This tends to solve most issues, including improper display of images, user-preferences not loading, and old versions of pages being shown.
No, we will not use JavaScript to set focus on the search box.
This would interfere with usability, accessibility, keyboard navigation and standard forms. See task 3864. There is an accesskey property on it (default to accesskey="f" in English). Logged-in users can enable the "Focus the cursor in the search bar on loading the Main Page" gadget in their preferences.
No, we will not add a spell-checker, or spell-checking bot.
You can use a web browser such as Firefox, which has a spell checker.
If you changed to another skin and cannot change back, use this link.
Alternatively, you can press Tab until the "Save" button is highlighted, and press Enter. Using Mozilla Firefox also seems to solve the problem.
If an image thumbnail is not showing, try purging its image description page.
If the image is from Wikimedia Commons, you might have to purge there too. If it doesn't work, try again before doing anything else. Some ad blockers, proxies, or firewalls block URLs containing /ad/ or ending in common executable suffixes. This can cause some images or articles to not appear.
How hard would it be to set up ORES for the Draft: space?
Hi all
So I've tried recently to find articles within the Draft: space from Wikipedia:Drafts to work on but I have found it very difficult because I have to go to each one individually to check what is in it.
I think that having a page with an ORES rating would make this much easier. E.g pages with an ORES rating of 0.9 are probably more likely to only need a small amount of work to get them to being publishable than an article with a rating of 0.4.
Would this be difficult to set up? I would be very happy to work on the documentation around it if someone technical could set up the actual rating system.
Hi John Cummings! I don't think it would be that hard. I'd like to see the Page Curation Tool start integrating ORES predictions. I think the most difficult work to make that happen is to make ORES predictions available inside of MediaWiki so tools like that can query them in a straightforward way. We've just finished work to make those predictions available. See Phab:T175757 (pending deployment). We'll be working to get that "draftquality" and "drafttopic" predictions in there soon too -- probably with a month. These predictions will allow you to quickly identify new drafts that are likely to be spam (G11), attack (G10), or vandalism (G3) (draftquality) and to categorize the rest by topic so you can review the drafts that align with your interests/expertise -- maybe even prioritizing the highest quality drafts first (wp10/articlequality). Once these are available in MediaWiki's internal databases, we need to find places to surface the predictions and to filter by them. Where would you want see the predictions surfaced? I think once we decide on a location to surface a prediction, we can build a better estimate of how much dev work it'll take. --Halfak (WMF) (talk) 21:49, 22 March 2018 (UTC)[reply]
Thanks very much @Halfak (WMF): and @SQL:. @SQL I never saw this before, very interesting :) @Halfak, this is very timely, I guess I'd really like to see Wikipedia:Drafts become a place for people to rescue drafts from drafts hell and get turned into articles, to become more of a place to collaborate on articles before (not only after) they are published. I think this will help less effort be wasted when people create a draft or when an Article for Deletion is moved to draft space. Do you have an example of what an ORES page might look like? I'm aware that there may be a very high volume of drafts so displaying them all may be quite a large page..... Thanks again both. John Cummings (talk) 15:33, 23 March 2018 (UTC)[reply]
I've been thinking about tagging draft talkpages with ORES wp10 predictions, and maybe draftquality as well. It wouldn't be terribly hard. SQLQuery me!03:25, 24 March 2018 (UTC)[reply]
SQL cool! Nice to see you experimenting with that. I think we need to find a good way to get this in front of people. E.g. if we could implement a sort-of backlog browser that could take advantage of some of these machine predictions and other filter criteria, we could help people to "rescue drafts from drafts hell and get turned into articles". John Cummings, I can't tell you how excited I am that you've got an eye for this too. I wonder if we could work with SQL to develop a sort-able list using the draft quality/wp10 models now. If we have a few people try to use it, that will help us better understand what is missing -- and what would be important for the draft hell backlog browser tool. :) I'm hoping to have the draft topic model deployed and well documented in time for the Wikimedia Hackathon in Barcelona. Either of you interested in attending or maybe just hacking with us remotely? --Halfak (WMF) (talk) 20:46, 26 March 2018 (UTC)[reply]
@SQL:, this looks amazing, one suggestion I have is to have table sorted by OK% as the default. Is is possible to have it updated more commonly than daily? Just thinking if it becomes popular then people may trip over each other? John Cummings (talk) 21:45, 26 March 2018 (UTC)[reply]
@SQL: super, is it possible to add definitions of the table columns at the top? I'm not clear what they all mean exactly. No idea how often they should update, is there something similar we could look at and see how often they update? John Cummings (talk) 09:52, 27 March 2018 (UTC)[reply]
@SQL: amazing work. My only other thoughts are about the order of the columns, I think the OK% column is probably the most useful so should go before the other % measurements (it appears to be a summary of the other percentages?). Just an observation that having one very long file name in the the Article column makes that line very long. These are all a bit of nitpicking rather than big things, I feel like its ready to be used when a short introduction is added to explain the purpose of the list, something like:
This list (generated by ORES) rates draft articles to help users find suitable articles to improve and then publish.
Any suggestions on the best way to integrate this into the Wikipedia:Drafts page? I'm aware its a very long list...perhaps a translusion with a scrolling section or a collapsable section?
Two things - one, I was thinking of removing "Vandalism" and "Attack" all together - they aren't really useful in this context. Two - maybe 'top 25 possible copyvios' / 'top 25 spam' / bottom (and/or top) 25 OK tables transcluded onto WP:DRAFTS? SQLQuery me!02:52, 29 March 2018 (UTC)[reply]
@SQL:, I think removing vandalism and attack are a good idea, I've been looking at the spam articles and many of them appear to be of notable subjects, just written in a tone that doesn't fit with Wikipedia so I'm not sure how useful that label is. Transcluding the top 25 would be a great idea, I guess if one gets popular you could upgrade to top 50... Let me know if there's anything I can do to help, once its live I'll be sure to let people know about it :) John Cummings (talk) 07:57, 29 March 2018 (UTC)[reply]
Hi all -- I'm Marshall Miller; I'm a new product manager on the Community Tech team with DannyH at WMF. As part of the follow-up to ACTRIAL, the Community Tech team is going to have some bandwidth over the next couple months to take on an improvement to the AfC process. I've been following along with the conversation on this talk page and elsewhere to learn about the biggest challenges facing AfC and the ideas for improving it.
The idea of integrating ORES scores or Copyvio checks in a streamlined fashion into the AfC workflow is one of the ideas being discussed on the talk page of that summary linked above, so I wanted to let you know in case you want to participate in that conversation over the course of the next week as we consolidate around some of the top ideas for improvements that WMF could help with.
Responded on the talk page. I am not an experienced articles for creation reviewer, so please be wary of some of my comments that might be misleading or counter constructive as a result. --Gryllida (talk) 04:02, 5 April 2018 (UTC)[reply]
the tables seem very useful. I want to track what actually happens to the articles there. Is there any automatic way of providing a historical record of what gets added/removed, or snapshots at intervals? (If not, I may do it manually). DGG ( talk ) 20:23, 11 April 2018 (UTC)[reply]
@SQL:, Wikipedia:Drafts looks great :), I just did a bit of testing with a few volunteers and we have a few requests of changes:
Please can the list update more frequently or have a manual update button (its not clear on how regularly the page updates).
