Jump to content

Wikipedia:Village pump (technical)/Archive 16

From Wikipedia, the free encyclopedia

This is the current revision of this page, as edited by MalnadachBot (talk | contribs) at 06:52, 19 February 2023 (Fixed Lint errors. (Task 12)). The present address (URL) is a permanent link to this version.

(diff) ← Previous revision | Latest revision (diff) | Newer revision → (diff)

See on our watchlists which pages are redirects

Wouldn't it be nice...

...if we could see on our watchlists which pages are articles and which pages are redirects ? (Sorry if it has been already discussed, I don't usually come here.) Have a nice day, Rosenknospe (talk) 10:58, 15 January 2008 (UTC)

This is an open bug in the bugzilla database: http://bugzilla.wikimedia.org/show_bug.cgi?id=2872 --TheDJ (talkcontribs) 11:56, 15 January 2008 (UTC)
Thank you. Forgive me for being a noob, but does that mean this functionality will be implemented some day ? Rosenknospe (talk) 11:59, 16 January 2008 (UTC)
Well, that depends on whether any of the developers - who all work on the software in their spare time - decide it should be prioritised. As categories now indicate which articles are redirects, I imagine that it wouldn't require too much work. Warofdreams talk 14:41, 17 January 2008 (UTC)

Infobox date format problem

Hi. I'm trying to fill in the date category on a template on this article which I've created, however, whichever format I use the article displays {{{date}}}. Can someone please tell me what I'm doing wrong? Thanks, TicketMan - Talk - contribs 13:02, 17 January 2008 (UTC)

The field is case-sensitive, so you needed "Date =", rather than "date =". I've fixed it. Warofdreams talk 13:12, 17 January 2008 (UTC)
Thanks Warofdreams! :-) TicketMan - Talk - contribs 13:16, 17 January 2008 (UTC)

Request for installation of Labeled Section Transclusion

Cross posted to Village pump (proposals). Please discuss there.

I figured this was the right place to get this discussion going. This is a request that the Labeled Section Transclusion extension be installed on the english wikipedia. This new extension would (it appears) have no conflict with existing extension and have several uses, especially for intermediate to advanced editors.

This extension allows for the selective transclusion of only part of a page or template (defined by <section> tags). Although this can partially be done with <includeonly> and <noinclude> tags, these two do not provide the full capabilities as provided by this extension. In my own editing experience, I have found several circumstances where these tags would have been useful:

  • In creating a Potal, the tags could be used to show a constantly updated 0-section of an article, minimizing the need for manual portal updates.
  • Also in a Portal, these tags would reduced the need for numerous sub-pages. All featured articles, pictures, or whatever could be listed on one page with selective inclusion tags.
  • Templates with intricate features that only create selective display on specific pages/prefixed pages could have their code drastically simplified and reduced.
  • "Configuration sub-pages" used by many community voting and/or maintenance centers could also be drastically simplified.

These are only a few examples of how this new extension could be used (please feel free to add more). I'm not all "hip" on the mediawiki lingo but from my basic coding knowledge, this does not seem to difficult to accomplish. Happy Editing! --omtay38 14:53, 17 January 2008 (UTC)

I would support that- many user pages have copied and pasted a section of editing rules- if only this could be transcluded. On talk pages it would be wonderful to transclude a bit of policy, or a chunk of the page under discussion. In writing geographical articles about a group of neighbouring villages- I could transclude a paragraph about the geology- it would also allow the transcluding of a paragraph that contained all the basic reference for the area. Think of the enhanced user experience and the disc space we could save! ClemRutter (talk) 19:49, 17 January 2008 (UTC)

Note: Cross posted to Proposals. Please discuss there. --omtay38 21:00, 17 January 2008 (UTC)

Something went wrong here. I moved Pat Ross (disambiguation) to Pat Ross. However, I somehow lost the edits there...they're not appearing in either location. It was previosuly a disambiguation page wtih two names on it. hbdragon88 (talk) 05:39, 18 January 2008 (UTC)

I undeleted and restored the disambig page as the latest revision. Feel free to finish moving however you were hoping to. — Carl (CBM · talk) 05:43, 18 January 2008 (UTC)
That's strange...it looks OK now, I guess. But before, Pat Ross was a looping redirect page. Thanks CBM. hbdragon88 (talk) 06:08, 18 January 2008 (UTC)

Category bot

Hi. I'm looking for bot script that automatically scans and moves pages from categories marked with {{categoryredirect}} to the correct category. I'm planning to run such a bot on id.wiki. I asked user:selket to share the bot script, but still no answer. He owns user:Selketbot which the only bot I know so far that run such a program. Can anyone help me? borgx (talk) 00:31, 18 January 2008 (UTC)

third door on the left, WP:BOTREQ βcommand 16:37, 18 January 2008 (UTC)

Many templates have "v • d • e" links in the upper left corner (Example). Would it be possible to show those links only to logged in users? To others, it's only confusing I think. --Apoc2400 (talk) 17:44, 18 January 2008 (UTC)

Not possible within the MediaWiki syntax. As a workaround, I guess one could theoretically edit the Tnavbar template to include a <span class="tnavbar">...</span> span. Then, add some javascript to common.js to hide the tnavbar spans when the user is not logged in. Theoretically. • Anakin (contribscomplaints) 20:56, 18 January 2008 (UTC)
I can understand wanting to supress the (e)dit link, but why hide the (v)iew and (d)iscuss links? EdokterTalk 22:49, 18 January 2008 (UTC)

missing file history

here is a strange case of a huge GIF with no visible upload history: Image:Granim.gif. dab (𒁳) 19:11, 18 January 2008 (UTC)

um, it has now disappeared, with no deletion log. This is passing strange.... dab (𒁳) 19:45, 18 January 2008 (UTC)
Deletion log is there:
  • 20:44, 18 January 2008 Dbachmann (Talk | contribs | block) deleted "Image:Granim.gif" ‎ (content was: '{{unknown}}' (and the only contributor was 'Dbachmann')) (Restore)
EdokterTalk 22:52, 18 January 2008 (UTC)

