This page contains discussions that have been archived from
Village pump (technical). Please do not edit the contents of this page. If you wish to revive any of these discussions, either
start a new thread or use the talk page associated with that topic.
This has been bugging me for a while. I have noticed that, in a number of wikitables, some cells do not accurately align with adjacent cells. In other words, the bottom line of a particular cell may appear 1px below the bottom line of the cell adjacent to it. Is this a browser issue or a Wikipedia one? FWIW I'm using Internet Explorer 11.--
Nevé–selbert02:14, 12 October 2017 (UTC)
It sounds likely to just be a browser issue in certain circumstances but you didn't post the page name and also failed to do it here. I have never seen the issue in Firefox. You should follow Aklapper's advice and make reproducible reports. The page name you are posting about is an absolute minimum. I found it at
List of Prime Ministers of the United Kingdom#Since 1721 after some searching but I really shouldn't have to do that. I cannot reproduce it in Internet Explorer 11.608.15063.0 on Windows 10. Details like zoom level, font size and window size can affect browser rendering. I did notice a similar issue with vertical borders around merged cells sometimes differing by a pixel in IE 11. I guess the browser uses a system with "local" cell border calculations without ensuring the table gets globally consistent borders.
PrimeHunter (
talk)
18:29, 12 October 2017 (UTC)
Many things might affect a small rendering detail like this. Skin could be one of them but I don't see the problem in MonoBook and IE11 with or without Compatibility View.
PrimeHunter (
talk)
22:45, 12 October 2017 (UTC)
Checked again, the lines no longer align in Compatibility View. I have tried tweaking the settings in
Internet options but to no avail. @
PrimeHunter: Which operating system are you using to run Internet Explorer 11? This might be a Windows 8 problem.--
Nevé–selbert02:07, 13 October 2017 (UTC)
I think it depends on the browser and not the Windows version. As mentioned, I see a similar issue with vertical borders in IE11 on Windows 10.
PrimeHunter (
talk)
17:27, 13 October 2017 (UTC)
PrimeHunter is right, this is a browser issue, more specifically an Internet Explorer issue. I was able to get my hands on Windows 10 yesterday and the lines aligned perfectly in Edge, but not in Internet Explorer 11.--
Nevé–selbert00:47, 15 October 2017 (UTC)
Template:Hiero
To save vertical space, is it possible to put two Egyptian hieroglyphs in
Template:Hiero side by side, similar to the default placing in
Template:Multiple image? I know it's possible via <hiero> </hiero>, but in that case it would give an impression of a single word or phrase, which is sometimes undesirable.
Brandmeistertalk12:36, 14 October 2017 (UTC)
The logs appear to indicate that pending changes protection was applied, meaning that registered editors can see live edits even before they have been moderated/accepted, but they should see them in a separate section of the history until they are reverted/accepted (
WP:PEND for more details). —
PaleoNeonate –
15:36, 15 October 2017 (UTC)
It was also semi-protected at the same time, as seen in the logs (the times are in
Eastern Time):
00:07, 15 October 2017 Samsara (talk | contribs) configured pending changes settings for Motu Patlu [Auto-accept: require "autoconfirmed" permission] (Persistent disruptive editing: via RfPP) (hist)
00:07, 15 October 2017 Samsara (talk | contribs) protected Motu Patlu [Edit=Require autoconfirmed or confirmed access] (expires 23:07, 14 November 2017) [Move=Require autoconfirmed or confirmed access] (expires 23:07, 14 November 2017) (Persistent disruptive editing: via RfPP) (hist)
But I can not edit it as IP-user. So you are right, it seems very stange that the IP could edit the article. Nothing has changed since the IPs edit.
Christian75 (
talk)
18:13, 15 October 2017 (UTC)
I was under the impression that IPBE allowed you to edit through autoblocks as well as through hardblocked IPs, and also I was under the impression that admins had IPBE by default. Which of those statements is wrong? Trying to find
MediaWiki:Autoblockedtext, I decided to get it directly, so I blocked my alt account, logged out, logged into it, tried to edit (and got the you-have-been-blocked message), logged out, got the autoblocked message, logged into my main account — and I was autoblocked! The message said Editing from Nyttend backup has been blocked..., as if I were still logged into the alt. I thus posted a request with {{unblock-auto}}, and after a few minutes
User:Floquenbeam came along and resolved the situation. The biggest annoyance here, moreover, is that as I was autoblocked from editing any pages other than my talk, I couldn't unblock my alt account, so I ended up being unable to undo my own block; when you post an unblock template, it looks rather silly to have the link to the blocking admin's page be black because the unblock is posted on the blocking admin's talk page :-)
So where did I go wrong? As an admin, shouldn't I be able to edit through an autoblock? I've been able to do that in the past; see
[1]. Several years ago, I tried to use my alt account to edit from a public computer, but I found that it was autoblocked, so I used the main account to request IPBE for the alt (I didn't get autoblocked), and once someone gave me IPBE for the alt, I went back to using it and didn't have further problems. That was almost eight years ago; did we have some sort of software change that takes IPBE away from the basic admin rights package?
Nyttend (
talk)
22:43, 13 October 2017 (UTC)
(
edit conflict) According to
Special:ListGroupRights, admins are exempted from IP blocks (search for ipblock-exempt, given to admins, bots and users with the IP block exemption user right). Also, the page states that this user right exempts the user from autoblocks.
The user right torunblocked, which allows you to edit using
Tor (anonymity network), is apparently only given to those who are specifically given the IP block exemption user right: admins and bots can't use Tor by default. --
Stefan2 (
talk)
22:53, 13 October 2017 (UTC)
Hm, cookie block; I figured those simply were another means of assigning autoblock, i.e. it's a normal autoblock that detects you by means of a cookie, not by means of your IP address. I wasn't using Tor; it was just the public wireless network at
Wendy's #2834, which uses
Comcast IP
75.75.127.168.
Nyttend (
talk)
23:07, 13 October 2017 (UTC)
It could be that it was a cookie block, especially if you were using the same browser. See also
phab:T5233. This may need some more experimenting. Generally it will be the same "person" that would be impacted by a cookie block, so being autoblock exempt (which is designed to prevent collateral damage for DHCP and rangeblocks) doesn't quite apply. —
xaosfluxTalk00:30, 14 October 2017 (UTC)
Yes, it was the same browser. So I could have just closed the browser and logged in with another one? That didn't occur to me. Also, what's DHCP?
WP:DHCP is red. And finally, is it intentional that you can't do things like blocks and protections when you're blocked? It seems rather pointless, since as a security measure it's not useful (a rogue admin can always unblock himself and continue to make mayhem until he's desysopped), and any kind of tool usage when you're blocked will get you in big trouble anyway, unless, like here, you're testing; it seems silly that removing a block that I myself had set would be impossible as long as I didn't get desysopped.
Nyttend (
talk)
00:36, 14 October 2017 (UTC)
PS, of course I know that this is a rare (and silly) situation, so I understand if fixing the problems I've identified here isn't considered to be worth the effort.
Nyttend (
talk)
00:44, 14 October 2017 (UTC)
Well, of course, the important thing is that important things such as deleting and protecting don't have to be bundled in with block (and more importantly unblockself). Therefore, the easiest way to deal with abusive accounts/problems in general is to simply block the user and not worry about handling rights in the short term. Of course, here on WP that doesn't really apply. --
The VoidwalkerWhispers00:55, 14 October 2017 (UTC)
If we got rid of unblockself, that would be fine with me; I just don't understand why that's an option and other stuff isn't (either all of them should be options, or none of them should be), since the rogue admin can't really be stopped without bureaucrat or steward assistance, but the confused self-autoblocking admin has to get help.
Nyttend (
talk)
01:06, 14 October 2017 (UTC)
The only thing admins can do when blocked is unblockself and editowntalk (if not also blocked). If you would have used a different computer or certain other browsers you would have a new set of cookies and would not be impacted by the cookie block anymore. —
xaosfluxTalk01:17, 14 October 2017 (UTC)
I have used Recent Changes for the past year, but the Recent Changes won't load properly when I put in my custom changes to fight vandalism, it just sit there and load for minutes. Is it a bug in the coding?
Gary "Roach" Sanderson (
talk)
16:14, 15 October 2017 (UTC)
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
When you edit a template with the visual editor it follows the template's format configuration on how to save it. You can now change the format in
more powerful ways if your wiki wants it.
[2][3]
The
TwoColConflict extension is a new way to solve edit conflicts. It makes it easier to copy and paste the relevant text to the text field. It has been a beta feature for the last five months. You can now
try out the new interface on a page where you will actually get simulated edit conflicts. You can
give feedback to the developers.
Problems
There is a problem with recent changes pages and your watchlist that show a large number of pages. Until it is fixed, Wikidata edits will not be shown on recent changes or in your watchlist on Commons or Russian Wikipedia. If necessary, Wikidata edits will be removed from all wikis.
[4]
Notifications are not working for some actions. For example some users don't get a notification when they are mentioned. The developers are working on fixing this.
[5]
Users on some wikis could not change their preferences. This has now been fixed.
[6]
Changes later this week
Editors and readers who still use Internet Explorer 8 on Windows XP will not be able to use Wikipedia from 17 October. Internet Explorer 8 on Windows XP can't connect securely to the wikis. When we allow them to do so it means that we get less security for everyone else. If you use Internet Explorer 8 on Windows XP you can install
Firefox 52 ESR instead. Around 0.1% of the traffic to the Wikimedia wikis comes from Internet Explorer 8 on Windows XP.
[7]
The
new version of MediaWiki will be on test wikis and MediaWiki.org from 17 October. It will be on non-Wikipedia wikis and some Wikipedias from 18 October. It will be on all wikis from 19 October (
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
17 October at 19:00 (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
18 October at 15:00 (UTC). See
how to join.
Hello, does anyone else experience this problem: Just an hour ago I logged in, went to recent changes, and the whole menu have disappeared (i.e. I can't choose how many pages to display (from 50 to 500)). Any help will be appreciated. Thanks.--
Biografer (
talk)
20:22, 12 October 2017 (UTC)
The English Wikipedia always gets new MediaWiki versions on Thursdays, and got one this week. I suspect Redrose64 is just guessing that it's related. I get a menu at
Special:RecentChanges but a couple of seconds after the rest of the page has loaded.
PrimeHunter (
talk)
23:06, 12 October 2017 (UTC)
@
PrimeHunter: and @
The Voidwalker: Your site already says that it's October 13. So no, apparently it have nothing to do with October 12 update. Will see if hiding it will do anything for me, if not, seek your suggestion and support. Thanks. PS: What would we do if it will not be fixed by October 19, shouldn't we just revert changes?--
Biografer (
talk)
02:20, 13 October 2017 (UTC)
Today being 13 October will have nothing to do with it. MediaWiki updates are typically rolled out once a week, on Thursdays, in the middle of the day (California time). If it is a MediaWiki problem, don't expect any change until Thursday 19 October.
@
PrimeHunter: I also see that Recent Changes is different, instead of nice quick "50 100 250 500" links that were present right from the start, I now have to wait for the JavaScript to stop churning, click a button (that is not present until the JavaScript finishes), then click another button. --
Redrose64 🌹 (
talk)
07:42, 13 October 2017 (UTC)
@
Trizek (WMF): You will be surprised, but, no. All of them are white coloured and where its light green its bright yellow (old version). The problem is, is that the whole filter menu is gone, and three dots (indicating loading screen) appears. With that thing absent, I can only view 50 pages at a time, and when I am done, I need to hit recent changes link to refresh it. I hope, now you can get a clearer picture of what's going on. And yes, I tried hiding List of abbreviations box, tried to log out/log in back, nothing. :(--
Biografer (
talk)
16:03, 13 October 2017 (UTC)
@
Biografer:, I didn't get your notification (they are broken). The interface you have (had, before you opt-out) is apparently the right one. At least, there is no interference from a script or a gadget. On recent changes with the new interface, you can still change the number of pages you want to view, like any function you had before. There is new functions as well, like a direct link to update the list of results without reloading the page each time. Please have a look at
the documentation to know more.
Trizek (WMF) (
talk)
16:32, 16 October 2017 (UTC)
Looking for small tasks+mentors for new contributors – got something in mind?
Hi everybody!
Google Code-in (GCI) will soon take place again – a seven week long contest for 13-17 year old students to contribute to free software projects. Tasks should take an experienced contributed about two-three hours and can be of the categories Code, Documentation/Training, Outreach/Research, Quality Assurance, and User Interface/Design. Do you have an idea for a task and could you imagine mentoring that task? For example, do you have something on mind that needs documentation, research, some gadget or template issues on your "To do" list but you never had the time, and can imagine enjoying mentoring such a task to help a new contributor? If yes, please check out
mw:Google Code-in/2017 and become a mentor! Thanks in advance! --
AKlapper (WMF) (
talk)
19:51, 16 October 2017 (UTC)
"14,882 users are trying this feature."
In "preferences – Beta features", each feature gets a "X users are trying this feature." statement. Are this users on enwiki or global?
Fram (
talk)
14:28, 6 October 2017 (UTC)
That's what I thought someone said awhile ago as well. I then already thought that the numbers given were suspiciously high. For the "Content translation tool", it says "58,320 users are trying this feature.", even though the tool is "restricted to editors that are extended-confirmed.", of which there are at the moment only 36774...
Fram (
talk)
14:54, 6 October 2017 (UTC)
It wasn't always limited to extended-confirmed and all users still see the option in preferences and can enable it there even if it doesn't currently work in their account.
PrimeHunter (
talk)
15:05, 6 October 2017 (UTC)
In addition to this, the number is only about those who opted-in deliberately: doesn't count those who chose "Automatically enable...".
Elitre (WMF) (
talk)
15:07, 6 October 2017 (UTC)
While I can buy Pimehunter's explanation, it seems extremely unlikely that more than 50,000 editors on enwiki, who don't have 'enable all beta features', have actually specifically enabled this feature.
Fram (
talk)
15:44, 6 October 2017 (UTC)
It's not. First, because it's been a tool in active development for a few years now, that's globally much appreciated and will soon celebrate the creation of article #250k. Also, the English Wikipedia is quite often the source of translations, so one needs to have Content Translation enabled here to start a translation into their own language.
Elitre (WMF) (
talk)
16:07, 6 October 2017 (UTC)
To clarify: That number reports all accounts that are currently opted in at this (local) wiki, according to the prefs table. So if someone creates an account, plays around with the prefs, and forgets the password, then that account will always be counted as "trying this feature". But if you actually do try the feature, but without ticking the individual button, then your account will not be counted as "trying this feature". It is, at best, an estimate. Given that >99% of accounts are not actively editing, the difference between "actually using" and "ticked the button at some point in the past" can be quite substantial.
Whatamidoing (WMF) (
talk)
22:32, 16 October 2017 (UTC)
Several people tried to ping me today, and nothing went through
[8][9][10]. My settings don't have anyone muted, and the templates/signatures seem to have been placed correctly. Also checked to make sure that my preferences have notifications on for mentions. Anyone have any ideas why these aren't going through?
TonyBallioni (
talk)
01:07, 11 October 2017 (UTC)
I've also noticed several pings did not reach me too. Is this another system change? Incidentally, are all of the failed mentions from {{u}}?
Alex ShihTalk01:24, 11 October 2017 (UTC)
@
Adam9007: It's not going to be the {{
u}} template. Templates do not, by themselves, trigger notifications: the required feature is a link to the user page. If a template – whether that be {{
u}}, {{
user}}, {{
replyto}} or similar – happens to generate a user page link, a notification will be sent exactly as if a bare link had been used, as I did in this post. --
Redrose64 🌹 (
talk)
16:56, 11 October 2017 (UTC)
I’m not receiving any pings from anyone either. For example,
Huon pinged me multiple times on article talk pages where they started discussions about an issue with which I am involved. However, I didn’t get any of those notifications, and I was only aware of the discussions because I was watching those talk pages. Additionally, I didn’t get any notifications when my edits were reverted on some pages, even though my settings are supposed to send me such notices. This is appearing to be a technical bug of some sort, because pings and other notifications were working just fine a few days ago. Also, when I get a talk page message, the “You have new messages” banner appears, in addition to a notification in
Special:Notifications. This is weird. —
EcstaticElectrical,
01:30, 11 October 2017 (UTC)
I also have this impression, because I did not change any related setting and the edit history did show that the signature was present in the two edits I mentioned (a requirement for successful notification). —
PaleoNeonate –
01:34, 11 October 2017 (UTC)
Yes, I used {{ping}} and you used {{re}}. I didn't get either. When the software told the sending account that a notification was sent, I doubt it's related to which template was used.
PrimeHunter (
talk)
01:54, 11 October 2017 (UTC)
It sounds like you are referring to the mention feature in general when it's working correctly. We were discussing why some notifications failed to go through when they satisfied all the requirements including a link to the user page. When something is partially broken, it's fair to speculate about which circumstances cause it to fail. I received several mention notifications today so it works in some cases. My guess is that the failures are independent of whether a template or which template was used.
PrimeHunter (
talk)
17:30, 11 October 2017 (UTC)
I have figured out what was causing this problem. 1367 users are affected on English Wikipedia, and they are not receiving any notifications (mentions or otherwise) from logged-in users (but notifications coming from anons should work). I have a fix lined up, but we probably won't be able to roll it out until Monday (most people don't work over the weekend, I usually don't either). In the meantime, you can work around the issue by making a change to your preferences. It can be any change, and you can even change it back, as long as you save in between. For example, change your language to Canadian English, save, change it back, save. Once you do that, you should start receiving notifications again. For a more detailed (technical) explanation and for a play-by-play of how I figured this out over the course of an hour, see
my comments on the Phabricator task. --
Roan Kattouw (WMF) (
talk)
00:06, 15 October 2017 (UTC)
@
Roan Kattouw (WMF): I have now made a pref change/un-change. Will I need to do this on all other wikis too?
Also, is there any possibility that these incorrectly-muted notifications could be re-sent? Since Monday 9 October 2017, the only notifications that I have received have been of just two kinds: (i) 'Your mention of Example was sent' and (ii) 'Example left a message on your talk page in "Thread"'. It's easy to find out who has thanked me in that period; it's all recorded at
Special:Log/thanks – but the notifications which I have really missed are those concerning: mentions elsewhere than my talk page; edit reverts; and notifications (of all types) from another wiki. I am certain that I have been reverted in the past six days, but I only know about those on my watchlist; similarly, I don't know who has mentioned me elsewhere than discussion pages on my watchlist. --
Redrose64 🌹 (
talk)
07:45, 15 October 2017 (UTC)
So I received my first ping today (since about a week ago) after changing a setting, saving, then changing it back, saving again. Lost notifications are not available in the log, however. I agree that it would be a very nice thing if when the bug is fixed the log could also be restored (assuming that it's still available somewhere in non-public logs)... —
PaleoNeonate –
16:21, 15 October 2017 (UTC)
@
Redrose64:@
PaleoNeonate: Yes, you would have to do that on every wiki. However, that issue should be moot quite soon, as my fix for this should roll out within the next half hour or so (I'll update here when it does). Sadly, muted notifications are not stored (since muting happens at send time) and I don't think they're going to be recoverable. --
Roan Kattouw (WMF) (
talk)
18:27, 16 October 2017 (UTC)
It took a bit longer than half an hour because we mistakenly believed that my patch broke other things, but as of about 10 minutes ago, my fix is live. --
Roan Kattouw (WMF) (
talk)
23:37, 16 October 2017 (UTC)
Hello again. I failed to find a magic word which expands to the current user's name in the documentation, like USERNAME (ideally not the signature). I thought of an unsubst trick with tildes, but tildes output the signature anyway, not the actual username (assuming that the signature includes a link to the user page, possibly that a fancy URL extraction trick would also be possible). Does anyone know of a simple trick to have a template automatically output the editor/substitutor's username as part of its results? Thanks, —
PaleoNeonate –
09:51, 17 October 2017 (UTC)
I don't think you can do this unless the template is actually on the users userpage, when you can use variations of PAGENAME. The full list of parser thingies is
Help:Magic words.
Dysklyver10:11, 17 October 2017 (UTC)
This is not possible, as it would mean that the page would be different for each user viewing it, which makes it uncachable. Wikimedia is very dependent on caching, therefore such functionality is not provided. —
TheDJ (
talk •
contribs)
11:11, 17 October 2017 (UTC)
If you want to permanently save the the name of a user making an edit then {{subst:REVISIONUSER}} will do it. In a template to be substituted you can for example write {{safesubst:<noinclude />REVISIONUSER}} to prevent your own username from being saved when you edit the template. {{REVISIONUSER}} without subst: will display the last editor of the page, or a user previewing the page. It is not possible to display the name of the user reading a page. We could make global JavaScript to modify certain code on a page after loading but it would only work for users with JavaScript and there are other issues. I don't support it.
PrimeHunter (
talk)
11:15, 17 October 2017 (UTC)
Yes I'd of course want it to be permanent (substituted at edit time). This may be what I was looking for, thank you very much, —
PaleoNeonate –
12:39, 17 October 2017 (UTC)
You must see it surely- it's doing it right now on this page, where the edit screen isn't a window it's a page. I'm crap about the lingo. Yeah soz about the ****, was just going slightly potty back there :) It loads ike V?E but displays mark-up rather than WYSIWYG. Here's a cap ------>
LOL thanks
Nihlus- I take it the nicely ready-ticked box "Automatically enable all new beta features" is the culprit!?! -or should I say the ******* culprit?! ;) —
fortunavelut luna14:43, 6 October 2017 (UTC)
Thanks
Elitre (WMF), in general I don't mind experimenting (got me into a lot of trouble when I was young!), but al I was trying to do was paste '{{subst:GAN|subtopic=Royalty, nobility, and heraldry}}' onto a talk, but instead got <code><nowiki>{{
subst:
GAN|subtopic=Royalty, nobility, and heraldry}}</nowiki>- see, loads of nowiki /code etc wrapped around it. It worked on the fifth attempt :p —
fortunavelut luna15:02, 6 October 2017 (UTC)
Just for general information: The WMF decided that they could reduce the number of editors by building a "wikitext mode" inside of Visual Editor, and getting us to eventually switch into it. It has performance problems and faulty previews because it's build inside of VE. Performance is severely affected by the side of the article, and seems to vary for different people. The last time I tried previewing
United States it took over 60 seconds. There was a Village Pump consensus to submit the performance and preview issues as deployment-blockers. The WMF has given no indication that performance can be improved to be comparable to the current editor, and they appear unwilling to provide genuine previews. Instead they want to eventually switch over article views to use the VE-rendering engine. That would theoretically fix most or all of the "preview flaws" of trying to use VE-itself as a preview for wikitext editing.
Alsee (
talk)
20:55, 6 October 2017 (UTC)
Wow. That's nuts. I really really really hope they don't go ahead with it — the normal editor's worked well for fifteen years, and if it's not broke, don't fix it.
Nyttend (
talk)
23:29, 13 October 2017 (UTC)
Ditto. A change like this probably would "reduce the number of editors", but probably not in the way that the WMF intended. –
Jonesey95 (
talk)
00:29, 14 October 2017 (UTC)
If you're interested in editing tools, then please see
mw:Editor for a long list of editing environments. Almost all of the tools currently on the list are supported by the Foundation. The Foundation plans to continue support indefinitely for almost all of these editing environments, including the 2003 (non-Javascript-requiring) wikitext editor and the 2010 wikitext editor (
Nyttend, that's probably the one that you describe as "the normal editor", but you can check the screenshots at
mw:Editor to be sure). There is a plan to drop Foundation support for exactly one (1) of those tools (and a volunteer is looking at converting it to a user script, in which case, it won't even go away – it'll just be someone else's problem when it breaks).
I happen to know about the plans for editing support, but in general, if you want to know the actual plans for any particular product, then please ask me. If I don't know, I'll try to find out for you. You can also ask any dev working on the product that interests you. Having a bunch of rumors floating around is not going to help anything.
Whatamidoing (WMF) (
talk)
22:24, 16 October 2017 (UTC)
MW 1.31 started recently. The removal has been delayed because of two bugs (so far; one was recently resolved, and the other is still open).
Nyttend, as I said above "There is a plan to drop Foundation support for exactly one (1) of those tools". The one editing environment that the Foundation will no longer support is the 2006 wikitext editor (=not 2003 or 2010, both of which will be supported for a long time to come). See
phab:T30856 if you want to follow that work.
Whatamidoing (WMF) (
talk)
15:20, 17 October 2017 (UTC)
Why does the right column of this table not align center?
Align left
Align center
good
bad
good
bad
good
bad
good
bad
Obviously I'm misunderstanding the "scope" parameter or something. But is there a way to do this other than setting the style of every cell in the column?
Help:Table doesn't give an answer. Brightgalrs (/braɪtˈɡæl.ərˌɛs/)[1]23:44, 16 October 2017 (UTC)
Yes, this is a limitation in html. Each cell in a column must be styled. It would technically be possible for MediaWiki to make a way to add the same styling to every cell in the generated html but that hasn't been done and maybe shouldn't be done due to complications like possible explosion of html size. Some tables use a
row template which adds styling.
PrimeHunter (
talk)
11:25, 17 October 2017 (UTC)
This is a limitation of wikitext not of html (see
https://phabricator.wikimedia.org/T2986). It can be overcome using sitewide CSS and nth type pseudo classes (although that will increase page size for everyone) or by something like
Help:TemplateStyles. 11:58, 17 October 2017 (UTC) — Preceding
unsigned comment added by
197.218.82.93 (
talk)
Týr (/ˈtɪər/;[1] Old Norse: Týr [tyːr]) is a Germanic god ...
I would expect the API response to include the IPA pronunciation, as they used to do.
This started happening recently, maybe a month ago. The second JSON snippet that I pasted is from 2017-01-27, so back then it was definitely working. Also notice that the first IPA (/ˈtɪər/, within //) is missing, whereas the second one ([tyːr], within []) is still included in the API response: perhaps it has something to do with the //. If necessary I can provide more examples and try to determine more accurately the date when this started happening. — Preceding
unsigned comment added by
2602:306:B89F:F850:61B0:46C5:989D:75 (
talk)
19:43, 17 October 2017 (UTC)
Notice the empty () (or (<span></span>) in the API response).
If the IPA is removed, the () or ; should be removed as well, otherwise the API response looks broken. However I'm not sure why it was removed in the first place, since it's useful and not intrusive. — Preceding
unsigned comment added by
2602:306:B89F:F850:61B0:46C5:989D:75 (
talk)
16:10, 18 October 2017 (UTC)
Twitter Cards / Facebook Open Graph
Both Twitter and Facebook have techniques where a website owner can specify metadata that allows a posted link to be expanded into a small preview. Such previews generally show an image, the page title, the publisher, etc. Many media companies, e.g. CNN, NYTimes, etc., embed support for both
Twitter Cards and Facebook's
Open Graph, which allows their content to be previewed when linked from either platform. Presently, Wikipedia supports neither.