Is it possible to link to longer lists than top 25 for each section? E.g Top 250? One of the testers wanted to start using the draft space as their place to create articles and getting to the Top 25 is a bit of a dark art.
The term 'OK drafts' is a bit confusing, is saying 'highest quality' and 'lowest quality' a good alternative?
@John Cummings: - Absolutely on all 3 points. It was updating daily, I've bumped that to ~every 6 hours. I can make some more subtables, possibly tomorrow. I don't think the OK% reflects the quality of the draft, but I'm not 100% sure. mw:ORES#Curation_support reads like it might be the inverse of the likelyhood of being CSD'ed for being spam/vandalism/attack. SQLQuery me!01:41, 14 April 2018 (UTC)[reply]
@SQL:, great stuff :), I'll think some more about the naming, maybe just 'highest rated', 'lowest rated' and 'potential spam'? Top 25 OK is very underwhelming :) John Cummings (talk) 13:33, 15 April 2018 (UTC)[reply]
Great chart but simply overwhelmed for transclusion on this explanatory supplement page arrived at by many many talks. Pls dont make our editors looking to derive serviceable information from the page to have to scroll though a huge list at the start of a parent type article. --Moxy (talk) 04:35, 21 April 2018 (UTC){{quotation }}[reply]
The irony is that Aaron is talking about this earlier in the thread with John Cummings but they don't appear to be aware of the other discussions on the same topics in in other venues where Marshall Miller is also leading some discussions. I'm never sure how the work at the WMF overlaps, or if indeed it does, or even if we are discussing these issues at too many different venues.. That said, the simple answer is that it should dovetail. This is one of the very reasons why the New Pages Feed should incorporate the flow of Drafts - because AfC doesn't have such a flow overview. So it's all a bit chicken-egg. AFAICS, many new page reviewers AFAICS don't appear to be bothered to use most of the excellent article meta information that is displayed in the feed, which is a great shame, so I'm not sure if ORES would help much. It might even lead them to making even more superficial checks on the new pages. Too much reliance on AI technology sometimes turns out to be a net negative to accurate productivity. That said, overall, NPPers do indeed do a somewhat better job than the AfCers - for now. With the correct training in these apps however, perhaps we would meet our goals, but even the tutorial at WP:NPP seems to be a challenge for some. On a more positive note, I believe there is finally a fully automatic replacement for the sorely missed Coren Search Bot under development which would help enormously.Kudpung กุดผึ้ง (talk) 05:42, 21 April 2018 (UTC)[reply]
@Moxy:, can you describe your concern here so that we can have a discussion about if it is possible to address this? Simply removing a useful tool people are using because you don't like it is disruptive editing, please leave the tool on the page until we have discussed. Thanks, John Cummings (talk) 15:33, 21 April 2018 (UTC)[reply]
Suggestion: we change the tables to collapsed so that they are easy to find and access but do not take up room for people who are not using them. John Cummings (talk) 15:37, 21 April 2018 (UTC)[reply]
Er, no, John Cummings, it's fairly standard practice - suggest you self-revert until there is consensus to include. Moxy is right: whether the tool is useful or not, it overwhelms the informational page, and it really isn't necessary to transclude it there at all for people to use it. Nikkimaria (talk) 15:43, 21 April 2018 (UTC)[reply]
I suggest something like this, it maintains all of the functionality of the uncollapsed version and addresses the concerns of taking up space on the page John Cummings (talk) 16:11, 21 April 2018 (UTC)[reply]
@Nikkimaria: Ah, sorry, forgot to change the section headings to <h4>, if the problem persists (I'm sure that's a way of avoiding it making section headings but it doesn't seem to have filtered through) I can just change it to big or something else. Other than the formatting issue I feel like this resolves the issue, what do you think? John Cummings (talk) 23:21, 21 April 2018 (UTC)[reply]
As mentioned above simply overwhelming for an information page. Guess I will have to make a new page WP:Finding drafts for this table and explain what the chart is about and more info about the topic. Will do this in the next few days. This does raise a concern.....do we have these walls of links on other info type pages. If so perhaps we should explain accessibility concerns about this on our MOS pages. It's disruptive to have a wall of links leading to editors not getting serviceable information..... we know statistically editors simply won't scroll through all that mess wall of text to get any information *Which_parts_of_an_article_do_readers_read....we structure our help pages in a certain format for a reason Moxy (talk) 01:40, 22 April 2018 (UTC)[reply]
Thanks for bringing my attention back to the conversation here, Kudpung. Aaron and I have been talking a lot about this in the last couple weeks, so he and I are in sync on this front. SQL and John Cummings -- what you have been putting together so far is really great, and the Community Tech team is thinking on our end what we can learn from it as we mull over the potential improvements to AfC summarized here, which sound similar to what you're doing. We'll be able to build some more detail around our thoughts this coming week as people come back from Wikimedia Conference. -- MMiller (WMF) (talk) 15:26, 21 April 2018 (UTC)[reply]
Thank you, Marshall for pointing me to this. Aaron, SQL, John Cummings, and Gryllida, these lists are really excellent. However, and as mentioned above by DGG, doing this post facto they represent a lot of work for anyone wishing to work through them manually. AFAICS, AfC still has no recognised work flow in the sense that one exists for WP:NPP as shown in the chart by Insertcleverphrasehere, which while complex and somewhat intimidating, has received positive comments by serveral users, or the more simplified one at File:New Page Review Flow Chart.jpg. IMO, the meta information provided by ORES would be very useful if it were highlighted in the Special:NewPagesFeed if submitted drafts were to be incorporated there with a user defined option in the preferences to select drafts only. More reason to consider migrating the functions of AfC into the NewPagesFeed and Curation UIs. Despite the resistance of Community Tech to prioritise the upgrades required for these UIs, IMO this would be a more rational deployment of WMF resources.Kudpung กุดผึ้ง (talk) 01:41, 22 April 2018 (UTC)[reply]
@Kudpung:, Marshall, Aaron, SQL, DGG, Moxy, and Gryllida are you happy for me to re-add the list in the form below until a longer term solution is potentially found through the suggestions above? This solution reduces the size of the list when people aren't using it to 2 lines and does not effect the table of contents. I hope this proof of concept helps to show the value of having something more developed and so make it more likely to have resources allocated to making it happen. John Cummings (talk) 07:23, 22 April 2018 (UTC)[reply]
I understand Moxy's concerns. My concern is more: Who will go to Wikipedia:Drafts and see and use the list even if it is collapsed? Perhaps it could be on a dedicated sub-page such as, just for example, Wikipedia:Drafts/ORES. That said, it would need to be linked to from AfC (and/or) the page patrolling tools. However, that would be the same parallel problem with CopyPatrol.Halfak (WMF) has said he would like to see ORES integrated into the Curation system, but I think he means into new arriving articles. I can certainly imaging that meta information being included in the entries in the Special:NewPagesFeed. So I ask again, what is the problem with giving AfC the opportunity to share that feed for their reviewing? It would seem to me to be a ready-made solution (bar the engineering). Kudpung กุดผึ้ง (talk) 08:53, 22 April 2018 (UTC)[reply]
Does ORES provide WikiProjects tags? If ORES was used to add WikiProject tags to any articles which do not already have them, that could be interesting. I am really interested in knowing what percentage of the main namespace is tagged with a WikiProject and what is not. --Gryllida (talk) 04:29, 24 April 2018 (UTC)[reply]
I don't know. If a bot were sent around to re-assesses all of the redirects as being redirects (and disambiguation pages as being dab pages), then you could probably calculate it. However, I think it's safe to assume that most articles are tagged, possibly 90% or even more. Clicking Special:Random ten times gave me ten tagged artices. Whatamidoing (WMF) (talk) 22:30, 25 April 2018 (UTC)[reply]
How do you transclude a lead only?