#ifexist:

Does this only follow branches that are necessary? I ask because {{archive list}}, which uses 100 #ifexist: calls, has not been used in archive boxes since the beginning of last month because of "server strain". I know that #if: had been changed in this way, but I was not sure if #ifexist: has yet. MrKIA11 (talk) 19:29, 18 January 2008 (UTC)

Under the new parser (soon to go live), it will. --Brion VIBBER (talk) 20:34, 18 January 2008 (UTC)
Is there some place where it says when it will go live, or do you know? MrKIA11 (talk) 20:36, 18 January 2008 (UTC)
m:Migration to the new preprocessor gives a date of 24 January. Tra (Talk) 20:42, 18 January 2008 (UTC)
January 24 is the goal. If you'd like to help test out the new preprocessor, please visit that page for details; bug testing is always needed. Cheers. --MZMcBride (talk) 20:58, 18 January 2008 (UTC)

Broken page?

Wikipedia:Tutorial (FormattingE67/sandbox - I'm seeing both text (and a history tab with one entry) and the message "Wikipedia does not have a project page with this exact name". It's not the page per se that bothers me - I found it via a search that virtually no one else is likely to do - but rather that the software really shouldn't be confused, as it seems to be. -- John Broughton (♫♫) 22:35, 18 January 2008 (UTC)

The page exists... but I've deleted it. Someone just pasted the code from the MediaWiki:Noarticletext message into the page. EdokterTalk 22:45, 18 January 2008 (UTC)
Thanks. I'll try to remember next time to not fall for the "fake message" trick. -- John Broughton (♫♫) 20:40, 19 January 2008 (UTC)

Judie Brown

Is it just my computer, or is every computer that, when accessing the Judie Brown page, the "Judie" in it gets highlighted in yellow? This is very bizarre. Editorofthewiki (talk) 22:27, 19 January 2008 (UTC)

Hey, it just did it after I saved this! Editorofthewiki (talk) 22:28, 19 January 2008 (UTC)

It just stopped. Wow. Editorofthewiki (talk) 22:37, 19 January 2008 (UTC)

I can't see any highlighting. What browser are you using and do you have any extra toolbars or extensions installed? Tra (Talk) 00:42, 20 January 2008 (UTC)
Never mind, it stopped. BTW, I use Internet explorer. Editorofthewiki (talk) 00:58, 20 January 2008 (UTC)
Is it possibly a result of searching for that word? Some browsers will highlight all occurrences of a word you searched for. --ais523 12:26, 20 January 2008 (UTC)

Bugs with + tab

I'm coming here to discuss a couple issues I've found with the + tab, to see if it's something I can fix by changing my CSS or a bug in the software.

The first annoyance is that #ca-edit doesn't get the class istalk when the magic word __NEWSECTIONLINK__ is used to add the comment tab on a non-talk page; I'm using that class to visually combine the edit and + tabs in my monobook.css. The second is the fact that there is no + tab on non-talk pages' "add comment" edit page; from the tabs, it appears one is viewing the page, not adding to it.

Does anyone have any suggestions/comments on this? Tuvok[T@lk/Improve] 06:04, 20 January 2008 (UTC)

Working WYSYWIG?

See here. Can we implement it? Easy table creation and editing, for starters, is a much wanted feature. --Piotr Konieczny aka Prokonsul Piotrus| talk 17:52, 15 January 2008 (UTC)

It appears to be running FCKeditor (I'm not sure which of the extensions mentioned at mw:Extension:FCKeditor it's running). That still has some issues; I filed some bugs against FCKEditor on their bug tracker a while ago and as far as I know they're still open. --ais523 18:01, 15 January 2008 (UTC)
Please read user:Cacycle/wikEd help#WYSIWYG and What you see is Wiki - Questioning WYSIWYG in the Internet Age why a true WYSIWYG editor might not be such a good idea in general (although the table editing is a cool feature). Сасусlе 06:14, 16 January 2008 (UTC)
I've tried those on other wikis, there not perfect and when it come to complex wiki code it just gets in the way. Also its not a proper extension you still have to do some hacking or else the preview button won't work. --Chris 09:49, 20 January 2008 (UTC)

Watchlist slow to load

My watchlist is very slow to load today - other pages aren't. Is there some sort of a fault? DuncanHill (talk) 14:36, 20 January 2008 (UTC)

Can't answer the question, unfortunately, but I have noticed a certain slowness. I thought it was just my browser being unusually slow today. Tuvok[T@lk/Improve] 15:06, 20 January 2008 (UTC)
I've noticed it too, and not just with the contributions pages. Any thoughts? Caknuck (talk) 18:13, 20 January 2008 (UTC)
I'm also getting more error messages today than I have in all the rest of my time here, nearly two years. Are the servers trying to blow up? Tuvok[T@lk/Improve] 18:20, 20 January 2008 (UTC)

Ipblock exempt proposal

Wider audience for commenting requested...


