Help talk:Magic words

{{REVISIONxxx:Foo}}

Is there any way to get revision information using the magic words for another page such as getting {{REVISIONUSER:Foo}} or {{REVISIONTIMESTAMP:Foo}} from [[Bar]]? This would be useful for keeping track of articles in a project for a list of pages. Technical 13 (talk) 15:44, 22 May 2013 (UTC)

No. This can be handled by a DynamicPageList extension, though. --Ciencia Al Poder (talk) 09:11, 23 May 2013 (UTC)

+1 for feature request -- sir KitKat 78.21.21.68 07:38, 15 October 2013 (UTC)

• Yes, this has worked for some time now. Guess no-one has patrolled this page in a while... Foo was lasted edited on (01:22:16 on December 21, 2024) by ([[User:]]). Technical 13 (talk) 19:53, 10 November 2014 (UTC)

_NOEDITSECTION_ on every page?

Hey there, I am using Mediawiki and I have problems with partly editing of pages, so I want to disable the "edit section" Button for the whole wiki. Is this possible?

If there was such a thing, I would think it would be listed on Manual:Configuration settings (alphabetical), which it is not. However, depending on which version of MediaWiki you are running you could just set the class (use to be .editsection but was recently renamed .mw-editsection on some wikis like en.wikipedia) for to display: none; in your MediaWiki:Common.css -- Technical 13 (talk) 12:48, 15 June 2013 (UTC)

I'd like to be able to do this too

I don't understand the answer saying that you set \<code\> display: none; \</code\> - where would you set that?

Isn't there a switch you can set in LocalSettings.php to change the behaviour so that, by default, all sections are displayed without a edit -- so you'd, instead, need to specify __EDITSECTION__ if you wanted a particular section to display an edit.

Fustbariclation (talk) 08:32, 10 January 2014 (UTC)

• As I said above, you would add:

.editsection {
    display: none;
}

or possibly, depending on the version of MW you are running:

.mw-editsection {
    display: none;
}

to your MediaWiki:Common.css. Technical 13 (talk) 19:46, 10 November 2014 (UTC)

__DISAMBIG__

Could someone please add the necessary documentation for __DISAMBIG__ (Extension:Disambiguator) here and, if you could be so kind, also at en:Wikipedia:Magic words? Thank you. -- Michael Bednarek (talk) 13:48, 11 July 2013 (UTC)

• I was under the impression that this page was reserved for core (or at least ships with core extension) magic words and I don't remember seeing Extension:Disambiguator being in that package. Is this going to be a new inclusion in the default package? Technical 13 (talk) 10:26, 17 July 2013 (UTC)
  • When I look up magic words on the EN Wikipedia, I often end up looking at Help:Magic words here on MediaWiki, because that page is linked there, e.g. at en:Help:Magic words. My impression, which may well be wrong, is that MediaWiki documents the MediaWiki software. If that is not so, feel free to remove the entry about __DISAMBIGUATE__ again; in that case, I wonder whether it would be better to remove the mentioned link to MediaWiki at "en:Help:Magic words" as well. -- Michael Bednarek (talk) 12:21, 17 July 2013 (UTC)

Number of files in Articles

Hi, Is there any way to be able to count the number of files (In my case: Images) shown in article/category? In other words, I want to have the number of occurrence of any [[file:...]] in given article.

TNX --82.80.126.166 13:01, 23 October 2013 (UTC)

Fullurle

The magic word {{fullurle:}} is not described in this page, and I honestly don't know the difference between that and {{fullurl:}}. Cainamarques (talk) 21:32, 6 December 2013 (UTC)

Displaying the article's author on the page

Is there any magic word for that? For example {{REVISIONUSER}} displays the username/IP of the last user, who edited the article. How to display the username/IP of the user, who created the article?--185.31.48.30 09:45, 17 April 2014 (UTC)

No. That would be an expensive operation, as MediaWiki should find the first revision of the current page. --Ciencia Al Poder (talk) 16:53, 17 April 2014 (UTC)

• Requested on bugzilla:18748 in May of 2009. Technical 13 (talk) 19:32, 10 November 2014 (UTC)

I am also interested in this magic word. Like {{ORIGINALAUTHOR}} or {{PAGEAUTHOR}} and to go with them, the creation date. {{CREATEDYEAR}} and {{CREATEDMONTH}} and {{CREATEDDAY}}. That would help me with my project. Thanks. — Preceding unsigned comment added by Pvodrazka (talk • contribs) 09:18, 9 April 2019 (UTC)

You can try with Extension:ContributionCredits, or Extension:CreditsSource --Ciencia Al Poder (talk) 09:18, 9 April 2019 (UTC)

Adding translatable docu to such a page is horrifying

This diff together with translate tags etc. causes fatals, so I gave up. Sorry for this. The standard nonsense error message "Fatal exception of MediaWiki exception" or what this is called will probably not help digging into the issue. --[[kgh]] (talk) 22:32, 4 July 2014 (UTC)

anchorencode

Did somebody change anchorencode ?

It now seems to strip italic markup, rather than encode it:

{{anchorencode:Glasgow Digital Library, ''Omond House''}} -> Glasgow_Digital_Library,_Omond_House

It just means that on en-wp, we have to clean up a few hundred articles where en:template:SfnRef (which uses anchorencode to generate anchors in footnotes) has anchors that are generated from italicised words. --RexxS (talk) 16:30, 6 July 2014 (UTC)

Variables that affect behaviour

Why are {{DISPLAYTITLE:title}} and {{DEFAULTSORT:sortkey}} listed under Variables when they do not "return information about the current page, wiki, or date"? Surely, since they "control the layout or behavior of the page", they should be listed at Behavior switches. This has come up at w:en:Help talk:Magic words#Three "behavior switches" that aren't. --Redrose64 (talk; at English Wikipedia) 10:14, 17 July 2014 (UTC)

I surmise it's because "switches" have a special __FORMAT__ that doesn't accommodate parameters, whereas DISPLAYTITLE: and DEFAULTSORT: require them. Of the "variables" listed on this Help page, those that take parameters use them only to specify the object about which the "variable" should return information. As DISPLAYTITLE: and DEFAULTSORT: use parameters to set environmental states rather than report them, they are certainly a special case. Maybe they (along with NOEXTERNALLANGLINKS:) should be described as "environment variables", "behavior variables", or "behavior functions", and be placed in their own section or subsection in the page.

⋮

My suggestion at this point would be to rename the current "Behavior switches" section to "Behavior settings", move the current content of that section into a subsection called "Behavior switches", and add a second subsection called "Behavior functions" where DISPLAYTITLE:, DEFAULTSORT:, and NOEXTERNALLANGLINKS: are explained. I would also suggest that the list of "general types of magic words" in the lede be rewritten accordingly, mentioning the differences in effect, as well as in format, of the various kinds of magic words.

⋮

As an aside, I note that the keywords called "variables" on this Help page would probably be more accurately described as "functions", as they are used to return values that report some aspect of the environment, rather than to store arbitrarily-assigned values. But I guess that's a whole 'nother ball o' wax. — Jaydiem (talk) 16:56, 17 July 2014 (UTC)

Pages in category + sub categories

Is there a deeper penetrating {{Pagesincategory that will also count the pages in the subcategories? That would be wonderful. Kmath87 (talk) 22:02, 8 August 2014 (UTC)

• Not built in to the parser, Kmath87, no. You will have to use a recursive template, some kind of E:Loops, Lua (scribunto if I spelled that right), or javascript via the api if you want to have a count of all pages in a category & subcats. Technical 13 (talk) 12:14, 9 September 2014 (UTC)

What is the opposite of 21010?

If I know the page ID number, how do I find the page?

I see that, if I have the page name, the 21010 will give me the number, I'd just like to do it the other way around.

197.155.4.118 09:08, 20 August 2014 (UTC)

• The answer to your question is page with an ID of 21010. Technical 13 (talk) 19:20, 10 November 2014 (UTC)

Missing magic words

It looks like the following magic isnt listed:
{{CURRENTMONTH1}} gives 12
{{CURRENTMONTH2}} gives 12
{{LOCALMONTH1}} gives 12
{{LOCALMONTH2}} gives 12
{{ARTICLEPATH}} gives /wiki/$1
{{ROOTPAGENAME}} gives Magic words
{{ROOTPAGENAMEE}} gives Magic_words
Christian75 (talk) 07:30, 30 September 2014 (UTC)

Maybe because the first four are considered the same as

• {{CURRENTMONTH}} gives 12
• {{LOCALMONTH}} gives 12

As of today, the last two have been included on the page. Bennylin (talk) 10:14, 10 December 2014 (UTC)

Magic Word '{{!}}'

When I created the template for this magic word, I was having an issue using it with tables when I passed them in parameters for templates. I was using the <noinclude /> syntax for specifying why I created the template. One thing to note is that the '|' character must come immediately after the closing tag. Otherwise, the white space messes up the parsing of the character when being used in tables; at least for what I was using it for. A newline simply doesn't cut it.
--Erutan409 (talk) 14:48, 5 October 2014 (UTC)

I have the same issue, have you found the solution?... Crystian

Proposition of text changes and errors.

• In {{CASCADINGSOURCES:page name}} page name is nor translatable.
• Where says:
  • "exactly as formatnum formats them with the wiki's locale" must be "exactly as formatnum formats them with the wiki's locale"
  • "are changed; formatnum will only transform" must be "are changed; formatnum will only transform"

--Jmarchn (talk) 06:26, 10 October 2014 (UTC)

Magic word print username??

There is some word that prints the username of the user who is seeing some page where is that MAGICWORD?-- Crystian Marquez

Crystian Marquez, Did you find any solution? --Mavrikant (talk) 16:34, 22 February 2015 (UTC)

Creation date

How can i get the "creation date" of a page? ... There exist some MAGICWORD?... or there exist a parser for get the creation date of a page?