I'm not really interested in arguing the merits either way, but I'm curious if Wikipedia's current lack of support is due to a conscious choice NOT to support such technologies, or is it more the case that no one has taken the time to try and do so?
Dragons flight (
talk)
11:52, 18 October 2017 (UTC)
@
Dragons flight: A little bit of both. There is reservations about adding tons of vender specific tags to all our HTML content, so generally, we tried to avoid that. We did implement all the fallback strategies that those platforms supported, so most of the time, it's pretty good actually (unless you want to share an imagepage). See also
phab:T56829 and it's many connected tickets. —
TheDJ (
talk •
contribs)
14:51, 18 October 2017 (UTC)
I'm not sure about those fallback strategies. Facebook does seem to generate a preview from Wikipedia content, but Twitter just seems to leave a bare link, e.g. most of the tweets at
[12]. It was because of the Twitter case that I actually thought to ask. Thanks for the Phabricator reference.
Dragons flight (
talk)
15:07, 18 October 2017 (UTC)
You don't happen to know if any of the existing extensions in this area are reasonably functional? (For potential use on a non-WMF Mediawiki installation.)
Dragons flight (
talk)
15:27, 18 October 2017 (UTC)
Currently, Mobile view of Wikipedia's Main Page doesn't show the Did you know section. I think this should be shown as a large number of readers visit Wikipedia from mobiles and tablets, I-pads etc and DYK's aren't being exposed to them. Also, some topics may share interests with mobile readers, like an Apple smartphone hook, which recieves less views on Computers but more from Mobiles and Tablets. What do you guys think about this ? Request a passing administrator to implement this if enough users agree.
31.215.114.79 (
talk)
11:37, 17 October 2017 (UTC)
Actually its quite simple, the existing main page layout can't be made responsive because Wikipedia has very limited global CSS which does not allow for it. The other option, that of changing the layout, has already been subject to wheel warring and bitter infighting. So nothing happens. And before you say we should have a main page just for mobiles, that can't be done without breaking the mediawiki software.
Dysklyver20:08, 17 October 2017 (UTC)
I recognise the situation of heaven-blocking "no consensus so no change" (me in lower level situations ;-) ). Does not this sort of point out that the Wikipedia consensus model needs improvement? -
DePiep (
talk)
21:41, 17 October 2017 (UTC)
I am messing about with conversions from {{
familytree}} to {{
chart}}. I recently filled in the missing tiles in chart so that chart can now handle all of the familytree tiles and more. For reasons that are not clear to me there is no one to one mapping of tile names between familytree and chart, but over a decade ago
User:GregU wrote a tool that may help (see
User:GregU/familytree.js).
I am messing about with a fairly complicated tree that appears in the article
Family tree of Muhammad (the first one under the section "Family tree"). I have already converted it by hand so this is a test. As can be seen by looking at the documentation that comes with the templates {{
familytree}} to {{
chart}} the tile letter that represents a dashed line is a Tilde "~". To represent a line of dashes GregU's tool produces a line of tildes in the ASCII representation. The problem is that if one then saves the ASCII representation, the tildes get converted into signatures.
As GregU no longer appears to be active,I can not ask her/him for help, so can any one tell me if there is a way to save thee or more tildes in a row without them being turned into signatures? --
PBS (
talk)
14:08, 18 October 2017 (UTC)
A self-closed nowiki tag inside will only work if there are at most four tildes. I don't know your exact use case but a single <pre>...</pre> around a whole block of text may also be an option. This creates a
monospaced font and preserves linebreaks like the below example.
PrimeHunter (
talk)
15:40, 18 October 2017 (UTC)
Use a proper svg capable program like abode illustrator or inkscape to convert it, basically the transparency/colors/code/format/specification/some thingies aren't supported by whatever mediawiki is using.
Dysklyver19:19, 18 October 2017 (UTC)
"Do you really want to remove page X from your watchtlist?"
Today, viewing my watchlist, I can click to remove a page. All fine, very professional. But hey. Randomly I must confirm, and even then we do not return to the page came from. -
DePiep (
talk)
01:55, 16 October 2017 (UTC)
@
DePiep: This is likely because the page had not finished loading Javascript, or because an error in your scripts had an error and then interrupted the rest of the Javascript initialisation. —
TheDJ (
talk •
contribs)
09:38, 16 October 2017 (UTC)
To be clear: viewing my watchlist I can click a linked "×", then the page is struck <s>-style in view, all right. In other moments (IMO randomly), by that same "×" click I must confirm in a page clicking "Yes", and then click to return to my Watchlist. This is the not funny route I complain about.
Firefox 52.4.0 (=latest IMO). I think there is more in TheDJ's note: very slow computer, so scripts may not be active. -
DePiep (
talk)
20:38, 19 October 2017 (UTC)
@
DePiep: "I can click a linked "×", then the page is struck -style in view i'm still wondering what is giving you that functionality... Cause i don't have it, and i can't find it either. :) —
TheDJ (
talk •
contribs)
21:33, 19 October 2017 (UTC)
In my Watchlist, it is in the second column (first column is: blue/green colored ball/triangle-to-unfold; third+ columns are the "m, b" indicators). It's a wikilink, so "×" has regular underline. When fast, the whole page line is struck and greyed out right away (good). When slow, the issue at hand, it leads to url
[13] (demo page infobox here; of course wrt my own watchlist). Could be related to a Preference I have set. Clear enough? -
DePiep (
talk)
21:52, 19 October 2017 (UTC)
Hey, I seem to a recall some years back these were moved to meta, and reading a description of the new way to handle them at that time, but I haven't thought about it much since. All the ones I added by the old method were automatically fixed -- I think? But that doesn't help me fix cases where they are already messed up.
Specifically, someone linked the Japanese article
ja:仏足石歌 from our article
bussokusekika, but our article was actually on the different (but related) topic covered in the Japanese article
ja:仏足石歌碑. I moved the page (the earlier title was technically inaccurate, as "bussokusekika" describes a poetic form not exclusive to the
bussokuseki-kahi).
But now I don't know how to change the link to ja.wiki to match the identical topic.
I have tried in two different places, one in main space and one in my user sandbox, both with {{toc right |limit=3 or 4}} and with {{toc limit | 3 or 4}} and neither work, yes I am setting the correct respective number of "=" in the section titles. Only a limit of 2 seems to work. Any ideas?
Eno Lirpa (
talk)
13:07, 21 October 2017 (UTC)
This question is about the Quarry SQL interface. If there's a better place to ask, please let me know.
A query I'm developing takes a long time to run. Is there any way to see the query plan without actually running the SQL? (The Explain button only appears once the SQL is running, and the EXPLAIN statement fails because I don't have the necessary table permissions.) If not, is there any way to stop a query once it's running and I can see from Explain that it's going to hog resources forever without producing useful results?
The specific query is
DabPairs, which aims to identify when Foo and Foo (disambiguation) are distinct dab pages (not redirects to each other) but a solution would also be of more general use. The specific problem I'm having is that MySQL refuses to use the name_title index on the page table, despite my specifying a constant P2.page_namespace and an expression for P2.page_title. I've tried using STRAIGHT_JOIN (which successfully forces the join order I'm aiming for) and FORCE INDEX (which fails, telling me that name_title isn't a key) but nothing seems to work.
Is it just me or the spacing between a hatnote and "From Wikipedia, the free encyclopedia" has indeed shrinked?
Twosamples. The upper spacing was rather identical to the bottom spacing between a hatnote and the article's first sentence several days ago (Windows Firefox, default 100% zoom).
Brandmeistertalk13:13, 20 October 2017 (UTC)
Hello! Is there any practical way in which we could have a word count of articles of a given Wikipedia? Maybe enwp is too big, but in euwp we don't have so many articles and this could be a great information on our quality.-
Theklan (
talk)
12:43, 22 October 2017 (UTC)
If you want a table showing the word count for all articles at the same time, I can't help you, but if you're only interested in the current article, the following can be added to
Special:MyPage/common.js:
It's not perfect (the last word in paragraph 1 will count as the same word as the first word in paragraph 2), but if you want a more accurate count, it quickly becomes a lot more complicated. It's also possible to filter out infoboxes or other arbitrary templates if you just throw more code at it, if that's what you want to do.
Nirmos (
talk)
13:36, 22 October 2017 (UTC)
/wikipedia/commons/8/80/Farewell_to_442402_Beckenham_Junction_to_London_Victoria_1Z43_approaching_Beckenham_Road_station_(32556548674).jpg
We could not find the above page on our servers.
Did you mean: /wiki/wikipedia/commons/8/80/Farewell_to_442402_Beckenham_Junction_to_London_Victoria_1Z43_approaching_Beckenham_Road_station_(32556548674).jpg
Alternatively, you can visit the Main Page or read more information about this type of error.
Pkbwcgs (
talk)
08:20, 7 October 2017 (UTC)
Do you have the media viewer option turned on? If so, you could try disabling it, in case that helps. The image shows up for me (including its large version). —
PaleoNeonate –
08:51, 7 October 2017 (UTC)
@
PrimeHunter: Both versions are not working. The error message comes after going to the image – there is an image link instead of the image. If I click on the image link, it leads to error 404 and the error message I put right at the top of this thread. If I am too vague, I will post a screenshot.
Pkbwcgs (
talk)
12:29, 7 October 2017 (UTC)
@
PrimeHunter: It still quite a bad problem. As soon as I clear the history, the images will work for 5 to 10 minutes and after that, the images will suddenly stop working. This problem suddenly happened on Friday and I still didn't get a solution.
Pkbwcgs (
talk)
14:36, 8 October 2017 (UTC)
@
PrimeHunter: The link you have given is working. However, the pictures are still not showing up. The main page in the error is linking to Wikimedia Common's main page. However, sometimes even opening Wikimedia commons is throwing up this error below:
Error
'Our servers are currently under maintenance or experiencing a technical problem. Please try again in a few minutes.
See the error message at the bottom of this page for more information.'
Again, I am being held up with technical difficulties on Wikimedia Commons. Is there any maintenance work going on?
Pkbwcgs (
talk)
19:56, 8 October 2017 (UTC)
Commons currently works for me but I have periodically seen such error messages for years. It's unrelated to the other problem. If the "Main Page" link goes to Commons then it sounds like you are at commons.wikimedia.org when you see the error message while you should have been at upload.wikimedia.org after you click an image to view the image directly and not just on a file page. But I cannot be sure because you still haven't answered what the address is. Maybe there is a problem with your browser. Do you have any browser extensions which may manipulate addresses?
PrimeHunter (
talk)
21:36, 8 October 2017 (UTC)
Which of the problems? "We could not find the above page on our servers" and "Our servers are currently under maintenance or experiencing a technical problem" are completely different. And upload.wikimedia.org is only a domain. Please post a full address where you see the problem, e.g.
https://upload.wikimedia.org/wikipedia/commons/7/70/Example.png which works for me. Copy-paste the address from your browser to be sure to get the precise address.
PrimeHunter (
talk)
01:04, 11 October 2017 (UTC)
404 Not Found
The resource could not be found.
File not found: /v1/AUTH_mw/wikipedia-commons-local-public.70/7/70/WrongExample.png
It's a different message and it has no link to Commons or anywhere else. Contructing a wrong link with the right directory per
mw:Manual:$wgHashedUploadDirectory gives me the same result:
https://upload.wikimedia.org/wikipedia/commons/b/bf/WrongFilename.png. Please confirm that your browser is still on the address https://upload.wikimedia.org/wikipedia/commons/7/70/WrongExample.png when you see "We could not find the above page on our servers", and a "Main Page" link going to Wikimedia Commons. Do you know where to see which address you are on? Can you try another browser?
File:Wikimedia error 404.png is only a partial match to the message you quoted in the first post, but the file is five years old and from another domain so that's not surprising.
PrimeHunter (
talk)
16:15, 11 October 2017 (UTC)
@
PrimeHunter: Sorry for the delayed reply. I can confirm that the web address is definitely upload.wikimedia.org. Most of the images are still not working.
Pkbwcgs (
talk)
19:07, 18 October 2017 (UTC)
@
PrimeHunter: Also, after a quick test, apparently this issue I am facing at the moment is on Google Chrome and not Internet Explorer. However, I still prefer Google Chrome to Internet Explorer. Will this be enough to help me out on my issue?
Pkbwcgs (
talk)
20:36, 22 October 2017 (UTC)
Hillary Clinton template
I've located an issue with {{Hillary Clinton series}}, but I can't figure out how to fix it and wanted to ask for some assistance. The issue is that unlike any other template in ‹The
templateCategory link is being
considered for merging.›Category:"Part of a series on" sidebar templates, this template is resulting in the direct inclusion in that category of every individual articlespace page that transcludes the template — but it's not supposed to be doing that, because that category is for the templates themselves, not for every mainspace article that has one of them on it. But I can't find the category code to determine how to fix it. Can anybody with more template experience take a look at this? Thanks.
Bearcat (
talk)
19:41, 22 October 2017 (UTC)
Purge removes the category from the bottom of the article, but it doesn't remove the article from the list on the category page. --
Redrose64 🌹 (
talk)
20:42, 22 October 2017 (UTC)
Assistance request: Enabling citoid templates on Wikiversity
Hi, I'm trying to enable Citoid templates on Wikiversity (
discussion). The
configuration process is slightly beyond my technical expertise. Would someone from Wikipedia be able to help me to make sure I don't break anything? Secifically, I need to enable the ability to autogenerate the {{
cite journal}} from a DOI and PMID code.
T.Shafee(Evo&Evo)talk06:22, 22 October 2017 (UTC)
A few weeks ago I changed the theme I am using in Preferences, but it has now changed back to Vector. I'm not sure what it was called, but I don't see any news about it being removed. Where did it go? (I see a RFC about adding "Timeless" skin, that started a couple days ago, maybe that's the one I was using, but that's for *adding* a skin, not removing one.) I'm confused. Thanks! —{{u|
Goldenshimmer}}|✝️|ze/zer|😹|
T/C|☮️|John15:12|🍂02:53, 20 October 2017 (UTC)
I see. Dunno why they fix what ain't broken, but whatever :P Fixed it by adding some conditional redirects to my browser, in case anyone else is wondering:
MinervaNeue is very similar to the mobile version (
example) of Wikipedia so you can also try that. It doesn't require browser tricks. In the desktop version of the site, click "Mobile view" at the bottom of a page. Some things like navigation templates are omitted from the mobile version.
PrimeHunter (
talk)
19:24, 20 October 2017 (UTC)
Thanks! I prefer the one with the proper skin over the mobile one I think though, it doesn't fold up sections for one thing. Thanks though! (I discovered my redirects above don't work right with anchor fragments in URLs, though; here is a fixed replacement (struck my earlier version):
It will be back Thursday!! A patch failed to merge before we deployed. It's still there but sadly not accessible via preferences. You'll need to use the api to enable it in the mean time if that's too long to wait. Sorry about that!
Jdlrobson (
talk)
14:48, 23 October 2017 (UTC)
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
You can use ccnorm_contains_any when you create an abuse filter. This can be used to look for multiple words or phrases within a string. It will find words where some characters have been replaced. You can read more
in the documentation.
[14]
Changes later this week
When you search in some languages the search function could use specific other languages if the first language didn't work. This is called a fallback language. This didn't work properly and created bad searches. The search index is being fixed to work better.
[15]
The
new version of MediaWiki will be on test wikis and MediaWiki.org from 24 October. It will be on non-Wikipedia wikis and some Wikipedias from 25 October. It will be on all wikis from 26 October (
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
24 October 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
25 October at 15:00 (UTC). See
how to join.
The lead states: "A tip of the day is a piece of advice that is given daily. For example, many computer programs, such as those in older versions of the Microsoft Office suite, will display a tip of the day when they are started."
Do you know of any major programs that provide a daily tip (upon startup)?
Some games do on their loading screens, particularly games like the Football Manager series that aren't aimed at traditional gamers and thus there's an assumption that a significant proportion of players will be unfamiliar with gaming conventions. I'm unaware of any major piece of non-game software that still does this (with the obvious exception of Wikipedia); some Microsoft programs used to back in XP and Clippit the Paperclip days but they discontinued it years ago. The computing sense is certainly not the primary use of the term, since I'd be willing to bet (sic) that well over 90% of usage of the phrase relates to sports betting. ‑
Iridescent22:24, 22 October 2017 (UTC)
What qualifies as a major software?.
gimp,
solidworks,
VMware,
Sketch Up,
Intellicad? These all have or had TotD. I would suggest however that TotD is more commonly associated with online help, support websites, betting sites, health, product, wellbeing websites rather than "on-startup" widgets these days. A simple google search for "tip of the day" shows that. Club
OranjeT23:13, 22 October 2017 (UTC)
You need |1= when the first unnamed parameter contains an equals sign (|text= would also work for this template). Otherwise it's interpreted as assigning the right hand sign of the equals sign to a parameter with the name on the left side.
PrimeHunter (
talk)
00:05, 23 October 2017 (UTC)
Is there a bot that can update links to Talk page topics once those topics get moved to an archive? Or does this have to be done manually? Thanks. SharkD Talk 14:45, 24 October 2017 (UTC)
Is it possible to list deleted user contributions from a given year/month, in the same way that current contributions can be listed? Obviously this only applies to admins.
Optimist on the run (
talk)
11:39, 24 October 2017 (UTC)
There is no interface to do it but you can click "older 50" on deleted contributions and manually change the timestamp in the url. The format is offset=YYYYMMDDhhmmss.
PrimeHunter (
talk)
12:23, 24 October 2017 (UTC)
A lot of Canadian radio stations are stuck at the moment with a dead external link on them, because a site that we regularly rely upon as an EL for Canadian broadcast outlets has changed its URL structure — all of the pages still exist, but the URLs have changed and the old ones go 404. Fortunately, there's a regular, entirely predictable format to the new URLs, so it should potentially be possible to just set a bot with a substitution rule loose on the affected articles — but I don't know whether there is a bot than can do this or not, or whether we're just going to have to make the changes manually.
So, is there a bot that can do the URL switches for us? If there is, I'm willing to provide the needed information to the programmer. Thanks.
Bearcat (
talk)
22:01, 24 October 2017 (UTC)
I know my bot can handle this. What's the old structure and what changes need to be made to it in order to comply with the new structure?
Nihlus22:03, 24 October 2017 (UTC)
Thanks. This is going to be slightly trickier for AM radio stations than it is for FM, and we're willing to look after those manually if it would be too complicated, but here's the basics: what we need is a recursive run through all radio stations in ‹The
templateCategory link is being
considered for merging.›Category:Radio stations in Canada and its subcategories.
For the URL format, I'm going to use the example of
CBO-FM: the old format is
http://www.broadcasting-history.ca/listings_and_histories/radio/histories.php?id=452&historyID=226 , while the new one is
http://www.broadcasting-history.ca/listing_and_histories/radio/cbo-fm. So what the bot would need to do is replace the word "listings" in the URL with the singular "listing", and replace "histories.php" and all of the data after it (the digits used to be the part that varied from station to station) with the actual call sign of the station (basically a straight substitution of the page title itself.) The digits in the old histories.php don't actually matter, and are relevant only in the sense that the bot needs to be programmed to strip that part of the URL regardless of what the numbers after the equals signs are.
For FM stations, this should be entirely straightforward, since "-FM" is both the format of their new URLs and the format we use in our page titles. So for these stations, all it would have to do is replace the histories.php data with the page name.
For AM stations, it's a little bit trickier. The new URLs at CCF all use -AM, whereas our articles vary depending on whether disambiguation is needed or not. So what the bot should do is, if a radio station has a title with an (AM) disambiguator in it, it should just add the call sign with an -AM suffix instead of (AM), and if the title is just four letters with no (AM) or -FM suffix at all, it should also add an -AM suffix.
We also have a few radio stations that have extra disambiguatory information beyond just "(AM)" or "-FM", such as for defunct stations whose call signs then got taken over by a different new station later on. Usually this is "(defunct)", but not always. For stations with that kind of title, it would be too complicated to program the right rule for a simple bot substitution — so for any radio station that has a more complex title than just "CXXX", "CXXX (AM)" or "CXXX-FM", just have the bot just skip those and we'll look after them manually. As well, I've already manually replaced the URL on CBO-FM, since that was the example initially brought to my attention. And, of course, the bot should obviously just skip any page that doesn't have a broadcasting-history.ca URL on it at all — and the category tree also includes some redirects, so it should skip those as well.
We'll look after the wonky-titled pages manually, like I said, and if the AM station rules are too complicated to program we'll take them on as well. But whatever the bot is able to help with, we'd be most grateful.
Bearcat (
talk)
22:27, 24 October 2017 (UTC)
So you only want to fix those links under "External links" sections? It might be best to create a template similar to other EL templates and then just replace the ones on the articles with the template. Then it could also make any future changes to their url structure easier. Perhaps it should be titled {{BroadcastingHistoriesCanada}} or some variant. I'll start working on the template and then we can look to have the bot do the dirty work.
Nihlus22:56, 24 October 2017 (UTC)
Browser tab HTML title doesn't have the free encyclopedia while other's do
I noticed that the HTML title data on simple english wikipedia has the free encyclopedia on
every article,
<title>John Doe – Simple English Wikipedia, the free encyclopedia</title>
while the title on the regular English Wikipedia
is just,
<title>John Doe – Wikipedia</title>
It might be understandable with the simple english wikipedia as they don't don't have the motto under the globe in the top left, but other wiki's like the Portuguese wiki, which has "the free encyclopedia" under the globe, also has it in the HTML title data
as well which contradicts the idea of it being on the HTML title when it isn't under the globe(albeit as far as I have checked ~all wiki's have it atleast on the main page. The german wikipedia
doesn't have the motto in the HTML title data, but does in the globe. So is it just seemingly random as to the use of the motto in the HTML, is there a reason for it, or is it something that can be fixed?(Also if this belongs better on the meta as it is crosswiki related please point me there)
-glove- (
talk)
00:45, 25 October 2017 (UTC)
Administrators can see the deleted page history of
MediaWiki:Pagetitle. It included "the free encyclopedia" from 20 August 2005 to 12 October 2016. Many other wikis probably copied this from us in that period.
PrimeHunter (
talk)
01:35, 25 October 2017 (UTC)
Thank you for the detailed response, also has this been proposed over at the simple english wiki to be changed? Probably wouldn't be be best as it isn't under the globe though but I was wondering if there was a discussion on it.
-glove- (
talk)
01:52, 25 October 2017 (UTC)
@
Th4n3r: In case that's not clear, each repetition of "2001:0db8:0000:0042:0000:8a2e:0370:7334" in the wikisource is a significant fraction of the maximum edit comment length, leaving very little for the undoer to write anything. It was changed from what you say it should be to the shorter wikitext.— Preceding
unsigned comment added by
104.153.72.218 (
talk)
22:27, 4 November 2017 (UTC)
Structured Commons newsletter, October 25, 2017
Welcome to the newsletter for Structured Data on Wikimedia Commons! You can
update your subscription to the newsletter. Do inform others who you think will want to be involved in the project!
Translation. Do you want to help out translating messages about Structured Data on Commons from English to your own language? Sign up on the
translators page.
The documentation and info pages about Structured Data on Commons have received a thorough update, in order to get them ready for all the upcoming work. Obsolete pages were archived. There are undoubtedly still a lot of omissions and bits that are unclear. You can help by editing boldly, and by leaving feedback and tips on the talk pages.
We have started to list tools, gadgets and bots that might be affected by Structured Commons in order to prepare for a smooth transition to the new situation. You can help by adding alerts about/to specific tools and developers
on the dedicated tools page. You can also
create Phabricator tasks to help keep track of this. Volunteers and developers interested in helping out with this process are extremely welcome –
please sign up!
Structured Data on Commons was presented at Wikimania 2017 in Montréal for a packed room. First design sketches for search functionality were discussed during a breakout session. Read the Etherpad reports of
the presentation and
the breakout session.
Sandra Fauconnier, Amanda Bittaker and Ramsey Isler from the Structured Commons team will be at WikidataCon. Sandra
presents Structured Commons there (with a focus on fruitful collaboration between the Wikidata and Commons communities). If you attend the conference, don't hesitate to say hi and have a chat with us! (
phabricator task T176858)
Team updates
Two new people have been hired for the Structured Data on Commons team. We are now complete! :-)
Ramsey Isler is the new Product Manager of the Multimedia team.
Pamela Drouin was hired as User Interface Designer. She works at the Multimedia team as well, and her work will focus on the Structured Commons project.
Partners and allies
We are still welcoming (more) staff from GLAMs (Galleries, Libraries, Archives and Museums) to become part of our long-term focus group (
phabricator task T174134). You will be kept in the loop of the project, and receive regular small surveys and requests for feedback. Get in touch with
Sandra if you're interested – your input in helping to shape this project is highly valued!
Jonathan Morgan and Niharika Ved have held interviews with various GLAM staff about their batch upload workflows and will finish and report on these in this quarter. (
phabricator task T159495)
At this moment, there is also an online survey for GLAM staff, Wikimedians in Residence, and GLAM volunteers who upload media collections to Wikimedia Commons. The results will be used to understand how we can improve this experience. (
phabricator task T175188)
Upcoming: interviews with Wikimedia volunteers who curate media on Commons (including tool developers), talking about activities and workflows. (
phabricator task T175185)
The multimedia team at WMF is gaining expertise in Wikibase, and unblocking further development for Structured Commons, by completing the MediaInfo extension for Wikibase.
Generating lists of articles created by a sockfarm
I help out a lot at
WP:COIN. Often one of the things we do is create lists of articles created by sockfarms, for review and cleanup. An example of such a list is
User:Bri/COIbox63. A new case created by Doc James today on COIN (we get 2-3 a week sometimes) made me think this might be better automated. I got as far as creating pagepile 11077 as a test, using a mix of crawling sock categories and manually entered untagged accounts, then got stuck. Is there a way to go forward and get a list of articles created by the accounts in the pagepile? - ☆
Bri (
talk)
20:26, 18 October 2017 (UTC)
I am trying to add Wikipedia's notes templates to my wiki (hosted by Wikia.com, aka Fandom). I believe it has the default MediaWiki pages, with the exception of those that were edited. Which MediaWiki pages are needed for the notes templates ({{Notelist}} and {{Efn}} and variants) to work properly? AChildOfGod(
Talk |
Contribs | Bible Wiki)19:55, 25 October 2017 (UTC)
I was looking at the article when I noticed that it was saying this at the top of the Oceania section:
...which came to power through a military coupin December 2006. The United Statessuspended $2.5 million in aid money pending a review of the situation...
