and perform a search"

When the usual page search box is used, it generates an url of the form:

 localhost/mediawiki/index.php?title=Special:MySearch&search=test&go=Go


This url itself has a "search" parameter to index.php, so it is called repeatedly, ending in "This web page has a redirect loop". How can I prevent this and get the url to load the Special:MySearch page?

96.50.92.10206:52, 25 July 2015

## How do I log in the user to two different domains?

I want the user to be logged into h4x0rs.net and 6c.nz at the same time. How do I make this work?

14:47, 24 July 2015

## CreateWiki (WikiFarm)

Hello, we at Orain have an extension called CreateWiki that allows us to create wikis. Before, we had this sort of "hack" where there would be a forced user-account created on the local wiki so the local permissions could be assigned. And I do not know why this doesn't work anymore. Here is the GitHub.

04:48, 16 July 2015

Ok, is there _any_ error message and what exactly does not work anymore? What does the log say?

23:05, 18 July 2015

@Florianschmidtwelzow: When I create a wiki it does get created but the local account of the user doesn't get created. There is no error message. this doesn't work.

14:07, 24 July 2015

 First page Previous page Next page Last page