I don't think that's possible by current Magic words. Bennylin (talk) 10:01, 10 December 2014 (UTC)

{{filepath:file_name.pdf}}

How to create a filepath for a specific page in a PDF or DJVU? It only accept one parameter, thumb size, and right now it only works for the first page of the PDF. Thanks. Bennylin (talk) 08:54, 10 December 2014 (UTC)

My current workaround on includes\parser\CoreParserFunctions.php (MW 23.1) is by adding the third argument ($argC) and one line of code. So now I could call the thumbnail for page 2 of the PDF {{filepath:file_name.pdf|100px||2}}. I'm not sure this is the best solution though. Bennylin (talk) 09:58, 10 December 2014 (UTC)

	public static function filepath( $parser, $name = '', $argA = '', $argB = '', $argC = '' ) {
		$file = wfFindFile( $name );

		if ( $argA == 'nowiki' ) {
			// {{filepath: | option [| size] }}
			$isNowiki = true;
			$parsedWidthParam = $parser->parseWidthParam( $argB );
		} else {
			// {{filepath: [| size [|option]] }}
			$parsedWidthParam = $parser->parseWidthParam( $argA );
			$isNowiki = ( $argB == 'nowiki' );
		}

		$parsedWidthParam['page'] = $argC;

Magic Word for IP address lookup

Are there any magic words to look up someone's IP address? 47.20.96.42 19:53, 28 December 2014 (UTC)

There is a special user right to get this data at all (checkuser or oversight, I always forget what is what). IOW, no. –Be..anyone (talk) 11:47, 29 December 2014 (UTC)

language magic word

It would appear that {{#language:<code>|en}} where <code> is an ISO 639-3, three letter language code, either doesn't work as it should or that the documentation, such as it is, that suggests that ISO 639-3 language codes should work, is incorrect:

{{#language:ara|en}} → ara

As a check, switching to ISO 639-1 works:

{{#language:ar|en}} → Arabic

So, the questions are:

1. Should {{#language:}} return a language name for three character ISO 639-3 language codes?
   1. If no, shouldn't the documentation state that {{#language:<code>|en}} only supports ISO 639-1?
   2. If yes, where is the error and how does it get fixed?

—Trappist the monk (talk) 19:09, 10 January 2015 (UTC)

1.1 no: The docu shouldn't say only, e.g., {{#language:tlh|en}} gives Klingon. Some ISO 639-3 three letter codes have older ISO 639-1 two letter codes. From the work on RFC 4646 (predecessor of RFC 5646) I recall that ar and zh were the worst cases. Check out section 2.2.2 in 5646 about "macrolanguage zh". Admittedly arb also doesn't work, maybe the logic is "support only the shortest code". –Be..anyone (talk) 07:28, 13 January 2015 (UTC)

I'm confused. We aren't talking about language sub-tags; the documentation for {{#language:<code>|en}} makes no mention of sub-tags. The documentation says that it "[returns the] full name of the language for the given language code" where 'language code' is vaguely defined as ISO 639-3. Because ara is a legitimate ISO 639-3 code (see at registration authority: ara) {{#language:ara|en}} should return Arabic which the language name assigned by the registration authority, shouldn't it?

—Trappist the monk (talk) 14:01, 14 January 2015 (UTC)

ISO 639 macrolanguage != language, if there is a problem it would be arb, not ara. –Be..anyone (talk) 08:45, 15 January 2015 (UTC)

Missing magic word

I just figured out that {{CURRENTMONTH1}} outputs the current month in a non-zero padded format (shown here: 12). The only problem is I am unsure of how to add it to the table without breaking the <translate> tags, as each one is numbered.

--KnightMiner (t/c) 03:33, 22 January 2015 (UTC)

• I just added it without a number. Seems to render fine. — {{U|Technical 13}} (e • t • c) 03:44, 22 January 2015 (UTC)

Failed to mark for translation

Failed to mark the page for translation:

Function: MessageGroupStats::clearGroup
Error: 1205 Lock wait timeout exceeded; try restarting transaction (10.64.16.27)

— Preceding unsigned comment added by Shirayuki (talk • contribs)

EXPENSIVE tag

What does it mean when a magic word is tagged as 'expensive' ([EXPENSIVE])? For example, the PAGESINCATEGORY magic word. In the description column, it says:

[Expensive] Number of pages (including subcategories and files) in the
given category. (Category:Help used for demonstration)

I had one thought of what the term 'expensive' means, which was that it ate up a lot of server memory? Or just in general, related to servers? Also, what makes some magic words 'expensive' and others not? Codyn329 (talk) 00:37, 12 March 2015 (UTC)

Check out Special:TrackingCategories and Category:Pages with too many expensive parser function calls, it's a tilt - game over effect. Articles MUST NOT be Turing Machines. Admittedly implementing the Ackermann function with templates is fun. –Be..anyone (talk) 17:34, 13 March 2015 (UTC)

urlencode

Is it possible to use this, to construct a link to an external webpage? An example would be greatly appreciated. --[[User:Bmrberlin| Bernd M.]] (talk) 11:26, 31 March 2015 (UTC)

There are three examples in the 2nd column of three, and there's a cheat sheet in the 3rd column: In essence urlencode always does the same thing, percent-encode Unicode, e.g., the input á u+00E1 is encoded as %C3%A1 for UTF-8 C3A1=C000+0300+0080+0021, with C=1100 for length two bytes, 3=0011 for some high bits, 8=10(00) for two bits indicating a tail byte, and 21=(00)10 0001 for six low bits. Putting hi+lo together that is (bits) 0011 10 0001 = (hex) E1 = u+00E1 as expected.

For the space character u+0020 you have to say what you want. Default or QUERY encodes spaces as + suited for, e.g., google queries. WIKI encodes spaces as _ as you would need it for a page#section_link. And PATH percent-encodes spaces as %20, suited for everything else, also known as URL-encoding. –Be..anyone (talk) 22:27, 31 March 2015 (UTC)

{{msgnw:xyz}} doesn't work with hidden comments

I try to include the unexpanded wikitext of a page (this) into another page (this). I use {{msgnw:Namespace:Pagename}}, but I have a problem with hidden comments in the original page. Some are included, others aren't! Can you help me? Thanks in advance. --FRacco (talk) 11:14, 9 May 2015 (UTC)

The visible XML-comment is within a gallery. All other XML-comments never make it into msgnw. Something isn't as it should be, please report the bug. –Be..anyone (talk) 08:27, 9 July 2015 (UTC)

Help! I'm completely stumped!

Hi, I was wondering if anyone can help me out or at least point me in the right direction as to where I can solve this problem. I have imported some pages and templates into my wiki from Wikipedia but they are not displaying correctly. In all the inboxes I get a bunch of errors saying "mw:Help:Magic words#Other"

For the life of me, I can't figure out what the problem is.

Can anyone help me out please.

Thanks

Here's a link: http://games.appipedia.com/wiki/Dizzy_–_The_Ultimate_Cartoon_Adventure — Preceding unsigned comment added by Finalcutbob (talk • contribs)

The problem is one of the included templates of the page. See [1]. --Ciencia Al Poder (talk) 09:31, 2 June 2015 (UTC)

Thanks for your help!

Help for translation

Hello, I recently made changes in this help page, but these changes are not transposed in the other languages. I guess this is related to the header saying "This page contains changes which are not marked for translation.", but I don't know how to mark my changes for translation. Anyone can tell me the way to do that? Best, Wikini (talk) 07:32, 5 June 2015 (UTC)

The message disappeared, I guess that someone made the changes needed. I again make changes in the English version and this message appears again. Best, Wikini (talk) 08:11, 9 June 2015 (UTC)

Behavior switch for auto-numbering section headings

Tracked in Phabricator
Task T33735

I suggest to create a new __NUMHEADINGS__ behavior switch to force auto-numbering the section headings. It would be useful in votes and polls, and in some structured pages. It should override individual user preferences in the page where it is included. Gustronico (talk) 02:19, 6 July 2015 (UTC)

Already filled on phabricator. --Ciencia Al Poder (talk) 19:05, 8 July 2015 (UTC)

PAGESINCATEGORY intersection

Is there anyway to do a PAGESINCATEGORY intersection with the current magicwords? E.g., incategory:"puzzle video games" incategory:"rareware games" in the enwp search bar shows the intersection between the two categories. I'd like to easily calculate the number of hits for that intersection. – czar 20:53, 3 August 2015 (UTC)

{{PAGENAME: }} and {{filepath: }} don't work well together with apostrophe in filename

Tracked in Phabricator
Task T18474

{{filepath: {{PAGENAME: File:Our country's report.pdf}} }}

doesn't return a filepath, although filepath does with only the filename and {{PAGENAME: }} does return the right filename.--AdSvS 07:29, 6 August 2015 (UTC)

AdSvS seems to be saying that something should be returned but it returns blank:

• {{filepath: {{PAGENAME: File:Our country's report.pdf}} }} →

The first problem is that File:Our country's report.pdf doesn't exist, and {{filepath:}} always returns a null string for non-existent files. If we try the same thing using the name of a file that does exist, but contains no apostrophes, such as File:ChameleonSkin.png we get the expected result:

• {{filepath: {{PAGENAME: File:ChameleonSkin.png}} }} → //upload.wikimedia.org/wikipedia/commons/f/ff/ChameleonSkin.png

Trying a file which exists and also contains apostrophes, such as File:Chameleon VisualEditor 'Insert Media Dialog' Z Index Issue.png, we get a null string:

• {{filepath: {{PAGENAME: File:Chameleon VisualEditor 'Insert Media Dialog' Z Index Issue.png}} }} →

If we test each part separately, we see

• {{PAGENAME: File:Chameleon VisualEditor 'Insert Media Dialog' Z Index Issue.png}} → Chameleon VisualEditor 'Insert Media Dialog' Z Index Issue.png
• {{filepath: Chameleon VisualEditor 'Insert Media Dialog' Z Index Issue.png }} → //upload.wikimedia.org/wikipedia/mediawiki/3/30/Chameleon_VisualEditor_%27Insert_Media_Dialog%27_Z_Index_Issue.png

i.e. it is something to do with the combination of the two functions. I think the actual problem is that {{PAGENAME:}} encodes certain characters, i.e. the apostrophe becomes the numeric entity '. --Redrose64 (talk; at English Wikipedia) 14:37, 6 August 2015 (UTC)

Thanks Redrose64 for the elaboration. That was indeed what I wanted to say and the filename was only an example. But your way is much, much better and I'll think about it when I run into strange things again! --AdSvS 15:51, 6 August 2015 (UTC)

And there's a bug for that! task T18474. This happens with other magic words as well. --Ciencia Al Poder (talk) 19:00, 9 August 2015 (UTC)

Thanks, that explains it, Brion added a link to the closed phab:T16779 eight years ago, and in phab:T16779 Tim Starling wrote six years ago: I don't see the need to take MediaWiki apart and put it back together again when you could just fix your broken #ifexist calls. Therefore I guess this is working as designed, you are not supposed to use minimally (PAGENAME) or fully (PAGENAMEE) encoded strings with filepath:. Suggestion, would…

{{filepath: {{#titleparts: {{PAGENAME: File:Chameleon VisualEditor 'Insert Media Dialog' Z Index Issue.png}} }} }}

//upload.wikimedia.org/wikipedia/mediawiki/3/30/Chameleon_VisualEditor_%27Insert_Media_Dialog%27_Z_Index_Issue.png

…work for you? The behaviour is documented for #titleparts and here as Warning. Be..anyone (talk) 20:40, 7 April 2016 (UTC)

Substituting PAGEID

The page id can be substituted with {{subst:PAGEID}} when editing an existing page. Can it be substituted when creating a page? GeoffreyT2000 (talk) 00:09, 26 December 2015 (UTC)

No, content is parsed before the new page is created, so it doesn't have page id at this stage. That's similar with REVISION variables. --Ciencia Al Poder (talk) 10:56, 26 December 2015 (UTC)

fullurl now outputs protocol specifier?

