home.social

#mediawiki — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #mediawiki, aggregated by home.social.

  1. My #MediaWiki medium heat take on #LuaModules is that we don't make them magic ✨️ enough because most modules are converted from wikitext.

    With substitution, Lua allows us to transform any user input into any output we want! So more templates should just accept more types of inputs.

    For instance, my ru.wikipedia.org/wiki/Шаблон:Г needs Wikidata item ID to work but it accepts 1) plain old page titles, 2) wikitext with links, 3) interwiki links, and 4) pasted URLs to Wikidata, and just transforms them.

  2. My #MediaWiki medium heat take on #LuaModules is that we don't make them magic ✨️ enough because most modules are converted from wikitext.

    With substitution, Lua allows us to transform any user input into any output we want! So more templates should just accept more types of inputs.

    For instance, my ru.wikipedia.org/wiki/Шаблон:Г needs Wikidata item ID to work but it accepts 1) plain old page titles, 2) wikitext with links, 3) interwiki links, and 4) pasted URLs to Wikidata, and just transforms them.

  3. My #MediaWiki medium heat take on #LuaModules is that we don't make them magic ✨️ enough because most modules are converted from wikitext.

    With substitution, Lua allows us to transform any user input into any output we want! So more templates should just accept more types of inputs.

    For instance, my ru.wikipedia.org/wiki/Шаблон:Г needs Wikidata item ID to work but it accepts 1) plain old page titles, 2) wikitext with links, 3) interwiki links, and 4) pasted URLs to Wikidata, and just transforms them.

  4. My #MediaWiki medium heat take on #LuaModules is that we don't make them magic ✨️ enough because most modules are converted from wikitext.

    With substitution, Lua allows us to transform any user input into any output we want! So more templates should just accept more types of inputs.

    For instance, my ru.wikipedia.org/wiki/Шаблон:Г needs Wikidata item ID to work but it accepts 1) plain old page titles, 2) wikitext with links, 3) interwiki links, and 4) pasted URLs to Wikidata, and just transforms them.

  5. My #MediaWiki medium heat take on #LuaModules is that we don't make them magic ✨️ enough because most modules are converted from wikitext.

    With substitution, Lua allows us to transform any user input into any output we want! So more templates should just accept more types of inputs.

    For instance, my ru.wikipedia.org/wiki/Шаблон:Г needs Wikidata item ID to work but it accepts 1) plain old page titles, 2) wikitext with links, 3) interwiki links, and 4) pasted URLs to Wikidata, and just transforms them.

  6. If you find a wiki page seemingly from a MediaWiki, but without a "View History" link, here’s a tip for how to find it.

    Example: the old HTML Working Group (WG) wiki at W3C
    * https://www.w3.org/html/wg/wiki/Main_Page

    Note the warning at the top: “This wiki has been archived and is now read-only.”

    Aside: many years ago, each group at W3C had their own MediaWiki install, in a different path and disconnected from both every other group’s wiki, and the top level general W3C wiki (http://www.w3.org/wiki). I pointed out this made no sense, since groups more often than not had overlaps in documentation, technology, practices and would of course benefit from wiki-syntax interlinking. This was never really fixed at an infrastructure level, it had to be fixed at a cultural level, convincing working groups one by one to use the general W3C wiki, and ignore any WG-specific wiki. The other apparent hazard of having a WG-specific wiki is they can be (and have been) shutdown or frozen (like the example given), breaking any chance of fixing typos, other minor errors, or creating redirects to where the work has been restarted subsequently.

    Note also the lack of a "View History" link as a tab top right adjacent to the content area (or anywhere else on the page).

    This is likely because the page has been archived as static HTML, preserving the markup and styling, but removing any editing or history browsing or diffing or any other form submission based features that would require a dynamic backend like MediaWiki.

    Here is how to find such a page’s history:
    1. Find the “last modified” date in the footer and note the year, month, day.
    2. Check for the page in the Internet Archive, e.g. prepend its URL with "https://web.archive.org/web/"
    3. If there’s (presumably) still no "View History" link, look at the year month day numbers in the URL — the first 8 digits after "/web/" E.g. "20260512" for this page as of today.
    4. Replace those 8 digits with the year, month, day noted in step 1, with 4, 2, 2 digit numbers respectively. E.g. "20150623" for this page. Press return to load that.
    5. You may have to "close toolbar" the (x) in the top right of the page content.
    6. Note the "View History" link, click it.

    Presto, view history on a former MediaWiki page that’s been archived to static HTML.

    This is my second #WASSUP #WASSUP2026 post, and first useful web tip.

    #wiki #wikiPage #MediaWiki #ViewHistory #W3C #HTMLWG #wikiArchived #readOnly #staticHTML #InternetArchive #WaybackMachine #LastModified

  7. Mein kleines Refactoring-Projekt der heimischen Infrastruktur zieht immer größere Kreise.

    Wenn ich den fetten Server nachts abschalten will, brauche ich 24x7 Ersatz in klein. Dafür habe ich gestern eine APU 1D4 mit #Debian 13 bestückt.

    Stellt sich raus, dass ich vor 10 Jahren einen #Datenlogger ebenfalls auf einer APU 1D4 installiert habe, der alle 5 Minuten 24x7 die Zählerstände des digitalen Stromzählers und vom #Wechselrichter in eine Datenbank protokolliert. Ach ja, langsam erinnere ich mich wieder.

    Die Software dafür läuft im wesentlichen seit dieser Zeit ununterbrochen (ausser Stromausfällen) in meinem hinterletzen Subnetz. Zum Glück hatte ich damals in einem #MediaWiki dokumentiert, was ich damals verbrochen hatte.

    $ grep PRETTY /etc/os-release
    PRETTY_NAME="CentOS Linux 7 (Core)"

    Da muss ich dann jetzt wohl auch mal ran...

    #EinmalMitProfis

  8. @firefoxwebdevs I believe this would reach its full potential if supported and implemented also for headers in @mediawiki transcluded pages/templates, and in situations with no front-end JavaScript, for example when computing the heading level inside @11ty includes.

    #JavaScript #HTML #11ty #BuildAwesome #MediaWiki