[01:19:52] yo, anyone know if special:mypage works with importscript? [01:20:34] Writ_Keeper: test it :) [01:20:55] yeah, I'm doing that right now; just wondering if someone knew already [01:21:02] save a few moments of itme [11:10:06] the last 3 messages at wikitech-l are just spam. Not sure if it's moderated or if this is the right channel.. [13:30:50] Reedy: Ping [13:47:06] does anyone know the combined size of all the raw enwp tarballs combined when extracted? [13:50:20] apergos: ^ [13:50:45] over 10T for the en wp history dumps [13:53:58] Oh, crap. [13:54:09] nice apergos ! [13:55:26] Quick question, test.wikipedia should handle my SUL account as any other wiki does..? [13:56:13] yes [13:56:28] p858snake|l: i though it wasn't in sul? [13:56:47] i'm pretty sure test.w is [13:56:49] test2 is :P I dont know about test. though [13:56:54] * addshore checks [13:57:04] i'm pretty sure it isn't :P [13:57:06] only the private wikis don't use centralauth [13:57:20] alright, i retract, it is [13:57:31] and the fishbowl wikis [13:57:36] (the size is one of the dump FAQs, see http://meta.wikimedia.org/wiki/Data_dumps/FAQ#How_big_are_the_en_wikipedia_dumps_uncompressed.3F ) [13:57:37] okay :) so, why does it never detect me as logged in? :> [13:57:45] in fact testwiki is the only wiki that explicitly has it marked as on, everything else goes to default [13:58:15] addshore: its in sul, but not in auto login [13:58:24] :< [13:58:30] its so annoying :P [14:02:01] addshore: the more sites in autologin, the longer it takes [14:03:24] :< [14:21:25] also anyone around that can grant permissions on http://test.wikidata.org/ ? :> [14:35:42] Last I checked test.wikipedia.org is served from pmtpa while everything else is in eqiad. Which causes problems for session sharing and the like, pmtpa and eqiad have separate memcached pools. So the little icon thing to do the SUL login doesn't necessarily work right there, nor does the new code for SUL2. [14:44:44] time to move it to eqiad? [14:54:36] MaxSem: Probably. And I'd guess the right people already know about it too, but it's probably not a high priority item. [15:04:17] Reedy: will the scripts for mass emailing and so on show up at https://git.wikimedia.org/tree/mediawiki%2Fextensions%2FSecurePoll.git/fede1376420c025b01c120f8f1f2b4938c2af50c/cli%2Fwm-scripts too? [15:04:28] Presumably [15:06:25] nice [15:06:30] aww "In general, you should ask SPI to set up the wiki at least a week in advance, rather than 36 hours like in 2011" [15:07:07] Reedy: So what are details on this election email thing greg-g asked me to help you with? [15:07:17] I've no ideae [15:08:35] Never done it before, so would be as blind to the issue as most people [15:10:05] Like https://wikitech.wikimedia.org/wiki/SecurePoll#Email_spam ? [15:10:28] The section already has todos [15:10:51] Now you only need to steal Andrew's key or get a root copy the scripts, or something? :P [15:11:49] The latter sounds a lot more ethical.. [15:11:49] Nemo_bis: Looks like they're world-readable on hume, actually [15:12:11] (I suppose that should get copied to terbium at some point) [15:12:43] (or better, added to git somewhere) [15:12:45] Ah, that makes things simpler [15:13:04] anomie: Reedy already copied some scripts to the SecurePoll directory I linked above [15:13:15] Nope.. [15:13:24] Those were new versions of the 2009 ones already in git [15:13:32] oh [15:15:27] Let me start the clone to tin and I'll see what's actually in my email from overnight [15:21:04] Is doSpam.php one of the scripts in Andrews ~? [15:22:13] Hope so, as I can't see it in our git repos [15:22:35] Yes. It looks like it will be easy to add "already voted" checking along the same lines as the "nomail list" check. Deduplication should be easily enough doable too. [15:23:00] The nomail list handling should be easy enough [15:23:12] even if we just copypaste the current list into a text file/similar [15:23:19] rather than getting text from a remote MW install etc [15:23:29] nomail handling is already in there, just copy the list from