[00:01:47] marktraceur: http://tools.wmflabs.org/jembot/wiki-php.txt [00:02:05] It's a bit of a mess :), the key function is apiedit [00:02:44] Hm [00:02:47] If you actually want to run it, another launcher script is needed [00:02:56] No, it's OK [00:03:24] Hm. [00:03:31] I get the same edit token - just + [00:03:58] By going to http://en.wikipedia.org/w/api.php?action=tokens&format=json [00:04:21] Ah [00:05:45] Not sure what's up with that [00:06:04] mw.org is borked too [00:06:20] jem-: You said other bots of yours are working fine, though? [00:07:00] Yes [00:07:20] https://es.wikipedia.org/wiki/Especial:Contribuciones/PatruBOT [00:07:25] Hm. [00:07:35] And my local wiki works fine for edit tokens at least [00:07:36] That one is run by me (not written) [00:07:40] So not sure what's wrong [00:07:59] https://es.wikipedia.org/wiki/Especial:Contribuciones/Invadibot is working also [00:08:14] https://es.wikipedia.org/wiki/Especial:Contribuciones/Jembot is my bot [00:09:01] jem-: Can you file a bug about this please? [00:09:44] Hum, if it's the way to go... Ok [00:10:08] (I should go to bed, but I hoped to fix it before that) [00:10:12] Yeah, people who might know more don't seem to be around [00:10:19] Ok [00:10:21] Hopefully overnight it can get fixed :) [00:10:30] Let's hope so [00:10:53] At least it seems the problem isn't just mine, is it? [00:15:09] marktraceur: Is it ok to add http://en.wikipedia.org/w/api.php?action=tokens&format=json ? [00:20:28] Yeah, it should be [00:21:10] jgonera: did you still want to look at something? [00:29:05] jem-, yt? [00:29:38] MaxSem: https://bugzilla.wikimedia.org/show_bug.cgi?id=64727 [00:29:43] marktraceur: [00:29:55] *nod* [00:30:20] jem-, have you tried logging in again? [00:32:14] MaxSem: I don't understand [00:32:28] The bot logs in again every time it is started [00:33:22] hmm [00:35:30] jem-: log out and then back in [00:35:59] Hmmm [00:36:06] Ok, I'll try [00:36:27] forces the server to clear you session/cookie info [00:37:52] I've logged in and out from my browser, and I've tried to run it from Labs, but Ok [00:38:07] (I've tried with the same results) [00:41:34] jem-: what does https://en.wikipedia.org/w/api.php?action=tokens&format=jsonfm give you> [00:42:38] Betacommand: "tokens": { "edittoken": "79098e237b8ed18177d0b0fe062a8f05+\\" [00:43:14] what about: [00:43:23] https://es.wikipedia.org/w/api.php?action=tokens&format=jsonfm give you> [00:43:55] (Ah, I see the problem with that was the https) [00:44:04] "edittoken": "ec2b512394afae9bb3b8c9cc57d54b89+\\" [00:44:15] then it looks like the problem is with your code [00:44:58] As the API is passing valid tokens to you [00:45:16] But not a valid edit token, just a login token [00:45:24] And it did work until 6 hours ago [00:45:49] jem-: the "edittoken" is a valid edit token [00:46:20] ¿The "+\"? [00:46:36] No, the "edittoken": "ec2b512394afae9bb3b8c9cc57d54b89+\\" [00:46:56] which means something in your code isnt working/passing the correct info [00:47:17] does your code use HTTP or HTTPS ? [00:47:24] for its connections? [00:47:38] I would say https, but let me check [00:48:10] "would say" is the root cause of a lot of bugs :P [00:52:25] jem-: What are you trying to do? [00:55:49] Well, it was http, now I'm trying to change to https but I get read errors (I guess that's why it was http) [00:57:09] jem-: go to your user preferences and uncheck the use HTTPS box [00:57:48] logging in via HTTPS and then using http to do stuff would cause session issues [01:01:49] What? [01:01:55] What's the actual issue? [01:01:59] You should generally be using HTTPS. [01:03:51] jem-: Are you writing your own edit code in http://tools.wmflabs.org/jembot/wiki-php.txt ? [01:04:05] You should probably be using a PHP framework for MediaWiki. [01:04:48] Gloria: getting invalid edit tokens is the issue that was brought up [01:04:54] I guess Snoopy is kind of a framework. [01:05:19] traced it back to logging in via HTTPS and