Abrir menu principal

UESPWiki β

UESPWiki:Administrator Noticeboard/Archive 36

< UESPWiki:Administrator Noticeboard
This is an archive of past UESPWiki:Administrator Noticeboard discussions. Do not edit the contents of this page, except for maintenance such as updating links.

Planet Elder Scrolls

The wiki needs to be checked for links from Planet Elder Scrolls. Not only did it move from Gamespy to IGN, which would've already broken the links, it actually closed down two years ago. That means links on a number of pages are broken and need to be replaced. I noticed it on Tes3Mod:Texture_Fixes and Morrowind:Recommended_Mods at least - there's probably several more affected. Digifiend (talk) 20:31, 29 November 2014 (GMT)

It looks like a new user, Gormadoc took care of this a couple of days ago. Thanks for reporting it, and thanks to Gormadoc for fixing them! Robin Hood  (talk) 16:19, 18 December 2014 (GMT)
No problem at all, it was interesting. The rest of the links are on user pages, user talk pages, are in some user comments on article talk pages, and Roleplaying is protected. The last is easy for anybody with higher permissions to fix, but what do I do, if anything, about the others? - Gormadoc (talk) 04:38, 29 December 2014 (GMT)
Not overly related to this, but what kind of protection is on Roleplaying? It doesn't say and there is no mention of who even placed the protection. I would think that for the user talk posts adding the deleted link thing would suffice, same for the user pages (I hate suggesting editing someone else's userpage, but it would fall under general link maintenance I think). ~ Ad intellige (talk) 04:43, 29 December 2014 (GMT)
It looks like it's not protected, but locked. I don't recall seeing that before. It may have been to prevent even attempting to edit the page and encourage editing of the other page. I hoped it would fall under link maintenance and would be acceptable, but I wanted to get confirmation from someone who has actually edited the site for a while. Thank you! - Gormadoc (talk) 04:53, 29 December 2014 (GMT)
It was a major problem article back in the day, so it is fully protected to discourage any further editing. I have removed the Planet Elder Scrolls link from the article in question. --AKB Talk Cont Mail 06:01, 29 December 2014 (GMT)
The "it's" in "For players who consider Fast Travel to be unrealistic, it's use can be minimized ..." needs to be changed to "its". :) Holomay (talk) 14:23, 29 December 2014 (GMT)
Okay, everything that was using the PES template has been changed. This task can be considered done! - Gormadoc (talk) 17:05, 29 December 2014 (GMT)

ESO Data Mining

I've been busy over the last month mining a variety of data from ESO and while its not completely finished it is working well enough to be let loose.

Two things I'd like to get from everyone:

  • Bugs and issues: I already know of a bunch of things but let me know of things anyways in case I've missed it.
  • Feedback/Ideas: I don't have a clear direction of where to use these utilities/data. Some of the data may be useful to include in the wiki page, hence the dumpMinedItems utility, but let me know what format/data is needed. I'm thinking the dynamic item tooltip might be useful to include within the wiki as well.
  • More Data: There is still a ton of data from the uespLog add-on that hasn't been completely parsed yet. If you need/want anything in particular let me know.

-- Daveh (talk) 21:44, 11 December 2014 (GMT)

One thing I've wanted since before the game, but haven't figured out how yet is to have it so hovering over an item link would result in a popup of the tool tip. That way we can have links to the items without having to have a separate page for the millions of items that are in the game (and thus breaking the wiki). If you have ever been on Wowpedia, they do something similar, plus they use mediawiki and not wikia so I hope that something like this could be possible? Jeancey (talk) 07:39, 12 December 2014 (GMT)
Making tooltips from it is an interesting idea actually. Having a quick look at Wowpedia (I've never been there before), it seems that they still have pages for all the items, but they have a javascript code that pulls the infobox from the article and displays it on the tooltip. Their help page on tooltips gives a bit of detail on how this works, and I assume the code required for it is in Common.js. --Enodoc (talk) 09:53, 12 December 2014 (GMT)
Yes, a popup tooltip is definitely possible. I'll test out some things next week. -- Daveh (talk) 13:22, 12 December 2014 (GMT)

() Update: Added a few parameters to the dump script to output a summary of item types:

Items that have different levels/qualities will show the ranges of various data in these tables. Html and wiki tables show the item's icon and have a link to the item data page. DumpMinedItems Documentation has been updated with these new parameters and all the type numbers that go with them. -- Daveh (talk) 16:09, 12 December 2014 (GMT)

Pop-ups are cool, but I still think we should make use of wikitables as well (at least for unique items). It'll be exciting to start incorporating this data into actual articles. Online:Sealed Urn was my attempt at manual extraction. —Legoless (talk) 16:42, 12 December 2014 (GMT)
I can get a similar output to that page using http://esoitem.uesp.net/dumpMinedItems.php?equiptype=3&armortype=1&fields=icon,name,armorType,level,armorRating,value,enchantDesc,trait,description&output=wiki&extrait (see here for that pasted into the wiki). The dump script has been designed to be as general as possible to make it easy to adjust output as needed. Still needs a few tweaks depending on what data is needed to be output. For example, I like the trait (X%) format on that page. It probably also needs a few more ways to select which items to output (simple search by item name or a list of itemIds come to mind). Updated with a new trait name display option. -- Daveh (talk) 17:52, 12 December 2014 (GMT)
Popup Test Page: Got it working with only minor changes to the existing code and seems to work well. Will likely have to create an extension to get it working in the wiki but this is not difficult. -- Daveh (talk) 19:51, 12 December 2014 (GMT)
Turned out easier than I thought which is strangely nice for a change. See User:Daveh/testlink. You may need to force-reload the wiki page to get the link popups to display properly. -- Daveh (talk) 21:25, 12 December 2014 (GMT)
Templating it might make it the most accessible, but yes, that's really nice to see! Jeancey (talk) 01:49, 13 December 2014 (GMT)
Not sure if it's just me but the popups seem to have stopped working. I can try and get this into a template, but that may be something that Robin or someone else could do better than me. --Enodoc (talk) 19:54, 14 December 2014 (GMT)

() I haven't really been following this closely, but I don't think we can do anything with a template unless the data is actually published directly on the wiki (i.e., not via a PHP script). Robin Hood  (talk) 20:01, 14 December 2014 (GMT)

(edit conflict) It's still working for me, but I'm noticing that it causes my entire system to freeze up for a few minutes upon first loading. Dunno if it's Chrome or my machine causing the problem, since it's an issue I encounter semi-frequently, but this is the first time I've seen it happened on UESP. —Legoless (talk) 20:04, 14 December 2014 (GMT)
It might just be that my js has stopped working for some reason, as some other functions seem to have disappeared too.
I thought maybe it would be possible to template it in the same way that <ref> became {{ref}}, and looking at it, that seems to be based on #tag:, but I wasn't sure how generalized that function can be for custom tags like <esoitemlink>. --Enodoc (talk) 21:53, 14 December 2014 (GMT)
I wasn't thinking of that, Enodoc. Something like that might be doable. If you want to do a proof-of-concept or even just a mockup template, we can see how it works and go from there. Robin Hood  (talk) 01:36, 15 December 2014 (GMT)
Template is done and seems to work. See {{Item Link}}. There may be some unforseen issues with it, but at base level it seems to be working. --Enodoc (talk) 11:02, 15 December 2014 (GMT)
Which browsers are people seeing JS issues in? Everything looks/works fine for me in Chrome/IE but Google Adsense is messed up in FireFox (ad code is displayed instead of the ad image). This doesn't seem to be related to this new plugin as disabling it has no effect. -- Daveh (talk) 20:13, 15 December 2014 (GMT)
I just tested on all five major PC browsers and the popups fail on all of them once the page is fully loaded. On Chrome, IE, Firefox, and Safari, the popups will load and display properly provided that the page is still loading and processing my custom JS. Strangely, on Opera, it simply never worked at all, but perhaps Opera doesn't allow scripts to run until it's completely finished loading the page? I dunno. I tried logging in as the bot, which has no custom JS, and I get no popups at all on either IE or FF (didn't try the others). Robin Hood  (talk) 21:05, 15 December 2014 (GMT)
They work fine while loading and when loaded on Chrome on PC for me. Jeancey (talk) 21:24, 15 December 2014 (GMT)
It's IE (11) and Firefox (33) that I'm having the trouble with, but only when I'm logged in. If I log out, everything works fine. --Enodoc (talk) 23:18, 15 December 2014 (GMT)

() Can people with Javascript issues try doing a forced reload (ctrl+F5) or clear your browser's cache to see if it fixes anything or if the problem persists? I've disabled the ESO item popup just in case it was the one causing problems. -- Daveh (talk) 21:24, 16 December 2014 (GMT)

It's working ok for me now; HotCat is back and other JS things are working properly again. the item popup must be conflicting with specific user-based JS from options in Special:Preferences or something, as it was only a problem when logged in. --Enodoc (talk) 22:37, 16 December 2014 (GMT)
The extension is re-enabled. Try doing another reload (or forced reload) and see if anything breaks this time. -- Daveh (talk) 16:30, 17 December 2014 (GMT)
Unfortunately it all broke again, sorry! --Enodoc (talk) 17:38, 17 December 2014 (GMT)
Same here. Without the extension it works. With that enabled, no JS :( Moinante (talk) 19:53, 17 December 2014 (GMT)
I removed a piece of logging code that I forgot was there that was causing issues on IE/FF. Try a forced reload of the page and see if that does anything. There may still be an issue with some logged out pages that have been cached by Squid until I clear/reset them. -- Daveh (talk) 14:27, 18 December 2014 (GMT)
Almost; there's now just a minor residual conflict. If the Gadget Convert the "sections" portion of the sidebar to be collapsible. is enabled, all JS works except the esoitemlink popup. If the Gadget (Sidebar) Replace labels for collapsible menus with links, condensing it a bit is enabled, no JS works at all. At the moment I'm running with neither of them, and everything else (HotCat, the Clock, grouped Recent Changes, and the esoitemlink popup) is working fine. (Admittedly I didn't test this thoroughly before, so they may have been causing the conflict all along.) --Enodoc (talk) 15:35, 18 December 2014 (GMT)

() Confirmed here as well. Disabling the sidebar code, popups started working again. I don't have much else that I use that involves JS after initial page load. Robin Hood  (talk) 16:01, 18 December 2014 (GMT)

I just noticed I have javascript again. It seems everything is working ok now. Moinante (talk) 21:50, 21 December 2014 (GMT)
It "should" be fixed now...I added back the jquery include for the item link code. MW should already have jquery included but it doesn't seem to work 100% if I don't include it. -- Daveh (talk) 18:27, 22 December 2014 (GMT)
Works for me. After re-enabling the sidebar code, popups still pop up the way they should. Robin Hood  (talk) 18:30, 22 December 2014 (GMT)
Same here. Everything seems to be working now. Thanks Dave! --Enodoc (talk) 18:48, 22 December 2014 (GMT)
Maybe not quite yet. I'm just noticing that only the first table on a page displays as sortable, and even that one isn't actually sortable (e.g., Skyrim:Marriage). Robin Hood  (talk) 03:47, 23 December 2014 (GMT)
I just noticed the image uploading preview doesn't work eihter Moinante (talk) 22:36, 23 December 2014 (GMT)

Edit Break 1 (ESO Data Mining)

JavaScript once again seems to be broken when the gadget (Sidebar) Replace labels for collapsible menus with links, condensing it a bit is enabled, and the popup again isn't working with Convert the "sections" portion of the sidebar to be collapsible. Maybe I should just stop using those gadgets... --Enodoc (talk) 09:37, 19 January 2015 (GMT)

Alternatively, we could fix the gadgets so they work. :Þ I'll look at them and see what's up. The popups were working with the sidebar fix the other day, but they seem to have stopped now. ARGH! It might just be something to do with the server upgrades. I'll poke around and see what I can find. Robin Hood  (talk) 17:41, 19 January 2015 (GMT)
Things seem to work now, though I did need to do a hard refresh for them to work. Please let me know if they work now for you as well. Robin Hood  (talk) 23:42, 19 January 2015 (GMT)
Yep, they do appear to be working again. Thanks! --Enodoc (talk) 00:16, 20 January 2015 (GMT)

Featured Article

That's the first time I've processed an FA nom. Can someone with more FA experience just look over it and make sure I got everything right. Thanks! Robin Hood  (talk) 20:51, 15 December 2014 (GMT)

Request for my 0RR -> 3RR

I am writing to request for a 3RR.

Why would I request for it? Because I have made approximately 100 edits the past few days which weren't pointless, improved articles and have taken things to the article talk page or a user's talk page when a contribution is partially reverted or fully reverted.

This is a request directed mainly towards the administrators to have a form of consensus about whether this request is appropriate. Dragon Guard  (talk) 16:06, 2 January 2015 (GMT)

Non-admins are still more than welcome to join in the consensus. Dragon Guard  (talk) 16:11, 2 January 2015 (GMT)
My suggestion would be a trial period at 1RR, then see how that goes. You've obviously improved over the time you've been here, but whether you're young or are on the autism spectrum or whatever else, there are still a lot of niceties of social interaction that you don't seem to understand. I would prefer to test the proverbial waters before restoring you to the standard 3RR. Robin Hood  (talk) 19:05, 2 January 2015 (GMT)
(edit conflict) If I may leave my own opinion as a semi-active patroller watching the RC, I'm a little cautious about him getting his 3RR returned to him. He's got his decent edits from time to time, but he tends to get reverted a decent amount himself, and I have a slight concern for him being able to judge and perform his own reverts, considering that he himself is not quite far enough beyond the threshold of what I personally find comfortable for editing standards. And, for the content editing he does perform, the ability to perform a reversion of another user isn't a hinderance necessarily.
If he was in good standing, I'd be willing to give him the benefit of the doubt and watch as he develops and matures as an editor, but right now he's not in good standing. He's better behaved than he was when he was being warned or blocked all the time, but he still has his moments that come off as condescending and inappropriate at times, which was a concern coupled together with his 1RR blocking, and as he's not developed quite enough in that aspect of his editing, I'm not entirely convinced that he will not return to using revert privileges as part of his argument.
Or, per RH70, I could see a 1RR trial being doable, and if that's the case, I have faith in the administration's decisions. -damon  talkcontribs 19:10, 2 January 2015 (GMT)
Before I dwell into your comments I would just like to say thank you. Thank you, one and all, for taking your time to respond to my request. I will now state what I think based on the number of statements you have put down.
@RobinHood70—I concur with your 1RR trial suggestion, and I have improved a lot since my last warning (back in 2013) from ThuumofReason, who told me, within the last couple of months, that there is still room for improvement but I have much better understanding of how things happen on the site. Regarding these "niceties of social interaction" (please specify these), I have only really been reminded about my comments which could be construed as "condescending", "rude" and "outright rude at worst". Just last monty he gave me a reminder of my condescending comments I have said recently, telling me to stop making them. He gave no diff links to them whatsoever and serious accusations require serious evidence. I take it was because of my first few sentences on the community portal starting from "What a pity" all the way to "references". I don't really see how "This is how I feel about these "guidelines" and anyone wanting to support them." is (if at all) bad. It was just stating that the previous thoughts were from myself and no-one else. Anyway, moving on, what do you mean by "proverbial waters"?
@Damon—I understand that you are cautious about me being given my 3RR privilege back to me, but I haven't reverted edits in ages and thus reverting the wrong edits probably isn't going to happen (you can almost be rest-assured of it, if you think of it that way). Trust me on this. Moving on: what is your definition of "good standing"? Is it fully knowing all the guidelines perhaps? As for the "moments" part, see the above paragraph on the ToR part. Regarding the 1RR, I would support that concept fully. Thank you for reading through this long reply and stating your thoughts on the matter. Dragon Guard  (talk) 20:29, 2 January 2015 (GMT)
There is no set standard on when someone returns to "good standing". It's a subjective thing, and it varies based on a case-by-case basis, though it's roughly consistent. If one were to merely given a warning on a one off time and then quickly corrected their behaviour and made it not an issue anymore, I personally would start overlooking the presence of the warning within a few days or weeks when interacting with that user. A user who got a block might be closer to a view months of being behaved and not having problems.
For your specific case, my personal opinion of "good standing" is based on the fact that in 2013 you had numerous warnings and blocks consecutively, which is generally concerning, and while you're greatly improving as an editor, you've still got a few odd moments that i'm not fond of where you can come off as combative or condescending, which was one of the reasons you were blocked some time ago. Were that behaviour to outright cease or nearly cease and stay that way for some time, I might generally disregard your warnings/blocks if I had to interact with you unless it was a serious matter. And, of course, that's just my opinion. Whichever administrator decides to make the decision on giving you a 1RR or 3RR verdict might have different ideas about how you'd have to be to get their favour. Have I answered this thoroughly enough? Or, do I need to clear something up or elaborate? -damon  talkcontribs 21:20, 2 January 2015 (GMT)
Not yet. From your perspective, please could you evaluate about why those comments on the CP were rude and condescending? Also, why did ToR give me a reminder on this behaviour, but gave no diff links to show evidence when serious accusations require serious evidence? Dragon Guard  (talk) 21:26, 2 January 2015 (GMT)

