Guild Wars 2 Wiki:Reporting wiki bugs/archive 8

From Guild Wars 2 Wiki
Jump to navigationJump to search

Key: Yes = Solved, No = Unsolved, No = Not a bug

No Animated Gifs

So I tried uploading a .gif, and am having no luck. At first I thought it was too big, so I made it smaller, then I moved it, and tried again under a different file name, still no dice. Not sure if it's something I did when creating the file or on the wiki's end but: File:User Doodleplex please work please.gif. - Doodleplex 03:30, 25 January 2017 (UTC)

Works fine for me, assuming it's supposed to have 7 frames 400ms each. Thumbnail on the video gives an error code though:

Error creating thumbnail: /usr/bin/timeout: the monitored command dumped core /var/www/sites/gw2w-en/includes/limit.sh: line 101: 11815 Aborted /usr/bin/timeout $MW_WALL_CLOCK_LIMIT /bin/bash -c "$1" 3>&-

Error code: 134

Works fine putting it in a page though --Gimmethegepgun (talk) 03:55, 25 January 2017 (UTC)
Here's the thing, if anything should have a bug, I would think it would be File:Nevermore drawn animation.gif, larger, and has more frames, whereas this one is much smaller and only has 7 frames. I'd like it to be a thumbnail image to be used on the legendary pages, but for some reason for this file, it's just not working, and I'm not sure if it's me or the wiki, because if it's me, I'd like to know what to fix. - Doodleplex 18:40, 25 January 2017 (UTC)

Yes Superior Sigil of Fire

Seems one of the queries there is pulling up stuff it shouldn't. SarielV 20 x 20px 16:41, 25 January 2017 (UTC)

My best guess is a hiccup due to assigning the page as a redirect. I'm not sure if properties would be passed on that way, so the edit may be a red herring. G R E E N E R 17:20, 25 January 2017 (UTC)
Like I said over at Template talk:Title achievement table, Faithful (skin) also has this error so I agree it's something to do with that edit to Faithful. —Azurem 17:28, 25 January 2017 (UTC)
This is related to having run robots at a time when the wiki was behaving in a very buggy-as-hell way, combined with poor oversight by myself.
I've blanked the relevant pages - and I will leave them blank for twenty four hours to try and clear the redirect cache. -Chieftain AlexUser Chieftain Alex sig.png 18:25, 25 January 2017 (UTC)
You might want to temporarily protect as sysop edit only on them from people editing them back to ensure it clears the cache. - Doodleplex 18:28, 25 January 2017 (UTC)
That would suggest I'm not grasping at straws here for a solution. (which I am) -Chieftain AlexUser Chieftain Alex sig.png 18:32, 25 January 2017 (UTC)
So, it didn't work at all. Ideas? If pywikibot was working, I'd delete all the redirects pointing to Title, purge Title, then restore the same pages. -Chieftain AlexUser Chieftain Alex sig.png 22:12, 26 January 2017 (UTC)

(Reset indent) The only thing I can think of is to list it manually for the time being until your bot is working or just don't have a list. Not all Superior Sigil pages have a list of weapons due to how many weapons had the sigil sometimes, for example Superior Sigil of Accuracy has a crazy long list. (Or option 3: we wait until Poke takes out his magic wand and fixes it) - Doodleplex 22:21, 26 January 2017 (UTC)

Since straight up deleting is not an option, would moving the pages without redirect to another name (say maybe under a user space) for 24h do the same thing? I'm not sure if move without redirect is equivalent to delete, but if it is, that might be a decent compromise? -Darqam 23:39, 26 January 2017 (UTC)
If a file is moved without leaving a redirect, it has essentially been deleted. I imagine the same applies for pages, though whether or not it will fix this, I dunno. - Doodleplex 23:42, 26 January 2017 (UTC)
Is there a reason why we need pywikibot for this? Why not disable all the redirects to Title with AWB and purge/null edit the page itself (manually) to update the properties, then undo the changes to the redirect pages again with AWB? This is the current list of redirects for Title.
Just so I understand this correctly from smw:Help:Inferencing#Equality of pages: redirects: all properties from the target page and from all redirects are shared. However, with the edit to Faithful, a bug(?) allowed the old properties to haunt the Title page and all redirects? spooky —Nefastu 01:08, 27 January 2017 (UTC)
So the good news is that (A) I've written an interface that can delete/restore/purge lists of pages (which will be useful to me anyway), and (B) the bug is fixed.
Ultimately the issue wasn't fixed by blanking + reverting, however it would have been fixed if I'd just straight up deleted the "Title" page, and then restored it. That's a learning point there.
I only realised that after deleting and restoring 170 pages. Seems fixed for the moment. -AWB Alex User AWB Alex sig.png 21:42, 27 January 2017 (UTC)
Thanks mr robot. -Chieftain AlexUser Chieftain Alex sig.png 21:44, 27 January 2017 (UTC)
どうもありがとう. G R E E N E R 22:22, 27 January 2017 (UTC)
I hope that's "Domo arigoto Mr. Roboto" in Japanese.... XD - Doodleplex
Alex, "Bouncer" didn't get restored, was that on purpose? - Doodleplex 23:08, 27 January 2017 (UTC)
Silly human -Chieftain AlexUser Chieftain Alex sig.png 23:36, 27 January 2017 (UTC)

No Bottle of Airship Oil icon missing from templates

I know this is not the place, but the icon for Bottle of Airship Oil seems to be linking to the wrong file in some templates, as seen in Fulgurite or Ley Line Spark, for example. The article - Bottle of Airship Oil - doesnt seem to have this problem. 46.53.200.113 11:18, 22 February 2017 (UTC)

This is the correct place, since it deals with bugs on the wiki. I can't, however, replicate your issue. Are you referring to the icon in the recipe list? —Ventriloquist 22:39, 22 February 2017 (UTC)
Can confirm, problem has miraculously disappeared. It definitely wasn't my browser cache problem, but oh well. 46.53.200.113 09:56, 23 February 2017 (UTC)

No .gif issues revisted

Was poking around, and I'm starting to wonder if .gif files and/or their thumbnails in general are starting to not work well on the GW2 wikis after the last update. I noticed fr:Fichier:IruleManik-Ménestrel.gif has a broken thumbnail and it was uploaded on the French wiki, I found de:Datei:Abkürzung Nicht so geheim.gif on the German wiki, there's my broken .gif above, and I was checking something on an older .gif and it broke as well as soon as I hit "refresh" (File:Bonfire animation.gif). - Doodleplex 19:16, 25 February 2017 (UTC)

Yes Expression errors still on the loose!

