[02:42:11] JD|cloud: Can't you simply disable it? [02:42:33] Elsie: that's kinda besides the point [02:42:37] Not really. [02:42:47] what's relevant here is that the community didn't actually want it [02:42:58] Not really. [02:43:03] There was unanimous support for it. [02:43:05] no [02:43:07] Yes. [02:43:09] no [02:43:11] just no [02:43:14] that's not how consensus is built [02:43:14] Well, yes. [02:43:25] Consensus is built by asking people if they'd object. [02:43:27] Nobody objected. [02:43:29] You can't elect a functionary with 4 supporting votes alone [02:43:33] Elsie: poorly advertised [02:43:42] and closed by proposer who had a COI here [02:43:49] I suppose it's a good thing we didn't elect a functionary, then? [02:43:50] and consensus can change [02:44:03] I'm not sure the closer matters if it's unanimous. [02:44:06] which it clearly has within a few minutes [02:44:19] * Jasper_Deng_busy will refuse to argue with Elsie over something that should be blatantly obvious here [02:44:32] Where should it have been advertised? [02:44:34] Meta:Babel? [02:44:40] Global site notice to all wikis? [02:44:45] I gave you the answer to that TWICE [02:44:53] once on-wiki, another time on Bugzilla [02:45:09] Stewards' noticeboard? [02:45:13] That seems a bit silly. [02:45:48] Anyway, I'm not sure a few stewards hating it makes it a bad idea. [02:45:55] it's not just a few stewards [02:45:59] The stewards were going to hate anything. [02:46:06] I'm not a steward, and I hate it [02:46:12] non-admins hate it too [02:46:13] Are you a global sysop? [02:46:17] I'm not a GS either [02:46:30] and it shouldn't matter [02:46:32] stewards are people [02:46:37] We think. [02:46:43] Did you have enhanced recent changes enabled? [02:46:45] they are members of this community like anyone [02:46:52] ... [02:46:55] Nobody said or suggested otherwise? [02:47:01] What did I miss? [02:47:02] after it was /forced/ on me [02:47:04] I don't understand "...". [02:47:09] Jasper_Deng_busy: So disable it? [02:47:17] I think most people appreciate the condensed format. [02:47:25] If you don't, don't use it. [02:47:28] T13|sleeps: Elsie clearly doesn't get that the support for that proposal wasn't sufficient to force it onto the broader community [02:47:44] What proposal? [02:47:47] Jasper_Deng_busy: Are you being forced to use it? [02:47:51] I /was/ [02:47:52] Jasper_Deng_busy: Is anyone else? [02:47:55] Jasper_Deng_busy: When were you? [02:48:04] When you noticed it had been enabled and you disabled it? [02:48:11] the moment it was enabled [02:48:18] w/o instructions on how to disable it [02:48:24] frankly it's a nuisance [02:48:25] Yes, we should add instructions. [02:48:31] I didn't use it for years. [02:48:35] I started using it recently. [02:48:40] Enhanced changes, that is. [02:48:42] It's nice. [02:49:04] Oh... [02:49:04] I'm not sure about the magnify thumb icon or whatever. [02:49:09] * Jasper_Deng_busy will note that he thinks Nemo_bis should refrain from doing these kinds of things in the future, it looks really bad after his being desysopped [02:49:19] He's going to be desysopped from Bugzilla next. [02:49:23] (!) [02:49:24] you don't close a proposal you make, simple as that [02:49:28] Jasper_Deng_busy: Are you admin on Meta-Wiki? [02:49:31] I'm not [02:49:32] Enhanced recent changes? [02:49:34] Lame. [02:49:44] T13|sleeps: Just keeping asking questions. We'll continue to ignore them. [02:49:53] I've been using it for years. [02:49:57] https://meta.wikimedia.org/wiki/Meta:Babel#Enable_CleanChanges [02:50:07] I am quite mad at how this proposal was implemented, just like enwiki watchlist [02:50:10] Jasper_Deng_busy: There are usually exceptions to that rule. [02:50:15] Elsie: quit pinging me [02:50:23] this wasn't one of them [02:50:29] Unanimous support? [02:50:32] After a week? [02:50:34] 4 people [02:50:36] is not enough [02:50:47] How many would you, umm, expect on Meta-Wiki? [02:50:49] in fact, for proposals that don't get more support, it usually takes more time [02:50:53] I would expect 20ish [02:51:05] Can you point to a similar change with 20 supports? [02:51:24] I think CodeEditor had about four supports (if that). [02:51:36] The namespace additions had... zero supports, I think. [02:51:43] Well, one, if you include the proposer. [02:51:49] Who also filed the bug, I believe, heh. [02:51:55] So this doesn't seem unprecedented. [02:51:56] * Jasper_Deng_busy really wants to stop arguing over this, it's blatantly obvious that this wasn't done