Jump to content

Commons:Village pump/Technical

Add topic
From Wikimedia Commons, the free media repository

Shortcuts: COM:VP/T • COM:VPT

Welcome to the Village pump technical section
Technical discussion
Village pump/Technical
 Bug reports
 Code review
Tools
 Tools/Directory
 Idea Lab



This page is used for technical questions relating to the tools, gadgets, or other technical issues about Commons; it is distinguished from the main Village pump, which handles community-wide discussion of all kinds. The page may also be used to advertise significant discussions taking place elsewhere, such as on the talk page of a Commons policy. Recent sections with no replies for 30 days and sections tagged with {{Section resolved|1=--~~~~}} may be archived; for old discussions, see the archives; recent archives: /Archive/2026/07 /Archive/2026/08.

Please note
 
SpBot archives all sections tagged with {{Section resolved|1=~~~~}} after 1 day and sections whose most recent comment is older than 30 days.

Solution idea for category name problem

[edit]

before we finally get a system that allows multiple / multilingual category names, i just had an idea to patch on top of what we have now.

what if in addition to showing the actual cat names, a tool shows additional names above them (like ruby text)? the additional names could be, that category's wikidata label (in a language chosen by the user), or defined on the cat page itself by using

{{en|...}}
{{fr|...}}
{{ja|...}}
...

it's a solution for people quarreling over whether a category name should be english or other native latin-alphabetic language. example: Category:FC Bayern München. with this tool, if the cat name is english, users can see the german name above so german users still know what it is; if the cat name is german, non german users can see the english (or any other language user chooses) name above.

it appears like this on a parent cat page containing it.

  FC Bayern Munich
> FC Bayern München (24 C, 88 F)

or

 FC Bayern München
> FC Bayern Munich (24 C, 88 F)

it's not just useful for multilingual name conflicts, but also if something has some very famous alternative names, e.g.

 Sears Tower
> Willis Tower (10 C, 35 F)

other considerations:

  1. if querying up to 200 pages every time someone opens a cat page is too much for the servers,
    1. the tool could query the page and show the alternative name only when the cursor hovers over the target.
    2. or there could be a cache set up on toolforge that stores cat pages queried and their alternative names, so if a page is queried recently, there's no need to query wikidata or commons again but the tool can directly fetch the cached data.
  2. this solution is ofc not compatible with current designs of hotcat, catalot, uploadwizard... which can only show the actual page name.

RoyZuo (talk) 15:20, 6 July 2026 (UTC)Reply

Closest thing I could think of would be to disable $wgRestrictDisplayTitle, but MediaWiki's manual page discourages that practice. Alternatively, the Translate extension could be used to create internationalization variations for categories, similar to what's already in place for pages in the Commons namespace. ForeverFlying (talk) 18:11, 5 August 2026 (UTC)Reply

So one option here, would be to do something like the following in the {{Wikidata infobox}} template:

{{DISPLAYTITLE:<span class="translated-title" title="TRANSLATED NAME OF CATEGORY FROM WIKIDATA HERE"><span class="translated-title-inner">{{FULLPAGENAME}}</span></span>}}

Then add the following CSS to either site CSS or your personal Special:MyPage/common.css

#firstHeading .translated-title::before {
  content: attr(title) " ";
}
#firstHeading .translated-title .translated-title-inner {
  font-size: 50%;
  color: grey;
}
#firstHeading .translated-title .translated-title-inner::before {
  content: '(';
}
#firstHeading .translated-title .translated-title-inner::after {
  content: ')';
}

Then we can get translated titles based on wikidata information on the category page.

As far as translating category names on individual pages, that is one of the potential benefits of showing categories in the information template (As i proposed at Commons:Village_pump/Proposals#Include_categories_and_depicts_in_the_{{Information}}_template), since we can potentially translate them in display. For example, see how the category Western Caucasus is shown in my example in english vs french.