Remember Drop rate research gives out errors in table? Well that's still an issue with some of the tables. See here or here for some examples. Sometimes it works. Sometimes not. Glad I'm not the one who has to Nordic-wiki-fu this... WormholeGo ahead and talk. ≈ 06:57, 13 March 2017 (UTC)

I'll take another look. Thanks for the heads up. It's related to the way SMW formats the results again (en: 10,000 (Unbound Magic)\n50 (Gold) vs fi: 10 000 (Unbound Magic)\n50 (Gold)) however I don't know yet what we can do about record properties. -Chieftain AlexUser Chieftain Alex sig.png 07:29, 13 March 2017 (UTC)
Okay applied another bandaid to Template:Coin this time. -Chieftain AlexUser Chieftain Alex sig.png 20:28, 13 March 2017 (UTC)

Yes Tables reporting vendors as historical

Using {{vendor list}} to display items purchased with an item (rather than currency) marks all as historical, even though they are not (example). No idea when the behavior started --Gimmethegepgun (talk) 23:27, 15 March 2017 (UTC)

I can help a little with the why and when. I asked here if it would mark vendors who were historical but who were showing up as not historical, so AFAIK it's more of this template causing the issue than a wiki bug. Unforunately I still dont' know enough code to fix it. =( - Doodleplex 23:49, 15 March 2017 (UTC)
Switched #if: for #ifeq:, problem resolved I think. -Chieftain AlexUser Chieftain Alex sig.png 00:33, 16 March 2017 (UTC)

Yes Armorsmith Vendor list showing up strangely

Every armorsmith (vendor) NPC table shows a string of code instead of the rarity/level/cost. It instead shows as: style="text-align:center" style="text-align:center" | - Khavanya (talk) 21:05, 11 April 2017 (UTC)

Thanks for the report. The issue was caused by my last edit to {{vendor table row result format}} (I didn't anticipate blank rarities would cause empty content hence the cell didn't appear and the formatting did instead) fixed. -Chieftain AlexUser Chieftain Alex sig.png 21:19, 11 April 2017 (UTC)

No Interactive Maps are Broken

I've tried looking at this on three different browsers and two different computers. All of the interactive maps don't show anything but blackness. Has anyone else noticed this? --72.182.237.74 03:05, 12 April 2017 (UTC)

Not our problem sorry, the SSL certificate for the tile and render server has expired on ArenaNet's end. They know about the problem so hopefully they can resolve this quickly.
The good news is that we have local jpeg maps too (top of each area page). -Chieftain AlexUser Chieftain Alex sig.png 06:35, 12 April 2017 (UTC)
And it's now been fixed. -Chieftain AlexUser Chieftain Alex sig.png 18:24, 12 April 2017 (UTC)
It's broken again. all the interactive maps i've tried to view today since about mid afternoon have shown nothing but a big black rectangle. ZianaSue (talk) 05:49, 5 June 2017 (UTC)
Please be specific. Which maps did you try, what's your browser, are you running any weird extensions? http://imgur.com/a/2yNh5 - Draconis Mons (large area article embed), Defeat Agent Xinn (an infobox dynamic map) and the world map widget are working for me at least. -Chieftain AlexUser Chieftain Alex sig.png 06:33, 5 June 2017 (UTC)

Yes Special:RecentChanges viewed with Chrome on mobile platforms

Hi guys, I've been figuring out the cause of a text display bug visible in Chrome on Android phones (I suspect safari mobile would have a similar issue). Screenshot of visual bug.

The problem boils down to Chrome automatically enlarging text elements so that viewers can click on them more easily (and was not caused by our gw2 CSS). This of course looks pants when the collapsible elements weren't being size-adjusted. The solution is outlined here, and resulted in my addition of the following line of html via MediaWiki:Vector.js:

 $('head').append('<meta name="viewport" content="width=device-width, initial-scale=1">');

Yes Weird look

So for some reason when I load out an page (not in edit mode) I get it looking like http://prntscr.com/hnfivq. For some reason edit pages look fine. This is happening on firefox 57.0.2 64bit. On chrome the whole look is wintersday-y. So my guess is loadout is not fully changed? -Darqam 14:01, 14 December 2017 (UTC)

Sounds like a problem on your end. Cleared cache and cookies? (note: I've tested appearance and found it working in chrome, firefox, edge, + firefox mobile) -Chieftain AlexUser Chieftain Alex sig.png 18:21, 14 December 2017 (UTC)
FWIW I'm seeing it too! --Stephane Lo Presti talk 19:08, 14 December 2017 (UTC)
Derp you're both running Vector aren't you. I can reproduce this bug. -Chieftain AlexUser Chieftain Alex sig.png 19:19, 14 December 2017 (UTC)
Should be fixed. [1] -Chieftain AlexUser Chieftain Alex sig.png 19:22, 14 December 2017 (UTC)
It's working now, thank you Alex! By the way, any plans to use the festive French wiki skin? (I'm just curious) --Stephane Lo Presti talk 19:23, 14 December 2017 (UTC)
Mostly just a lack of time, I hadn't gotten around to doing it. Given the choice of Htg17 + Idaida's images that you provided over email, I went with Idaida's - simply because I was having trouble picking out exactly what Htg17's images were supposed to be of with relation to gw2. -Chieftain AlexUser Chieftain Alex sig.png 20:14, 14 December 2017 (UTC)
Thanks for the update, I'll let the artist know that their art is over there :) --Stephane Lo Presti talk 20:18, 14 December 2017 (UTC)

Yes Waypoint template broken?

Most of the waypoints show up as

Error: No waypoint named "Restoration Refuge Waypoint" found. Check if it has been added to the area.