the wiki into the file named on line 9 [15:23:37] aha, great [15:23:59] Would you mind committing any and all script there into git as is please? [15:24:08] Already-voted could be queried the same way from the database, or done as a file just like that. [15:25:36] Reedy: Ok [15:26:10] Then we can submit and merge that as is, and modify them as necessary [15:29:17] anomie: Reedy yeah, we're all going blind on this one :) (re: email) [15:29:27] g'morn, btw [15:29:59] I wonder if it'd be quicker for me to drive to Maastrict and go and kidnap Andrew.. [15:32:42] Reedy: how far to Maastrict? :) [15:33:08] I could be there before your working day is over [15:33:22] :) [15:39:21] !g 68411 | Reedy [15:39:21] Reedy: https://gerrit.wikimedia.org/r/#q,68411,n,z [15:40:09] I was about to say that's a lot of code [15:40:22] Crap [18:07:15] where is the mediawiki-config repo gone? this link worked yesterday: https://git.wikimedia.org/commitdiff/operations%2Fmediawiki-config/af0ad405d9ac2a4f876df3f8df1b3402868a101a [18:07:34] Works fine today [18:07:39] Raymond_: it's still there afaik [18:07:42] also WFM [18:07:51] gitblit has been… having issues [18:08:48] Raymond_, WFM [18:08:55] greg-g, Reedy: Since it looks like Reedy already finished the wmf7 deployment a few hours ahead of schedule, would anyone object to me deploying a bugfix to the Thanks extension within the current Platform window? [18:08:55] MatmaRex: hmmm yes... coming back now... [18:09:02] kaldari: Nope [18:09:15] Have to move one wiki to rebuild l10n cache [18:09:22] I'm likely to be done in 5-10 minutes though [18:09:30] cool [18:10:04] kaldari: be on irc when merging changes, we both submitted the same fix to Thanks at almost the same time :) [18:10:14] lol [18:10:14] (and funnily, they both merged) [18:10:26] Have I mentioned I hate Internet Explorer :) [18:11:02] don't we all ~ [18:12:29] Here's a wiki [18:12:31] There's a wiki.. [18:13:00] That reminds me of the Llama song. [18:13:05] everywhere a wiki wiki [18:13:21] http://www.albinoblacksheep.com/flash/llama [18:13:26] wiki wiki duck. [18:13:30] doesn't have the same ring to it. [18:13:43] isn't it "wiki wiki goose"? [18:24:44] Bsadowski1: Precisely [18:24:52] llama llama duck [18:29:29] LeslieCarr: isn't SF about the sourdouh? (which was what I had last time I was there) [18:37:18] Reedy: Heh, we had the same idea with 68469 and 68477 [18:46:16] kaldari: I resubmitted with a less nasty array structure [18:46:26] sweet [18:46:28] (why are you not on -dev now? :/) [18:46:40] because it's too noisy for me [18:47:15] I found myself closing the channel on a daily basis so it wouldn't distract me, so now I just don't join :P [18:47:53] which probably defeats the purpose I admit [18:48:21] Yeahh [18:48:39] marktraceur: I still need to find reviewers for all your OSM changes :( [18:48:44] It means I can't say "kaldari ^", I either have to assume you know which patch I mean or link it manually [18:48:56] Ryan_Lane: Oh dear - JS reviewers? That's a tough un. [18:49:00] luckily I always know ;) [18:49:00] yep [18:49:12] MatmaRex: Care to help Ryan_Lane review some JavaScript changes to OpenStackManager? [18:49:23] You know, in case you're not busy enough [18:49:38] heh [18:49:40] marktraceur: i have no idea what OpenStackManager is, just so you know :P [18:49:47] it's just JS reviews :) [18:49:54] against api calls [18:49:56] MatmaRex: It's what we use on wikitechwiki to run Labs, basically. [18:49:58] the rest I can review [18:49:59] i can read thecode if you want me to :P [18:50:04] Really basically. [18:50:33] https://gerrit.wikimedia.org/r/#/c/65557/ https://gerrit.wikimedia.org/r/#/c/65421/ https://gerrit.wikimedia.org/r/#/c/60783/ https://gerrit.wikimedia.org/r/#/c/65269/ [18:50:45] Oh wow I didn't realize how many changes I'd made. [18:50:50] I made one of them :) [18:50:54] but you fixed it [18:50:59] Reedy: Lemme know when the coast is clear. Just gotta update the extension in tin and sync 1 file. [18:51:12] I've not done anything for 20 minutes or more [18:51:14] feel free [18:51:18] thanks :) [18:51:35] greg-g: ^ [18:51:49] * greg-g nods [18:52:30] Ryan_Lane: marktraceur: add me as a reviewrr? [18:52:43] same reviewer name as irc? [18:52:52] yeah [18:52:54] cool [18:52:59] or close enough at least [18:53:31] done [19:15:16] Reedy: I made the change to wmf7 fine, but it looks like the wmf6 branch on tin is dirty. Should I reset it? The local changes are to the Release Notes and GuidedTour submodule. [19:19:24] Yeah [19:22:25] Reedy: error: unable to unlink old 'RELEASE-NOTES-1.22' (Permission denied) [19:35:27] Reedy, kaldari: Would it screw up anything you're doing if I run l10nupdate against testwiki? I suspect !g 67531 may have partially broken it, and I want to see for sure so I can hopefully have a fix in time for the lightning deploy slot this evening. [19:35:41] I'm not doing anything related to deployment [19:35:43] Feel free [19:38:42] anomie: I think it needs running everywhere tbh [19:39:43] Reedy: It runs every day at something like 2am UTC. I'll check the log and see if there have been errors. [19:41:48] I mean after/for the wmf7 deploy [19:42:01] I do wondr if scap should really do it whne necessary [19:45:34] The whole thing is rather slow. scap does a quick version that doesn't check for non-en messages that were updated between $branch and master, just uses the cache from the last full run. [19:45:55] And when you first branch, there shouldn't be many (if any) of those. So. [19:49:57] what's new in busy town? [19:51:02] Technical_13: https://en.wikipedia.org/wiki/Special:RecentChanges [19:51:17] LeslieCarr: it's a mystery.. [19:53:59] anomie: fine with me [19:54:12] What category does bugs about bugzilla go in in bugzilla? [19:54:20] kaldari: Good, because I already did it ;) [19:54:24] :) [19:55:02] Tools? [19:55:03] Reedy: still can't do a reset on wmf6 on tin [19:55:06] Technical_13: wikimedia/bugzilla? [19:55:10] error: unable to unlink old 'RELEASE-NOTES-1.22' (Permission denied) [19:55:10] fatal: Could not reset index file to revision 'HEAD'. [19:55:15] Technical_13: If it's about our installation, https://bugzilla.wikimedia.org/enter_bug.cgi?product=Wikimedia component "bugzilla". If it's something upstream, probably upstream. [19:55:16] Technical_13: Wikimedia product, Bugzilla component [19:55:29] thanks [19:55:29] Reedy: any idea how I can fix that? [19:56:06] greg-g: Can a brotha get a lightning deploy window today? Want to deploy a change to UploadWizard to fix Opera [19:56:20] Only two files changed, so [19:58:45] marktraceur: do the submodule bump and we'll add it to E3's deployment [19:58:52] Right on [19:59:49] marktraceur: fo sho [19:59:57] marktraceur: put it on ze calendar, plz [20:00:38] greg-g: Evidently ori-l is going to be kind and help out, so no need [20:00:53] well fine [20:02:25] greg-g: did i commit a faux pas? [20:02:46] ori-l: Haven't updated my deploy-branch clone of mediawiki in a while, so I'll be a few minutes [20:02:49] ori-l: no, I just wanted to give marktraceur a hard time for no good reason [20:02:56] Yeah he does that sometimes [20:03:01] greg-g: I'm on board. [20:03:07] I'll go pelt him with things later [20:03:17] You, however, are in range now [20:03:26] * greg-g is working from the Bernal library today! ha! [20:05:19] greg-g, ori-l: I merged an update to Thanks extension to the wmf6 deployment branch, but was not able to deploy it due to tin having dirty branch + permissions issues for wmf6. It should show up as a pending change to pull if you're doing a wmf6 deployment. See the backscroll for details. [20:06:18] maybe you'll have more luck than I did [20:06:32] hrm, looks like Reedy didn't have luck with it :/ [20:06:57] wmf7 branch is fine though [20:06:59] on tin [20:07:02] ping someone in ops to get the permission issue taken care of? what's the diff on the release notes/etc? is it important? [20:07:13] no [20:07:20] kaldari: k, thanks [20:07:32] but there's also a submodule diff for GettingStarted [20:07:42] that shows as a local diff [20:07:49] not sure what that's about [20:07:58] maybe ori knows [20:08:37] * ori-l knows. [20:09:12] Help Your country get to the top in the virtual world of E-sim. Become a businessman, military leader or a mighty president. Join for FREE now: http://secura.e-sim.org/lan.3355/ [20:09:57] wow, nice, I always wanted to be a military leader! [20:10:07] ori-l: anyway if you manage to pull the Thanks update for wmf6 let me know. It only needs a single sync-file. Doesn't need scap or anything. [20:10:39] Do I get to wear a uniform? [20:10:49] kaldari: k. spagewmf is handling our ordinary deployment items. i was going to do mark's afterward, and i'll see if i can do anything with thanks after that. [20:10:59] tho if reedy couldn't fix it i doubt i'd be able to do anything [20:11:06] I hadn't tried [20:11:19] ah, k. will look then. [20:11:26] ori-l: https://gerrit.wikimedia.org/r/68548 {{done}} [20:11:49] marktraceur: for the future, you should let jenkins merge [20:11:58] just CR +2, don't touch verified [20:12:02] no big deal, tho. [20:12:42] reedy@tin:/a/common/php-1.22wmf6$ git diff [20:12:42] reedy@tin:/a/common/php-1.22wmf6$ [20:12:50] kaldari: Try and sync whatever file you need now [20:13:02] Sync extensions/UploadWizard/includes/specials/SpecialUploadWizard.php and extensions/UploadWizard/resources/mw.LanguageUpWiz.js should do it [20:13:17] kaldari Yes, go ahead, still prepping [20:13:28] ori-l: I generally do, I saw kaldari had self-V'd earlier with his Thanks change, I was following that [20:13:30] kaldari: i'll do it [20:15:02] ori-l: too late, I'm already done :) [20:16:07] synced the Thanks fix. I'll let you guys handle the UploadWizard stuff [20:19:19] greg-g: Did anyone sign up for Lightning Deploy today yet? I'm hoping to get in a backport of !g 68546. [20:19:56] anomie: it's yours [20:20:12] anomie: mind adding it to [[Deployments]]? [20:20:32] anomie: also, any movement on the election email stuff? [20:20:34] greg-g: Already doing so ;) [20:20:37] rock [20:22:35] greg-g: For election stuff, Reedy and I have put the scripts from 2011 into Gerrit and we have some patchsets in progress for cleaning them up and adding the deduplication and already-voted filtering. https://gerrit.wikimedia.org/r/#/q/project:mediawiki/extensions/SecurePoll+status:open,n,z [20:22:45] awesome [20:23:16] thanks anomie and Reedy [20:23:42] who was responsible for the new login? Jorm? [20:24:01] Responsible how? [20:24:05] And you can talk to me. [20:25:15] heh [20:25:58] just wanted to thank them for a good job. [20:26:16] Ah. Then you want to talk to Steven Walling and S Page, really. [20:27:26] the only thing that seems to be missing is a link to ACC on the /signup page if the wanted name is 'too close' [20:27:42] spagewmf: ^ [20:27:43] No tailgating? [20:28:34] is there any plan to add the name availability mockup you have jorm? [20:28:46] Not as far as I know. [20:28:50] that looked real nice to me [20:29:45] T13|mobile: aww thanks, the check is in the mail :) What do you mean "ACC"? [20:30:01] account creation something [20:30:15] a local en.wiki process iirc [20:30:19] account creation [20:30:39] there is a link on the login but not on the signup [20:31:37] #wikipedia-en-accounts [20:42:04] Argh [20:42:26] greg-g: I may need to sneak into the lightning deploy after all, I failed to realize that commons was running wmf6. [20:43:00] So testwiki's Uploadwizard works great, not so much for commons [20:43:05] * greg-g nods [20:43:09] marktraceur: coordinate with anomie [20:43:18] Will do [20:43:35] anomie: I only need to sync two files, shouldn't take long [20:43:47] I'm going to head out soon, Rowan's 1.5 year checkup is this afternoon. I'll be online afterwards, but spotty/non-existent for a while [20:44:24] greg-g: Good luck, see you later [20:44:42] "this kid is so dirty, do you ever bathe him?" [20:44:43] .... [20:45:12] marktraceur: Ok. Unless you're making changes to i18n in wmf7, we probably won't even conflict anyway. [20:45:50] I am not [21:05:06] legoktm, MatmaRex, Reedy: I hope you guys didn't mind me cc you in that wm-bot bug earlier. if it was improper let me