MediaWiki-l January 2010

mediawiki-l@lists.wikimedia.org

81 participants
75 discussions

Mysql ISAM vs InnoDB
by Jim Hu 05 Dec '11

05 Dec '11

$wgDBtransactions gets set to true if using InnoDB tables. Is there an advantage to using InnoDB tables? The disadvantage is that with MySQL there is a file, ibdata1, that seems to grow endlessly if InnoDB tables are used. See http://bugs.mysql.com/bug.php?id=1341 We're wondering if we should just convert everything to MyISAM. Any thoughts? ===================================== Jim Hu Associate Professor Dept. of Biochemistry and Biophysics 2128 TAMU Texas A&M Univ. College Station, TX 77843-2128 979-862-4054

4 4

Custom namespaces and discussion pages
by Snowolf 24 Jan '11

24 Jan '11

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi all, I've created some custom namespaced on one of my wikis, Botwiki (previously known as pywikipedia). I've put these lines in my LocalSettings.php file: - --- #Custom namespaces $wgExtraNamespaces = array(100 => "Manual", 101 => "Manual talk", 102 => "Python", 103 => "Python talk", 104 => "Php", 105 => "Php talk", 106 => "Perl", 107 => "Perl talk", 108 => "AWB", 109 => "AWB talk", 110 => "IRC", 111 => "IRC talk", 112 => "Other", 113 => "Other talk" ); $wgContentNamespaces[] = 100; $wgContentNamespaces[] = 102; $wgContentNamespaces[] = 104; $wgContentNamespaces[] = 106; $wgContentNamespaces[] = 108; $wgContentNamespaces[] = 110; $wgContentNamespaces[] = 112; - --- However, I have a big problem: when I go to a page in one of these new namespaces (not the discussion, the main ones), for example http://botwiki.sno.cc/wiki/Perl:Copyright_Violation_Bot , I found the red link to the discussion page. It's right, as there is no discussion page for that article. But if you click on it, it brings you to http://botwiki.sno.cc/w/index.php?title=Perl_talk:Copyright_Violation_Bot&a… correct, of course. But have a look of the article and discussion tabs: they are both red! The first, "article", leads to http://botwiki.sno.cc/w/index.php?title=Perl_talk:Copyright_Violation_Bot&a… when it should lead to http://botwiki.sno.cc/wiki/Perl:Copyright_Violation_Bot and the second, "discussion", leads to http://botwiki.sno.cc/w/index.php?title=Talk:Perl_talk:Copyright_Violation_… , when it should lead to http://botwiki.sno.cc/w/index.php?title=Perl_talk:Copyright_Violation_Bot&a… . It's the first time I deal with custom namespaces :-( but I have some ideas of what it can be. Can the problem be with the $wgContentNamespaces settings? So it detects everything as ns0? (don't think so). Or can it be the fact that I haven't used an underscore in the $wgExtraNamespaces definition? Snowolf -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.7 (MingW32) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFGWhk7sdafW5NQMtERAuX+AKDQ7QLNjXv9cu+ZbSLXidMzgi/vNgCaA7VT +VTgR3iI/BI7FVDqcyRZVJ0= =a4yP -----END PGP SIGNATURE-----

2 2

Easy creation of page sequences
by Jean-Marc van Leerdam 28 Jul '10

28 Jul '10

Hi, I have been trying to create a maintainable way of grouping pages together and allowing readers to page through them in sequence. Most samples I've seen use templates that require you to supply the 'previous' and 'next' page. This however results in three page edits to insert a new page between two existing pages and does not guarantee that your prev-sequence is identical to the next-sequence... Being a programmer, that is way too much duplicate and error-prone work for me ;-) There must be a better way to do this and I was hoping to solve it with a plain MW installation (with ParserFunctions and DynamicPageList at the moment). Given an 'index' page that holds a list of all page titles in the preferred order, isn't it possible to create a template that selects the correct previous and next page title given the current page title? I get stuck in getting the correct lines from the index page. DPL can select based on section name (= page title), but then the contents of that index section must be the prev/next links themselves: =First Page= Prev [[Second Page|Next]] =Second Page= [[First Page|Prev]] [[Third Page|Next]] =Third Page= .. etc. This works, except that it is still a lot of duplication of page names (but the edits are contained in a single page, big plus). I hoped to simplify the index page by creating a template that writes the section header and prev/next links, but then DPL no longer recognizes the sections :-( Apparently DPL 'sees' the page text before the templates are called ({{Page|Prev page|Page title|Next page}}: {{Page||First Page|Second Page}} {{Page|First Page|Second Page|Third Page}} {{Page|Second Page|Third Page|Fourth Page}} Basically my questions are: 1. Am I completely off track here? 2. Can DPL be coerced to evaluate templates before looking at the page 3. Can DPL (or another extension) select the text from a section before/after a matched section? 4. Is it possible to determine the section sequence number given the section name (so 'Second Page' results in 2, allowing me to use DPL to retrieve the name of section 2 -/- 1 and 2 + 1 to create the prev/next links? Apologies for the long post, hopefully someone can point me to some good resources (I've been to Meta, Wikibooks, Medawiki.org but could very well have overlooked something there as the amount of info is a bit overwhelming and it is difficult to judge how up to date it is). -- Regards, Jean-Marc