attempting to edit via HTTP [01:05:23] Gloria, I began 4 yours ago and I feel comfortable knowing the code I use, besides my bot integrates IRC functions with the same interface [01:05:38] Sure. [01:07:42] jem-: Okay, to answer the general question, edit tokens include "+\" at the end of them intentionally. [01:07:55] It's a feature to prevent bad scripts and bots, or something. [01:08:04] You'll sometimes see +\\ because the slash is being escaped. [01:08:40] Gloria: that was the complete token that he was getting, which means he wasnt logged in [01:08:42] Yes, but that's at the end [01:08:50] Yes [01:09:11] And now it seems that with https a I don't even get a response [01:09:25] Betacommand: What? [01:09:49] I must give up for now, I must wake up in 5 and a half hours [01:10:01] Gloria: when your not logged in your edit token is "+\" [01:10:31] that was the complete token that jem- was getting [01:10:44] (Yes, and I get [result] => Success after the login) [01:11:16] Thank you everyone for the help, I'll keep on trying tomorrow, and I'll read if you have any more ideas [01:12:04] Betacommand: Ah, thanks. [01:12:30] The point is I would like to know positively if something was done around 18:50 h UTC yesterday to make http not work [01:13:19] Gloria: since he was using a third party library to handle the login/cookies I thought the easiest solution would be to not use HTTPS, instead of hacking at the library [01:13:37] Everyone should generally be using HTTPS. [01:13:38] jem-: that shouldnt have been working for a looog time [01:13:46] Unless there's a very good reason not to. [01:14:09] HTTP login still works fine. [01:14:12] Or should. [01:14:29] Gloria: it does [01:14:33] I imagine most of my scripts are still using it. [01:14:38] Though they really ought not be! [01:14:56] Gloria: the forceHTTPS cookie is a real bitch at times [01:15:28] I haven't had any issues with it, but I use some library that someone else maintains. [03:21:52] wikipedia is down [03:22:45] cortexA9: What do you mean? [03:22:58] en.wikipedia [03:23:05] doesnt work for me [03:23:18] https://en.wikipedia.org/wiki/Lochry%27s_Defeat works for me. [03:24:06] ok solved [03:24:36] Glad to hear it. [09:05:18] Hi again, I'm still stuck with my problem: https://bugzilla.wikimedia.org/show_bug.cgi?id=64727 , any help would be greatly appreciated [09:17:16] Ok, now I'm positively sure it is related to the 1.24wmf2 deploy: in es.wikinews it stopped 29 Apr, 18 h UTC, also when it was deployed [11:04:30] Still no one there to help? [11:11:07] sigh echo's message bar never resets when it should [11:18:42] jem-: to help with what? Repeating the question is welcome for people who have recently joined. [11:19:49] andre__: I know, I was waiting for an answer, thanks :) [11:19:58] The problem is https://bugzilla.wikimedia.org/show_bug.cgi?id=64727 [11:20:51] jem-, my best guess would be asking anomie when he goes online [11:21:27] Hmmmm [11:21:34] I take note, thanks [12:06:59] andre__: I noticed I missed your Bug Triage announcement last week. [12:07:22] andre__: Would you mind cross-posting those announcements to Wikitech-ambassadors in the future? [12:07:31] twkozlowski, that's no problem, there will be another one, as we have about 14000 open tickets left. :P [12:07:40] twkozlowski, let me try to remember that, ok [12:07:42] andre__: I'd like to feature them in Tech News so it reaches more people [12:07:48] oh yay [12:07:53] that's nice! [12:49:12] @seen keegan [12:49:12] twkozlowski: Last time I saw keegan they were quitting the network with reason: Ping timeout: 246 seconds N/A at 5/2/2014 7:37:14 AM (5h11m57s ago) [13:19:25] Hi, anomie [13:20:04] I have been told I should talk about https://bugzilla.wikimedia.org/show_bug.cgi?id=64727 with you [13:24:27] anomie ? [13:25:56] jem-: Let me look [13:26:19] Thanks [13:27:01] Right now I can see it has something to do with the way I handle cookies, I was using the Snoopy class [13:28:27] I