() Damon phrases my thoughts exactly. You're beginning to make progress, but you're not at the point where I have enough faith in your edits to support a removal of the restrictions. A 1RR trial run would be feasible, but even that makes me feel a bit uncomfortable at the moment. Until you demonstrate that you're capable of understanding our procedures on par with other editors, I wouldn't feel comfortable with a 3RR on your account. Zul se onikaanLaan tinvaak 01:48, 3 January 2015 (GMT)

Also, I did not say that you have a much better understanding of how things happen; I told you that you were acting more politely in discussions than you have been in the past, like Damon said. However, your overall handling of discussions still leaves much to be desired, as do many of your content edits. Understanding why something is rude isn't really the kind of thing that can be explained; either you understand it or you don't, but it's an important thing to be able to understand it if you want to make real progress in how you handle disputes and discussions. Zul se onikaanLaan tinvaak 01:56, 3 January 2015 (GMT)
Edit wars were long ago...a long time ago...I'm not going to be doing that type of stuff again. And why do I need a better understanding of discussions for a 3RR when it has nothing to to with reverting? Maybe if you specify what these things are I need to understand in discussions (bullet point format maybe) are I will learn from these mistakes. Dragon Guard  (talk) 09:06, 3 January 2015 (GMT)
The issue I see here is that you continue to take a condescending tone in discussions, which can result in further disagreements between users and was, if I recall correctly, the entire reason your 0RR was enacted. -- RNM|T 10:29, 3 January 2015 (GMT)
Thank you for the reply, Nocte. If I remember correctly, AKB put me on 0RR after conversing with some others. Dragon Guard  (talk) 11:01, 3 January 2015 (GMT)
There have only been four other users stating their thoughts. Some more feedback would be appreciated. Dragon Guard  (talk) 17:13, 3 January 2015 (GMT)
While I typically would stay uninvolved in these sort of discussions, I believe that 1RR would be the best option, as it allows you some flexibility to prove that you understand the proper way for a discussion to occur, without resorting to simple reverts. Jeancey (talk) 19:52, 3 January 2015 (GMT)
Yes, I agree. But maybe we should let Krusty drop by in this discussion and see what he thinks—after all—he put me on 1RR in the first place. Dragon Guard  (talk) 20:28, 3 January 2015 (GMT)

() Since Krusty hasn't really been active since August/September, it's unlikely that he'll be around to contribute. That being said, I too think that a 1RR is a good idea for the time being. •WoahBro►talk 20:52, 3 January 2015 (GMT)

OK, a 1RR it is then, since most people seem to think it's best. But which admin will place me on 1RR (i.e., leave a notice on my talk page to tell me I'm officially on 1RR)? Yes, WoahBro. Krusty (and many others) have been inactive recently. Maybe they flew south for the Winter? Haha :)! Dragon Guard  (talk) 21:07, 3 January 2015 (GMT)
And maybe when they do leave me a message regarding 1RR, they should tell me all the things I have a lack of understanding in, in a wikitable, with two headers: occupational (when editing pages) and communication (my social niceties). An example would be
Communication
Don't derail posts about unrealistic suggestions with the goal being to get my own way (such as that investigator guard and NPC slider discussion).
Occupational
Don't add full stops to proper-sentence captions that I am not sure about.
What do you think? Dragon Guard  (talk) 21:16, 3 January 2015 (GMT)
I think if you have not already determined where you need to improve, your 0RR should not be changed at all. -- RNM|T 08:57, 4 January 2015 (GMT)
Agreed. The entire reason you were placed on a 0RR and given those blocks was because you never read the policies you've been linked to, and you still expect other editors to explain things in great detail to you step by step. We've been very patient with you, but our job is to assist, not spoonfeed. I'm sorry if that sounds a bit harsh, but if you can't take the time to try and improve yourself without needing other editors to explain every single thing to you, you're not ready for any lifting of sanctions. An understanding of policies must come before, not after. Zul se onikaanLaan tinvaak 14:02, 4 January 2015 (GMT)
Very good points. However, I think a trial 1RR is a good compromise. If DG can show that he can abide by the policies and guidelines, then maybe he can have the 3RR reinstated.
Also, this is the Administrator Noticeboard. This doesn't work like the CP trying to get a consensus. Just because a consensus is reached here, doesn't mean anything really has to happen. The admins can take into consideration the opinions of the community, but it is ultimately their decision what happens here. •WoahBro►talk 16:55, 4 January 2015 (GMT)

() Of course. I wasn't trying to imply that I had any final say in making the decision, just wanted to voice my concerns. I'm not exactly reassured, but I'm willing to go along with it if the admins decide to do a trial 1RR. Zul se onikaanLaan tinvaak 18:17, 4 January 2015 (GMT)

So...all users: basically...should we end it here and put me on 1RR...or shall we wait for more feedback from others (e.g., Legoless, AKB, etc)?
This is how I remember things: after a month of not following the policies I was given a warning by Silencer. One month later, he gave me another warning for still not following policy, although I had improved a very little bit. Very shortly after, I was blocked for two whole days by Krusty for refusing to follow instructions, who said I was now under 1RR. The next month, I was blocked for a month by AKB for the same reasons as the last warnings and blocks, but he was preparing it before the "Dragon Guard" topic went underway on the AN, apparently. AKB stated he blocked me for still not following policy. Shortly after my block expired he said put me on 0RR following further discussions with others, by the email feature presumably? In November I was given a warning by Thuum for bad etiquette towards others. I have not received a warning or a block since. This seems like I am not disruptive anymore I take it, otherwise someone would've blocked me by now?
I have been contributing a lot to the MWOP. And for constructive edits towards the project, everyone's work matters when the goal of the UESPWiki is to be the ultimate number one site of information on the Elder Scrolls. Dragon Guard  (talk) 18:28, 4 January 2015 (GMT)
As WoahBro said, this is the Administrator's Noticeboard. Asking all users for consensus doesn't matter. All editors are welcome to give opinions (and the ones interested in doing so have already), but at the end of the day this is a disciplinary hearing to wrap up sanctions placed on you, and it is solely up to the administration whether or not to do anything about it, and as a disciplinary discussion, they can make the decision taking the words of senior editors into account or completely disregard it and make their own call. The best thing to do right now would be to just leave this conversation alone, let it sit and simmer for a bit, and await an administrator's verdict on what they think is appropriate. -damon  talkcontribs 19:06, 4 January 2015 (GMT)
OK thanks. Dragon Guard  (talk) 19:10, 4 January 2015 (GMT)
To answer a few of DG's points in this conversation, "social niceties" includes anything to do with having polite conversation. You still obviously have some difficulty in that area. I'm thinking in particular of phrases like the one you used in the Gstaff discussion, which you mentioned above, "What a pity; your concerns are invalid". That's definitely not a polite thing to say to someone. Even the phrase you suggested as being not bad could be taken as such, where the "how I feel about...anyone wanting to support them" makes it sound like you see them as being invalid. Sometimes, you're not even being rude, you're just not understanding whose job it is to do what. "OK, a 1RR it is then" was not your call to make. As others have mentioned, that's for the admins to decide.
The phrase "testing the proverbial waters" just meant that it would be a trial. ("Proverbial" just means that I'm not being literal; it was unnecessary since, obviously, there are no actual waters to test.)
A table outlining specific concerns would be impractical. I know you're looking for concrete examples, and I've given you a couple above, but social etiquette isn't something you can really list in a table. It comes more from experience and interacting with people. If I'm right in my suspicions that you're on the autism spectrum (like me and one or two others here, btw), it takes a lot more experience and interaction than it does for most people, and you'll probably still never quite get it right. But that's getting a bit off-topic for this discussion. You're welcome to use the e-mail feature to contact me if you'd like to discuss that further. As for the occupational stuff, I'd say not to worry too much about that part. That's mostly a matter of time, though it's always good to review policies or ask about things if you're at all unsure about anything.
Lastly, while most people seem to be in favour of a 1RR trial, there has been some opposition as well. If an outside admin is reading this, I'd appreciate a call on this. If there's no response in the next day or so, though, I'll go with the 1RR trial discussed above. Robin Hood  (talk) 19:46, 4 January 2015 (GMT)
While I have noticed some improvement, I do not feel comfortable at this time removing 0RR. When I do, I will contact you to let you know that I am willing to make the change. --AKB Talk Cont Mail 20:22, 4 January 2015 (GMT)
So here's the situation now everyone: a 3RR is out of the question. A 1RR maybe. I know making a list of 1RR "supporters" and "opposers" would serve no purpose...but a majority seem to want to give me the benefit of the doubt with it. @AKB: I think different; if I'm reading correctly are you (kinda) taking control here—saying you can remove 0RR without a consensus? Because I don't think that is a very wise decision to make...suddenly putting me on 0RR, telling me it is (simultaneously) removed without letting anyone else know wouldn't be what I would do. Actually...maybe I won't get so involved in site discussions (including the AN, CP etc) anymore if I have difficulty being civil. Dragon Guard  (talk) 21:03, 4 January 2015 (GMT)

() I'm not sure you entirely understand how administrative actions work. Administrators cannot revert each others decisions without an agreement that it should be undone. As I was a part of the decision to impose 0RR in the first place (although it was universally supported by every administrator with whom I discussed it), my opposition here kills this since this has very minor support and due to Robin being unwilling to make the change if any admins opposed it. I said that by the time I feel you do not need to be on 0RR, I would contact you, because that would mean that I would feel that it is clear enough that you should be leaving 0RR for me to propose the change. You shouldn't have to ask to leave 0RR, we should be making the suggestion that it should happen. --AKB Talk Cont Mail 00:17, 5 January 2015 (GMT)

Oh all right then. Dragon Guard  (talk) 00:35, 5 January 2015 (GMT)
Thank you for understanding. As I said before, when I contact you about this in future, that will mean that I will have seen enough of a change in your performance on the UESP that most other users will likely see it as well. --AKB Talk Cont Mail 00:39, 5 January 2015 (GMT)
I've already discussed this with AKB, but directing DG here first was my fault. I misremembered how the 0RR came into effect—I had thought it was more of a group decision than specific to one administrator. If I'd realized that, I would have suggested you take it up with AKB first, then only bring it here if you disagreed with his decision. (Although it seems likely you would have disagreed, so this discussion would probably have happened anyway, just under a slightly different set of circumstances.) As AKB said, admins don't overturn one another's decisions unless there's a really good reason and strong support for the change. That's obviously lacking here. Robin Hood  (talk) 03:02, 5 January 2015 (GMT)
So the 0RR remains as arranged? Dragon Guard  (talk) 20:15, 5 January 2015 (GMT)
Yes. Robin Hood  (talk) 21:03, 5 January 2015 (GMT)
Look at it this way then: if my performance stops me from getting past 0RR (such as etiquette, even in discussions that don't concern reverting) then why doesn't everyone that has a lack of understanding get put on 0RR, as in the case of people who get blocked for misconduct (elliott, Silencer, rpeh etc)? Why is this a big deal where I can't even have a trial? Please can I have an explanation as to why the 1RR trial is not being allowed? Dragon Guard  (talk) 20:08, 7 January 2015 (GMT)

Edit Break 1 (Request for my 0RR -> 3RR)

Every situation is looked at individually. Comparing your situation to that of three other users is not helping your case, especially since two of them are currently blocked from the wiki. Additionally, this discussion is pretty much purposeless now - AKB has made it clear that he is not comfortable with removing your 0RR, and administrators do not typically overturn each others' decisions without significant grounds to do so. It is very much like AKB said: You shouldn't have to ask to leave 0RR, we should be making the suggestion that it should happen.-- RNM|T 20:40, 7 January 2015 (GMT)

Really? Well others clearly thought the opposite of AKB. Take WoahBro: he believes a trial 1RR would be OK, but a 3RR would only be OK when I understand the policies/guidelines better. This "significant grounds to do so" part is other believing a trial 1RR would be suitable, but not everyone; such as ToR, who would feel uncomfortable about it. And also, @AKB: what do you mean by when you see an increase in my performance? Does that mean when you see me acting normally like everyone else or just when I've made some more constructive edits, like how I'm working on Morrowind/Bloodmoon NPCs in my sandboxes? Dragon Guard  (talk) 22:57, 7 January 2015 (GMT)
Some other admins think I should be on 1RR—even Jeancey, in his comment well above about saying he wouldn't get involved in these discussions. That would be at least two admins supporting a 1RR, and at least one opposing
RobinHood70 and Jeancey (maybe others).
Vs.
AKB (maybe others). Dragon Guard  (talk) 23:02, 7 January 2015 (GMT)
I wonder what Legoless thinks of a 1RR? He hasn't dropped by here. Dragon Guard  (talk) 23:06, 7 January 2015 (GMT)
(edit conflict) Look, I'm sorry but this conversation just simply isn't going anywhere. Like Nocte said, this discussion is pretty much purposeless now. If the other admins felt strongly enough to challenge AKB's decision, they would've by now. •WoahBro►talk 23:09, 7 January 2015 (GMT)
I just wanted to drop by and give my thoughts on this situation. Personally i think it would be fine to give you a 1RR trial run, but AKB doesn't think you're ready yet. When he thinks you're ready then he'll let you move up to 1RR, 2RR or even 3RR. It really isn't a huge deal not being able to revert things and most of the time reverting isn't (Except for things like vandalism or a noticeable error on the page) necessary. If an editor really feels like a change needs to be made, they can take it to the talk page and propose it. You've been doing a great job from what i can tell recently, but others have been around a lot longer than me and know you better than me. I have to agree with what AKB said which is "You shouldn't have to ask to leave 0RR, we should be making the suggestion that it should happen." Keep up the good work and eventually you will be able to move up in the Revert rule stuff. Lorenut (talk) 23:10, 7 January 2015 (GMT)
Well...let's hope AKB changes his mind when he hears about my work on NPC articles. I've been having so much fun and satisfaction from writing them myself. Some very important NPCs (Darius, Carnius, Falx etc). He should be proud of me making the wiki a better place for everyone (and Jeancey too for my tireless contributions to the Morrowind and Bloodmoon namespaces and the MWOP). Sure, there's no rush to get it done; we're not obliged to do it by a certain time, but it's when I take time out my day, going in CSList while playing the game. Now that is some awesome work. Ain't it? If AKB won't change his mind, I'll work harder and harder. Dragon Guard  (talk) 23:20, 7 January 2015 (GMT)

() DG: Please don't try to set admins against each other. Whether or not that was your intent, that's the perception you're giving by bringing up who said what and trying to turn this into a vote. Even if I'd remembered that AKB was the one to impose the 0RR and this discussion had happened the way it should have, there's obviously no clear consensus to lift your 0RR. Since that's the case, the 0RR stands. Now, I suggest you drop this discussion, as it's very clearly irritating people. Robin Hood  (talk) 23:33, 7 January 2015 (GMT)

Irritating people? I see Lorenut supporting 1RR. Dragon Guard  (talk) 16:11, 8 January 2015 (GMT)
It's irritating because it is obvious that nothing is going to happen here and you keep prolonging a dead topic. Just please stop. •WoahBro►talk 16:14, 8 January 2015 (GMT)
Well this topic might be over. But I can't say I am too impressed—the development of the discussion was ruined. I guess maybe I should listen and do as you's say, and close this discussion. Thanks all for trying to get me a 1RR. Your call, AKB. Dragon Guard  (talk) 17:23, 8 January 2015 (GMT)
I may support it, but it's obvious that a lot of people don't, so please just let it go. Lorenut (talk) 18:34, 8 January 2015 (GMT)
Yes. I will drop this discussion. A very good idea. Dragon Guard  (talk) 18:49, 8 January 2015 (GMT)

() Dragon Guard, do not post to this topic again, for any reason. First, it's not "my call", it's quite clear from this that you definitely do not need be removed from 0RR. First, you failed to cite any reason as to why you needed to revert edits. Every discussion of why you should have a sanction/edit restriction/ban lifted should begin with what you plan on doing with that ability once you have it. While you have followed proper procedure by not reverting back, now that you're on 0RR, that may only be because we are completely forcing you to do so. Did you intend to try patrolling, reverting plain vandalism? You never made that clear in the discussion above.