I've run into a bit of a mystery...
This works:
{{#section:Donald Trump|Lead text}}
This doesn't work:
{{#section:Physical geography|Lead text}}
<section begin=Lead text />
...
<section end=Lead text />
At Help:Magic words, search for "#section" to find a link to documentation. I hope transclusions like this are rare because it is very unlikely that the same text is really useful in two places, and it is likely that future edits will either break the transclusion or make it inappropriate. Johnuniq (talk) 23:43, 14 April 2018 (UTC)[reply]
But it picks up hatnotes too. Getting closer though. I'd like a method that doesn't require marking up the source page being transcluded, and if it is needed, makes it as minimally intrusive as possible. — The Transhumanist00:08, 15 April 2018 (UTC)[reply]
Galobtter, Can this technique be extended to transcluding the content of a short description template?, Something in the line of {{#invoke:String|match|pattern={{short description|= {{#lsth:Water}}}} · · · Peter (Southwood)(talk): 08:04, 17 April 2018 (UTC)[reply]
Hmm, @Pbsouthwood:{{#invoke:String|match|pattern=<div class="shortdescription.-</div>|s = {{#lsth:Water}}}}. Not running the code here, since it adds the short description to the village pump (that can be fixed, depending on your use case- what are you trying to do this for?) Galobtter (pingó mió) 08:47, 17 April 2018 (UTC)[reply]
Galobtter, there are uses as an annotation in some lists, such as indexes, where transcluding the short description after the link would be very informative. All that would be needed is the text from the unique short description in an article. Not needed for disambiguations etc where the same short description applies to several articles/pages. I have an idea that this would be helpful for creating automated indexes, possibly for outline lists too, maybe also as an option in category lists. Really anywhere that a short description after a link might be useful. Could also have applications in hovercards. This sort of application would make the short descriptions really useful inside Wikipedia instead of just being a pain in the butt forced on us to avoid being described by Wikidata.· · · Peter (Southwood)(talk): 09:35, 17 April 2018 (UTC)[reply]
Any auto-hyphenation yet
I wonder if yet any way to auto-hyphenate long words on browsers, as done 40 years ago by soft hyphens (versus "required hyphens"), to now wrap long words on narrow windows or smaller hand-held devices. See Danish sports page: "1927–28 Danmarksmesterskabsturneringen". I have used "<wbr>" to auto-wrap non-hyphen text, but want to auto-hyphenate to append "-" in each mid-word wrap. No hurry on this. -Wikid77 (talk) 00:49, 18 April 2018 More: As I recall, the soft-hyphen "­" worked fine on either computer browsers or mobile phones, but there were concerns for some screenreader devices to handle Wikipedia pages. However, I was thinking the newer screenreaders could handle (skip) the "­" characters when processing the hyphenated text. Alternatively, I wonder if soft hyphens were to be limited to one per long word, as a compromise to reduce problems in screenreaders, that limit might allow more use of soft hyphens in WP text. -Wikid77 (talk) 04:06, 18 April 2018 (UTC)[reply]
i think the primary concern is readability of the wikitext honestly. Anyway, hyphenation is primary a layouting problem, and with HTML that makes it a responsibility of the browser, not of wikipedia. It seems that browsers still dont recognize syllables in arbitrary languages and thus still dont support this as a layouting option. Then again, i feel that people are increasingly confusing the web with a typesetting platform sometimes. Its not, the web is designed to be dynamic, not to be an optimal typesetting language. Thats what latex is for. You cant even dictate what font or what size of rendersurface a person is using, so dictating hyphenation is even harder and shy’ing every syllable of every word seems an excessive solution to a minor problem to me. —TheDJ (talk • contribs) 06:24, 18 April 2018 (UTC)[reply]
The relevant W3C Working Draft was last updated on 22 August 2017 - just 8 months ago. It should not be considered stable, and is not a finished work - that won't occur until it is promoted to W3C Recommendation. Nutshell: don't rely on it. --Redrose64 🌹 (talk) 13:28, 18 April 2018 (UTC)[reply]
Just wanted to put a word in that I definitely noticed, as have some of my colleagues and friends. I understand that in the case of something similar to "Danmarksmesterskabsturneringen" that soft hyphens would be helpful, but hyphenating words of all lengths across line breaks really impacts the readability of an article. I'm not speaking with any knowledge about the different technical solutions to this, just purely feedback on how it affects some more casual Wikipedia users. Peloneoustc03:23, 25 April 2018 (UTC)[reply]
That sounds good to me. I'll keep an eye out on this page, so if the conversation or feedback about this moves to another location, I'd appreciate a link here to that. Peloneoustc03:45, 26 April 2018 (UTC)[reply]
When editing, I'm now seeing two mysterious extra checkmarks just below the proper check boxes. They first appeared a few days ago, but vanished within minutes. Now they're back: what are they for? MonoBook, Opera 36, Windows XP. --Redrose64 🌹 (talk) 20:58, 23 April 2018 (UTC)[reply]
I have difficulty trying other browsers. I can't upgrade from XP since the hardware won't be able to handle anything more sophisticated, so I'm confined to using XP-compatible browsers. Safari hasn't been updated since about 2012; IE8 is rejected by MediaWiki; Chrome hasn't been updated since April 2016. Firefox still gets the occasional update but is horribly slow - three minutes to follow a link and six minutes to go back again, which is prohibitive for watchlist checking. That leaves Opera. --Redrose64 🌹 (talk) 23:31, 23 April 2018 (UTC)[reply]
If you have hardware that old, have you tried Ubuntu or another Linux OS? Their requirements should be significantly less than Windows (which is in the realm of 4 GB RAM just for the OS right now...). --Izno (talk) 00:19, 24 April 2018 (UTC)[reply]
In that vein, I've had a lot of luck using Neverware Cloudready (Chromium OS - turns your laptop / desktop into a chromebook / chromebox), booted from a thumbdrive in older machines. SQLQuery me!00:29, 24 April 2018 (UTC)[reply]
@Redrose64: I looked into this, and it was a temporary issue when deploying a new version of MediaWiki – on last Thursday and last Monday, for about five minutes each, Wikipedia was using incompatible versions of HTML and CSS for the checkboxes, causing this weirdness on MonoBook (and causing the checkboxes to be very slightly misaligned on Vector and other skins). I wrote a longer description on the Phabricator task: phab:T192951#4155800. I think we understand why this happened and it shouldn't happen again. (Thank you for including the screenshot with your bug report, it was actually surprisingly helpful!) Matma Rextalk22:54, 24 April 2018 (UTC)[reply]
Also, while Opera 36 is probably the best browser you could be using if you still have to use Windows XP (and these weren't the cause of this problem), I really would recommend upgrading your hardware. You can get a decent used laptop (e.g. Thinkpad T500) for $100 or less, it should comfortably run Windows 7/10 (I have one, use it for testing the sites on old hardware sometimes). Matma Rextalk23:04, 24 April 2018 (UTC)[reply]
@Matma Rex: Please disregard the above – I figured out that the persistence of the problem was caused by the line $(".oo-ui-checkboxInputWidget").removeClass("oo-ui-checkboxInputWidget") in my common.js page. I've removed this line and everything is now appearing as it should. Also, my thanks to @TheDJ: for suggesting that .oo-ui-checkboxInputWidget .oo-ui-checkboxInputWidget-checkIcon.oo-ui-iconElement-icon could be the problem, which prompted me to try doing this! Double sharp (talk) 16:30, 28 April 2018 (UTC)[reply]
Checkboxes obscuring start of text
Seems to be a recent development: when I edit (standard wikitext mode) the two checkboxes in front of "this is a minor edit " and "watch this page" are now too big in all directions, which is most noticeable because they now obscure part of the "T" and part of the "W" of these two texts. Fram (talk) 13:36, 25 April 2018 (UTC)[reply]
Thanks. Safemode=1 puts it just next to the boxes (though still extremely close. When I enlarge my screen to 200% (I normally work at 100%), I can't see any space between the left side of the T or W and the box, which isn't visually attractive. But, like I said, it is no longer obscured in safemode, so some gadget must cause the issue (I have no css or other fancy stuff added, only enabled or disabled things in preferences, most recently the "show the wikidata description" one). Fram (talk) 13:53, 25 April 2018 (UTC)[reply]
We have been warned many times over the last year or so, but HTML Tidy seems to have been turned off. This means that pages with unclosed (or badly-closed) markup may now show large amounts of corruption. Examples:
It seems to have been inadvertently enabled in all wikimedia wikis and certainly caused https://phabricator.wikimedia.org/T192855, and probably other reports there. It will probably be reverted as soon as the developers notice it. It makes the case for simply leaving it as is considering that there hasn't been that much of "drama" related to the early deploy.