Screenshot: https://i.imgur.com/tNWU2RH.png

Even on waypoint template page: https://i.imgur.com/t0Is3WC.png --109.227.124.100 19:49, 25 December 2017 (UTC)

When I browse the properties of the area pages the sub-objects (where all the waypoint data is stores) are all missing until I cache the page again. A solution is to edit the area page and save it immediately without any changes. This is called a null edit and does not appear on the recent changes page but it does force the wiki to cache all the data and allow the waypoint template to request it again. I believe this also happens automatically once every 24 hours. I have no idea why this has happened. J.Tesla (talk) 22:41, 25 December 2017 (UTC)

In turn this requires the addition of a css rule to MediaWiki:Vector.css in order for the page to display properly:

html {
    min-width: 800px; /* chrome mobile fixes */
}

I've initially applied this "fix" to Vector only because it might have unintended side-effects, and I'd like to test it for a bit. You can see the difference for yourself on mobile chrome using the following two links:

-Chieftain AlexUser Chieftain Alex sig.png 21:26, 10 July 2017 (UTC)

I ended up reverting this change as it wasn't thorough enough, probably needs a different set of styles altogether. -Chieftain AlexUser Chieftain Alex sig.png 06:27, 26 July 2017 (UTC)

Yes AutoWikiBrowser inconsistently broken

For the last month, half the time when I've tried using AutoWikiBrowser, after pressing "Start" it completely fails to make a single edit. I'm wondering if there's something wiki-side causing this. -Chieftain AlexUser Chieftain Alex sig.png 06:26, 26 July 2017 (UTC)

Not that it helps much, but I didn't have a problem with it today, nor a few weeks ago. - Doodleplex 06:57, 26 July 2017 (UTC)
Just had it do a run to see if any render images didn't have a licensing header. It changed 51 files with no issues, and I didn't notice anything odd in regards to how fast or slow it was going. - Doodleplex 21:54, 26 July 2017 (UTC)
I've installed some trashy "microsoft network monitor" tool to see if I can fail to make the edits. it worked just now, i'll try and break it tomorrow. -Chieftain AlexUser Chieftain Alex sig.png 23:20, 26 July 2017 (UTC)
Could have been related to my high packet loss. Closed. -Chieftain AlexUser Chieftain Alex sig.png 07:29, 9 November 2017 (UTC)

Yes Duplicated subobjects

All the items from Season 1 Memory Box—Flame and Festivals are duplicated on Season 1 Memory Box—Scarlet vs. Lion's Arch. You can see that Mini Molten Firestorm is apparently contained in both boxes even though that isn't true. A quick blank didn't help but I'm not so familiar with the wiki anymore. Tyndel (talk) 13:48, 28 August 2017 (UTC)

Okay, deleting the page fixed the issue, at least for the Mini and a few other pages I checked. —Ventriloquist 14:56, 28 August 2017 (UTC)

Yes Prices API 503 error "not active"

https://api.guildwars2.com/v2/commerce/prices?wiki=1&lang=en&ids=41746,41747,41748,41749,41750,41751,41752,41753,41754,41755,41756,41757,47901,47902,47903,47904,47905,47906,48924,48925,48926,48927,48928,48929,49525,49526,49527,49528,49529,49530,64198,64199,64200,64201,64202,64203,65164,65165,65166,65167,65168,65169,67284,67285,67286,67287,67288,67289,67991,67992,67993,67994,67995,67996,68673,68674,68675,68676,68677,68678

returns

{

   "text": "API not active"

}

Being called from

https://wiki.guildwars2.com/wiki/Celebratory_Dye_Pack

The API has been temporarily disabled to fix issues that arose from the launch of PoF. It should be back Monday. 185.189.14.230 17:49, 23 September 2017 (UTC)
Monday: API is working again!

Yes The Map bonus reward page no longer works - Error

On the following page, I can no longer view the table with the rewards. There is the following error.

https://wiki.guildwars2.com/wiki/Map_bonus_reward/profit


Commerce API unresponsive. Status: error

Error: 503

See the last comment above. - Doodleplex 19:34, 23 September 2017 (UTC)

Yes Weapon set table seems to be broken for the Elonian and Sunspear sets

Edit: Issue appears fixed after User:Greener deleted and re-edited the page. See Sunspear_weapons and Elonian_weapons or https://imgur.com/LEaJumz and https://imgur.com/WK6dxpN. The tables are misaligned, and for the Sunspear weapons one, it kept Suti's Hexbreaker on the page even after I removed it from the set. - Divinebaboon (talk) 21:25, 24 September 2017 (UTC)

No Wiki API does not return interwiki results

Using the help page here, I've followed the example at the bottom to https://wiki.guildwars2.com/api.php?action=query&prop=iwlinks&titles=Main%20Page. There should be three results here (DE/ES/FR), but there are no results at all! For comparison, here is wikipedia's. Ideas? -Chieftain AlexUser Chieftain Alex sig.png 20:22, 9 October 2017 (UTC)

Okay de:User:Think pointed me to https://wiki.guildwars2.com/api.php?action=query&prop=langlinks&titles=Main_Page - this works fine. Apparently mw:API:Iwlinks (interwikis) doesn't also include mw:API:Langlinks (language interwikis) -Chieftain AlexUser Chieftain Alex sig.png 20:27, 9 October 2017 (UTC)
  1. https://wiki.guildwars2.com/api.php?formatversion=2&format=jsonfm&action=query&titles=Guild_Wars_2_Wiki:Sandbox&generator=links&gpllimit=10&prop=langlinks
  2. https://wiki.guildwars2.com/api.php?formatversion=2&format=jsonfm&action=query&titles=Guild_Wars_2_Wiki:Sandbox&generator=links&gpllimit=20&prop=langlinks
And depending on how many results you ask for, it may or may not return the language interwikis... which sucks. (check out Agrak Kraal) -Chieftain AlexUser Chieftain Alex sig.png 20:54, 9 October 2017 (UTC)

No Constant reflowing of the text in the talk page

Can be seen here:
https://wiki.guildwars2.com/wiki/Talk:Mistfire_Wolf

The screen alternates between the following 2 states, with a delay of about a second:
https://imgur.com/a/06Ib0

Very distracting, making the page almost unreadable.