That's not exactly what you are suggesting, but perhaps a step in the right direction. Bawolff (talk) 05:44, 21 August 2026 (UTC)Reply

Flickr2Commons

[edit]

Not all OK with Flickr2Сommons, sadly. I've uploaded File:Sunset in Vernazza 2018.jpg and File:Sunset in Hehuanshan, Taroko National Park.jpg, they have been transferred without date and {{Flickreview}} tag. I've added them manually. It is possible to use an old version? Юрий Д.К. 09:34, 8 July 2026 (UTC)Reply

@1Veertje. RoyZuo (talk) 13:45, 8 July 2026 (UTC)Reply
sorry, should have tested more. Submitted a patch that should fix this Vera (talk) 19:25, 8 July 2026 (UTC)Reply

Batch uploading audio files – bot or not?

[edit]

Hello! I want to upload a large amount of audio files, approximately 33 000, to Commons (pronunciations of Estonian words, if you're interested). For the sake of my mental and physical health, I don't really want to click the Upload Wizard 33 000 times to upload each file individually. There's gotta be a more efficient way to do that, perhaps it's a job that a bot can do? I know there are tools other than the Upload Wizard, which can be used to upload many files at once, but there's a few additional catches which lead me to think it's a job better left to a bot or a user more experienced here than me, who would know how to tackle these issues and not end up in a mental hospital in the process.

Firstly, the files need to be renamed. Currently they are just numbered, for example psv_08693.wav. Their names need to be reformatted to Et-{{{term}}}.ogg. Fortunately, there is, at least for one portion of the audio files, a .txt file which connects every one of those numbers to the corresponding terms. I imagine that a bot wouldn't have a problem with reading that and renaming the audio files accordingly. The other portion of the audio files also have an additional file, but it's in the .ndx format, which I can't open, so I'm not sure if it can be used in the same way, hopefully it can. The fun doesn't end there though – not only do the file names need to be converted from a numeric value to a text value as per the .txt file, but these text values in turn need even more changing. Nothing too extreme though, mostly just converting some symbols, like ` to *, with some exceptions, such as the symbol being removed altogether for words of 4 or less letters (it's more complicated than that but I will not go into it). The reason being that these audio files are intended to be used on Wiktionary, and the template that converts Estonian text to IPA uses different symbols when respelling, and the audio files should, ideally, be equal to the respelling in the first parameter of the pronunciation template. But it will also allow manual input of the audio files, so if you deem making this shortcut possible to be too complicated and time-consuming, then it's not the end of the world either. A couple of examples: psv_08693.wav → t`el'k → Et-tel'k.ogg, psv_01396.wav → hilja → Et-hilja.ogg (no changes needed), psv_00855.wav → el`ekter → Et-el*ekter.ogg, psv_00447.wav → `asf`alt → Et-*asf*alt.ogg, psvalg_1721.mp3 → k`un'stn`ik → Et-k*un'stnik.ogg (here, again ideally, delete the symbol if the term ends in -ik, but as I said, no problem if that's too complicated, and there are a million exceptions to this anyway), psvalg_2109.mp3 → organisatsi`oon → Et-organisatsi*oon.ogg, psv_00931.wav → enآ´ergiline → Et-en"ergiline.ogg (´ becomes "), psv_00028.wav → abi+elu → Et-abi-elu.ogg (+ becomes -, compound words), psv_07746.wav → sini+m`ust+v`alge → Et-sini-must-v*alge.ogg (components of a compound word are treated individually, so the rule of removing ` from 2-4 letter words applies). Every other symbol other than letters, spaces, and those symbols mentioned, should be removed. A final note: some forms are exactly the same and have the same pronunciation, therefore they have only one file for all of the forms in question, but the .txt file lists them multiple times, although with the same file name, just on different lines.

The second issue is one of categorization. The files need to be categorized as [[Category:Estonian pronunciation|<first letter of the term>]] or [[Category:Estonian pronunciation|<term>]] (I think it's the same). This can probably be solved with some non-back-breaking code, but it should still be done by someone who knows what they're doing.