A proposal has started to allow established or trusted editors to edit via Tor, or other anon proxy. This discussion is located at

talk page

The proposed policy in its “needs to be worked on” form is located at

project page Mercury at 20:42, 20 January 2008 (UTC)

#ifeq adding a paragraph when 'if not'

At Template:Citytrain Station I have a simple parser function (which I took inspiration from Template:Infobox Company's intl element): If the element closed in a Citytrain Station article equals 'yes', then a black table row is inserted to show the train station is closed. If not, then nothing. Works great here (and similarly here when it is not a closed station). However, I later find that when the template is used in a closed station page without a photo/image in it (the element immediately after/under), such as here, it inserts <p><br /></p> outside of the nested table, being before the second (bottom) one it is within, causing a large whitespace. Thanks in advance those who can take a look at this and see what needs fixing.  SEO75 [talk] 03:49, 21 January 2008 (UTC)

Fixed now? Gimmetrow 05:07, 21 January 2008 (UTC)
Fantastic, thank you very much! :)  SEO75 [talk] 05:53, 21 January 2008 (UTC)

Deletion restrictions for pages with long histories

A couple times a year somebody does something like trying to delete the Wikipedia:Sandbox, which reaaalllly bogs down the server due to the large number of revisions. While there are warnings about this, I'm hacking in some limits which will restrict such deletions to keep the system from falling over accidentally. --Brion VIBBER (talk) 23:12, 16 January 2008 (UTC)

Thanks Brion, I was starting to get the DTs from wiki-withdrawal! —Elipongo (Talk contribs) 23:16, 16 January 2008 (UTC)
Limit is initially set at 5000 revisions. Error message is ugly at the moment, will be improved shortly. --Brion VIBBER (talk) 23:17, 16 January 2008 (UTC)
Thans, Brion. Do we need to make any plans for the unlikely event one of these pages does need to be deleted? Somebody on IRC mentioned something about bcrats-only. Or maybe it'd be rare/never enough to warrant grabbing developer attention. – Luna Santin (talk) 23:26, 16 January 2008 (UTC)
Bureaucrats are there to set user flags and rename accounts, nothing more. There is no need for such a page to ever be deleted. (And if it did, there would be no urgency to - a dev can just be asked) Majorly (talk) 23:31, 16 January 2008 (UTC)
Regardless of who does it it will bring everything crashing down. I think the limit could probably be raised to 10000, I have done a 8697, a 5133, and a 5098 before, without destroying anything. Prodego talk 01:45, 17 January 2008 (UTC)
Over on the mailing list the discussion is whether it would be a good idea to delete/recreate the sandbox on a daily basis to keep it from getting too unwieldy. —Elipongo (Talk contribs) 23:33, 16 January 2008 (UTC)
Well I suggest you tell the mailing list to bring it here so the whole community can be involed in the discussion. Ryan Postlethwaite 23:59, 16 January 2008 (UTC)
(Point taken, but I submit that any given talk page is no more "the whole community" than that mailing list is... —Steve Summit (talk) 00:27, 17 January 2008 (UTC))
Done! It's just that for a while we couldn't discuss it here... :-) —Elipongo (Talk contribs) 00:25, 17 January 2008 (UTC)
Drat, that means we can no longer witness the bold madness of deleting Wikipedia:Votes for deletion.  ;-) As I recall, that locked the database for more than 15 minutes. Dragons flight (talk) 00:38, 17 January 2008 (UTC)

This new limit is not entirely clear to me: has it been made impossible to delete articles with more than 5000 revisions, or will we see an additional warning before the article is deleted? I understand that we need to keep Wikipedia from falling over, but my concern is that this might make deleting BLP violations from high-profile articles difficult to impossible. Is that a valid concern, or is there no cause for concern? AecisBrievenbus 00:58, 17 January 2008 (UTC)

The software won't let you delete anything over 5000 revs now. If there's a BLP violation that needs deleting from the logs, then until revs deleted gets commited, we'll have to go through oversight. Not ideal, but at least it's a solution. Ryan Postlethwaite 02:31, 17 January 2008 (UTC)
Will this limit be hard-coded into every download of MediaWiki now? If so, is it configurable? The thing is, I'd much rather not have this feature on my wiki. :) I can see how it's useful for Wikipedia, though. Is it just for en.wp? Tuvok[T@lk/Improve] 05:13, 17 January 2008 (UTC)
Configurable in localsettings. The permission is bigdelete; assign it however you want :) – Mike.lifeguard | @en.wb 13:15, 17 January 2008 (UTC)
And by default, the limit is not in place. That is, on mediawiki installs that don't explicitly specify a limit, you can delete as large of articles as you want. AmiDaniel (talk) 17:42, 17 January 2008 (UTC)
Thanks, both of you! Good to know it's not on by default, and to know it's configurable in case I need it somewhere in the distant future (my entire installation has only 3,000 revisions). Cheers! Tuvok[T@lk/Improve] 22:05, 17 January 2008 (UTC)

Copied over from ANI

Well, thank goodness someone finally got around to implementing this limit! A good side-effect will be that talk pages and user pages with more than 5000 revisions will need to be properly debated (well, the talk pages debated at least) and then deleted, rather than deleted on the quiet with a "right to vanish" request. As a matter of interest, if there is a need to delete revisions of a large page, or the whole page, how should we do this? Request it somewhere? Have a "Requests for bigdelete access" process to rival "Requests for rollback" in pointlessness? Carcharoth (talk) 23:53, 16 January 2008 (UTC

Actually, per WP:BEANS, keep the list quiet for a while? Carcharoth (talk) 23:57, 16 January 2008 (UTC)
Yes, I don't think it wise to go creating a list of easy targets. Woody (talk) 00:00, 17 January 2008 (UTC)
If only there was some way to decide what the community thinks of this new user right... EconomicsGuy (talk) 00:02, 17 January 2008 (UTC)