Secondly, there was your immediate action to turn this into an argument after I confirmed I was not comfortable with you being off of 0RR yet, with a promise that I would consider this in the future, and see that this happens on my own volition. Even before I said this shouldn't happen, Damon, Thuum, and Rosa all voiced in to express some form of disapproval. But after I did, you failed to accept my answer that this wasn't going to happen, Robin's answer that this wasn't going to happen, Rosa's answer as to why this wouldn't happen, WoahBro's answer as to why this isn't going to happen, Lorenut's explanation for why this shouldn't happen, and Robin's second answer as to why this shouldn't happen. You even tried to start pitting the administrators against each other, a group who usually decides things as close to unanimously as we can can would not disregard my opinion on this subject, even if they supported you tentatively before, they would not have after I made my opinion on the subject clear.

Lastly, if you wanted to convince me that you deserve to be off of 0RR, a decision I made, why would you argue with me? Do you think I appreciate you saying "oh all right then" to me after I explain why this won't happen, only for you to proceed to argue for this for another three days? You accepting my decision was the first step to getting off of 0RR, it wasn't a roadblock, but the way off of it. Your failure to accept it is the single biggest detriment to your case going forward, to be clear. If you had just ended this at "oh all right then", you'd have been off of 0RR in maybe February, but the fact that a simple administrative decision had to become such a fight does you no favors. --AKB Talk Cont Mail 19:07, 8 January 2015 (GMT)

  Do not add to this discussion any further. --AKB Talk Cont Mail 19:07, 8 January 2015 (GMT)

New Servers

Renewal time for all the UESP's servers is approaching in February. The current servers have worked well for the past 4 years but looking at the current offerings and prices it looks like we can get significantly beefier servers at a lower cost:

Item Current New
CPU i3
i7 (db2)
i5
Dual Xeon E5620 (db1/db2)
RAM 4 GB
8 GB (db2)
8 GB
24 GB (db1/db2)
Disk 300 / 500 GB 1000 GB
RAID Only on files1/db1/db2 All
Cost/Year $15k $10k

For most of the server switch-overs it will (or should be) completely transparent with no down time and only minimal read-only time. I'll update this thread as I switch servers and any issues can be brought up here as well. I hope to start switching over later this week. -- Daveh (talk) 21:13, 14 January 2015 (GMT)

I'll be switching squid1 to use the new content1 server shortly. No downtime or interruption is expected. Note that I won't be changing content1's DNS entry for a few days at least. The new content1 can be accessed at 67.205.78.100 should anyone need it. -- Daveh (talk) 17:26, 16 January 2015 (GMT)
I was curious if it was working, so I tried logging in and got the enable cookies error. Obviously I have cookies enabled, but it appears that content1 doesn't work for logging in. Is that intentional? ~ Ad intellige (talk) 16:55, 19 January 2015 (GMT)
It should work as it worked for me. I'll try again once the new content1 is back up (see below). -- Daveh (talk) 22:33, 24 January 2015 (GMT)
Change of plans. The service provider is setting up a completely new cluster with all the new servers. I'll be reverting back to the old content1 in order to move the new content1 to this cluster. -- Daveh (talk) 22:33, 24 January 2015 (GMT)

All my settings, ranging from my Recent Changes view count, to my miscellaneous gadgets, to my wiki signature, etc, were all mysterious reset a few hours ago, and I had to go back to my preferences and fix them. Is this a random side-effect of the routine maintenance, and are other people needing to go to their Preferences page to fix things, or is it just me? -damon  talkcontribs 20:27, 25 January 2015 (GMT)

This shouldn't be anything related to the server migration unless you happened to be using the new content1 explicitly (either by IP address or by modifying your hosts entry). If anyone else has experienced this please let me know. -- Daveh (talk) 14:33, 26 January 2015 (GMT)

Content1 -- Will be switching over to the new content1 shortly. Will also be switching over all domains to the new content1 this morning. Please let me know of any issues that pop up regarding the wiki or any of the other sites (forums, blog, maps, mobile wiki, etc...). -- Daveh (talk) 14:35, 26 January 2015 (GMT)

In Chrome only, I'm not getting any of the RSS/blog icons (the RSS bullet is actually a completely empty bullet). All other browsers are working as expected. Are any other Chrome users noticing this or is it just me? Robin Hood  (talk) 15:38, 26 January 2015 (GMT)
What are you referring to, RH? I'm on Chrome and haven't noticed anything out of the ordinary. •WoahBro►talk 15:52, 26 January 2015 (GMT)
See :File:User-RobinHood70-Sidebar.jpg. The left side is what I see in Chrome; the right side is what I see in IE. I was logged in as the bot, just to ensure that there was no custom Javascript or anything like that coming into play. Robin Hood  (talk) 16:11, 26 January 2015 (GMT)
You're using an adblocker ~ Dwarfmp (talk) 16:36, 26 January 2015 (GMT)
Oh, bugger! :Þ Now that you remind me, I'm pretty sure we've run into this before at some point. Okay, problem solved. (On further research, if anyone else is using AdBlock, the problem comes from the "Fanboy's Annoyances" filter, which I just turned on recently.) Robin Hood  (talk) 16:38, 26 January 2015 (GMT)
Fixed issue with the esomap due to database permissions. -- Daveh (talk) 17:14, 26 January 2015 (GMT)

Content2 -- Will be switching over to the new content2 this morning. Let me know of any problems with the wiki or specifically the wiki search. -- Daveh (talk) 14:15, 27 January 2015 (GMT)

Content3 -- Will be switching over to the new content3 server shortly. -- Daveh (talk) 17:20, 28 January 2015 (GMT)

Fixed an issue with NFS locking service that caused semi-random slowdowns/timeouts for logged in Wiki users. -- Daveh (talk) 14:27, 29 January 2015 (GMT)

Squid1 and Files1 -- Will be switching over both today. Squid1 is simply a DNS change from the old to the new server which will propagate slowly over the next day. Switching to the new files1 shouldn't require any interruption but you may have issues loading newly uploaded images until the DNS change to images.uesp.net is recognized on your network (new images will be uploaded to the new server but images.uesp.net may still point to the old server). This can be temporarily fixed by manually adding a hosts entry in your OS to 174.142.163.104 for images.uesp.net. -- Daveh (talk) 14:45, 29 January 2015 (GMT)

Db1 and Db2 -- Will be switching over these last two sometime this morning. This will require a short period (~5 minutes) of read-only time for the Wiki and other services to be safe. I'll post here just before and after the read-only period. -- Daveh (talk) 13:30, 30 January 2015 (GMT)

Wiki write-lock happening now. -- Daveh (talk) 13:40, 30 January 2015 (GMT)
Wiki is back to normal. If I did everything right we should be running from the new db1/db2 now. -- Daveh (talk) 13:44, 30 January 2015 (GMT)
Had to switch back to the old db1/db2 as the wiki's weren't correctly setup to write/load from the new database servers. Doing some database fixes and will try again shortly. -- Daveh (talk) 14:43, 30 January 2015 (GMT)
Attempt #2 happening soon...wiki will be locked shortly. -- Daveh (talk) 15:06, 31 January 2015 (GMT)
Wiki unlocked. Should be running on the new db1/db2 if I did things right this time. -- Daveh (talk) 15:11, 31 January 2015 (GMT)
The only problem I can see is that the ESOlog isn't updating anymore. I assume this is due to the server switch? Jeancey (talk) 18:35, 31 January 2015 (GMT)
I'll check it out this afternoon to see what's happening with it. -- Daveh (talk) 19:44, 31 January 2015 (GMT)
It was simply missing the cron script to update it daily (I haven't completely finished copying the cron scripts to all servers yet). I've manually updated it and set it to update daily now. -- Daveh (talk) 21:01, 31 January 2015 (GMT)
I also no longer have permission to connect to DB1 and DB2. Robin Hood  (talk) 19:11, 31 January 2015 (GMT)
When you say "permission" do you mean shell login, mysql login from an external computer, or something else? I haven't setup the new firewall to permit MySQL access from an external computer like the old firewall yet. -- Daveh (talk) 19:44, 31 January 2015 (GMT)
Sorry, I meant MySQL. I don't think I even have shell permissions on DB1 and DB2, or if I do, I don't know about it. :) No rush. Robin Hood  (talk) 20:56, 31 January 2015 (GMT)

() Dave, if it's not been addressed yet, it's possible that Special:Webchat is having some issues, potentially related to all the updates that have been happening. On the Forum's Xertion IRC channel, a user reported that they couldn't access the wiki's IRC through the Webchat link, and I responded to a similar issue sent in via our Facebook page's messaging just a few minutes ago. I can't access it either, though it's not an issue for me, as I use an actual IRC client for all of my IRCing. It just returns the message "The server refused the connection.". Damon(talkemail) 21:22, 1 February 2015 (GMT)

qwebirc communicates over port 9090, which may be blocked by the service provider or the firewall. If for whatever reason you can't unblock port 9090, I believe you can change the port in the config files, then update the link used by the wiki extension I wrote. Dave, let me know if you need any help with that stuff. • JAT 06:22, 3 February 2015 (GMT)
My omission there...I had half-transferred it but didn't complete the setup on the new server for it. It should be running now and let me know if there are further issues with it. -- Daveh (talk) 12:41, 3 February 2015 (GMT)

Excluding Talk/User pages from Special Pages

I couldn't really think of a good subject without being long winded, sorry. Anyways, is it possible to have subpages not show up for anything other than just page list stuff (Long pages, oldest pages etc) in the Maintenance Reports section? Additionally, can userpages and all talk pages be excluded from the same list? I have yet to find a valid reason for them showing up, since they often times will only clutter it and leave entries that shouldn't be altered. ~ Ad intellige (talk) 16:17, 18 January 2015 (GMT)

Unfortunately, the MediaWiki software just isn't setup to filter almost any of the Special page results. Could it be done? Yes, with some additional programming. The problem is that that programming would need checked and possibly re-adjusted every time we update the wiki software. Just as with the Unused Templates we discussed, the software just isn't there yet. While not ideal either, the more future-proof solution would be to use on-wiki workarounds to trick the wiki into removing these results wherever possible. Those, at least, can readily be deleted or updated by anyone without requiring any back-end PHP programming that only a couple of users can do. Robin Hood  (talk) 20:03, 18 January 2015 (GMT)
Hmmm... I tried something to remove a dead end page from a list, but AFAIK it doesn't work. Is there any way to remove the dead end pages without adding a link in weird spots? And for the unused stuff, what do we do for them? (I would ask about the wanted pages, but that can wait a year or so until we have ESO mostly done) ~ Ad intellige (talk) 16:52, 19 January 2015 (GMT)
{{Dead-end pages exception}}—it's still a link in a weird spot, but at least it's an invisible link in a weird spot. :) I just tried it on Online:EsoExtractData and it worked fine. Whether or not it should be used on the remaining Daggerfall pages I'm not sure. At least the one I was looking at looked like it had names on it that might be useful to link, but I'm not acquainted with Daggerfall enough to know if creating pages for them would be a meaningless exercise. As for unused stuff, the only way is to actually use it, but we can set up a specific page or set of pages for things like that. Robin Hood  (talk) 17:34, 19 January 2015 (GMT)
I never considered using divs haha. I've yet to see any possible links, but its been so long since I played, I can't remember for sure. ~ Ad intellige (talk) 17:52, 19 January 2015 (GMT)

Moving SR:Thieves Guild (place)

Can an admin merge/move the pages? The merge proposal has had no opposition for around three years, and apparently no one got around to merging them. ~ Ad intellige (talk) 23:22, 28 January 2015 (GMT)

As far as I know, it's just a matter of manually merging any relevant information, then converting the original to a redirect (or perhaps deleting it). Merging page histories would be very confusing. Robin Hood  (talk) 23:39, 28 January 2015 (GMT)
I think the (place) is to be merged, but the (faction) page is to be moved to just Thieves Guild, overwriting the disambig. -- Hargrimm(T) 23:58, 28 January 2015 (GMT)
The disambig is still important, since 'Thieves Guild' appears on an actual map marker. —Legoless (talk) 00:02, 29 January 2015 (GMT)

IP Block Lengths

Dwarfmp just brought up a good point on my talk page about perma-blocking IPs and since we've pretty much all done it at some point or another, I thought I'd open a larger discussion of it here. The general idea, as I understand it, is that IPs should never be permanently blocked unless they're open proxies (excluding TOR nodes, which vary over time and are already taken care of by TorBlock in any event).

I'm pretty sure the admins all know this already, but for anyone else who might be wondering, the idea is that anybody at all could be editing from an IP and even tomorrow, a legitimate editor could come along on the same IP that was used by a vandal today and find themselves blocked for (from their POV) no good reason. With IPs tending to be somewhat more stable now that cable and DSL are the norm (which can sometimes stay with the same address for months at a time), it's not unreasonable to impose longer blocks. Since we're not as high-traffic as Wikipedia, I generally use a year when I'm blocking a persistent IP spammer/vandal if there are no legitimate edits by similar IP addresses...shorter if there are. Naturally, however, mistakes have been made over the years or admins just weren't aware of the issue at the time, as can be seen here.

I'd like to start pruning our perma-blocked IPs, but before I do anything, I think we should be sure that we're all on the same page and if we're not, then discuss any issues with it. Then, there's the question of how to go about the pruning. I'd like to suggest temporarily giving HnB block rights, then having it go through and re-block any non-proxies (based solely on the description, because there's no other easy way to tell) with a length of a year from the original block. If it's been more than a year, then unblock. After that, we can always reinstate anything that it misses if the address/range become problematic again. Alternatively, if we'd rather take a hands-on approach, we can all just chip away at the list over time, but there are roughly 2000 entries, which is...rather a lot.