Thirdly, the files are either in .wav or .mp3 format. They need to be converted to .ogg. This doesn't seem like a nightmare on its own, I could probably do that by myself, but combined with all these other things it doesn't bring me much optimism.

So in conclusion, I'm looking for somebody who is willing to and able to help me with uploading these files. It would be fantastic if somebody could do it for me, and if not, then could at least explain to me the necessary steps, as I genuinely have no clue. I understand if this seems like a lot to ask, we can do compromises on the file naming part, if necessary, but I really hope somebody here has the necessary skills and readiness to take on this task in some format, either with a bot or not with a bot. Joonas07 (talk) 09:37, 12 July 2026 (UTC)Reply

 Comment MP3 is now accepted. - Jmabel ! talk 18:02, 12 July 2026 (UTC)Reply
Does require having the autopatrol right, which you will need to request (for you or a bot) —TheDJ (talkcontribs) 18:59, 12 July 2026 (UTC)Reply
Great, that's one problem less. But I'm still looking for someone who would generally know what they're doing, especially if some kind of autopatrol rights have to be requested, it's better to have a user who has some kind of prior experience. Seeing as nobody has answered me here yet, could you maybe suggest some users that I could contact directly and ask if they are ready to do it? Apologies if I'm being intrusive, but I really need help here. Joonas07 (talk) 09:36, 15 July 2026 (UTC)Reply
Are the files other than the actual audio small enough that you can pack them all into an archive (eg. a zip file) and upload them to somewhere we can look at? I think seeing the real data would help us prepare for this task.
It intimidates me that we'll have to derive the upload filenames according to conventions of dictionary in a foreign language that I'm not familiar with. So you'll probably have to explain these transformation rules and proofread the derived filenames, possibly in multiple passes until we get all the names right.
Also please tell us about the legal parts, as in licenses for copyright and database rights. I know this can be annoying, but I think anyone who wants to help will want to be really sure about them before uploading tens of thousands of audio files. Though of course we won't upload all of them at the same time, instead we'd upload a smaller selection first and then allow users some time to tell us what we did wrong. – b_jonas 03:45, 22 July 2026 (UTC)Reply
Sorry for not replying, fortunately, I found a user already who can help me, but thanks for offering! Joonas07 (talk) 16:19, 28 July 2026 (UTC)Reply
But I'll keep you in mind, in case it won't work out with that other user after all. There's the issue of bot permissions and when it comes to converting the names, I'm not sure if it can just be done with find and replace. So I might still contact you at some point, if you are okay with that. Joonas07 (talk) 16:30, 28 July 2026 (UTC)Reply
For the record, all the files including the index are packed into three zip files, and the files are CC-BY-4.0, so we can use them for educational purposes. Joonas07 (talk) 16:37, 28 July 2026 (UTC)Reply
Wait, “psv_00931.wav → enآ´ergiline → Et-en"ergiline.ogg”? There's a letter from arabic script in the middle of that one! I hope there's a semi-reasonable explanation for this one, like it's an old database from back when Unicode wasn't as prevalent as now and some characters got replaced by inappropriate ones during the conversion. – b_jonas 05:07, 22 July 2026 (UTC)Reply
Yes, I noticed that too, but I don't know where it came from. Seems like it appears only before the symbol ´, so probably some Unicode stuff like you said. Joonas07 (talk) 16:28, 28 July 2026 (UTC)Reply

Outdated pictures of the day

[edit]

I've been seeing reports that the picture of the day is sometimes showing POTDs from previous days. Are the caches for the Main Page not always getting updated? Clarinetguy097 (talk) 16:47, 21 July 2026 (UTC)Reply