Using Chrome 61.0.3163.100, Win7, 64
Screen size: 1920x1080, full screen. Could be related to me using 2 screens, but unlikely.

--84.0.244.101 18:59, 10 October 2017 (UTC)

Forgot to add, in the screenshots, look at the 4th comment in the "Hounds of Balthazar?" section. In one state, it cuts the sentence after the first word, in the second, it pushes it into the next line. Also there is a scroll bar in one of the states. --84.0.244.101 19:05, 10 October 2017 (UTC)

It looks like it's an issue specifically related to using Chrome, as the same issue(basically the page scrolling nav thing appears and vanishes constantly) does not occur for me in Firefox nor Internet Explorer(checked with dual screens as well). I'll note that I have no special features added onto my chrome, so it's definitely not triggered by anything that isn't with the core browser, but unfortunately I can't figure out more than that to determine if it's something on our end or a Chrome hiccup. - Doodleplex 19:15, 10 October 2017 (UTC)
Nice I can reproduce it in Chrome... weird as hell. -Chieftain AlexUser Chieftain Alex sig.png 19:48, 10 October 2017 (UTC)

Bugs related to the server upgrade

Yes Favicon

It looks like the favicon is missing when the wiki is viewed from "https://wiki.guildwars2.com/". -Chieftain AlexUser Chieftain Alex sig.png 20:12, 8 November 2017 (UTC)

Could this be a local caching issue? I can't repro the issue. Justin Lloyd (talk) 21:22, 8 November 2017 (UTC)
It seemed to fix itself shortly after the URL problem was addressed. I reckon this can be considered resolved. -Chieftain AlexUser Chieftain Alex sig.png 07:29, 9 November 2017 (UTC)

Yes ".wiki.guildwars2.com" to ".wiki-en.guildwars2.com"

Apparently making any edit whilst logged in to .wiki. redirects you to .wiki-en. afterwards. Do we have to edit the wiki from the -en address now? -Chieftain AlexUser Chieftain Alex sig.png 20:14, 8 November 2017 (UTC)

This has been fixed. Please confirm whether you're still seeing the issue. Justin Lloyd (talk) 21:22, 8 November 2017 (UTC)
Possibly related, but https://www.guildwars2.com/ is now sending me to https://wiki-de.guildwars2.com/wiki/Hauptseite . G R E E N E R 21:46, 8 November 2017 (UTC)
I am currently unable to reproduce that redirect error. Have you purged your DNS and browser caches and perhaps trying to reproduce the error with another browser? Also, did you mean wiki.guildwars2.com and not www.guildwars2.com? Justin Lloyd (talk) 22:33, 8 November 2017 (UTC)
Just got back home. Problem seems to have been fixed. The problem had been with www.guildwars2.com, not the wiki. G R E E N E R 05:03, 9 November 2017 (UTC)

Yes Logged in

I seem to being having issues not being logged in on random pages, and after I edit things I seem to get logged out again. Using Firefox currently. - Doodleplex 20:16, 8 November 2017 (UTC)

I had the same issue. I think it has to do with the url force change in the subject above. Had to log-in to the 'new' url to get things working. -Darqam 20:17, 8 November 2017 (UTC)
Hopefully this is fixed with the correction to the wiki/wiki-en redirect issue. Let us know if it persists. Justin Lloyd (talk) 21:22, 8 November 2017 (UTC)
Seems to be fixed, thanks! - Doodleplex 22:35, 8 November 2017 (UTC)

Yes HTTP to HTTPS redirect missing

A while ago, I set up a custom search within my browser to search on the GW2Wiki from the address bar via gw2w <my search term>. When I created that custom search, I set the query URL (for unknown reasons) to use http as the protocol instead of https. When I tried to edit a page some minutes ago, I was not logged in due to not using the https protocol. I don't recall ever running into this problem before the server upgrade, so I assume this appeared with the upgrade. Example link for a quick test.Nefastu 23:47, 11 November 2017 (UTC)

Our engineer identified the issue and thinks he has a fix that he may be able to deploy next week. I'll keep you posted --Stephane Lo Presti talk 00:16, 1 December 2017 (UTC)
I forgot to update this thread when we posted the fix last week. Let us know if this is still happening! --Stephane Lo Presti talk 22:13, 12 December 2017 (UTC)

Yes Pearl Conch recipes appear on the wrong Inscription pages

I was looking at the page for Commander's Orichalcum Imbued Inscription and I noticed that the Used-In recipe lists at the end included the Minstrel's Pearl Conch (in addition to the correct Commander's Pearl Conch). I checked the Minstrel's Orichalcum Imbued Inscription page and found that its Used-In lists included the Seraph Pearl Conch. Checking the Seraph page , I found the Sinister Conch.

Upon examining the main Pearl Conch page, I noticed that the series of pages including the wrong additional recipe follows a pattern: In the current order of the page list, starting with Sinister and going down through Seraph, each Inscription page contains both the correct recipe, and the recipe of the next item in the list. The cycle ends at Seraph and resets back up to Sinister, so the top few and bottom few Conches aren't caught in the loop. PaladinOne (talk) 00:52, 18 November 2017 (UTC)

I gave the Pearl Conch page a bit of a kick by removing those items and subobjects. Looks like they're okay atm. G R E E N E R 01:05, 18 November 2017 (UTC)
It seems the issue has returned. Same set of erroneous items. (I checked it two weeks ago, after you fixed it, and at that time it was fixed. It appears to have un-fixed itself.) PaladinOne (talk) 07:08, 4 December 2017 (UTC)
I'm not seeing these problems. I recommend clearing out your browser cache of the affected pages and check again. SarielV 20 x 20px 07:18, 4 December 2017 (UTC)
Just tried looking at Trailblazer's Orichalcum Imbued Inscription in 3 different browsers, 2 of which have never been to the GW2 wiki, and also purged the wiki page cache. Issue still present. Were you maybe looking at one of the Conches that isn't affected? PaladinOne (talk) 08:03, 4 December 2017 (UTC)
I see this too. When I paste {{#ask:[[Has ingredient::{{{1|{{#titleparts:{{BASEPAGENAME}}}}}}}]]}} onto the page it's showing Pearl Conche#recipe15 (Trialblazer) and Pearl Conche#recipe16 (Valkyrie). I've checked the properties of recipe 16 and it is correct. Perhapse someone with more SMW knowledge can look into this? J.Tesla (talk) 23:24, 4 December 2017 (UTC)

Yes Session hijacking login

I got the following email from Innythia:

I seem to be having logging-in issues using Overwolf Browser (in-game browser) - Error message: There seems to be a problem with your login session; this action has been canceled as a precaution against session hijacking. Go back to the previous page, reload that page and then try again.