properly; removing the adminship-on-other-wikis requirement [02:52:10] Anyway, I agree with you about advertising how to disable it. [02:52:12] I'll do that now. [02:52:15] the right thing to do is to merge the patch disabling it [02:52:22] and then call a broader RfC on this [02:52:23] Seems silly. [02:52:34] you simply don't mess with such a prominent part of the UI without broad consensus [02:52:46] /especially/ making it default for all [02:53:14] https://meta.wikimedia.org/wiki/Broad_consensus [02:53:26] https://meta.wikimedia.org/wiki/Prominent_UI_components [02:53:55] Looks like it was poorly excuted to me... [02:54:00] Anyways.. [02:54:02] ZZZzzZzzz [02:54:02] * Jasper_Deng_busy thinks you're just trying to cover up the obvious fact that this was poorly executed and should be redone properly [02:54:28] Jasper_Deng_busy: I'm not sure what you think I have to gain. [02:54:39] Nemo started a discussion in the appropriate place. [02:54:42] And waited a week. [02:54:45] didn't advertise it properly [02:54:51] didn't ask an uninvolved admin to close it [02:54:55] According to some standard you made up this evening. [02:54:56] Okay... [02:54:58] didn't make announcement [02:55:01] I'll bite. [02:55:02] T13|sleeps: Good night. [02:55:03] these are universal standards [02:55:11] Announcement where? [02:55:13] Meta:Babel? [02:55:16] no [02:55:18] watchlist notice [02:55:23] was necessary for this [02:55:29] Is Meta-Wiki's watchlist notice ever used? [02:55:30] What you would have to gain is a consensus you clearly didn't have. [02:55:31] Like, ever? [02:55:35] yes... [02:55:46] Really? For what? [02:55:52] site notice was used for global bans [02:56:03] https://meta.wikimedia.org/wiki/MediaWiki:Watchlist-details [02:56:07] That's the same page en.wiki uses. [02:56:09] also, quit trying to use other things in the past from excluding what should be done [02:56:13] I'm the only person to have ever edited it. [02:56:26] Has anybody ever done a watchlist notice on Meta-Wiki ever? [02:56:26] that it wasn't used in the past doesn't mean it couldn't have been used for this change, which really has no analog [02:56:37] except the enwiki watchlist thing (you know how that ended up) [02:57:05] I don't know what you're talking about. [02:57:09] Regarding en.wiki's watchlist. [02:57:19] I was the person who began watchlist notices, as I remember it. [02:57:19] enabling "Mark all pages as visited" [02:57:25] Oh. [02:57:31] Okay. [02:57:35] How does that relate to this? [02:57:47] It was also an ill-advertised proposal that annoyed the community [02:58:03] So this change has no analogies, except for the ones you come up with? :-) [02:58:22] I can't think of any other ones, but things like global bans, global abusefilter etc. are simply not of this nature [02:58:42] the bottom line is that this should've been advertised regardless of previous things [02:58:47] because it was a very prominent UI change [02:58:50] Global discussions require global input. [02:58:56] this wasn't a global discussion [02:58:58] If this change had been made to all Wikimedia wikis, I'd agree with you. [02:59:02] and those were policies [02:59:04] in question [02:59:05] And it didn't have a global scope... [02:59:07] not user interface [02:59:17] It only applied to a single wiki. [02:59:22] not my point [02:59:25] So it only had discussion on that specific wiki. [02:59:26] Okay. [02:59:27] What's your point? [02:59:45] It was poorly executed: ill-advertised, proposer didn't seek neutral closer [02:59:50] insufficient consensus [02:59:58] (really, could Nemo not have waited a month??) [03:00:25] * Jasper_Deng_busy hates "lack of opposition = consensus" when the !votes are unrepresentative of the whole community [03:00:57] Silence is consent. :-) [03:01:01] no [03:01:04] It's always been that way. [03:01:06] not when it was in the dark [03:01:12] In the dark? [03:01:13] * are unrepresentative of the whole community* [03:01:19] Did Nemo use invisible ink on Meta:Babel? [03:01:26] he might as well have [03:01:32] Most community discussions never involve the whole community. [03:01:35] because that was the overall effect [03:01:38] well this is one that should've [03:01:39] If they did, we'd never, ever get anything done. [03:01:55] you don't need the entire community but you need a representative sample [03:01:59] I doubt it would have made any difference if Nemo had waited a month. [03:02:01] 4 people simply ain't it [03:02:17] The only difference would have been that the