  • Here is an example where both POTD and MOTD are affected: the POTD and MOTD for 11th July are repeated in archive.org from 11-15th July: 11th, 12th, 13th, 14th, 15th July. Only on 16th July did POTD and MOTD get updated. --Tagooty (talk) 15:42, 22 July 2026 (UTC)Reply
    i vaguely remember that @Schlurcher's bot purges the main page every day after midnight. sorry if wrong mention.
    2 potential problems:
    1. not sure when exactly the purge happens. might it help if it's done a few minutes after 00:00 instead of exactly midnight?
    2. still, i have a vague impression that pages quite often look different to logged in and logged out users, probably due to how the website is cached and delivered by wmf.
    RoyZuo (talk) 13:16, 23 July 2026 (UTC)Reply
    @Clarinetguy097, @RoyZuo, @Tagooty: RoyZuo remembers correctly SchlurcherBot purges all the different language versions of the main page every day, around 1am. However, SchlurcherBot was down during time period noted. I did not see this before, but it seems this is a faily critical task to do, so maybe we need a second bot, that does the same or another backup solution. --Schlurcher (talk) 15:08, 23 July 2026 (UTC)Reply
    thx a lot.
    i agree that having a 2nd bot to double tap is essential.
    but, maybe wmf should also purge the highest traffic pages? for pages that are frequently visited, or contain important info (such as legal stuff or time-sensitive stuff), maybe wmf should purge them so they make sure pages delivered to users are up to date?
    maybe a page should be automatically purged whenever it hits a certain view count? purged every 100k-th visit. or purged daily/hourly when a page accumulates 100k visits over 24h/3d/7d...? just some random ideas. RoyZuo (talk) 16:45, 23 July 2026 (UTC)Reply
    Thanks for the explanations. Given the time-sensitive nature, I think the best solution is for the bot that updates POTD and MOTD at midnight UTC to purge immediately. Leaving the task to another bot that usually, but not always, runs 1 hour later is risky and imprecise. Tagooty (talk) 04:02, 25 July 2026 (UTC)Reply
    see if the problem gets worked on. RoyZuo (talk) 11:12, 25 July 2026 (UTC)Reply

Automatic categories for Template:Translate

[edit]

Scribunto now can return a list of languages translated in an SVG file (T259018), and so it's possible to automate the addition of files to Category:SVG by language (as requested in T407647). I'm not quite sure where the best place would be for this feature (maybe in Module:FileMedia, although that doesn't really deal with categories; then called from {{Translate}}), but something like the following should be possible:

    local langs = mw.title.new( 'File:Tulunadu.svg' ).file.languages
    local cats = ''
    for _,langCode in pairs( langs ) do
        local langName = mw.language.fetchLanguageName( langCode, 'en' )
        cats = cats .. '[[Category:' .. langName .. '-language SVG]]\n'
    end
    return cats

Pinging @Booksmurf who's working on related stuff.
Sam Wilson 02:07, 22 July 2026 (UTC)Reply

I have created this module and its doc in relation to:
https://meta.wikimedia.org/wiki/Community_Wishlist/W528 @docjames
https://commons.wikimedia.org/wiki/Module:Translate
As per Sam's comment, I do not know where would be the place place for this, so if any admin could help figuring out the next step or giving recommendations, it would be greatly appreciated!
Thank you Booksmurf (talk) 09:27, 1 August 2026 (UTC)Reply

Last month in technical changes

[edit]

Following Wikimania 2026, another overview of technical changes in the media support corner of MediaWiki, covering activity since June 17th.