Inside the table code was text not belonging to a cell and therefore displayed outside the table. This can give odd results like the below example where the source has a space but some code cleanup process removes it from the generated html.
Xaosflux, the issue posted by MrHumanPersonGuy was fixed with this revert of the Foreign relations article as mentioned above by Optimist on the run. Or do you mean that the result of PrimeHunter's example doesn't look like:
@
Anomalocaris: One possibility is page archiving. You might have come to this page when this version was current, or perhaps an even earlier one. If you did't reload the page for some hours before attempting to edit the section that you mention, this archiving edit may have occurred in the meantime. This will have altered the section numbering for all sections from "
IP was somehow able to edit a semi-protected page" onwards, so that each one now had a true section number one less than the number used by the "[edit]" links in that old version. So using one of those out-of-date "[edit]" links will have caused editing of the section below. --
Redrose64 🌹 (
talk)
09:16, 26 October 2017 (UTC)
Yes, section edit links don't name the section but just have a url ending in a section number like &action=edit§ion=32. If the wrong section is edited then you can try to
bypass your cache or
purge the page. Or just click another edit link like you did, or manually change the number in the url.
PrimeHunter (
talk)
09:39, 26 October 2017 (UTC)
Hi everyone, I used to be a pretty active Wikipedia contributor for about a decade, but took a long hiatus where 1) forgot my password, 2) lost access to the recovery email I have on file.
I'd love to resume contributing to Wikipedia, and would much prefer to use the old account I have rather than the new one, but I can't use the traditional recovery method. Is there any way admins could grant me access to my (still existing) account if I provide my expired recovery email address or anything of the sort?
Ah, now I realize that piping the IPv6 number consumes too many characters, and avoiding that is worth the ugliness when lacking a better solution. Has there been any thought about making MediaWiki handle this better? When noticing this, it jumped out as looking plain wrong, even though it's a sensible workaround. @
PrimeHunter. --
Pipetricker (
talk)
16:53, 17 October 2017 (UTC)
Mostly this would be fixed with longer edit summaries (
phab:T6715), the work for which has mostly been done in
phab:T166732, but deploying it to Wikimedia is invasive and will occur very carefully. That deploy process is tracked at
phab:T166733 and likely will take a few more months. —
TheDJ (
talk •
contribs)
09:07, 18 October 2017 (UTC)
Context-specific editor not allowing capital letters
Hi, I have suffered in my editing for about a week, since the rollout of new editor interface which I am using now. Very frequently, but not always, when I try to insert a capital letter or any character which involves the shift key (including the colon character, angle brackets, squiggly brackets at least on my keyboard), the cursor jumps to the left/beginning of the paragraph. It is exasperating! This is using Chrome browser in a Windows 10 environment, with whatever javascript or whatever user-specific settings that I have and don't particularly know about. What is going on? It seems worse editing in the Imicrosoft edge browser. It doesn't happen much in new text entry; it is horrible when I am trying to edit a reference or other template call. Help! --
doncram22:52, 20 October 2017 (UTC)
Thank you! It makes sense that I was one of relatively few editors experiencing this via my apparent previous choice to enable beta features, because otherwise WikiProject NRHP and other places would have blown up with complaints. I see that going to that preferences section also allows me to contribute feedback about the "New wikitext mode", which i will do. --
doncram00:01, 21 October 2017 (UTC)
Patrolling new articles, I came yesterday across this article:
Serapeum of Alexandria. It was started as a redirect in February, and then started as an article yesterday (by moving text from another article with an attribution, but this is irrelevant for the story; the article is not problematic in any sense). I was checking for copyvio, which is part of my workflow when I deal with NPP, and I found that the article was actually indexed by Google. If I am right, it was indexed by Google because the redirect, which existed here for 8 months, allowed the article to bypass the 3 month threshold of not being indexed by search engines. Again, if this is correct, this is a bug, not a feature. I would propose to count the three months from the date of an actual creation of the article in the case the preceding history only includes redirects (and possibly blanking redirects, just in case).--
Ymblanter (
talk)
07:25, 21 October 2017 (UTC)
There's no technical difference between a page that exists with the content of indicating a redirect and the content of an article. What you propose is therefore not easily implementable. But as allways feature requests can be filed in phabricator. —
TheDJ (
talk •
contribs)
09:18, 21 October 2017 (UTC)
Thank you. I believe we need consensus here that the feature is needed (or at least there is no consensus against implementing it) before we file a phabricator ticket.--
Ymblanter (
talk)
22:05, 21 October 2017 (UTC)
You don't need to get written proof of consensus in advance. If the relevant team believes that it's necessary or appropriate, then they'll tell you.
Whatamidoing (WMF) (
talk)
16:50, 26 October 2017 (UTC)
Tool to detect SPA-owned articles?
I wonder if there is any way to create a tool that would detect articles that were either created by SPAs and/or are heavily trafficked by them. I think such a tool would be of great utility to fighting paid editing. Pinging
Smallbones in case he has any thoughts on the subject. Perhaps one of our preexisting tools already have that capacity and I'm not aware of it.
Coretheapple (
talk)
17:18, 26 October 2017 (UTC)
Search intitle: doesn't work right with quoted strings that include a space
intitle:"word1 word2" (notice the space between the words) returns results that do not have the search string in the title.
It seems to happen for any search string with a space in it.
So, intitle:"Milan" produces only items with Milan in their titles, while intitle:"in Milan" includes items which has this in the title, but also includes items that have this only in the article contents and not in the title. This seriously waters down the search results.
This happens to me too. Rather annoying. It seems contrary to spec too? If my recollection is correct it was working correctly, or at least much better, two to three months ago?
Eno Lirpa (
talk)
00:46, 22 October 2017 (UTC)
I think it's because "in" is a common word which is ignored in some search contexts. I get 400 hits on both intitle:"in Milan" and "in Milan" intitle:Milan. The latter seems to be the actual search performed for both. intitle:"word1 word2" works as expected in my tests with two uncommon words. For example, intitle:"Inter Milan" gives me 56 hits, and all have the quote in the title.
PrimeHunter (
talk)
01:25, 22 October 2017 (UTC)
Si on oublie le "/wiki"... the message is surprising
Everyone knows that you can often (depending on what characters are in the title) construct the URL for a Wikipedia article by simply prefixing http://en.wikipedia.org/wiki/ to the title. And if you forget the /wiki part and write something like http://en.wikipedia.org/Paris, the server will kindly give you a page that links to the correct page, saying:
From this I also wonder if maybe there are other error messages that the servers produce that are similarly in English when they should be in other languages. --
69.159.60.147 (
talk)
23:16, 26 October 2017 (UTC)
Using popups, when I hover over any "hist" link on my watchlist, it shows every single edit to every article without exception as being a minor edit. Now I'm wondering how long it's been doing this and if I've only just noticed. When I load a history page, it's all normal. Anyone else?
RivertorchFIREWATER15:36, 26 October 2017 (UTC)
@
Rivertorch: I’ve been doing some long overdue maintenance on nav popups and this was a result of that. Should be fixed now. You might see some more minor breakages the coming months as I beat this behemoth of 2003–2007 javascript code into submission. —
TheDJ (
talk •
contribs)
19:30, 26 October 2017 (UTC)
Yes, it seems to be fine now. Thanks for responding. More importantly, thanks for your work in maintaining what is surely the most useful and functional tool around.
RivertorchFIREWATER03:42, 27 October 2017 (UTC)
Notifications stuck at "1"
Log in, get a blue "4" for notifications. Click on it, I get the annoying dedicated "notifications" page instead of the useful dropdown (no idea why I get one or the other, seems to happen randomly). I tag as "read" two enwiki notifications, tag as read the first non-enwiki notification, and the second non-enwiki one disappears from the list and can't be clicked. The counter is now stuck at a grey "1".
Now, if I click that "1", I get the dropdown with "More notices from another wiki", and when I click "view 1 notice" I get the name of the wiki (Mediawiki) but no page to visit. I can click the small blue button at the top right, and the grey "1" disappears ... for about a second, and then it reappears. Rather annoying. Am I the only one with this issue, or is this some regression / bug with the most recent Thursday deployment?
Fram (
talk)
07:10, 27 October 2017 (UTC)
I made
this edit to a fully-protected article, not realising said article had been fully protected. Since the edit could potentially be seen as admin abuse (using ability to edit to impose my edit where a non-admin could make said edit or revert it), would it be possible to tweak HOTCAT so that it brings up the edit page on trying to save, thus alerting the editor that the article is fully protected. This would give the editor the chance to bail out of the edit and go to the article talk page. For the record, I reverted the edit after discussion at the talk page. It was not my intention when making the edit to use my tools to gain an advantage.
Mjroots (
talk)
16:24, 27 October 2017 (UTC)
These two articles:
mile run &
two mile run had wrong interwiki link to
Farsi. I have fixed the link in wikidata and it shows up correctly in Farsi wiki but on the English wiki it's still showing up wrong even though when you click on edit and go to wikidata page the links there are correct.
Bardia90 (
talk)
18:41, 28 October 2017 (UTC)
After using & editing Wikipedia for several years using a tablet set to "landscape" mode, I've just noticed that when two pictures are included next to each other, they instead stack one above the other. For example, the images of the band members at
HIM (Finnish band)#History appear on my screen with Ville Valo above Mige, even though the caption says "Ville Valo (left) and Mige (right) formed His Infernal Majesty in 1991". It's a minor niggle, but is there any way with the mobile version to detect which mode is being displayed and toggle between "left / right" and "top / bottom"?
~dom Kaos~ (
talk)
19:00, 28 October 2017 (UTC)
The example uses {{Multiple image}} with a common caption in the footer parameter. I have mentioned the issue in the documentation.
[20] My mobile tests with individually specified widths gave "line wrapping" of the images when the total number of image pixels in a "line" exceeded 320 − 4×(number of images). I don't know how consistent this is but for me it seemed independent of window width and zoom. I only tested the mobile version in a desktop browser.
PrimeHunter (
talk)
22:17, 28 October 2017 (UTC)
Rollback edit marked minor
Interestingly this was marked as minor (and the result of me clicking rollback). I noticed that ClueBot's revert was also marked minor. Is this a bug? Thanks, —
PaleoNeonate –
18:08, 27 October 2017 (UTC)
I've been writing scripts to view and to automate the development of outlines, and I was wondering if you would take a look to see if there is anything I could be doing better (in my programming approach, programming style, etc.). So that other programmers can follow along with how the source code works, I've provided extensive notes on the scripts' talk pages.
So far, there is:
User:The Transhumanist/OutlineViewAnnotationToggler.js – this one provides a menu item to turn annotations on/off, so you can view lists bare when you want to (without annotations). When done, it will work on (the embedded lists of) all pages, not just outlines. Currently it is limited to outlines only, for development and testing purposes. It supports hotkey activation/deactivation of annotations, but that feature currently lacks an accurate viewport location reset for retaining the location on screen that the user was looking at. The program also needs an indicator that tells the user it is still on. Otherwise, you might wonder why a bare list has annotations in edit mode, when you go in to add some. :) Though it is functional as is. Check it out. After installing it, look at
Outline of cell biology, and press ⇧ Shift+Alt+a. And again.
User:The Transhumanist/RedlinksRemover.js – strips out entries in outlines that are nothing but a redlink. It removes them right out of the tree structure. But only end nodes (i.e., not parent nodes, which we need to keep). It delinks redlinks that have non-redlink offspring, or that have or are embedded in an annotation. It does not yet recognize entries that lack a bullet (it treats those as embedded).
It is my objective to build a set of scripts that fully automate the process of creating outlines. This end goal is a long way off (
AI-complete?). In the meantime, I hope to increase editor productivity as much as I can. Fifty percent automation would double an editor's productivity. I think I could reach 80% automation (a five-fold increase in productivity) within a couple years.
There's more:
User:The Transhumanist/StripSearchInWikicode.js – another script, which strips WP search results down to a bare list of links, and inserts wikilink formatting for ease of insertion of those links into lists. This is useful for gathering links for outlines. I'd like this script to sort its results. So, if you know how, or know someone who knows how, please let me know. A more immediate problem is that the output is interlaced with CR/LFs. I can't figure out how to get rid of them. Stripping them out in WikEd via regex is a tedious extra step. It would be nice to track them down and remove them with the script.
@
Redrose64: Here is fine. I posted it separately there because they would consider it in another context, that being in regard to bot policy. Like, how much automation is acceptable. I want to make sure I don't step on anybody's toes. The Transhumanist09:46, 26 October 2017 (UTC)
@
The Transhumanist: where you use local storage.getItem() or setItem() you should always wrap that in try catch, as it can fail at any moment (even if you checked previously). This can be due to the browser running out of storage space for the domain, or because the browser is running in privacy mode or with an ad blocker extensions or something. Also, your new RegExp() calls should be lifted outside of the for loops, so that they aren't continuously recreated. For wpTextbox1.value, realise that sometimes the content might be managed by an editor (The syntaxhighlighting beta does this for instance). We use the
jquery.textSelection plugin to abstract way from these differences. Don't check document.title, check mw.config.get( 'wgTitle' ) or mw.config.get( 'wgPageName' ). And when you use mw.util.addPortlink, you have to ensure that the mediawiki.util plugin is loaded already, which you can do by using
mw.loader.using. —
TheDJ (
talk •
contribs)
14:47, 27 October 2017 (UTC)
I will be sure to use these techniques, to the best of my ability. I've copied your post to some of my script talk pages so I won't forget. Thank you. The Transhumanist08:03, 29 October 2017 (UTC)
Firefox on XP
I am not sure where I should post it so I post it here. From my understanding, IE on XP will no longer be able to connect to Wikimedia sites soon.
My parents are using Firefox with latest patch on their XP system – old computer, so stuck there. However, Wikimedia server erroneously detects the browser as IE and refuses connection. Is there anyone can help in this? Thanks.
SYSS Mouse (
talk)
01:47, 23 October 2017 (UTC)
SYSS Mouse, your problem isn't what you think it is. You problem is in the words "old computer, so stuck there" A computer that runs Windows 10 will cost you less than $100 including shipping. See
[21].
Seriously, running XP is a terrible idea. Support for Windows XP ended April 8, 2014. Microsoft no longer provides security updates for the Windows XP operating system. Antivirus programs ignore malware that only runs on XP. More and more applications and websites simply do not run on Windows XP or the old browsers that run in XP. It is only going to get worse. --
Guy Macon (
talk)
06:49, 23 October 2017 (UTC)
I don't believe there's any attempt to actually check for Internet Explorer. If you're seeing
this message, it's detecting the browser using an obsolete cypher for the HTTPS connection and assuming that's IE on XP. The bottom of that error page (unfortunately underneath the TL;DR in several different languages) has some more details:
We have removed support for the legacy 3DES cryptographic cipher, which your browser software relies on to connect to our sites. This is usually caused by using Internet Explorer on Windows XP, but could also be caused by other ancient browsers or user agents, or could be interference from corporate or personal "Web Security" software which actually downgrades connection security.
How old the computer is? What is the actual configuration? It may be possible to update to Win 7 with some hardware upgrades.
Ruslik_
Zero20:20, 23 October 2017 (UTC)
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
You can now block users from emailing you through the Wikimedia wikis.
[22]
Changes later this week
The
new version of MediaWiki will be on test wikis and MediaWiki.org from 31 October. It will be on non-Wikipedia wikis and some Wikipedias from 1 November. It will be on all wikis from 2 November (
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
31 October 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
1 November at 16:00 (UTC). See
how to join.
{{cite conference|last1=Buaban|first1=Jesada|date=August 2016|script-title=ความทรงจำในดวงแก้ว: ความทรงจำที่แปรเปลี่ยนไปเกี่ยวกับวัดพระธรรมกายภายใต้ปริมณฑลรัฐบาลทหารปี พ.ศ. 2557–2559|trans-title=Memory in Crystal: Changing Memory on Dhammakaya Movement under the Umbrella of Military Junta 2014–2016|url=http://www.academia.edu/28222034/Memory_in_Cristal_Changing_Memory_on_Dhammakaya_Movement_under_the_Umbrella_of_Military_Junta_2014-2016|format=pdf|conference=The science of remembering and the art of forgetting 2nd conference|language=th|location=Songkhla-Nakharin University, [[Songkhla]]|publisher=Southeast Asian Studies Program, [[Walailak University]]|page=3|url-access=registration|conference-url=https://memoria2016-psu.blogspot.com/}}
With more experiments in my sandbox, it seems that the lint error happens with any {{
cite}} template with parameters |script-title=(non-ascii + ascii)|url=any|url-access=registration. It won't happen if |script-title= is all non-ascii. It won't happen if |title= is used instead of |script-title=. This is demonstrated with several examples in
User:Anomalocaris/sandbox/Lint Test. —
Anomalocaris (
talk)
08:58, 20 October 2017 (UTC)
And in case anyone is interested, I got rid of the lint error in
Global Buddhist Network by changing |script-title= to |title= and nesting the Thai-ascii title with two apostrophes before and two apostrophes after to reverse the italicization of the title. —
Anomalocaris (
talk)
09:54, 20 October 2017 (UTC)
I would suggest that such fixes are inappropriate. Yeah, 'unfixed' there is a problem but once the issue is remedied in the cs1|2 module suite, the problem goes away; but, 'fixed' cs1|2 templates in wikitext will, almost certainly, never be restored to their correct form. This error in the module suite is not fatal, there is time to remedy it properly. I think that you should undo your 'fix'.
Trappist the monk: I have reservations, but I suppose that this bug will be fixed before Wikipedia goes (to make it possible for Wikipedia to go) HTML5 compliant, so I undid my 'fix' as you suggested, and I left a note there <!--script-title generates Lint error: Misnested tag with different rendering in HTML5 and HTML4 here -->, because the lint error is localized to {{reflist}}, which makes it hard to find; it could be in any <ref>.
Anomalocaris (
talk)
06:58, 26 October 2017 (UTC)
You left in the italic wiki-markup which you should not do; |script-title= disables the cs1|2 template's automatic italics but does nothing about editor-inserted italics. I think that you should also remove the hidden comment because, just like the 'fix', the comment will remain in the wikitext for who knows how long after the next version of the cs1|2 module suite goes live.
The cause of the misnesting has been removed from
Module:Citation/CS1/sandbox. Here is the {{
cite conference}} sandbox rendering of your citation (do not use this in article space):
'"`UNIQ--templatestyles-00000033-QINU`"'<cite id="CITEREFBuaban2016" class="citation conference cs1 cs1-prop-foreign-lang-source">Buaban, Jesada (August 2016). <span class="id-lock-registration" title="Free registration required">[http://www.academia.edu/28222034/Memory_in_Cristal_Changing_Memory_on_Dhammakaya_Movement_under_the_Umbrella_of_Military_Junta_2014-2016 <bdi>ความทรงจำในดวงแก้ว: ความทรงจำที่แปรเปลี่ยนไปเกี่ยวกับวัดพระธรรมกายภายใต้ปริมณฑลรัฐบาลทหารปี พ.ศ. 2557–2559</bdi>]</span> [''Memory in Crystal: Changing Memory on Dhammakaya Movement under the Umbrella of Military Junta 2014–2016''] <span class="cs1-format">(pdf)</span>. [https://memoria2016-psu.blogspot.com/ The science of remembering and the art of forgetting 2nd conference] (in Thai). Songkhla-Nakharin University, [[Songkhla]]: Southeast Asian Studies Program, [[Walailak University]]. p. 3.</cite><span title="ctx_ver=Z39.88-2004&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Abook&rft.genre=conference&rft.btitle=%E0%B8%84%E0%B8%A7%E0%B8%B2%E0%B8%A1%E0%B8%97%E0%B8%A3%E0%B8%87%E0%B8%88%E0%B8%B3%E0%B9%83%E0%B8%99%E0%B8%94%E0%B8%A7%E0%B8%87%E0%B9%81%E0%B8%81%E0%B9%89%E0%B8%A7%3A+%E0%B8%84%E0%B8%A7%E0%B8%B2%E0%B8%A1%E0%B8%97%E0%B8%A3%E0%B8%87%E0%B8%88%E0%B8%B3%E0%B8%97%E0%B8%B5%E0%B9%88%E0%B9%81%E0%B8%9B%E0%B8%A3%E0%B9%80%E0%B8%9B%E0%B8%A5%E0%B8%B5%E0%B9%88%E0%B8%A2%E0%B8%99%E0%B9%84%E0%B8%9B%E0%B9%80%E0%B8%81%E0%B8%B5%E0%B9%88%E0%B8%A2%E0%B8%A7%E0%B8%81%E0%B8%B1%E0%B8%9A%E0%B8%A7%E0%B8%B1%E0%B8%94%E0%B8%9E%E0%B8%A3%E0%B8%B0%E0%B8%98%E0%B8%A3%E0%B8%A3%E0%B8%A1%E0%B8%81%E0%B8%B2%E0%B8%A2%E0%B8%A0%E0%B8%B2%E0%B8%A2%E0%B9%83%E0%B8%95%E0%B9%89%E0%B8%9B%E0%B8%A3%E0%B8%B4%E0%B8%A1%E0%B8%93%E0%B8%91%E0%B8%A5%E0%B8%A3%E0%B8%B1%E0%B8%90%E0%B8%9A%E0%B8%B2%E0%B8%A5%E0%B8%97%E0%B8%AB%E0%B8%B2%E0%B8%A3%E0%B8%9B%E0%B8%B5+%E0%B8%9E.%E0%B8%A8.+2557%E2%80%932559&rft.place=Songkhla-Nakharin+University%2C+Songkhla&rft.pages=3&rft.pub=Southeast+Asian+Studies+Program%2C+Walailak+University&rft.date=2016-08&rft.aulast=Buaban&rft.aufirst=Jesada&rft_id=http%3A%2F%2Fwww.academia.edu%2F28222034%2FMemory_in_Cristal_Changing_Memory_on_Dhammakaya_Movement_under_the_Umbrella_of_Military_Junta_2014-2016&rfr_id=info%3Asid%2Fen.wikipedia.org%3AWikipedia%3AVillage+pump+%28technical%29%2FArchive+160" class="Z3988"></span> <span class="cs1-visible-error citation-comment"><code class="cs1-code">{{[[Template:cite conference|cite conference]]}}</code>: </span><span class="cs1-visible-error citation-comment">Invalid <code class="cs1-code">|script-title=</code>: missing prefix ([[Help:CS1 errors#script_parameter|help]])</span>
Trappist the monk: I don't see it as a bug, I see it as an inconvenience. What the linter sees is code after much processing. That processing includes converting all of the {{
cite}} templates and moving all of the <ref>s to {{
reflist}}. There is a burden on the user. Such lint errors can be found by manually removing <ref> tags while leaving their contents in place; if necessary copying the text from a talk page edit window to an article edit window, and using the LintHint tool at
User:PerfektesChaos/js/lintHint. With these steps, it is possible to locate any lint error buried in a <ref>. —
Anomalocaris (
talk)
02:22, 31 October 2017 (UTC)
You got many helpful answers last time. Imagine going to a doctor and saying "My foot has something wrong with it." How much help would you get with just that description, especially if the doctor asked you followup questions that you did not answer?
If you want help, you need to provide information that is required for troubleshooting. As you are aware, this problem was difficult for all or most of us to reproduce. Once again, you have not provided a link to an article, you have not provided a specific OS version, and you have not provided a specific browser version. Please do that. Some of the people who helped you in the thread linked above listed specific articles and browser versions that were working just fine.
Here's what a helpful report would look like: "I have Windows 10 with all of the latest patches (as of October 27), and I am using Internet Explorer 11.608.15063.0. When I look at
Chris Janson#Extended plays, the cells in the "Details" column of the table have a bottom border that is one pixel lower than the cells to the right and left. When I look at the page in Microsoft Edge (version XXXX), Firefox (version XXXX), and Chrome (version XXXX), the cells look fine. Can someone help me diagnose this problem?"
When you are experiencing a problem that other people are not normally experiencing, it is helpful to provide enough information that people might have a fighting chance to reproduce the problem themselves. Also pinging
TenPoundHammer, who might be able to provide a good bug report for us. –
Jonesey95 (
talk)
23:04, 27 October 2017 (UTC)
@
Jonesey95: I agree that I should have been more detailed in my query, but I should note that I did open
a Phabricator report with a screenshot of the problem I was experiencing. Based on your template of a helpful report, I can say that I have Windows 10 with all of the latest patches (as of October 31), and I am using Internet Explorer 11.674.15063.0. When I look at
List of Prime Ministers of the United Kingdom#Since 1721, the cells in the "Name" column of the table have a bottom border that is one pixel lower than the cells to the right and left. When I look at the page in Microsoft Edge (version 40.15063.674.0), I am experiencing the same issue but with a different set of cells in the same section. Can someone help me diagnose this problem? I haven't checked Chrome yet, and similar to Edge the problem may be with a different set of cells. Thank-you though for the feedback, which I will certainly take into account in future.--
Nevé–selbert15:20, 31 October 2017 (UTC)
Still the same issue...
So, considering that nobody notified me of either removal or not removal of autopatrol flag, I decided to check out Google Chrome (as some of your advise here) in order to see notifications better. To my surprise, I still can't see mark this page as patrolled button! So, I don't see it in either Internet Explorer or Google Chrome (but I do see capital N when a new page being created. So, what's going on and how to solve it? Any help will be appreciated.--
Biografer (
talk)
23:39, 30 October 2017 (UTC)
I don't patrol new pages so I don't know whether anything changed. Currently I have a "Mark this page as patrolled" link on new yellow pages at
Special:NewPages when
Wikipedia:Page Curation/Help#Curation Toolbar is completely hidden (first minimized on arrow and then hidden on X). When the toolbar is visible I instead have a checkmark in the toolbar with hover text "Mark this page as reviewed". Tested in Firefox and Internet Explorer 11.