know and I won't do it again. just did it since you saw the bug and could probably clarify better than I. [21:05:46] nah, its fine :) [21:05:48] T13|mobile: nah, it's okay [21:05:56] kk :) [21:05:57] if anybody would mind, they'd un-cc themselves [21:06:12] that was what I figured. [22:20:52] kaldari: Another UW fix if you're up for more review [22:21:49] kaldari: can't do it today, gotta run to a meeting in a bit [22:46:10] hoo: to be precise about https://bugzilla.wikimedia.org/show_bug.cgi?id=49123 , even with rcid I don't see any patrol link [22:46:25] in short I'm unable to patrol new pages, what to do? [22:47:10] Nemo_bis: rcids are no more [22:49:38] hoo: so how do I patrol a page? [22:49:55] Nemo_bis: There should be a patrol link on the bottom [22:50:03] I don't have the rights there, so I can't test it [22:50:21] there isn't any link [22:51:15] mhm [22:51:19] might be broken then [22:52:00] Works on mediawiki.org [22:52:09] (patrolling worked for me at pl.wp an hour ago) [22:52:41] MatmaRex: that one still has wmf6 ;) [22:53:26] hoo Nemo_bis something was wrong with PageTriage/NewPagesFeed on enwiki earlier today, see screen shot at bottom https://saucelabs.com/jobs/ccf73f903d3e42b7a377cbd1a797a5b1 [22:54:13] :/ [22:54:27] that's old I think [22:58:25] * bawolff definitely did test that patrolling patch locally, it really should work :) [22:58:44] oh crap [22:58:54] I think I forgot something :/ [22:59:03] Nemo_bis: Does TWN have RC patrolling enabled? [23:00:08] Nemo_bis: ^ [23:01:02] graar [23:01:57] yes, they do (based on the patrol log) [23:02:54] bawolff: Ok, going to push a hotfix in a second [23:03:01] ok [23:03:22] I just want to test it very briefly [23:03:32] No testing allowed. [23:03:34] I'm guessing its the whole $wgRCPatrol is supposed to automatically imply $wgUseNPPatrol in a confusing way [23:04:01] bawolff: Yes... I forgot this :/ [23:04:07] yeah me too [23:04:49] ok, patch works [23:05:06] anomie: Right, are you deploying? [23:05:28] marktraceur: Getting ready to. What was it you were going to do again? [23:05:45] I just have a patch to wmf6 that I need to sync two files from [23:05:49] Fixes UW in Opera [23:07:00] Go ahead and do that [23:07:05] KK [23:07:33] bawolff: https://gerrit.wikimedia.org/r/68588 [23:07:46] reviewing [23:07:48] We should get that deployed ASAP [23:07:54] marktraceur: i'm behind you, so let me know when done. [23:08:00] KK [23:08:02] and of course prior to Monday or hell will break loose :P [23:08:21] * marktraceur waits for Jenkins to merge [23:08:23] It has +2'd, though, so [23:09:27] bawolff: Shall I cherry pick to wmf7? [23:10:07] Should probably check with the someone who actively touches the wmf branches first [23:10:22] * hoo eyes anomie [23:10:26] I'm not sure if there's some procedure for backporting to a wmf branch [23:10:34] or if one can just do it [23:10:49] I've did that before, you just need someone with shell to +2 it [23:10:54] * I've done [23:11:32] I've only ever backported things to the RELX_Y branches [23:11:33] bawolff: https://wikitech.wikimedia.org/wiki/How_to_deploy_code is the howto. The big deal is to not actually merge it until you're ready to deploy it. Otherwise someone else will be trying to deploy something else and get upset with you. [23:12:15] anomie: Well given I don't have deploy access, I guess I'll just leave the merge step to somebody else :) [23:12:29] (or actually all the steps since I imagine hoo will make the patch) [23:13:04] cherry picked the wrong change first... had to reset and stash ... hang on :P [23:13:09] hoo: do you want +sysop or do you have enough test wikis? [23:13:20] bawolff: Oh, ok. I take it this is something that is major breakage for users in wmf7? [23:13:28] Nemo_bis: I probably have enough :P [23:13:48] anomie: It would break new page patrol for wikis in certain (possibly common) configuration [23:13:59] oh common [23:14:06] I just cherry picked that wrong change again [23:14:41] anomie, ori-l, done, will be testing on commons soon, I'll let you know if it works, ping me if