{{fullurl:page name}} is described as:

A protocol-relative path to the title. This will also resolve interwiki prefixes. Note: Unbracketed (plain) protocol-relative links are not automagically linked.

But in my MW 1.26 the output is a complete URL including "https:" and is automatically turned into a link!

Is this a new behaviour, or is it specific to my configuration?

--Ahmad Gharbeia أحمد غربية (talk) 11:01, 6 April 2016 (UTC)

I guess you should also have a protocol-relative $wgServer for it to work as described. --Ciencia Al Poder (talk) 19:43, 6 April 2016 (UTC)

Gender magic word

In my opinion Gender magic word should work in Help namespace too. I created a task on Phabricator. --Dvorapa (talk) 15:34, 13 June 2016 (UTC)

__NOGALLERY__ question, feature request

This tag is used in wikipedia mainly to avoid copyright problems in image categories, but unfortunately it makes it difficult to check which images need improvement. My question: is there a way to override the __NOGALLERY__ tag without removing it? If not, would it be possible to add such a feature? (e.g.- append "ignoreNogallery" to the url, or something similar). Thank you --Benstown (talk) 04:27, 31 August 2016 (UTC)

DISPLAYTITLE Example

While not clear from the current Help page description, DISPLAYTITLE may be used to creatively format page titles with <span> tag style. One example of this is hiding the namespace and base page name for subpages. The following code displays subpage name only.

{{DISPLAYTITLE:<span style="font-size:0">{{NAMESPACE}}:{{BASEPAGENAME}}/</span>{{SUBPAGENAME}}}}

Dave Braunschweig (talk) 13:58, 15 September 2016 (UTC)

Add magic word like REVISIONDATE

Whilst creating a content page i am missing the possiblity to create a simple revision date. The method i am using now is as follows: {{REVISIONDAY:Wielen}}-{{REVISIONMONTH1:Wielen}}-{{REVISIONYEAR:Wielen}} It would be much easier to have something like {{REVISIONDATE}} in stead — Preceding unsigned comment added by Gharryh (talk • contribs)

You have {{REVISIONTIMESTAMP}} which you can format with {{#time: d-m-Y | {{REVISIONTIMESTAMP}} }} --Ciencia Al Poder (talk) 10:22, 30 December 2016 (UTC)

Your solution wont work,it just shows the timestamp Type Faiverly Faiverly 20161230151037 }}

— Preceding unsigned comment added by Gharryh (talk • contribs)

It pretty much works: 07-12-2024 --Ciencia Al Poder (talk) 16:14, 30 December 2016 (UTC)

How to create dynamic dates (ie: {{CURRENTDAY}} + 1 = Tomorrow's date)

Does anyone know of a way to do this?

MikeDarling (talk) 04:17, 24 February 2017 (UTC)

UPDATE: I found a solution on Wikipedia. {{CURRENTMONTH1}}/{{#expr:1+{{CURRENTDAY}}}}/{{CURRENTYEAR}} produces: 12/22/2024 (which is tomorrow's date based on the server time). There's more that can be done too. See https://en.wikipedia.org/wiki/Wikipedia:Date_math for more info.

MikeDarling (talk) 04:42, 24 February 2017 (UTC)

Your solution has a flaw, because the last day of the month will give a non-existent date. You should get a date as a whole, add 1 day to it and get back the text representation of it. See Help:Extension:ParserFunctions: {{#time: m-d-Y | +1 day }} gives 12-22-2024 --Ciencia Al Poder (talk) 10:47, 24 February 2017 (UTC)

Thanks! I knew there had to be a better way. MikeDarling (talk) 22:09, 27 February 2017 (UTC)

Is the version 1.7 in the Namespaces section correct?

| <code>{{SUBJECTSPACE}}</code> <br /><code>{{ARTICLESPACE}}</code>
| {{SUBJECTSPACE}} <br />{{ARTICLESPACE}}
| Name of the associated content namespace
| {{MW version|version=1.7|compact=y|comment=and after}}
|-
| <code>{{TALKSPACE}}</code>
| {{TALKSPACE}}
| Name of the associated talk namespace
| {{MW version|version=1.7|compact=y|comment=and after}}

Should the version really be 1.7 for these 2 items? — Preceding unsigned comment added by 72.22.12.62 (talk • contribs)

Why not? Here are the release notes for 1.7 that prove it. --Ciencia Al Poder (talk) 20:13, 30 March 2017 (UTC)

Fullurl:page name

Hello.
When I create a template with this magic word - fullurl, it return an error massage. It say there is a mistake in the title. What can be the prublem? Betmidrash (talk) 14:18, 4 May 2017 (UTC)

What is the title and what is your code? Matěj Suchánek (talk) 11:58, 5 May 2017 (UTC)

Hi. What do you mean? The problem is not on the sites of WikiMeda, it's happen in my private wiki. When I use the magic-word: FULLURL in some template, and try to enter the URL, it say that there is a mistake in the title...

Betmidrash (talk) 08:03, 14 May 2017 (UTC)

urlencode:<syntaxhighlight>...</syntaxhighlight>

Hello!

I'm trying to pass a code snippet to a template. The code snippet uses <syntaxhighlight>...</syntaxhighlight>. I would then link to urlencode that code snippet, but urlencode:<syntaxhighlight>...</syntaxhighlight> doesn't work - presumably because syntaxhighlight somehow pulls the text out of the processing flow? Any ideas how this could be achieved? I.e. something like:

Template:ShowCodeMakeUrl

{{{1}}}
myserver?q={{uriencode:{{{1}}} }}

which is called as

{{ShowCodeMakeUrl|<syntaxhighlight>
...
</syntaxhighlight>}}

It works fine in terms of displaying the code, but urlencode doesn't handle the <syntaxhighlight>... Thanks! Bjohas (talk) 13:56, 5 May 2017 (UTC) (Also see: https://www.mediawiki.org/wiki/Topic:Tpz21dqvqgkokhfl)

• I did not fully understand your plan.
  • What is a “code snippet” and where is the problem?
• Yes, your assumption is right: syntaxhighlight pulls the content out of the source text around.
  • The parser will put the stuff between <syntaxhighlight> tags aside and transclude it later into the HTML document.
  • It is not possible to wrap something into <syntaxhighlight> tags first and process the content later. The other way around works.
• I fail to guess your expectation, is there a link supposed to be clickable, or shall the URL be displayed as code?
  • Note that for security reasons no external server must contribute code, but internal pages only and may be transcluded.
• As far I read your intention try something like that:
  • Displaying generated URL, simply use nowiki, why coloured syntaxhighlight?

{{#tag:nowiki|{{ShowCodeMakeUrl|...}}}}

• A clickable link with coloured syntaxhighlight as linktext, got passed a pipe-escaped code snippet, but no line breaks:

• Template programming of ShowCodeMakeUrl:

[http://example.org?q={{uriencode:{{{1}}}}} {{#tag:syntaxhighlight|{{{1}}}|lang="php"|inline=1}}]

• Transcluded with:

{{ShowCodeMakeUrl|way[highway][name="6th Avenue"];node(w)->.n1;way[highway][name="West 23rd Street"];node(w)->.n2;node.n1.n2;out meta;}}

Greetings --PerfektesChaos (talk) 03:46, 6 May 2017 (UTC)

I'll try to summarize all requirements we have:

- Source code (actually "Overpass QL" language) is passed to template as a parameter

- Source code is rendered with syntax highlighting, preserving any new line characters

- An additional icon with hyperlink is generated, hyperlink includes source code as URL parameter. All newline characters need to be preserved!

- If you click on the hyperlink, you'll see exactly the same source code as shown on the wiki page.

See this page with the template in action: https://wiki.openstreetmap.org/wiki/Overpass_API/Overpass_API_by_Example

Bjohas wanted to find out some way to get rid of the sometimes insane escaping, like ({{((}}bbox{{))}}) instead of ({{bbox}}). This would have the advantage, that a user can just copy & paste some source code as a template parameter and doesn't have to bother with escaping.

For even more details: see https://github.com/tyrasd/overpass-turbo/issues/288

Mmd-osm (talk) 06:51, 6 May 2017 (UTC)

Template:OverpassTurboExample is already using the means I suggested above.

• Note: source is called syntaxhighlight today.
• Therefore, {{#tag:source|{{{query}}}|lang=c is already present like I proposed {{#tag:syntaxhighlight|{{{1}}}|lang= above.
• Template programming functionality is as best as possible yet.

Your real problem is easy escaping of {{bbox}} since those curly brackets are colliding with Wiki syntax.

• The solution is quite easy:

{{OverpassTurboExample|query=<nowiki>node["amenity"="post_box"]({{bbox}});</nowiki>}}

• The enclosing nowiki-tags when providing the query= parameter will be stripped off before their content is passed as {{{query}}} inside the programming.
• However, they will prevent {{bbox}} getting evaluated too early. It is intended that this should happen in advance usually. When returning, their remainders won’t get expanded any longer.
• Should work. Enjoy.
• Entire code should be able to be wrapped this way without touching anything inside.

The basic idea when this thread has been opened was fine: enclosing the program code into something to prevent from evaluation.

• However, syntaxhighlight is a very powerful thing that does a lot of analysis and produces sophisticated output with error messages and highlighting. That is parked outside the wikitext and merged later when the HTML document is composed.
• 15 years old nowiki has the same disabling effect, but on low level, just hiding its content once from parsing, and bye.