have been testing with libcurl also, http and https, same results, but with the curl program I've managed to get the edit token [13:29:51] It seems making $Snoopy->setcookies() and passing that $Snoopy object to the next call was enough until now, but since 1.24wmf2 it is no more (?) [13:30:24] Sorry, the boss called, I must leave in a minute [13:30:41] I'll read later any suggestion you can make, thanks [14:00:06] * twkozlowski found another interesting patch title for his collection: "Nicified some tests" [14:00:49] jem-: Hmm. It looks like gerrit change 94614 has changed things so all API logins are done as if the "Keep me logged in" checkbox is unchecked where it was previously as if the checkbox was checked. That may have something to do with it; what cookies are you sending with your request? [15:38:05] Oh no. [15:38:12] * twkozlowski just enabled Compact Personal Bar. [15:43:15] http://tools.wikimedia.pl/~odder/screenshots/CompactPersonalBar.png [15:43:19] Is this ^^ supposed to happen? [15:46:46] http://tools.wikimedia.pl/~odder/screenshots/CompactPersonalBar2.png [15:46:50] This is also very uninviting. [16:41:23] LOL odder [16:42:14] * Nemo_bis is already getting sick of BetaFeatures [16:43:02] Nemo_bis: Oh? [16:44:36] Shhh. [16:44:45] Nemo_bis, what? [16:45:26] Did you shhh me? [16:49:10] Nemo_bis, oh right. [16:49:35] Nemo_bis, oh hey I broke it already [16:52:56] aww [16:53:11] Krenair: are you working on that thing? [16:53:18] no [16:53:41] Ok or I'd demand my old Krenair back. :) [16:55:24] Nemo_bis, to be fair I think it's one of my gadgets breaking things [16:56:03] busy on VE stuff now, might look later [17:14:16] Nemo_bis: More of an auch that a LOL for me :-( [17:14:29] Very badly broken, methinks, but meh. [18:22:33] hey all; after looking at the code for extension:renameuser, am I right in concluding that there's no upper limit to contribs for an account that is to be renamed? [18:25:44] Writ_Keeper: sure; except that it often breaks then they're too many :) [18:26:03] how many edits are you talking about? [18:27:09] well, the editor that led me to this as 52,000 or thereabouts [18:27:35] but the point is more that the page at Meta suggested that renameuser wouldn't work on accounts with more than 50,000 [18:27:45] and that doesn't seem to be the case [18:28:05] (and also, I don't know where that number came from anyway, as when there was a limit, it looks like it was 25,000?) [18:28:21] used to be 6800 edits it seems [18:28:35] where do you see that? [18:28:40] https://noc.wikimedia.org/conf/InitialiseSettings.php.txt [18:28:49] # So that I can rename users with more than 6800 edits -- TS [18:28:49] # Removed as obsolete -- 2009-03-05 BV [18:28:50] # 'developer' => array( 'siteadmin' => true ), [18:29:53] anyway, just tell the users the process may require a bugzilla request and some more days of waiting if things happen to go bad [18:30:47] sure, sure [18:31:02] though I'm kinda trying to learn some of this stuff [18:31:06] I was looking at https://git.wikimedia.org/commitdiff/mediawiki%2Fextensions%2FRenameuser/fddf47a516b00103aaa76589df878c38d9a5ec5f [18:31:26] is there another place where an edit limit could've been implemented, other than in the extension? [18:48:49] Krinkle: I was thinking it would be useful to add a global JS var to core indicating whether localStorage is working or not (by actually trying to do a localStorage.setItem and localStorage.removeItem). Do you think that's a good idea? If so, where should such a test be run from? mediawiki.js? [18:49:08] thedj: ^ [18:49:18] kaldari: no [18:49:28] kaldari: probably something like jQuery.support [18:49:33] but not in their space of course [18:49:35] lazy-init [18:50:04] We can cherry-pick from modernizr (avoid re-inventing the wheel) [18:50:23] add a new module though, don't add any more built-ins to mediawiki.js [18:50:42] I checked moderizr and all they do is a set and remove in a try statement. [18:51:08] kaldari: What's the value though? [18:51:16] It can reach capacity or get disabled at any time [18:51:22] so you still need to account for errors in your app [18:51:39] yes, totally agree [18:51:48] unless its for lazy-loading a polyfill, you can't use it for naything [18:52:20] Though I suppose you could stop trying if you know it's full/disabled (it's more likely to get disabled over time them to get enabled) [18:53:28] looks like modernizr sets it as a property called localstorage [18:54:03] as a boolean [18:54:14] yeah, the test function is executed once, and then cached [18:54:23] at least that's how I'd recommend doing it [18:55:44] support.hasLocalStorage = function () { var has = ( true || false ); support.hasLocalStorage = function () { return has; }; return has; } [18:56:11] jQuery does this for various support tests it has (e.g. the kidn that involve creating a temporary element and testing something) [18:56:25] kaldari: Again though, what would you use it for? [18:56:52] I think in our case, we should make sure localStorage doesn't become full as a normal part of business. [18:57:17] I haven't seen it myself, but if what you experienced is going to be more common, we should find a way to reduce the mw.loader.store footprint to only certain modules. [18:57:36] Can we move to #wikimedia-dev? [19:02:01] MediaWiki, you son of a bitch [19:02:12] "As an anti-spam measure, you are limited from performing this action too many times in a short space of time, and you have exceeded this limit. Please try again in a few minutes." [19:15:04] anomie: Thanks for the help, I've read the bugzilla page and the solution already works for http://en.wikipedia.beta.wmflabs.org/, but now I'm trying a workaround until Monday and I'm stuck trying to edit with curl (program) using the edit token [19:15:53] jem-: Sorry I can't get it deployed before Monday, but we don't do deployments except in major emergencies in Fridays or weekends. [19:16:19] anomie: I understand and I don't pretend so [19:17:35] But I would like to have some kind of workaround, my bot has jobs every few minutes and people is already asking [19:23:14] Anyone around willing to help with editing using curl or any other "quick" idea? [19:57:31] Anyone around willing to help with editing using curl or any other "quick" idea? [20:25:10] https://li.wikipedia.org/wiki/Plaetje:Wiki.png [20:25:19] Nice logo they have, and they refuse to change it. [20:27:08] I see many iterations of improvement! [20:30:00] Nemo_bis: Even so, I'm seriously considering ignoring them and just forcing them to get rid of this ugly logo. [20:30:22] Mine is so much prettier! [20:31:09] <^d> I colored a new logo today. [20:31:11] <^d> We should use it. [20:31:30] <^d> http://instagram.com/p/ngCRh3m0Z9/ [20:32:22] We could use it as default beta.wmflabs.org logo. [20:32:54] Though it would probably be a copyright violation, unless dino siluettes are PD-ineligible. [20:33:07] Anyway, you should propose li.wiki to adopt it. [20:33:27] <^d> It wasn't in a book, it was just in a pile of coloring sheets in the office I was in. [20:33:38] <^d> Ostensibly for children, but I was bored and there were crayons. [20:33:45] If you want I have some dozens spare colouring books here that I can bring you. :) [20:34:25] ^d should have a look at the Vietnamese logos if he likes to draw with bright crayons [20:35:01] https://upload.wikimedia.org/wikipedia/vi/archive/b/bc/20130218000840!Wiki.png [20:35:21] <^d> Aww, that's cute [20:35:45] That's officialy The Best Wikipedia Logo Of All Time [20:35:57] Knowledge is guarded by pitons! A big moral lesson. [20:36:14] *pythons [21:11:05] Reedy: Sorry for the repeat, I've lost your response (if any). Is there any reason for me not to disable (and subsequently clean up) your 'samreed' login? [23:58:06] kaldari: let me know if you need to push anything for the special:newfile thing. [23:58:20] also, why it's showing up just now on commons is confusing to me