I told them to bring it up here, but they never did, so I figured the issue fixed itself. However I just had the same issue trying to login to the French wiki on Firefox. I tried logging in on Google Chrome and had no issue, so it might be browser specific or it might be the whole "links changed and now we have to hunt to find the secure login page". - Doodleplex 20:44, 29 November 2017 (UTC)

Same PC right? mw:Topic:Tvx6bgsxmwktdye9 - this suggests that it's related to https and http - possibly your cookie still has http instead of https. Did you clear your cookies after your last set of connection issues with the wiki? -Chieftain AlexUser Chieftain Alex sig.png 20:52, 29 November 2017 (UTC)
If this is related to the http/https issue reported above, we may have a fix coming next week. --Stephane Lo Presti talk 00:17, 1 December 2017 (UTC)
Yeah same PC Alex, but the ironic thing is several days ago, I was able to login just fine to the French wiki, and now I can't, so it was a bit strange. Also good to hear Stephane, can't wait. =) - Doodleplex 00:22, 1 December 2017 (UTC)
Alright so now I've got this issue. If I login over an HTTPS connection, and then visit any page on the wiki using HTTP - I won't be logged in. Trying to login under HTTP when I already have an HTTPS login active throws the error. I don't think this used to happen. -Chieftain AlexUser Chieftain Alex sig.png 10:37, 4 December 2017 (UTC)
Based on what Alex said, I clicked on the link to the French wiki from the recent changes page, put "https://" in front of the link, and bam, I was logged in. So it would appear our interwiki links only go to the HTTP pages, not the HTTPS. Can this be changed somehow, or will the above mentioned fix solve this as well? - Doodleplex 17:44, 4 December 2017 (UTC)
Currently all interwikis are sending people to the http location, as per Special:Interwiki. I recall coming across info stating that interwikis can have the http or https removed from the start of the link, which will then ape the users current setting when creating the link. Can't find that documentation atm, but it's one of the adjustments I made over at the gw1 equivalent.
Let me know if there are objections, and I can get around to it later tonight or tomorrow if need be. G R E E N E R 19:01, 4 December 2017 (UTC)
A day late, sorry. I've updated all of the interwikis to refer to the user's preference for http or https, rather than force them to go to the http version (as per this information).
I should also point out that the language wikis do not strictly enforce using https (i.e. you can access via http://wiki-fr.guildwars2.com/wiki/Accueil) while sites like wikipedia cannot be (i.e. http://www.wikipedia.org/ will always send you to https://www.wikipedia.org/). Possibly something to look into? G R E E N E R 07:01, 6 December 2017 (UTC)

Did this behavior change after the https fix we pushed last week? --Stephane Lo Presti talk 22:13, 12 December 2017 (UTC)

Well you've prevented reaching the wiki over http... so yes I think this message should largely be avoided since we'll never reach the http bit.
(new bug 1) Interestingly if you take an url to a usepage, e.g. "https://wiki.guildwars2.com/wiki/User:Chieftain_Alex", and then remove the "s" from https, http://wiki.guildwars2.com/wiki/User:Chieftain_Alex the wiki redirects to the same page with a slash on the end of it - a title that doesn't exist.
(new bug 2) Similarly, if you take a title with an apostrophe in it, e.g. "Talk:Repair the fort's walls by bringing rubble to the supply officer", https://wiki.guildwars2.com/index.php?title=Talk:Repair_the_fort's_walls_by_bringing_rubble_to_the_supply_officer and remove the "s" from https, it generates a bad title error (it tries encoding the pagename bit of the url again). This appears to be the same bug which also breaks the "/wiki" search functionality from ingame. I've added a workaround but it would be good to fix it on your end instead. -Chieftain AlexUser Chieftain Alex sig.png 22:50, 12 December 2017 (UTC)
Hey Alex, since we've applied a few fixes related to http, I was wondering if you knew whether there were still issues we need to take a look at? Thanks --Stephane Lo Presti talk 20:08, 15 January 2018 (UTC)
Yeah all seems good now. --Chieftain AlexUser Chieftain Alex sig.png 22:28, 15 January 2018 (UTC)

Yes AWB connection issue

As of today I can't connect to the wiki with AWB. I got the same message from what happened last year at this time, except this time, it's not giving me a reason in the error description. =< Doesn't look like there's been anything on AWB's end, no changes recently or complaints. - Doodleplex 18:38, 6 December 2017 (UTC)

Seems to be fixed now. Not sure if it was the below or not, but all is well. - Doodleplex 23:23, 8 December 2017 (UTC)

Yes /wiki (command line arguments) not parsing HTML codes

Using the in-game /wiki command will launch a browser window, but HTML characters are not parsed correctly (for example, %5B%26AgE%2BTgAA%5D instead of [&AgE+TgAA] for Hall of Monuments Portal Stone), therefore, no matches are found by the search engine. --NIN37 05:44, 7 December 2017 (UTC)