Greetings --PerfektesChaos (talk) 17:01, 6 May 2017 (UTC)

Thanks a lot for the detailed feedback! We have tried the nowiki approach before. Unfortunately, it has one major shortcoming: the source code is no longer passed on to the URL parameter: https://wiki.openstreetmap.org/wiki/User:Mmd/TurboExample - so our initial requirement If you click on the hyperlink, you'll see exactly the same source code as shown on the wiki page. is no longer fulfilled using this approach. Not sure if there's some way to get around this limitation? Mmd-osm (talk) 07:29, 7 May 2017 (UTC)

• Sorry, my approach has the same pitfall like syntaxhighlight, since nowiki also gets some kind of escaping and is not available for a parser function. I never noticed that yet.
• Then, the only way to relieve your burden a bit is as follows:
  • Create an auxilary template:bracketed (or any other name).
  • Implement that as follows:
    <onlyinclude>{{<nowiki />{{{1}}}<nowiki />}}</onlyinclude>
  • Transclude

{{OverpassTurboExample|query=node["amenity"="post_box"]({{bracketed|bbox}});}}

• At least slightly more readable.
• The last remainig syntax with no extra template would be:

{{OverpassTurboExample|query=node["amenity"="post_box"]({<nowiki />{bbox}<nowiki />});}}

Good luck --PerfektesChaos (talk) 15:02, 7 May 2017 (UTC)

Many thanks for the suggestions - that's helpful! The developer of https://github.com/tyrasd/overpass-turbo has included an 'escaping mechanism' that essentially does this https://github.com/bjohas/mediawikiencode. Many thanks. Bjohas (talk) 10:43, 8 May 2017 (UTC)

On the letter-case of Variables

Hi, I'm recently working on a parser on wikitext, and have read this page so that I can decide whether a "transclusion" is actually a variable or parser function.

Variables, per definition, are uppercase words surrounded by double braces, so {{PAGENAME}}, {{SERVERNAME}} are variables. On the contrary, I believe, parser functions (esp. those starting with #) are case-insensitive.

However, some "variables" are suprisingly case-insensitive, for example, {{sErverName}} will still be parsed to www.mediawiki.org; while others, e.g. {{PAGENAME}} are not (e.g. Template:PAgeName). Though I can hard-code a list for these built-in variable and parser function names, I just couldn't figure out why some "variable"s should be case-insensitive, as it's against the definition in Help:Magic words.

Another problem that confuses me: in Help:Magic words, it's said that behavior switches, variables, and parser functions are three kinds of magic words, but in Manual:Magic words, magic words are defined as "… a technique for mapping a variety of wiki text strings to a single ID that is associated with a function. Both variables and parser functions use this technique." I was thinking about a name to indicate either variables or parser functions, while to distinguish from behavior switches, because variables and parser functions share the similar traits (e.g. the first argument is started with colon, not pipe). Now I'm using something like "Magic Template" but not sure whether it's suitable. Any suggestions? Thank you!

--CXuesong (talk) 08:00, 12 May 2017 (UTC)

{{formatnum:}} result doesn't include comma separator in mywiki

The result value of {{formatnum:123456789000}} in Burmese Wikipedia (mywiki) shows as "၁၂၃၄၅၆၇၈၉၀၀၀" without comma separator. To include group separator in mywiki, how should I do?

code	result on mywiki	expected value
{{formatnum:987654321.654321}}	၉၈၇၆၅၄၃၂၁.၆၅၄၃၂၁	၉၈၇,၆၅၄,၃၂၁.၆၅၄၃၂၁
{{formatnum:၁၂၃၄၅၆၇၈၉၀၀၀}}	၁၂၃၄၅၆၇၈၉၀၀၀	၁၂၃,၄၅၆,၇၈၉,၀၀၀

Ninja✮Strikers ^«☎» 03:51, 6 July 2017 (UTC)

I've been digging in the code and I can't find why it's not formatting it correctly. languages/messages/MessagesMy.php looks correct. API request for easy testing. I suggest you to open a task for this --Ciencia Al Poder (talk) 09:49, 6 July 2017 (UTC)

Username

Is there anything that can detect a character's username? For example, if your name was UserGuy1, the code {{USERNAME}} or something similar would become UserGuy1. Is there anything that can do this? 96.48.149.66 07:25, 29 July 2017 (UTC)

You may try Extension:GetUserName but I'm not sure if this extension will work in current versions of MediaWiki and it also breaks caching. --Ciencia Al Poder (talk) 15:34, 30 July 2017 (UTC)

{{NUMBEROFPAGES}}: bug

On pl wiki and (probably) some others {{NUMBEROFPAGES}} doesn't count new pages, but counts deleted pages. Sławek Borewicz (talk) 19:45, 20 February 2018 (UTC)

I can confirm half of what you are reporting. There's been a lot of page creation and deletion lately on plwiki, so I was able to monitor the statistics (more or less) "live" as pages were created and deleted. What I saw is this:

1. the page count decreased when pages were deleted (and, as far as I could tell, it seemed to be decreasing by the correct amount each time),
2. the article count decreased only sometimes when pages were deleted (which would be expected if some deleted pages counted as articles and some didn't),
3. the article count increased but the page count did not change when an article (with links) was created, and
4. the page count did not increase when a new page was created outside of the main namespace.

The last two things do seem to be a bug and not just due to lag. I watched for more than 25 minutes and never saw the page count go up despite multiple page creations in that time. I guess this deserves a Phabricator task (i.e., bug report). - dcljr (talk) 21:17, 20 February 2018 (UTC)

Turns out, this was a known bug that has been fixed. Camping out at pl:Special:RecentChanges and pl:Special:Statistics for a while, I was able to verify that page creation is now increasing the page count. - dcljr (talk) 01:07, 22 February 2018 (UTC)

Thx (for the info). Sławek Borewicz (talk) 16:55, 22 February 2018 (UTC)

You forgot equal

Tracked in Phabricator
Task T91154

You forgot to document {{=}}. (seems to be a template currently, but should be a magic word)

Two hours and a headache. Alexis Jazz (talk) 10:21, 23 July 2018 (UTC)

No, it’s not forgotten. This page documents things that are magic words, not those that should be. {{= }} is not a magic word. You can propose it to became one on Phabricator, and, if implemented, it will be documented here. —Tacsipacsi (talk) 12:55, 23 July 2018 (UTC)

Templates, not magic words

Some variables listed, such as {{ROOTPAGENAME}}, are templates, and not magic words.TheConqueror4712 (talk) 00:55, 8 October 2018 (UTC)

I don’t think so. It’s used on this page, but Help:Magic words doesn’t appear on Special:WhatLinksHere/Template:ROOTPAGENAME. Looks like StructuredDiscussions thinks MediaWiki doesn’t support it, but it does. —Tacsipacsi (talk) 07:19, 8 October 2018 (UTC)

JS to make "wgRevisionId" a magic word

Hi,

I'm using Extension:Approved Revisions [1] and I have a need to be able to display the revision ID of the page revision currently being viewed. From the MW Manual page on Javascript [2] I see that the variable "wgRevisionId" contains this information and is exposed to mediawiki for use, but it is not clear to me how to do so. I would like to create a magic word [3] called Template:VIEWEDREVID that outputs the revision ID of the revision being viewed such that it can be used in other parser logic

For example:

{{#ifexpr:( {{VIEWEDREVID}} < {{APPROVEDREVID}} )
|if true
|if false
}}

Can someone please help me identify the way to make a magic word that contains the value of "wgRevisionId"?

Thank you! -Rich User:revansx

[1] https://www.mediawiki.org/wiki/Extension:Approved_Revs

[2] https://www.mediawiki.org/wiki/Manual:Interface/JavaScript

[3] https://www.mediawiki.org/wiki/Help:Magic_words

---

Let's concentrate all discussions in one place: Extension talk:Approved Revs#Help making "wgRevisionId" a magic word --Ciencia Al Poder (talk) 09:09, 25 October 2018 (UTC)

CURRENTWEEK/YEAR when 31 Dec is on a Monday

Hello from Monday, 31 Dec 2018, where CURRENTWEEK returns 1 and CURRENTYEAR returns 2018, resulting in nonsense when used together. I'm guessing that CURRENTWEEK is handled like ISO week dates but CURRENTYEAR is Gregorian, resulting in weird behavior on this edge case. -PFWOz (talk) 19:02, 31 December 2018 (UTC)

Dug through some 10-year-old bug tickets and the workaround is to use {{#time}}. T18838:

Use {{#time:o/W}} combination instead of {{#time:Y/W}} or {{CURRENTYEAR}}/{{CURRENTWEEK}}.

This should be made much clearer in the documentation, which suggests {{#time}} for more thorough time formatting, but not for improved correctness. Some details on how these magic words are calculated would help when trying to determine whether that's actually necessary. -PFWOz (talk) 19:07, 31 December 2018 (UTC)

False example for PAGENAME encoding of apostrophe

Help:Magic words#Page names shows an example which is currently false:

Warning:

For example, if page name is "L'Aquila", the following code will produce the string "Numeric char encoding": {{#switch:{{PAGENAME}} | L'Aquila = No translation | L = Not OK | L'Aquila = Entity escaping | L'Aquila = Numeric char encoding }}

The code actually produces "No translation". Numeric char encoding to &#39; does happen but #switch (same with #ifeq) will match &#39; to any of ' (a pure apostrophe), &apos; and &#39;. The first match in the switch is chosen. "Show changes" on {{subst:PAGENAME:L'Aquila}} will show that {{PAGENAME}} produces L&#39;Aquila if page name is "L'Aquila". It can cause problems in other situations. PrimeHunter (talk) 12:33, 12 January 2019 (UTC)

It is a wiki...

Anyway, when it was written it was 100% right, but things in the parser have changed since then, which is easy enough to prove by installing an older mediawiki version, or finding a public wiki, e.g. on wikia's wikis it shows "L'Aquila". The only way to keep the example always true is to make it dynamic by identifying a parser function or lua function that will always match its behaviour, and transcluding that.

The rest of the comment is rather irrelevant, as far as the misguided "switch" parser function is concerned. It mostly follows the rules of programming languages, if multiple values match then it picks the first one, and that's clearly documented even for non-programmers.

Also, the warning comes with bug reports, so if any of them gets fixed it affects the example. So technically it isn't false.

197.218.85.83 11:07, 13 January 2019 (UTC)

Other magic words

Where can I find magic words such as {{GENDER}} and {{SHORTDESC}}? CaiusSPQR (talk) 01:31, 23 January 2019 (UTC)

These are included with Extension:Wikibase. Took me a couple hours to figure out {{SHORTDESC}}, but it seems you need to add $wgWBClientSettings['allowLocalShortDesc'] = true; to your localsettings.php after setting up Wikibase. This does add "shortdesc" to the list of parser function hooks in Special:Version, but I have yet to see it actually work as {{SHORTDESC:}}, is still recognized as a template on my own wiki for some reason. Somebody should really document this properly somewhere. - ElementalLagomorph (talk) 22:52, 29 January 2019 (AEDT)

Get number of subpages of a mainpage

I have a page with multiple subpages attached to it. Is it possible to get the count of subpages and display that information? Extarys (talk) 05:06, 19 February 2019 (UTC)

I don’t know of any core function, but I should be possible using an extension (I don’t know whether such extension exists currently). —Tacsipacsi (talk) 18:22, 19 February 2019 (UTC)

{{DEFAULTSORT:}}

Does an empty defaultsort make sense or can it be deleted? I found it on may pages in Commons[2]. --Aschroet (talk) 05:07, 5 April 2019 (UTC)

You see that with empty argument it has no effect – instances on Commons likely resulted from a malfunctioning bot. This behaviour certainly should be documented. Incnis Mrsi (talk) 11:56, 7 April 2019 (UTC)

Behavioral switches -> directives?

