[16:39:57] /!\ AᎢТⲚ᛬ Thіѕ cһannel һas ⅿo⋁еԁ to іrc․freenode．ᥒᥱt #/jഠin /!\ [18:11:39] apergos: do you feel up to reviewing and eventually merging the initial BackupDumper MCR patch? [18:11:57] which one is the initial one? [18:12:02] the 'main slot only' one or...? [18:12:03] https://gerrit.wikimedia.org/r/c/mediawiki/core/+/443608 [18:12:10] the one we both worked on [18:12:34] i think if we both agree that the latest patch set is good, it can be merged [18:12:40] or do you think it needs another pair of eyes? [18:12:43] yep I will re-read and test the heck out of it, no I should not merge it I think, we'll have to nag someone else when it's ready to go [18:13:18] but that someone else will just be someone with +2 and general mw chops, nothing more [18:13:36] I mean we don't need someone who knows all about exports or whatever, it's fine [18:13:54] ok, give a +1 when you think it's fine, and find someone to poke, then :) [18:13:58] yep [18:14:27] the new ones are next in my queue as soon as that goes (where by 'goes' I mean 'I give a +1 and pass it on') [18:14:42] anomie is the only one oficially has time reserved for mcr work, but you may be able to get a quick review from someone else [18:15:46] I'll find someone ;-) [18:16:05] thanks for all the work on this [18:17:32] da nich für, as we say in German ;) It's what I do! [18:17:34] We (CPT) had a brief conversation about the ownership of the dump script code. In my mind, XmlDumpWriter and WikiExporter should be owned by CPT, but anything in maintenance/ should probably be owned by SRE. Do you think that's a good split? [18:18:01] which is cpt, sorry? [18:18:28] "core platform team" [18:18:54] and the way I think of it is that I (because when you say 'sre' realistically you mean me, there's no one in sre that's going to take ownership of any php code in mw) [18:19:35] sort of keep a close eye on or shepherd along stuff as far as the maintenance scripts, and sometimes touch things in the exports/ directory, but [18:19:39] duesen__: How'd it wind up that no one else is officially working on it anymore? [18:19:43] always always get someoen else to review and sign off [18:20:27] I've just recently revived the workboard for uh [18:20:54] mediawiki-import-or-export out of desperation but I don't want to 'own it' [18:21:01] anomie: tgr and addshore were borrowed, to get sdc out the door. that's done. so the two of use are left. [18:21:07] I can try to watch it and poke people and occasionally take things [18:21:39] those have not had any owner since [18:21:59] well since there was a mw core team way back in the day, and then they and the maintenance dump scripts were never owned by anyone [18:22:15] except maybe theoretically by tfinc for about 5 minutes after brion handed dumps to him [18:22:18] yea, that's the problem [18:22:25] they need to be owned by someone [18:22:40] I guess I'd better hear what 'owning' means [18:22:58] and sre (read: you) run them. that means that sre (that is: you) need to sign off on any changes. [18:23:22] I am ok with being one of the people to sign off [18:23:24] owning means all changes go through you, but you are also responsible for reviewing any patches against that code, and addressing any urgent bugs in that code. [18:23:30] that's why I brought back the workboard so I can follow that stuff [18:23:37] but that can never be one person alone. at least two, better three people. [18:23:43] but I should not be *the* reviewer [18:23:50] there's too much of mw innards I don't know [18:24:02] the problem is that nobody has *more* expertise in this than you do :) [18:24:10] and tbh I cannot track that, because much of my effort is spent on all the python or c or other infra around dumsp [18:24:26] well they ahve more expertise in the rest of how mw works [18:24:39] and what php style looks like today and which functions we use to do this and that with the db and etc [18:24:56] so I can be one of the people that must sign off [18:25:00] but we need another [18:25:18] they don't have to have speciaalized import/export knowledge [18:25:26] they do need to have general mw code nknowledge [18:26:29] yea, i understand what you mean [18:26:36] now if you give me 4 people any one of whom know enough and would be willing to be pinged for a last review (and I can be sure I'll find 1 of 4 that has time) [18:26:58] then I could do 'owner' in the sense of making sure things get moved along, testing, my part of 'reviewing' [18:27:10] ideally, this should be boards/teams, not individuals [18:27:18] and finding one of those general reviewers to make sure that it's right as far as mw code generally [18:27:38] well if a team wants to give its name to me that's fine too :-D [18:27:56] as long as when I go looking there's going to be *someone* available [18:28:41] CindyCicaleseWMF: *ping* Following up on xml dump code ownership with apergos above. Let me know if you read the backlog, otherwise i'll have to copy all of that