discussion would've been archived by a bot. [03:02:20] silence is not consent in this situation because people were not informed [03:02:35] They weren't informed on Meta:Babel? [03:02:37] * Jasper_Deng_busy doesn't like how you think this could not have been executed any better [03:03:08] Well, if you think people should advertise these discussions in specific places, you ought to write that down. [03:03:17] not in specific places [03:03:26] I'm not sure how Nemo or anyone else is supposed to know what your standards are. [03:03:29] but one hardly-watched forum doesn't suffice [03:03:31] For advertising a discussion of this nature. [03:03:34] my standards are the stewards' [03:03:52] if a proposal is jarring, you post it to everyone [03:03:55] whatever it takes for that [03:03:55] The stewards don't act on requests involving four people? [03:04:02] ... [03:04:17] I think most admins are elected like that. [03:04:20] And bureaucrats. [03:04:21] steward requests aren't usually consensus-requiring except requests for rights... [03:04:40] So the stewards' standard is not having consensus? [03:04:43] Unless it's a right change? [03:04:45] rights * [03:04:50] huh? [03:04:57] for SRG [03:05:03] Oh, ugh. [03:05:04] You said we should follow the stewards' standards, I think. [03:05:05] and SRCU the consensus is already there [03:05:19] I'm not sure what SRCU is. [03:05:33] Elsie: the bottom line is that it was not done properly, no matter how you put it, no matter what previous practices were. [03:05:38] and should thus be reverted. [03:05:46] I cannot put it any better. [03:05:58] quit telling me that it could not have been advertised better [03:06:10] or that 4 people was enough for such a sweeping change [03:06:29] * Jasper_Deng_busy sighs and gets back to work [03:06:47] It wasn't done properly according to the standards you made up this evening. [03:06:53] I'm shocked we didn't live up to them. [03:07:25] They're not /my/ standards. They go without saying. If done properly, so many users wouldn't have been annoyed. [03:07:41] * anomie files https://bugzilla.wikimedia.org/show_bug.cgi?id=54090 [03:08:55] JD|cloud: So many users? Heh. [03:09:02] About as many as have supported have complained. [03:09:07] Maybe four or five? [03:09:14] * JD|cloud won't argue that [03:09:44] more users were unhappy in -stewards [03:17:55] https://meta.wikimedia.org/wiki/Template:CleanChanges [03:22:33] Jasper_Deng_busy: Any chance you're a translation admin? [03:22:38] no [03:22:41] Damn. :-( [03:22:44] * Jasper_Deng_busy hates the translate extension [03:22:51] Yeah, I'm not a big fan either. [03:22:56] It'd be cool to get that notice translated. [03:23:10] https://meta.wikimedia.org/wiki/Special:Watchlist [03:23:15] It looks kinda trippy there. [03:23:19] It's getting two borders. [03:23:22] might want to slim it down [03:23:22] I kind of like it. [03:23:36] You should be able to edit it. It's just a normal template. :-) [03:23:38] Feel free. [03:23:50] It's a bit big and wordy. [03:27:59] God, Translate is so broken... [04:12:40] [07:59:40 PM] (really, could Nemo not have waited a month??) <-- would anything ever get done that way? [04:12:46] "No one objected, lets wait a month" [04:42:02] legoktm: let's * [04:42:17] I'm much more concerned with how broken Translate is. [04:42:20] :/ [04:42:25] "TUX" [04:42:31] Who's TUX? [04:42:42] Translation User Experience [04:43:06] Every time I try to update a page for translation, I get that blue/green Wikimedia error. [04:43:13] heh [05:08:53] Elsie: on what wiki? [05:18:40] Nemo_bis: Meta-Wiki. [05:18:54] I have four bugs to file against the extension. [05:18:58] I'll do it later. [05:23:34] Ok thanks. [05:24:08] Elsie: note that there was a deployment yesterday (or today your TZ... no yesterday for you too) [05:24:37] hi! is it possible to block some external links? we get a lot of spam to a particular domain, and it would be nice to just block it completely (on sv-wikt) [05:26:51] skalman12: yes, ask a sysop to edit the local spam blacklist [05:27:05] I am a sysop - where's the spam blacklist :) [05:27:14] MediaWiki:Spam-blacklist [05:27:33] https://sv.wikipedia.org/wiki/MediaWiki:Spam-blacklist [05:27:43] each line its a regular expression [05:27:53] hah, it's a very new thing for sv.wiki I see [05:28:52] Nemo_bis: I'm on sv-Wiktionary :) but thanks anyway [05:30:05] skalman12: oh, sorry, I missed the t; should get breakfast [05:31:02] Elsie: https://bugzilla.wikimedia.org/show_bug.cgi?id=53769 and added