PrimeHunter (
talk)
10:32, 31 October 2017 (UTC)
@
Biografer:autopatrol means that your edits are automatically marked as patrolled. Usage of the "mark as patrolled" button is part of the
New pages reviewer toolset. N just means "New page" and has nothing to do with patrolled status. If you would like to become a patroller, read through
WP:NPP. —
xaosfluxTalk13:09, 31 October 2017 (UTC)
Two absolutely contradictory replies, but I think I will go with @
Xaosflux: on that one (considering that I wasn't aware of October 2016 change). I personally am interested in reviewing newly created sandboxes ant/or talkpages (majority of content to review I see it that way). But, because of my block two days ago, that probably will hinder me from reviewing anything for a time being... Let me know if its OK to join the reviewing team. Thanks.--
Biografer (
talk)
17:13, 31 October 2017 (UTC)
@
Biografer: I suggest starting at
Wikipedia:New pages patrol/School if you have already read the overview. Note any editor can follow and edit new pages and recent changes, but extra access is needed to "mark" the page as patrolled. The general guidelines for that access are listed at
Wikipedia:New pages patrol/Reviewers. If you feel you meet the guidelines or have a reason they should be excepted, post at
WP:PERM to request access. Please note, frivolous perm requests may delay future access requests. —
xaosfluxTalk18:02, 31 October 2017 (UTC)
Also, talkpages and sandboxes are pretty much the lowest priority things that would need patrolling (articles are the highest). —
xaosfluxTalk18:40, 31 October 2017 (UTC)
WP 1.0 bot just broke non-ASCII links
User:WP 1.0 bot maintains tables of article assessments, among other things.
Today its daily update broke the links to categories of a WikiProject whose name includes an en-dash, see
[23]
According to
MW:Help:Extension:Linter/obsolete-tag, The deprecated tag <font color=x size=y face=z> should be replaced by <span style="color: x; font-size: y; font-family: z;">. True! But there's a strange difference in the behavior of <font color=x>[[Wikilink]]</font> and <span style="color: x;>[[Wikilink]]</span>. Look at these examples:
<span style="color: #00FF00;">[[Green]]</span>: Green
<span style="color: #00FF00;">[[Green]]</span>: Green
Viewing the source, we see that the examples with <font> insert font code inside the href, viz:
<font color="green">[[Green]]</font> converts to <a href="/info/en/?search=Green" title="Green"><font color="green">Green</font></a>, but
<span style="color: green;">[[Green]]</span> converts to <span style="color: green;"><a href="/info/en/?search=Green" title="Green">Green</a></span>
Why are these methods of markup not treated identically? How should editors deal with this discrepancy when trying to de-lint articles laden with <font> tags?—
Anomalocaris (
talk)
05:03, 23 October 2017 (UTC)
If you use the proper Hex value for Green and stop trying to use Wikilinks in your examples they are identical. Also, your fourth example tries to close a <span> with a </font>.
Result with plain text:
<font color="green">Green</font>: Green
<span style="color: green;">Green</span>: Green
<font color="#008000">Green</font>: Green
<span style="color: #008000;">Green</span>: Green
But what about the wikilinks?
Result with red Wikilink:
<font color="red">This contains a [[red]] Wikilink</font>: This contains a
red Wikilink
<span style="color: red;">This contains a [[red]] Wikilink</span>: This contains a
red Wikilink
<font color="#FF0000">This contains a [[red]] Wikilink</font>: This contains a
red Wikilink
<span style="color: #FF0000;">This contains a [[red]] Wikilink</span>: This contains a
red Wikilink
Result with green Wikilink:
<font color="green">This contains a [[green]] Wikilink</font>: This contains a
green Wikilink
<span style="color: green;">This contains a [[green]] Wikilink</span>: This contains a
green Wikilink
<font color="#008000">This contains a [[green]] Wikilink</font>: This contains a
green Wikilink
<span style="color: #008000;">This contains a [[green]] Wikilink</span>: This contains a
green Wikilink
Result with blue Wikilink:
<font color="blue">This contains a [[blue]] Wikilink</font>: This contains a
blue Wikilink
<span style="color: blue;">This contains a [[blue]] Wikilink</span>: This contains a
blue Wikilink
<font color="#0000FF">This contains a [[blue]] Wikilink</font>: This contains a
blue Wikilink
<span style="color: #0000FF;">This contains a [[blue]] Wikilink</span>: This contains a
blue Wikilink
Result with uncolored Wikilink:
This contains a [[color|uncolored]] Wikilink: This contains a
uncolored Wikilink
Result with red Redlink Wikilink:
<font color="red">This contains a [[XredX]] Wikilink</font>: This contains a
XredX Wikilink
<span style="color: red;">This contains a [[XredX]] Wikilink</span>: This contains a
XredX Wikilink
<font color="#FF0000">This contains a [[XredX]] Wikilink</font>: This contains a
XredX Wikilink
<span style="color: #FF0000;">This contains a [[XredX]] Wikilink</span>: This contains a
XredX Wikilink
Result with green Redlink Wikilink:
<font color="green">This contains a [[XgreenX]] Wikilink</font>: This contains a
XgreenX Wikilink
<span style="color: green;">This contains a [[XgreenX]] Wikilink</span>: This contains a
XgreenX Wikilink
<font color="#008000">This contains a [[XgreenX]] Wikilink</font>: This contains a
XgreenX Wikilink
<span style="color: #008000;">This contains a [[XgreenX]] Wikilink</span>: This contains a
XgreenX Wikilink
Result with blue Redlink Wikilink:
<font color="blue">This contains a [[XblueX]] Wikilink</font>: This contains a
XblueX Wikilink
<span style="color: blue;">This contains a [[XblueX]] Wikilink</span>: This contains a
XblueX Wikilink
<font color="#0000FF">This contains a [[XblueX]] Wikilink</font>: This contains a
XblueX Wikilink
<span style="color: #0000FF;">This contains a [[XblueX]] Wikilink</span>: This contains a
XblueX Wikilink
Result with uncolored Redlink Wikilink:
This contains a [[XuncoloredX]] Wikilink: This contains a
XuncoloredX Wikilink
Guy Macon: I fixed the fourth example above where I closed a <span> with a </font>.
The two methods of markup, which should be identical, are identical when the modified text includes no wikilink at all, or more than a wikilink. When it consists of a wikilink, no more and no less, the two methods of markup diverge:
I suspect what you are seeing here is caused by the legacy Tidy system rewriting <font>...</font> tags to lie inside <a>...</a> (link) tags, while not doing the same for <span>...</span> tags in the same context. These sorts of anachronistic inconsistencies are among the reasons why Tidy is going to eventually be removed and replaced with a more modern solution. The Linter tool gives users an opportunity to find and fix instances of these nasty old Tidy behaviours. — This, that and the other (talk)09:06, 23 October 2017 (UTC)
A template that colors Wikilinks breaks the bluelink/redlink behavior that users expect. We shouldn't have templates that do that. --
Guy Macon (
talk)
09:41, 23 October 2017 (UTC)
If coloring of links is really necessary (for example, to achieve white link text on a dark background) you can use [[target|<span style="color: #0088FF;">text</span>]]:
text — This, that and the other (talk)10:03, 23 October 2017 (UTC)
@
This, that and the other: On another note, there are 14,350 templates which use an opening or closing <font> tag. A Lua module adding the <span>...</span> might be necessary to replicate the colouring-of-links behaviour in case the links are on blue backgrounds and wouldn't be visible without the coloration or something.
Jc86035 (
talk)
10:28, 23 October 2017 (UTC)
In addition, there are some contributors who haven't updated their signatures in years, so the signature leaves font tags everywhere. Leaving aside existing uses which can be bot-corrected, does the software allow the signature definitions to be forcibly fixed or does this have to be changed manually by each of the editors in their preferences?
Jc86035 (
talk)
10:37, 23 October 2017 (UTC)
I discovered this while fixing lint errors on talk pages.
I assume that most uses of <font> to alter Wikilink colors occur in user signatures on talk pages.
I believe that in article space, Wikilink colors should never be overridden, and any templates that do this should be modified.
I believe that this nonstandard behavior of <font color=...>[[Wikilink]]</font> should be removed.
If it is to be removed, Wikipedians should be notified, especially those whose signatures rely on this nonstandard behavior.
I don't think it's a great loss to have existing signatures on a plethora of talk pages appear in the default link colors even if the author intended a different color.
That said, if someone wants to create a tool that facilitates users having their past postings on talk pages edited to move font coloring (preferably expressed using <span style=...> markup) into the wikilink, that is OK.
Regardless of whether anything is done about this, all users using <font> in their signatures should be informed that this code is non-compliant with HTML5 and may cause problems later, and they should change their signature to <span style=...> markup.
Thanks everyone for this new Tidy discovery. We'll introduce a new high-priority linter category to track this .. will probably call this tidy-font-bug to identify pages and templates that will be affected by this Tidy inconsistency. Will have to play with a bunch of different font wikitext snippets and/or read Tidy source code to narrow this down.
SSastry (WMF) (
talk)
15:46, 23 October 2017 (UTC)
I don't think this was documented quite a decade ago, as it was noted in Phabricator on May 11, 2010, 1:45 AM. Unfortunately, nothing was done about it, and some Wikipedians depend on this bug for their signatures to have their desired font colors, while other Wikipedians stumble into it and are perplexed when <font> markup properly recoded into <span> markup behaves differently. Decisions need to be made. We need to systematically inform users with signatures that depend on the bug that they should upgrade their signature to <span> markup so that they will be HTML5 compliant, and also, if there is the intention of fixing this bug eventually, they should upgrade their signature to <span> markup so their future signatures will not change their display color when the bug is fixed. —
Anomalocaris (
talk)
08:53, 24 October 2017 (UTC)
The whole signature system was a blunder, and results in users wasting time dealing with such things. MediaWiki developers probably realized this a long time ago. For one thing, it needlessly pollutes the link tables, and can easily be used to break a whole page by leaving an unclosed tag, and finally telling the software what it already knows (who added what / where) by signing makes no sense. The system would have probably been order of magnitudes better if it used a tag extension, e.g. <sig userid=123/> or simply if custom signatures were disallowed, but that's water under the bridge. This is an easy to fix technical quirk, blocked by social issues.
A better use of developer's time is to outright ban the use of font tags (and possible other needless tags) when the newer tool is deployed. Rather than making it obsolete, render it inert by removing it from the output. That way users don't have to do anything outside content namespaces, and old signatures can remain unchanged as the relics that they are. Anyway, here's a related fun fact (
https://phabricator.wikimedia.org/T67747) that has probably existed for as long as the displaytitle has existed. Oddly enough this isn't even detected by the linter probably because it ends up being interface text:
The issue related to the links means that everything aside from links (and other possibly other elements, e.g. text within tables, borders, etc) will be hidden. Funny thing is that when the font issue is fixed, it will make it be possible to hide all content of a simple enough page, and even affect text elements of a preview.11:30, 24 October 2017 (UTC) — Preceding
unsigned comment added by
197.218.88.243 (
talk)
We now have
Special:LintErrors/tidy-font-bug to detect this. It is starting to populate and as expected, it is primarily filled with talk page entries, likely because of user signatures, but as of now, it has one article namespace entry from a railway station page.
SSastry (WMF) (
talk)
15:55, 31 October 2017 (UTC)
A discussion at
Template talk:Convert#Dots concerns the fact that {{convert}} has many units which use middot in symbols such as kW·h (convert uses the Unicode character, not the html entity). However, sdot should be used. I'm thinking of changing all convert's units to replace middot with sdot. Is there any reason that should not happen? I'm wondering about old systems which might not display sdot properly.
Johnuniq (
talk)
04:24, 1 November 2017 (UTC)
"Multiplication dot" is definitely the right thing in terms of how things should be displayed. I have no info though about issues with old systems.
Jeh (
talk)
05:18, 1 November 2017 (UTC)
Edit conflict: just paste my text below, please
Once more I ran into an Edit conflict situation, and I received the Resolution Screen Of Death: I could not recognise nor even find my own text. (Had to solve it indirectly).
Why does not ec-solver provide this basic solution: Put your text paragraphs below, and add '(ec)' marker. -
DePiep (
talk)
19:49, 22 October 2017 (UTC)
(ec)
TheDJ (So I should solve it manually, by opening this page for edit etc., but the problem really is: find my own text in that new ec-screen) late sign, because of ec's; -
DePiep (
talk)
20:43, 22 October 2017 (UTC)
Or, give a single text overview! And use three background colors: color-1: OK text (blank), color-2: the other changes text (yellow?), color-3: my text (blue?). (or use more colors for situations, e.g. samep-word conflicts; you'll know about these conflicts). Point is: one text list. HTH. -
DePiep (
talk)
20:12, 22 October 2017 (UTC)
RIGHT NOW I am about to loose my next post here because of new ec resolution screen. [I need to do what, TheDJ? Bad luck, retype my post?]. Also, why cannot software put my post right below a preceding "ec" post? -
DePiep (
talk)
20:06, 22 October 2017 (UTC)
The ec-screen opens with two columns, the right one fuzzed, illegible. Hidden text, many colors: it says I've done something wrong. But what? Then, still without being able to discern texts, I have to click choices on how I want to resolve my error (flashing colors around). Like a monkey, I want my peanut: my own text, in the page. So I push buttons, see what happens. I've never reached the final button that says: "Yes, this is what I want to sign & save".
Then I discovered this: My new text is in the lefthand side column. Never met this before. By nature: lefthand=old, righthand=new. -
DePiep (
talk)
20:38, 22 October 2017 (UTC)
@
DePiep: With the old edit conflict thing, you also got two boxes: the top one is for the page as it stands after the other person's edit, and the bottom one has your edit in it. But of course, whichever type of edit conflict interface you have, you can use the browser's "back" button to return to the pre-conflict edit box. There, you can copy your text to clipboard, back out one more level and edit again, this time pasting instead of typing. --
Redrose64 🌹 (
talk)
20:45, 22 October 2017 (UTC)
At least, that was a "Old and recent vs You"-text presentation. I could figure that out. Do you see it is the interface I am complaining about? Today I can not discern any more the crucial before-newlysaved-yournew variants. Let alone the solving steps (how to proceed). And don't forget: many ec's (talkpages!) can be resolved by the button I proposed above: ec? Add my post below!. -
DePiep (
talk)
20:58, 22 October 2017 (UTC)
Oh and also don't forget (2): I'd have lost my carefully crafted (time=ec!) texts, had I not circumvented the new ec-screen. -
DePiep (
talk)
20:58, 22 October 2017 (UTC)
BTW. There is a
simulation for edit conflicts now. The german team has been trying to do some testing with edit conflict resolution, but they are having a very hard time gathering good feedback, because people encounter the screen so infrequently (and often then focus on using it, not giving feedback about it). So they hope that with the simulation, they can gather better feedback. See also:
meta:WMDE Technical Wishes/Edit Conflicts/Feedback Round Test Page —
TheDJ (
talk •
contribs)
22:10, 22 October 2017 (UTC)
@
DePiep: I'm aware this could be considered victim-blaming to some degree, but if you don't want potentially buggy features why have you enabled a beta feature? The whole point of beta testing is that you're running potentially buggy scripts to see if they're useful and to identify bugs. ‑
Iridescent22:18, 22 October 2017 (UTC)
Good point. Didn't know I had activated this beta. Deal?: me victim behaviour, sc-interface team accept frustration. Bottom line: there is some serious non-personal sense in the experience I described. -
DePiep (
talk)
22:31, 22 October 2017 (UTC)
@
DePiep: it is really helpful to hear what was confusing for you when using the new screen (although I'm sorry that it was a frustrating experience). So thanks a lot for the detailed description you gave! About the "add my post below" suggestion: Leaving technical aspects aside, I agree that this would work for many ec cases on talk pages. But it wouldn't work for cases in the article name space. As mentioned above, the feature is in beta testing to gather feedback and find open bugs. So, if anyone else would like to support us by testing the feature and telling us what works well/what doesn't, it would be great! See the links given by
TheDJ in this section for more info on how to test the feature without having to wait for a real edit conflict. We plan to keep the test page up at least until Nov 9. As long as we have this feedback round running, we will adress bugs only, and then look into all the feedback and discuss the next steps. Thanks! --
Birgit Müller (WMDE) (
talk)
12:40, 27 October 2017 (UTC)
That's not true. navboxes are filtered out of the html served by the mobile version and can thus not be made visible with CSS. This is due to the enormous amounts of bytes this saves, as well as because it creates pages and pages of footer links. The Obama article for instance manages to dump 1560 links at 230 kB, which is a navigation nightmare. Especially, since it will require you to scroll horizontally (on for instance iOS, where scrollbars are not visible) —
TheDJ (
talk •
contribs)
21:28, 25 October 2017 (UTC)
So puzzle is to make 230 kb static and fetch them once
"Obama page" + "Obama navigation template1"
Don't beat me, but I would go with iframes :-)
It is wrong to use "transclusion" ideas and reject frames completely: we don't have a better replacement for them
See also
phab:T124168 on showing navboxes in mobile skins and
phab:T112987 on storing infoboxes and navboxes separate from articles.
There's also the age-old question of whether people who aren't editors actually use navboxes very much. Navboxes haven't been shown on mobile for years. There's been a lot of user research done in the last few years. I've never yet seen a research report that includes any comments about any reader missing navboxes when they switch to their mobile device.
Whatamidoing (WMF) (
talk)
17:49, 26 October 2017 (UTC)
@
Whatamidoing (WMF): If the WMF did a research project and determined through some means that desktop readers never or very rarely use navboxes, would the WMF remove them from Vector as well?
Jc86035 (
talk)
10:22, 27 October 2017 (UTC)
I doubt it, though there are serious issues there, they are manageable. But
this is how that looks on mobile. I consider that damaging.. As in, people who encounter this, will go away instead of engage with Wikipedia. And no one has ever shown me a design to make that a usable piece of content on a smaller form factor. I have seen a few ideas, but almost all of them require major engineering resources, and for me personally, the added value of navboxes is simply not worth it compared a whole lot of other stuff that the paid developers are working on. —
TheDJ (
talk •
contribs)
11:11, 27 October 2017 (UTC)
If you think this is important please do work on it. I'm just saying, it's harder than most people make it out to be, with many more edge cases, a level of browser support that is far higher than most people are used to and a group of users that is more critical than is pleasant. —
TheDJ (
talk •
contribs)
16:40, 1 November 2017 (UTC)
What happened to my watchlist, and how do I restore it?
On my main machine, which is Windows 10, I get the same result on Chrome, IE, Edge, and Firefox. I see the same thing on other Windows 10 and Windows 7 machines (tested with Chrome only).
I have not enabled the new "filters for edit review".
I dislike the new scheme because the diff and hist links are in variable places depending on how long the pagename is. I do like having the edit time in a column that's always in the same place, but since one never clicks on that, there is little UI benefit to it.
Jeh (
talk)
19:32, 31 October 2017 (UTC)
Did you disable that for me? 'cause it doesn't seem to be enabled now. And my WL still looks funny.
Jeh (
talk)
00:29, 1 November 2017 (UTC)
@
Jdforrester (WMF): I tried enabling it, Save – the WL is indeed now "enhanced". Then disabled it, Save – the "enhancements" went away but the format of the line items is still as in my first ("bad") linked picture above.
btw I tried looking at it on a couple of browsers/machines from which I have NEVER accessed WP before, so I don't think it's a client-side cacheing issue.
Jeh (
talk)
00:35, 1 November 2017 (UTC)
@
Jeh: Whoops, sorry, yes, the other change-how-watchlist-results-look preference, as
PrimeHunter spotted. We really need to clean up the set of all these preferences so that it's more obvious which does what! That should fix it.
Jdforrester (WMF) (
talk)
01:08, 1 November 2017 (UTC)
Yes, that did it, thank you! (It's not at all obvious to me, not just that those options imply a different watchlist display, but why the different display is preferable for either of those options.. but that's not important. FWIW if it were me I would move the time display in the "regular" watchlist to the left of the editor's name so it lines up in a neat column – that aspect of the "strange" WL display, I liked.)
Jeh (
talk)
05:12, 1 November 2017 (UTC)
Though I agree that there are positives about each of the layout options, I think changing any of them to take the 'good' things from the others would be pretty disruptive to users. Because we don't want to get in the way of the brilliant volunteers dedicating their time to fixing articles, mentoring new users they find in Recent Changes, and improving things on their watchlists, we'd not want to make changes to people's experiences of such a core workflow. Consequently, any improvements in this area would mean adding a fourth layout option, which means a slower site (more code), more bugs (more things to go wrong), and more confusing site (more options to pick between for users). Right now we have no plans to do this – and if we were, it'd be a quite a large scale process to ensure it was as significant an improvement as possible, which would mean less time spent increasing the performance, fixing bugs, and all the other stuff to which we're committed.
Jdforrester (WMF) (
talk)
17:24, 1 November 2017 (UTC)
Converting en dashes to hyphens in date ranges without affecting the source text
There is
a long-standing discussion at
Template talk:Cite DNB over whether it would be possible for hyphens to appear as en dashes without affecting the source. To quote my edit request of 18 October:
Per
wikisource:Page:Dictionary of National Biography volume 01.djvu/19, the original source
uses the en dash and not a hyphen for date ranges, i.e. ABBOT, GEORGE (1562–1633) not ABBOT, GEORGE (1562–1633). Please can editor modify this template so that hyphens that appear before and/or after an integer appear automatically as en dashes, in the same way hyphens are automatically converted at {{
cite book}} with |pages=, i.e. {{cite book |title=Foo |pages=1-2}} appearing as Foo. pp. 1–2..--
Nevé–selbert 19:41, 18 October 2017 (UTC)
Now I'm not exactly sure over the practicality of this proposal, but if someone has an idea how to work this, I'd very much like to hear it.--
Nevé–selbert15:35, 31 October 2017 (UTC)
The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
And why does the target site matter more than the actual source itself and its unedited transcription? Your opposition makes absolutely no sense. Only Wikisource doesn't bother with the ndash. Why repeat their errors here? These are date ranges that are part of the title, yet the fact that the original source uses the ndash doesn't matter? Bizarre.--
Nevé–selbert19:02, 1 November 2017 (UTC)
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
Language links when topics don't map one-to-one
I'm really confused about how to add language links on wikipedia when two wikis have slightly different topic divisions.
Today, as I sometimes do, I was trying to use the language links to figure out how to say something in another language (
Civet, specifically), but the link was not there so I searched up the english word on the other wiki and found that there the article for the animal and its Latin family were the same. I went to wikidata to try to link "ジャコウネコ科" to "Civet" but it told me that the link for "ジャコウネコ科" was already occupied at "Viverridae" so I needed to "merge the items" and now I'm so confused on what I'm supposed to do....
SougonNaTakumi (
talk)
06:26, 1 November 2017 (UTC)
I would love to learn otherwise, but unfortunately I think the Wikidata scheme is simply not set up to handle anything other than a 1-1 mapping between languages.
My canonical example for this is
retinol versus
vitamin A. In English Wikipedia, we have separate articles for retinol, the specific molecule, and vitamin A, which includes nutritionally equivalent compounds. When I worked on this many years ago, there were lots of languages that had only one article. Sometimes it had a title cognate to retinol, and sometimes to vitamin A.
It seems to me that both choices are defensible, and do not need to be uniformized between languages. But then the single article in the languages with just one ought to have two English interlang links appear, and both en.wiki articles should have an interlang link to the single article in the other language.
Now, I just spot-checked it, and in the cases I checked, the other languages have the same division as in English. I only checked a few, so I don't know whether it's the case for all of them. But I see no reason it should have to be, just to make the interlang links tidy. Our software ought to support many-one (or possibly even many-many) links between languages, when the semantic ranges of the existing articles don't exactly line up. But if it does support them, I don't know how to do it. --
Trovatore (
talk)
07:03, 1 November 2017 (UTC)
Vitamin A is the general topic,
retinol is a subtopic; if a specific wiki doesn't believe that the latter deserves its own article (due to lack of information on that wiki, perhaps they're strongly against permanent stubs), then merging the topic into the vitamin article would be appropriate. And the problem with Civet is that, as far as I can tell, it's not an official taxon – it's a group of genera from a single family, but not more closely related to each other than to other genera in the same family.
עוד מישהוOd Mishehu07:37, 1 November 2017 (UTC)
Well, retinol is not purely a subtopic of vitamin A; there's presumably stuff to say about it also from a purely chemical viewpoint, though I suppose that's never going to be of as much interest as the biological function. But maybe more to the point, I think the "letter" assignments for vitamins are not as much used in some other languages as they are in English, so some languages naturally saw fit to have the general article at (the cognate of) "retinol". I suppose you could put a link between :en:Vitamin_A and :otherlang:Retinol, and have no link from :en:Retinol to otherlang.wiki, but it would be counterintuitive and likely a constant battleground. --
Trovatore (
talk)
07:54, 1 November 2017 (UTC)
(That's all I know about this.) ... I guess you'll have to look at the bottom of the articles in that category, either in wikitext editing mode or in visual editing mode, to find the Wikidata extra item number, and then look up that item in Wikidata and follow links from there, to figure it out. --
Pipetricker (
talk)
10:48, 1 November 2017 (UTC)
1.The Japanese Wikipedia article "ジャコウネコ科" is specifically about the scientific family, and the article name translated black in English is also "Family of (species name in Japanese)". Information for Civet have been mentioned in the article but only as form of "Some species in the family are being called as Civet because of their ....", so I don't think it is apporipate for it to be added into the wikidata item for civet, although linking it via "interwiki extra" would be useful and helpful. 2. There are a few possible translation for Civet in Japanese, includes: ジャコウネコ(Lit. meaning musk cat, same as the name of the family name, but the name is also used to specifically referred to some species in the family too) シベット (phonetic translation of "civet") 麝香猫(Kanji expression of the concept of musk cat). 3. For general wikipedia user, in situation like this, they can probably look at the "wikidata item" link on the left sidebar and see if there are any relevant description or translation of the entry name in language they are looking for. However in this particular case it doesn't seems to help. Also, I think maybe that wikidata entry link can be moved down to the "other languages" section in the left sidebar instead of staying at where it currently is to increase visibility? 4. I do see there are instances where multiple wikilinks to same wiki or/and same item in multiple wikidata entry being useful. For instance, the word
Kamikaze is generally being used to describe all the suicidial attacks performed by Japanese military during WWII but in reality they are just one of the many
Japanese Special Attack Units . The two English Wikipedia article have more or less seperated them accordingly, but in many other languages' wikipedia version like Spanish/German/Arabic wikipedia, there are only one article for both concepts. I think it would be helpful if visitors from all those other wikipedia are presented with a choice when they are clicking interwiki link to English/Japanese/Russian wikipedia to see which of the two concepts' article were they looking for.
C933103 (
talk)
20:10, 1 November 2017 (UTC)
When we take artificial groupings of animals, there frequently won't be a good translation. For example, the word "Kangaroo" comes from the name of a specific kangaroo species in
Guugu Yimithirr, a language spoken in Australia; in this language, there is no word for Kangaroo. When taking any official taxon, we can at least take the taxon name and use it, translate it, or transliterate it; for non-taxon groupings, we can't even do this.