into an email to you :) [18:29:15] I can otherwise write a on paragraph blurb that explains how things look from over here [18:29:42] if I spent a lot more time with mw core code I would be less reluctant (but even then a second set of eyes is, I believe, better) [18:29:43] apergos: well, I don't know any team other than CPT that would fit the bill. But on the other hand, CPT doesn't have the resources to effectively own all of core. So we are trying to share the load a bit. [18:29:44] anyways... [18:29:48] sure [18:30:01] Perhaps having official co-oweners or domain experts for cetain areas would help [18:30:02] and I'm not trying to give you folks more work or anythiing [18:30:33] just want to make sure I don't do something stupid form lack of awareness or solid knowledge of the current code base [18:30:51] co-owners is not a bad plan [18:31:00] apergos: i understand :) I'm just trying to finda home for code that currently has *no* owner. [18:31:03] I think you could probably get more people to give their time [18:31:16] if you market it that way :-) [18:31:24] maybe [18:32:12] apergos: for reference https://office.wikimedia.org/wiki/Core_Platform_Team/Meeting_Notes/2018-11-13#TEC13:_Code_Health [18:32:48] oh good [18:33:05] if I had more free cycles I'd follow along, this is a good discussion to be having [18:33:16] (and I've been one of the loudest complainers at certain times) [18:33:58] so to be clear if we go the co-ownership route, I would expect this to apply to stuff in include/exports too [18:34:02] xmldumpwriter etc [18:34:37] yes, i agree [18:35:19] this mean I would be setting some informal expectations on how long before I look at a patch, and makign sure that my manager knows some of my time will go to this on e a regular basis [18:35:31] thus guaranteeing that there will be time available [18:35:37] there's supposed to be a concerted effort to sort out stewardship... https://www.mediawiki.org/wiki/Code_stewardship_reviews [18:35:55] apergos: that would be excellent :) [18:36:25] anomie: do you happen to have the link to the big "who ownes which code" list? I can't find it [18:36:44] duesen__: https://www.mediawiki.org/wiki/Developers/Maintainers ? [18:37:07] anomie: right, that one, thanks! [18:37:16] in theory I somehow 'own' (not really at all but whatever) mwdumper and some other thing too because 'there is no one else' [18:37:31] well it would be nice to sort out anything Im ight be involved in and get it formalized [18:38:06] 'maintenance scripts' [18:38:19] maintained by... two people who are either gone or on other teams. yeah [18:38:56] oh yeah 'activeabstract' I sort of (not really by any stretch of the imagination) 'own' that too, ugh [18:39:04] apergos: maybe but yourself somewhere next to "export" on that table. I suppose stewardship will have to fall to cpt. [18:39:43] yeah I'll try to figure out things I touch and put 'Ariel &?'' on them [18:40:19] also honestly giving someone(s) ALL of the maintenance script is just unworkable [18:40:22] *scripts [18:40:39] how about *all* of core? ;) [18:41:23] hey [18:41:29] NotMyProblem :-P :-D [18:41:37] :P [18:41:39] ah I'l fix the 'datasets-general' thing [18:41:56] that's no longer accurate and a lot of other folsk are producing datasets/generating their own tasks in that project any more too [18:42:51] if you do wind up sending an email or need this discussion in a compact format someplace, ping and I'll do so [18:44:35] apergos: i just copy&pasted this to an email to cindy. and forgot to cc you. will fix, sorry [18:44:43] heh ok [18:46:39] welp. that's a long irc log :-D [18:47:00] I wil follow up on the email to avoid any more of it! [18:49:45] apergos: anyway, thanks for working with me on getting the dumps ready for mcr! [18:50:00] yep! you know I want it ready [18:50:12] i'll try to touch up the follow up patches next week. [18:50:17] great [18:56:26] anomie: what do you think? https://gerrit.wikimedia.org/r/c/mediawiki/core/+/473780#message-19eaf91bcad5525e649d9d1b16412b6eee7bf0c5 [18:59:04] duesen__: Replied there [18:59:33] thanks [21:03:06] duesen: I plan to get back to MCR patches (not sure if it's "official" but no one told me not to) but I want to follow up on some security stuff first. Feel free to ping me if there's anything urgent though. [23:37:56] Hello [23:38:44] Im having issues creating my wiki acc. [23:39:29] My IP have been ban for excessive attempts [23:39:55] Sol3PrDG: which wiki? [23:40:19] generally the rate limiter will let you try again after 5 minutes [23:40:35] Wikipedia @1st time [23:40:50] Ok [23:41:47] You can also ask at https://en.wikipedia.org/wiki/Wikipedia:Request_an_account [23:42:07] Should I try again in 5 minutes so? [23:42:17] Thanks for the link! [23:42:57] Just waiting the five minutes is probably the fastest method [23:43:56] Ok [23:44:06] Thank you very much [23:48:15] I've followed your 2nd solution :) [23:48:31] Thanks bawolff! [23:51:14] Glad it worked