I haven't been had a need to do /wiki recently but I think this may have been caused by the same patch as when ANet are trying to fix the https/http bug. -Chieftain AlexUser Chieftain Alex sig.png 07:39, 7 December 2017 (UTC)
I cannot confirm this NIN37, searching chat links still works as expected. The game properly URL-encodes the chat link to %5B%26AgE%2BTgAA%5D which is absolutely the correct behavior when sending it as a query argument to the wiki. And going to the search url with that query argument will then also work properly. poke | talk 08:03, 7 December 2017 (UTC)
There was a similar report from another user, and these correlate with my edits to the Special:Interwiki where I removed the hardcoding of http on redirects (as per this response to a different problem).
I do not know if the /wiki function uses the en:, de:, fr:, or es: codes, but if it does, that may be the source. Also, given that I haven't heard if my edits solved Doodle's problem, we could err on the side of caution and revert my interwiki edits. G R E E N E R 12:56, 7 December 2017 (UTC)
I haven't had any issues since the change, other than my above bot issue of my bot not being able to connect to the wiki now. =x - Doodleplex 17:22, 7 December 2017 (UTC)
I returned the three language interwikis back to hardcoding http in response to the /wiki function (again, I have no idea if they are related). As for the AWB issue, is there any usage of the remaining sites when working with your bot? If so, we can return those interwikis back to http as well. G R E E N E R 19:19, 7 December 2017 (UTC)
I don't know what you changed, but I just changed AWB's default settings to connect to the https variant of this wiki and it does seem to be working now. Not sure if whatever you just changed back was the problem or if it's the http problem, I don't know, I'm just happy to have my bot back. Additionally, the interwiki links still do seem to work correctly btw and go to the https variant of the other wikis, so perhaps all is well? *goes off to hug J.A.R.V.I.S.* - Doodleplex 19:32, 7 December 2017 (UTC)
"the interwiki links still do seem to work correctly btw and go to the https variant of the other wikis" That would not have been me, as I returned them to the previous http. I'm sure there are tech savvy people out there saying, "Please stop changing things," so I'm going to leave things as they are (i.e. http for the interwikis, user preference for the other links). G R E E N E R 20:05, 7 December 2017 (UTC)
moved from Talk:Ascended Salvage Tool#[&AgEUJgEA]

I'm using Chrome. I'm brought to this page (https://wiki.guildwars2.com/wiki/?search=%255B%2526AvpvSQEA%255D) and it just stays there without re-directing. It looks like this: https://cdn.discordapp.com/attachments/295628621631062016/388693549106987008/unknown.png
--175.156.243.239 14:10, 8 December 2017 (UTC)

Yep it's broken for every chatlink returned from ingame. -Chieftain AlexUser Chieftain Alex sig.png 23:21, 8 December 2017 (UTC)

Is there anything else I can do to solve the problem or do we have to wait for ANet to fix it? --175.156.243.239 03:47, 9 December 2017 (UTC)

Symbol Encoded
[ %5B
& %26
+ %2B
= %3D
] %5D
The characters are already converted by the time they reach the search box from which the chatlink script operates. We could apply a patch to the script locally to replace sequences of text but I would really really like ANet to fix it properly and our patching it might hamper their debugging. -Chieftain AlexUser Chieftain Alex sig.png 13:27, 9 December 2017 (UTC)
Actually ANet can use their stage-test-wiki to debug the issue. I've applied a temporary fix. here. --Chieftain AlexUser Chieftain Alex sig.png 17:32, 9 December 2017 (UTC)

(Reset indent) We've deployed a fix for this and I tested [&AgE+TgAA] successfully from the in-game chat. Let us know if you see any other oddities of that kind. Thanks --Stephane Lo Presti talk 16:33, 9 January 2018 (UTC)

Thanks for applying that. -Chieftain AlexUser Chieftain Alex sig.png 17:56, 9 January 2018 (UTC)

Bugs related to the server upgrade

No Incomplete page loads

So I've had this 3 times now, example: http://prntscr.com/h7yg08. The page is no longer loading, nor does it 'give up after a long time'. It kind of just stops there after 3-4 seconds. I forgot to look at console when this happens if anything appeared, but will next time. I also got http://prntscr.com/h7yg6f when reloading the shown page (this has only happened once so far). -Darqam 20:31, 8 November 2017 (UTC)

I'm not sure about this one, as I don't understand that error message. Can anyone else provide some further insight about this? Justin Lloyd (talk) 21:22, 8 November 2017 (UTC)
I believe it's a 416 error (https://developer.mozilla.org/en-US/docs/Web/HTTP/Status/416), which could make it a one time 'my problem' issue. Until I get it once more or someone else reports it. I'm going to assume (t least on my end) that the error was a one off thing. The incomplete page load seems to also be gone, so this might have all been some browser screwery on my end. -Darqam 13:13, 9 November 2017 (UTC)
https://http.cat/416 -Chieftain AlexUser Chieftain Alex sig.png 18:10, 9 November 2017 (UTC)
Also everything is horribly slow. -Chieftain AlexUser Chieftain Alex sig.png 01:15, 11 November 2017 (UTC)
That isn't just me? Thank god. I thought I was going crazy. I was all set to start deep cleaning files.--Rain Spell (talk) 02:24, 11 November 2017 (UTC)

I know these reports are almost 3-weeks old but I wanted to know if this still happening? --Stephane Lo Presti talk 00:16, 1 December 2017 (UTC)

Pressing edit on some pages right now is taking over a minute to load the editing window: https://imgur.com/a/1n62J
Also receiving 503 Gateway timed out, and 503 Backend fetch failed errors too, in addition to not loading some of the CSS files, making it look ugly as hell like the sections below. --Chieftain AlexUser Chieftain Alex sig.png 11:46, 28 January 2018 (UTC)
Alex, the 503s/Varnish errors from yesterday morning were an issue with one specific server and that issue has been resolved. Justin Lloyd (talk) 13:13, 29 January 2018 (UTC)

No Odd page loading/messed up UI

Idk if this is just a normal bug or one from the move. Was loading a page and it took 5s instead of 2. Looked like this and the lower half looked like this. I'm using Firefox. I could click on some links, but others (the bug notice) wasn't working/where it said it was. Refreshing fixed the problem, and I have no idea how it would be replicated, but it happened. --Rain Spell (talk) 20:05, 9 November 2017 (UTC)

Occasionally getting this issue along with slow page loads. I get the feeling that they're both somehow correlated due to high traffic, but that's just my opinion and I have no way of confirming my theory. Seeing as everyone here is having issue in Firefox (including myself) is it possible it could be browser-specific? Sythe 16:34, 13 November 2017 (UTC)