And while we're pruning anyway, we should probably re-enable e-mail for any IPs with blocked e-mails, so they can readily mail an admin or Daveh if they need to. Thoughts? Robin Hood  (talk) 05:37, 24 February 2015 (GMT)

As a long-time sufferer of dynamic IPs, I agree with this sentiment. —Legoless (talk) 11:31, 24 February 2015 (GMT)
The bot is ready to go on this. If nobody has any objections, I'll ask Dave to give HnB temporary admin rights later tomorrow. It should be noted that the current version of HnB doesn't support blocking, so I'm calling HnB 2.0 into service for this one, which is still in development. Logging hasn't been developed in that version yet, so the bot'll just go at it with none of the usual log page edits. Apart from that, it should be fairly normal and, as always, you can edit its talk page to make it stop, if necessary. Robin Hood  (talk) 03:59, 26 February 2015 (GMT)
And done. The only permanently blocked IPs now are tagged as proxies. Our list of active IP blocks just got a whole lot smaller. Robin Hood  (talk) 17:55, 27 February 2015 (GMT)
So I take it that we should block open proxies permanently? ~ Dwarfmp (talk) 06:28, 28 February 2015 (GMT)
That's been our historical practice, to reduce sockpuppetry and IP-hopping vandalism. We can always change that practice if there's reason to, though. Robin Hood  (talk) 07:21, 28 February 2015 (GMT)

Request for archive semiprotection

On Oblivion talk:Sanguine/Archive 3. Sorry for the late request! Zul do onikaanLaan tinvaak 10:50, 4 March 2015 (GMT)

The bot's on it. Robin Hood  (talk) 15:01, 4 March 2015 (GMT)

Rangeblock

I'm not knowledgeable when it comes to rangeblocking, but there's need of one. See here for a history, with the latest vandalism made by an almost identical IP that was blocked for repeated vandalism ~ Dwarfmp (talk) 09:48, 23 March 2015 (GMT)

I've blocked the range for a year. Given that the edits were nearly a year apart to the same page, this person is obviously fairly persistent, so the block may need to be renewed. Robin Hood  (talk) 12:08, 23 March 2015 (GMT)

UESP Forums

I know there is something of an artifical divide between "UESP wiki people" and "UESP Forum people", but really, I don't think that a link to an official UESP Forum should count as an external link! It seems that my link to the UESP Official Guilds Forum from the Official Guilds wiki page gets trapped by the spam filter. Which is quite hilarious considering it's already on the page just a little way further up!

Please fix, because any link to forums.uesp.net from a registered wiki editor is extremely likely to be legitimate! baratron (talk) 19:25, 4 April 2015 (GMT)

The {{forums}} template might allow you to get around that for now. —Legoless (talk) 19:36, 4 April 2015 (GMT)
(edit conflict) It probably ought to be whitelisted, sure. Definitely an oversight if there are some users who are triggering the filter on UESP things. In any case, on a wiki formatting note, we generally don't link to the same thing twice in the same page, because it's just redundant to do so. Is the link on top not going to be sufficient in some way that there needs to be a second link? Or would that be linking to a specific locale on the forum? A link that only goes to the forum home page would probably be pointless and redundant, although there's not a problem with linking to other places on there. -damon  talkcontribs 19:46, 4 April 2015 (GMT)
I don't see any record of it in the abuse log, so I'm assuming it's some other anti-vandalism process like the spam blacklist. I've tried to trigger it with one of my alternate accounts, but can't seem to do so. Can you provide any details about what exactly it said, just so that I can be sure I'm fixing the right thing. Thanks! Robin Hood  (talk) 20:07, 4 April 2015 (GMT)

Elder Scrolls Soundtrack KickStarter

I received an email last week from Alex concerning a KickStarter related to the recording of ES music using an orchestra. Since he was unfamiliar with our wiki I'm posting his e-mail below for him:

I am running an exciting project that is about to launch on Kickstarter (1st May), and am hoping you would be able to help us out. We are trying to raise funds in order to record highlights from all the Elder Scrolls soundtracks with a full professional orchestra (City Of Prague Philharmonic Orchestra) as well as a full professional choir.

We have contacted the game studio who have given us their full backing to carry out this project and we have the relevant licences to re-arrange and record any pieces from the Elder Scrolls games series in place.

If we can successfully fund this project, we will be releasing over an hour's worth of the best Elder Scrolls music on CD and digital download, along with newly commissioned artwork based in the Elder Scrolls world.

We are currently looking for Elder Scrolls based websites such as yours to help us promote the Kickstarter project when the time comes, in the form of news feeds, social media links etc. In return we would be able to offer a link to your website under the 'Partners' section on our Kickstarter page as well as promotion from our other social media outlets.

If you would like to know more about the project or get UESP involved please let me know, Thanks.

Alex Gobbett

I don't necessarily have any issues with posting a news article and mentioning it on our social feeds but would like more details on the project, namely the legality aspect of re-recording the ES soundtracks. I'll direct Alex to this post so he can answer any questions/comments anyone might have. -- Daveh (talk) 18:09, 14 April 2015 (GMT)

Interesting suggestion, and (assuming it has Bethesda's backing) a worthwhile endeavor in my personal opinion. I could definitely see us using social media pages to mention this, and a news post may be an option as well, though I would also like more details before committing my support. Due to the nature of crowd-funding, I would feel a lot more comfortable promoting this project if, for instance, we could get some kind of verification that Bethesda personally supports this idea, or some idea of how much support the project already has. Zul do onikaanLaan tinvaak 20:35, 14 April 2015 (GMT)
No legal advice to give, but it would be fantastic if we could get confirmation from BGS that he has the appropriate clearances for this project. Other than that, if a guy's got an Elder Scrolls dream, I wish him all the best, so why not mention it. I probably wouldn't campaign for it, just publicize its existence, but that's just semantics.
Perhaps it would be prudent to widen the scope of a news entry to TES Kickstarters in general? It would cover Alex's project almost entirely and up front, of course, but could also include some minor tidbits at the end just to cushion things a bit. I just noticed that Jeremy Soule's The Northerner, started and funded over two years ago now, is not yet completed. A search turned up 8 projects which were (substantially) related to the Elder Scrolls, including Soule's symphony, and they were funded with a success rate of 25%. Insignificant RevisionsThreatsEvidence 03:15, 15 April 2015 (GMT)
Hi all. First of all thanks for posting and discussing this. There's a lot I could write about the project here, but I will try and keep it brief and to the point and answer any specific questions you have (hopefully the rest of the questions will be answered in the Kickstarter video and description).
Regarding the legality of performing, recording and re-arranging The Elder Scrolls music, its pretty straightforward and happens all the time, you apply for a licence to the relevant agency that handles that companies audio licensing and they will approve or deny you a licence to arrange and perform the music. In our case we have been given the green light on all the Elder Scrolls music, providing we have the funding to pay for it. The only sticking point we have here is obtaining the licences for Arena and Daggerfall which are owned by a separate company. The other part of the licensing/copyright is with Bethesda and the use of names, images etc. They have agreed in principle to allow us to go ahead with the project and use the Elder Scrolls name (after all every sale we make gives them a cut of the profit) and we are currently going through the details of exactly what we can and cannot do in terms of imagery etc. In any case, for the final product we will be commissioning original artwork that has the look and feel of the Elder Scrolls world.
Back to the music, my partner and I have a combined 20 years experience in the Music industry in many different roles, composing, recording, mixing music for film, games, bands etc. and we will personally be transcribing by ear each and every note that will be played in this project. We have been working on mockups of the final scores and re-arranging music where we think it could benefit (I am happy to play/discuss some pieces privately if anyone is interested). Once we have secured the funding for the project, we will be working on getting all the arrangements finalised and recorded by a world class orchestra and choir as well as any solo voices/instruments parts that there may be.
Obviously, this project isn't going to be cheap, which is why we are trying to get as much support from the Elder Scrolls community as possible. We are finalising our reward structure for Kickstarter at the moment, but it will include CD's, digital downloads, Limited Edition CD's and a few extra collectible items we are working on.
If you have any more questions, feel free to ask. Sorry if i have not posted this correctly, this is my first post here. Alex--Agobbett (talk) 14:54, 15 April 2015 (GMT)
I would have assumed the Arena/Daggerfall music passed into the public domain after both games were released for free. Might be worth looking into that again. —Legoless (talk) 16:50, 15 April 2015 (GMT)
Not a lawyer, but I'm pretty sure that's not how public domain works. At least in the US, works do not fall into PD until I think 100 years after the death of their creators. (I think it used to be 75 years?) Either way, most of the creators are still very much alive, so it doesn't apply unless they've explicitly released the content as PD. Free to play is not the same thing. I imagine it won't be too difficult to get the rights if you can get them for the later games, but these things can get thorny. (At any rate, MW was where the music for this series started getting good anyhow. The earlier games were a little too 8-bit for my tastes.) — TheRealLurlock (talk) 21:55, 18 April 2015 (GMT)
Correct, sadly... Seems the assets are still the intellectual property of Bethesda (or this mysterious third party composer?). Shame though, I'd pay for an orchestral rendition of the shop theme. —Legoless (talk) 22:40, 18 April 2015 (GMT)
(edit conflict) Actually, originally, it became in the public domain upon the dead of the creator, or 50 years from the release of the thing in question, whichever came first. That changed like.... 80 years ago, so it isn't important now, but originally it made sense, now its just silly :P. In case this wasn't off topic enough, the reason disney keeps rereleasing their old movies in new formats (like snow white, little mermaid were "remastered") was because the copyright was about to expire and by rereleasing it, they renew the copyright on it. Jeancey (talk) 22:41, 18 April 2015 (GMT)
(edit conflict) True, they will not be Public domain. We are not going to be doing a lot with the Arena/Daggerfall music, we want to put it in as a kind of nod to where it started and to give the themes a true representation with a full orchestra since they are well written tracks (just a shame it was all midi sounds back then). Interestingly, the Arena theme tune is used in the ESO soundtrack - bonus points if you can spot which track! (Agobbett (talk) 09:42, 21 April 2015 (GMT))

Second patroller nomination

Hey folks! I have a question about patroller re-nomination. It's been a month since my first nomination, which took me by surprise. I know that the official policy is to wait three month before nominating again, however, there is an opinion that the cooldown period can be shortened in this case. What do you think, may I nominate myself now or should I wait another two months? Thanks.  ~Shuryard (talk) 18:29, 18 April 2015 (GMT)

I believe the intent of the wording on the Patroller page was to avoid a user gaming the system by withdrawing when the voting was going against them, then claiming that the nomination wasn't rejected. Given that you withdrew due to a change in plan, I don't see a problem with you running again. If others agree, I'd also suggest that we change the wording on the Patroller page to reflect that intent (e.g., "rejected, or withdrawn while there were opposing votes, ..."). If others disagree, well, another two months isn't all that long. :) Robin Hood  (talk) 20:42, 18 April 2015 (GMT)
I have no issues with Shuryard running for patroller again before the three month waiting period has ended. As for creating a rule regarding this in general, I have no strong feelings. I think the situation doesn't come up too much and could be handled on a case by case basis like how we're doing for this nomination. However, the current three month waiting period does seem silly and perhaps counterproductive to enforce with no exceptions if the nomination appeared to be supported unanimously before it was withdrawn by the nominee. Forfeit (talk) 22:12, 18 April 2015 (GMT)
I supported Shuryard the first time, and so have no objections to this. As a response to forfeit, however, I think that unless there is an emergency need for Patrollers, 3 months after a failed or going south nomination is okay, but 3 months after you were nominated by someone and declined it, even though it looked to succeeded just doesn't make sense. I think that a rule change would reflect the intent of the rule and prevent cases like this where the rule is blocking someone it was never meant to block. Jeancey (talk) 22:41, 18 April 2015 (GMT)
Rule change seems in order to me too. —Legoless (talk) 22:44, 18 April 2015 (GMT)

Dev Wiki Reset

I'd like to reset the dev wiki, both database and files, in testing and preparation for a wiki upgrade. If there's anything there that needs to be saved or backed up just let me know. -- Daveh (talk) 11:31, 1 May 2015 (GMT)

Dave: For our reference, there's a file in my user folder on content3 called "modified.txt" that lists all files modified since January 28, 2014 (the day after the dev wiki was installed). I've removed the images folder from the results, which was just thumbnail re-calculation and such. Robin Hood  (talk) 17:54, 1 May 2015 (GMT)

Wiki Upgrade to 1.22.15

I've reset the dev wiki to a copy of the wiki from this morning and re-updated it to v1.22.15. Since we worked out most of the bugs last year hopefully there shouldn't be too much left to fix. Feel free to take a look and post here any issues. Barring any major issue I'd like to perform the actual upgrade sometime this week. This will require the wiki to be set to read-only for 1-2 hours (I'd like to have less read-only time but it involves lost writes if the upgrade fails for whatever reason). -- Daveh (talk) 15:28, 19 May 2015 (GMT)

Assuming no news is good news, I'm scheduling the upgrade for this Friday morning around 8am EST. I'll add posts here when the wiki is locked/unlocked for writing and when the upgrade is completed. -- Daveh (talk) 13:09, 20 May 2015 (GMT)
I took a quick peek around the dev wiki and everything looked fine to me, but I wasn't exactly thorough. As usual, I should be online by no later than about 1:00 EDT (17:00 UTC) on Friday if there's anything that needs adjusting after the update. Robin Hood  (talk) 17:35, 20 May 2015 (GMT)

Wiki is Read-Only -- Will be set to read-only for the next 10-15 minutes to perform a complete database backup in preparation for the upgrade. -- Daveh (talk) 12:02, 22 May 2015 (GMT)

Done -- The backup is done and the wiki read-only mode has been turned off. The backup will be loaded into a temporary database over the next hour in preparation for the upgrade shortly after that. -- Daveh (talk) 12:19, 22 May 2015 (GMT)
Small hiccup due to changing files underneath Apache/PHP/MediaWiki which it didn't like. Should be fixed now. Some pages may need to be reloaded to take effect. Small delay in upgrade as I have to restart the database restoration for the temporary wiki. -- Daveh (talk) 13:20, 22 May 2015 (GMT)

Wiki is Read-Only -- Set back to read-only mode for the upgrade. Technically the live wiki will be using an hour-old backup during this time so this edit won't be visible until after the upgrade. Estimated time is 30 minutes if all goes well. -- Daveh (talk) 14:40, 22 May 2015 (GMT)

Upgrade Done -- Upgrade is finished and wiki has writes turned back on (assuming everything works when I try to save this edit anyways). Will be tweaking and clearing caches over the next few hours which hopefully won't break anything. Please add any issue related to the upgrade below, including minor display/edit issues. -- Daveh (talk) 14:59, 22 May 2015 (UTC)

Resolved Upgrade Issues

  • Slow Page Loads/MetaTemplate Messages -- Old pages are being slow to load and display a partial error message at the very top of the page: string(30)"MetaTemplateSaveData::savedata". New pages show a little box saying "Your edit has been saved" every time you load the page. —Legoless (talk) 16:04, 22 May 2015 (UTC)
Looks like some of this may be due to pages recreating all their thumbnails. So on pages with a lot of images it will be slow the first time. Can you give an example of a page with the MetaTemplate error...I haven't seen that particular one myself. -- Daveh (talk) 16:51, 22 May 2015 (UTC)
I tried loading some random pages, it seems that indeed the first opening of some pages is very slow, the MetaTemplate message appears, and some fonts and divs go wild, but in these cases, after reloading everything seems to fix itself. --Vordur Steel-Hammer (TINV1K) 17:05, 22 May 2015 (UTC)
It happened on Lore:The Brothers of Strife, which has no images. —Legoless (talk) 17:25, 22 May 2015 (UTC)
Regarding the slow page loading, I'm thinking/hoping I was merely too zealous with the clearing of the various caches and it is a temporary slowdown. Am continuing to monitor and check things to see if there's other reasons for the page load times. -- Daveh (talk) 17:53, 22 May 2015 (UTC)
The server loads are back to normal so I'm guessing the page load issue was due to the resetting of the caches. If any MetaTemplate issue shows up again let me know where and what page. -- Daveh (talk) 18:26, 22 May 2015 (UTC)
(edit conflict) I got the MetaTemplate error string(20) "ParserAfterTidy hook" string(30) "MetaTemplateSaveData::savedata" on some diff pages, like this one. It seems to have gone now though. The error was right at the top, partially hidden by the logo. --Enodoc (talk) 18:28, 22 May 2015 (UTC)
I'm still noticing the MetaTemplate error on some user pages (User:Forfeit and User:Legoless are two examples) and sporadically on other pages. For the other pages, reloading the page typically fixes the issue. However, reloading the two user pages I mentioned will still have the MetaTemplate error pop up. Forfeit (talk) 19:10, 22 May 2015 (UTC)
The MetaTemplate messages are my fault. They were a leftover from trying to figure out the issues with saving which I forgot to remove. They're completely harmless (other than altering page layout slightly) and do not indicate an error. I'm working on removing them all now. Robin Hood  (talk) 19:21, 22 May 2015 (UTC)
  • Custom CSS --- It was a purely aesthetic change I had made, so ultimately it can be gone and not matter at all, but my monobook.css page, which only had a CSS tweak that rounded out the wiki's sharp corners, isn't doing its job. -damon  talkcontribs 17:55, 22 May 2015 (UTC)
We moved to a custom skin "UespMonobook" to make it easier to track changes and do upgrades. If you copy/move your old CSS to a "UserName/uespmonobook.css" sub-page it should work. -- Daveh (talk) 18:36, 22 May 2015 (UTC)
  • Recent Changes -- minor issue
Duplicate text, "Track the most recent changes to the wiki on this page." I think this happened last time too. Silence is GoldenBreak the Silence 19:05, 22 May 2015 (UTC)
Ah yes, should be fixed now. -- Daveh (talk) 19:18, 22 May 2015 (UTC)
It's working fine for me at the moment. If you haven't tried already, can you do a hard refresh and see what happens? Also, do you get an error or just a blank page? Robin Hood  (talk) 20:14, 22 May 2015 (UTC)
Blank page. A hard refresh loads it, but a normal load after that goes back to the blank page. Silence is GoldenBreak the Silence 20:20, 22 May 2015 (UTC)
I think that's probably a caching issue. Dave, can you try purging any relevant caches and/or restart memcached and we'll see where we stand? Robin Hood  (talk) 22:14, 22 May 2015 (UTC)
Unless it is a more widespread issue I'm going to wait until Monday rather than risk a page load issue during the busy weekend like we did after the upgrade. -- Daveh (talk) 18:30, 23 May 2015 (UTC)
I'm also having issues with getting Special Pages to load, and refreshing doesn't seem to do anything. Zul do onikaanLaan tinvaak 14:36, 24 May 2015 (UTC)