@Redrose64:, @Galobtter:, @Whatamidoing (WMF): Click on the edit link from the lint errors listing which will take you to the regular edit form with the wikitext section that is the source of the problem. You can then replace the ?action=edit part of the URL with ?action=parsermigration-edit and that will both give you the side-by-side preview and also continue to highlight the problematic wikitext section. For example, here is that link for the basketball season page. In this case, the problem is with '''{{Nowrap|[[Galatasaray Medical Park (men's basketball)|Galatasaray Medical Park]]''' {{flagicon|TUR}}}} where the quote tags are improperly nested across the nowrap-span. As you can tell, in some cases, there is no rendering difference since the fixups by Remex and Tidy are effectively the same, but this is hard to determine automatically, so we flag all potential cases. SSastry (WMF) (talk) 22:29, 24 April 2018 (UTC)[reply]
You can also get quick access to this side-by-side comparison through your prefs. Go to Special:Preferences#mw-prefsection-editing and choose "Enable parser migration tool". Then look for a new option in the sidebar. NB that today this is in the "General" group at the top, and that during the next deployment train, it will move to the bottom of the page, in a new section called "Developer tools". Whatamidoing (WMF) (talk) 22:23, 25 April 2018 (UTC)[reply]
Two thoughts:
Most of the noticeable problems were on talk pages. The editors working on this problem are mostly focused on the mainspace. It might be a good idea to double-check your talk page and/or any other talk pages you care about. Just add ?action=parsermigration-edit to the end of the regular (/wiki/ style) URL, scroll down, and see if anything looks broken. Also, I want to thank User:Anomalocaris in particular for individually contacting editors whose sigs were broken, starting months ago, because getting those fixed will reduce the number of problems on active discussion pages such as this one.
When this change happens (in about a month or two), any resulting formatting problems won't be visible until the page is re-parsed. This means that the problems will become visible after someone edits the page, when the page is purged, or when the cache expires. I expect, as a result of that delay, that people are going to see problems when someone makes an unrelated edit, and that this page is going to see questions along the lines of "I fixed this simple typo at the top, and now something's wrong at the end of the page, and reverting doesn't fix it". Whatamidoing (WMF) (talk) 19:03, 26 April 2018 (UTC)[reply]
mw-datatable no longer highlights
Is there a reason why cells are no longer highlighted when hovered over in tables using the mw-datatable attribute? Per Help:Table#mw-datatableWhen a cursor hovers over the table, that row over which the cursor is on will be highlighted. Sadly this no longer works and there is no highlighting. 78.18.91.209 (talk) 18:20, 24 April 2018 (UTC)[reply]
Maybe it's the OS I'm using. I'm currently viewing Wikipedia in desktop mode on an iPad running iOS 10. Before yesterday whenever I tapped a cell it would be highlighted, yet that no longer happens for some reason. 78.18.91.209 (talk) 10:37, 25 April 2018 (UTC)[reply]
@Xaosflux: thanks. I added my comment at phaprecator task. I wrote here just to make sure that there is no local mediawiki page need to be modified according to thia issue. Regards--مصعب (talk) 17:45, 26 April 2018 (UTC)[reply]
Neither of these replies have helped in any way. The way that I read the linked pages is that: either both of these MassMessage posts should have triggered the tag; or neither of them should have. How can it be that one triggers yet the other doesn't? --Redrose64 🌹 (talk) 18:16, 26 April 2018 (UTC)[reply]
The two diffs are posted to different user pages; one of the users could've triggered a few more conditions than the other, based on their username or edit count or whatever the filters on enwiki happen to check against. The condition limit takes into account short-circuit evaluation, so if a hypothetical filter has a condition that says (article_text rlike 'Phi.*') & <other stuff>, then the Whispyhistory edit will only accrue one condition (the rlike) and no more, because the rlike comparison failed and the filter can stop there, whereas the Philafrenzy edit will keep going, since it matched. Writ Keeper⚇♔18:29, 26 April 2018 (UTC)[reply]
IP edit requests
There's a recent spate of seven consecutive malformed edit requests at Talk:Grand Theft Auto: San Andreas from various IPs. All of them are empty and as such weren't implemented. My assumption is that anons don't know how to properly file an edit request, but could someone check what it's all about? Brandmeistertalk22:36, 26 April 2018 (UTC)[reply]
The IP's are from around the World so I guess nothing special is going on. If you click the "View source" tab or try to edit a protected page without permission then there is a "Submit an edit request" button. The form of the empty requests with dated headings show they all clicked the button and then saved without writing anything. This is quite common, it happened during nearly three months and the article had 117,383 views in the past 30 days alone, and the top has a box with an "improve this article" edit link which leads to the edit request button for IP's. Some editors just revert empty edit requests. PrimeHunter (talk) 23:26, 26 April 2018 (UTC)[reply]
Some blank requests are people trying to edit the article, but not understanding, others are trying to get around the protection (it is not uncommon for an IP who has left a blank SPER to delete the pp template, hoping that doing so will grant them access). With many people's attention spans being so short, the instructions (seen here) are too long, the IP just wants to make an edit, not read a lot of instructions. I tend to leave a short instruction for the first/latest, empty SPER, which, sometimes, leads to a proper request. - Arjayay (talk) 09:43, 27 April 2018 (UTC)[reply]
Howdy all, I use WP:EPH to assist with edit request which when clicking Remove Request states: "Are you sure you want to completely remove this edit request from the page? In general, this should not be done to good-faith requests, even if they are blank, unless they are duplicates of other requests by the same user, etc." in a popup. So that's the line I follow. I feel so long as pages have an archive it'd be no harm to give them the template x/y response and if really necessary remove them in bulk at a later date. Personally I think it is better to help that 1 in 5 that genuinely wanted to improve an article for the cost of a little more talk page spam. — IVORKDiscuss12:50, 27 April 2018 (UTC)[reply]
@Arjayay:it is not uncommon for an IP who has left a blank SPER to delete the pp template, er, how can they remove the template if they can't edit the article? Please provide diffs of where this has happened.