Are you two still experiencing slowness on the wiki after the past few weeks? --Stephane Lo Presti talk 00:16, 1 December 2017 (UTC)

I'm still experiencing slowness (especially when editing. Example: I added one link on the Fractal Console page (+18 characters) and it took a good 7 seconds to finish loading. Loading the edit page is usually fast.) I haven't had any more odd page loads though. Oh, and I updated to the new Firefox a few days ago as well, so it should be up to date. --Rain Spell (talk) 20:14, 1 December 2017 (UTC)
Sorry to hear that Rainspell! Is this slowness consistent, meaning that you experience it every day or every week, or something that occur occasionally? If you could provide us with more data (article you were editing, with corresponding date and time), it may help us with troubleshooting this. And if anybody else experiences this slowness, more data would be helpful. --Stephane Lo Presti talk 22:13, 12 December 2017 (UTC)
Earlier today was particularly slow - about 2 hours previous to this message. -Chieftain AlexUser Chieftain Alex sig.png 22:32, 12 December 2017 (UTC)
I've been experiencing it consistently with pretty much every single article I edit or create. Creating a page is slightly faster, but moving/changing names on things take 2-3s longer than normal. It's only since the move, as far as I know. Were the servers moved really far from where they were before? It used to be it'd only take 5s to load a page if it had a lot of calls, but editing any page from Lesser Call of the Wild to Jingo to Black Lion Weapon Arsenal takes 5-7s. Maybe I was just spoiled before, but having a consistent load time of 5-7s for editing pages drives me batty. My loading times for navigating to pages is within a second, however. As for date/time. Since it's every edit: here ya go:(--Rain Spell (talk) 03:38, 13 December 2017 (UTC)

This comment seems to have moved a bit to loading times but I'd like to add that my UI is doing the same thing as pictured above, a lot of information is being pushed to the bottom of the page in a single column. I'm using Google Chrome and it started when the PoF graphics were put in. It's been getting progressively over time as well it used to have the search bar up top where it normally should be but now even that is being pushed down to the bottom as well. Some of the pages are loading better than others but I'm not sure why this is happening at all, I have the latest version of Chrome and it used to work fine, on top of that my dad uses Internet Explorer and none of these UI problems are happening for him. This is what my pages usually look like but unfortunately no amount of refreshing has fixed this so I've solved the replication issues because it's always happening for me and it's reaching the point of becoming unusable. SherlyDame (talk) 17:30, 21 January 2018 (UTC)

Hum. Having logged into the wiki on Chrome (v63, 64 bit, Windows 10) I'm not finding any immediate issues with the appearance. I know with firefox it sometimes drops the connection when downloading the stylesheet and that renders like Rainspell's first image above - maybe Chrome does something worse.
Can you please confirm that (i) you've cleared your local browser cache to make sure you're fetching the most recent revision of the stylesheets (CTRL + F5), (ii) cleared your cookies (Chrome ... > Settings > Advanced > Clear browsing data), (iii) which wiki skin are you using (Special:Preferences > Appearances > Skin), (iv) does the clock show up at the top of the page? (if not, indicates javascript bug) --Chieftain AlexUser Chieftain Alex sig.png 18:09, 21 January 2018 (UTC)
Oh wow! I feel dumb, I just did the first thing and cleared the cache, it's totally fine now! I rarely do any clearing but it's never presented a problem like that before so I never think about it. Thanks for the help! SherlyDame (talk) 15:27, 25 January 2018 (UTC)
Lol don't worry about it. I would hope that we have enough users using a variety of browsers to be able to detect bugs quickly enough to at least identify them, if not sort them out. --Chieftain AlexUser Chieftain Alex sig.png 18:07, 25 January 2018 (UTC)

(Reset indent) To revisit the original bug reports here, I occasionally get the page looking like Rainspell's link (https://snapr.pw/i/41323e68a3.png). The apparent cause on this occasion, reading the developer console in firefox, is that https://wiki.guildwars2.com/load.php?debug=false&lang=en&modules=startup&only=scripts&skin=monobook failed to load. -Chieftain AlexUser Chieftain Alex sig.png 18:03, 26 January 2018 (UTC)

Yes Increased ping time with bots

I used to run a deletion script which would just burn through a complete list in a few seconds flat, it seems to have a minimum delay of 5 seconds now. Is that intended? -Chieftain AlexUser Chieftain Alex sig.png 13:09, 18 November 2017 (UTC)

Hey Alex, it's not something we specifically set. Is this still happening? --Stephane Lo Presti talk 00:16, 1 December 2017 (UTC)
Yes, it happens with regular editing accounts too - there is a five second delay before any edit is submitted. (e.g. if i press submit now at 17:27:22, the edit only appears to process at 17:27:27). This doesn't seem like a lot but we used to be able to push through 100 bot edits per minute no problem (same routine would take 8 minutes now). -Chieftain AlexUser Chieftain Alex sig.png 17:28, 4 December 2017 (UTC)
Sounds like the same issue reported above. Do you have more data about the bots so we can look into logs? --Stephane Lo Presti talk 22:13, 12 December 2017 (UTC)
The circa five second delay between submitting edits and the corresponding page appearing is pretty consistent.
(i - now) I raised this issue above when deleting Erasculio's userpages - see these log entries. If you look you can see I've deleted roughly 12 pages per minute - this was using a deletion API script.
(ii - previously) Compare with these log entries earlier in October when I was running the same script to delete orphaned files, and was getting >150 pages per minute.
The script is the same so I'm guessing rate-limit or throttling somewhere else. -Chieftain AlexUser Chieftain Alex sig.png 22:43, 12 December 2017 (UTC)
Going to add on to this a bit. It would appear the edit/saving the edit is still very fast: I just did some testing while fixing some item stub tags and hit "save" with my bot, then opened the page right away and it was already updated, which would be about the same speed it would have shown me as being "saved" before the move. However, on the bot's end it was still showing as "saving". It's the same for regular editing as well, I just edited the sandbox, and had that page open on a separate tab. As soon as I hit save, I refreshed the other page, and it showed me my edit on the page, however the original page was still showing as "saving". So the edits are going through at the same speed, but the pages are just slow to show it after editing I guess? - Doodleplex 22:54, 27 December 2017 (UTC)
(Reset indent) When reading pages that I have not just edited, they load acceptably fast. The bit which is slow is (i) press edit until the source comes up, (ii) submitting edits until the new page comes up. I have renamed the topic title since it is not just limited to bots. -Chieftain AlexUser Chieftain Alex sig.png 22:36, 15 January 2018 (UTC)
This seems to have been fixed at some point. Getting good ping-times with AWB. -Chieftain AlexUser Chieftain Alex sig.png 21:05, 12 February 2018 (UTC)