() I've restarted memcached so try a page reload (preferably a full forced reload, CTRL+F5) and see if that does anything. If not let me know some more details on what browser/OS you are using and what page(s) are broken and how. -- Daveh (talk) 12:08, 25 May 2015 (UTC)

still getting blank page "Failed to load resource: the server responded with a status of 500 (Internal Server Error)" Safari 7.1.6 OSX 10.9.5 - caches cleared from my end, Cntrl-F5 does nothing. --Rhynchelma (talk) 12:01, 26 May 2015 (UTC)
I assume you are logged in when you get the error? Do you have any custom gadgets or CSS/JS? Do you get the same error when logged out? I don't have a Mac myself so can't directly test it on that platform. -- Daveh (talk) 12:43, 26 May 2015 (UTC)
Same logged in or out. No custom stuff. Works fine with Chrome, may be a Safari issue. --Rhynchelma (talk) 12:50, 26 May 2015 (UTC)
Just tested this for a bit and am getting the same issue as you. Chrome works fine but Firefox/Safari (using BrowserStack) gives me a 500. It also seems to work if you bypass the Squid cache directly via http://content1.uesp.net/wiki/Special:SpecialPages, which is strange as when you are logged in the Squid cache should do nothing. Will try a few things and see what works or not. So far as I can see it is only the SpecialPages that is broken. -- Daveh (talk) 12:57, 26 May 2015 (UTC)
It should be fixed now so give it a try and let me know. I'm unsure what the exact issue was but Squid didn't like loading that page so I explicitly set it to not cache SpecialPages. Strangely it had no problem with the indirect link (/w/index.php?title=Special:SpecialPages). -- Daveh (talk) 13:35, 26 May 2015 (UTC)
Works fine now. Thanks. --Rhynchelma (talk) 13:38, 26 May 2015 (UTC)
  • Diff colours have changed. This may have been a version difference rather than a bug. We should consider what colours work best for us (i.e., keep the new ones, revert, or something else altogether). Robin Hood  (talk) 20:17, 22 May 2015 (UTC)
That's a version difference, I think. The same thing happened on Wikipedia. --Enodoc (talk) 21:36, 22 May 2015 (UTC)
If that's a version difference, can that be changed? I think the new colors are significantly less clear than the old ones, which makes checking the differences much slower and difficult. --Vordur Steel-Hammer (TINV1K) 22:03, 22 May 2015 (UTC)
Fairly easily, yes. It's the same basic concept as the CSS changes I just did for the preferences and login pages. The only tricky part is figuring out what the old colours were, assuming we want to stick with those. Robin Hood  (talk) 22:09, 22 May 2015 (UTC)
Fable Wiki is still on 1.19, so a diff page there may help identify the old colours. --Enodoc (talk) 00:39, 23 May 2015 (UTC)
I've used a mix of the new style with the old colours, and added some that are more similar to our skin. I'm not 100% satisfied with the results, but it's certainly easier to read than the pale blue. We can do the backgrounds in solid colours rather than just the borders, if that works better for people, but I thought I'd try to respect the intent of the new style, at least for a start. Like I said, I'm not completely happy with it, so I'm open to suggestions if anyone has any. Robin Hood  (talk) 02:33, 23 May 2015 (UTC)
Personally I would prefer a solid background, as it makes it easier to see at a glance whether something was added or removed. There's a suggestion in my common.css if anyone would like to try it (it's the td.diff- lines). --Enodoc (talk) 11:21, 23 May 2015 (UTC)
Gallery descriptions have the same background colour as the wiki. Silence is GoldenBreak the Silence 23:28, 23 May 2015 (UTC)
() That looks like it's the default behaviour for a gallery. Did we used to have it looking more like the thumbnails? It doesn't look like in the site CSS file, but they did make significant changes to galleries in 1.22, so it may well be working differently now. Robin Hood  (talk) 05:14, 27 May 2015 (UTC)
  • Categories. Pages with no categories have no box from which to input them manually or with hotcat without opening the page to edit (the talk page here for example). Silence is GoldenBreak the Silence 20:20, 22 May 2015 (UTC)
Other gadgets are not working, like drop down menus. Silence is GoldenBreak the Silence 20:37, 22 May 2015 (UTC)
The ability to add categories to pages without any has been restored, at the cost of adding categories to pages that already have some. Silence is GoldenBreak the Silence 23:17, 23 May 2015 (UTC)
I just updated to the newest version of HotCat, which was the only change. I don't use HotCat, myself, but I see the button to add categories correctly (was looking at Oblivion:Quests). What page(s) are you looking at for no categories and what page(s) are you looking at for those that do? I don't normally use HotCat, so it's possible I'm just misunderstanding. It's also possible that this is another cache issue or whatever else it is that's hitting Modify Sidebar (added in this edit), for example, where it just seems to work or not, randomly, from one page to the next. Robin Hood  (talk) 23:26, 23 May 2015 (UTC)
Very odd, I just opened the page at the top of the RC, Online:The Colored Rooms, which doesn't have a +. I also can't do it on this page (the admin board) the category pages for the page linked, SR:Nord, LO:Atmora, OB:Potato, MW:Magic (random pages), but the option is there on OB:Quests, and also OB:Main Quests, but not Whitmond Farm, Daedric, or Master Training (all linked from that page). I'm using the news pages to check the non-categorised pages. I've purged and bypassed browser on all those pages, as well as content3. Silence is GoldenBreak the Silence 23:51, 23 May 2015 (UTC)
Noticeably, all the pages that HotCat doesn't work on are the same ones on which Modify Sidebar isn't working. I'm not sure what to make of it. There could be some other script breaking and causing everything to fail (like the sidebar code used to) or it could be something else entirely. I'll keep investigating when I have time. Robin Hood  (talk) 01:35, 24 May 2015 (UTC)
() We have some progress, at the expense of breaking a number of things like collapsible tables. The problem appears to be coming from the site's global Javascript. I just blanked it and every last page mentioned above is now working just fine. Over the next little while, I'll experiment in my own Javascript page and see if I can figure out exactly where the problem is coming from. Robin Hood  (talk) 04:10, 24 May 2015 (UTC)
Okay, I'm now officially confused. After blanking, testing, and restoring our site Javascript essentially unaltered, everything works fine...at least for me. I see a few possibilities: Dave did something behind the scenes at some point, the act of blanking that page fixed something internally, the fix is only temporary, or possibly we're just suffering from collective insanity. We'll have to wait and see what happens, then go from there. Robin Hood  (talk) 04:31, 24 May 2015 (UTC)
Fixed on all those pages, and also talk pages, which I noticed after my last message. The only gadget not working now for me is the drop-down menus. Silence is GoldenBreak the Silence 11:20, 24 May 2015 (UTC)
  • Preferences. The skin behind the preferences when using UESPMonoBook skin does not match, and the sidebar is not the same. Silence is GoldenBreak the Silence 20:20, 22 May 2015 (UTC)
The log in page appears to have the same issue as well. Forfeit (talk) 20:29, 22 May 2015 (UTC)
This is now partly fixed. It was a deliberate change in MediaWiki to help prevent page spoofing and other types of CSS attacks. I didn't consider that a major issue for us, so I switched the setting to allow customization once again. I'll change the gray backgrounds shortly, which is just a standard CSS change. Robin Hood  (talk) 21:47, 22 May 2015 (UTC)
  • Page title - on this page "Editing UESPWiki:Administrator Noticeboard (section)" seem to be lacking a few pixels in height causing the tops of the characters to be slightly clipped. Safari 7.1.6 on Mac OS 10.9.5. Exactly the same on Google Chrome (latest build). --Rhynchelma (talk) 23:12, 22 May 2015 (UTC) Edited to standout
I'm not seeing this in Chrome on a PC, but I'll download Safari and give that a shot later tonight. Is anyone else seeing font-clipping in the title? If so, what browser and OS? Robin Hood  (talk) 02:38, 23 May 2015 (UTC)
Neither am I. —Legoless (talk) 13:50, 23 May 2015 (UTC)
I am seeing this. You can see on the "S" in particular on UESPWiki that it's been cut off. I'm running Chrome. OS X 10.9.5 as well, if OS X is what will turn out to be the common denominator in this. Here's a screenshot for reference. This was an issue before I re-rounded my corners, so it's not my own CSS tampering that's caused this.-damon  talkcontribs 15:48, 23 May 2015 (UTC)
It would seem then to be an OSX problem. Anyone who is not running 10.9.5 want to add anything? --Rhynchelma (talk) 16:29, 23 May 2015 (UTC)
Working fine since your changes of a few mins ago, thanks. --Rhynchelma (talk) 19:02, 23 May 2015 (UTC)
Hadn't gotten around to updating this while I was investigating other issues. I increased the line-height slightly, but since I'm not experiencing the problem myself, I can only guess at the correct value. If anyone still sees clipping (after a hard refresh), let me know and I'll increase it some more...or any other patroller/admin can too. Robin Hood  (talk) 20:13, 23 May 2015 (UTC)
  • Add New Section - The "[Add New Section]" link at the bottom of pages is huge and bolded. —Legoless (talk) 15:09, 23 May 2015 (UTC)
I don't see anything that's changed in the code or styles, but it's hard to be sure. What did it look like before? Robin Hood  (talk) 20:13, 23 May 2015 (UTC)
As far as I remember, it looked like plain text before. —Legoless (talk) 20:49, 23 May 2015 (UTC)
I just told it to use the same style as an [edit] link. How does that look? Robin Hood  (talk) 22:40, 23 May 2015 (UTC)
It looks better, though there's a space between 'section' and the closing bracket ~ Dwarfmp (talk) 12:57, 24 May 2015 (UTC)
That was part of the original design. There's a space between the bracket and text on both sides. It can be removed quite easily, if we want to. Robin Hood  (talk) 21:23, 24 May 2015 (UTC)
Oh, but there was no space in between each bracket, I guess you fixed that. As long as it's even! ~ Dwarfmp (talk) 01:03, 25 May 2015 (UTC)
  • Falmer Font - As brought up here, the Falmer alphabet font isn't showing up on Lore:Falmer Alphabet, but it is on other pages. I checked around the other alphabet pages and it isn't happening there. Maybe a little slow to load, but it still shows. I really don't know anything about this wiki upgrade tech nonsense, so I have no clue if this is just something specific to this page, a deeper issue, or possibly even an issue unrelated to the upgrade. •WoahBro►talk 04:49, 26 May 2015 (UTC)
Fixed. Apparently, that hadn't been copied over to the development server before we upgraded, so it never got copied back once we did. It's there now, though I had the bizarre experience that it refused to "reload" in Chrome, but if I opened it on a new tab, it would show up. Thought I'd mention it in case anybody else has this problem. Robin Hood  (talk) 07:17, 26 May 2015 (UTC)
  • Drop-down menus gadget - I mentioned it above but its still 'broken'. Can it be enabled on UespMonoBook skin? Silence is GoldenBreak the Silence 18:29, 26 May 2015 (UTC)
Sorry, I've been trying to look into this, but hadn't found the time. Looking at it now. Robin Hood  (talk) 19:51, 26 May 2015 (UTC)
Just adding our monobook skin didn't work, so I updated to the latest version and that seems to have done it. Let me know if you have any further problems with it, as I don't use it myself. Robin Hood  (talk) 20:11, 26 May 2015 (UTC)
It looks fine. The first edit could have been missing a colon to work, but some extra usability might have been added in all those bytes. Silence is GoldenBreak the Silence 20:21, 26 May 2015 (UTC)
Yeah, I noticed that after I'd started importing the updated code and tweaking it, but I figured I was well into the process of upgrading at that point and might as finish the job. Robin Hood  (talk) 21:58, 26 May 2015 (UTC)
  • Edit Count pages - All aside the "all time" count (day, week, month, year) seem to be broken; they're just white pages. ~ Alarra (talk) 00:55, 27 May 2015 (UTC)
This is nominally fixed, but the Squid cache hasn't updated yet. You can go directly to one of the content servers for all your urgent filtered user edit count needs. ;) Robin Hood  (talk) 05:03, 27 May 2015 (UTC)
  • Edit Lead Section overlaps protection icons. I realize I'm probably the one who will end up fixing this, but I'm too tired right now, so just noting this for later. If someone else wants to have a crack at it before I get to it, go for it! I suspect it's something simple. Robin Hood  (talk) 03:27, 28 May 2015 (UTC)
There's a big gap between the Edit link and the Featured Article star. Silence is GoldenBreak the Silence 23:27, 3 June 2015 (UTC)
Any specific page, or all of them? What browser are you using again? I could swear you've mentioned it recently, but I can't find it. I'm seeing the expected gap (about the width of a T) on Sheogorath. How big is it for you? Robin Hood  (talk) 00:16, 4 June 2015 (UTC)
Firefox. All of them. It's about the size of a tab (__________) between the link and the star. It's not there on other browsers so its a Firefox issue. Silence is GoldenBreak the Silence 00:36, 4 June 2015 (UTC)
Okay, thanks! I haven't installed FF on this computer yet, but I'll do so tomorrow and look into this. Robin Hood  (talk) 01:28, 4 June 2015 (UTC)
The issue turned out to be that in Firefox, the icons hadn't always finished rendering when the gadget was trying to figure out where to put the button. I've added a half-second delay before the button is displayed, which fixed the issue in all of my tests. Let me know how it works for you. Robin Hood  (talk) 03:24, 5 June 2015 (UTC)