More seriously, can anyone delete pages with more than 5000 revisions? When this is needed, what should be done? I've heard somewhere that oversight is less server intensive, but I don't think oversight should be used purely because it is the only option for removing (eg. libellous) revisions from a page history for pages with 5000+ revisions. Carcharoth (talk) 00:05, 17 January 2008 (UTC)

end of copied section
Actually, is there a reason why whenever we need to take out a single revision from a page history we can only do selective un-deletes, after a complete delete, and not just a direct selective delete in the first place? Fut.Perf. 00:23, 17 January 2008 (UTC)
At the minute it's not possible. The developers have been working on a system that would allow us to delete individual revisions (and some wiki's have it implemented already (I use it on the medcom wiki)), but it's just too buggy at the minute for en.wiki. I'm sure it will come in time. Ryan Postlethwaite 00:46, 17 January 2008 (UTC)

Can we please have the ability to delete individual revisions - like we do with the restore function? ViridaeTalk 00:48, 17 January 2008 (UTC)

If you look above, i've explained it's just too buggy at the minute. I hope someone can fix that ASAP because it would be a great tool to have. Ryan Postlethwaite 00:50, 17 January 2008 (UTC)
Excellent. I hope somebody does fix that ASAP. Deleting single revisions is such a pain right now. Especially if one is trying to delete a revision from an article that already has deleted revisions. -- Gogo Dodo (talk) 00:56, 17 January 2008 (UTC)
If people do want it, I could file a bug, it might hurry things up a bit. Ryan Postlethwaite 01:03, 17 January 2008 (UTC)
It seems readily apparent that anything to get this process moving would be great. Its only a matter of time (minutes maybe even!) before someone figures out how to exploit this... --Jayron32.talk.contribs 02:22, 17 January 2008 (UTC)
There's a bug filed already, bug 3576. Been open since late 2005. (I'm not sure how to interpert it, but, it looks like it was closed and re-opened at some point.) SQLQuery me! 04:38, 17 January 2008 (UTC)

Move the sandbox?

There was some discussion on the list, that I thought would be good to bring here, too. IIRC, the idea was to have a bot move the sandbox (to, say Wikipedia:Sandbox/Delete or Wikipedia:Sandbox/DeleteXXX like we do at WP:ACC) nightly, and re-create it, while tagging the old sandbox for deletion. SQLQuery me! 04:45, 17 January 2008 (UTC)

Yes, absolutely - is there any argument against doing this? We could always get a bot to move it, provided it's not been edited in (say) the past five minutes. Warofdreams talk 13:16, 17 January 2008 (UTC)
I think this is a good idea. J Milburn (talk) 18:35, 17 January 2008 (UTC)
Good idea, but invite the Sandbox-resetting bots to do the job. They currently have reset time periods; define the move/deletion parameters. -- SEWilco (talk) 18:47, 17 January 2008 (UTC)
Sounds good, but double-check that a move like this won't cause locks like deletion would. – Mike.lifeguard | @en.wb 20:00, 17 January 2008 (UTC)
It will not. Prodego talk 02:51, 18 January 2008 (UTC)
The proposed bot move would not cause the massive locking problem. The moves should happen often enough that the article history for each version will be relatively small. When the archived tagged-for-deletion Sandbox articles are deleted, the servers will be as busy as for any article with a history of that size. So... how often should the Sandbox be archived? Are there any relatively quiet times? -- SEWilco (talk) 03:18, 18 January 2008 (UTC)
Is the current Sandbox the dangerous one? Developers should be invited to perform the first archival and deletion. -- SEWilco (talk) 03:18, 18 January 2008 (UTC)
No, the Sandbox has been cleared out - the dangerous one is /Archive. Warofdreams talk 13:49, 18 January 2008 (UTC)
I'll note that the sandbox is move-protected; this has only traditionally been the case due to its size, but unless this is changed any such bot would have to have the sysop bit. —Random832 16:28, 18 January 2008 (UTC)
Or just the 'protect' bit. I think there's a case to be made to split the admin bits for bots, at least; it could also be set up so any admin could demoveprotected the bot if it went wrong. --ais523 16:32, 18 January 2008 (UTC)
Just unprotect it, moving won't cause any DB problems. Prodego talk 16:49, 18 January 2008 (UTC)
Don't people expect the Sandbox to exist? It might be better for it to not be moved to random places. Particularly because people might create an article in it then move it to article space along with irrelevant editing histories. -- SEWilco (talk) 17:00, 18 January 2008 (UTC)
So have the bot go back to the move-created redirect and recreate the sandbox with the standard headers, etc. That would just be the first edit to the new sandbox. Seraphimblade Talk to me 08:09, 19 January 2008 (UTC)
I was actually concerned about new articles acquiring irrelevant edit histories from preceding Sandboxers, not about the missing Sandbox itself. -- SEWilco (talk) 16:40, 21 January 2008 (UTC)

Triple entries in page history

Why do the edits to User talk:RickK older than c. June 21, 2005, appear in triplicate? - dcljr (talk) 22:33, 17 January 2008 (UTC)