No PoF interactive maps broken

The interactive maps for all 5 PoF map pages aren't working. It just shows a ~1/2 inch height black bar instead of the map. Domain of Istan is full size and shows the icons but the terrain map doesn't reach it so it's surrounded by black. I sampled some other maps from Core, HoT, and LWS3 and all the ones I checked are working fine --Gimmethegepgun (talk) 19:16, 2 December 2017 (UTC)

It's the API that is broken. For some reason when LWS4 E1 hit, the other maps got removed from the API. https://github.com/arenanet/api-cdi/issues/586 -Chieftain AlexUser Chieftain Alex sig.png 20:52, 2 December 2017 (UTC)

No Error 503 Backend fetch failed

Was editing a line of text on the Of Wiki of Gold page. Went to hit "Show Preview" and got the following page (text only):

Error 503 Backend fetch failed

Backend fetch failed
Guru Meditation:

XID: 73074799

Varnish cache server

I went back and tried hitting preview again and it worked fine. Just a cache issue?-

Oh, Browser is Firefox, newest version afaik.-Rain Spell (talk) 03:03, 9 December 2017 (UTC)

I'm getting this occasionally. -Chieftain AlexUser Chieftain Alex sig.png 13:30, 9 December 2017 (UTC)
Getting it again. Refreshing or pressing back and clicking the same link works, but it takes a good 10-15s. I think it's happening more on pages I've left open for an hour and come back to. Also, apparently others are experiencing similar. :
Error 503 Backend fetch failed

Backend fetch failed
Guru Meditation:

XID: 784826374

Varnish cache server

--Rain Spell (talk) 23:43, 12 January 2018 (UTC)

Yes Problems with random images on the Wiki?

moved from Talk:Main Page

I've been having problems seeing random images on this Wiki and I believe it's somehow related to how thumbnails are generated/cached in Google Chrome (like if you have a 100x100 pixel image and resize it to 50x50 pixels on a page).

  • Exhibit 1 - The images for Euryale and Street Rat don't appear for me on the Biography page
  • Exhibit 2 - If I click on the link for Street Rat, the link works just fine and I can clearly see the image
  • Exhibit 3 - If I then click on the BACK button in my browser, I can now see the Street Rat image

However, if I then click on the Euryale link to see the image, then go back - I can see the Euryale image but not the Street Rat image. And the images that don't work appear to be random per computer. For example, my work PC (Windows 7 Pro 64-bit, Google Chrome browser) doesn't show Euryale & Street Rat while my home PC (Windows 10 Home 64-bit, Google Chrome browser) does, but with other images not being displayed.

If I use Firefox to view the Biography page, I don't seem to have this problem... but I'd like to stick with Google Chrome if possible. Any idea how to fix this? The preceding unsigned comment was added by Kencussion (talkcontribs) at 22:28, 8 January 2018‎ (UTC).

Hi Kencussion, did this problem happen to you before, or is this something you noticed recently? Does a page reload fix the issue? Thanks --Stephane Lo Presti talk 16:49, 9 January 2018 (UTC)
I'm not sure, but I just started noticing this problem a couple weeks ago. I've tried reloading the page, clearing the browser's cache - still having the same problem though. It doesn't seem to happen on my iPhone (even when using Chrome Browser), but I noticed it DOES happen on my Ubuntu box (Ubuntu 16.04 64-bit) when I use the Google Chrome browser - again it works fine in Ubuntu when I use Firefox. So it may just be an issue with Google Chrome on desktop PCs? Kencussion (talk) 19:49, 9 January 2018 (UTC)
At this point, we've been able to reproduce the issue and it looks like it's a Chrome-specific issue related to caching. It's a complicated problem to troubleshoot given its nature and it may take a while to investigate. Because this seems to be a fairly controlled issue (a workaround is to use Firefox and we haven't heard tons of report about it) and we want to push a few updates to the MediaWiki software, we'd like to first focus on these updates, and only after that go back to this investigation. It may take a while, so feel free to keep us updated to let us know if the issue is affecting a lot of people or being more widespread. Thanks --Stephane Lo Presti talk 21:00, 9 January 2018 (UTC)
Keep pushing until you reach MW 1.29! -Chieftain AlexUser Chieftain Alex sig.png 22:06, 9 January 2018 (UTC)
We've got 1.27.4 in test right now and hope to get it live very soon. We're then moving to 1.30 because it will take some time to iron out the new update procedure (this could also come with an upgrade of Semantic MediaWiki to 2.5 or 3.0). That being said, given the specificity of this issue, we're not sure that software updates are going to help here --Stephane Lo Presti talk 22:34, 9 January 2018 (UTC)
That's good to hear - there are so many annoying bugs that get fixed by the time MW reaches 1.30 (e.g. currently it downloads the CSS files twice, and sometimes javascript errors prevent scripts working) -Chieftain AlexUser Chieftain Alex sig.png 23:10, 9 January 2018 (UTC)
Sorry for the lack of updates on this front. We've focused most of 2017 on the move to the different platform and knew that we'd have to skip version updates, which could unfortunately impact editors the way you describe. Once we're past the upgrade to MediaWiki to 1.30, we should go back to a more regular pace of updates --Stephane Lo Presti talk 23:25, 9 January 2018 (UTC)

(Reset indent) We deployed a fix and this seems to have fixed the issue. Let us know if you still see it, or any other related problem. Thanks! --Stephane Lo Presti talk 20:06, 15 January 2018 (UTC)

It seems to have fixed the issue for me - thanks! Kencussion (talk) 20:10, 15 January 2018 (UTC)