you to cc [05:31:30] anyway, thanks a lot! I'll look at other wikis and figure out what to do [08:56:46] hello [11:05:40] apergos: hello [11:05:47] hey [11:06:07] * apergos looks around for a parent1337 [11:06:32] pinged [11:06:59] ok [11:13:26] ok [11:13:26] I couldn't get it to build on precise [11:13:27] precise has a 3.6 gcc, which doesn't take the same stdc++ flag [11:14:20] I put in the appropriate changes to the cmake file [11:14:27] but... [11:14:51] /home/ariel/incrementals/Diff/Changes/../../DumpObjects/../DumpException.h:25:38: error: expected ‘;’ at end of member declaration [11:14:51] /home/ariel/incrementals/Diff/Changes/../../DumpObjects/../DumpException.h:25:47: error: ‘override’ does not name a type [11:14:51] /home/ariel/incrementals/Diff/Changes/../../DumpObjects/../DumpException.h:18:7: error: looser throw specifier for ‘virtual UserException::~UserException()’ [11:14:51] 3.6? the one i'm using 4.8 [11:14:51] and lots of other fun things [11:14:51] er 4.6 sorry [11:14:51] and i think 4.6 doesn't have support for C++11 [11:14:58] > check_cxx_compiler_flag(-std=c++0x LIBCXX_HAS_STDCXX0X_FLAG) [11:14:59] > check_cxx_compiler_flag(-std=c++11 LIBCXX_HAS_STDCXX11_FLAG) [11:14:59] > if(LIBCXX_HAS_STDCXX11_FLAG) [11:14:59] > add_definitions(-std=c++11) [11:14:59] > elseif(LIBCXX_HAS_STDCXX0X_FLAG) [11:14:59] > add_definitions(-std=c++0x) [11:15:00] > endif() [11:15:40] no, that's the best we can do, but that's what all our boxes run (precise), and they have 4.6, that's the latest package [11:16:03] oh, that could be a problem [11:16:06] uh oh [11:16:37] ok well if this turns out to be a Big F-ing Deal and needs a serious rewrite, it jus tmeans that testing is put off for awhile [11:16:51] i'll look into it and try to build it with 4.6 [11:17:00] ok [11:17:34] if fixing the build is going to take several days, put it off, get your last tweaks in and do the docs [11:18:34] it's hard to say, i don't know what level of support for the new features does 4.6 have [11:18:40] ok [11:19:01] what would be required to get 4.8 there? [11:19:17] also, wouldn't it be possible to compile on another computer? [11:19:31] well I built on my box here (4.7 something, it was fine) [11:19:40] but that doesn't run on a precise host [11:19:54] there's an explicit check for the gcc version [11:20:18] but it would mean a backport of a pile of crap that might not even build, so practically speaking... [11:20:27] we'd be waiting for the next move to an lts [11:20:59] http://gcc.gnu.org/projects/cxx0x.html [11:23:42] that doesn't look that bad, i use dectlype in few places, but rewriting that shouldn't be complicated [11:24:07] I can give you the error output if that's helpful [11:24:33] if you don't have a precise vm sitting there... I can also pull and build if poked to do so [11:25:49] the vm i'm using runs on Ubuntu 12.04, i think that's the number for precise [11:26:03] (i got the new gcc from an unofficial repository) [11:26:08] ah :-D [11:26:11] tsk tsk [11:26:42] hello apergos [11:26:48] what's wrong with that? [11:26:57] hello (in a meeting) [11:27:13] hello svick [11:27:16] nothing wrong for you, just not something we would be likely to do for our production hosts [11:27:26] right [11:27:44] what is the backbone of wikimedia foundation ? [11:28:28] what are your thoughts about the metadata compression? [11:29:55] for trwiki, out of the 2.3 GB, 500 MB are comments, so i think i should focus on that first [11:30:17] that's a lot of comment [11:32:04] assuming 14M revisions (per [[tr:Special;Statistics]]), that's ~35 bytes for comment per revision [11:32:39] which sounds reasonable, since they're not compressed at all at the moment [11:32:53] 35 bytes isn't so much [11:33:41] you figure most edits that aren't page creation or revert are going to have the section name and some short text [11:34:20] (worse on wikis with multibyte character encoding) [11:34:43] right, but i'm not sure how to take advantage of that [11:35:01] I don't think that's something for your code to look at [11:35:09] let the compression take care of it [11:35:26] i think i'll try compressing them in groups, just like i do now for revision texts [11:36:03] yes, you [11:36:18] 'll likely get the biggest gain that way [11:39:39] there are other some smaller things i could do with the compression, but i guess they will gain at most something like 100 MB on that dump [11:39:55] eh, leave those til much later I'd say [11:40:04] right [11:40:19] if you have a todo list