  • A big new development: WMF has created a new dedicated Media Platform Team on July 1st, to own how we store, manipulate, stream, and display media going forward — Thumbor, file storage, and thumbnailing infrastructure all fall under it now. Their initial focus is thumbnail infrastructure sustainability.
  • The hCaptcha support added to UploadWizard has now also found its way to the Special:Upload page and there have been a few follow-up fixes, like a proper loading indicator T429322, T429287.
  • Failed uploads in UploadWizard can now be retried individually per file instead of only being able to retry the whole batch. An improvement contributed by @Amdrel. T410980.
  • The "Other resolutions" list on file pages, broken since the thumnbnail size limitations that I mentioned in the previous editions, should now be matching the actual available sizes. T401668
  • @Mimurawil added a new category suggestion framework for UploadWizard. This works towards Wishlist item W526. Geo- and full-text-based search suggestions are in active development T427047, T428065, T428066.
  • On mobile, you can now double-tap the left or right third of the video player to skip back or forward 10 seconds, similar to many video apps. This was wishlist item Wishlist item W468 and added by @Brooke Vibber. T432729.
  • I myself completed the changes for WebVTT subtitle support. Next week, it will be possible to create a Timed Text page with the .vtt file extension and you can use advanced styling and positioning of subtitles beyond what the SRT support already allowed for. T116154. If you want to help update the Timed Text documentation, that would be appreciated btw !
  • There has been some work on 3D models, with @Amdrel driving the latest changes. Thera are patches out for review across the extension, its metadata handling, and the thumbnailing backend. Without support of the foundation and people reviewing the patches, it is however hard to land these changes, so if you can help in any way with testing or reviewing, that is appreciated. T246901, T419711, T187844.
  • There were several Wikimedia Commons related projects presented at the Wikimania 2026 Hackathon closing ceremony, which you can watch back on Youtube

——TheDJ (talkcontribs) 12:33, 26 July 2026 (UTC)Reply

Can't wait to see the rollout of textured meshes on Commons :3 --PantheraLeo1359531 😺 (talk) 19:15, 1 August 2026 (UTC)Reply

Thumbnail sizes

[edit]

Still doesn't appear to be consistently matching available thumbnail sizes. See, for example, File:Seattle - Rainier Avenue overpass, 1954 (22651065004).gif. It offers a 1,920 × 1,481 pixel thumbnail of this 1,400 × 1,080 pixel image. - Jmabel ! talk 06:46, 28 July 2026 (UTC)Reply

Language missing

[edit]

I don't know where to reach out about this, so I'll do it here.
It is not possible to create a VTT file through the TimedText page (example) when there already is an SRT for the same language. It doesn't come up in the list. I suspect it is the same the other way around too. It also isn't a translation, so the text is technically wrong. Thanks. // Kakan spelar (talk) 16:16, 6 August 2026 (UTC)Reply

Newsletter?

[edit]
[edit]

I searched for "english national football team" in Commons. The URL is https://commons.wikimedia.org/w/index.php?search=english+national+football+team&title=Special%3AMediaSearch&type=image. To my surpise, the result set includes File:Lions Family Portrait Masai Mara.jpg which makes some sense in that the logo for the English national football team is three lions, but the metadata for the lions family portrait file shows nothing about English football that I can see. What might have happened with the search tool that caused this result? I'm wondering if the search tool leverages data from additional sources beyond the search text string and the image metadata. ↠Pine () 19:20, 26 July 2026 (UTC)Reply

I think the search uses Wikidata to try to match up words with concepts. My guess is that since "english national football team" is one of the aliases of Wikidata item England men's national association football team (Q47762), and so is "Three Lions", that those are considered completely equivalent terms by something within the search engine, and that picture has the text "three lions" in its description. That's a guess; I'm not sure where (or if) it's documented exactly how the search finds synonyms or related concepts. — PeterCooperJr (talk) 17:18, 31 July 2026 (UTC)Reply
@PeterCooperJr: thanks for the comment. That may be what happened. See mw:Help:MediaSearch, mw:MediaSearch and Commons talk:Media search. By the way @JTweed-WMF: you appear to be a member of the newly-created mw:Media Platform Team, and perhaps can answer my question authoritatively as well as the unanswered questions and comments on those talk pages. Thanks, ↠Pine () 17:27, 1 August 2026 (UTC)Reply
Thanks for the ping, I love this example and have passed it onto the right team internally. Search is not part of the new Media Platform team scope, but it is a problem space that I'm familiar with from a previous role so it's been interesting to take a look. In this case, the guess above is correct. The "three lions" term is coming from Wikidata and being included as a synonym. The reason that the image of lions is appearing so high in the results is because it's a previous picture of the day, which gives its relevance score a considerable boost. This is a good example of two good things that improve search in the aggregate (synonyms and a quality image signal) causing issues for one particular search. Optimising search is really hard and often things that fix one query will perform worse for others. The Search team is always trying to find the balance that gives the best results overall across all queries, but are now aware of this query and have taken this example on board. JTweed-WMF (talk) 14:26, 12 August 2026 (UTC)Reply