you need something [23:14:45] not that common on wmf [23:15:05] commonswiki, probably [23:15:08] anomie: https://gerrit.wikimedia.org/r/68589 [23:15:41] cherry-pick of https://gerrit.wikimedia.org/r/68588 [23:15:42] Nope, there are no wikis in wikimedia with $wgUseNPPatrol false but $wgUseRCPatrol set to true [23:15:46] * anomie starts l10nupdate. If anyone is waiting in line for something that isn't changing i18n files or doing a full scap, you can probably go ahead [23:16:10] bawolff: Are you sure? test2wiki seems broken as well [23:16:22] hoo: hmm, not according to the config file I'm reading [23:16:54] Looks like we got more bugs, then [23:17:17] hoo: are you sure its not PageTriage interfering on test2? [23:17:25] bawolff, hoo: Yeah, it looks like only huwiki, ruwiki, sqwiki, and ukwiki have $wgUseNPPatrol false. [23:17:35] bawolff: Ah, that could be [23:17:46] anomie: yeah, and none of those have $wgUseRCPatrol on, so we should be safe [23:17:53] Damn, that extension is ... [23:18:40] ori-l: What is it you're updating? [23:19:13] adding extension to wmf6 + config change sync [23:19:20] anomie: ^ [23:20:07] bawolff: abandoned the change ... [23:20:09] anomie: do you need more time? [23:20:54] You were right, it's the PageTriage extension... that one is the most confusing thing I saw in ages [23:21:47] ori-l: l10nupdate just finished against wmf6, running against wmf7 now. You could probably do your thing to wmf6 as long as you don't trigger another l10nupdate before the first one finishes (note scap does a (quick) l10nupdate, but sync-dir and sync-file are ok) [23:23:40] bawolff: I hope the PageTriage guys are aware of the fact that many still use Special:NewPages and just hiding the patrol link from core for style reasons isn't this clever [23:25:48] I guess https://gerrit.wikimedia.org/r/#/c/67872/ is supposed to not hide it totally [23:27:27] Oh, yes... I don't have the extension isntalled so I can't test this [23:27:39] but this certainly needs to get into wmf7 as well [23:27:56] I start to slightly dislike these fast deploy cycles [23:28:38] hoo bawolff I am surprised to find recent changes to PageTriage, that has been stable for some time now [23:29:07] I more develop with a wait 3 months for review anyhow, so I don't find faster deploy cycle effects me that much ;) [23:29:14] chrismcmahon: We've changed core a lot, so that they had to adapt these changes [23:30:07] hoo: I have some cross-browser tests for NewPagesFeed but none for Special:NewPages so I would not have caught the problem. I'll put that on my test backlog [23:30:42] chrismcmahon: You should... NPP is only deployed on enwiki while the core features are used on almost every wiki (including enwiki) [23:33:37] Who needs tests for enwiki? When things break they have no problem telling us [23:33:46] [23:35:43] bawolff: somebody said recently "sufficiently advanced testing is indistinguishable from monitoring" :-) This is something greg-g and I are going to be talking about really soon [23:35:52] hoo: There's actually slightly more time (same number of working days, but you gain a weekend in there: MTW before and TFM now) between the testwiki deploy and the non-Wikipedias deploy as there was with the old deployment cycle. And the same number of working days (but you lose the weekend) between the non-Wikipedias and enwiki too (WTFM vs MTWT), for that matter. [23:37:35] anomie|awayish: mhm, I haven't thought of it this way... doesn't sound bad :) [23:37:49] chrismcmahon: That really depends on if the sufficiently advanced testing also yells at you or not :P [23:38:23] bawolff: very true. there is a signal and noise issue to be solved also [23:38:45] hoo: The big difference is that we lose the gap between enwiki and other Wikipedias, and the gap between the other Wikipedias and the start of the next cycle. [23:40:05] anomie|awayish: The gap between other WPs and the next cycle was totally superfluous IMO, and the one between en and the other ones wasn't this useful either (if you break the other WPs you already did it wrong)