somewhere on wiki you could add those to 'optional-later' [11:42:00] i don't really have anything like that, there is TODO.txt in the repo, but i use that mostly for short-term stuff [11:42:30] ok [11:42:40] well having a running list somewhere for 'stuff to do after gsoc' might be good [11:42:54] not necessarily stuff you have to do but that oughta get done/checked/tested eventually [11:42:58] yeah, i guess [11:43:16] by somebody [11:44:42] apergos: the after gsoc checklist should also be more than just one entry of "make sure someone touches it" [11:45:18] don't worry, there's already a few todos [11:45:26] actual todos, that's why I want such a list [11:45:39] my goal is to get it into production. so [11:45:44] though maybe you have a better idea about that than me; the big things i can think of that belong there are code review and integrating with your python scrips [11:46:29] sure, it's the small stuff that we're lilely to lose track of (or I will) [11:48:48] I wonder how hard it would be to have a python library that would be able to read a page and its revisions from a new format dump... mainly a question of the compression library bindings [11:48:53] anyways that's way off track [11:50:06] well, that's what i envisioned originally: that i would write a library for which you could easily write binding to any language [11:50:17] later :-D [11:50:23] see? a todo list would be great! [11:50:27] yeah [11:50:41] but it's not a todo list for you, it's a todo list for the project [11:51:01] anyways... [11:51:38] about the precise build, if as you get into fixing it, you see it will take days and not just 'today', put it off til afterwards [11:51:48] otherwise, I'll be watching the commits :-D [11:53:34] ok [11:54:23] i started the TODO list at https://www.mediawiki.org/wiki/User:Svick/Incremental_dumps/TODO, but i couldn't think of any other things to add there [11:54:35] so feel free to edit it, if you think of something [11:54:43] apergos: We could set up a section in bugzilla for it! [11:54:51] *runs away before people hurt me* [11:54:52] not yet p858snake|l [11:55:00] but we will [11:55:01] (maybe some things from User:ArielGlenn/Dumps new format (deltas, changesets) belong there?) [11:55:09] probably! [11:55:36] what are the things you were thinking that might gain you only 100mb in compression? [11:56:08] because 'see if it's worth it to...' and a list of those, would be good [11:56:37] adding index for users, so user names aren't repeated (but users currently take only 130 MB) [11:56:47] add it in there! [11:57:14] and compressing indexes using varints (since they mostly consist of consecutive integers) [11:57:51] ok, i will [11:58:49] if there are ideas you want to check out, or you want them checked out by someone eventually, even if you don't know they are going to work out finally, this is where to put them [11:58:49] keep 'em coming [11:58:49] :-) [11:59:55] i have few other ideas, but they'll help even less than those two (shaving 1-3 bytes per revision) [12:00:20] apergos you are in a meeting ? [12:00:21] hmm 500 million revs = 1500 million bytes [12:00:29] yes cortexA9, in a meeting [12:00:33] ok [12:00:38] right [12:00:52] for a downloader, might be worth it... [12:01:14] why torrents have 0 seeders ? [12:01:25] so, it looks like i have enough work until monday, and i'll start with documentation on tuesday [12:01:28] anything else? [12:01:46] nope, good to go [12:01:57] ok, see you monday [12:02:01] talk to you then [12:02:23] cortexA9: I have n idea, we don't manage those torrents (thank goodness, we already have plenty to do) [12:02:59] always 0 seeders apergos [12:03:04] we need to do something. [12:03:17] *you* need to do something :-) [12:03:23] me ? [12:03:24] you're in a much better position than I am [12:03:28] yes, you [12:03:40] i don't have the file. [12:03:42] who uses those torrents? who seeds them, starts them up? [12:03:44] i can't seed. [12:03:58] find out, find out why seeding stops [12:04:09] cortexA9: if you're talking about the burnbit torrents, then 0 seeders is not a problem, since there are also the webseeds [12:04:13] getting the file is easy, download it (preferably from a mirror) [12:04:23] yes i talking about that. [12:04:31] it foesn't kill the torrent but it does mean downloads are slow [12:04:55] oh 2 seeders. [12:05:01] right now [12:05:22] if you want to work on health of the torrent seeding community, this is a great service, [12:05:32] something I can't really take on [12:05:39]