Hmm. There seems to be some sort of ripple going through the space-time continuum, here... Earlier today, just before I left the above comment, the effect was happening with edits made on or before June 22nd. When I left the comment, it was happening with edits on or before June 21st. Now it's edits on or before June 20th. In case the effect disappears entirely, here's what I'm seeing right now:
  • ...
  • (cur) (last) 23:04, 20 June 2005 Bradeos Graphon (Talk | contribs) (undo)
  • (cur) (last) 23:03, 20 June 2005 Craigy144 (Talk | contribs) (come back...) (undo)
  • (cur) (last) 22:59, 20 June 2005 Calton (Talk | contribs) (undo)
  • (cur) (last) 22:57, 20 June 2005 Hoary (Talk | contribs) (Take a break, and then come back.) (undo)
  • (cur) (last) 22:44, 20 June 2005 Duk (Talk | contribs) m (Reverted edits by Sir Grazingship to last version by Lucky 6.9) (undo)
  • (cur) (last) 22:40, 20 June 2005 Sir Grazingship (Talk | contribs) (undo)
  • (cur) (last) 22:37, 20 June 2005 Lucky 6.9 (Talk | contribs) (undo)
  • (cur) (last) 18:02, 20 June 2005 Everyking (Talk | contribs) m (Reverted edits by 205.217.105.2 to last version by Jake007) (undo)
  • (cur) (last) 18:02, 20 June 2005 Everyking (Talk | contribs) m (Reverted edits by 205.217.105.2 to last version by Jake007) (undo)
  • (cur) (last) 18:02, 20 June 2005 Everyking (Talk | contribs) m (Reverted edits by 205.217.105.2 to last version by Jake007) (undo)
  • (cur) (last) 18:00, 20 June 2005 205.217.105.2 (Talk) (test6) (undo)
  • (cur) (last) 18:00, 20 June 2005 205.217.105.2 (Talk) (test6) (undo)
  • (cur) (last) 18:00, 20 June 2005 205.217.105.2 (Talk) (test6) (undo)
  • (cur) (last) 17:47, 20 June 2005 Jake007 (Talk | contribs) (Hello again) (undo)
  • (cur) (last) 17:47, 20 June 2005 Jake007 (Talk | contribs) (Hello again) (undo)
  • (cur) (last) 17:47, 20 June 2005 Jake007 (Talk | contribs) (Hello again) (undo)
  • (cur) (last) 15:30, 20 June 2005 Calton (Talk | contribs) (→Netoholic crusade) (undo)
  • (cur) (last) 15:30, 20 June 2005 Calton (Talk | contribs) (→Netoholic crusade) (undo)
  • (cur) (last) 15:30, 20 June 2005 Calton (Talk | contribs) (→Netoholic crusade) (undo)
  • (cur) (last) 14:38, 20 June 2005 Lucky 6.9 (Talk | contribs) (undo)
  • (cur) (last) 14:38, 20 June 2005 Lucky 6.9 (Talk | contribs) (undo)
  • (cur) (last) 14:38, 20 June 2005 Lucky 6.9 (Talk | contribs) (undo)
  • (cur) (last) 00:31, 20 June 2005 DS1953 (Talk | contribs) m (→GAP Project) (undo)
  • (cur) (last) 00:31, 20 June 2005 DS1953 (Talk | contribs) m (→GAP Project) (undo)
  • (cur) (last) 00:31, 20 June 2005 DS1953 (Talk | contribs) m (→GAP Project) (undo)
  • ...
As you can see, every edit made before 18:02 (U.S. Central Time), 20 June 2005, is showing up three times. The page versions in each triple are one unit apart — e.g., the edit(s) made by DS1953 at 00:31, 20 June 2005, are numbered 15562586, 15562587 and 15562588. The "diffs" among these versions show no change in page content (example, between 15562587 and 15562586 — although I don't know if this diff will show the same thing in a few minutes), as one would expect if they were all referring to the same edit. Very strange. - dcljr (talk) 01:09, 18 January 2008 (UTC)
It was deleted 3 times within 2 minutes and restored.[1] I guess it was restored in triplicate for some reason. PrimeHunter (talk) 01:28, 18 January 2008 (UTC)
Um, how could a page be deleted repeatedly without the page being restored in between deletions? According to the log, it was only restored once, after the three deletes. - dcljr (talk) 00:41, 22 January 2008 (UTC)
That was way back in 2005, so maybe the old version of the software had a bug that allowed users to delete nonexistent pages. - FISDOF9 01:12, 22 January 2008 (UTC)

The edit history of Pat Ross also appears in triplicate too for unknown reasons. hbdragon88 (talk) 00:18, 19 January 2008 (UTC)

Some of these repeats may derive from the history merge. If two pages were related enough to get merged later, it's not implausible that someone might have edited the pages in quick succession at some time before the merge. Also, some delete/restore actions appear to get logged without having any effect, though I'm not sure why. Gimmetrow 01:19, 22 January 2008 (UTC)

Length of references

Is there any way of reducing the length of external links that are used as references. I use the Atlas of Canada as references in several articles but the links to the maps make it hard to edit the page. To see what I mean take a look at Henik Lake in edit mode. The two links account for more than half the size of the article, 5,740 bytes with and 2,200 bytes without. Obviously I can't use tinyurl but there should be a way around it. Thanks. CambridgeBayWeather Have a gorilla 05:35, 20 January 2008 (UTC)

In Firefox, the long URls just go off the screen to the right, so editing really isn't a problem. -- John Broughton (♫♫) 14:48, 21 January 2008 (UTC)
True but using Firefox can cause viewing problems if not checked with IE before saving. CambridgeBayWeather Have a gorilla 15:25, 21 January 2008 (UTC)

strange bug on block logging

resolved - fixed in r30040

While doing unrelated testing on block lengths (incidentally, I've added "2 weeks" to the list of choices, since people seem to love fortnights, and this way a 2-week block will show up as "2 weeks" in block logs no matter how it was entered), I discovered something bizarre: My alternate account is, presently, blocked. I have logged in and attempted to edit to verify But it has a pristine block log. Can anyone guess how this happened? —Random832 21:56, 21 January 2008 (UTC)

Your IP might be blocked, and ipblock-exempt prevents you from feeling that block's effects. But your alternate account, not being an admin, is blocked. You would need to check the "Block originally applied to" section of the message to be sure. Prodego talk 22:24, 21 January 2008 (UTC)
Like I said, this was something I did. You guessed incorrectly - anyway, what happened was, I blocked my alternate account _in lowercase_. It doesn't show up in the block log, but it shows up in my log, and (until I unblocked it), in ipblocklist. —Random832 22:55, 21 January 2008 (UTC)
mentioned it on IRC and it's been fixed now. —Random832 23:19, 21 January 2008 (UTC)