עוד מישהוOd Mishehu08:11, 2 November 2017 (UTC)
Percentage bar
Hi, I wondered if there was a way to programme a # tag list of numbers of articles in a list to automatically update the percentage bars as the
Wikipedia:WikiProject Women in Red/The World Contest proceeds see the bottom. What I would need is somebody to write a script which a] Reads the number of articles currently on the list and calculates the percentage of 2000 articles as each entry is added. b] Is able to detect when the day ends UTC + 0 time and able to read what the article count was each day and calculate each day what percentage of 66 articles have been created? At least something which will do a] automatically anyway. Anybody? ♦
Dr. Blofeld15:41, 2 November 2017 (UTC)
Wikidata Football results templates
In the course of
an acrimonious discussion the templates
Template:2017 J1 League table with WD and
Template:2017–18 Premier League table with WD were sent to
TfD on grounds they they were being used to import spam from Wikidata (at, for example,
2017 J1 League). I do not start to understand the techniques used but I think that changes on Wikidata were directly affecting the data emanating from these templates. Is that roughly correct? At the TFD discussions it has been remarked that these templates are similar to
Template:2017 J1 League table and
Template:2017–18 Premier League table (which are not up for deletion) but I can't follow what is happening with all these. Do they update immediately also or is some update operation required either to the templates or to the articles transcluding (substing?) them? To see relevant wikidata changes (with "Show Wikidata edits in your watchlist" in my preferences) would I need the templates or the articles watchlisted or would neither of these suffice? I'd like to get some sort of grasp of what is going on to form an opinion as to whether this is a good thing or not.
Thincat (
talk)
13:27, 2 November 2017 (UTC)
@
Thincat: The "with WD"-ones require people/bots to
update the count in the Wikidata information and the other 2/older ones require people/bots to
update the information in the template. The 'with WD'-versions will be easier for bots to keep up to date, the older template one requires more human labour. There is little inherent difference between the two templates informational content and their usage, other than Wikipedia edit vs Wikidata edit. To see changes to either you would need to have the templates (and ALL subtemplates used by them) on your watchlist, just watching an article won't show the changes for either of the variants of the templates. For the "with WD" variants, you would also need to have "Show Wikidata edits in your watchlist" enabled. But relatively speaking information like that isn't watched much to begin with. All of them are only watched by 2-5 people. All of them use subtemplates with even fewer watchers (like {{Fb team FC Tokyo}} with 0 watchers). The primary concern of the deletion discussion seems to be the source of the information, which is suspected to be an attempt at spam and self promotion. I personally think that could also be an honest mistake resulting from not understanding wikipedia and the concept of sourcing. —
TheDJ (
talk •
contribs)
14:17, 2 November 2017 (UTC)
Good grief. Thank you. It's fortunate I'm not interested in football results and I hope nobody else is either. Perhaps someone could warn me when the travels of 19th-century explorers are included on Wikidata.
Thincat (
talk)
16:54, 2 November 2017 (UTC)
Proper article, but it's all infobox!
Sorry if this is the wrong place to post this, but can anyone fix
this article, or direct me to the appropriate forum? I've had a play with the infobox and the brackets, but I can't seem to find a solution. Thanks. LugnutsFire Walk with Me19:39, 1 November 2017 (UTC)
Something has changed and I can't figure it out. I should be able to create a page in some xx.wikipedia, then click on "Edit links," add the new page in that language, then see the link in all the other Wikipedias. However, it's not happening for me now.
This new version of the source editor is horrible and I cannot deal with it! It's super slow and I'm unable to copy/paste anything while editing on an iPad/iPhone. Is there anyway to rollback to the previous version?--◂
épinetalk♬10:46, 3 November 2017 (UTC)
I cant make a account because of this error There seems to be a problem with your login session; this action has been canceled as a precaution against session hijacking. Go back to the previous page, reload that page and then try again. Its on the latest version of Google Chrome on Windows8 — Preceding
unsigned comment added by
71.169.165.176 (
talk)
10:13, 2 November 2017 (UTC)
i tried but it also happened i wonder if its the browser as we use the same browser and operating system on the computers 1 is a desktop and another is a laptop — Preceding
unsigned comment added by
71.169.165.176 (
talk)
22:25, 2 November 2017 (UTC)
@
A Den Jentyl Ettien Avel Dysklyver: You'd need an RfC; the other projects for which it's been added all have had RfCs to enable it. I assume it will be relatively inconsequential and will pass easily, but you, or the proposer (if not you), should explain that nothing changes for the people who don't like those sorts of newfangled things.
Jc86035 (
talk)
15:38, 18 October 2017 (UTC)
RfC: Enabling the Timeless skin
The following discussion is an archived record of a
request for comment. Please do not modify it. No further edits should be made to this discussion.A summary of the conclusions reached follows.
It is proposed that the Timeless skin available at
mw:Skin:Timeless is added to the en-wikipedia configuration as a optional skin which could be enabled in user preferences. There are currently 5 skins available in preferences. This will be the sixth.
The current default skin that controls how most people see Wikipedia is Vector; this will not change.
Unless an individual user wants to use the Timeless skin and enables it in their preferences, it will not be visible or change the user interface in any way. This skin has been tested and is currently enabled as a non-default preference on other Wikipedia including the French language Wikipedia, Wiktionary (fr), Wikinews (fr) and on the MediaWiki wiki.
Please don't oppose this just because you don't like it; you never have to know it is there, it will not affect any other skins or appearance settings. It is a completely optional choice, new users will still use Vector, and everyone who doesn't change their settings will still use Vector (or their other skin choice such as Monobook).
>
Dysklyver15:55, 18 October 2017 (UTC)
I don't see any reason not to support this implementation, but given my limited knowledge of anything
MediaWiki-related, I'll keep an eye here and consider any opposition duly. — fourthords |
=Λ= |16:48, 18 October 2017 (UTC)
Support – it's a nice look, better than most existing options, so as long as it won't affect anyone who doesn't decide to use it, why not enable it here.
Ravensfire (
talk)
16:53, 18 October 2017 (UTC)
Sure why not? also I don't see why this is on CENT. Its about enabling an optional skin, not really something that needs sitewide advertisement. If no one objects, I'd suggest removing it.
TonyBallioni (
talk)
23:11, 18 October 2017 (UTC)
Go for it. Adding it as an optional opt-in preference setting is low-impact for the most part, so it can probably be done right away without waiting the typical RFC discussion period.
Mz7 (
talk)
09:20, 19 October 2017 (UTC)
Support This is a modern UI skin which would likely get more users to interact with Wikipedia + it is done through the preference so it should have little impact as the default UI won’t be changing yet.
Paladox (
talk)
01:36, 29 October 2017 (UTC)
Comment: We don't usually allow RFCs on VPT, but this seems like the right forum to me. Given that it's on CENT, we might want to consider a short RFC period (like two or three days) if this looks like SNOW in favor of support. –
Jonesey95 (
talk)
03:29, 19 October 2017 (UTC)
It seems that all users who have commented here are in favour. 0 are opposing.should we close this as support for enabling this skin through the preference? (It’s been more then a week)
Paladox (
talk)
01:35, 29 October 2017 (UTC)
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
It's the top Google result for me on Pe'er Tasi. The article is 94 days old so indexing has only been allowed for 4 days per
Wikipedia:Controlling search engine indexing#Indexing of articles ("mainspace"). We don't control how quicly Google starts indexing after it becomes allowed. Maybe it happened after your post. Their current cache was made after your post because it says: "It is a snapshot of the page as it appeared on 4 Nov 2017 11:09:30 GMT".
PrimeHunter (
talk)
13:31, 4 November 2017 (UTC)
ehm, compare how often? For one time, it's probably best to just compare manually, more regularly.. well that's exactly why people built wikidata, because parsing wikicode is hard. I guess a relatively easy way to compare them, is to check for the presence of the link
using the api, but it's gonna be fragile. —
TheDJ (
talk •
contribs)
14:17, 3 November 2017 (UTC)
I would like to do the comparison more than 1 time. I would like to know the pages available in the manual list and not available in wikidata generated list to add them to wikidata list. Doing it manually is not something I am looking for. -- Pankaj Jain Capankajsmilyo(
talk·contribs·count)15:39, 3 November 2017 (UTC)
I don't see any sane way of doing this. It's just too difficult to try to programatically exclude all the links in
List of Jains that you don't want to compare with
User:Capankajsmilyo/List of Jains. If you were to create a sanitized version of
List of Jains in your own user space that only contains the links you're interested in, we can talk. Of course, this does mean that you would have to create a new sanitized list every time you wanted to compare the two pages, so there's still a significant maintenance burden even if someone were to write a script for this.
Nirmos (
talk)
08:09, 4 November 2017 (UTC)
There's a list comparer built into
WP:AWB (I'm not sure if you have to be AWB-enabled to get to it; I don't think so). If you use the Links on page option for both lists, it seems to work reasonably. -
Jarry1250Vacationneeded13:15, 4 November 2017 (UTC)
You can copy and paste these tables fro Firefox to Excel and then use functions there. However I prefer Unix text commands, then comm -3 file1 file2 will show up the lines that only appear in one list or the other. However note the lists have to be sorted and in the same format for this to work.
Graeme Bartlett (
talk)
23:00, 4 November 2017 (UTC)
Hi all. I would like to request help to fix an error that I (now) see a lot when using
User:Kephir/gadgets/rater (
js and
css): "Cannot read property "taskforces" of undefined" For example (See screenshot), on the article
Tsaone Macheng but note that the error is present on many pages not just this. Also, I suspect the error is related to {{WPBannerMeta}} or {{WikiProject banner shell}} since nothing changed in the script's code recently. I would really appreciate it if someone could fix this tool for me. Thanks in advance. --
Meno25 (
talk)
15:41, 4 November 2017 (UTC)
Apparently timing out edits followed with SSL connection error
This occurred to me before occasionaly when attempting to fix broken signatures in the past, but this is another case. Interestingly, I can usually edit without problems, but whenever I encounter this bug if I try the same edit again the same issue occurs. If I try to make an edit at
Batak, adding {{clear}} at the end of the See also section and removing the {{reflist}} parameter, this now occurs. This also occurs if I try "preview" or "show changes". The HTTP form is sent, a delay occurs waiting for a the server response, then I get a generic "Secure connection failed: The connection to en.wikipedia.org was interrupted while the page was loading." error. It appears like if the server thread or process servicing the request crashes or dies. Thanks, —
PaleoNeonate –
23:35, 3 November 2017 (UTC)
Would it be possible to authorize the bot that adds links to versions archived at archive.org to all those references, before the sites finally go dark?
Geo Swan (
talk)
17:25, 5 November 2017 (UTC)
I obviously can't provide evidence for this, but two days ago both
User:Home Lander and I could only find one contribution by
MissesX (
talk·contribs) which led to us asking her for her old account name as told us to look at her other edits. Hopefully this is just a glitch that won't repeat itself, but I thought I'd mention it here in case others have seen it.
Doug Wellertalk17:01, 5 November 2017 (UTC)
From what I recall, when I first looked at the user's contribs, I saw only
this one and left her a message. After receiving a message from
Doug Weller on my talk page (think it was the next day), I went back to the contribs and saw all of the previous ones, and have been able to see all of them since.
Home Lander (
talk)
18:22, 5 November 2017 (UTC)
Can someone help me out? Please access
on this diff, then click on next edit and tell me what you see? I get taken to a Twitter status post window (See
here) with the following text pasted into the field:
@Bernstein SODOMIZED BY THE #GNAA
For every Tweet, @Wikipedia will donate $1 to @HIVGov #WikiCares
Miiiight have had something to do with the EFF's Privacy Badger extension. Disabled it and it seemed to go away, but again, oddly, it went away on another browser where I had not disabled it.
Cyphoidbomb (
talk)
14:46, 24 October 2017 (UTC)
Um, is the original problem still there? When I access "on this diff" I get an innocuous diff, and when I click on next edit I get another innocuous diff. —
Anomalocaris (
talk)
20:26, 5 November 2017 (UTC)
Instead of blocking webhost and colocation IP addresses, what about treating them like dynamic IP ranges, and creating templates for the big colocation companies (e.g. {{OVH}} and {{DigitalOcean}}) a la {{SingNet}} and {{HughesNet}}? Here's a mock-up of the OVH template:
Welcome!
Interested in becoming a regular contributor to Wikipedia? Create an account!
This
IP address, Village pump (technical)/Archive 160, is part of a pool of addresses registered to OVH France, a cloud computing service owned by
OVH Inc. and is shared by multiple users. Comments left on this page may be received by other users of this IP and appear to be irrelevant.
If you are an unregistered user operating from this address, note that this is not the IP address of your device. Village pump (technical)/Archive 160 is the IP address of a
colocation server that communicates between your browser and the
Wikimedia Foundation servers. This and other servers are shared among an large number of users.
In the event of
vandalism from this IP address, abuse reports can be sent to its network administrator for investigation.
Review contributions carefully if blocking this IP address or reverting its contributions. Applying an indefinite block to this IP is not appropriate, because the IP address may later be reallocated to a different, unrelated, and unsuspecting user. The best approach is to determine the CIDR IP numerical address range and make a higher-level decision to
soft-block the entire range. Caution is advised when applying a range block as other Wikipedians, both registered and otherwise, may be affected.
Network administrators, to monitor this IP address for vandalism, can
subscribe to a
web feed of this page in either RSS or Atom format.
I noticed that noone answered yet. Do we really systematically block those? I know that some large OVH blocks are so problematic that various corporations outright block them (I'm not sure about Wikipedia). In any case, why not a single more generic template similar to {{Shared IP}} with a parameter to specify the hosting provider, like that template's |organization= parameter? A possible advantage of provider-specific templates may more easily permit to build statistics, though. The tagging system is also imperfect: We have no bots (that I know) designed to systematically tag talk pages and restore them when deleted (which would also imply an easy to maintain and machine-readable database of ranges and type/descriptions)... —
PaleoNeonate –
20:44, 5 November 2017 (UTC)
Maybe this is better suited for
WT:UTM? I'm not sure what is meant by "instead of blocking" the colos. As a CU I can confirm they're a real pain and I'll typically hardblock them on sight (which is very frequent). I'd agree there are far too many providers to have a template for each. The only place templates are useful is in block messages, and for that we use the generic {{Colocationwebhost}}. --
zzuuzz(talk)20:56, 5 November 2017 (UTC)
Upright parameter is not currently scaling up images AFAICT
Hi. The [[upright=1.16]] parameter is not scaling up. See the Charles Manson article, and this image. It is stuck at the native image size. I would like to scale it up to be the same width as the others. This is not the first time I have stumbled on this. Ping me back. Cheers! {{u|
Checkingfax}} {
Talk}10:59, 4 November 2017 (UTC)
if I recall, |upright= won't stretch an image beyond it's size. in this case the file is 189 pixels wide, so you can't go beyond that size with upright.
Frietjes (
talk)
15:01, 4 November 2017 (UTC)
(ec) re Frietjes: That's two explanations. I tried some settings (with/out thumb, frameless, "none"): did not give an answer. My old problem understanding
WP:EIS: interacting parameters. While, there must be an logical explanation. -
DePiep (
talk)
20:42, 4 November 2017 (UTC)
Hi,
Frietjes. I get it. Thanks. It would be nice if we could scale up on the fly. It would also be nice if 'upright' did not scale down by 25% by default. No sense to that one. Cheers! {{u|
Checkingfax}} {
Talk}20:57, 4 November 2017 (UTC)
Actually, that was its only original intent. To downsize portrait oriented (aka upright) images, to not be uneasily high at it's default width, compared to landscape images. The behavior we use it for is an accident of history. —
TheDJ (
talk •
contribs)
21:57, 4 November 2017 (UTC)
Conclusion please. Is it: "upright=1.x" (>1) has no effect when source size is <220px" [*or users agent setting]? -
DePiep (
talk)
22:20, 5 November 2017 (UTC)
Template:UK National Archives ID and Wikidata
On
George V, the template {{UK National Archives ID}} used in the external links section generates a bad url. This seems to be because the template pulls
Wikidata:Property:P3029, which for
George V contains two UK National Archives IDs. The broken url is generated with one of the IDs and a comma.
What's the best way to address this? Should the Wikipedia article use two National Archives templates, specifying the two National Archives IDs rather than depending on Wikidata? Should the template be generalized to handle multiple National Archives IDs returned by Wikidata? --
Worldbruce (
talk)
19:20, 5 November 2017 (UTC)
Override the Wikidata call by specifying the ID. All templates that call Wikidata allow for (or should allow for) local override of Wikidata calls. –
Jonesey95 (
talk)
19:54, 5 November 2017 (UTC)
Individual cases can be handled like that but a more general solution in the template would be better.
wikidata:Q269412#P3029 has two values F41939 and F257261. {{UK National Archives ID}} only works for a single value. It says {{#property:P3029}} which on
George V produces comma-separated values F41939, F257261, resulting in the broken link [https://discovery.nationalarchives.gov.uk/details/c/F41939, F257261 "Archival material relating to George V"]. The correct links are
https://discovery.nationalarchives.gov.uk/details/c/F41939 and
https://discovery.nationalarchives.gov.uk/details/c/F257261. Coding the template to produce two or more links is messy. How should the result be formatted? Maybe it should just trim down to the first value and only link that. In this example the first link has a lot more content but I don't know whether that is common.
PrimeHunter (
talk)
20:50, 5 November 2017 (UTC)
Use {{#statements:<P-number>}} instead. It is a newer parser function (not "new" though--it's been a thing for a year and a half or so) that generates the data you want. It should work for most properties, but I do not know how it treats qualifiers off hand.
Reference this version of "George V". --
Izno (
talk)
22:35, 5 November 2017 (UTC)
A question was asked back in July 2017 on the talk page
Template talk:Tree list#First branch for second level?. As no one has answered that question, I thought I would ask the same question here in the hope of finding an answer.
I am in the process of building a large
Ahnentafel genealogical numbered tree using {{
Tree list}} while I can do it if all the leaves of the tree hang below the subject:
Hello @
PBS: I also have similar question (I'm struggling with
similar kind of module). Although I don't know the answer, at least I suppose this is not the template problem, but CSS related. I think messaging to
MediaWiki talk:Common.css would be better. Because "treeview" style in that CSS creating the tree view. I'm paraphrasing PBS's question as follows for other new readers. --
Was a bee (
talk)
13:24, 2 November 2017 (UTC)
Question
Is it possible to create following kind of tree with CSS?
Branchs are growing not only toward bottom-right, but also toward up-right direction.
We have a bunch of those tournaments templates which, while not coded great, could plausibly be mirrored image'd around the vertical axis. Those also probably need a redesign given how painful they are to deal with. --
Izno (
talk)
13:48, 2 November 2017 (UTC)
32. Prince Ernest I, Duke of Saxe-Coburg and Gotha
64. Prince Francis, Duke of Saxe-Coburg-Saalfeld
128. Ernest Frederick, Duke of Saxe-Coburg-Saalfeld
256. Francis Josias, Duke of Saxe-Coburg-Saalfeld
(9 generation)
Although this {{clade}} does wrap the names I think the example below makes the point that trees with more than six generations will start to cause formatting problems on small screens and printers:
Ancestry of Elizabeth II
1. Elizabeth II of the United Kingdom
2. George VI of the United Kingdom
4. George V of the United Kingdom
8. Edward VII of the United Kingdom
16. Prince Albert of Saxe‑Coburg and Gotha
32. Prince Ernest I, Duke of Saxe‑Coburg and Gotha
64. Prince Francis, Duke of Saxe‑Coburg‑Saalfeld
128. Ernest Frederick, Duke of Saxe‑Coburg‑Saalfeld
256. Francis Josias, Duke of Saxe‑Coburg‑Saalfeld
257.
129.
65.
33.
17.
9.
5.
3.
With the tree layout it only comes number of generations (in this case 9) + the length of the last entry (49) making less than 60 characters wide. Something that will fit on most screens and all printers without changing the font.
1. Elizabeth II of the United Kingdom
2. George VI of the United Kingdom
4. George V of the United Kingdom
8. Edward VII of the United Kingdom
16. Prince Albert of Saxe-Coburg and Gotha
32 Prince Ernest I, Duke of Saxe-Coburg and Gotha
64. Prince Francis, Duke of Saxe-Coburg-Saalfeld
128. Ernest Frederick, Duke of Saxe-Coburg-Saalfeld
Hello. Before an hour there was a file in a web site. Now they have removed it. Is there a way to find it? The file was only in the web for some hours.
Xaris333 (
talk)
13:09, 6 November 2017 (UTC)
There is probably no way to get the file now but we could look for a cached copy at other websites like Google and the Internet Archive if you say what you know about the file and where it was, the former url if you know it. If you viewed the file then there may also be a copy in your own browser cache on your computer. In that case, do not visit the site again in the same browser. Preferably, don't use the browser at all until you get instructions from people knowing the browser. This is better suited for
Wikipedia:Reference desk/Computing.
PrimeHunter (
talk)
13:34, 6 November 2017 (UTC)
Editing box inserting bogus hard linebreaks (MacOS X, Chrome, about a month now, only on this site)
For the last month or so I've been experiencing a problem (in Chrome, on MacOS X), where if I enter a space character at the end of a displayed "line" (i.e. right before a soft-wrap) in the editing window, a hard linebreak ends up being inserted when I save, often resulting in crap like this:
Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore
magna
aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat.
where the second line, which should not be a real line at all, is treated as <pre>...</pre> text. What's happened is a hard line break was inserted after "magna", causing the "[space]aliqua" to start a new hardcoded line. This most often happens when the displayed "line" in the editing box soft-wraps at something not alphabetical, such as ]] or the like. Driving me nuts. Generally the only way to fix it is to save the broken version and go back in and re-edit it. I get hit by this around 1–5 times per day. Never happened before, in 12 years.
Yes, I've also run into it being done by other editors with increased frequency recently, though it's not all over the place. —
SMcCandlish☏¢ >ʌⱷ҅ᴥⱷʌ< 23:55, 4 October 2017 (UTC)
Would
youtwo please turn off the gadget, and then try to trigger this? Since you have the gadget in common, it'd be helpful if we could quickly rule that out, or know that the gadget needs an update.
Whatamidoing (WMF) (
talk)
17:35, 5 October 2017 (UTC)
I've turned off the beta gadget. As a test I turned off the regular one, and left the beta one on, and did not actually get any syntax highlighting at all, so the beta one seems to have issues anyway. I've run with the regular one for ages (before noticing there was a beta replacement) without this issue, so I'll try this change first. —
SMcCandlish☏¢ >ʌⱷ҅ᴥⱷʌ< 19:52, 5 October 2017 (UTC)
It's still happening. So, I have no turned off the non-beta syntax highlighter. This is going to be really painful. —
SMcCandlish☏¢ >ʌⱷ҅ᴥⱷʌ< 00:15, 7 October 2017 (UTC)
@
Whatamidoing (WMF): Update – I've run for a week or so without syntax highlighting (either version), as requested, and it's still happening, under the same conditions as initially reported. And additional one (or a variant, really) is that it often happens when double-spacing after a . (i.e., another case of a space after a non-alphanum character). I find that re-sizing the browser window can make it go away, and sometimes doing "Show preview" will as well. —
SMcCandlish☏¢ >ʌⱷ҅ᴥⱷʌ< 03:50, 18 October 2017 (UTC)
Thank
you for the update. Which editing environment are you using? There are screenshots of each at
mw:Editor if you're not absolutely certain.
Hairy Dude and
Cabayi: same question for you. A complete list of operating system/web browser/skin/editor would be the most useful (e.g., this comment is being added via macOS 10.12/Safari 11/Vector/2017 wikitext editor).
Whatamidoing (WMF) (
talk)
16:47, 18 October 2017 (UTC)
I am working right now in the wikitext editor, in Chome on a mac, with the browser window at about six inches wide. If I now add a link with a window that is about a foot wide. If I now now add something something long
@
Jytdog: The formatting observed above occurs because there's a carriage return after the word "long" and the next line starts with a space. If you remove the space and/or the carriage return, it formats normally.
DH85868993 (
talk)
21:20, 9 October 2017 (UTC)
I am not a moron and there is no carriage break. Something changed in wikitext editor in the last month or so. What is there, is two spaces. This happens with just one space as well, just before some sort of bracketed text like this. I end up having to change the width of the browser window to move the juxtaposition of the long thing with what is just before it, to be in the middle of the window instead of at the end, and then I can resolve it. But something broke in the wikitext editor to make this something i now have to stop and juggle to fix instead of just working like it should. And yes i fucking know how to fix it. The problem is that it happens at all. The wikitext editor is inserting some kind of carriage break itself. I noticed there have been other problems with wrapping on this page. I have not seen this described here.
Jytdog (
talk)
21:23, 9 October 2017 (UTC)
I am experiencing this as well and had opened a thread below about this,
here. I am using chrome on a mac and this error is completely about wrapping and is dependent on the width of the browser window. I have no gadgets enabled, just the plain wikitext editors.
Jytdog (
talk)
21:38, 9 October 2017 (UTC)
btw, what you see there is an example of this. I had first saved the comment above in
this diff, and i edited it a bit further in
this diff, and when I saved that, you can see that my signature didn't wrap correctly and a hard stop was added by the wikitext editor. this never used to happen.
Jytdog (
talk)
21:42, 9 October 2017 (UTC)
I've reported a ticket about this now, but i'm still not entirely sure what combination of settings is triggering this, so if someone could clarify that, then maybe someone can fix it. —
TheDJ (
talk •
contribs)
22:41, 9 October 2017 (UTC)
btw i just logged in with firebox (most recent version for mac) and i cannot make this bug replicate this in firebox.
Jytdog (
talk)
03:45, 10 October 2017 (UTC)
Something I have noticed (not new, it's always been that way) is when I want to move a sentence into a new paragraph. I position the cursor in front of it and press the space bar until the sentence starts at the beginning of the next line in the edit window. Then when I move the cursor back to the end of the new first paragraph and press "return" to format the second paragraph the start of the second paragraph shoots into the middle of the line and it renders in <pre> </pre> mode. A carriage return would surely just add a single space (and I don't see why it would add any space anyway). Repositioning by use of the delete key and/or space bar does not return the rendered text to normal. Is this another manifestation of the same issue?
80.5.88.70 (
talk)
14:22, 6 November 2017 (UTC)
Is Google translate incompatible with wikitext?