Tech News: 2026-31

[edit]

MediaWiki message delivery 18:46, 27 July 2026 (UTC)Reply

TSE-Dados-Abertos

[edit]

Hello,

I'd like to know if it would be possible to update the TSE-Dados-Abertos to improve the categorization. As it stands now, the main category have 18,915 files, with the categories by party and by year having to be added manually.

To empty the main category, could it be possible for the template to work like {{TSE-Dados-Abertos|party=(party acronym; example: MDB, PT, PDT...)|year=(electoral year)}}, with an warning message displayed if any or both options are missing? It would also be interesting if the template could link to the divulgacandcontas.tse.jus.br (as the source of the political party), as we can see here, for 2022. This could also help if this batch upload proposal ever happens. Erick Soares3 (talk) 16:22, 28 July 2026 (UTC)Reply

Redesign of photo navboxes

[edit]
Deutsch: Hi! Ich habe länger darüber nachgedacht, und möchte es jetzt gerne zur Aussprache bringen. Die Navigationsboxen von Fotos eines Jahrzehntes, wie bei Category:2019 photographs of Bavaria, kommt mir etwas klein vor. Wäre es eventuell besser, die Schriftgröße zu erhöhen und die Box über die ganze Breite der Weißfläche zu ziehen oder zumindest in eine zweite Zeile zu erweitern? Bin gespannt auf eure Meinungen. Grüße!


English: Hi! I've been thinking about this for a while, and I'd now like to bring it up for discussion. The navigation boxes for photos from a specific decade, like the one at Category:2019 photographs of Bavaria, seem a bit small to me. Would it perhaps be better to increase the font size and stretch the box across the entire width of the white space, or at least expand it to a second line? I’m curious to hear your thoughts. Best regards!

--PantheraLeo1359531 😺 (talk) 19:20, 1 August 2026 (UTC)Reply

@Reinhard Müller made a series of super neat templates Template:Category description. RoyZuo (talk) 09:27, 6 August 2026 (UTC)Reply

Tech News: 2026-32

[edit]

MediaWiki message delivery 19:43, 3 August 2026 (UTC)Reply

There was a problem during the HTTP request: 429 Too Many Requests

[edit]

I've never encountered this message before. Just trying to upload some of my photos from Flickr as normal, but I'm getting cut off at around 50 photos. Normally I have no issues, but now it won't let me upload anything. Is there a way to increase the limit of photos I can upload? PascalHD (talk) 16:28, 4 August 2026 (UTC)Reply

There is a related discussion here: Commons:Upload Wizard feedback#There was a problem during the HTTP request: 429 Too Many Requests. I also have encounter this same error couple days ago when I was trying to upload Flickr images, but I'm not sure what is the reason for this problem. Thanks. Tvpuppy (talk) 18:03, 4 August 2026 (UTC)Reply

Cleanup of bad pages created by broken template

[edit]

I just wanted to notify that due to a missing pipe in the template Commons:Deletion requests/Mass deletion request/header/i18n some bad pages got accidentally created such as:

I already fixed the template but Commons:Deletion requests/Mass deletion request/header/i18n/en and Commons:Deletion requests/Mass deletion request/header/i18n/de need to be updated accordingly.