Outstanding Upgrade Issues

  • SearchLog Extension -- SearchLog is generating "mysql_real_escape_string(): Can't connect to local MySQL server through socket...". The extension should probably use the MediaWiki built-in database classes for interfacing with the database instead of hard-coded queries. -- Daveh (talk) 16:00, 22 May 2015 (UTC)
Note: this discussion refactored to separate two different issues - see next topic for what was split off
In layman's terms, are you saying its broken? Silence is GoldenBreak the Silence 18:55, 22 May 2015 (UTC)
Ah...no, this specific thing just affects the search log which merely logs all the search terms and the number of results found. It doesn't affect the actual search results. -- Daveh (talk) 19:16, 22 May 2015 (UTC)
  • Search - suggestions only show main namespace
For me if I just enter a term, eg vampire, it offers no pages (OB:Vampire, MW:Vampire). Adding the prefix makes it shows up but it needs the full OB:V before working. Instead it offers to search for terms containing what I've just typed. Silence is GoldenBreak the Silence 18:55, 22 May 2015 (UTC)
I'll look into the issue you mention (unsure if it was an old bug or not). -- Daveh (talk) 19:16, 22 May 2015 (UTC)
The thing Silencer mentioned seems to be new, as I would always get a dropdown even if I didn't include the namespace before. Post-upgrade, it directs me to the search results page if there's no namespace, resulting in an extra click (plus some scrolling). Not wiki-breaking, but mildly annoying nonetheless. likelolwhat talk lulzy to me 22:16, 4 June 2015 (UTC)
We may have a bit of progress on this. There was a setting missing from LocalSettings.php. I've added that now, but that on its own doesn't appear to have fixed anything. The search indexes may need to be rebuilt. I tried doing so myself from the lucenesearch folder on content3, but this doesn't appear to have succeeded. I don't want to play too much and mess anything up, so I'll let Dave take it from here. Robin Hood  (talk) 05:16, 9 June 2015 (UTC)
It seems that $wgEnableMWSuggest was deprecated in v1.20 so it doesn't do anything. I had to dig into it pretty deeply to find the cause but it seems that the search suggestion box now uses a hard-coded namespace setting of 0 as set in line 179 of resources/mediawiki/mediawiki.searchSuggest.js. Since there appears to be no way to adjust this even in the latest MW version I've gone ahead and just manually hardcoded our default namespace search set into it. You'll have to do a forced page reload but it seems to work now. The downside of this is that it ignores the namespace search settings of logged in users now. I'll look into a more permanent solution that doesn't involve manual editing of a MW file. -- Daveh (talk) 14:59, 9 June 2015 (UTC)
It's working for me logged in, and it has given me back Skyrim namespace to search in, though not Online. If you mean the preferences section to choose which namespaces search looks in is not working, I have "search all" ticked and nothing else. Silence is GoldenBreak the Silence 17:40, 9 June 2015 (UTC)
I'm not sure why Online is not showing up in the suggestions since it shows up fine in the search results and it should be the same index. Your search preferences will still affect the normal search but not the search box suggestions (at least until I find a permanent solution). -- Daveh (talk) 18:45, 9 June 2015 (UTC)
I'm also not seeing Online suggestions. Another issue with search too, is that the search results are not finding a lot of the more recently added pages (mainly ones added since Update 6 came out - for instance, Shara's House and Kirstenreth; heck, even Justice itself. These pages have been around for a few months.) Some of the results also contain out-of-date text, such as when you search "Crown Store": it says it's a "planned" in-game store in the results, while the page itself no longer includes the word "planned". ~ Alarra (talk) 00:23, 10 June 2015 (UTC)
Very strange...I've rebuilt the search index using a database dump from yesterday and the issue with Online results is still there. For example searching for "skyshard" shows no results in the search box suggestions but an actual search gives the Online:Skyshard page among others. Something, somewhere is interfering with the Online search results and I just have to figure out where. -- Daveh (talk) 12:58, 12 June 2015 (UTC)
For reference, if it helps, search suggestions for the Online namespace have never worked; it's not just since the upgrade. It's also not just Online pages. It appears to be every new page since about January 2013. Looking through some DB namespace pages, the last page I found that is in the suggestions is DB:Tel Mithryn (tower), created 8 Jan 2013. The page DB:Factions A, created 2 Feb 2013, is not in the search suggestions. Looking at UESPWiki:Upgrade History/2013, there are a few changes to search-based things that took place in that time period (as well as the update to 1.19.3). --Enodoc (talk) 13:42, 12 June 2015 (UTC)
There were some issues with the incremental search index updater that I assumed were related to the v1.22 update. I figured a complete index update would fix that but apparently not and I'm not sure why. Unfortunately, while Lucene appears to be the one used by Wikipedia its installation and usage is not well documented so I'm left to probe the guys of MediaWiki and try things to see what works or doesn't. -- Daveh (talk) 14:48, 12 June 2015 (UTC)
  • Page Saved Keeps Popping Up -- Mentioned elsewhere but adding it here now that I've noticed it as well. To duplicate:
  1. Edit/save a page
  2. Navigate to another page
  3. Go back to the page you just editted
  4. The "Page saved" popup shows up for a few seconds
If you go back to the page after a minute or two you don't get the popup any more. I'm unsure if this is the expected MediaWiki behaviour or might be a bug in v1.22 already fixed in a later version. -- Daveh (talk) 18:32, 22 May 2015 (UTC)
  • Redirects - I'm not sure this is related to the update but some redirects are not working properly. Lore:Guar, Lore:Gnarl, Lore:Giant Scorpion, and Lore:Troll, are all linking to some part of the Bestiary page but not their sections. This is inconsistent though, they all link properly from here, but not from the redirect page itself (refresh on the redirect page don't click the #link). For me Guar goes to Gourami, Gnarl to Giant Wasp, Giant Scorpion to Giant, and Troll to Travally. Silence is GoldenBreak the Silence 17:30, 24 May 2015 (UTC)
I'm also seeing this; seems to be related to having images on the page. For instance, it happens to a small extent on People D, where there's a few images, but not at all on People N where there's no images. So I assume it's some issue with setting the anchor location without images loaded, and then once they load and push around the text layout, the anchor spot is 'out of date'. -- Hargrimm(T) 13:20, 26 May 2015 (UTC)
That theory appears to hold water, the lower a thing is on a page the further away it appears to be. The thing is that redirects that are lower than the broken ones work, and the same redirects work properly from some locations but not others. Silence is GoldenBreak the Silence 18:29, 26 May 2015 (UTC)
For "People D", it's just hitting the bottom of the page, so it can't go any lower. Or at least it is for me. That's intended behaviour and it's always worked that way. Some of The Silencer's other links, though, like Giant Scorpion, can't be explained that way. I'm not getting the issue on Chrome. I'm wondering if this is another browser-specific issue that's just started appearing in 1.22. Robin Hood  (talk) 19:51, 26 May 2015 (UTC)
  • Blank pages - Similar to the Special Pages issue, I'm intermittently getting a blank page at Skyrim:Books. Robin Hood  (talk) 01:56, 3 June 2015 (UTC)
With some investigation, I found that we were getting this error (though in a different part of the MW code). As stated in that thread, a temporary workaround is just to give PHP more time to display the page, which I've done. In the long term, there could be a few issues that need to be looked at, including:
  • Caching not working as (well as) expected. I think this is the most likely issue, as caching should render the remaining points almost moot, as I understand things. Also, the fact that the same pages sometimes render almost instantly while other times timing out suggests that this might be the case. This is almost certainly tied in with the slow page loads issue and is something Dave would have to look at.
  • Some part of the {{Book Normal}} template may be an issue. Candidates include most of the MetaTemplate functions, but particularly #load, since, in my testing, pages like Skyrim:Places—which uses a lot of #loads via {{Place Summary}}—also sometimes took a long time to load.
  • General code bloat may be causing templates to take longer to render, therefore causing pages with a lot of templates to timeout when not cached for whatever reason. There's little we can do about this if this is the case. I think this is unlikely, as Morrowind:Morrowind uses a lot of {{Bullet Link}} templates (which are very straight-forward and don't use #load at all) and that never took more than a couple of seconds to display in my testing. Still, it can't be entirely discounted. Robin Hood  (talk) 03:26, 11 June 2015 (UTC)
  • Slow page loads (45-120+ secs) - Still getting slow page loads and page edits intermittently, not always on rarely accessed pages. Robin Hood  (talk) 01:56, 3 June 2015 (UTC)
  • UTC - I think the wiki switched back to "UTC" instead of "GMT" when labelling server time. Personally I prefer the name GMT. —Legoless (talk) 21:45, 17 June 2015 (UTC)
That's an internal change to MediaWiki, as it's actually still set to "GMT" in the settings. I'm guessing they changed the programming to use a consistent name for the same timezone, and in this case, they picked UTC. Robin Hood  (talk) 23:00, 17 June 2015 (UTC)
That's weird, pretty sure it changed to GMT after an update a few years back. Guess it goes in cycles? —Legoless (talk) 23:02, 17 June 2015 (UTC)
It did, but that was Dave's fault. :) We had had the time zone as "UTC" originally, but then when we upgraded, somehow, the setting got lost, and we briefly saw EST (or EDT?) in use. Someone noticed that and mentioned it to Dave, and when he restored the setting, he used "GMT" instead of "UTC". Apparently, nobody noticed the change (or at least didn't care enough to mention), so it stayed "GMT". Robin Hood  (talk) 23:05, 17 June 2015 (UTC)

ESO Log Data

I've been spending most of this week working on the ESO log and item data. Among the notable changes are:

  • Various bug fixes and performance tweaks.
  • More tables/fields are searched.
  • Mined item data summary can be browsed directly.
  • Addition of Set Summary data.
  • Update of mined item data (done).
  • New -- Updated esodata.uesp.net with all the updated LUA API related information.
  • New -- Updated the ESO map with new and updated icons (Dye Station, Fence, Guild Vendor, Dungeon, Solo Dungeon, Group Dungeon, Raid Dungeon).

If there are any specific things that you'd like to see worked on related to the uespLog, the parsed data, and item links just let me know. The purpose of the data was to help bring data from the game into the wiki but sometimes I don't have a good idea what data is needed and in what format. -- Daveh (talk) 15:22, 3 June 2015 (UTC)

There was a small layout problem with the item links going off the page if placed too far to the left (see here). —Legoless (talk) 15:35, 3 June 2015 (UTC)
The item dump and parsing has finished. Below are a quick summary of the items removed/added (I didn't track changes as almost all items changed last update):
  • Deleted Items (83)
  • 1386
  • 26246
  • 27382
  • 32311
  • 44575
  • 44584
  • 44634
  • 44710-44713
  • 44716
  • 45017
  • 54238-54240
  • 54389-54393
  • 54638-54640
  • 54855
  • 54871
  • 55429-55430
  • 56932
  • 57694-57695
  • 57834-57848
  • 58745-59487
  • 59733
  • 59943
  • 60998
  • Added Items (712)
  • 61028-64709
If anyone sees any issues with the display item data just let me know. -- Daveh (talk) 11:55, 5 June 2015 (UTC)
It seems like these items were deleted/updated from the database. Would it be possible to restore them? In general, it's probably not good to delete old items, since we're currently cataloguing them for posterity. —Legoless (talk) 13:35, 6 June 2015 (UTC)
Those are a bit tricky since technically the game didn't delete them, they just edited the existing name and changed its name/stats. Right now our ESO item database doesn't have a "history" capability but I'd like to add one so we can keep old item data like this. Note that I still have the old item data, it just isn't accessible by the item link database at the moment. I'll think about it next week and see what I can come up with. -- Daveh (talk) 20:55, 6 June 2015 (UTC)
No rush if the data's still around. —Legoless (talk) 21:10, 6 June 2015 (UTC)
Another thing that needs restoring is items that were converted to the Collections system, such as this. —Legoless (talk) 16:15, 17 June 2015 (UTC)

Google Malware Notice

If you've tried to download anything from the site today you may have noticed that Chrome/Firefox may have not allowed you to due to a malware notice. The reason is that Google thinks that File:1.2_Oldblivion_SI.zip contains malware and has flagged the site as potentially dangerous.

I don't have specific knowledge about Oldblivion_SI but scanning it online does reveal that some AV engines think it is a threat. I suppose it could contain some malware/virus but it is also possible that the nature of the program being a DLL injection that it creates a false positive. I haven't been able to find any reports of Oldblivion causing issues.

The next question is what to do with it. I've requested a review of the report by Google but chances are they'll trust the AV reports and keep the warning on the site. Simplest solution is to remove the file but it may still be a useful mod for some users. Another possible solution is to encrypt/password protect the ZIP file assuming that will stop Google from scanning inside it.

Thoughts/comments? -- Daveh (talk) 19:59, 3 June 2015 (UTC)

For now, I would say let's remove the file. The number of users who would need something like that at this point are probably in the dozens. --AKB Talk Cont Mail 20:05, 3 June 2015 (UTC)
Googling, I found the same version of the file quickly enough on FilePlanet. So, it is available elsewhere, at least for now. I'd still rather keep a copy available to users somehow, though, since we are pretty much the go to place for Elder Scrolls stuff. Personally, I'd go with either encoding it or keeping it offline but still offering it on request (which, of course, requires that you or another volunteer be willing to respond to such requests...I'd be willing, provided the requests aren't excessive). Robin Hood  (talk) 20:25, 3 June 2015 (UTC)
We could delete the file from the Wiki but move the actual file to another place and only give the link to users on request. If we don't include the actual link anywhere on the site Google can't "find" it and should be satisfied. I've already copied the ZIP to a few places so the wiki file for it can be deleted anytime if desired. -- Daveh (talk) 20:51, 3 June 2015 (UTC)
I've been poking around in it and have found a problem with the oldblivion.exe. It doesn't extract properly, and when forced to it auto-deletes. I've found a work around and repacked it so it extracts properly, but it still has two sites detecting threats. If someone wants to have a look and see if that helps. Permission to delete at the same time as the original. Silence is GoldenBreak the Silence 21:21, 3 June 2015 (UTC)
I've backed up your version as well (same filenames under content3.uesp.net/oblivion/) so we're good to delete them. -- Daveh (talk) 21:41, 3 June 2015 (UTC)
This is addressed on the FAQ: "My antivirus detects a "HEUR/BACKDOOR.Generic" virus in Oldblivion.exe / This is a false alarm. The approach Oldblivion uses to interact with Oblivion can be exploited maliciously, so some anti virus programs will consider it a trojan. Oldblivion is NOT one." Would it be possible to implement a click request, i.e. only offering the download link through a non-trawlable embed? —Legoless (talk) 22:06, 3 June 2015 (UTC)
I think so, but I'd go ahead and delete the two files now and I'll look into something else later this week when I can figure out what might work. -- Daveh (talk) 22:21, 3 June 2015 (UTC)
(edit conflict) Actually, that solution probably wouldn't help our case as far as truthworthiness goes. Instead, it would probably be best to offer the download link in plain text without a hyperlink ("uesp (dot) net") and with plenty of warnings. If we're still gonna be hosting it, it might also be smart to mark the directory as private so Google's spiders don't look at it even if they figure out the plain text URL. Something like Dropbox would also be an option. —Legoless (talk) 22:24, 3 June 2015 (UTC)

() I've gone ahead and speedy deleted them, per this discussion. While they normally wouldn't qualify, I think this could readily be considered an "emergency" situation, since we don't want great big warnings on everybody's downloads. Robin Hood  (talk) 00:06, 4 June 2015 (UTC)

As expected, Google basically ignored my initial request to ignore the file and flag it as a false positive. Since the file in question has now been removed I've requested another review and we'll see how that goes. The problem with Google is that it is hard to impossible to get an actual human there to take time and understand a unique situation like this. -- Daveh (talk) 01:38, 4 June 2015 (UTC)
Fixed -- The site is no longer flagged as hosting malware, although Google took their sweet time to get it fixed (it took two more requests on my end). I'll look into the final solution for hosting the file in question next week. -- Daveh (talk) 12:40, 6 June 2015 (UTC)

Range Block Requested

Two almost identical IPs vandalized the site repeatedly: 46.60.252.96 and 46.60.252.97. Someone should probably teach me how to do this safely, that could work too. ~ Dwarfmp (talk) 13:56, 4 June 2015 (UTC)