Here's the original Portuguese:
Eu edito o suficiente para me sentir no direito de questionar, o que não significa, vejam bem, que isso seja um ataque a quem concorda ou aos editores que se empenharam em fazer tais alterações (Joalpe). Também queria deixar registrado meu repúdio a qualquer tentativa de substituir as infocaixas locais por infocaixas da wikidata. Todos os exemplos de usos dessas infocaixas automáticas foram em artigos pequenos onde não mais que meia dúzia de elementos são alimentados. Me pergunto o quão prático seria ficar alimentando infocaixas como aquela que está em El Greco ou mesmo algumas ainda maiores. E quando essa informação precisar ser atualizada, o que acontece às vezes, sempre precisaremos ficar indo na wikidata? Discordo de tudo isso.--Rena (discussão) 04h29min de 12 de outubro de 2017 (UTC)
Here's what Google Translate makes of it:
I edit enough to feel right to question, which does not mean, mind you, that this is an attack on anyone who agrees or editors who have made a commitment to make such changes (Joalpe). I also wanted to register my repudiation of any attempt to replace the local infocaixas by lobbies of the wikidata. All examples of uses of these automatic infocaixas have been in small articles where no more than half a dozen elements are fed. I wonder how practical it would be to feed infoca- tions like the one in which does not mean, mind you, that this is an attack on anyone who agrees or the editors who have made such changes (Joalpe). I also wanted to register my repudiation of any attempt to replace the local infocaixas by lobbies of the wikidata. All examples of uses of these automatic infocaixas have been in small articles where no more than half a dozen elements are fed. I wonder how practical it would be to feed infoca- tions like the one in which does not mean, mind you, that this is an attack on anyone who agrees or the editors who have made such changes (Joalpe). I also wanted to register my repudiation of any attempt to replace the local infocaixas by lobbies of the wikidata. All examples of uses of these automatic infocaixas have been in small articles where no more than half a dozen elements are fed. I wonder how practical it would be to feed infoca- tions like the one in All examples of uses of these automatic infocaixas have been in small articles where no more than half a dozen elements are fed. I wonder how practical it would be to feed infoca- tions like the one in All examples of uses of these automatic infocaixas have been in small articles where no more than half a dozen elements are fed. I wonder how practical it would be to feed infoca- tions like the one inEl Greco or even some even bigger ones. And when that information needs to be updated, what happens sometimes, will we always have to stay on the wikidata? Symbol declined.svg I disagree with all this .-- Rena (discussion) 04:29, 12 October 2017 (UTC)
"El Greco" is a wikilink. You will see that when Google Translate reaches it it doesn't put it in but repeats the preceding text four times before entering the wikilink and carrying on to the end. Why is this?
80.5.88.70 (
talk)
11:20, 6 November 2017 (UTC)
Google is incompatible with
wikitext in the sense that it doesn't know which parts of the text is code and should not be translated. You should have said how you used Google Translate. I could not reproduce your result until I found your text is from
pt:Wikipédia:Esplanada/propostas/Utilização do Wikidata em infoboxes (3ago2017) and I entered the url in Google Translate. That means Google is not working on the wikitext (the source text when you edit) but on the rendered page. I don't know why Google Translate copies some of the translation. I don't get copies for other tested Portuguese Wikipedia texts with wikilinks. "El Greco" is not translated because Google correctly guesses from the capitalization that it is a proper noun. We also say
El Greco in the English Wikipedia. Google translates "el greco" as "the Greek".
PrimeHunter (
talk)
12:42, 6 November 2017 (UTC)
El Greco is presumably Spanish for "The Greek". The Portuguese equivalent is O Grego. Maybe Google was foiled by the change of language. Were all of the wikilinks you tested on pt:wp Portuguese ones or were some in other languages?
80.5.88.70 (
talk)
14:03, 6 November 2017 (UTC)
My tests included pages at
pt:Special:WhatLinksHere/El Greco with the same El Greco link in running text, e.g.
pt:Toledo. Regarding your example, if I enter the quoted sentence "Me pergunto o quão prático seria ficar alimentando infocaixas como aquela que está em El Greco ou mesmo algumas ainda maiores." in the box at
https://translate.google.com/ and select Portuguese to English then I get: "I wonder how practical it would be to keep feeding infocaels like the one in El Greco or even some bigger ones." If I change it to "el greco" but keep the sentence then I oddly get: "I wonder how practical it would be to keep feeding infocaels like the one in El Greco or even some even bigger ones." Here "El Greco" is the same as before but a second "even" is added near the end. My former test was with "el greco" alone which gives "the Greek", while "El Greco" alone stays as "El Greco".
PrimeHunter (
talk)
17:31, 6 November 2017 (UTC)
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
You will no longer see the patrol log on
Special:Log unless you specifically select it.
[25]
The
new version of MediaWiki will be on test wikis and MediaWiki.org from 7 November. It will be on non-Wikipedia wikis and some Wikipedias from 8 November. It will be on all wikis from 9 November (
calendar).
Meetings
You can join the technical advice meeting on IRC. During the meeting, volunteer developers can ask for advice. The meeting will be on
8 November at 16:00 (UTC). See
how to join.
The Community Wishlist Survey is the process when the Wikimedia communities decide what the Wikimedia Foundation
Community Tech team should work on over the next year.
The Community Tech team is focused on tools for experienced Wikimedia editors. You can post technical proposals from now until November 20. The communities will vote on the proposals between November 28 and December 12. You can read more on the
2017 wishlist survey page. /
Johan (WMF) (
talk)
20:15, 6 November 2017 (UTC)
Currently edit summaries for moves read (Jenks24 moved page
Example Link 1 to
Example Link 2: [reason]). I would prefer that the username be removed from the summary so it reads (
Example Link 1 moved to
Example Link 2: [reason]). This is because there are only a limited number of characters available in edit summaries and the current format needlessly uses a lot more of them than necessary, resulting in chopped off reasons (see for example my recent contributions). The username is unnecessary because you are either looking at the history page where you will see the username next to it or you are a browsing a user's contributions in which case again you already know the username. Plus you don't get caught out by username changes. I'm also fairly confident this is how page move edit summaries used to be set up prior to 2012 or 2013.
My questions for VPT are: can this be changed at enwiki or do I need to create a Phab ticket? If it can be changed locally, how? Thanks,
Jenks24 (
talk)
06:46, 4 November 2017 (UTC)
Thanks
PrimeHunter. I was hoping it would be able to be done here but suspected not. Do you think that it would be still be worthwhile me creating a Phab ticket or does it not seem fixable even for the devs on their end?
Jenks24 (
talk)
09:44, 5 November 2017 (UTC)
There was a ticket about this a few years ago (on Bugzilla), after the change which put the username there in the first place. I'm pretty sure it was closed as WONTFIX for reasons I thought were absurd at the time, but I wouldn't even know where to begin looking for it and my recollection could be wrong.
HJ Mitchell |
Penny for your thoughts? 15:29, 5 November 2017 (UTC)
Oh, wow. I actually commented on that VPT thread supporting a revert, I'd forgotten all about that. That Phab task makes sad reading, it seems your recollection is correct HJ. Frustrating that this is still an issue five and a half years down the track. Do you guys think opening a new Phab task would have any chance of success, or would it just get closed as duplicate of T36961? Has anyone ever had any success trying to get technical changes made by going through the community engagement team instead?
Jenks24 (
talk)
16:04, 5 November 2017 (UTC)
In this kind of situation (age of the old task, reasons for decline not apparently making sense), I would recommend writing a new task, referencing the old, with a suggested solution (mine would be that there should be two messages, not one -- I'm not sure why that obvious solution was not pursued). --
Izno (
talk)
16:10, 5 November 2017 (UTC)
(
edit conflict) @
Eric: I can't even find any cases of "This edit may have problems and should be reviewed" in my watchlist (which means that I can't use the uselang=qqx method) so I expect that you have
some settings that are different from mine. What might they be? --
Redrose64 🌹 (
talk)
14:54, 7 November 2017 (UTC)
Thanks for all the prompt replies! I'll put in an edit request at that page.
Redrose64: In my watchlist settings, the only one I see that might be related is Highlight likely problem edits with colors and an "r" for "needs review", which I do not have checked. But my watchlist legend displays as it appears in my discussion post linked above.
Erictalk15:09, 7 November 2017 (UTC)
Yes, nobody said that it was in the legend box. I've set a lot of boxes display:none; when they are the same on every page, take up unnecessary space or (most of all) use JavaScript thus delaying page load or causing things to skip around during or after load. --
Redrose64 🌹 (
talk)
16:29, 7 November 2017 (UTC)
In
this article, the
notes section has no content even though notes are included throughout the article. Whereas, in
this article, which uses the same code,
the notes do appear. Why do the notes not show in the first article but do in the second? I get the same result on Chrome, and Safari. Thank you.
Sandbh (
talk)
03:40, 7 November 2017 (UTC)
Image revisions not properly loading on Google Chrome?
I uploaded newer versions of
File:Too Much single 1989 by artist BROS.jpg and then
File:Bananas, Beaches and Bases.jpg. I still see old revisions of files. On Google Chrome, I tried reloading the browser over and over, yet I still see old revisions. I recently upgraded Google Chrome, yet I still see old revisions. However, on Microsoft Internet Explorer 11, the cache is cleared once I reload the IE11 every time. Is there something wrong with Google Chrome, or is the MediaWiki software struggling to load newer image revisions on Google Chrome? What about Firefox, which I don't have? --
George Ho (
talk)
07:54, 8 November 2017 (UTC)
Almost forgot, on Chrome, I even clicked the "Purge" button as well as the UTC gadget clock link (or whatever you call it). The issue still persists. --
George Ho (
talk)
07:56, 8 November 2017 (UTC)
I'm getting the following exception in my window, in debug and non-debug mode:
Error: Unknown dependency: jquery.badge Error: Unknown dependency: jquery.badge
at sortDependencies (https://commons.wikimedia.org/w/load.php?debug=true&lang=en&modules=jquery%2Cmediawiki%7Cmediawiki.legacy.wikibits&only=scripts&skin=vector&version=1b46sjy:11984:12)
at sortDependencies (https://commons.wikimedia.org/w/load.php?debug=true&lang=en&modules=jquery%2Cmediawiki%7Cmediawiki.legacy.wikibits&only=scripts&skin=vector&version=1b46sjy:12029:7)
at resolveStubbornly (https://commons.wikimedia.org/w/load.php?debug=true&lang=en&modules=jquery%2Cmediawiki%7Cmediawiki.legacy.wikibits&only=scripts&skin=vector&version=1b46sjy:12064:7)
at Object.load (https://commons.wikimedia.org/w/load.php?debug=true&lang=en&modules=jquery%2Cmediawiki%7Cmediawiki.legacy.wikibits&only=scripts&skin=vector&version=1b46sjy:12961:17)
at https://commons.wikimedia.org/wiki/File:Coccinellidae(Ladybird)-2.jpg?debug=true:258:63
at startUp (https://commons.wikimedia.org/w/load.php?debug=true&lang=en&modules=startup&only=scripts&skin=vector:9056:15)
at HTMLScriptElement.script.onload (https://commons.wikimedia.org/w/load.php?debug=true&lang=en&modules=startup&only=scripts&skin=vector:9081:3)
I've written a user script called
StripSearch.js. It creates a menu item that toggles the removal of details from Wikipedia's search results on and off, to view the results as a bare single-spaced list of page names.
What I'd like to do is fork StripSearch.js to create
SearchSuite.js and add some more toggles (as menu items).
The first one I would like to add is a menu item for turning the sorting of search results on/off. When it sorts, it will do so whether or not the search results have been stripped yet.
Another one will insert/remove wikilink formatting (bullets, and double square brackets around each page name), in stripped mode only, for ease of copying and pasting the links into lists. (Not everyone will want the link delimiters displaying all the time – hence, the menu item).
And more.
(Both of the above mentioned features are included in
StripSearchSorted.js, but operate by default without on/off switches.)
What do I need to think about in approaching this script writing project?
I think the basic approach to this would be to set classes for elements you want to show/hide, because then the code for the menu items themselves becomes really easy: when clicked, you just hide/show the relevant classes, i.e. searchSuite-foo-show and searchSuite-foo-hide for elements that should be shown/hidden when foo is activated (and then reversed when foo is turned off). Sorting is going to more tricky to do/undo, but basically you would store data about the original order, then write functions like sortAlpha and sortOrig that would operate on the containing element, replacing it's contents with sorted contents. For a simpler example, to sort the links in the toolbox, you could do
I'm not really sure, just be aware anything one feature might change has to be accounted for when coding the other features. - Evad37[
talk17:09, 9 November 2017 (UTC)
Redirect to a module
Hello. Is there any way to add a redirect to a module within the Module namespace, or otherwise create a shortcut to or alternate name for one? —
Jaspet12:18, 8 November 2017 (UTC)
It's a couple lines of code. Why though? Modules are not really meant to be used outside templates or other modules. --
Izno (
talk)
14:31, 8 November 2017 (UTC)
Basically no, you could have the 'old' module make calls to the 'new' module, but that is just messy. What need would there be for this? —
xaosfluxTalk14:55, 8 November 2017 (UTC)
The reasoning is the same for redirects to templates. To have shortcuts and alternate names really has three functions: it allows for more concise ways of linking to the target (compare {{R}}), it gives the option of using a slightly more specific/relevant/accurate name for the task at hand (there is a number of reasons for this when it comes to templates, such as naming one particular function of the whole), and it decreases the chances of error resulting from an incorrect name (e.g. due to capitalisation where there need not be). I'm surprised the same practice doesn't exist for modules, considering it's ubiquitous within the Template namespace (and virtually every other namespace with perhaps the exception of File). —
Jaspet02:34, 9 November 2017 (UTC)
You can make a pseudo-redirect with return require('OTHER_MODULE_NAME'), which will load the code from the other module, but won't work as a redirect when viewed.
{{repeat|p|3}}ery (
talk)
02:37, 9 November 2017 (UTC)
Regarding why a module needs to redirect, it's not unusual to want to rename a module to better reflect its purpose as it evolves, or just because you thought of a more apt name, but you don't want to break anything using the old name.
isaacl (
talk)
03:44, 9 November 2017 (UTC)
You can't use the #REDIRECT [[...]] technique because that only works for Wikimarkup pages. Templates are Wikimarkup, so it works for them; modules are not. It's exactly the same problem as with .css and .js pages: redirection cannot be used with these either. --
Redrose64 🌹 (
talk)
21:46, 9 November 2017 (UTC)
Article title search defaults to User page search
Don't know exactly when started happening, but recently any search argument I enter in the upper-right search box defaults to prefixes "User:" and "Wikipedia:" (e.g. if I enter "Hillary", instead of article titles beginning "Hillary", pages beginning "User:Hillary_" are displayed instead). I went to Preferences but could find no option to stop those prefixes from being added to my search argument. How do I get back to being able to see article titles beginning w/ my search argument? --
IHTS (
talk)
11:55, 10 November 2017 (UTC)
(
edit conflict)@
Ihardlythinkso: Go to
Special:Search. Below the search input box there may be a number of checkboxes, one for each namespace; if this is not present, click "Advanced". In that list of namespaces, click on the empty checkboxes for the namespaces that you do want, and also click on the filled checkmarks of those that you don't want. Then click the "Remember selection for future searches" checkbox, enter any search query, and click Search. --
Redrose64 🌹 (
talk)
12:12, 10 November 2017 (UTC)
Thx! p.s. I'm sure now what happened. I checked "Remember" recently, thinking *that* search box was separate (& differently maintained) from the upper-right corner search box. (My assumption was wrong, thus my confusion.) Thx again. --
IHTS (
talk)
17:40, 10 November 2017 (UTC)
"User-friendly" editing tools damaging references
From time to time I stumble on an article where references have been replaced by blue superscript markers of the form [[<article name>#cite note-NNN|<sup>[NNN]</sup>]].
This search currently lists 31 articles where footnote 1 has this form. According to the edit summaries,
this edit was made with the Visual Editor,
this one with the 2017 source editor. This damage is difficult to put right, as the articles don't immediately look broken, and get edited for months afterwards. Readers are left with an article where the blue footnote markers either don't go anywhere or link to the wrong footnote. Can we try to identify how these good-faith editors are using these editing tools, and fix the tools so that this kind of damage does not occur? --
John of Reading (
talk)
08:58, 10 November 2017 (UTC)
There are actually a surprising number of deliberate references of this sort, almost to the point of looking self-referential in a sense.
[31] --
Izno (
talk)
15:08, 10 November 2017 (UTC)
Those are different, though. They do look deliberate, and they don't have the "sup" tags. How are they supposed to work reliably, though? Is there a bot that maintains the links when "cite-note-23" in a separate article becomes note 24 because someone added a new reference? See the phrase "critically acclaimed" in
Howard Hughes, for example, which points to ref 23 in
The Aviator (2004 film). –
Jonesey95 (
talk)
16:33, 10 November 2017 (UTC)
This looks like an unfortunate user error. Ref tags, when you're reading a page, don't actually contain the citation; they're just HTML links to an anchor on the page, with a little superscript character formatting. If you copy the little blue clicky number in an article that you're reading it, then you are actually copying the link to the anchor with the superscript formatting (that is, copying the little blue clicky number itself, not the citation that the anchor links to). As a result, when you paste an HTML link into VisualEditor, then you get a link – the [1], rather than the ref>{{cite web|url=www.example.com...</ref>.
As a practical matter, those editors should be reminded that if they want to copy refs from their sandboxes or other articles, they either need to copy the wikitext (which they can paste into the visual editing mode, as it will convert it all for them) or open their sandbox in the visual editing mode and copy the footnote from there. (This information is already included in most of the materials for educational programs, but not everyone is a student, and not every student always remembers this step.)
Whatamidoing (WMF) (
talk)
17:05, 10 November 2017 (UTC)
Is there some warning we could pop up that could deter editors from damaging articles in this way, or maybe a bot that could track these edits, notify editors about them, or revert them? –
Jonesey95 (
talk)
18:34, 10 November 2017 (UTC)
Strange editing problem with new interface
Since the new source editing interface (not sure what it's called and honestly don't really care) was introduced recently, I have frequently run into a weird problem when trying to edit. Sometimes (it's not that uncommon, actually, and it just happened to me a few seconds ago), when I'm trying to add or delete content on a page, it will add or delete the content several spaces away from where my cursor is, and therefore where I intended to add/delete it to/from. This means I have to try to add/delete content a few spaces ahead/behind where I actually want it to be. Does anyone know why this frustrating bug might be happening, and has it happened to anyone else?
Everymorning(talk)21:47, 10 November 2017 (UTC)
I have exactly the same problem, on both chrome (latest version) and Firefox (really old version). It generally occurs near the bottom of a long page, the cursor is slightly further off the further down the page I go. it occurs for me, only on the vertical plane. horizontal position is accurate.
Dysklyver09:55, 11 November 2017 (UTC)
"There was an error connecting to the database. This is most likely a temporary condition. Please try again in a few minutes. If the problem persists, please contact User:CBM on enwiki. The error message is: Unknown database 'p50380g50494_data' WP 1.0 bot revision 541, updated Sat, 6 Dec 2014 by theopolisme"
I went to
User:CBM talk page, but answer to previous query says CBM is no longer involved in maintaining that wiki-system. Is there any one out there who can fix this? Variations of this table are used by many Wiki-Projects.--
Orygun (
talk)
19:57, 5 November 2017 (UTC)
In addition, the process to update Assessment tables is broken. Update project data – Wikipedia Release Version Tools at url tools.wmflabs.org/enwp10/cgi-bin/update.fcgi. —
JoeHebda • (
talk)20:28, 8 November 2017 (UTC)
@
Optimist on the run and
Explicit: I rather think that somebody should have verified with the bot operators and those who maintain the mass message system that this was feasible before a bot was sent in to
make all the category amendments on user talk pages. These users may suddenly find that they are receiving messages that they thought that they had blocked. I see nothing at the CFD to indicate that such checks had been made, nor is there any notice at
WP:BON. --
Redrose64 🌹 (
talk)
10:05, 11 November 2017 (UTC)
Thank you for doing that. I admit I wasn't expecting things to move quite so quickly once the discussion closed, but you're absolutely right that I should have added a link to the nomination at
WP:BON from the outset. Lesson learned for next time. --
Black Falcon(
talk)04:49, 12 November 2017 (UTC)
The new category name doesn't make sense. It's not just Wikipedians who can opt-out, it's any talk page. It just happens to be that so far it's just users who have opted out.
Legoktm (
talk)
05:26, 12 November 2017 (UTC)
Starting about a week or so ago, with some frequency when I go to mark my watchlist as being taken care of, I get a database error. The most recent one said:
A database query error has occurred. This may indicate a bug in the software.[WfaDTQpAICoAAC8eu7IAAAAS] 2017-10-30 01:41:33: Fatal exception of type "Wikimedia\Rdbms\DBQueryError"
The watchlist is properly cleared anyway, and this doesn't happen every time, but I thought I'd report it just in case it was significant in some way.
Yes, I've also been seeing similar errors much more frequently. Mine usually occur when refreshing my watchlist. Here's the latest: [WfclJwpAAEUAAEEpVwsAAADO] 2017-10-30 13:13:07: Fatal exception of type "Wikimedia\Rdbms\DBQueryTimeoutError"
older ≠
wiser13:17, 30 October 2017 (UTC)
No, I don't either. although I'll note that this error has not occurred for the past couple of days. Before that it was happening with some frequency.
older ≠
wiser09:54, 8 November 2017 (UTC)
That's really not all that helpful, as once I clear the watchlist, which generates the database query error, the database is cleared, and there's no way to check anything as suggested by the page you've pointed us to. What I don't understand is that both Bkonrad and I have quoted an error code here – shouldn't that tell you where the problem is?
Beyond My Ken (
talk)
00:45, 13 November 2017 (UTC)
Beyond My Ken, I'm most of time that kind of issues are related to gadgets or scripts that are broken. That may be the case for you too, so I advice you to use
that link to explore your watchlist and then clear it. Do you still have that issue? How many items do you have in your watchlist (on average)?
Trizek (WMF) (
talk)
10:06, 13 November 2017 (UTC)
Question about oversized photos on my Wikipedia Sandbox page
If I expand the photos on my Wikipedia Sandbox page to a ridiculously large size, does that detract from anyone else's Wikipedia experience? I was just experimenting but I also don't want to be a bandwidth hog. Thank you!
Beauty School Dropout (
talk)
05:35, 13 November 2017 (UTC)
I'm not sure if this completely answers your question, but If you don't specify a size while using thumb, user preferences should determine the size they should have (and a decent default should be presented to unregistered users). I also suggest reading
MOS:IMAGES, the style guide for images. I hope this helps, —
PaleoNeonate –
07:32, 13 November 2017 (UTC)
I read the question to mean that if I markup a huge image size like [[File:Example.jpg|10000px]], will that make the servers create an image of that size (server load), and is that huge image then downloaded to my computer (network bandwidth)?
And the answer for jpg and png is no in both cases, but for
SVG images a png of the specified size is generated and displayed (and thus downloaded).
Latest tech news from the Wikimedia technical community. Please tell other users about these changes. Not all changes will affect you.
Translations are available.
Changes later this week
The diff you see when you compare two different versions of a page has changed on
MediaWiki.org and
the test wiki. This is to make it easier to find a text change in a moved paragraph. It will hopefully soon come to more wikis. You can
report bugs in Phabricator.
[33]
A new user group on Commons will be able to upload
MP3 files. The plan is to have this user group from 17 November.
[34]
Wikis using Flagged Revisions will get the
New filters for Edit Review by default on the recent changes pages. It will be possible to opt-out in user preferences.
[35]
The
new version of MediaWiki will be on test wikis and MediaWiki.org from 14 November. It will be on non-Wikipedia wikis and some Wikipedias from 15 November. It will be on all wikis from 16 November (
calendar).
Meetings
You can join the technical advice meeting on IRC. During the meeting, volunteer developers can ask for advice. The meeting will be on
15 November at 16:00 (UTC). See
how to join.
Future changes
Support for uploading and viewing 3D models is coming soon to Wikimedia Commons. The feature will support the
.STL file format. You can see an example
on the test wiki.
[36]
I would like to tidy up references on various articles. I can readily generate a list of the articles I want to look at. Is there any way of retrieving the references for these articles in bulk, for example like a search result.
Eno Lirpa (
talk)
11:20, 10 November 2017 (UTC)
is matching most hatnotes correctly, but runs into a problem with:
{{About|the church in [[Aleppo]], [[Syria]]|other churches with the same name|Church of the Holy Mother of God (disambiguation){{!}}Church of the Holy Mother of God}}
the matched string is truncated at the end of the embedded template:
{{About|the church in [[Aleppo]], [[Syria]]|other churches with the same name|Church of the Holy Mother of God (disambiguation){{!}}
How can I tweak the regex to include the embedded template when scanning to the end of the hatnote template? Thanks,
wbm1058 (
talk)
18:28, 10 November 2017 (UTC)
That regex isn't very efficient. In my testing I found that maybe a third of my test cases weren't working, and it was a bear for me to find out why as the ones that failed seemed kind of random. After much frustration, I finally checked the error code (I know, good coders should always check for errors ;) and found it was returning PREG_JIT_STACKLIMIT_ERROR for the ones that failed.
Error codes The new PREG_JIT_STACKLIMIT_ERROR constant introduced with PHP 7.0.0 – I guess I either write a more efficient regex or disable the PCRE JIT via php.ini (pcre.jit=0). Running that in the AWB regex tester I see that it's stacking them up one character at a time. Someone's twitter
discussing a similar problem: "It was creating a capture group for every single character in the string and blowing out of memory" – Help writing a more efficient regex would be appreciated.
wbm1058 (
talk)
04:06, 15 November 2017 (UTC)
Another thing that can be done is limiting backtracking via a technique called "unrolling the loop". The end of the pattern that matches the contents before the closing "}}" is of the form (?:caseA|caseB)*, where caseA is [^}] – anything but an opening brace – and caseB is \{[^{]|\{\{[^{}]+\}\} – either an opening brace followed by non-opening brace, or a pattern matching a template inclusion without any nested templates. This can be transformed into caseA*(?:caseBcaseA*)*, which can be thought of matching caseA first, then multiple repetitions of (one caseB plus zero or more caseA's). By eliminating the alternation, the pattern can only match one way, and backtracking is eliminated. For this particular case, since case B is an alternation itself, it could in turn be unrolled into case C and D, with an end result of caseA*(?:caseC*(?:caseDcaseC*)*caseA*)*. But the problem is the pattern within the first set of parentheses consists of a sequence of zero-or-more matches, which introduces a new source of backtracking. So the regexp would have to be tweaked to avoid this (the pattern in the parentheses needs to be anchored with some patterns that don't have a zero-or-more quantifier). However I suggest starting with the single unrolled loop first.
isaacl (
talk)
04:54, 15 November 2017 (UTC)
I'm guessing you're not actually directing this question to me? I haven't looked specifically at uses of the About/For/Redirect templates, but for some other templates, it is common to have line breaks before the pipe character that separates template arguments.
isaacl (
talk)
05:12, 15 November 2017 (UTC)
Right, while it would be very unusual to put line breaks in a hatnote template, it's not syntactically illegal. More common is placing multiple hatnote templates on a single line.
wbm1058 (
talk)
17:40, 15 November 2017 (UTC)
First I tried just making it non-capturing (grouping-only) parenthesis: (?:···) – that cleared the results in the AWB regex tester, but had no noticeable effect on my PHP test results.