Does behavioral switches shouldn't be named directives? They seem similar. MarMi wiki (talk) 19:36, 13 June 2019 (UTC)

Help:Magic_words#Localization - plural description should be more descriptive

@Shirayuki: - I think that plural description should be more descriptive:

• Is Form 1 a singular form of plural (then some more examples would be handy), or it's just a singular (only 1 single item)?
• Is it only for words like is/are, or it's for singular (form 1, number 1)/paucal (form 2 or 3, numbers ending with digit 2-4)/plural (form 2 or 3, numbers ending with digit 0-1 [without number 1] and 5-9) forms of nuons (like Russian example may suggest [don't know Cyrillic/Russian that well])? (Numbers are for Poland).
• Form 1 has numbers 1, 21, 31,... What about 11?
• Does it even work on non-english sites? From my tests (based on is/are examples) it seems that it doesn't ({{plural:5|is|are}} should return empty value, but it doesn't). Okay, it works on Russian wikipedia, but in strange way: {{plural:5|is|are}} returns are, {{plural:5|is|are|third}} returns third. Shouldn't first return empty value?

MarMi wiki (talk) 00:11, 14 June 2019 (UTC)

As I wrote in edit summary, the description came from translatewiki:Thread:Portal_talk:Ru/Plural_changes_in_many_languages, so "21, 31, ..." should NOT be removed. It is written about some languages (ab, av, ba, etc.).

For example, English ordinal numbers have similar rule. 1st, 11th, 21st, 31st, etc. -Shirayuki (talk) 11:57, 14 June 2019 (UTC)

But what if it's not an ordinal number? Then it will work wrongly (in Polish): 1 szablon (template), 11 szablonów (templates), 21 szablonów (templates), 31 szablonów (templates), etc. Unless it's meant only for ordinal numbers (then there should be two magic words: ordinal and plural)? MarMi wiki (talk) 12:12, 14 June 2019 (UTC)

I think that plural description needs to specify if it's only for general broad singular/plural differentiation (jeden szablon/one template, wiele szablonów/many templates) or more detailed singular/plural differentiation based on numbers (1 szablon/template, 2 szablony/templates, ..., 4 szablony/templates, 5 szablonów/templates, ..., 9 szablonów/templates, etc.). MarMi wiki (talk) 13:10, 14 June 2019 (UTC)

Maybe someone knows where I can see couple of uses of this magic word in actual action? It would help to determine if it's fine as it is, or it needs some changes for my language.

Another note: is it choosing the mode depending on parameter count? Ex. 3 parameters (number + 2 words) - English (general broad) mode, 4 - Russian (more detailed) mode?. MarMi wiki (talk) 13:59, 14 June 2019 (UTC)

Summary of magic word plural after some tests:

• It's for simple one/many differentiation (English mode) AND for singular (form 1, number 1)/paucal (form 2 or 3, numbers ending with digit 2-4)/plural (form 2 or 3, numbers ending with digit 0-1 [without number 1] and 5-9) forms of nuons (given numbers may be wrong depending on given word) ("Russian" mode)
• It works on Polish wikipedia (but numbers given in documentation for Form 1/2/3 doesn't match - they should be translatable too)
• Used mode depends on number of parameters: number + 2 parameters (words) - simple English mode, with number + 3 parameters - advanced "Russian" (adapted to Polish) mode. MarMi wiki (talk) 14:49, 14 June 2019 (UTC)

And I think Russian example needs to be updated - it doesn't match the order of numbers in Form 1/2/3. Russian example is correct - form 2 is in the third parameter (not counting the first number), and form 3 is in the second. Form order doesn't match the parameter order, which may be a little confusing. MarMi wiki (talk) 15:48, 14 June 2019 (UTC)

Moved the Russian special rule into Note template. It is not about English, Polish, and Japanese. -Shirayuki (talk) 02:29, 15 June 2019 (UTC)

@Shirayuki: I find it confusing too. Form 1, form 2 and form 3 are supposed as {{PLURAL:$1|form 1|form 2|form 3}}. The note box for Slavic languages has a link to Translatewiki explaining that form 2 and form 3 were interchanged in 2014, but when I try it w:ru:Special:Permalink/103697654 they work in the previous order. Have I missed anything? --Vriullop (talk) 10:42, 5 December 2019 (UTC)

I think I found an explanation. The three forms are usually referred in papers as singular (1st form in nominative singular), plural (2nd form in genitive plural) and paucal (3th form in genitive singular). The paucal form is usually the last one as it is less common. But the parser function uses another order {{PLURAL:$1|singular|paucal|plural}} that is not properly documented, it can only be deduced from the examples in Russian by native speakers. --Vriullop (talk) 11:49, 6 December 2019 (UTC)

Template name

Is there a magic word similar to {{PAGENAME}} that doesn't change when it's transcluded as a template? For example, say this magic word was called {{TEMPLATENAME}}. If I created a template called {{Title}}, containing something like "This template is called {{TEMPLATENAME}}", it would transclude as "This template is called Title". I can't use {{PAGENAME}} for this, because it automatically changes to the page title. Gfdgsgxgzgdrc (talk) 00:31, 16 June 2019 (UTC)

@Gfdgsgxgzgdrc: I don’t know of any, but why would it be useful? {{PAGENAME}} is useful because the template can query the name of the page it’s used in—the template’s creator doesn’t know where the template will used. But the title of the template you’re just editing is known to you, so you can simply hardcode it. —Tacsipacsi (talk) 23:27, 16 June 2019 (UTC)

You can put {{subst:PAGENAME}}, that will be substituted on page save. --Ciencia Al Poder (talk) 09:35, 17 June 2019 (UTC)

SUBPAGENAME : PAGENAME should be replaced by BASEPAGENAME for assertion to be coherent with display

Hi all, In paragraph https://www.mediawiki.org/wiki/Help:Magic_words#Page_names we use PAGENAME . On the FR translated page I get:

SUBPAGENAME fr Titre de la sous-page ("bar" pour "Aide:Titre/foo/bar"). Si aucune sous-page n'existe, la valeur de {{PAGENAME}} est renvoyée.

If I look at PAGENAME I read 'Magic words/fr' which is not coherent since I expect only 'Magic words' (as for EN pages).

Then Translations:Help:Magic words/125/fr should be:

SUBPAGENAME 	fr 	Titre de la sous-page ("bar" pour "Aide:Titre/foo/bar"). Si aucune sous-page n'existe, la valeur de  {{BASEPAGENAME}}  est renvoyée.

=> Can anyone substitue the value of $code please using BASEPAGENAME instead of PAGENAME  ?

FR version displays:

PAGENAME 	Magic words/fr 	Titre complet de la page (incluant tous les niveaux des sous-pages) sans l'espace de noms.

BASEPAGENAME 	Magic words 	Titre de la sous-page de niveau immédiatement supérieur sans l'espace de nom ("Titre/foo pour "Aide:Titre/foo/bar")...

SUBPAGENAME fr Titre de la sous-page ("bar" pour "Aide:Titre/foo/bar"). Si aucune sous-page n'existe, la valeur de {{PAGENAME}} est renvoyée.

Thanks.

Christian Wia (talk) 09:07, 5 July 2019 (UTC)

I personally think the current situation is fine, and even more illustrative than then english page, since you can actually have different values for SUBPAGENAME, BASEPAGENAME, etc. --Ciencia Al Poder (talk) 10:49, 8 July 2019 (UTC)

page is in category X

I'm searching for a magic word or a function that can tell if a page or file is in a specific category. I have not found anything so far. --D-Kuru (talk) 10:36, 25 August 2019 (UTC)

@D-Kuru: This is because it could have unpredictable results in edge cases. Let’s say we have a such function named #ifincat. What should the below code produce on the page “Page”?

{{#ifincat:Page|Category|<!-- do nothing -->|[[Category:Category]]}}

If the page is in the category, the function returns nothing, so the page doesn’t get placed in the category, so the function returns the code placing it in the category, so it’s in the category, so it doesn’t get placed in it… —Tacsipacsi (talk) 15:06, 25 August 2019 (UTC)

@Tacsipacsi: Thanks for the fast reply!

I don't think that you need a specific output on the page itself (unlike eg. PAGESINCAT). I would have used just the return value (be it a boolean or a string) as some kind of the opposite of a template that includes a category. The use in my project would have been to check if file A.jpg is in Category:Hello and Category:Goodbye and if this is the case include a template, a warning, a category or whatever.

Your example would indeed be bad. You could display a warning though, when the function is used like that (like there is for template loops).

--D-Kuru (talk) 15:25, 25 August 2019 (UTC)

This could also be interesting as tool as such.

If you want to improve your photograohy skills and you want to see what images User:B has taken with camera Y and lens Z that are considered Qualiy Images or Fetured Pictures --D-Kuru (talk) 15:52, 25 August 2019 (UTC)

@D-Kuru: Do you want to use this feature on Commons or your own wiki? I don’t think WMF would welcome a such function because of its edge cases, but you can use PetScan for intersecting Commons images (and that requires no template at all on the file description page!). —Tacsipacsi (talk) 17:54, 25 August 2019 (UTC)

@Tacsipacsi: I would use this for a project on Wikimedia Commons. PetScan looks quite nice (even I'm not quite sure how to use it), but I can not include this on a page where I can automatically tag certain images.

I already thought that this is not a very popular function or it would already exist. Because of the lack of knowledge how I could do that myself I have to see if there already is something like that. Short of a function like that existing, I have to manually edit the file description pages - which is actually the kind of manual work I want to avoid with this. --D-Kuru (talk) 19:49, 25 August 2019 (UTC)

@D-Kuru: I don’t know how do you want to use it on Commons, but the template should be on the page to start with, and building it in a widespread template is a big no performance-wise—such templates’ edits may slow down Commons, if only the 0.1% of the files break (e.g. with “too many expensive parser functions” error), that’s tens of thousands of files, and the edit may be visible even in WMF’s electricity bills due to reparsing all 55 million files. And I haven’t thought of it before, but I’m not sure it’s possible at all to access a page’s categories reliably before processing the page entirely, at which point no parser function can do anything. So I think this feature cannot be implemented at all without rewriting MediaWiki, which is hardly a possibility. 🙂 —Tacsipacsi (talk) 21:26, 25 August 2019 (UTC)

Please document SHORTDESC

Please add documentation for SHORTDESC. Thanks. I see a mention of it in task T184000, but I don't know where to find official documentation (aside from this page, where it is missing). Jonesey95 (talk) 19:45, 6 October 2019 (UTC)

@Jonesey95: This page documents only MediaWiki core’s magic words, not extensions’ ones. SHORTDESC is provided by Extension:Wikibase Client, so it should be documented there. —Tacsipacsi (talk) 20:47, 6 October 2019 (UTC)

Magic word that checks how many pages a template is on?