However, as explained above WP:EPH states: "Are you sure you want to completely remove this edit request from the page? In general, this should not be done to good-faith requests, even if they are blank" so the current advice is mutually contradictory. - Arjayay (talk) 20:56, 27 April 2018 (UTC)[reply]
IP edit requests have become common in the past few years often with POINTy or otherwise vandal-like or trolling aspects. IMO the best solution is to delete them rather than littering the talk page which encourages more "requests". -- GreenC20:44, 27 April 2018 (UTC)[reply]
Agreed. Please delete empty or suspected trolling edit requests. Look at it from the perspective of the other person: they see a button, they click it, they don't actually have a constructive proposal so some of them close the tab page and we see nothing, while others publish their thoughts. Such junk attracts other fluff and should not be encouraged. Johnuniq (talk) 00:08, 28 April 2018 (UTC)[reply]
Yeah that is what I do, just litters the talk page for no benefit, it is a test edit and thus is reasonable to revert. I do usually leave an explanation in the edit summary and revert the edit so they get a ping with the message. Galobtter (pingó mió) 11:21, 28 April 2018 (UTC)[reply]
Is it just my computer or is there a bug on Wikipedia. I've been noticing black squares appearing now & then on articles during both editing & viewing. GoodDay (talk) 00:54, 27 April 2018 (UTC)[reply]
GoodDay, next time you see one like that, please link the article here, along with where the black squares are appearing. By any chance, have you been reading any articles that deal with foreign people or places that use a non-European language? Mathglot (talk) 01:09, 27 April 2018 (UTC)[reply]
They appear randomly & disappear randomly. Sometimes flickering into different places on the same article, each time I press a key. GoodDay (talk) 01:41, 27 April 2018 (UTC)[reply]
I don't understand those terminologies. If anyone else has noticed these things, then I'll know it's Wikipedia related, not just my computer. GoodDay (talk) 02:15, 27 April 2018 (UTC)[reply]
This is also happening to me on bn.wiki (see right). Browser google chrome, latest version, windows 8. This appear randomly. --আফতাব (talk) 20:33, 27 April 2018 (UTC)[reply]
Could you open the article "Anybody Seen My Baby?" and try clicking on the "Bielorussian (taraškievica)" interwiki (the first one)? When I click on it, I am being sent to a page that doesn't have a question mark in its title (to a page that doesn't actually exist). But when I go to the corresponding Wikidata item and click on the same interwiki ("be_x_old") there, it works correctly. --Moscow Connection (talk) 13:57, 27 April 2018 (UTC)[reply]
Something seems off...we should link directly to the correct wiki location, not to the redirect. Also, the redirect should retain the full page title, not chop it off (I bet it's treating the ? as start of a query string). FACE WITH TEARS OF JOY[u+1F602]16:20, 27 April 2018 (UTC)[reply]
Hi. When I search in Google, find a Wikipedia article, instead of the English version, I'm taken to Google Translate for "Terjemahan". -SusanLesch (talk) 14:48, 27 April 2018 (UTC)[reply]
User:Moscow Connection, Thank you! Then the mysterious language may be Indonesian. My husband's machine on the same connection doesn't do this so TheDJ might be right, too. It happens for any search, like "Minneapolis." -SusanLesch (talk) 21:51, 27 April 2018 (UTC)[reply]
Is there any way for PAGESINCATEGORY to show how many pages that appear in both of two categories?
Hi
Is there any way for PAGESINCATEGORY to display pages that appear in both of two categories? I don't mean adding the sums of two categories together, I mean showing only the pages that appear in Category A and Category B.
John Cummings specifically asked for PAGESINCATEGORY so I suspect he wants wiki code which automatically updates it in a wiki page. This is not possible. PrimeHunter (talk) 21:46, 27 April 2018 (UTC)[reply]
I'm not sure if {{PAGESINCATEGORY}} will do what I think you are attempting and I couldn't find an existing template doing it. If I did understand you, however, I worked up a template that might do some of what your seem to be trying to do (hopefully). I've used Category:1711 births and Category:1711 deaths for a demonstration and they plug in to the template with the following syntax: {{Ct2|1711 births|1711 deaths|mode=pages}} and render as:
@John Cline:, sorry for the late reply, thanks so much for all the effort :) I didn't explain myself very well, I'm just looking for something that will simply give me a number rather that a list (see the In Progress section on this page. I'm sure your work will be useful for many people, I encourage you to turn it into a template :) John Cummings (talk) 09:06, 30 April 2018 (UTC)[reply]
Same here as well. I am using an up-to-date Chrome 66 and Windows 10, 64-bit. However the other Twinkle functions still works, it's just only the rollback links that are missing... theinstantmatrix (talk) 20:59, 27 April 2018 (UTC)[reply]
For me is it is missing in Wikitext view but is back (in a new form) in VisEd mode. I'm using Monobook skin. It seems to have started with the introduction of the new "browse history" widget. Was this introduction trailed somewhere? It passed me by if it was. The good news is the browse history header has buttons making it easy to switch to VisEd. SpinningSpark21:20, 27 April 2018 (UTC)[reply]
If I turn off "Visual differences" in preferences and purge the cache then the problem goes away (but the "browse history" feature stays on). SpinningSpark22:53, 27 April 2018 (UTC)[reply]
I turned off "Automatically enable all new beta features" and "Visual differences" and the issue is resolved for now. It appears that this issue is being looked into. Classicwiki (talk) If you reply here, please ping me.23:58, 27 April 2018 (UTC)[reply]
I'm sorry about this, I didn't anticipate that adding the "header" to visual diffs (which we did last week) would cause issues with other scripts. I've just proposed a patch to do this differently (in a way that should interact better with other tools like Twinkle), but it might take a day or two before we get it reviewed and deployed. In the meantime, if you want to keep using visual diffs and Twinkle, note that the "[restore this version]" link is actually present, but only when you switch to the "Visual" tab (using the buttons in top-right corner of the diff). Sorry! Matma Rextalk21:20, 30 April 2018 (UTC)[reply]
Some portals would like to show an excerpt from a page chosen from a category. Is there any way for a module or template to list the members of a category? title:getContent() on a category page just returns the "This is a list of foos" blurb, not the actual members. mw.site.stats.pagesInCategory counts the pages but doesn't reveal the titles. getContent doesn't work on Special:RandomInCategory, and transcluding it just produces a link rather than the text. Surely we don't need to write a bot to periodically copy the category page to a standard page (which we can then read easily)? Certes (talk) 01:32, 28 April 2018 (UTC)[reply]
@Certes: You can get a listing of category members via Special:WhatLinksHere/Category:Name, if it is restricted to namespace 0 (articles). This can be done via transclusion, e.g. {{Special:WhatLinksHere/Category:Australia|namespace=0|limit=200}}, and so is available in Lua via frame:preprocess(). That's as far as I got in Module:Sandbox/Evad37/randomInCat, but it should be possible to do some pattern-matching magic to extract a random title from the whatlinkshere list. - Evad37[talk]04:39, 28 April 2018 (UTC)[reply]