Then I went with the suggested "
unrolling the loop" solution: \{{2}(About|For|Redirect)\s*\|+[^{]*(?:\{[^{]|\{\{[^{}]+\}\}[^{]*)*\}{2} did the trick! Stacklimit errors gone! Thanks!
One of these days maybe I'll get around to reading that O'Reilly book "Mastering Regular Expressions" that I bought ~5 years ago.
wbm1058 (
talk)
17:40, 15 November 2017 (UTC)
The regex I use to find templates (in JavaScript) is /\{\{\s*(.+?)\s*(\|(?:.|\n)*?(?:(?:\{\{(?:.|\n)*?(?:(?:\{\{(?:.|\n)*?\}\})(?:.|\n)*?)*?\}\})(?:.|\n)*?)*|)\}\}\n?/g if that helps anyone - Evad37[
talk15:09, 15 November 2017 (UTC)
Ordered list with multiple items in same order
Ordered list in wikipedia can be created by using the "#" symbol like this:
Item 1 Score 997
Item 2 Score 993
Item 3 Score 990
Item 4 Score 990
Item 5 Score 984
However, as indicated in the above example, there are situations where some items should be ranked the same in an ordered list. Is there any syntax that would allow the use of two "3." in an ordered list?
C933103 (
talk)
19:09, 15 November 2017 (UTC)
According to
Help:List#Specifying a starting value, you can use a hybrid of wiki-markup and HTML to set an arbitrary number for an entry in a list, and then the list continues consecutively thereafter:
Item 1 Score 997
Item 2 Score 993
Item 3 Score 990
Item 4 Score 990
Item 5 Score 984
Item 4 is coded as:
#<li value="3">Item 4 Score 990</li>
If you wanted to have Item 5 become "5." (the fifth in order, rather than having the fourth rank), you could use that same technique to assign it as value="5".
DMacks (
talk)
19:17, 15 November 2017 (UTC)
I see, thanks. So, if it is a 100-member list, then all the remaining 95 items will need to be manually assigned with the value too?
C933103 (
talk)
19:45, 15 November 2017 (UTC)
You only have to manually assign when you want a number to be non-consecutive with the previous. So if you wanted:
Item 1 Score 997
Item 2 Score 993
Item 3 Score 990
Item 4 Score 990
Item 5 Score 984
Item 6 Score 980
only the boldfaced ones need to be assigned manually because those are the entries whose values are non-continuous from the previous (Item 4 is #3 rather than #4 after the previous #3; Item 5 is #5 rather than #4 after the previous #3). Item 6 is automatic #6 after the previous #5.
DMacks (
talk)
20:06, 15 November 2017 (UTC)
Using Wikimedia maps instead of geohack?
I've experimented a bit with the
sandbox version of our coordinates, in order to make use of the wikimedia maps, instead of geohack. There are some examples in the
testcases and in order to get a bit more realistic impression I temporarily added it to
this article, which also allows it to pull the geoshape. From the map viewer, you can still select alternative map services (including geohack actually). For maps from the moon and mars, we fallback to geohack. Several other wiki's already use this, the biggest being the GermanRussian Wikipedia. I'm wondering if people think it is worth pursuing this further. Please leave your feedback. —
TheDJ (
talk •
contribs)
13:36, 14 November 2017 (UTC)
Please, no. Our task is to be useful to readers, not to be ideologically pure. I get that there are some people who are fanatically open-source-only and resent the fact that we give prominence to commercial entities, and I equally get that there's a legitimate desire to "keep it in the family", but for actual readers who click links because they want to know more about the topic—as opposed to
spherical cow readers who are only interested in seeing where the subject is on the map—Google and Bing, and even to a lesser extent OpenStreetMap, are orders of magnitude more useful.
To stick with your example,
Wikimedia Maps shows me where Stadskanaal is.
Google Maps shows me where Stadskanaal is, shows me how I get to Stadskanaal if I want to see it for myself, allows me to see that the Museumspoorlijn STAR is nearby so if it's pouring with rain when I arrive I know there's still something to do, allows me to see which shops are in the area so I know whether to bring my own food and drink or not, lets me check the train and bus times so I know when to set out and return, and allows me to see other people's photos of the area so I have an idea what I'm seeing. Given that even a company with the resources of Apple has struggled for years to compete with Google when it comes to mapping that people actually want to use, I have no realistic hopes that Wikimedia Maps will ever be a viable alternative. Wikipedia/media is an important part of the internet ecosystem, but we need to not follow de-wiki down the "ourselves alone" blind alley they've been pursuing for the last few years and recognize that sometimes, other organizations do things better than us and it's a service to our readers to point that out when it's the case. ‑
Iridescent22:30, 14 November 2017 (UTC)
Iridescent, both of the options being considered offer direct links to Google Maps. It's in the list under "Global services" for the one, and under "External services" for the other.
WhatamIdoing (
talk)
23:07, 14 November 2017 (UTC)
Sure, but the existing setup has links to potentially useful mapping services prominently displayed, with the worthless Wikimedia Maps relatively hidden in plain sight down the right (I've no idea if the WMF have ever researched this, but I'd be willing to bet that well over half the visitors to the Geohack link don't even notice that it's there), whereas the proposed alternative is a huge Wikimedia Maps splodge with some tiny links hidden in miniature type down the right. (And I know you're aware that links in this position don't get noticed, given that "information on the right of the screen tends not to be noticed by readers" has been the WMF's prime argument in their years-long crusade to impose a maximum display width.) ‑
Iridescent23:16, 14 November 2017 (UTC)
As noted, geohack would remain available from this new UI though, we would just not enforce every person through it. One more click. —
TheDJ (
talk •
contribs)
19:41, 15 November 2017 (UTC)
One more click isn't a problem. However, taking the geohack option should allow the localisation aspects (region parameter of {{coords}}) to be preserved. This doesn't seem to be happening in the region:GB case in
testcases (but I am always out of my depth at WP:VPT!).
Thincat (
talk)
08:30, 16 November 2017 (UTC)
I think this is a good step forward, particularly since it's moving from a short-term(!) hack on a temporary server to something that's properly integrated with Mediawiki. For those that aren't aware: this is a regularly-updated copy of the OpenStreetMap dataset, which we can use like this because it's freely licensed. Thanks.
Mike Peel (
talk)
23:57, 15 November 2017 (UTC)
@
TheDJ: I would like it to be enabled, but there are still quite a few issues with the map style. I filed a few Phabricator bugs more than a year ago but they haven't been addressed (e.g. the map still doesn't render light rail tracks, and still incorrectly renders stop positions when they aren't supposed to be rendered at all). Furthermore, the map doesn't render the eastern half of Sydney Harbour for some reason (it displays properly in the osm.org style so no idea what's up with it).
Jc86035 (
talk)
07:01, 16 November 2017 (UTC)
@
Jc86035: they haven't been addressed because the community isn't showing it's support for the effort for the intermediate progress. As such the foundation has scaled back on it. It's a common theme where we demand perfection and as a result just simply end up getting nothing. —
TheDJ (
talk •
contribs)
07:49, 16 November 2017 (UTC)
@
TheDJ: Would the WMF ever consider further improving or expanding the stylesheets at this point? (I suppose it wouldn't be that hard for me to figure out for myself how to copy the railway=tram styling to railway=light_rail, though.)
Jc86035 (
talk)
11:14, 16 November 2017 (UTC)
@
Jc86035: They'll review and approve patches if you submit them, but that's about it. What surprises me, is that there isn't even a single proposal among the community wishlist dealing with maps, a sharp contrast with previous years... Honestly, compared to previous years I find the quality of the wishlist proposals to be significantly lower all around... Let's hope people will make the most of this opportunity in the last 3 days they have available.... —
TheDJ (
talk •
contribs)
13:25, 16 November 2017 (UTC)
Is there a way to search for pages which 1) have a certain word (or pattern) in their title, and 2) we created within the past N days? --
RoySmith(talk)13:42, 16 November 2017 (UTC)
For one yes. There is more
here. If you search for intitle:"word or phrase" in the regular search you should get what you are looking for. Unfortunately there is no way to filter by time as far as I know.
This tool can be used to search for regular expression patterns. (But again, nothing about time). --
The VoidwalkerWhispers21:23, 16 November 2017 (UTC)
I really don't like not having the
WP:OBOD, so shortly after they introduced Notifications, someone helped me to create
User:Nyttend/monobook.js with code restoring the OBOD; it basically just calls
User:Writ Keeper/Scripts/orangeBar.js. Normally it works, but it's intermittent; why would that be the case? For example, some minutes ago I got the OBOD because of a message from
User:Sbmeirow, so I viewed it, went to another page, got another OBOD, viewed the message, went to another page, and suddenly here's the
You have new messages notice without the OBOD. I would expect that the OBOD either would never work or would work all the time, especially in a matter of a few minutes; this intermittency just doesn't seem to make sense. I'm running IE 11.
Nyttend (
talk)
23:42, 16 November 2017 (UTC)
Can anyone help me make a complex geologic time scale-related template?
I'm trying to make a
navbox template that can easily be filled in with a region's
stratigraphic units that are manually inputted into variables by editors. The problem I'm facing is that there are a huge number of time units on the geologic timescale that spreads the content over a wide area of the template. Is there any way to make it so that only the units that have had content inputted for their respective variables show up in the template?
Abyssal (
talk)
14:18, 16 November 2017 (UTC)
In August
Whatamidoing (WMF) said that a workround to the 1000 limit would be available in about a month. Can we have an update on progress on this workround?
I understand that a partial workaround is in the works, to let you filter out changes you've already looked at. (So you read the most recent 1000, then hide the ones that you've already read, and see the next-most-recent 1000, and repeat – you'll never miss any, but you won't see 1,0001+ on the screen at the same time.) However, this probably won't be available for approximately one month. In the meantime, please reset your watchlist numbers to something greater than 0.
It's live now.
You need the new filter system enabled. Click by the "hamburger" icon where it says "Filter changes (use menu or search for filter name)". Tick the box for "Unseen changes" (only). Wait (and wait) for it to filter the list.
Whatamidoing (WMF) (
talk)
19:13, 17 November 2017 (UTC)
No I have not got that enabled. Tried and looks better but there is problem in that it only marks it visited if you actually open the page. Most of the time a page is only visited using pop-ups so that also needs to be marked as visited for this to be useful.
Keith D (
talk)
21:01, 17 November 2017 (UTC)
Possible to set timestamp of pages-visited on watchlist?
Special:Watchlist somehow tracks a datestamp of the last time "Mark all pages visited" was clicked. Is it possible to set it to some point other than the time at which that button is clicked? If I'm away for a few days, I have a long list of pages on my watchlist that have been changed. I browse the list from the oldest change I hadn't seen, but might not have time to get as far as "now". But with popups and such, I might not actually "read" every article. I'd like to update the mark to how far I got, as if I clicked mark-all-pages-visited at some arbitrary other point in the watchlist. Is this possible? Or is there some other way to keep a "pages changed since [time I choose] based on watchlist list" rather than just a certain period of time before now?
DMacks (
talk)
03:37, 16 November 2017 (UTC)
I created a proof of concept type script at
User:The Voidwalker/setWatchlistTo.js. The one you click is marked as unread, as well as all above it. The implementation is a bit ugly, and creates the dreaded page-jumping (hence proof of concept). If you have suggestions for making it look better, they would be most welcome. --
The VoidwalkerWhispers00:58, 17 November 2017 (UTC)
Special:Watchlist doesn't track the time at all. There is a field for the "notification timestamp" on every watchlist entry; whenever a page is edited, every watchlist entry for the page where this field is empty gets filled in with the time of the edit. If the field is non-empty, Special:Watchlist displays the page as having new revisions, and the history for the page displays all revisions with the field's timestamp or newer does similarly. All "Mark all pages visited" does is clear the field for all pages on your watchlist.
If you make use of the indications on the history page to see all revisions of the page since you last viewed it, you would not want to use a script like the one The Voidwalker made since it would reset the field to indicate some revisions had been seen when they actually hadn't. On the other hand, if you ignore the history page and re-read the whole page whenever your watchlist says it has been updated, The Voidwalker's script would probably be fine for you.
Anomie⚔22:12, 17 November 2017 (UTC)
Ah, that gives me an idea. I could easily modify the existing script to only set the timestamp for individual pages. The downside to that being you would have to mark everything individually, but I could set up check boxes.... Anyway, good to know Anomie, thanks. --
The VoidwalkerWhispers23:57, 17 November 2017 (UTC)
Special:WhatLinksHere/49 says "
49 (transclusion)". This means the page "transcludes" itself. It's caused by {{PAGESIZE:{{FULLPAGENAME}}|R}} in {{Short pages monitor}}. {{PAGESIZE:X}} apparently counts as transcluding X.
Special:WhatLinksHere/37 also says "
37 (transclusion)" after a
null edit. Maybe PAGESIZE hasn't always counted like this. I didn't null edit
60 if others are curious. There is a similar situation with {{#ifexist:X}} which causes a WhatLinksHere entry (as a link and not a transclusion) even though no link is produced. I guess it's about which database operations are performed.
PrimeHunter (
talk)
15:49, 16 November 2017 (UTC)
The reason these links are stored is to do with caching. When the page "49" is edited, its size may change, so any pages that include or use the size of that page will also need to be updated (this occurs as part of the job queue, which does not guarantee instantaneous updates). Same with page existence. — This, that and the other (talk)09:39, 17 November 2017 (UTC)
But when the page "49" is edited, the software already knows that "49" will have to be updated. There's no need to remember whether a page tests its own size. --
John of Reading (
talk)
07:24, 18 November 2017 (UTC)
But when the page "49" is edited
Only if one ignores the fact that the size of the page can change without being edited. The way chaotic way that wikitext was designed leads to a lot of inception-like madness. It is an unsolvable problem as long as page properties depend on another page, and as long as reading & editing are convoluted. The only thing developers can do is mask it or kill all functions that cause it.11:03, 18 November 2017 (UTC) — Preceding
unsigned comment added by
197.218.82.255 (
talk)
Thank you for all the helpful replies. It seems that the Wikimedia software is confusing "Page X has a clickable wikilink leading to page Y" with "The appearance of X depends on the content of Y, so X may need purging if Y changes". Given an infinite budget we'd rewrite that bit, but it doesn't sound as if that will be happening this week. At least the problem is more manageable once you understand it.
Certes (
talk)
11:15, 18 November 2017 (UTC)
It sounds like you are confusing links and transclusions at WhatLinksHere. When WhatLinksHere says "
49 (transclusion)", it is not in any way a claim that the page has a wikilink to
49. "(transclusion)" means it is a (false) claim that the page
transcludes49 in the same way as templates can be transcluded. Click "Hide transclusions" on WhatLinksHere to hide pages which are only listed due to transclusions and not wikilinks. For example,
Special:WhatLinksHere/Template:Citation needed has a huge number of article entries with "(transclusion)", so those articles say {{citation needed}} (or use a redirect like {{fact}}) but they don't say [[Template:citation needed]]. Most articles are not transcluded anywhere. You could say two things:
1) The Wikimedia software is (deliberately) confusing "Page X has a clickable wikilink leading to page Y" with "Page X uses {{#ifexist:Y}} to test whether page Y exists".
2) The Wikimedia software is (deliberately) confusing "Page X is transcluding page Y" with "Page X uses {{PAGESIZE:Y}} to get the size of the wikitext in page Y".
"
49 (transclusion)" at
Special:WhatLinksHere/49 is an example of 2). An example of 1) could be created if a page with no link to
49 says: {{#ifexist:49|There is a page called "49"|There is no page called "49"}}. 1) is documented at
Help:What links here#Overview: The parser function
#ifexist: causes a listing in "What links here" among the normal links even though no link is produced. 2) appears to be undocumented.
PrimeHunter (
talk)
19:58, 18 November 2017 (UTC)
The logs
[38] show
User:Leprof 7272/Everipedia was moved from
Everipedia 20 February 2017 including the page history at the time. Two pages have later been created and deleted at
Everipedia but all deleted edits there are from March 2017 and seem unrelated to
User:Leprof 7272/Everipedia apart from being about the same subject. In a brief comparison I found no sign of content being copied between the pages so I see no need for a history merge if deleted edits of
Everipedia are restored. If content is later copied between the pages then it can just be noted at the time of the copying per
Wikipedia:Copying within Wikipedia.
PrimeHunter (
talk)
16:17, 19 November 2017 (UTC)
Visual Editor lacks a Cite field for naming references
Hi, y'alls. In Visual Editor there is no field in the Cite creator for Reference name. e.g. - to render: <ref name = "hamburger steak"> in the dressed citation template.
Would it be easy to add that field? It is very common. Right now it is kludgey to add it after the fact, or by toggling to the text editor. Ping me back. Having fun! Cheers! {{u|
Checkingfax}} {
Talk}02:01, 20 November 2017 (UTC)
Through a Google search, I stumbled upon
this website, which looks like a carbon copy of Wikipedia under a different domain name. It looks exactly the same and up to date on any page, so it's apparently not a mirror site cached on some server but rather a proxy to Wikipedia. At first I suspected it was something benign, like for people living in parts of the world where they are denied access to Wikipedia or something, but upon making searches and failing to acquire absolutely any information on the site, I grew suspicious.
It is under
.ng, the Nigerian ccTLD, but
Whois reveals it was registered in April 2016 by someone with a Turkish address and a
mail.ru account. The source is essentially identical to Wikipedia's except Google Analytics code is embedded in the header. Accessing the site via a non-secure address redirects to HTTPS, but the certificate is one issued by a different organization than one that Wikimedia uses (I have no idea how relevant this is, but still).
Is this some kind of phishing site? And even if it was innocuous in nature, it concerns me that it was included matter-of-factly in Google results when it is completely indistinguishable from the actual Wikipedia in the eyes of someone who doesn't always pay attention to the address bar, AKA everyone.
WP:MIRROR attempts to deal with this. I thought I had seen a way to report to WMF legal when blatant problems such as using trademarked logos occurred, but I can't see it.
Johnuniq (
talk)
22:44, 18 November 2017 (UTC)
Thanks. Of course if I had tried the obvious
WP:TRADEMARK I would have found that. I reported the website (which includes other languages) to the legal email address advised as a trademark violation.
Johnuniq (
talk)
21:54, 19 November 2017 (UTC)
I have a question about fixing broken Wikilinks. I wasn't sure where to ask it, so I'd figured I'd ask it here. There used to be a subsection in
Japan Shogi Association titled "Apprentice school", and links to this section have been added to various other articles as
Japan Shogi Association#Apprentice school. The relevant content in the section, however, was recently moved to another article
Professional shogi player (more specifically
Professional shogi player#Apprenticeship. So, I would like to know if there's a quick way to fix all of these links at once by sort of redirecting them to the new article/new section, or do all the individual links need to be manually corrected. There were other subsections in
Japan Shogi Association#Players which were also moved to "Professional shogi player", so there other links which need correcting as well. However, the method for correcting them all should be the same, so once I find out the best way to fix one set, the others can be fixed in the same manner. --
Marchjuly (
talk)
02:29, 20 November 2017 (UTC)
a quick way to fix all of these links at once by sort of redirecting them to the new article/new section, or do all the individual links need to be manually corrected The latter, unless you've actually targeted a redirect with those links (and that redirect's title could just as easily redirect to the new location of the content). For example, were the content of
Agahnim to be moved elsewhere (say,
The Legend of Zelda), then you could simply change the text of Agahnim from #REDIRECT [[Characters in The Legend of Zelda series#Agahnim]] to e.g. #REDIRECT [[The Legend of Zelda#Agahnim]]. Otherwise, you have to hunt down all the links and fix them manually. Someone at
WP:Bot requests can probably take care of this if you have an exact list of links that you're looking to fix. --
Izno (
talk)
02:51, 20 November 2017 (UTC)
Thanks
Izno and
PrimeHunter for the input and the extra info. I believe I am the one who added most of the links in question, so fixing them manually it not a huge deal. I just wasn't sure if there were others floating around in articles which I didn't add. I didn't know about {{search link}}, but that seems more than sufficient to track them all down. --
Marchjuly (
talk)
04:06, 20 November 2017 (UTC)
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
If you use the Chrome web browser on
Android you can see a download icon on the mobile website. You can download a formatted PDF. It will work in other mobile browsers in the future.
[39]
The abuse filter now has a function called get_matches. You can use it to store matches from regular expressions – one of each capturing group. You can read more
in Phabricator.
Problems
Last week's MediaWiki version didn't come to all Wikipedias because of a database crash. It will be on all wikis on 20 November.
[40][41][42]
Changes later this week
There is no new MediaWiki version this week.
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
21 November at 19:30 (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
22 November at 16:00 (UTC). See
how to join.
Future changes
Language converter syntax will no longer work inside external links. Wikitext like http://-{zh-cn:foo.com; zh-hk:bar.com; zh-tw:baz.com}- must be replaced. You will have to write -{zh-cn: http://foo.com ; zh-hk: http://bar.com ; zh-tw:http://baz.com }- instead. This only affects languages with Language Converter enabled. Examples of such languages are Chinese and Serbian. This will happen next week.
[43][44]
Hi all, the Visual Editor keeps stripping whitespace from Infoboxes, like
here with the {{Infobox settlement}} template. Some editors like to line up the equals signs so that all the data lines up nicely, other editors don't like extra whitespace. Whatever the case, this is basically a local consensus issue. I reported this Visual Editor activity
here, which didn't get any useful attention, then opened a case at
Phabricator, but didn't get much in the way of help.
I'm not a technical type, so I'm totally ignorant on any technical aspect of this issue, but my general perspective is that VisualEditor shouldn't be using inexperienced editors as shills to perform maintenance tasks that you'd expect of experienced AWB users. A volunteer at Phabricator suggested that the templates could be modified so that the status quo is preserved. They wrote:
If you wish to change this instruction, you can go to [the template], and where it says
},
"format": "block"
}
you can replace it with
}
}
So, short story, if this is something that any of you find problematic, (and I might be a lone crusader here) it would require a fix. I guess that would mean identifying the problematic templates and then removing the format preference? This had been happening with {{Infobox film}} and {{Infobox television}} but AussieLegend was kind enough to fix those. Thanks,
Cyphoidbomb (
talk)
04:53, 19 November 2017 (UTC)
Removing "format": "block" is the equivalent to changing it to "format": "inline" for newly added infoboxes. It would be better to use something like "format": "{{_\n| ______________ = _\n}}\n" for an infobox. See
mw:Help:TemplateData#Custom formats. —
JJMC89 (
T·C)
07:13, 19 November 2017 (UTC)
I have no confidence in my technical ability. This is something I was hoping someone here could address, but maybe I'm alone in thinking that forcing Visual Editor users to make questionable (and potentially consensus-violating) changes without their knowledge is a bad thing.
Cyphoidbomb (
talk)
03:22, 22 November 2017 (UTC)
New user script needs testing
I've been working on
User:Enterprisey/reply-link, which is a user script that adds a "reply" link after signatures on a discussion page; when you click a link, a form comes up that lets you type your reply without having to go through an edit window. I'm looking for brave editors to test it out and find some bugs. This script could be very useful for new editors or people who don't want to wade through pages of markup to find the one comment they're replying to. (Indentation and signing happen automatically!) Thanks!
Enterprisey (
talk!)
09:42, 21 November 2017 (UTC)
Installed! This is a really interesting idea – one that might bridge the gap between those editors who want to contribute to discussions but are intimidated by wikitext editing. This was something that Flow attempted to resolve, but has not succeeded yet. This might just be an alternative solution.
(Replied using the new script.)Mz7 (
talk)
09:52, 21 November 2017 (UTC)
Great idea. There was a script back in the day that allowed in-line editing without having to load the whole edit window but it stopped working. Regards SoWhy10:16, 22 November 2017 (UTC)
I forgot to read that signing works automatically, kinda weird. Also, maybe you can add an option to auto-reload the page after adding a comment. Also, if I do reload but came to the page from a different section link, it will send me to that section again and not the section I was replying in. Regards SoWhy10:20, 22 November 2017 (UTC)
Good idea regarding the URLs – it's been a feature I've wanted since the beginning, but it turns out to be a little harder than I thought.
Enterprisey (
talk!)
10:42, 22 November 2017 (UTC)
ORES 'draftquality' data for English Wikipedia is now available in ores_classification table
Hey,
'draftquality' is a model in English Wikipedia which tries to predict if a new article is likely to be vandalism, spam, attack page or it's okay. "OK" in this context means it doesn't need new page patrollers immediate attention but still there might be reasons to get the page deleted (for example, notability). In a more wikipediholic term, this model predicts if articles need to be tagged with
SD#G3,
CSD#G10, or
CSD#G11.
This model has been available for a while now but recently we started to store these data in mediawiki and now it's available in its full capacity for patrollers.
I made two example queries,
this is the probability scores of draftquality classes, keep in mind that oresc_class values map to
the values defined in the extension. And in
this one, I made list of new page creations that are likely to be spam. You can increase the confidence threshold if you think it has too many false positives or you can change the oresc_class to 3 if you want to get page creations that likely to be vandalism.
Is it possible to suppress the edit link in a section? I'm working on template that includes a heading (on a small sister project) and that link opens the template, not the page section on which the link appears. Andy Mabbett (Pigsonthewing);
Talk to Andy;
Andy's edits19:01, 23 November 2017 (UTC)
I think you can also do it with wiki markup if it's produced by something instead of being directly in the source at the start of a line, e.g. {{#if:1|==Heading==}} (untested example).
PrimeHunter (
talk)
00:03, 24 November 2017 (UTC)
Table borders on mobile site
This has been bothering me for a little while now: tables are not appearing correctly on the mobile site. It is particularly evident when a table is not as wide as the screen on a device. There is an extension to the table filled with white space where the border of the table is half as wide as the rest of the table. It looks like this:
Yes, this is a known issue. It's a side effect of making tables work in when they are wider than the screen. There doesn't seem to be a good CSS way to handle this unfortunately. I've been playing with some JS hacks myself, but has some other downsides (reflows of the content). —
TheDJ (
talk •
contribs)
10:15, 24 November 2017 (UTC)
I fixed a wrong colspan in your example.
[46] Did that help? It looked right to me both before and after with a desktop browser at the mobile version but it varies how browsers react to wrong rowspan and colspan.