Does anyone know how to do this? — Preceding unsigned comment added by Pomegranatecookie (talk • contribs) .

No, there's no magic word for that. You can get rough idea of the number however through Special:WhatLinksHere and using namespace selector. If it's on Wikipedia that you want count this you can use https://tools.wmflabs.org/templatecount/ — Preceding unsigned comment added by Ammarpad (talk • contribs) 21:19, 19 January 2020 (UTC)

If your wiki has CirrusSearch installed, then you can get an exact count by searching with hastemplate:"Template:MYTEMPLATE" and selecting all namespaces. --Bdijkstra (talk) 08:35, 21 January 2020 (UTC)

Querying preprocessor values during page processing

Is it possible for a template or module to query the following values (taken from "view page source" of en:Wikipedia:Template_limits oldid=925577618 a few minutes ago)? If so, how?

Preprocessor visited node count: 613/1000000
Post‐expand include size: 52139/2097152 bytes
Template argument size: 234/2097152 bytes
Highest expansion depth: 10/40
Expensive parser function count: 6/500
Unstrip recursion depth: 0/20
Unstrip post‐expand size: 3942/5000000 bytes
Number of Wikibase entities loaded: 0/400
Lua time usage: 0.085/10.000 seconds
Lua memory usage: 2.52 MB/50 MB

A "limit-aware" module could, for instance, revert to simpler logic at the expense of functionality if it detected that the page calling it was approaching any of the above limits. Davidwr (talk) 21:37, 25 February 2020 (UTC)

New magic word?

I think the cirrus search function filemime: is really neat. I have not seen it as a magic word however. Have I missed something? If not, where can you suggest new magic words? Honestly I think all cirrus search functions filemime:, filetype:, incategory:, hastemplate:, etc. should have a magic word pair, if they already do not.Jonteemil (talk) 00:20, 14 March 2020 (UTC)

MIME is accessible from Lua, see Extension:Scribunto/Lua reference manual#File metadata. Lua can generally access more things, so I’m not sure if it will ever be added directly to wikitext. —Tacsipacsi (talk) 00:25, 14 March 2020 (UTC)

@Tacsipacsi: Okay, thanks. What would I write at c:File:Flag of Sweden to output image/svg+xml?Jonteemil (talk) 16:32, 14 March 2020 (UTC)

Okay, I found it at c:Module:File, thanks!Jonteemil (talk) 16:58, 14 March 2020 (UTC)

Avoid section numbering for just one page

In my preferences I have included to show section numbers. But on some pages (e.g. multi-column) this is disturbing the layout. Could there possibly exist a __NOSECTNUM__ construct to overrule the preferences setting for one single page? Geertivp (talk) 13:43, 12 April 2020 (UTC)

What preference is it? Does it come from an extension or is it a gadget? For single-page customizations, TemplateStyles would work. --Matěj Suchánek (talk) 09:17, 13 April 2020 (UTC)

It is a plain MediaWiki (global/local) setting:

Preferences -> Appearance -> Advanced options -> Auto-number headings -> Enable/disable

@Matěj Suchánek: Thanks for your Extension:TemplateStyles hint.

I implemented your recommendations -- it worked like a gem -- see: wmbe:Wikimedia Belgium and wmbe:Template:CSS/Nosectnum.css - Geertivp (talk) 12:27, 13 April 2020 (UTC)

Brackets

Is there a PAGENAME variant which removes brackets?--Launchballer (talk) 23:36, 1 June 2020 (UTC)

Most brackets are already not allowed in Page title, I think only round brackets are. It would be easier to give you better answer if you explain exactly what you want achieve. Ammarpad (talk) 07:44, 27 July 2020 (UTC)

enwiki and other wikis use w:Template:PAGENAMEBASE (custom template, not a magic word, don't confuse with BASEPAGENAME). Matěj Suchánek (talk) 09:31, 27 July 2020 (UTC)

Could {{FULLURL:}} be modified to take more than one parameter?

All the url functions seem to only take one parameter. Is there any way of having more, whilst still using magic words? User:WT79 (please use my enwiki talk page) 14:58, 26 July 2020 (UTC) (I am not watching this page, so please ping me if you want my attention.)

@WT79: The query_string is not constrained to only one parameter. So you can build a complex query with many parameters. Example using four paramaters, {{fullurl:en:Wikipedia:Sandbox|oldid=969629111&action=edit&section=1&safemode=1}} gives: https://en.wikipedia.org/wiki/Wikipedia:Sandbox?oldid=969629111&action=edit&section=1&safemode=1 Ammarpad (talk) 07:39, 27 July 2020 (UTC)

@Ammarpad: Thanks; din't think of using &s in the parameter. User:WT79 (please use my enwiki talk page) 08:07, 27 July 2020 (UTC)

PAGENAME vs. #invoke:WLink|getArticleBase

Hi,

please can anyone tell me is there any difference between {{PAGENAME}} and the LUA-Module {{#invoke:WLink|getArticleBase}}? I can't spot it and I can't find anything about a comparison via Google/Bing, too.

Thanks! — Preceding unsigned comment added by Semmelrocks (talk • contribs) 22:11, 2 November 2020 (UTC)

{{#invoke:WLink|getArticleBase}} removes text in parentheses from the page name, whereas the {{PAGENAME}} doesn't. For example, you can see from de:Wikipedia:Lua/Modul/WLink/Test#getArticleBase that {{#invoke:WLink|getArticleBase|WLink (library)}} produces WLink, but {{PAGENAME:WLink (library)}} produces WLink (library) * Pppery * _{it has begun} 22:50, 2 November 2020 (UTC)

@Pppery Thanks a lot!--Semmelrocks (talk) 15:30, 3 November 2020 (UTC)

User-related magic words

Is there some equivalent of {{CURRENTUSER}}? The functionalities I'm looking for are:

• Is the current user viewing this page logged in?
• What is the username of the current current user viewing this page?

Are there any magic words or equivalent that can return this in wikimarkup? T.Shafee(Evo﹠Evo)^talk 06:47, 13 December 2020 (UTC)

There was once an Extension:USERNAME but it has been archived and nobody uses it, because it's not compatible with caching. --Ciencia Al Poder (talk) 10:28, 14 December 2020 (UTC)

MediaWiki:Group-user.css can be used to specify styles that apply only for logged-in users. You can define e.g. a .noanon and an .anononly class, hide .noanon in MediaWiki:Common.css, and hide .anononly and forcibly show .noanon in Group-user.css. This is quite limited, as it can only affect the end result, you can’t compute further depending on the logged-in state, and probably it doesn’t work on mobile, but at least it exists. Depending on your precise requirements, it may or may not be useful. —Tacsipacsi (talk) 19:50, 18 December 2020 (UTC)

Formatnum outputs extraneous spacing

Tracked in Phabricator
Task T277760

When {{Formatnum }}, or any wrapper template, receives a non-numeric argument, the page is correctly added to Category:Pages with non-numeric formatnum arguments. However, I noticed that the template also outputs what appears to be a newline at the end of the text. This does not show or pose any problems for rendered articles but appears as extraneous spacing when using the Visual Editor. Check my sandbox (in Visual Editor mode) as an example. Inexperienced users who see this often try to "fix" what looks like a double-space by removing the actual space that follows the template use. Can this be fixed? I could not immediately find the source code for the magic word, if it is even viewable publicly. Regards, IceWelder (talk) 20:09, 9 March 2021 (UTC)

Hello @IceWelder: . It looks like you're using the commas on a magic word. We recommend you to remove that. It is automatically adding commas. For the 4. option, it is not a good idea to put there in a "million" word in formatnum. Exclude the word from the "million" word, so that category will be removed. Thanks. —Baran ^☪︎ 03:37, 10 March 2021 (UTC)

@Baris6161TURK: I am aware that the usage is technically incorrect. However, such errors appear numerously on enwiki, and the large majority of users who see the aforementioned double-space would not know how to fix it properly. As I see it, this problem needs fixing on both sides: Correction of the magic word output to stop the issue from occurring to new users, and cleanup of the template arguments on enwiki to correct the template usage. IceWelder (talk) 09:02, 10 March 2021 (UTC)

The code responsible for noticing non-numeric input and marking it as such is here, but I have no clue why this space appears. Maybe it’s a bug in VisualEditor/Parsoid rather than in the parser function. —Tacsipacsi (talk) 15:02, 10 March 2021 (UTC)

It might be a bug in Parsoid depending on how it processes the template output. For example, the generated HTML for a proper out looks like this:

<p id="mwBQ" class="ve-ce-branchNode ve-ce-contentBranchNode ve-ce-paragraphNode">||<span about="#mwt2" typeof="mw:Transclusion" data-mw="{&quot;parts&quot;:[{&quot;template&quot;:{&quot;target&quot;:{&quot;wt&quot;:&quot;formatnum:22.0&quot;,&quot;function&quot;:&quot;formatnum&quot;},&quot;params&quot;:{},&quot;i&quot;:0}}]}" id="mwBg" class="ve-ce-leafNode ve-ce-focusableNode ve-ce-mwTransclusionNode ve-ce-focusableNode-focused" contenteditable="false">22.0</span>||</p>

While the improper one looks like this:

<p id="mwCA" class="ve-ce-branchNode ve-ce-contentBranchNode ve-ce-paragraphNode">||<span about="#mwt3" typeof="mw:Transclusion" data-mw="{&quot;parts&quot;:[{&quot;template&quot;:{&quot;target&quot;:{&quot;wt&quot;:&quot;formatnum:22,000,000&quot;,&quot;function&quot;:&quot;formatnum&quot;},&quot;params&quot;:{},&quot;i&quot;:0}}]}" id="mwCQ" class="ve-ce-leafNode ve-ce-focusableNode ve-ce-mwTransclusionNode" contenteditable="false">22,000,000
</span><link rel="mw:PageProp/Category" href="https://en.wikipedia.org/wiki/Category:Pages_with_non-numeric_formatnum_arguments" about="#mwt3" id="mwCg" class="ve-ce-leafNode ve-ce-focusableNode ve-ce-mwTransclusionNode" contenteditable="false">||</p>