Viewing and editing Wikipedia while running scripts

An issue at ANI involves whether an editor can view and/or edit Wikipedia while running a script. Could people here advise on how easy it is for this sort of inability to arise, and how easy it is to get around? The idea is that editors should be able to respond to concerns and queries while running such scripts. This was an image deletion script running for 2 hours after a manual check. The thread is here. Any opinion on whether such "script-blindness" situations are common, and what the guidelines/workarounds should or could be, would be welcomed. Thanks. Carcharoth (talk) 22:29, 21 January 2008 (UTC)

There is no reason you shouldn't be able to view and edit Wikipedia with a script running. Prodego talk 22:34, 21 January 2008 (UTC)
To be clearer, it is possible this involves a single "task" that the script starts running at a rate of several a minute for two hours. If that was set going under the account, can you still view and edit Wikipedia (eg. in another browser window or tab) under the same account? I don't know, cos I've never used a script (tabbed stuff is as far as I've got). I am going on a single assertion made by someone in that thread (Phoenix-wiki). If that turns out to be wrong, this will be embarassing, as I've asked lots of people now... Carcharoth (talk) 22:44, 21 January 2008 (UTC)
Now my signature no longer works. (That happened on another talk page. But I signed it here and it worked.) But I have turned off Greasemonkey. Mattisse 23:06, 21 January 2008 (UTC)
This may be a Browser, OS, or Usage-style issue. Under WinXP with Firefox (don't know about IE, etc., but would guess the same applies), Twinkle's pop-up confirmation box locks out concurrent activity in its associated browser window (which may be the only browser window open and/or the only user task running at the time). Other wikipages which might be open for editing in other tabs of that browser window are inaccessible until the confirmation box closes. -- Boracay Bill (talk) 23:30, 21 January 2008 (UTC)
I've pointed this out back in the TWINKLE thread. Hey - does this multi-threading with see also work, or is it just annoying? Carcharoth (talk) 00:23, 22 January 2008 (UTC)
That's not the problem here, but that twinkle can eat up all your memory and cpu if you are unlucky, once api-edits get online, much of that problem will be gone. Currently there is a "bug" in probably all JavaScript engines, that I had to prevent by effectivly disable the garbage collector, at the moment all gathered data will be kept in memory until the whole script has finished. There could be a possibillity to break that bond, but would require some enginering. What I'm actually do is to add all XMLHttpRequests to an static array, because if they goes out of scope, they will disappear, and all but the last request made will go out of scope. AzaToth 00:33, 22 January 2008 (UTC)
Sounds... fascinating! :-) Carcharoth (talk) 01:11, 22 January 2008 (UTC)
And utterly incomprehensible, too! Avruchtalk 02:25, 22 January 2008 (UTC)

Locator map CSS problem

For some reason IE and Firefox aren't using the same coordinates when setting location points on Template:State_parks_of_Minnesota_map. I have no idea about CSS, (I stole it from Template:State_parks_of_Pennsylvania_map]) Can anyone help? -Ravedave (talk) 05:19, 22 January 2008 (UTC)

Bug in HTML tag: div dir="rtl"

I just read the article Mesha Stele, and there is a problem in the way it displays. The article uses HTML "div dir="rtl" tag for a section of hebrew transcription of the text, here is a copy of part of the article code:

== Text ==

The text, in [[Moabite language|Moabite]], transcribed into modern [[Hebrew alphabet|Hebrew letters]]:

<div dir="rtl"> <pre> 1. אנכ. משע. בנ. כמש.. . מלכ. מאב. הד 2. יבני | אבי. מלכ. על. מאב. שלשנ. שת. ואנכ. מלכ

... MORE LINES OF HEBREW HERE ...

33. ---------[ויש]בה. כמש. בימי. ועל[...]. משמ. עש 34. -------------- שת. שדק | וא </pre> </div>

== Translation ==

''I am Mesha, son of Kemosh[-yatti], the king of Moab, the Dibonite. My father was king over Moab'' ...

Notice that only the hebrew text is inside the div tag, the problem is that the "Translation" header of the following section is displayed in my browser justified to the right, as if inside the div tag. The next line after the header is displayed correctly on the left.

I'm using English Windows XP Professional, Interent Explorer 7. I have hebrew support installed.

Here in my sandbox I made a copy of the problematic part of the article, you can see the bug occures. I tried in another sandbox, you can see that adding a line of text "ADDED LINE OF TEXT", just before the "Translation" header, makes the bug silent. Itaj Sherman (talk) 17:22, 21 January 2008 (UTC)