11 16

Re: [Mediawiki-l] Please convince me to make my software open source (3)
by Giuseppe Briotti 27 May '10

27 May '10

Hi all, hi Bernard... probably, my response is a bit late, but I spent time reading (hopefully!) carefully all the answers... Thus, my 0,02$: I think that probably the correct question is: "Please convince me to make my business model" :-) I agree with others that probably, the SaaS business models require closed source. Thus, probably, no choice... IMHO the open source software implies a different business model, something similar to "partnership". Good luck, honestly, Giuseppe Briotti. -- Giuseppe Briotti g.briotti(a)gmail.com "Alme Sol, curru nitido diem qui promis et celas aliusque et idem nasceris, possis nihil urbe Roma visere maius." (Orazio)

4 4

wiki family and shared resources
by Mij 07 Feb '10

07 Feb '10

Hello everyone, We want to implement a MW-based WikiFamily with a reserved "pool" instance, very similarly to Wikipedia, but we're encountering some problems with the media sharing. We followed this process: 1) we started from - three language instances - containing about 1000 images altogether - configured with Wikipedia-style URL scheme: http://<lang>.domain.com/wiki/<Node> 2) we installed and set up one "pool" wiki instance, according to "Scenario 5" of http://www.mediawiki.org/wiki/Manual:Wiki_family 3) we modified maintenance/importImages.php to import automatically from any language instance to the "pool" all images along with metadata (description, author etc) The pool contains all medias of all language instances, and the "Upload file" link in the toolbox of each wiki points to the pool's Special:Upload correctly. The relevant configuration for the language wikis is as follows: === $wgUploadNavigationUrl = "http://pool.domain.com/wiki/Special:Upload"; $wgUseSharedUploads = true; $wgSharedUploadPath = 'http://pool.domain.com/wiki/File:'; $wgSharedUploadDirectory = '/var/www/wikis/wiki-pool/images/'; $wgHashedSharedUploadDirectory = true; $wgEnableUploads = false; === However, trying to enforce the "pool fallback lookup" documented http://www.mediawiki.org/wiki/Manual:$wgUseSharedUploads we tried moving away the instance's own images/ directory. Doing this, all images result into a broken link. We suspect the problem is in a misconfiguration of $wgSharedUploadPath , which isn't quite clear from the WikiFamily example or its own documentation, but cannot track down the problem more specifically. Notably, we observed that none of the includes/filerepo/* files is loaded. Can anyone help out with nailing down this problem? Thanks!

2 6

Making only a part of the content visible for a group of users
by Maximiliano Milicich 02 Feb '10

02 Feb '10

Hello, We've recently installed a wiki for IT dept usage but came to the conclusion that it would be very useful to let end users see some of the articles. The thing is we don't want them to see all the content mainly because it would confuse them. For instance, if they search for "Accounting" we want to provide definitions and troubleshooting on the Accounting system that might help them, but not every server list, SQL code and tech spec that we have. We've been looking around and found a couple of options based on namespaces, which we're still not sure will work the way we need (for one thing, limiting search results for end users) and seem kind of hard to maintain. The simplest option seems to be keeping two separate wikis, but I don't want to disconnect the end users documents from our wiki because it's valuable info for us too. What we're thinking about is having both wikis but maintaining them in sync with some batch process that could export and import all articles that belong to a certain category or even backup and restore the entire database and afterwards delete non-end user articles. That way, the end user wiki would be read-only and get updates every day. In both cases we'd have to know how to deal with images (seems easy with the backup/restore option but no so much with the export / import). Do you think we're on the right path with any of these? Any other suggestions? Here is our version info: Product Version MediaWiki 1.15.1 PHP 5.2.10 (isapi) MySQL 5.0.22 Thank you in advance!!