The template had been broken since it got localized in 12 June 2026 by @Matrix, see Special:Diff/1230195948/cur & Special:Diff/1230196854. After checking all the deletion requests subpages from June till August I don't think any extra cleanup would be needed.


It's moon (talk) 14:05, 6 August 2026 (UTC)Reply

Moved the last one and deleted it. - Jmabel ! talk 20:30, 6 August 2026 (UTC)Reply
The next needed step is to mark Commons:Deletion requests/Mass deletion request/header/i18n for translation. I don't actually know how to do that. - Jmabel ! talk 20:36, 6 August 2026 (UTC)Reply
Someone has already requested this in Commons:Translators' noticeboard#Mark translation for Deletion requests/Mass deletion request/header/i18n.
According to Commons:Preparing a page for translation, it appears you need translation admin rights, which appear to be separate from regular admin. It's moon (talk) 23:28, 6 August 2026 (UTC)Reply
It looks like the translated subpages have been updated (Special:Diff/1257622398 & Special:Diff/1257622393) :)) It's moon (talk) 21:48, 7 August 2026 (UTC)Reply
Thanks to It's moon for noticing and fixing it, my bad :) —Matrix(!) ping onewhen replying {user - talk? - uselesscontributions} 11:46, 11 August 2026 (UTC)Reply

Tech News: 2026-33

[edit]

MediaWiki message delivery 20:42, 10 August 2026 (UTC)Reply

Data file too big to edit

[edit]

I've come across the strange case of Data:McGraw electric railway directory company observations.tab. DavidLevinson was able to upload it and later add categories, but my attempt to move it from Category:Rail transport in the United States to Category:Tram transport in the United States failed with the error The text you have submitted is 3,123.549 kilobytes long, which is more than the maximum of 2,048 kilobytes. David thinks it might be due to how Commons expands JSON. Does anyone have any way to make the category change - and are there any guardrails we can put up to prevent large data files becoming effectively uneditable? Pi.1415926535 (talk) 02:54, 11 August 2026 (UTC)Reply

I filed a request to increase the tabular data limit, but it wasn't successful (Commons:Village_pump/Proposals/Archive/2026/04#Increase_data_maximum_of_(tabular)_data_pages) --PantheraLeo1359531 😺 (talk) 19:46, 12 August 2026 (UTC)Reply
in theory it should have not been possible to do this in the first place. the file will become editable once enough content is removed from it to bring it below the limit. Bawolff (talk) 17:07, 15 August 2026 (UTC)Reply

Some strange changes in header menu like "nominate this image, copy, etc...

[edit]
Something seems to have gone wrong here.

.

The old version was located in a separate bar, as can be seen here.

Good day! I have once again noticed that something was unnecessarily changed regarding the display of files here. It used to be displayed on a separate line, but now it only appears under a single dot, and if you're lucky, the full menu actually shows up. Why was this changed, and for what reason? I’m not a programmer, just an admin, and I’m a creature of habit; but if something doesn't work properly, it shouldn't have been implemented in the first place. Best regards, זיו「Ziv」For love letters and other notes 14:09, 13 August 2026 (UTC)Reply

Confirmed. Fortunately, it happened just twice and appeared temporarily two days ago, then it was fine again and now it doesn't work properly... --Achim55 (talk) 15:54, 13 August 2026 (UTC)Reply

"File:HTML tooltip.png" fails to render if requested as a thumbnail

[edit]

I noticed that File:HTML tooltip.png is rendered as a blank/broken image if the image is used with the "thumbnail" parameter included, no matter if it's viewed on desktop or mobile device.

For some reason, it leads to the Too many thumbnail requests for failing image error message if visited via the browser's context menu, no matter what the requested thumbnail size is.