No, actually, after looking into it a little. Returns a strip marker instead of the list, and per the reference manual can longer access the list: mw.text.unstrip no longer reveals the HTML behind special page transclusion, <ref> tags, and so on as it did in earlier versions of Scribunto.Galobtter (pingó mió) 06:45, 28 April 2018 (UTC)[reply]
Galobtter beat my reply but I'll post it anyway. That's interesting because putting that into Special:ExpandTemplates shows a bunch of HTML. By contrast, trying {{Special:WhatLinksHere/Category:Australia|namespace=0|limit=200}} in ExpandTemplates shows merely a single line:
[[:Special:WhatLinksHere/Category:Australia]]
However, the following shows that a module gets a strip marker rather than anything useful:
{{#invoke:dump|dumphtml|1= {{#categorytree:Countries in Africa|hideroot|namespaces=Main}} }} →
You can put that in a template and it works just the same, I believe that some WikiProjects do this - possibly using the alternative syntax <categorytree hideroot=on namespaces=Main>Countries in Africa</categorytree>. A third valid syntax is {{#tag:categorytree|Countries in Africa|hideroot=on|namespaces=Main}}. --Redrose64 🌹 (talk) 11:27, 28 April 2018 (UTC)[reply]
The original question is about how to get a random article from a category (list the category members in a lua readable way), hence trying to get the output readable in lua so a random member can be selected. Galobtter (pingó mió) 11:31, 28 April 2018 (UTC)[reply]
Thanks for all the replies so far. #categorytree looks promising but I can't make it give me the titles within a module for further parsing. frame:preprocess returns the string "nil" (not Lua's nil value), and frame:expandTemplate gives an error (quite reasonably, since it's not a template). Attempting to preprocess a dummy page containing "<categorytree...", or to expandTemplate a template with that text, return the string "nil" too. Further suggestions are very welcome. Certes (talk) 12:24, 28 April 2018 (UTC)[reply]
This is another use-case for an interface between the MediaWiki API and Lua, even if only some API entry points. My only suggestion is a cron bot to post the data where Lua can pick it up. This 2014 thread and this 2016 comment by User:Mr. Stradivarius confirms Lua does not have access to category members. -- GreenC14:33, 28 April 2018 (UTC)[reply]
Can someone have a look at Category:1710s and click on the subcategories for each year. For me, on all the years, they expand to a message, saying: "Problem loading data. Please wait a moment and try again." The portion in italics is blue linked, but dead for me. Thanks.--John Cline (talk) 05:49, 28 April 2018 (UTC)[reply]
Everything works for me. All the categories are there, I can browse them without any problem whatsoever. --Moscow Connection (talk) 06:15, 28 April 2018 (UTC)[reply] Update. I didn't understand the topic starter's message correctly. Yes, I can browse the subcategories, but if (instead of opening them) I just click on the arrows to expand them, I, too, see "Problem loading data. Please wait a moment and try again." --Moscow Connection (talk) 06:51, 28 April 2018 (UTC)[reply]
I'm trying to connect to MySQL at Toolforge via HeidiSQL, but after filling login form I get error "Can't connect to MySQL server on 'localhost' (10061)". But I don't see anywhere any reference to localhost. Probably I simply did something silly. Some help needed :) --Edgars2007 (talk/contribs) 16:43, 29 April 2018 (UTC)[reply]
You set up the SSH tunnel incorrectly. You want your computer <-> toolforge bastion <-> wiki replica. In the SSH tunnel you attempted to connect to wiki replica directly, which isn't possible and the enwiki.analytics.db.svc.eqiad.wmflabs will not resolve. Specify a bastion host instead, like login.tools.wmflabs.org with port 22 for ssh. --Zhuyifei1999 (talk) 14:42, 1 May 2018 (UTC)[reply]
As for why it's connecting to localhost but failed, the ssh tunnel is supposed to port-forward the port 3306 of enwiki.analytics.db.svc.eqiad.wmflabs to a local port on your computer, which you specified as 3306 as well. i.e. it will open a port 3306 on your computer (localhost) and relay all traffic to and from wiki replicas through the ssh connection. Since the configuration for ssh tunnel is broken, the tunnel cannot be set up, and the port will not open. The program will therefore fail to connect to the port. --Zhuyifei1999 (talk) 15:00, 1 May 2018 (UTC)[reply]
I notice that Wikimedia software doesn't play midi files. What would be the likelihood of adding an extension which converts the file to ogg/mp3 on the backend, allowing modern browsers to play the file without a download? Magog the Ogre (t • c) 18:30, 29 April 2018 (UTC)[reply]
I understand that having to open sound files separately is also a problem for Wikivoyage, Wikibooks, and Wikiversity contributors. Just being able to play the audio file in a phrasebook or other language-learning content, without having to open a separate page, would be very handy. Maybe something to remember for next year's m:Community Wishlist? Whatamidoing (WMF) (talk) 17:55, 30 April 2018 (UTC)[reply]
MIDI files ain't audio files, but 'sequence' files (think of them as 'advanced' sheet music); there is no 'canonical' translation from a set of instructions for musical instruments to an audio waveform. Yes, certain softwares are able to do the synthesis and play the music (TiMidity++ is one such software in the FOSS world), but given the complexity, it's best to store both the 'instructions' (for what notes on what instruments are being played) and the waveform (for a canonical audible representation of the music).
Something very odd appears to be happening. The initial discussion was at WT:CS1. It looks to me as though MediaWiki has become somewhat schizophrenic in how it handles <bdi>...</bdi> in wikilinks. The cs1|2 templates use <bdi>...</bdi> to wrap titles for languages that use non-Latn scripts. <bdi>...</bdi> conveniently isolates unspecified text direction and allows the browser to determine how to correctly handle the wrapped text. This functionality in the cs1|2 templates is not new; it was introduced 11 October 2014.
As I began writing this, the page source for Help talk:Citation Style 1 had this:
<p><a rel="nofollow" class="external text" href="http://cpc.people.com.cn/BIG5/64162/64172/85037/85038/7104292.html">"Lǐ Dàzhāo de Zuòyòumíng" <bdi lang="zh">李大釗的座右銘</bdi></a></p>
I then clicked the edit link for the above-named section, clicked Show preview, right-clicked View page source, and got this from the source:
<p><a rel="nofollow" class="external text" href="http://cpc.people.com.cn/BIG5/64162/64172/85037/85038/7104292.html">"Lǐ Dàzhāo de Zuòyòumíng"</a></p>
<bdi lang="zh">李大釗的座右銘</bdi>
Then to be perverse, I clicked Publish changes (without changing anything), fetched a new page source where I found this:
<p><a rel="nofollow" class="external text" href="http://cpc.people.com.cn/BIG5/64162/64172/85037/85038/7104292.html">"Lǐ Dàzhāo de Zuòyòumíng"</a></p>
<bdi lang="zh">李大釗的座右銘</bdi>
The first version above is the desired version. I'm pretty sure that this isn't a browser bug (I'm using latest version Chrome) because the html is so obviously different in the three inspections which leads me to suspect something in how the Wiki source is rendered into html.
I checked the content of the <bdi>...</bdi> just to be sure that there is nothing odd encoded there. There isn't. All of those characters belong to Unicode CJK Unified Ideographs (U+4E00–U+9FEA).