Notice the linebreak after "22,000,000" that is still within the span tag. I see the code for Parsoid is located on GitHub but I wouldn't know where to start looking for something like this. Maybe we can get a Parsoid dev (@Arlolra?) to also check on this? IceWelder (talk) 22:23, 11 March 2021 (UTC)

You can also file a bugreport to the relevant project --Ciencia Al Poder (talk) 15:41, 12 March 2021 (UTC)

Since we haven't identified the exact cause yet, I wouldn't know where to file the bug report. IceWelder (talk) 16:58, 15 March 2021 (UTC)

The source of the newline is here. It comes from this commit introduced for T72196. If it's causing a problem, a task should be filed against Parsoid. Thanks Arlolra (talk)

I filed T277760 for this Arlolra (talk)

@Arlolra: Thank you very much! Does the mobile app also use Parsoid or is that a separate bug? IceWelder (talk) 14:48, 18 March 2021 (UTC)

Same bug, the mobile apps use Parsoid's output. Arlolra (talk) 14:55, 18 March 2021 (UTC)

What does the wikitext look like here? It seems perhaps the extra space is being added between the output of formatnum and the category tag, when the "pages with non-numeric formatnum arguments" category has to be added? cscott (talk) 20:23, 15 March 2021 (UTC)

@Cscott: In the VE, it renders as:

||22||
||22.0||
||22,000,000 ||
||22 million ||

The spacing between the numbers and the trailing pipes in the latter two should not be present. You can also still check my enwiki sandbox in VE to replicate it. IceWelder (talk) 12:50, 16 March 2021 (UTC)

Slight update: A user reported that they also encountered this behaviour through the iOS Wikipedia app. I thus downloaded the Android app and I could reproduce the error there. Either the VE and the mobile apps make the same mistake or the error really does lie within formatnum itself. IceWelder (talk) 08:07, 17 March 2021 (UTC)

Ok, this should be fixed now Parsoid/Deployments#July 13-15: V0.14.0-a7 as part of 1.37.0-wmf14

You might need to ?action=purge a page to clear the cached result though Arlolra (talk) 15:59, 16 July 2021 (UTC)

NUMBEROFCATEGORIES

• Would like to see {{NUMBEROFCATEGORIES}} available as a magic word. Nicole Sharp (talk) 18:50, 13 March 2021 (UTC)
  • I got it working using {{PAGESINNAMESPACE:14}} and $wgAllowSlowParserFunctions = true; but might still be good to have a shortcut magic word. Nicole Sharp (talk) 20:19, 13 March 2021 (UTC)

How can I assure that DEFAULTSORT really sorts as expected?

On my wikis, I intensely use DEFAULTSORT for articles about persons. But every now and then I see that they aren't properly sorted in their assigned categories. New articles or those with edited DEFAULTSORT keys are sometimes inserted at the end of a Character list on the relevant Category pages, like you may be able to see e.g. here on my wiki. What am I doing wrong or is there any script or bot to resort articles according to their default sorting? (I'm not talking about proper sorting of e.g. German umlauts.) --Ulf Dunkel (talk) 10:56, 17 April 2021 (UTC)

What specific pages look wrong sorted? From a quick look I only see those: Hannes Moser and Moses Maimonides, but both are correct according to the DEFAULTSORT magic word. --Ciencia Al Poder (talk) 13:05, 17 April 2021 (UTC)

Please take a look at the Category:Male on my wiki once more. In almost every letter section, there are newer articles listed which are not sorted correctly. In A e.g., you will find Mikael Aktor and Simon Almkor sorted behind Bertrand Auvert. In B, you will find Gary Burlingame before Zachary Levi Balakoff, Felix Bölter (DEFAULTSORT here: "Boelter, Felix"), Shmuley Boteach, and Jordan Michael Bresson. It seems as if new articles or those with updated DEFAULTSORT values are sorted behind the last fitting entry for the alphanumerical sorting. So if there is a Balakoff behind Burlingame for any reason, the sorting mechanism drops the newer articles Bölter, Boteach and Bresson behind Balakoff, not between upper entries of the B section.--Ulf Dunkel (talk) 13:22, 17 September 2021 (UTC)

That's odd and reminds me of some problems I'm having, though maybe in your case, it's a job queue that's not been resolved or certain jobs have been aborted. See https://www.mediawiki.org/wiki/Manual:Job_queue. Cavila 14:19, 17 September 2021 (UTC)

Thank you for hinting me to cron jobs, but I haven't installed any which didn't come with the normal MediaWiki installation. It seems as if the sortKey is only computed after the page source has been changed. I tried this with various articles and saw that it adds a table entry page_props.pp_propname = "defaultsort" to the database. But that doesn't resort the pages properly. I also wonder why the parameter page_props.pp_sortkey is NULL in all rows. I have no idea how to fix this sorting issue.--Ulf Dunkel (talk) 14:56, 17 September 2021 (UTC)

@Ulf Dunkel: As long as page_props.pp_value is non-empty, this table should be okay. Perhaps Manual:$wgCategoryCollation and Manual:updateCollation.php can help you. --Matěj Suchánek (talk) 08:12, 11 October 2021 (UTC)

Possible typo: "character and phase"

[3], shouldn't this be "character and phrase"? MarMi wiki (talk) 15:22, 4 June 2021 (UTC)

I removed the dubious "character and phase". Feel free to re-add if this indeed was correct. Taylor 49 (talk) 15:05, 17 March 2024 (UTC)

Collapsible TOC

Is there an equivalent to the __TOC__ magic word that will produce a collapsible table of contents, e.g. in the same manner as the {{Collapse top}} template but a magic word and for TOCs only? –Tommy Kronkvist (talk), 15:07, 23 June 2021 (UTC).

@Tommy Kronkvist: Tables of contents are always collapsible, at least using the default Vector skin. Whether they’re initially collapsed or not depends on user preferences (even for logged-out users) and you can’t control it, although you can wrap it in an initially collapsed outer box so that it’s collapsed for all users on page load, see e.g. m:Template:TOC hidden. (It’s not a magic word, though.) —Tacsipacsi (talk) 15:16, 23 June 2021 (UTC)

@Tacsipacsi: Thanks: guess I have to have a look in my Vector skin- and related preference settings. Thanks for the link as well. –Tommy Kronkvist (talk), 15:22, 23 June 2021 (UTC).

Self-closing tag

It is possible to generate a self-closing tag? For example, <ref name="Einstein1905" />

Currently, when I enter null content, the separate closing is added:

• {{#tag:ref||name=Einstein1905}} → <ref name="Einstein1905"></ref>

The code functions as expected BTW, not a bug. -DePiep (talk) 19:16, 28 June 2021 (UTC)

I wonder why you think the code works fine? When I try your example, it drops an error: "Cite error: Invalid <ref> tag; no text was provided for refs named Einstein1905".--Ulf Dunkel (talk) 13:22, 17 September 2021 (UTC)

Of course it won’t work on its own, you need to define the content somewhere. For example:

Lorem ipsum{{#tag:ref||name=Einstein1905}} dolor sit amet.{{#tag:ref|{{Cite journal|last=Einstein|first=A.|date=1905|title=Zur Elektrodynamik bewegter Körper|journal=Annalen der Physik|volume=17|pages=891–921}}|name=Einstein1905}}

correctly renders as

Lorem ipsum dolor sit amet.

Tacsipacsi (talk) 22:05, 17 September 2021 (UTC)

References

{{#language:xx|zgh}}

{{#language:ary|zgh}} gives Moroccan Arabic instead of ⵜⴰⵄⵕⴰⴱⵜ ⵜⴰⵎⵖⵔⵉⴱⵉⵜ, and {{#language:he|zgh}} gives Hebrew instead of ⵜⴰⵄⵉⴱⵔⵉⵜ, same thing for some other language names are untranslated or translated incorrectly. How can this be fixed? And where?--Brahim-essaidi (talk) 21:46, 2 October 2021 (UTC)

noexternallanglinks

Where is the right place to put {{noexternallanglinks}} in this manual? (more info: Wikibase/Installation/Advanced_configuration#noexternallanglinks) --Valerio Bozzolan (talk) 17:04, 31 October 2021 (UTC)

{{=}}

Tracked in Phabricator
Task T91154

This looks also like magic word now, document it? Liuxinyu970226 (talk) 10:51, 16 November 2021 (UTC)

When I write =, Template:= shows up on the Templates used in this preview list. So no, it’s not a magic word (at least not on mediawiki.org). —Tacsipacsi (talk) 20:39, 16 November 2021 (UTC)

No, not yet. --Matěj Suchánek (talk) 11:10, 20 November 2021 (UTC)

@User:Matěj Suchánek: Now it is, since year 2022. Please delete Template:=. Taylor 49 (talk) 02:53, 1 December 2024 (UTC)

Template:! also exists, even though it has existed as a magic word for a decade. I think they are still useful for documentation, also they continue to be used on a few pages through redirects (as well as in some Flow posts directly – silly Flow makes the “parser cache” never expire, because it reinvents the wheel as always and doesn’t actually use the parser cache). —Tacsipacsi (talk) 17:35, 1 December 2024 (UTC)

Indeed all Template:! Template:= and Module:No globals should get deleted on all wikis, otherwise the related changes to MediaWiki do not make sence. You want {{=}} to be a parser function in order to get rid of Template:=. If you want to keep "Template:=" together with all its overcomplicated testing, documentation, categorization etc infrastructure anyway, then there is no need to make {{=}} a parser function. Wiki is getting overintricate and unmaitainable. Taylor 49 (talk) 17:28, 3 December 2024 (UTC)

What overcomplicated testing, documentation, categorization etc infrastructure? Neither template has any testing infrastructure, and documentation and categorization uses just the standard infrastructure – {{Documentation }} exists anyway, Category:Workaround templates contains more than just these two templates.

The main gain of these parser functions is that they work out of the box on new and small wikis, not that we can delete two established and still-used templates on an established and large wiki. The wiki won’t get overintricate and unmaitainable by not doing something (if we don’t do anything, things stay as overintricate and unmaintainable as they were, not more), and these templates aren’t intricate or unmaintainable anyway. —Tacsipacsi (talk) 20:12, 4 December 2024 (UTC)

{{=}} is not "used" anywhere at all unless you write {{Template:=}} for some reason such as faking usefulness. And new and small wikis still frequently create Template:! Template:= Module:No globals with the only reason "many wikis have those". About "get overintricate and unmaitainable by not doing something" ... new "features" are being added daily, so "not doing" is ultimately not the current case. And it's more diffucult to find out why something is broken if it can be broken on 100 places, compared with say 20 years ago when there were only two locations to memorize and check. Taylor 49 (talk) 15:48, 7 December 2024 (UTC)

I didn’t state {{=}} is used; I stated Template:= is used. As a redirect target, for example.