FWIW, this doesn't occur on my Firefox. Algebraist 17:46, 21 January 2008 (UTC)
Kludged. A slightly disconcerting Internet Explorer bug, possibly one up for consideration in the CSS. You put style="width: 100%;" into the <div> tag. I got the idea from here. See my changes. Confirmation for Internet Explorer 6 would be nice. x42bn6 Talk Mess 17:49, 21 January 2008 (UTC)
Without doing anything else, my initial reaction is that the code is bad. Try <div style="direction: rtl;"> instead. EVula // talk // // 19:41, 21 January 2008 (UTC)
Turns out either method is valid HTML and XHTML, although may have a slightly different interpretation. dir="rtl" is equivalent to style="direction: rtl; unicode-bidi: embed;". In either case, except for Internet Explorer bugs, it should revert to the previous direction at the end of the div.
I checked in Internet Explorer 6. It's broken exactly as Itaj Sherman says in the original revision. x42bn6's kludge has fixed it. Absolutely no problems whatsoever in Firefox with either revision (surprise surprise). • Anakin (contribscomplaints) 20:08, 21 January 2008 (UTC)
Does anyone think this could be added to one of the CSS files on Wikipedia as a "fix" for Internet Explorer? x42bn6 Talk Mess 17:15, 22 January 2008 (UTC)
How would we get a stylesheet rule to apply to divs with the dir attribute set? Or would it work to simply apply the width:100% to *all* divs? • Anakin (contribscomplaints) 19:57, 22 January 2008 (UTC)
I don't know, there might be some advanced CSS I don't know about that might help. The problem lies in the <pre> tag, it seems. I understand it might be awful CSS but putting pre { width: 100%; } might fix it. x42bn6 Talk Mess 21:56, 22 January 2008 (UTC)

Tables that feature both sortable columns and merged cells

In making a table, if one is to use the "rowspan" feature to merge multiple cells within the same row, the unfortunate consequence is that the table is no longer able to be sortable. Here is an example of what I mean: if I make the following table using both features, sorting the table makes things go crazy. Try it yourself.

Year Col 1 Col 2
2007 2 1
4 8
2008 5
2009 7

Does anyone have any idea of how to implement both? Although I don't know much programming this kind of stuff, this is how I would guess it would have to be done. Instead of specifying in the table code to "rowspan", the table could automatically megre identical cells, only after the table has been sorted. So that each time it is resorted, and the order changes, the cells will rearrange, and any matches within the same row will be displayed as merged. Of course, that could be an option, and not the default. Is that possible? Or is there a better way of doing it? Drewcifer (talk) 20:37, 21 January 2008 (UTC)

Neat! Acts somewhat like a sliding puzzle. Gimmetrow 07:07, 22 January 2008 (UTC)

MediaWiki Search - How does it work?

Hi there,

I'm really sorry if this is the wrong page...but I have nowhere else to turn, so I hope you'll forgive me. LOL.

I am a postgraduate student (and advocate of Wikipedia) doing research for a project at the Centre for Digital Library Research at Strathclyde University on Wikipedia/Web 2.0. With no technical expertise I am having trouble finding out the answer to the following question, so if anyone could help me it would be much appreciated!

Why, if I search for the string "gdl.cdlr" does this article:

http://en.wikipedia.org/wiki/Daldowie

show up in the search results, but not this one:

http://en.wikipedia.org/wiki/James_Maxton

?

As far as I can see, "gdl.cdlr" is part of what I call 'hidden text' behind the clickable link, and looking at the source code I see little difference; gdl.cdlr is not tagged as metadata; nor is it given as a keyword.

How does the MediaWiki search software work, and why am I am observing this effect?

Yours, Kathleen Menzies (kmenzies@cis.strath.ac.uk) —Preceding unsigned comment added by 86.142.45.46 (talk) 22:01, 21 January 2008 (UTC)

Well the links were not done in exactly the same way (note: they have been edited on the Daldowie article since this question was posted); on the Daldowie article they were done as [[http://example.com]] and on the James Maxton one they were done as [http://example.com Link description]. I'm guessing that given that it is more likely to be useful to people to search text in the article than obscure links within the article's code, in the second case the MediaWiki software didn't index the URL because it had a nice link description to index instead. In the first case, there wasn't a description, so it had to index the URL as is. Just a guess though. You see it doesn't search the active articles, but uses a search index that is compiled from the articles every so often. If it's useful to you the software that runs Wikipedia is open source, written in the PHP programming language, and available for download from http://www.mediawiki.org/. The MediaWiki site also has a good bit of technical documentation available. If you ask your question on the support desk there they may have a more detailed answer. Hope this helps. • Anakin (contribscomplaints) 14:40, 22 January 2008 (UTC)


Thank-you; that is immensely helpful! :)

Kathleen M. —Preceding unsigned comment added by 86.142.45.46 (talk) 20:06, 22 January 2008 (UTC)

Re: Procaine

hey i'm new to editing, but i found a controversial statement which textbooks indicate is wrong that needs to edited by someone who's completely sure.

It's regarding pharmacology of Procaine.

It says Procaine is a vasoconstrictor, but actually it's a vasodilator. If this could be put on a discussion so that relevant people can moderate it, that'd be great.

Thanks —Preceding unsigned comment added by Wanderingpistachio (talkcontribs) 06:12, 22 January 2008 (UTC)

Template for track listing

User:Juhachi and I have been playing around with the idea of making a template for track listings similar to how Template:Episode list works. (User talk:Ned Scott#A proposition). See Template:Track list and Template:Japanese track list. The basic benefit of this is that it allows parts of the data to be meta tagged, which makes future maintenance easier, and would even allow for the data to be machine readable. I'm not aware of any track listings that contain a large amount of different fields, but if there are then the labeling would also make entering the data easier. The templates also use a hack from the episode list template that makes a table cell pop out if a field is listed, but blank. This is to encourage people to fill in incomplete lists without having to place a placeholder to keep the cell open. Thoughts? -- Ned Scott 22:57, 22 January 2008 (UTC)

See Wikipedia talk:WikiProject Albums#Template for track listing for main discussion thread.