Comparison:

[[File:HTML tooltip.png]]:

[[File:HTML tooltip.png|thumb]]:

Is this a software bug that needs to be fixed, or do I have to update (to an image with different content) and upload a new version of the image with larger resolution, or should I do something else, like uploading a new image (like I mentioned before on this paragraph) with different file name and replacing all usage of the older one with the newer one? Exploraz (talk) 00:14, 16 August 2026 (UTC)Reply

According to this snapshot on the Wayback Machine of the downscaled thumbnail image URL, the server returned the following error message: Error while trying to fetch the image: Failed to convert image b"convert: no images defined `png:-' @ error/convert.c/ConvertImageCommand/3229.\n" Exploraz (talk) 06:03, 16 August 2026 (UTC)Reply
Looking at the error message, I think that includes an excerpt of an error message produced by en:ImageMagick.
Indeed, when I processed that image with magick convert HTML_tooltip.png -resize 250 250px_HTML_tooltip.png, an identical but more complete error message shows up:
convert: IDAT: invalid distance too far back `HTML_tooltip.png' @ error/png.c/MagickPNGError/1308.
convert: no images defined `250px_HTML_tooltip.png' @ error/deprecate.c/ConvertImageCommand/3375.
According to this GitHub issue coment, running pngfix --out=HTML_tooltip_fixed.png HTML_tooltip.png does indeed made magick convert HTML_tooltip_fixed.png -resize 250 250px_HTML_tooltip_fixed.png happy.
I'll overwrite it with the fixed image then.
Exploraz (talk) 06:24, 16 August 2026 (UTC)Reply
It turns out I can't overwrite it by myself with my current user rights (I see others used to be able to do that without autopatrol, but that got changed). If any autopatrols are running a Linux system, feel free to do that. —Exploraz (talk) 06:46, 16 August 2026 (UTC)Reply
@Exploraz: I've marked that file page to allow overwriting. Please let me know when you are done with it so I can revert that. - Jmabel ! talk 20:36, 16 August 2026 (UTC)Reply
✓ Done -- Thanks for helping, @Jmabel! It took about 3 minutes for the server-side cache to catch up, which has already passed now. Exploraz (talk) 05:02, 17 August 2026 (UTC)Reply
Turns out that this is not the only instance where this would happen in Commons:

Upload fails

[edit]

(copied here from Commons:Help desk) I no longer seem to be able to upload anything, either via the Upload Wizard or via the 'Upload a new version' command. I use Private Relay for privacy, so have to manually unhide my IP address to edit here. That worked fine until recently, but no longer seems to – in the upload wizard I get a message that says "The server responded with an invalid JSON document. This could be a problem with the API, or you could be using a proxy server that prevents you from uploading files"; in the 'Upload a new version' dialogue I just get shunted back to a blank upload page. Has something changed, or is there a way round this? Thanks, Justlettersandnumbers (talk) 08:02, 17 August 2026 (UTC)Reply

I've tried switching off Private Relay. That allows me to edit normally, without needing to unhide my IP, but doesn't allow me to upload. Any help welcome. Thanks, Justlettersandnumbers (talk) 08:05, 17 August 2026 (UTC)Reply
I've had nonexistent IP blocks be resolved by clearing cache and cookies. Try logging in in incognito with Private Relay off? JayCubby (talk) 20:18, 17 August 2026 (UTC)Reply

Tech News: 2026-34

[edit]

MediaWiki message delivery 21:00, 17 August 2026 (UTC)Reply

[edit]

politique de confidentialit'e ~2026-45131-71 (talk) 07:57, 19 August 2026 (UTC)Reply

Uploads do not appear permanently on my Watchlist

[edit]

My uploads do not appear permanently on my Watchlist (only for "1 year") though my Global Preferences are indicating they should. What do I do wrong or is this a hiccup in the system? JopkeB (talk) 09:13, 23 August 2026 (UTC)Reply

Edit requests for abusefilter admin and interface admin

[edit]

Can someone deal with this editrequest that has been open for a month and also this requested change to the abusefilter 103 ? —TheDJ (talkcontribs) 14:55, 24 August 2026 (UTC)Reply