Step 1: For each set of numbers from the right that you want to include in the range, set it to 0. So, using the above example, you'd get 46.60.252.0.
Range blocks are based on the number of bits in the IP address. So, the base is 32 bits (8 bits per number). Given that:
Step 2: For each of the numbers you changed to 0 in Step 1, subtract 8 bits from that 32. In this case, it's only the last number that's changing, so 32-8=24. Add a slash and that number to the end of the address, so: 46.60.252.0/24. That's the number you would enter in place of the regular IP address.
Step 3: Do the remainder of the block as normal, except don't leave a block message on the talk page, because the wiki misinterprets that and treats it as a subpage of the base IP. (Besides which, you wouldn't want to leave blocked messages on hundreds or thousands of pages at once anyway.)
Normally, you'll do this for the last number (/24) or the last two numbers (/16), but there's nothing preventing you from getting fancy and calculating the exact number of bits required to block the range and zeroing only those bits. For example, to just block the two addresses mentioned, you could do 46.60.252.96/31, but that would be a silly way to block just two addresses. If you want to be sure you're covering the entire range, or you just don't feel like doing the math, you can also go to APNIC or the RIPE database and search for the IP. On the results screen, look towards the bottom for "route" and that'll give you the entire range to block for that ISP. In the very rare event that the number is less than /16 (meaning you're blocking more than 65,535 addresses at once) and you actually want to block the entire range, you have to split it up into /16 blocks for each possible combo of the first two numbers, as that's the most a range block allows you to do at once. Clear as mud? Robin Hood  (talk) 16:06, 4 June 2015 (UTC)
Haha, wow. You lost me at 'bits'. I've been reading this over a few times, I'm having trouble understanding it. If you want to e.g. block the last two numbers, is it supposed to be 46.60.0/16? And I don't really get that example of blocking only those two addresses. And something unrelated, how long should I block them for? ~ Dwarfmp (talk) 00:36, 5 June 2015 (UTC)
If you're not familiar with converting numbers to binary, it's probably too complicated to get into here, so your best bet is to stick with the /24 and /16 that are the most common, or see if Help:Range blocks makes more sense to you. I've never been that good at explaining technical things.
You've got the right idea with the /16, but there should always be four numbers in the IP address, so it would be 46.60.0.0/16.
As for how long you should block them for, it depends on how big of a range you're using and whether there have historically been valid edits in that range. A /16 where there have been other edits in that range should probably start at a week or so, and increment gradually from there if the vandalism persists. A /24 where there haven't been any edits in the range apart from the vandalism can be blocked for a year with no worries. You can check for other edits in the range by turning on the "Allow /16 and /24..." gadget. That will accept the /16 and /24 range formats above, or if you're used to wildcards, it'll also accept things like 46.60.*. Just punch it into the "User contributions" screen and it'll show you all of them (e.g., [1]).
Never block a range for more than a year and always leave e-mail and account creation enabled since, obviously, range blocks can potentially affect a lot of legitimate editors. Robin Hood  (talk) 03:19, 5 June 2015 (UTC)
Thanks. I only ever get one edit when I look it up like that, don't think it's searching as it says, or maybe it takes REALLY long. But we know there are more, so I don't know what's up with that ~ Dwarfmp (talk) 13:52, 5 June 2015 (UTC)
Hmmm...yeah, you're right. I hadn't noticed that. It certainly should pick up the edits mentioned in the original post. Maybe that's another gadget that needs an update? I'll have a look around. (Update: updating the tool fixed it.) Robin Hood  (talk) 16:44, 5 June 2015 (UTC); Update: 17:01, 5 June 2015 (UTC)
Not sure if it would help or not, but when I need to block an IP range at Fable Wiki I use this Rangeblock Calculator. --Enodoc (talk) 22:43, 5 June 2015 (UTC)
I still haven't blocked them by the way, I can't help but feel I'd block legitimate people for uncommon or rare vandalism. Just thought that having them blocked for even a short while would be an indicator for a repeated offender. I guess it's not really all that important, but hey I learned something for once, which I should have probably known already because I've been an admin for a good while. Thanks for all the info ~ Dwarfmp (talk) 14:07, 6 June 2015 (UTC)

ON Page Creation

I seem to remember back when ESO launched, there were a bunch of temporary policies put in place, including that page creation in the Online namespace was restricted to some degree, although I can't remember by how much. Would it be necessary to do that again, due to the impending console release, or is there no need for it this time? Most things that need to exist should exist already, but I'm not sure if that's reason for or reason against doing it again. --Enodoc (talk) 13:56, 8 June 2015 (UTC)

I'm tempted to leave it be. We didn't see nearly as much editing as was expected for the PC release. —Legoless (talk) 14:05, 8 June 2015 (UTC)
Plus there are quite a number of missing pages, including redlinked quests. Online is nowhere near fully created. Silence is GoldenBreak the Silence 14:06, 8 June 2015 (UTC)

Unblock Self Right

Currently, administrators have the ability to unblock themselves if they're blocked. This has both good and bad sides to it, so I wanted to bring it up to see if there was a consensus to keep that in place or to change it. The problem occurs in the event of a rogue or hacked admin. Currently, if another admin blocks them, they can just go unblock themselves and go right back to what they're doing. Having the right to unblock themselves also allows an admin to test-block themselves if they need to for whatever reason, then undo it immediately rather than waiting for it to expire. On the flip side, however, if admins don't have that right, then an administrator who gets blocked will stay blocked until the block is removed...but then a sufficiently knowledgeable and fast rogue/hacked admin can block every other admin on the site, and I believe only bureaucrats (currently only Dave) and in a more roundabout way, server admins (currently only me) would be able to reverse that.

I'm not advocating for either side, but since either way can affect our ability to deal with rogue or hacked admins, I thought I'd bring it up to see how people feel. This is clearly not a pressing problem, as I think we've only ever had one admin unblock themselves that wasn't for testing purposes, but it makes more sense to me to debate the matter when it's not a problem than having the debate while there is a rogue/hacked admin issue occurring. :) Robin Hood  (talk) 20:26, 10 June 2015 (UTC)

So if the unblock self rights were taken away and an administrator would become blocked, would they not be able to block/unblock others users while they were blocked? Just seeing if I'm understanding the issue you mentioned with the hacked admin blocking all the other admins correctly. Forfeit (talk) 21:43, 10 June 2015 (UTC)
That's correct. Once you're blocked, without the unblock self right, you can't take any actions on the wiki at all except editing your own talk page. Robin Hood  (talk) 22:16, 10 June 2015 (UTC)
I'm not sure I see a problem here. If a rogue admin unblocks themselves, a Bureaucrat (Daveh) can revoke their adminship and thus remove their ability to block anyone. Likewise if they go blocking all the other admins. Surely that kind of behavior would get Daveh's attention (especially since most of us know ways of contacting him outside of UESP. The only potential downfall I see is if a hacked admin account blocks all the other admins and then goes and immediately starts deleting files. (Page deletes can be easily undone, but not files.) However, I believe even this would be reversible since the entire site is duplicated on multiple servers (though it would probably be a pain to fix). At any rate, I don't see either as a particularly likely scenario. As you said, we've only ever had one (possibly two?) admin unblock themselves in the history of the site. It's not like Wikipedia where there are thousands of admins. We have a considerably fewer people to keep track of, and the odds of any of us doing serious damage (even if someone hacked our account) before being stopped are pretty low. Preventing someone from unblocking themselves would not really change much in that respect. — TheRealLurlock (talk) 22:35, 10 June 2015 (UTC)
I just tried deleting/undeleting images on my test wiki, which is also on 1.22, and it all worked fine. Was this an issue on older versions of the wiki, maybe? Also, you got me curious, so I checked all current and former admins and there's only one case of self-unblocking that wasn't a test.
As far as the issue goes, now that I give it some more thought, I think it's probably safest in the long run to leave the right there, since being able to unblock yourself to reverse a rogue/hacked admin's damage is probably more important than the rogue admin unblocking themselves. Robin Hood  (talk) 23:16, 10 June 2015 (UTC)
If only I had known you could unblock yourself when I accidentally blocked myself instead of the vandal, yeah... But yes, it seems best to keep this feature, I can hardly imagine a rogue admin, possible but not likely ~ Dwarfmp (talk) 14:19, 12 June 2015 (UTC)

Possible plagiarism

I believe that the Alien Species wiki has plagiarized some of our content. Their Sload article is pretty much copied straight from our article (specifically the biology section) on it, Tsaesci is also copied directly from our page and Ka Po' Tun is copied as well. Here is the category for all their TES related pages. I don't know if there is anymore copied content, but I don't feel like going through 32 more pages. Lorenut (talk) 11:50, 5 July 2015 (UTC)

No, it is not allowed. Well, technically, in this case, it would be allowed, if they had acknowledged us as the original creators of those articles. They use CC-by-sa as well, but they are not fulfilling all of the requirements to have shared those articles. --AKB Talk Cont Mail 17:30, 5 July 2015 (UTC)
Update: An admin has gone through and added the required stuff, so I figured I'd update you on the situation. Lorenut (talk) 18:34, 6 July 2015 (UTC)
I don't see the point in making a new section for this, so I'm just going to post it here. The Maormer page on TESWiki has some plagiarized content. It's specifically in the history section of the article. It's almost identical to our article on the subject. Lorenut (talk) 17:28, 10 July 2015 (UTC)
After checking both pages' histories, it looks like both pages underwent overhauls on 18 June 2013. The users have different usernames on the wikis, but it's reasonable to believe that they are the same person, in which case I'm not sure this would classify as plagiarism because it is their own work after all. •WoahBro►talk 18:24, 10 July 2015 (UTC)
I just checked the userpage of the person who edited the article on TESWiki and you are correct. The person specifically mentions his/her account by name on this site and it's the same one that edited the article here. Lorenut (talk) 19:00, 10 July 2015 (UTC)

Redundant Images

A question has come up recently about what exactly constitutes a "redundant image" for the purposes of our Deletion Policy. The policy currently states that redundant images can be speedily deleted "when there exists another copy that is of the same or better quality and of the same image file format". We've been applying this policy a little arbitrarily (sometimes also ignoring the same-format requirement, though that's easy to miss), so I want to bring it up for discussion. Some questions that come to mind are:

  • Does this only apply to absolutely identical images (apart from resolution), or is it sufficient that an image is nearly identical or at least roughly the same shot? Or do the terms "redundant" and "better quality" mean that the new one is better and is intended to replace the old one in all the same places, regardless of the actual image (i.e., if someone should really have uploaded over the old image, but failed to do so).
  • What if nearly identical images are uploaded by different authors?
  • What if absolutely identical images are uploaded in different formats?

Answering my own questions, I think that it should only apply to perfectly identical images or those that are "functionally identical", for lack of a better term (i.e., different only in terms of cropping or other minor issues of that nature). Authorship makes a difference to me: if two different authors upload the same shot, we shouldn't speedily delete them as their may be lighting or colour issues that could be deemed better or worse between them. Finally, on the question of format, I think it should only be taken into account when it's actually important—for example if the image is intended for use as an icon and transparency is a concern. If an image is otherwise identical, I'd say the version in the non-preferred format (usually anything other than JPG) can be speedily deleted. Anyone else have any thoughts on this? Robin Hood  (talk) 02:07, 14 July 2015 (UTC)

I've never noticed that line about the "same image file format" before, and as far as I'm concerned it needs to go. We've been speedy deleting re-uploaded PNGs for years, and I don't see any reason an identical image in the wrong format would need to go through the usual deletion process. —Legoless (talk) 03:04, 14 July 2015 (UTC)
I agree with your opinion. I've been deleting my own images with "replaced by better image" when I made a mistake, or changed something slightly, or I simply took another shot in the game of the same content. Speedy is good for identical, or near identical (something like cropping and uploading under a different name, or changing the format) uploads by the same or different authors. Anything else should probably be prodded instead, unless I'm missing another scenario I'm not thinking of now ~ Dwarfmp (talk) 16:56, 14 July 2015 (UTC)
The intent of a speedy deletion is to delete an image that shouldn't have any objections to that deletion. I don't see anyone objecting to a deletion because of image format, so I definitely agree with legoless that it should go. Jeancey (talk) 17:01, 14 July 2015 (UTC)

Slow Page Loads

This is technically an issue left-over from the upgrade to v1.22 a few months ago but it is a new and large enough issue to warrant a new section. Some of the slow page loads around the time of the upgrade were seen to be due to thumbnail regeneration. At the time it was assumed to be a temporary issue and once thumbnails had been created it would disappear. However, the slow page load issue has persisted, even though intermittently, and some investigation shows it is the same issue. This seems to be more of an issue with pages with many images/icons on it. The first page load is slow and subsequent page loads after that are OK, at least for a while and then you get another slow page load.

Profiling data for these pages on first load (or when purging the page) look similar (full profile log):

  Start request GET /wtest/index.php?title=Online:Draconic_Power
  ...
  File::transform: Doing stat for mwstore://local-backend/local-thumb/7/75/ON-icon-skill-Dragonknight-Dragon_Leap.png/48px-ON-icon-skill-Dragonknight-Dragon_Leap.png
  BitmapHandler::doTransform: creating 48x48 thumbnail at mwstore://local-backend/local-thumb/7/75/ON-icon-skill-Dragonknight-Dragon_Leap.png/48px-ON-icon-skill-Dragonknight-Dragon_Leap.png using scaler im
  BitmapHandler::doTransform: Transforming later per flags.
  File::transform: Doing stat for mwstore://local-backend/local-thumb/7/75/ON-icon-skill-Dragonknight-Dragon_Leap.png/64px-ON-icon-skill-Dragonknight-Dragon_Leap.png
  ...many more lines...
  File::transform: Doing stat for mwstore://local-backend/local-thumb/d/d1/ON-icon-skill-Dragonknight-Inhale.png/48px-ON-icon-skill-Dragonknight-Inhale.png
  BitmapHandler::doTransform: creating 48x48 thumbnail at mwstore://local-backend/local-thumb/d/d1/ON-icon-skill-Dragonknight-Inhale.png/48px-ON-icon-skill-Dragonknight-Inhale.png using scaler im
  BitmapHandler::doTransform: Transforming later per flags.
  File::transform: Doing stat for mwstore://local-backend/local-thumb/d/d1/ON-icon-skill-Dragonknight-Inhale.png/64px-ON-icon-skill-Dragonknight-Inhale.png
  FileBackendStore::getFileStat: File mwstore://local-backend/local-thumb/d/d1/ON-icon-skill-Dragonknight-Inhale.png/64px-ON-icon-skill-Dragonknight-Inhale.png does not exist.
  BitmapHandler::doTransform: creating 64x64 thumbnail at /home/uesp/www/wtest/images/tmp/transform_e6d56efbd3d5-1.png using scaler im
  BitmapHandler::doTransform: returning unscaled image
  ...
  Profiling data
  Name                                                                         Calls         Total          Each             %       Mem
  -total                                                                           1      7500.083      7500.083       100.000%  47225808  (     7500.083 -     7500.083) [0]
  Parser::internalParse                                                           67      7267.480       108.470        96.899%  13828149  (        0.616 -     6885.161) [29285]
  MediaWiki::main                                                                  1      7236.824      7236.824        96.490%  14306539  (     7236.824 -     7236.824) [24092]
  MediaWiki::performRequest                                                        1      6996.978      6996.978        93.292%  11771717  (     6996.978 -     6996.978) [22222]
  MediaWiki::performAction                                                         1      6995.122      6995.122        93.267%  11628009  (     6995.122 -     6995.122) [22201]
  Article::view                                                                    1      6994.832      6994.832        93.263%  11616052  (     6994.832 -     6994.832) [22189]
  Parser::parse-WikitextContent::getParserOutput                                   1      6971.854      6971.854        92.957%  10714452  (     6971.854 -     6971.854) [21993]
  Parser::parse                                                                    1      6971.853      6971.853        92.957%  10714433  (     6971.853 -     6971.853) [21994]
  Parser::replaceInternalLinks2                                                  133      6240.923        46.924        83.211%   2595425  (        0.022 -     6236.272) [4569]
  Parser::replaceInternalLinks2-image                                             66      6222.354        94.278        82.964%   1645565  (       44.437 -      211.638) [3135]
  File::transform                                                                198      5889.116        29.743        78.521%   1305328  (        0.304 -      139.120) [1788]
  TempFSFile::factory                                                            102      5666.845        55.557        75.557%    141984  (       38.778 -      138.712) [0]
  Parser::braceSubstitution                                                     1604      3438.901         2.144        45.852%  29494967  (        0.082 -      324.206) [103548]
  Parser::braceSubstitution-pfunc                                               1498       731.916         0.489         9.759%   7593625  (       -0.000 -       46.009) [22243]
  Parser::callParserFunction                                                    1060       725.037         0.684         9.667%   4888320  (        0.011 -       46.001) [21183]
  ...