PrimeHunter (
talk)
10:23, 24 November 2017 (UTC)
Attention All Lua Programmers :)
I don't actually know what that means :) but having received some very kind advice / assistance
elsewhere, and having been directed here, can anyone do something like that? Happy Mid-Week Mash-Up, Everybody. Cheers! —
fortunavelut lunaRarely receiving (many) pings. Bizarre.18:36, 22 November 2017 (UTC)
Here is what I had in mind when giving the above-linked reply:
The request is to make a template that emulates the behavior described at
Template:Location map#User selection of multiple maps, where a selector can be used to switch the display between multiple images. The simplest form of such a template would be something like {{
multiple image display|imageA|imageB}} where the selector switches the display between
File:imageA and
File:imageB. Bonus points for making the template accept any number >=2 of filenames.
(
edit conflict) Thanks again
Tigraan, sorry I have now spread the same discussion over multiple boards-
WP:FORUMSHOP, anyone?! ;) - but another reply
back at WP:HD from the excellently useful
Frietjes has said something extremely simalr- and in my
sandbox, I can see how it should look- and this is basically exactly what I've been after! Brilliant. Obviously, I can't get the images into the bloody code, but-! :D that's natural justice, I suppose! —
fortunavelut lunaRarely receiving (many) pings. Bizarre.19:42, 22 November 2017 (UTC)
I think that's done it :) Haven't got time now to perfect it, but will do tomorrow- thank you VERY MUCH for everyone's help! —
fortunavelut lunaRarely receiving (many) pings. Bizarre.20:15, 22 November 2017 (UTC)
@
Frietjes: Yes mate, it's looking good at
Nicholas Exton :) tell me, you wouldn't happen to know how to move the box to the left-hand side and make the text flow around it by any chance? Thanks again! —
fortunavelut lunaRarely receiving (many) pings. Bizarre.15:52, 23 November 2017 (UTC)
One field is "NA" - ie no field applies (usually because the article is a redirect). Unfortunately that means the template incorrectly states "This redirect has not yet been associated with a particular anatomical discipline." (eg here
Talk:Peripheral nervous systems). Any ideas why? --
Tom (LT) (
talk)
00:07, 20 November 2017 (UTC)
Because that page is a redirect? :D The relevant code is
|TF_1_TEXT = This {{pagetype|{{{class|}}}}} {{#switch: {{lc:{{{field|}}}}}
|animal|embryo|gross|meta|micro|NA|neuro|organs|systems=has been classified as relating to {{#switch: {{lc:{{{field|}}}}}
...
|NA = [[:Category:Anatomy articles about NA|NA]
...
}}.
In it, you will see that the page type (which relies on the class, and presumably has some detection for redirects) sets the text "redirect" while NA sets the "has been..." text. Is this behavior undesirable? --
Izno (
talk)
02:16, 20 November 2017 (UTC)
Heh,
Izno, I understand that the page is a redirect. However on all pages with the banner you can set a discipline (Eg micro) and the talk page banner will say "has been classified as relating to microanatomy"... this works even on redirects
However, for all pages marked as discipline 'NA', the talk page banner will state "...has not yet been associated...". This is the bit which is confusing me. --
Tom (LT) (
talk)
03:45, 20 November 2017 (UTC)
I too was confused - when I wrote that part of the documentation in February 2015. I considered fixing the template, but decided to leave it alone and document the behaviour as best I could. Note particularly the lines
These categories are intended just as administrative categories. There's a big difference between unassessed (ie, haven't been assessed) and assessed as NA (ie, not relating to one of the set categories). --
Tom (LT) (
talk)
10:16, 21 November 2017 (UTC)
The template uses {{#switch: {{lc:{{{field|}}}}} ... }} and so when |field=NA is passed in, it's actually testing the value na which has no match (the tested options have the uppercase NA), so the default actions apply for each switch statement. I overlooked this when doing the doc in 2015. As a start, we can do this. --
Redrose64 🌹 (
talk)
11:44, 21 November 2017 (UTC)
Mediawiki question for Wikiversity copy of a Wikipedia article
The
radiocarbon dating article makes use of the {{chem}} template, mostly like so: 14 C. On Wikipedia, if you resize the browser window until one of the instances of this is at the end of the line, the 14 and the C are kept together; you don't get a line break between them. There's a copy of the radiocarbon dating article on Wikiversity as a WikiJournal of Science submission (
v:Radiocarbon dating), and in that copy linebreaks can sometimes show up between the 14 and the C. Is there some difference between the Mediawiki versions here and there that's causing this? Or is there some difference in the way {{chem}} is coded? Thanks for any help with this.
Mike Christie (
talk -
contribs -
library)
21:32, 25 November 2017 (UTC)
The two templates produce the same code with class="chemf nowrap". The difference in rendering is because the English Wikipedia has this in
MediaWiki:Common.css:
/* Prevent line breaks in silly places:
1) Where desired
2) Links when we don't want them to
3) Bold "links" to the page itself */
.nowrap,
.nowraplinks a,
.nowraplinks .selflink {
white-space: nowrap;
}
.nowrap pre {
white-space: pre;
}
I think that user:BetaCommand and User:DragonsFlight (sorry, I don't know the exact spelling) might have such tools. IIRC there's one on the toolserver right now. --
70.51.45.76 (
talk)
07:13, 26 November 2017 (UTC)
Currently
Tor users cannot create accounts due to the global block of Tor IP addresses, and the block message – "Your IP address is in a range which has been blocked on all wikis" (not specific to account creation) – does not mention that they can ask someone to create an account for them, or that they can request IP block exemption. Furthermore, the userpage link to the blocking Meta user is broken. Would it be possible to indicate to them that they can request an account be created for them, in a less oblique way than the contact address for stewards being shown? (The block message is not localized here, since I can't find it in a search of the MediaWiki namespace.)
Jc86035 (
talk)
05:12, 26 November 2017 (UTC)
Is it possible to make a separate message for account creation, or could the display of different text be achieved with a {{#switch:{{PAGENAME}}|…}}?
Jc86035 (
talk)
06:23, 26 November 2017 (UTC)
@
Jc86035: It doesn't show up in Special:Search because the page hasn't been modified from the MediaWiki's default message, thus the page hasn't been created yet. Graham8707:51, 26 November 2017 (UTC)
Blocked proxy messages have always been ugly, but believe it or not they've actually improved. I'd suggest we need a major review of these block messages, by the people at
WT:UTM. --
zzuuzz(talk)14:42, 26 November 2017 (UTC)
Filtering automated edits from my user contributions?
When I use some automated tool to do something like close an AfD, it'll often end up performing dozens of edits on my behalf (updating talk pages, removing templates, cleaning up orphaned links, etc). When I look at my user contributions list, that's all just noise. Is there some way to get a look at just the manual edits I've performed? --
RoySmith(talk)15:47, 26 November 2017 (UTC)
See
User talk:Doug Weller/Userboxes - the relevant mass message log (I think) seems to suggest it wasn't sent and also has entries saying things such as "User:Altruism/My Userboxes was skipped because target was in a namespace that cannot be posted in". In fact that page didn't exist until it got the message.
Doug Wellertalk17:23, 26 November 2017 (UTC)
The traceroute tool linked from the toolbar of an IP contributions page is broken, it chews for a bit and then just redirects to the homepage of dnstools. Can we replace the link with a tool that works, or remove the link? I noticed this months ago, and the fact it hasn't come up yet makes me think that maybe we don't need a traceroute link.
Ivanvector (Talk/Edits)
00:26, 27 November 2017 (UTC)
See
this request I made a bit ago. They apparently changed how their URLs were formatted; now it seems they don't accept direct links. I would support either removing it or just adding something more helpful (whether it's traceroute or something else). I have no suggestions for what that might be though.
Nihlus01:02, 27 November 2017 (UTC)
Can anyone tell me if there are any means of counting the links used globally in en.WP citations by domain (or rather sub-domain), excepting archiving sites? I hope to use such data to ensure that the top, say, two or three hundred top-cited domains are included to populate
my script that ensures that parameters are correctly used within citations. -- Ohc ¡digame!13:51, 19 November 2017 (UTC)
Is there an instructions/warnings page anywhere for this function? As I found out a long while ago, changing the content model of a page to JS or CSS automatically protects it against edits by users other than admins (and if the page in question is in userspace, the owner of said userspace). Correspondingly, changing the content model of a User.js page to wikitext loses that protection (which User.js pages have by default).
Jo-Jo Eumerus (
talk,
contributions)
15:35, 26 November 2017 (UTC)
I had a look through the archives and couldn't find anything about this, so I'm hoping someone here can help. I have a "Page size" link in my sidebar; I no longer recall where I got it from. It used to work fine, but a month or two ago I noticed that it only gives partial information. E.g. on
ice drilling I get:
Near the bottom of
User:Mike Christie/vector.js, you should see two javascript files from User:Dr pda. I recommend removing "prosesizebytes.js" and replacing it with "prosesize.js". That works for me. You can see the full line of code that calls the script near the top of
User:Jonesey95/vector.js. As always, use User scripts at your own risk. –
Jonesey95 (
talk)
19:43, 27 November 2017 (UTC)
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
You can now test the new
advanced search function beta feature on mediawiki.org. It makes it easier to use some of the special search functions that most editors don't know exist. It will come to German and Arabic Wikipedia this week. It will come to more wikis later.
[47]
You can now upload large files with the
Internet Archive upload tool. Previously you could not upload files larger than 100 MB.
[48]
The
new version of MediaWiki will be on test wikis and MediaWiki.org from 28 November. It will be on non-Wikipedia wikis and some Wikipedias from 29 November. It will be on all wikis from 30 November (
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
28 November at 19:30 (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
29 November at 16:00 (UTC). See
how to join.
I think I have completed a minimal set of edits to both templates in order to preserve the talk page text but remove the automatic categorization. Null edits to about 4,000 pages will be required in order to empty the categories. –
Jonesey95 (
talk)
15:08, 27 November 2017 (UTC)
I did not revert your edits. Please check the edit history.
You (or others) are welcome to adjust the templates in a way that meets the needs of the discussions linked above. Your changes re-added automated categorization to the templates, which was undesirable. Perhaps |category=no is the correct, albeit exceptional, solution in this case, since complete removal of automated categorization is the exceptional outcome that is desired here. –
Jonesey95 (
talk)
01:15, 28 November 2017 (UTC)
Just a guess here, but there are 646 transclusions of {{V0.5}} and approximately 646 pages listed under "T" in ‹The
templateCategory link is being
considered for merging.›Category:WikiProject banners with formatting errors, so I'm guessing that {{WP1.0}} is probably OK as is, but some small change needs to be made to {{V0.5}}. Pages sorted under "T" in that error category are described as "a banner using |category= in article talk space", which is not enough detail for me to understand how to fix the problem. Maybe someone else can grok that description.
Hope you all had a wonderful Thanksgiving! Hello fellow editors! I am interested in installing
mw:Extension:PdfHandler on my mediawiki.
Problem is that one of the prerequisites is Ghostscript:
https://www.ghostscript.com/download/gsdnld.html I am not sure how to install this on my mediawiki. Any assistance would be most helpful. Peace!
@
Imaginelenin: This is a help page for Wikipedia. We don't have PdfHandler. The MediaWiki software has a help page at
mw:Project:Support desk. Ghostscript is general software and not made for MediaWiki. I don't know how PdfHandler interacts with it but the first thing I would try is to just install Ghostscript normally on the same computer and see whether it works.
PrimeHunter (
talk)
21:47, 27 November 2017 (UTC)
It seems that in Minerva skin there is a selector that nulls the border of a cell belonging to the row that is the last child.
Ruslik_
Zero21:04, 21 November 2017 (UTC)
TheDJ yes, I know that the trigger for the bug is the usage of two classes, but it's not clear to me why this is a problem only on mobile. I have a
version of the module in the sandbox without the infobox class, which would also resolve the issue. I don't know if there are any other boxes using both classes at the same time.
Frietjes (
talk)
14:32, 24 November 2017 (UTC)
We really shouldn't be mixing classes that have conflicting styles like that though. This creates behavior that developers won't support. It's like saying "Please paint this car red and green", expecting the car to be delivered with 50-50 red and green striped, and then being surprised because someone painted it yellow :) —
TheDJ (
talk •
contribs)
18:59, 24 November 2017 (UTC)
TheDJ tried adding it to
Module:Album ratings and it does nothing for mobile view (as far as I can tell). my guess is that there is some special javascript or something for the infobox class which moves the infobox down and allows it to float right, where all other floating content is disabled on mobile. figuring out exactly what is going on is beyond my time budget at the moment.
Frietjes (
talk)
17:09, 28 November 2017 (UTC)
Auto-signed template(s) duplicating signature
Template:Welcome-image is automatically signed - it includes the ~~~~ right in the template - but this causes the username to be duplicated when Twinkle is used. Likewise, a non-auto signed template like
Template:Welcome-delete does not cause the the duplication. See
Special:PermanentLink/812662639 for comparison. I'm not experienced with templates and don't want to accidentally break the script, but I think the ~~~~ should be removed from this and any others that include it.
Home Lander (
talk)
04:05, 29 November 2017 (UTC)
A bit of a puzzle
I don't know if this is the right place to ask this, but I have a bit of a puzzle. A few days ago I was thanked for an edit I made, but the user who sent them doesn't seem to exist. AFAIK you have to be logged in as a registered user to thank someone, but when I followed the link there was no user page there, nor any associated talk page, and I couldn't find anything at Special Contributions either. Any idea how this might have happened? I'm stumped.
Xyl 54 (
talk)
23:49, 27 November 2017 (UTC)
@
Xyl 54: Someone can have an account with no user page, talk page, or contributions. When you went to his purported user page did it say "User account 'Kdslfjsdofisjdf0s9djflksdf' is not registered."? ―
Justin (koavf)❤
T☮
C☺
M☯00:45, 28 November 2017 (UTC)
Yes, that was the one. I checked the link again (
it was here); it didn't say he(?) wasn't registered, but looking at the “verify that “Bernjam” exists” link
shows an account created on 30 June. So you're saying it's never been used except this once? Curiouser and curiouser! (Lucky me!). Anyway thank you for shedding some light on this,
Xyl 54 (
talk)
23:41, 28 November 2017 (UTC)
@
Xyl 54: never used to edit, some readers make accounts just for reading, to use the watchlist (and nerds might use the watchlist RSS feed :D ). Full disclosure: I've used that before... —
xaosfluxTalk03:22, 29 November 2017 (UTC)
I want to fetch the "user" table Wikipedia has. I know Wikipedia limits the fields we can view but the ones open are just fine for me.
I understood such large queries will not work in "QUARRY" and also when looking in the
dumps I could not find any SQL or XML dump of that table. Do you know how can I get my hands on it?
REVISION TABLE
I am also looking for "revision" table. Again- the same problem with large queries :)
So I looked in the dumps and found the following:
when searching "revision" I got to 2 dumps that I think might help me- "Recombine first-pass for page XML data dumps" and "First-pass for page XML data dumps"
What is the difference between them? they both state at the beginning "These files contain no page text, only revision metadata."
Also, when looking into these sections the files there are divided to 3: stub-meta-history, stub-meta-current, stub-articles.
What is the difference between them in the means of what fields they include? of what sections in the site and of what time-frame?
My goal is to fetch for every user his whole edit history including- on which pages, how long was the edit, and the date of the edit.
Sorry if this post is not in his appropriate place, I'm new and was forwarded here from the QUARRY discussion.
There is no such XML dump of the user table. For queries against it, you'll want to use Quarry or the labs replicas directly. Or maybe the API. As far as revisions go: those options are all also available, as are the XML dumps. "stub-meta-history" is probably the one you want.
FACE WITH TEARS OF JOY[u+1F602]19:40, 29 November 2017 (UTC)
I've been trying to create some tables on a Wikiproject page and finding it frustrating because I can't use Visual Editor. I've been trying to work out why VE isn't enabled on VE and have been looking at ways to enable it.
From what I can find out, the reason it isn't enabled is that it shares a namespace (Wikipedia:) with many other kinds of pages which don't play well with VE and that the only way to enable VE would be to move Wikiproject pages to another namespace (e.g Wikiproject:).
Some questions:
Is the information I've found correct?
What would be the process of moving Wikiproject pages to their own namespace?
What process should be used to get community feedback on whether this should happen?
Can anyone see any downsides of making VE available on Wikiproject pages?
As far as I understand, VE is not for talk-like pages and developers won't adapt it as Flow is supposed to serve talk spaces. Wikipedia: is a mixture between two purposes which makes it problematic to use VE there. Perhaps you can convince developers to allow it on WPspace with the understanding that people won't use it on talk-like pages there.
Jo-Jo Eumerus (
talk,
contributions)
14:59, 20 November 2017 (UTC)
Thanks
Jo-Jo Eumerus, I'm sorry for not being clear, I didn't mean for Wikiproject talk pages, but for the organisational pages that require tables and other parts that are hard to do in source editor.
John Cummings (
talk)
15:52, 20 November 2017 (UTC)
Jo-Jo Eumerus, this is because the Wikipedia: namespace includes pages that have issues when editing with VE, my workaround for this is to move all Wikiproject pages to a separate namespace so that VE could be enabled. I'm trying to understand the process to make this happen.
John Cummings (
talk)
16:41, 20 November 2017 (UTC)
That would be a proposal to create an entire new namespapce, where we put "pages" is up to us. Note, the "Wikipedia:" namespace is already the canonical "Project" namespace. —
xaosfluxTalk18:16, 20 November 2017 (UTC)
Not sure what all the fuss is about. It is trivial to get visual editor working in the Wikipedia namespace without any development at all, e.g.
[50] . Just append ?veaction=edit to the URL. It would take a few minutes to create a usercript to make a button show up for those that want it in that namespace.19:17, 20 November 2017 (UTC) — Preceding
unsigned comment added by
197.218.84.101 (
talk)
I believe that the result depends upon whether you have the beta feature for the 2017 wikitext mode enabled. If you do, then you don't get wikitext; you get raw HTML, which is A Very Bad User Experience. This is a particular problem for the use case in question, because the people who most struggle with wikitext are the same ones who benefit most from a visual editing experience and from not needing to learn two different toolbars to be able to use a talk page.
At the English Wikipedia, the basic sticking point is that the team is never going to enable VisualEditor for any namespace that contains ANI (or any similarly large and high-traffic discussion page).
Whatamidoing (WMF) (
talk)
18:34, 21 November 2017 (UTC)
I just updated to Firefox 57.0, and it does not have NoScript available on it. Is this a fluke or did Mozilla do away with NoScript? There appears to be no explanation.
— Maile (
talk)
14:07, 15 November 2017 (UTC)
@
Xaosflux,
Isaac,
Redrose64, and
SoWhy: Still somewhat confused. One of the above links originally had a message that a new NoScript would be ready by the end of last week, but it wasn't. This is the thread I've been following:
1, and it's getting so I don't know what they're talking about. Can somebody give this to me in simplified terms? Is the Firefox 57.0 Quantum release of NoScript still in the works? When and if it's ever ready, will we Firefox users have to find it and reinstall, or will it likely be an option we'll find under Firefox Tools/Add-ons?
— Maile (
talk)
16:47, 21 November 2017 (UTC)
@
Maile66: The Quantum release for NoScript is available for download
here. Try downloading it manually if the addon is not updated automatically (using the "direct download" link). If you need more extensions, you can always downgrade to Firefox 52.5, which is the
ESR supported until Mid-2018. Regards SoWhy16:52, 21 November 2017 (UTC)
I see a lot of links to update stuff. On add-ons "Find a Replacement", it redirects to
Double NoScript or NoScript 10, whichever they're calling it. So ... it's either (a) isn't ready yet; (b) was downloaded by some user who thinks it is a disappointment, or (c) NoScript just working again - all by itself - on some user's browser and is wonderful. Again, this is confusing. NoScript and Firefox need to get their coordination together.
— Maile (
talk)
19:25, 21 November 2017 (UTC)
You can always install Firefox ESR or
Waterfox (you will probably want to, given the direction in which Mozilla is heading), it will load your Firefox profile automatically, no need to import bookmarks, passwords etc. There is also
Pale Moon, but I don't remember if the process is automatic.
93.142.92.15 (
talk)
03:51, 30 November 2017 (UTC)
An editor tried to submit a draft for review following the instructions. They said it didn't work. I tried myself, hoping to see some sort of error message but I don't. Can someone help identify the problem?
Isn't opening it up in "edit" view going to do that for you? Are you talking about doing that for a mass amount of pages at once? I guess knowing your end goal would be helpful here.
Killiondude (
talk)
18:06, 28 November 2017 (UTC)
I meant programmatically, such as with
WP:Parser functions, such as the example given for "((#IFEXISTS:(((X)))|A|B))" ; if coded onto a page, can wikicode (preferentially) or LUA determine the existence of the use of the REFTAG system on a page? --
70.51.46.255 (
talk)
09:31, 1 December 2017 (UTC)
Building mobile-safe templates help
When developing templates, I often meet effects in mobile view that should be taken care of. For example: I know that mobile has no show/hide button (it's "show" always). Other effects I can only learn by trial and error. Is there a place or process where this knowledge is available? Like, WP:MOBILETEMPLATES, or H:MOBILEEFFECTS? -
DePiep (
talk)
19:52, 1 December 2017 (UTC)
Quick notice / canvassing: Everyone who wants to see
WP:AALERTS grow in the future should support this. Potential benefits include better support / better stability / more frequent code updates, more workflows covered, and deployment to other Wikipedias in other languages. And it could possible even be better integrated with watchlists and stuff. Headbomb {
t ·
c ·
p ·
b}21:10, 1 December 2017 (UTC)
As I mentioned to
Ronhjones, it would be nice if the notification was extended so that we'd be able to thank editors how do tasks that appear on
Special:Log. If there's no technical reason that this cannot be implemented, please do so. Thanks! nagualdesign02:05, 25 November 2017 (UTC)
I was going to say Phabulous! but I see that it was first requested almost 4 years ago and is yet to be implemented. Oh well. Thanks for letting me know. nagualdesign21:14, 25 November 2017 (UTC)
Hi all, I'm hoping someone will be able to help me with a question on creating specific pages. Some background, I'm trying to confirm that a recent proposal by ArbCom is technically possible. Basically, I want to find a way for a user to create a subpage of a page which doesn't exist which is preloaded with a form. As an example, I want a user to be able to create the page
Wikipedia:Arbitration/Requests/Case/Example/Request where they chose what Example will be. The /Request page they create should be preloaded with
Wikipedia:Arbitration/Requests/Case/preload. I had thought about using an
inputbox but, from what I can see, it can't be used to create a subpage of a custom named subpage. Does anyone know how this goal could be achieved? Thanks, Callanecc (
talk •
contribs •
logs)
02:13, 2 December 2017 (UTC)
<InputBox>
type=create
preload=Wikipedia:Arbitration/Requests/Case/preload
placeholder=Enter your case here
prefix=Wikipedia:Arbitration/Requests/Case/
editintro=Template:Arbitration request editintro
buttonlabel=Create new case
</InputBox>
Callanecc wants /Request to automatically be added to the name given in the input box. I don't think that is possible. You can use default= to prefill the box but the user has to keep it while writing the case name:
I use an {{Unbulleted list}} in a data entry of {{infobox}}. In mobile view, it appears as being a bulleted list with the bullets hidden (that is: indented with whitespace). In regular text this could make sense, but in this tabular list the indenting is undesired (one would expect and want left-alignment in the data column). Interestingly, when the page is loaded, the bullets do actually appear for a moment. Is there a good way to solve this?
I have a preference set to preview an article lead when mouse-hovering a wikilink. OK. Nicely, the title is bolded in that preview. Now this works with
Copper, but not with
Mercury (element), that is
Mercury, the title word "Mercury" is not bolded.
The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
Now that we are in donations season, we get a lot of queries at OTRS from people who respond to the banner, make a donation, and are surprised that the banner doesn't go away.
While I typically tell them that they can register an account and turn off the fundraising banner, some don't wish to do that, and I would like to explain why it isn't the case that the banner automatically goes away.
I think the answer is that we would have to place a cookie to suppress the banner and we prefer not to do that, but I'd like to provide a more comprehensive answer. Is it easy to do and we just don't choose to do it because we'd rather that they register?
Is it possible to do with cookies, but we have a policy not to use cookies except for people who register?
If someone could help me craft a standard response that explains why the banner doesn't go away and whether this is something that we cannot do or simply choose not to do, I'd appreciate it.--
S Philbrick(Talk)21:15, 1 December 2017 (UTC)
Off the top of my head, they should disappear. We do send (some) cookies to anonymous users, just not as many as for logged in. In testing just now anonymously, when I dismissed a banner I definitely got some "centralnotice_hide_fundraising" cookies issued to me. Maybe this is a bug? I have a feeling @
Whatamidoing (WMF): might know ;-). Alternatively, ask on
meta?
FACE WITH TEARS OF JOY[u+1F602]01:29, 2 December 2017 (UTC)
@
Sphilbrick: that being said - the client has to accept our cookie to actually dismiss it. Some browsers (like any in 'private browsing mode') don't keep cookies after you close them. —
xaosfluxTalk15:00, 2 December 2017 (UTC)
After thinking about it a little more (which I probably should've done earlier), I realize that while we do get a number of queries to OTRS (and we are backlogged again so struggling to catch up), if they were no cookie-based suppression of banners occurring at all we would have been absolutely flooded as opposed to the volume we've been seeing. I'm not convinced we are only seeing the small number of people who may have cleared out the cookies for other reasons or may have turned off accepting cookies or some other issues, such as the one suggested by xaosflux/
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
Archive a url with pdf
Hello. Do you know a site where I can archive a url that is an online pdf? I was using www.webcitation.org but sometimes is not working (the problem is not the robot.txt).
Xaris333 (
talk)
14:02, 2 December 2017 (UTC)
There is a very long-standing problem whereby pages reached via a redirect are hours or days stale, when the directly-accessed page is up-to-date. It seems to be nothing to do with local caching. It seems to be a problem at the Wikipedia end. It would be nice if this could be fixed.
109.145.27.241 (
talk)
00:52, 2 December 2017 (UTC)
Durable examples do not exist for obvious reasons. You can create and test examples as well as me or anyone else in the obvious way, i.e. by making a change to a page and the comparing it as accessed directly and as accessed via a redirect. Results are not always consistent, but in some cases the access via redirect is stale by many hours.
109.145.27.241 (
talk)
03:40, 3 December 2017 (UTC)
The issue only affects users who are not logged in. When logging out I currently see "This page was last edited on 3 December 2017, at 03:41" at the bottom of
WP:VPT. The displayed content is indeed that version although there have been 13 edits since. I see the latest version at the real title, or when logging in and viewing the redirect.
PrimeHunter (
talk)
21:27, 3 December 2017 (UTC)