3 4

Re: [Mediawiki-l] Image files too big for the thumbnail program
by Jan Steinman 01 Feb '10

01 Feb '10

> From: Ron Laufer <gonzoron(a)hotmail.com> > > I've had a couple images uploaded to my wiki that are so big, it > seems they make my thumbnail program crash. Are you using ImageMagick? We've had no problems using the biggest files a 12 megapixel camera can produce. ---------------- An American is a person who demonstrates against a new power plant, then goes home and flips on all the lights, turns up the air conditioner, puts a tape in the stereo, opens the refrigerator door, plugs in the coffee maker and sits down to see if the television cameras caught him protesting. -- Wendell Trogdon :::: Jan Steinman, EcoReality Co-op ::::

3 2

api.php action=edit problem badtoken
by Tim McKenna 01 Feb '10

01 Feb '10

Hi, I went through the sample steps in the API example<http://www.mediawiki.org/wiki/API:Edit_-_Create%2526Edit_pages#Example> on using the api to edit a page: 1. login your bot: ( I just used my admin login) - it succeeds 2. get a token - it suceeds 3. post to api.php - this is the form I use<http://pathboston.com/forms/mw/append21.html>- after I get the token I replace the +\ with %2B%5C Alas, I get this error: <?xml version="1.0"?> <api> <error code="badtoken" info="Invalid token" xml:space="preserve"> my form code is: <form id="form1" name="Update" method="post" action="../../hum310/api.php?action=edit"> <label>title: <input type="text" name="title" id="textfield" value="Talk:Main_Page" /></label><br /> <label>section: <input type="text" name="section" id="textfield" value="new"/></label><br /> <label>summary: <input type="text" name="summary" id="textfield" value="Hello%20Worldwide"/></label><br /> <label>text: <input type="text" name="text" id="textfield" value="Hello%20everyone!"/></label><br /> <label>watch: <input type="text" name="watch" id="textfield" /></label><br /> <label>basetimestamp: <input type="text" name="basetimestamp" id="textfield" value="2008-03-20T17:26:39Z"/></label><br /> <label>token: <input type="text" name="token" id="textfield" value="96a50d37e6c48e6d0def3dd64dc9450d%2B%5C" size=50/></label><br /> <input name="" type="submit" value="send" /> </form> <a href="http://pathboston.com/hum310/api.php?action=query&prop=info|revisions&intok…">get token</a> Thanks, Timothy S. McKenna Humanities teacher Parkway Academy of Technology and Health, a Boston Public High School 1205 VFW Parkway, Boston, MA 02132 rm404 tim(a)sitebuilt.net, http://sitebuilt.net exp. course news on twitter, id: mrtmckenna class wiki sites: http://pathboston.com/hum Humanities 3 http://pathboston.com/poets class159 (857)498-2574 (mobile), (617)524-0938 (home)

2 1

Please help to make a rewrite rule
by Michael Wozowoy 29 Jan '10

29 Jan '10

Hello! I have Mediawiki 1.15 installed with Apache 2.2. I have created some virtual namespace (similar to categories) called Rubric. The articles can belong to one ore more rubrics. I want to use this namespace instead of Categories by some reasons. And I need to show these Rubrics' content. For example: http://mysite.com/wiki/Rubric:Economy/ should open a list of articles under Economy rubric. I've created a special page which does that. So, now I can open http://mysite.com/wiki/Special:Rubric/Economy and view Economy's content. It works, no problem. But I want to get rid of Special namespace in the URL! Can I do that? I've tried several RewriteRules in the apache's .htacces, but failed so far... The last try is: ============================================================================= RewriteEngine On RewriteRule wiki/Rubric:(.*)$ wiki/Special:Rubric/$1 RewriteRule wiki/(.*)$ w/index.php?title=$1 [L] ============================================================================= It does not work. When I'm opening http://mysite.com/wiki/Rubric:Economy/ this rule still opens "Rubric:Economy" instead of "Special:Rubric/Economy" and says "there is no text on the page blah-blah-blah..." What do I do wrong? NOTE: I need an internal redirect. The external one works, but that's not what I want. I mean, this one: RewriteRule wiki/Rubric:(.*)$ wiki/Special:Rubric/$1 [R,L] works great of course, but it results in that the Special namespace presents in the URL. Best regards. Michael Wozowoy e-mail: strannik(a)cnt.ru

1 0

Re: [Mediawiki-l] Style pages based on category, possible?
by Jean-Marc van Leerdam 29 Jan '10

29 Jan '10

Bernard, On 28 January 2010 14:58, Bernard(a)bernardHulsman.nl <bernard(a)bernardhulsman.nl> wrote: > Jean-Marc, you have to use > http://handleiding.wikiation.nl/MediaWiki:Common.css . That is where it is > defined. It is no standard MW feature. There is no LocalSettings involved. The Common.css is only the part where you define the background color depending on the category you use. I.e. you specify that .cat-Timecode-future results in a background: #fff7e7 That part is entirely clear to me, I use Common.css extensively. But in my mediawiki installation (1.13.3, and neither on mediawiki.org which uses 1.16alpha), there are no cat-[category name] classes being output. I see a ns-xxx and a page-yyy part in the body class, but no cat-zzz part. That is what confuses me. -- Regards, Jean-Marc -- . ___ . @@ // \\ "De Chelonian Mobile" . (_,\/ \_/ \ TortoiseSVN . \ \_/_\_/> The coolest Interface to (Sub)Version Control . /_/ \_\ http://tortoisesvn.net

2 1

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

MediaWiki-l January 2010