While the ultimate cause and solution for the issue hasn't been found a few things noted so far are:

  • Some of the thumbnails looked for actually exist already but are not being found for whatever reason.
  • Some of the thumbnails do not exit and are not being created.
  • It is assumed the image data is being stored in the cache (memcached) for some time before being purged/overwritten and causing another slow page load.
  • There are no obvious issues or errors noted (no Apache errors, no NFS errors, no write/read permission issues).

I'm assuming that something in MediaWiki between 1.19 and 1.22 was changed regarded thumbnails but haven't been able to narrow it down yet. I'll continue to look into things this week and hopefully have it solved soon. Ideas or comments welcome. -- Daveh (talk) 18:31, 22 July 2015 (UTC)

Could we put in a request to MediaWiki for help? I think our custom extensions have made getting help hard in the past, and I'm not seeing anything obvious that would be the issue from a quick glance at other MediaWiki issue reports (mind you, I'm tired and not reading very well right now). Possible idiot question, but shouldn't we also be able to load the page, even if the images do not? From what I understand, a page can load separately from its images, that's why on terrible connections you can see pages load well before its images come up. How are the images both failing to load, and stopping the page from loading as well? --AKB Talk Cont Mail 09:25, 23 July 2015 (UTC)
If my prodding around doesn't reveal anything I'll post a question to the MW tech mailing list. A few more findings however. It seems most (or maybe all) of the slow pages are on the Online skill pages. I can't find any slow page loads in other namespaces and icons in those namespaces appear to be created normally. Does anyone know of any none-Online page that is consistently or intermittently slow lately? I might suspect the family of Template:ESO_Active_Icon templates, are there any similar templates in other namespaces I can test? -- Daveh (talk) 13:59, 23 July 2015 (UTC)
I don't think there are any similar templates; the parent template {{Image Overlay}} only seems to be used for the ESO Active/Passive/Achievement Icon templates. Could it be that the culprit is the (rather excessive) number of template parameters that are #saved by {{Online Skill Summary}} on those pages for loading elsewhere? If so, sorry, that was my fault; I was hoping to cut down on the amount of duplicate info that needed to be changed following a major update by pulling the skill info from the pages straight to the skill lines pages. --Enodoc (talk) 17:20, 23 July 2015 (UTC)
I'll keep my eyes open, but I don't think it's just Online skill pages that are the problem, as I have very few Online pages in my watchlist. Enodoc may have a point, though, perhaps it's any page that #saves a lot of data. From what I've seen of the code, Metatemplate could use a rewrite to modernize it, as it was written way back on MW 1.10, I think, and a lot of new things have come into play since then. My PHP coding skills probably aren't up to that complex of a task, but if anyone else has the skills and time, I'd be happy to help out however I can. But I suppose we should make sure that's the problem first. :) Robin Hood  (talk) 20:05, 23 July 2015 (UTC)
A little more info...if I include an image like File:ON-icon-skill-Dragonknight-Dragonknight_Standard.png as a normal thumbnail it is created normally if it doesn't exist (i.e., I delete the thumbnail manually on the server). But when viewed within a page like Online:Ardent_Flame the thumbnail is not created if it doesn't exist on a normal page load. It is created if you purge the page however. I assume the thumbnail is created on demand and kept in the cache for a while. I'm not too familiar with the Online skill icon templates but I'll poke around in them a bit to see how they work and why it would circumvent the normal thumbnail creation.
The slow page loads could well be a multiple-issue problem. Once I fix this particular one I'll keep looking and see what else I can find. Examples of slow pages, whether consistent or intermittent, are useful. -- Daveh (talk) 21:11, 23 July 2015 (UTC)
The Contraband pages (especially the larger ones like S) usually take quite a while to load/preview/save for me. They're all full of icons too. Don't know whether this is related to this issue. Contraptions (talk) 10:49, 25 July 2015 (UTC)

() Just for reference, I just had longish page loads on Lore:Khajiit Names opening the three most recent diffs (1, 2, 3). #1 loaded fairly quickly, but the other two took about 20-30 seconds. Noticeably, there are no icons on the page and no #saves. Given that it wasn't the huge delays I've seen on some pages (> 1 min), I'd venture a guess that we have more than one issue. Of course, it could just be a fluke, but I thought I should mention it in case it's not. Robin Hood  (talk) 18:03, 30 July 2015 (UTC)

99% of the time everything loads fine, but I just tried to open a few pages to patrol and had to wait 2-3 minutes before they loaded while I could load things normally on non-UESP sites. Then suddenly they loaded and I could get to here just fine. Just something strange I thought I'd mention. •WoahBro►talk 19:16, 11 August 2015 (UTC)

ESO Imperial City Update

With the PC release of the Imperial City today there are a bunch of things to do. My immediate list has:

  • Dump all the new images/icons and upload them
  • Update maps
  • Dump and update the item data
  • Actually play the game (hopefully)

It will probably take a few days to get everything done. If there are other admin related things to do related to the update just post them here. -- Daveh (talk) 15:15, 31 August 2015 (UTC)

Patch notes are up and available here. -- Daveh (talk) 16:00, 31 August 2015 (UTC)
Files extracted and uploaded to:
Each ZIP has a .diff.txt text file showing the changes/additions/removals for the files. The maps have been uploaded and added with the relevant ones at:
If there are other files people need or you are looking for a specific one just let me know. -- Daveh (talk) 16:51, 1 September 2015 (UTC)
For reference, these uploaded maps, which were replaced in-game in PTS 2.1.2 by the large sewer overview map and therefore are no longer present in the data files, are still entirely accurate, and so are still valid if we want to use them on the respective articles:
File:ON-map-Imperial Sewers Ebonheart 03.jpg
File:ON-map-Imperial Sewers Daggerfall 03.jpg
File:ON-map-Imperial Sewers Aldmeri 03.jpg
File:ON-map-Imperial Sewers Hub.jpg
File:ON-map-Imperial Sewers Ebonheart 02.jpg
File:ON-map-Imperial Sewers Ebonheart 01.jpg
File:ON-map-Imperial Sewers Daggerfall 02.jpg
File:ON-map-Imperial Sewers Daggerfall 01.jpg
File:ON-map-Imperial Sewers Aldmeri 02.jpg
File:ON-map-Imperial Sewers Aldmeri 01.jpg
Enodoc (talk) 22:08, 1 September 2015 (UTC)
I think it would be a good idea to keep using them, as long as we mark them as 'unused'. The overview map is clunky enough to use in-game where you can zoom in, but trying to use it on the individual wiki articles wouldn't really be practical. We already use old maps on the subzone pages. —Legoless (talk) 02:59, 2 September 2015 (UTC)
I've added some new map icons and updated a few on the map so far. The item dump is also finished and complete along with a few tweaks/fixes to get item links displaying better. I'm working on a skill dump from the game as well but this will take a bit more time to get usable. -- Daveh (talk) 13:36, 8 September 2015 (UTC)
Updated the ESO data/API at esodata.uesp.net/100012/ for the new API version. -- Daveh (talk) 16:05, 8 September 2015 (UTC)

ESO Skill Data

I've been working on getting the skill data mined out from ESO and have most of it done. The relevant links from esolog.uesp.net/viewlog.php are:

  • Mined Skills -- All the 36,000 mined skills from the game. Most are internal and probably not of much use.
  • Skill Lines -- The relevant player skills parsed for each skill line (e.g., Templar::Aedric Spear). Use the right-hand View Skills to view all the individual skills in each line.
  • Skill Tree -- The mined skills parsed into each skill line including its ranks and morphs.

It is not 100% complete but is relatively close...the one main thing I'm missing are the progression for the all the passives which doesn't appear to be available from the game API. There also appears to be a few passives missing for some skill lines.

The main purpose of this is to help update all the ESO skill pages, some of which are grossly out-of-date. Comments and bug notes on this welcome. -- Daveh (talk) 14:54, 16 September 2015 (UTC)

Site Search

I've been trying to fix the site search for a while now with no luck and have moved on to replacing the old (and apparently deprecated) Lucene search with ElasticSearch/CirrusSearch that MW has moved to. Fortunately, it was much easier to setup and get working than Lucene over on the Dev Wiki, although a complete index time takes a full day to do unless I can figure out a way to make it faster (fortunately, its a one-time thing).

It seems to work fine with the following notes/observations:

  • Page edits are immediately reflected in the search results.
  • All content namespaces are searched by default.
  • Seems speedy enough even on the dev wiki (in theory it should be a little faster on the main wiki).
  • Prefix searches (results that show up when typing) currently seem to require the namespace. So typing "vampire" wouldn't show any results but "skyrim:vampire" or "sr:vampire" would show the relevant page titles. The final search results after you submit a query work normally whether you specify a namespace or not. There doesn't seem to be any obvious way in the current CirrusSearch version to get the prefix search to use all namespaces but we are a few versions back. I'd like to get MW upgraded to 1.25/1.26 in the next few months and see if that changes anything and then look at custom changes if needed.

Feel free to play with it and let me know of any issues. I'd like to get it setup and running on the main wiki next week. I should be able to get it to index all pages, test it on content3, and then roll it out on content1/2 without any site interruption. If there are issues it is a simple matter to roll back to the only half-broken Lucene search. -- Daveh (talk) 15:22, 18 September 2015 (UTC)

This is only slightly related to the search thing, but the ability to search only books or only npcs in the ESOLog is fantastic :). Thanks! Jeancey (talk) 20:53, 18 September 2015 (UTC)
Great! Glad you like it too...most useful to find certain things. I've been slowly working on EsoLog adding simple features like this. -- Daveh (talk) 01:34, 19 September 2015 (UTC)
I'm glad this is getting fixed. If we can save predictve searches without prefixes, that'd be ideal, as it really does open up the site I feel. It can be a bit distressing if when you type Vampire, you don't get any results up right away. I kind of feel we'd do better with predictive searches turned off until this is resolved, so users don't feel like their search query is incorrect. I don't know, how do you guys feel? --AKB Talk Cont Mail 22:07, 21 September 2015 (UTC)
I'll look into seeing if I can get predictive searches working "as expected" without too much hacks. Disabling it should be easy if needed. -- Daveh (talk) 18:44, 22 September 2015 (UTC)

Update -- I'm in the process of installing and creating the live search index using ElasticSearch/CirrusSearch. The site search is currently still using the old LuceneSearch and I'll post here when I do the switch either later today or tomorrow, depending on how long the reindex takes. -- Daveh (talk) 18:44, 22 September 2015 (UTC)

The live site is now using the ElasticSearch index. Let me know of any problems here and we can revert back to the Lucene search index if there are any serious issues. A couple of quick tests seems to show it working and updating well. The prefix/predictive search also seems to work better than on the dev wiki for some reason (e.g., vampire gives an assortment of results in various namespaces). If that needs improving or tweaking let me know. -- Daveh (talk) 23:44, 22 September 2015 (UTC)
Everything seems to be going well. The server loads on files1 (which hosts the ElasticSearch index) were fine last night. I've noticed a few oddities in the search results, like "skyrim books" not showing the obvious Skyrim:Books page at the top of the search results, but relatively minor. Time permitting I'll come back and see if results like this can be better tweaked for our MW layout. -- Daveh (talk) 15:52, 23 September 2015 (UTC)

Slow Page Loads (Revisited)

Another issue I've been trying to solve since we upgraded to v1.22 and have managed to make a little progress. A lot of poking around in the depths of MediaWiki code and profiles seem to indicate two main causes:

  1. Slow temporary image creation
  2. Complex/deep templates

Note that the slow pages loads in question affect only logged in users and only on the first page load. After that the parsed page is cached for around a day unless you do a page purge (?action=purge) or edit.

The slower pages in question, like Skyrim:Books and all the Online:Skills have both problems. The image problem was linked to having the temporary path ($wgTmpDirectory) set to the image upload directory on the NFS share. File creation on a NFS share seems to take 50ms which isn't too long until you consider many dozens of images being created in one page load. Changing $wgTmpDirectory to a local directory fixes this problem and makes most of the Online:Skills pages much quicker to load (from ~10 seconds to ~2 seconds).

This still leaves the complex template issue which seems to be the reason that Skyrim:Books still takes ~30 seconds to load the first time around. This is because MediaWiki has to load and parse each of the 470 included articles and any templates each of those use. At the moment it takes so long to load I can't get profiling data but will continue to try and see if its fixable without having to go back to a simpler page content with duplicated data. -- Daveh (talk) 16:13, 23 September 2015 (UTC)

Finally got some profiling data on Skyrim:Books which demonstrates the issue with complex templates:
    Function = Parser::braceSubstitution
    Calls = 35319
    Time = 90664.886 ms
    Average Time/Call = 2.567 ms
If I understand the data correctly, this means there are 35319 template brackets that have to be loaded, parsed, and variables substituted. I'm going to try upgrading the dev wiki to 1.25/1.26 and see if there happens to be a large speed improvement by some chance but I think we have to look at redoing the structure of at least this page and any others like it. -- Daveh (talk) 16:36, 23 September 2015 (UTC)
It should be pretty simply to change all instances of {{icon}} to use regular embedded images, although obviously for editing purposes having a template is much handier. —Legoless (talk) 18:56, 23 September 2015 (UTC)
Either you just changed something or caching kicks in here as well. I just loaded Books and it took about 40 seconds, but then subsequent reloads only took a couple of seconds. Even purging, it only took about 10-12 seconds to reload. In any event, the structure of the template could probably be improved a bit as well. I'll play around in a sandbox and see if there's anything I can do anything that'd make a significant difference. If page loads stay low, it could be an issue, but we'll see what happens. Robin Hood  (talk) 19:19, 23 September 2015 (UTC)
The ESO skill pages seem to load fine now even on a purge (2-3 seconds) so I'm not too worried about them. The books page is a concern along with any similar page taking more than 10 seconds to do a full purge/parse. After you load it once the page is saved in the object cache (memcached) for 1 day (I think) so subsequent loads will be faster. You can force a "slow" load by purging the page.
I've found another small thing to increase performance...use memcached to store the session keys. Right now the session keys are stored on a NFS share and PHP locks the session file while a page is being loaded. This prevents other asynchronous loads from happening at the same time (like the skin/javascript requests that go through MW). Changing to use memcached should help a little bit in some cases and much of performance tweaking is finding a bunch of little things that add up to a lot. The only thing from stopping me from doing it right now is that our editable maps use the session files so I'll have to add code to let the maps access the session via memcached as well. -- Daveh (talk) 19:27, 23 September 2015 (UTC)
The #load is clearly the biggest issue with the template, but I've tried to optimize other portions of it anyway. While it's difficult to tell with such wide variations in timing on my end, I believe there's at least a bit of an improvement in load times now. Robin Hood  (talk) 21:17, 23 September 2015 (UTC)

() Session Note -- It was relatively easy to switch the maps over to use the sessions in memcached so I'll be changing that on the servers sometime soon. This may/likely log you out and require you to log back in but this is a one-time thing. A couple quick tests show a noticeably speed improvement on page reloads...from 5 seconds down to 2 seconds. The server is currently spending those 3 seconds just waiting for the session file lock to clear. If there happens to other side-effects from this change just let me know. -- Daveh (talk) 21:29, 23 September 2015 (UTC)

I've made the change to have session data stored in memcached. I can confirm that you'll need to re-log in once. All the map's edit features should be working but if you run into issue with it or anything else let me know. This small change has definitely speeded things up a bit. A reload of the main page went from ~5 sec to ~2 sec. -- Daveh (talk) 23:41, 23 September 2015 (UTC)


Prev: Archive 35 Up: Administrator Noticeboard Next: None