The parser is parsing it just fine. Tidy is screwing around with it based on some other broken code elsewhere. Open the section with parsermigration-edit and you see that the right hand side is rendering correctly, and that is with Remex. When I parse this page locally with Parsoid with linting enabled, I see 3 linter complaints:
So, my speculation is: if you fix the missing-end-tag for the p-tag on the page around character offsets 36381 - 36601, my guess is you should probably see expected behavior with Tidy as well. Or, wait for Tidy to be switched (end of June). SSastry (WMF) (talk) 14:31, 30 April 2018 (UTC)[reply]
I confess that I don't understand the meaning of the linter output. However, I was able to glean sufficient information to lead me to a fix at Module:Citation/CS1/Configuration/sandbox for an extraneous </close> tag. That particular error has been in place since 14 October 2015. So, thanks for the clue.
I left this out: Even with the linter errors fixed higher up in the page, when editing the section, those errors are out of scope for an html rendering via Show preview. For the examples above, I was editing the section, not the whole page. This leads me to think that the linter errors are not the cause of the <bdi>...</bdi> problem because the linter did not find any errors in the section.
You are correct. The broken behavior is back - which is weird. Parsoid's output looks good. So, you can reduce it down to a single external link. And, if you keep hitting 'Show Preview' repeatedly, in some cases, the rendering looks good, and in other case, it doesn't. So, this almost feels non-deterministic. I tried this on metawiki (which has Tidy) and I see the same behavior. I tried this on mediawiki (which has Remex) and I don't see this behavior. This is mysterious indeed and I am not sure why Tidy would do different things for different runs. Will file a ticket just to record this (just in case this is not Tidy-specific), but, it is unlikely we will try to fix it if we know it is Tidy-specific since Tidy is going away in 2 months. SSastry (WMF) (talk) 16:49, 30 April 2018 (UTC)[reply]
Latest tech news from the Wikimedia technical community. Please tell other users about these changes. Not all changes will affect you. Translations are available.
Recent changes
The parameter for unpatrolled edits in recent changes filters changed name. You might need to update saved filters and links. [2]
Problems
We are migrating wikis from Tidy to Remex. Because of a bug the 250 wikis which do not yet use Remex were switched on 23 April. This is two months early. This meant that pages with broken wikitext showed wrongly to readers. The bug was undone the next day. You can help fix broken wikitext to avoid this problem when your wiki switches. Tidy will be removed on all wikis before July 2018. You can follow the process on Phabricator. [3]
When an administrator blocks someone they will have a calendar they can use to choose when the block ends. This is to make it easier to pick a specific date. [5]
You can soon turn on the Performance Inspector in the Editing section in your preferences. It shows information about the performance of pages. This could be the size of modules in the page, how many CSS selectors are defined on the page and how many are used, or the size of the images on the page. This tool is intended to help editors fix pages that load slowly. [6]
There is a new abuse filter function called equals_to_any. You can use it to check if its first argument is equal (===) to any of the following ones. For example you can use it to check if the page's namespace is amongst a given set of values in a more compact way than you could earlier. You can read more on mediawiki.org.
The new version of MediaWiki will be on test wikis and MediaWiki.org from 1 May. It will be on non-Wikipedia wikis and some Wikipedias from 2 May. It will be on all wikis from 3 May (calendar).
Meetings
You can join the next meeting with the Editing team. During the meeting, you can tell developers which bugs you think are the most important. The meeting will be on 1 May at 18:30 (UTC). See how to join.
You can join the technical advice meeting on IRC. During the meeting, volunteer developers can ask for advice. The meeting will be on 2 May at 15:00 (UTC). See how to join.
Future changes
The Wikimedia Cloud Services team is working on a new project called Toolhub. The goal is to make it easier for Wikimedians to discover software tools they can use. You can leave feedback on the talk page or email jharewikimedia.org to leave private feedback.
All wikis with fewer than 100 high-priority linter errors in all namespaces will switch to use the Remex parsing library. This is to replace Tidy. It will happen on 2 May. Other wikis will be recommended to switch soon when they have fixed the errors that must be fixed. Wikibooks wikis with fewer than 100 high-priority linter errors in the main namespace will switch on 9 May. Tidy will be removed on all wikis before July 2018. [7][8][9]
Location of Wikimania 2018—Cape Southern Sun Hotel
English Wikipedians can now use the mapframe function to embed maps right on a page. The map at right (which shows the location of this year’s Wikimania conference) demonstrates the functionality. Mapframe was requested by the English community via RfC last year and the request was reconfirmed this spring. If you you’re new to mapframe, this Kartographer help page shows how to use it.
Expect more significant new features in the coming weeks, with the implementation of the long-awaited map internationalization features. After internationalization, maps will display in the content language of the wiki they’re published on. Until then, they will continue to display in the language of the territory mapped. (You can experiment with internationalized maps on testwiki now—here’s a page of examples.)
The credits aren't watermarked into the image (see this image or this one). They're positioned on top of the bottom-right corner of the image by the browser, obscuring a corner of the image. If you download the image, it won't contain credits or anything like that. --Roan Kattouw (WMF) (talk) 04:48, 1 May 2018 (UTC)[reply]
Oh, I thought you meant the placement/floating issue of the map compared to the other elements. If you want an area highlighted you need to use a line or shape feature rather than coordinates. And |plain=yes will set frameless mode, removing the outer frame and caption. Or if you want more complicated things (Template:Maplink#Things_this_template_can't_do), you might have to use the <mapframe>...</mapframe> tags directly. - Evad37[talk]04:41, 1 May 2018 (UTC)[reply]
This is getting closer:
{{maplink|frame=yes|plain=yes|frame-width=250|frame-height=280|type=shape-inverse|zoom=7|frame-lat=29.185|frame-long=76.225|stroke-width=3|id=Q1174}} - Evad37[talk]04:54, 1 May 2018 (UTC)[reply]
Added a smaller stroke-width. Before adopting into infoboxes, the template should really support adding map data from multiple objects – e.g. an outline of an area plus one or more markers at various points. To do so in a sane way really require it to be rewritten in Lua - Evad37[talk]09:30, 1 May 2018 (UTC)[reply]
What is the easy way to localize a name in OSM? Is there any plans? Many places doesn't have Bengali localization, so last week i tried to have add some Bengali name. After so many steps i finally managed to add a translation & i took one minute to complete just for one place. আফতাব (talk) 22:21, 1 May 2018 (UTC)[reply]
Are you requesting a new gadget with this feature? You can write the namespace and a colon in the search bar, e.g. wp:searching where wp: is an alias for Wikipedia:. You can click "Advanced" in search results pages to select one or more namespaces with a click. The below in your common JavaScript will add a "Search" link to the top of pages, going directly to Advanced search with "Article" selected. PrimeHunter (talk) 11:01, 1 May 2018 (UTC)[reply]
Is anyone aware of the default implementation of Navigation popups for unregistered users? We've had a couple of complaints to OTRS in the last few days and I'm not aware of any discussion or announcement saying that the function would be enabled automatically. Can anyone point me to the discussion where this was agreed upon, so that I can direct OTRS enquiries to the correct information? Cheers, Yunshui雲水08:36, 1 May 2018 (UTC)[reply]