Edit conflicts because of sig

I keep getting edit conflicts with myself when posting comments on talk pages. The edit conflict seems to happen because ~~~~ turns into my sig — when it shows the diff, "My version" is just the comment I wrote with ~~~~ at the end, while the "stored version" is the same commetn I wrote, with my sig at the end. Why's that?--Phoenix-wiki 20:26, 20 January 2008 (UTC)

It probably happens because parallel edits are automatically merged by line, as long as there are no conflicts. So, you can go to a page, edit several parts, save, go back, edit some more, save again, and it won't mind (same for multiple users). If however you added a signature, that line may conflict the next time you try to save it, as the signature has expaded once already (so you're changing a line already changed). The solution is usually to get a newer edit window and newer wpEdittime time. If the wpEdittime is newer than the last edit, then no conflict checking is done.
If you aren't recycling edit forms, then maybe the server times are desynched. ^_^ --Splarka (rant) 08:34, 21 January 2008 (UTC)
Are you using some sort of library that checks for edit conflicts by downloading the page again after each edit to check that the edit worked? If so, you won't be able to use signatures or substitute templates. The solution to that problem is to get a better library. — Carl (CBM · talk) 14:28, 21 January 2008 (UTC)
It doesn't happen all the time, just some of the time.-Phoenix-wiki 20:19, 23 January 2008 (UTC)

IP user claims checkuser ability:

See here: [2]. Is this techincally possible? How so? If so, what do we do about it. There is a relevent thread at WP:ANI also, but I thought I would ask here to see if this were a bluff, or if this was for real. Any ideas? --Jayron32.talk.contribs 02:22, 22 January 2008 (UTC)

It is not possible for them to have access to the CheckUser tool in MediaWiki, but there may be other nefarious ways of gaining such information. Most likely explanations:
  1. bluff/lucky guess
  2. that user has accidentally made a logged out edit in such a way as to link the IP and the account; this user noticed it and grabbed the IP
Do not panic; this is very likely to be a non-issue. – Mike.lifeguard | @en.wb 03:35, 22 January 2008 (UTC)
I haven't edited logged out, I suspect it was a bluff to be honest. I agree that they have no access to the checkuser tool, but could someone create one of their own? Woody (talk) 03:47, 22 January 2008 (UTC)
The CheckUser tool is an extension, available here. It reads the recentchanges table in the database and can compare IP addresses in the rc_ip field. Other tools could be made to analyze that data, however, it is access to that data which is relevant (and highly restricted). --MZMcBride (talk) 04:39, 22 January 2008 (UTC)
There is no way of accessing that data through the wiki except via CU (so far as I know; I'm not a dev). – Mike.lifeguard | @en.wb 04:56, 22 January 2008 (UTC)
So the information isn't included in Wikimedia database dumps? If it isn't, then most likely, somethings been compromised somewhere. Nwwaew (Talk Page) (Contribs) (E-mail me) 05:08, 22 January 2008 (UTC)

There was a bug in the API that leaked autoblocked ips but went live about a month ago. (Does the threatened user have any blocks about that age?) Or there could be a SQL injection somewhere, it's not like our database schema is hidden in an obscure place. If this is genuine, then SQL injection is the most likely cause. MER-C 05:12, 22 January 2008 (UTC)

If this is genuine, is this a Bad ThingTM??? Wouldn't such exploits be open to abuse, and shouldn't they be closed or followed up? --Jayron32.talk.contribs 05:28, 22 January 2008 (UTC)
FYI, I have never had any blocks, auto or otherwise against my account. Woody (talk) 10:41, 22 January 2008 (UTC)
There are no SQL injection vulnerabilities in MediaWiki, I don't know why MER-C would think that such a rare and dangerous type of bug is the most likely cause. The anonymous user could have just posted an external link to a webserver under his control to Woody's talk page and correlated the referrer with the IP when Woody clicked on it. -- Tim Starling (talk) 12:29, 22 January 2008 (UTC)
That is the most likely cause, probably sirpeterscott.com if that is the case. That site has the same philosophy as the IP who has been editing the Kingston University page. See Talk:Kingston University#Protected and below for information on the dispute. Woody (talk) 12:37, 22 January 2008 (UTC)
I was just coming to the same conclusion. It has a JavaScript hit counter on it which logs the referrer and user agent for each visitor. -- Tim Starling (talk) 12:46, 22 January 2008 (UTC)
If that is the case, is it a problem worth fixing. Does it pose a threat to users privacy? It could always be blacklisted. Woody (talk) 12:51, 22 January 2008 (UTC)
Whenever you visit a website, usually that website logs your IP address anyway or at least the IP address that is currently yours at the time. It's not really an issue if a webpage logs you IP address - you may have seen those viral images that tell you your IP address, browser, location, etc. x42bn6 Talk Mess 12:56, 22 January 2008 (UTC)
It's just how the Internet works. You give out your IP address to every site you visit. There's no way to find out what those sites will do with that IP address. Use Tor if it's a problem for you. -- Tim Starling (talk) 12:58, 22 January 2008 (UTC)
That was why I asked the question, it seems quite normal to me. I don't have anything to hide and the only thing that alarmed me was the possibility of someone having unauthorised access to checkuser tools, which I thought was a very wild possibility. It was just a user who was losing an argument, trying to force his way out. Seems resolved to me. Thanks for your help. Woody (talk) 13:01, 22 January 2008 (UTC)
But how is the external site linking the IP address to User:Woody? The site has a list of IP's that connected to it. None of them say Woody, it shouldn't be in the referring URL and there shouldn't be any cookie leaks. So how is it done? Franamax (talk) 00:08, 23 January 2008 (UTC)
But if the site doesn't have that many visitors and you have some info about the intended victim such as their location it's possible to put two and two together. Theresa Knott | The otter sank 06:16, 24 January 2008 (UTC)
Why not ask? I asked in a completely non-threatening way how he does what he claims he can, so don't yell at me for the comment. Prodego talk 00:44, 23 January 2008 (UTC)
I got a response, and yep, thats how he says he does it, setting a trap link. Prodego talk 20:26, 23 January 2008 (UTC)