For the copying, I don’t think deletion is the best solution. If we delete Template:=, which currently exists on 280 wikis, it’ll exist on 279 wikis – still “many wikis have this”, people will simply copy it from another one. If we instead put a huge warning on it (which has already happened for Template:!, and could happen for Template:= as well), people who try to copy it from mediawiki.org will hopefully stop instead of copying it from somewhere else. —Tacsipacsi (talk) 21:35, 7 December 2024 (UTC)

How to set precision with {{formatnum}}

I see that the {{formatnum}} template was replaced by the {{formatnum}} magic word. With the template I could set the precision of a number, but I don't know how to do it with the magic word.

How do I format a number to limit it to e.g. 2 decimal places?

i.e. 10.11111 -> 10.11

Any help would be appreciated. - 196.10.113.11 08:55, 25 November 2021 (UTC)

{{formatnum:{{#expr: (10.11111) round 2 }}}} → 10.11. --Matěj Suchánek (talk) 09:14, 28 November 2021 (UTC)

List of available options for int?

{{int:lang}} gives me "en". what other things can i put in place of "lang"? where can i find the full list? RZuo (talk) 22:36, 9 April 2022 (UTC)

Any MediaWiki message. This includes the software-defined messages listed on Special:AllMessages (on mediawiki.org, it’s about 27,000 messages) as well as custom messages like lang or cat-browser (see all pages in the MediaWiki namespace; theoretically any message could be included with {{int:}}, but pages like gadgets’ codes make little sense). —Tacsipacsi (talk) 01:11, 10 April 2022 (UTC)

How to access the section number?

This is section number 78 of the talk page, as shown by the url under the link edit source:

https://www.mediawiki.org/w/index.php?title=Help_talk:Magic_words&action=edit&section=78.

Can I get this number via a magic word? --GrandEscogriffe (talk) 19:47, 19 April 2022 (UTC)

Not yet: phab:T3605. --Matěj Suchánek (talk) 09:18, 22 April 2022 (UTC)

#ask with [[Property::{{PAGENAME}}]]

Hello !

On "this page", I want to display all the "Contact" which have "this page" in their Template.

We have :

Contact de test in the Contact category, with

{{Contact

...

|WorkGroup=This page

...

}}

So I tried with {{PAGENAME}} :

{{#ask: [[Category:Contact]] [[CommunityOfInterest::{{PAGENAME}}]] |?Function }}

Which should ask this :

{{#ask: [[Category:Contact]] [[CommunityOfInterest::This page]] |?Function }}

But it does not work.

Does anyone have an idea ?

Thank you ! Ypermat (talk) 14:28, 6 July 2022 (UTC)

tag: confusing example

See #Miscellaneous, #tag:. The example is:

"Example for the Cite extension's ‎<ref> tags:

{{#tag:ref|Citation on Magic words. |name = "multiple"}} → " &tc.

At the same time, there is a red warning text:

"... You must write {{#tag:tagname||attribute1=value1|attribute2=value2}} to pass an empty content. No leading or trailing space of the text content are permitted between the pipe characters || before attribute1.

This seems contradicting: in the example, there is a space in "words. |name", at exactly the forbidden place. Is there something I am missing? DePiep (talk) 12:06, 26 October 2022 (UTC)

• The warning is I think specifically just about empty content, as leading/trailing space is preserved in the content. It is telling you that you must use || rather than | | if you want the content to be empty. For example, "{{#tag:span||}}" correctly produces "" (empty content) while "{{#tag:span| |}}" produces " " (non-empty content consisting of a space character). --Clump (talk) 12:57, 26 October 2022 (UTC)

Version at Meta has been soft-redirected to here

The version of this help page at Meta (i.e., m:Help:Magic words) has been deprecated in favor of this page. See the last "useful" revision of that page, which contains some things that might need to be merged into this page. (See also the corresponding talk page over there.) - dcljr (talk) 08:31, 5 March 2023 (UTC)

Do note that Help pages on Meta are under an differently license from help pages on MediaWiki.org (see Project:PD help), so wikitext can't be directly copied from there to here. * Pppery * _{it has begun} 14:45, 5 March 2023 (UTC)

Basename without filename extension

Mention how

• File:blabla.jpg could link to
• File:blabla 2.jpg via a magic word.

It seems a new e.g., GUTNAME template would need to be invented, that strips off the filename extension:

• [[:{{FULLGUTNAME}} 2.jpg]]

See also task T6421. Jidanni (talk) 06:42, 6 March 2023 (UTC)

What are you actually suggesting here? If you're not referring to a currently existing feature that needs to be documented (better or at all) on this help page, then you're commenting in the wrong place. - dcljr (talk) 22:56, 6 March 2023 (UTC)

OK then it is a feature request, like the many seen above on this discussion page. Jidanni (talk) 03:18, 13 March 2023 (UTC)

Feature requests can be discussed at Project:Support desk or m:Tech and then requested formally at the Phabricator. - dcljr (talk) 00:03, 14 March 2023 (UTC)

Here's his Support Desk topic: Topic:Xeaqr0k6zp1q6dsh. Jonathan3 (talk) 23:40, 18 March 2023 (UTC)

I don't see why this requires a magic word. Just use a lua module to parse the text and extract the substring up to the last . in the filename ? —TheDJ (Not WMF) (talk • contribs) 15:10, 14 March 2023 (UTC)

Well I read w:Wikipedia:Lua and it doesn't seem to work like {{PAGENAME}} that will even work for users not logged in viewing it. Well, OK, please put an example below. Thanks! Jidanni (talk) 00:22, 18 March 2023 (UTC)

No support questions. ... Right.

>"Please do not post support questions here."

Really!!

O well. Guess I just have to swallow my "{{NAMESPACE}}:{{PAGENAME}}" =?=> "{{{Self}}}", debugging related, question than. MvGulik (talk) 09:24, 10 June 2023 (UTC)

Support questions can be asked at Project:Support desk. - dcljr (talk) 03:22, 11 June 2023 (UTC)

This page suddenly disappearing from Google Search

No idea why, but this page has suddenly dropped out of google search. I don't see a __NOINDEX__ on it... in the past, whenever I typed "mediawiki" and "magic words" this help page was always the #1 hit, now the "Manual" namespace page shows up instead, or Wikipedia and other site versions, despite different browser/users. In other words, something is causing this page to drop out of Google Search. I'll bookmark it instead, but I'd imagine this would be a pretty widely used page. FrozenPlum (talk) 17:56, 14 September 2023 (UTC)

Yep, same thing with "Help:Templates", which has also dropped out of Google Search. For some reason, some pretty key pages are gettig de-indexed by Google. This is going to be super annoying. FrozenPlum (talk) 18:03, 14 September 2023 (UTC)

Both help pages do have "Indexing by robots: Allowed" [4][5]. --Matěj Suchánek (talk) 09:54, 17 September 2023 (UTC)

PAGENAME does not respect DISPLAYTITLE

If you are using {{DISPLAYTITLE}} to format a page name to have a leading lower case character (or really any formatting at all) the {{PAGENAME}} does not obey the display rule.

I think this is a bug.

47.186.29.164 47.186.29.164 01:12, 14 October 2023 (UTC)

This is not a bug, but a missing feature. However, if you want that behavior, there's an extension that provides it. See Extension:Display Title. Ciencia Al Poder (talk) 16:14, 14 October 2023 (UTC)

I wouldn’t even call it a missing feature. In many, probably most, cases, {{PAGENAME}} is used to get the page title for further processing (finding a subpage with the same name, its root page, using it in a conditional expression etc.). In these cases, any formatting is more harmful than useful. In addition, what if the wikitext fragment containing {{PAGENAME}} is processed before the wikitext fragment containing {{DISPLAYTITLE:…}}? It would lead to inconsistent results. —Tacsipacsi (talk) 13:39, 15 October 2023 (UTC)

Indeed, adding such a "feature" would cause a huge mess. One could add a separate parser function "{{READOUTTITLE}}" ... but let me guess that some could object this idea, since it could be abused as a replacement for the malicious Extension:Variables. Taylor 49 (talk) 14:58, 17 March 2024 (UTC)

Add a new magic word although if it is expensive

I suggest that the magic word "{{TRANSCLUSIONCOUNT}}" can be added, I'm sure it's quite expensive, what that function does is display the amount of transclusions it makes to itself, return 0 if is displayed on the source page, but return 1 if displayed from a page that transcludes the source page, return 2 if displayed from the page that transcludes the previous one, and so on, so it is possible to use it with "{{#ifexpr}}" to mainly prevent the main template containing that magic word from being transcluded a certain number of times 95.143.193.15 19:43, 4 October 2024 (UTC)

Usage of class="noresize"

The tables on this page appear to have recently been wrapped in class="noresize". I'm not entirely sure what the goal of doing this is (I imagine it's some kind of mobile support, although no information was provided in the edit summaries), but on Firefox on desktop it causes the tables to be enclosed in divs with scrollbars, with a pointless large block of space underneath them. It makes scrolling through this page substantially more inconvenient. I will revert this change, but if @Jdlrobson could explain what the purpose of these changes are, perhaps there is a way to apply the desired functionality without breaking the page on Firefox. SnorlaxMonster (talk) 06:12, 19 October 2024 (UTC)

The noresize class is documented here: Manual:Interface/IDs_and_classes and is in the process of being applied automatically by MediaWiki, so I've added this with that in mind.

Without it, tables overlap the sidebar on Vector 2022 and break the mobile site at certain resolutions as they are too large for the page: https://phabricator.wikimedia.org/F57626109 - impacting accessibility of important links

The issue with Firefox is new, looks like a cosmetic issue rather than a functional issue (I agree it's broken but inaccessible links is a greater bug than a block of whitespace at bottom of the element) and looks like it could be an upstream browser bug to me. Let's file a ticket and report that on Phabricator. Jdlrobson (talk) 16:38, 19 October 2024 (UTC)

URL data matching mobile or desktop?

fullurl, canonicalurl, etc. dont produce mobile view links while clicking them on mobile. is there a method to match output to the domain? if viewing on xx.m.yy.org the link should be the same and not xx.yy.org . RoyZuo (talk) 13:20, 29 November 2024 (UTC)

I don’t think so – I’m pretty sure the parser cache is shared between desktop and mobile, so we have to pick a domain name to be saved there. (Normal internal links work because they contain no domain name at all, just domain-relative /wiki/….) By the way, the whole mobile domain thing is a big hack – ideally, we’d have software that works greatly on small and large screens, with keyboard, mouse and touchscreen, and we wouldn’t have a separate mobile domain at all. But we’re far from that. —Tacsipacsi (talk) 22:04, 30 November 2024 (UTC)