Do you guys have any plan to maintain, especially refactor, some of the modules? A keyword is "code smell", and then complexity… Some of the code is even worse than mine. Jeblad (talk) 21:06, 1 May 2018 (UTC)[reply]
The same could be said of many templates, noticeboards, and articles. The problem with refactoring a module (or a complex template) is that unless the time is spent to get fully immersed in defining the problem and solution, it is very easy to overlook gotchas, and testcases are usually absent or insufficient. At any rate, feel free to nominate a couple of modules that are used as needing refactoring. Johnuniq (talk) 00:58, 2 May 2018 (UTC)[reply]
Featured content icons
I noticed a recent change to the icons for featured content. In most places, including the top-right corners of articles, the icon in use is this: (this matches the first gallery icon below), but in some places it recently changed to this: (this matches the second gallery icon below). There is a clear difference between the two when placed side-by-side. This would not be a problem, except the change has not been implemented across the board (see the table in Template:icon). The commons links for the files themselves are here:
In my opinion, there should be consistency with which featured icons are used and to that end, I believe the former icon above should replace the latter icon above, as the former has a higher definition and looks better in icon form. Jackdude101talkcont04:20, 2 May 2018 (UTC)[reply]
I also think that the icons should be consistent, although this has nothing to do with the resolution of the images, since they're SVGs and can be resized virtually infinitely; I have no strong preference for either icon. Jc86035's alternate account (talk) 10:54, 2 May 2018 (UTC)[reply]
In that case, agreed - if there is no clear difference between the two, but the former is used more widely then we should standardise on it. Richard061218:59, 2 May 2018 (UTC)[reply]
Reference archiving bot
Are there any bots dedicated to archive URLs in reference templates such as "cite web", in a platform in the like of Web Archive? Has this been tried? Leefeni de Karik (talk) 03:21, 3 May 2018 (UTC)[reply]
As I sit here reading a notification that there has been a failed attempt to log in to my account, I find myself wondering how it is the site I spend the most time on is the only site I spent much time on that doesn't have multi-factor authorization. I'm quite sure this has come up before, but apparently I don't know a good search term for the VP archives. Now, I have a pretty strong password, but it's unsettling nonetheless, as it most certainly was not me who tried to log in. — Rhododendritestalk \\ 13:34, 3 May 2018 (UTC)[reply]
Thanks. As I'm not an admin, it doesn't look like there's anything I can do at this point. Reading over the meta page, however, I wonder what reason there is for this to be seemingly more involved than other systems. Why not just use the texted code like everyone else? I'm sure that would require a little investment from WMF, but this sure seems like a good use of its cash... — Rhododendritestalk \\ 13:50, 3 May 2018 (UTC)[reply]
Also, though it's not something typically offered with this sort of mechanism, it might also be useful to determine the IP that tried to log in. In this case, the notification came shortly after my first edit of the day, this revert. With an unregistered user, there's not a whole lot of point except as an additional reason to block, but it makes me think that registered users might do it, too, in which case that information would be important if confirmed via checkuser. Getting into a whole can of worms there, I know. — Rhododendritestalk \\ 13:56, 3 May 2018 (UTC)[reply]
@Rhododendrites: if you really really really want it, you can request tester access at meta SRGP - there is no good way to manage these and you could lose access to your account forever. Some recovery options are being discussed on some of the tasks I linked to in the phab workboard. As far as the logon notifications, much discussion is in place regarding improving these and including more information that may be useful (see phab:T174388). — xaosfluxTalk14:24, 3 May 2018 (UTC)[reply]
Rhododendrites, I just also got one of those (for the first time in 11 years). Does WP record the IP of these attempts? If so, is this information available for discussion at a venue like this? Why should anyone try to do this when it's pretty obvious I'm logged in and editing? Perhaps it's some kind of stupid bot?| Thanks. Martinevans123 (talk) 14:51, 3 May 2018 (UTC)[reply]
This type of notification has only been available for about a year. Right now it is informational only, bots try to hack accounts all the time. — xaosfluxTalk15:09, 3 May 2018 (UTC)[reply]
Do they try to hack my account all the time? Or have they tried all the time in the last year? Or are you saying that bots won't trigger this warning? Thanks. Martinevans123 (talk) 15:23, 3 May 2018 (UTC)[reply]
Note, the bots I refer to are not "wikipedia bots" - these are malicious parties on the internet; and they try to hack lots of accounts all the time. I got a failed logon notification myself today. For the most part, you don't need to do anything on these - if you find the notifications annoying, you can turn them off in preferences. — xaosfluxTalk15:34, 3 May 2018 (UTC)[reply]
Understood. From your last reply it seems like you are assuming the notifications that all three of us got were triggered my malicious bots. I'm quite happy that I don't need to do anything (although if I start to get one every day, I might do something anyway). I'm just wondering how often other folks get these notifications, whether IP's are recorded and if anyone does anything as a result. Sorry is this has careered off topic. Thanks. Martinevans123 (talk) 15:45, 3 May 2018 (UTC)[reply]
A lot of other people have received this notification, it doesn't seem to stop, even on frwiki we receive a notification from enwiki. See [12]. Lofhi (talk) 16:09, 3 May 2018 (UTC)[reply]
Similarly to TenOfAllTrades, I've just got my first such notification ever where it hasn't been me on a different device. Is this sort of rate common or is something going on? Red Fiona (talk) 17:37, 3 May 2018 (UTC)[reply]
Who did what to the source editor?
The character icons are all changed. I don't think the Template Wizard people are responsible because I have been following the implementation of that tool, and this change is more recent. L3X1◊distænt write◊14:13, 3 May 2018 (UTC)[reply]
I've edited this page a lot, which can be seen at [13]. But it seems to me that something in the 2 "top ten" piecharts is somewhat off, my % doesn't match my piece of pie. Any good reason for this? Gråbergs Gråa Sång (talk) 17:35, 3 May 2018 (UTC)[reply]
I believe the percentages are percent out of total — so for example, you made 51 edits out of a total of 103, or 49.51& — but the pie charts are out of the top 10 — so 51 out of 90, or 56.67%. If you mouseover the pie chart slices, they display numbers (51, 9, etc.) not percentages. It's a little weird, perhaps, but it certainly makes the charts nicer. ~ Amory(u • t • c)18:02, 3 May 2018 (UTC)[reply]
Done SQLQuery me! 16:22, 27 March 2018 (UTC)
Thank you I found and fixed them. --Redrose64 🌹 (talk) 23:23, 24 April 2018 (UTC)
We are migrating wikis from Tidy to Remex. Because of a bug the 250 wikis which do not yet use Remex were switched on 23 April. This is two months early. This meant that pages with broken wikitext showed wrongly to readers. The bug was undone the next day. You can help fix broken wikitext to avoid this problem when your wiki switches. Tidy will be removed on all wikis before July 2018. You can follow the process on Phabricator. [3]
The new version of MediaWiki will be on test wikis and MediaWiki.org from 1 May. It will be on non-Wikipedia wikis and some Wikipedias from 2 May. It will be on all wikis from 3 May (calendar).
wikimedia.org to leave private feedback.
Fixed [10], the problem was with {{climate chart}} not floating properly and the mapframe being a bit too wide. - Evad37 [talk] 04:07, 1 May 2018 (UTC)
(this matches the first gallery icon below), but in some places it recently changed to this: 
(this matches the second gallery icon below). There is a clear difference between the two when placed side-by-side. This would not be a problem, except the change has not been implemented across the board (see the table in Template:icon). The commons links for the files themselves are here:
