[02:22:28] TimStarling: ping [02:22:46] hello [02:22:56] TimStarling: have you looked at http://pecl.php.net/package/v8js at all? [02:23:02] no [02:24:17] TimStarling: I was thinking about playing around with Extension:Scribunto a little bit [02:24:41] TimStarling: I was wondering about the feasibility of trying to add a JS engine [02:24:50] I can make a branch for you in gerrit [02:26:05] TimStarling: yeah, that would be great [02:26:05] my finger's on the button, do you want it? [02:26:08] ok [02:26:25] name... v8js? [02:29:36] TimStarling: yes, please [02:29:49] done [02:29:54] TimStarling: thanks [02:35:54] Ryan_Lane: http://svn.php.net/viewvc/pecl/v8js/ [12:54:12] <^demon> guillom: I just finished creating your WP-Victor and WMBlog repos in gerrit and importing all of the history. [12:54:59] hi ^demon: did you figure out why automerging broke? [12:55:29] <^demon> No, I haven't the faintest clue why. I dug through the error logs but didn't see anything. [12:55:43] <^demon> We'll throw some extra debugging in today to see why it worked for core but failed for extensions. [12:56:30] I guess someone had fun accepting all those [12:58:39] <^demon> I did about half of them. [12:59:12] locate includes/DefaultSettings.php | wc -l [12:59:12] 19 [12:59:19] not very efficient [17:12:13] Good morning, RoanKattouw, rsterbin, DarTar, auroraeosrose, mlitn! Are we ready to deploy the new feedback page for Article Feedback today? [17:12:21] Morning [17:12:31] I'm reviewing the last two revisions [17:12:45] One submitted late yesterday afternoon and one submitted as I was on my way over here [17:13:17] (That is, if Gerrit is using my local timezone. Not sure if it does that) [17:13:34] (Indeed it is) [17:13:40] yes it is [17:13:44] they're both very small things ;) [17:13:45] Thanks, RoanKattouw. Note that we only need to focus on https://gerrit.wikimedia.org/r/5863 for today's deployment (not https://gerrit.wikimedia.org/r/5870, which DarTar proposes we abandon). [17:14:21] Wait, 5870 was bad? [17:14:24] I just merged that [17:14:36] And of course, all of auroraeosrose's new revisions should be deployed today too. [17:14:36] ? [17:14:57] DarTar says we don't need 5870, only 5863, after all. [17:15:16] OK I'll revert 5870 [17:15:24] yeah, we don't need it [17:15:27] hello [17:15:36] It's not that 5870 was bad from a code standpoint, it's just that it's unnecessary for DarTar's purposes. [17:15:39] i appologize for the delay [17:16:12] Hello yoni_omniti. RoanKattouw is now reviewing the final revisions from auroraeosrose submitted this morning. [17:16:18] OK [17:16:20] 5870 is already n [17:16:23] *already in [17:16:26] Should it be taken back out? [17:16:26] perfect, thank you [17:16:32] hey [17:16:53] hi rsterbin [17:17:04] RoanKattouw: do you need me to do the revert, or is that on your end? [17:17:18] I'd rather revert 5870 unless the cta_edit suffix causes any conflict [17:17:31] as per my mail [17:17:35] DarTar: it's not going to cause a conflict [17:17:46] ok let's drop it then [17:17:59] drop the change or drop the cta_edit? [17:18:20] hi 20% people [17:18:25] * sumanah checks who that is [17:18:36] drop the change and stay with the current naming scheme [17:19:03] OK, reverting 5870 [17:19:04] roan's already merged it, so it's easier to drop it than to leave it [17:19:05] the only change we need is 5863 [17:19:10] ok, then [17:19:17] gwicke: preilly: and Aaron who isn't in yet. [17:19:23] DarTar: ignore me, Roan's already got it [17:19:30] yep, thanks [17:19:33] :) [17:19:35] auroraeosrose: what was that last revision? were you able to address the issue with the default relevance filter ranking? [17:19:45] hey preilly, per our conversation on Tuesday, do you have some thoughts on tooling you'd like to work on? [17:19:54] fabriceflorin: yes, that's what the last revision is for [17:20:07] that and fixing my stupid so the config variable for cutoff is actually used in the query ;) [17:20:27] maybe work on dippy-bird a bit, or gerrit metrics, or bz-gerrit integration, or -- how do you like working with JavaScript? :) Tomasz has a project you could do re MediaWiki JS tools to help managers report better [17:20:33] rsterbin: 5863 should fix the "[object-Object]" string in the log, right? [17:20:36] auroraeosrose: cool, that's what I thought. And I'm glad we're able to use the cut-off in this release too! [17:20:44] gwicke: how was your 20% day today? [17:20:47] sumanah: I'm going to work on Scribunto today [17:21:00] DarTar: i don't know what you're talking about [17:21:19] preilly: reviewing stuff from volunteers? [17:21:20] rsterbin: it's from a mail I sent earlier, hang on [17:21:25] sumanah: quite good, finally got Parsoid to expand Barack Obama without running out of memory [17:21:44] sumanah: no adding a JS engine to it [17:21:48] gwicke: that's great news, but how did it fit within https://www.mediawiki.org/wiki/Wikimedia_engineering_20%25_policy#Scope ? [17:21:51] gwicke: Yay! [17:22:03] > enwiki ext.articleFeedbackv5@2-[object Object]-edit_tab_link-click 20120426172122 1 42ryZnVBaMt4peJup1ssYWdVFwtgvYdhu [17:22:04] preilly: hmm, I'm looking at https://www.mediawiki.org/wiki/Wikimedia_engineering_20%25_policy#Scope to see if that fits [17:22:06] RoanKattouw: 77 seconds even [17:22:11] rsterbin: ^^ [17:22:15] DarTar: please, don't make us wait for last minute changes today. All these issues have to be addressed the day before deployment. [17:22:30] it's not a change, it's a clarificaiton of what 5863 covers [17:22:44] DarTar: Good. [17:22:51] sumanah: I might continue on 20% tomorrow [17:23:15] had to do urgent Visa stuff this morning [17:23:27] gwicke: there's certainly a bunch of Gerrit merge requests from volunteers piling up.... or you could work on documentation for Parsoid if there's a doc backlog [17:23:43] DarTar: that's never shown up in my clicktracking tests. i was getting "ext.articleFeedbackv5@2-option1-edit_tab_link-click" [17:24:03] sumanah: oh yeah, I actually did some documentation today ;) [17:24:25] preilly: I think it's cool that you want to work on Scribunto, but I'm not directly seeing how it falls in scope. [17:24:31] it's in the production logs and looking at 5863 I understood it was a fix for it [17:24:32] gwicke: cool! [17:24:41] if somebody could review https://gerrit.wikimedia.org/r/#change,5174, I could close the bug belonging to it [17:24:59] hi AaronSchulz, do you have 20%-related plans today? [17:25:30] DarTar: if that's what's showing up in the logs, it's something different happening in prod [17:25:35] sumanah: do you have a query for parser patches in gerrit? [17:25:47] preilly: maybe, if I could request, it looks like a lot of people are looking forward to your Git & Gerrit training, so maybe you could give a little of that some thought today as part of your 20% time [17:25:53] sure [17:26:11] AaronSchulz: code review I presume? or something else? [17:26:26] can you quickly explain what 5863 fixes? [17:26:30] gwicke: sorry, I do not, yet, but RoanKattouw might be able to help you with regex magic [17:26:32] gwicke: project:mediawiki/core file:^includes/parser should work I thnik [17:26:33] DarTar: or it could be browser related [17:26:37] Or something along those lines [17:26:48] RoanKattouw: cool, thanks! [17:26:49] regular cr [17:26:49] Ah, make that file:^includes/parser/.* [17:26:59] RoanKattouw: yep :) [17:27:27] rsterbin: looking at the diff I understood it fixes exactly this issue with edit-related events that didn't get the correct experiment_id prefix [17:27:38] DarTar: 5863 makes the edit tab/section events use the stage 3 prefixes instead of the stage 2 ones [17:28:13] isn't this exactly the same issue I reported above? [17:28:19] sumanah: just noticed that the session bug was re-opened yesterday: https://bugzilla.wikimedia.org/show_bug.cgi?id=35900 [17:28:36] until now all edit tab/section events have this [object-Object] string instead of the Stage 3 prefix [17:28:40] DarTar: if there's weird object stuff instead of a string, that's not an issue with it using the stage 2 prefix — "[object Object]" would *never* be right [17:28:40] DarTar: In the future, please ask all these questions the day before deployment, not on deployment day. This is making everybody wait unnecessarily. [17:29:22] DarTar: i vote in favor of deploying this fix and seeing if it happens again. [17:29:42] ok, I figure gwicke and AaronSchulz know what they're doing for 20%, and preilly knows my request :) [17:29:42] rsterbin: sounds good to me, it never happened in Stage 2 [17:29:45] I support rsterbin's recomendation. [17:31:19] !g 5863 [17:31:20] https://gerrit.wikimedia.org/r/#q,5863,n,z [17:32:20] fabriceflorin, yoni_omniti, rsterbin, auroraeosrose: OK, we should be all good on testwiki now. Code deployed, schema changes done, cache rebuild is in progress but is past English [17:32:24] RoanKattouw: there are quite a few sql things in this deployment for all the new featured stuff - need the new table items from alter.sql, then run the relevance_score.sql to populate relevance for existing entries, then run the filter_counts.sql (and those two scripts must be run in order) - there's nothing to get out of order in the alter script this time [17:32:25] So please hit test.wp.o and test [17:32:33] auroraeosrose: I did all of those [17:32:33] ok [17:32:36] sweet [17:32:42] Ran filter_counts and relevance_score [17:32:55] in order? [17:32:57] And a combined ALTER for all the new columns [17:33:01] filter has to be run after [17:33:07] Rawr [17:33:10] or the numbers won't be right [17:33:14] OK I'll run filter again :S [17:33:17] because I was populating a new field ;) [17:33:19] sorry [17:33:21] I learn something every week :) [17:33:23] No worries [17:33:25] LOL [17:33:30] OK done [17:33:39] RoanKattouw: Thanks for taking care of the SQL changes, much appreciated! [17:33:40] BTW, I'll put this in an e-mail too, but: [17:33:55] From now on, all schema changes must be vetted by our DBA, Asher Feldman [17:33:55] project:mediawiki/core file:^includes/parser/.* doesn't seem to work: operator not permitted here: file:^includes/parser/.* [17:33:57] i'm not getting any feedback at all on test [17:34:24] Technically this has been the rule since Tuesday, and we got a free pass this time. But not next time [17:34:27] sounds fine RoanKattouw, as long as there are no huge features added though db changes should be done [17:34:29] So submit your schema changes early [17:34:32] KO [17:34:38] gwicke: Maybe it was path: ? I forget [17:34:55] according to the documentation file: would be correct [17:35:00] hah [17:35:17] RoanKattouw: Thanks for the point about getting DB changes approved by Asher Feldman, and submitting schema changes early. Duly noted. [17:35:38] rsterbin: I'm not getting anything in relevant [17:35:45] but elswhere I am [17:35:49] yeah [17:35:49] we may be hitting the cutoff thing [17:35:55] cutoff is -5 [17:36:02] i just marked two comments helpful and they started showing up [17:36:03] which is VERY small [17:36:05] hmm- even the example file: query from the documentation fails [17:36:40] why would comments with no activity fall beneath the −5 cutoff? [17:36:42] yeah rsterbin, must be cutoff, I started pinging some stuff helpful and it's in relevant [17:36:52] because of the way the population script is written [17:37:04] it's applying a default - whatever to old feedback with no activity [17:37:14] ah [17:37:22] RoanKattouw: did file: work for you at some point? [17:37:39] I haven't tried it myself [17:37:45] RoanKattouw: I just told Kaldari about the new rule to have DB schema changes approved by Asher, and he is concerned that this could be a bottleneck in the future on projects like PT. Is there another person besides Asher that could review these? [17:37:57] RoanKattouw: k [17:38:08] just so we have a backup in case Asher isn't available [17:38:11] robla: https://bugzilla.wikimedia.org/show_bug.cgi?id=35900 popped up again [17:38:40] since most activity on enwiki is untouched, and it'd be better if the default filter wasn't empty, perhaps the cutoff should be lower than the value assigned to old untouched feedback? [17:38:43] rsterbin: I do not know of any use cases that would cause a comment with no activity to be below -5. [17:39:12] arroraeosrose? [17:39:30] fabriceflorin: when I wrote the population script, you asked me to assign a -10 to all old feedback with no activity to push it down [17:39:41] then we added cutoff, which is by default -5 [17:39:52] so it's not just pushed down now, it's cut off [17:43:22] auroraeosrose: I'm afraid I don't remember this request to assign -10 to all old feedback, and did not write it in the feature requirements page (there id a paragraph that talks about eventually having some form of declining relevance score, but it was pushed back this feature idea indefinitely). Is this a one-time thing? Or are you automatically assigning -10 each time a post goes past a certain date? [17:43:39] one time thing for inital population [17:43:51] because all the posts there needed to have a relevance_score applied [17:44:00] new stuff starts at 0 [17:44:36] RoanKattouw: 5863 is not yet in production, right? [17:46:34] Only in test [17:47:21] k [17:47:30] auroraeosrose: Hmmm. A one-time thing is not as bad. But are you saying that ALL the posts up to today will not show up on production? That would not be acceptable, because we need to test with existing content. [17:47:54] * robla looks at 35900 [17:48:23] aurorarose: Have we already done this change on production? or just on testing? Is that the SQL that RoanKattouw just ran? [17:48:30] just on testing [17:48:44] oh shoot - hey RoanKattouw we need a change in localsettings [17:48:51] there's a new right [17:49:07] rsterbin: the event prefixes for edit tab and section edit links I am getting from testwiki look good to me [17:49:21] great [17:49:22] auroraeosrose: OK let's have it [17:49:32] What's the situation with the -10 thing? [17:49:41] http://pastebin.com/02HrrV1A [17:49:47] aurorarose: Whew! I don't think we should run this change on production. I want to be able to sort through all the existing feedback on a range of articles, so we can work with existing comments, rather than have to wait for new comments to arrive. [17:49:48] hexmode: let's take the bug 35900 discussion over to #wikimedia-operations [17:49:54] k [17:49:54] fabriceflorin: we have to run the script [17:50:02] we can change that - value to be less then -5 [17:50:04] the cutoff [17:50:12] so like -3 or something [17:50:28] auroraeosrose: Could we run the script to assign a zero value to old comments instead? [17:50:46] that will make them the same as new comments then - and yes we can do that [17:51:09] RoanKattouw: pastebin is for the user rights stuff [17:51:21] robla: do you know when 1.20wmf02 is branched? [17:51:53] will be* [17:51:58] auroraeosrose: I am worried that lowering the cutoff would not remove items that have been flagged from the list. [17:52:20] fabriceflorin: .... I meant the "push down" value can be above the cutoff [17:52:25] for the one time run script [17:53:09] but still something a bit negative so we're sure that new items will appear above anything currently posted [17:53:13] but I can just use 0 as well [17:53:49] auroraeosrose: user rights applied on test [17:54:15] Nikerabbit: current plan: Monday morning [17:54:25] auroraeosrose: I like your suggestion of giving old posts a slightly negative value, but would suggest something even milder, like -1 (instead of -3 you propose). [17:54:28] (for some value of "morning") [17:54:39] that's fine [17:54:43] robla: Branch cut you mean? [17:54:51] yeah [17:55:39] RoanKattouw: can you change that -50 to -1 in relevance_score.sql or do you want me to do a commit/review for it? [17:55:43] it'll probably be Reedy cutting the branch. [17:55:49] rsterbin: the only event I don't see show up in the log from testwiki is for clicks on the "view source" tab for protected pages. It's definitely non critical but I was curious if it got implemented [17:56:18] DarTar: ooh.. no, that didn't [17:56:23] could you open a ticket? [17:56:31] I think we should just drop it [17:56:35] auroraeosrose: I'll live hack it for now but please do push it through the regular system when you have time [17:56:41] we can work without it [17:56:47] your call [17:56:58] Hi guys, the tools just started showing up for me on testing, at this URL: http://test.wikipedia.org/wiki/Special:ArticleFeedbackv5/Golden-crowned_Sparrow (not sure why I couldn't see them before). So I am just now starting to test. [17:57:07] RoanKattouw: no problem, I'll do the change [17:57:17] OK [17:57:28] I'm actually not entirely sure if just doing s/-50/-1/g is good enough [17:57:29] if you have a chance could you rerun it on test then? and then the filter script? [17:57:32] that's change things [17:57:33] For sure [17:57:50] Do I just change SET af_relevance_score=-50 to -1 or are there other things that need to change? [17:57:57] that's it RoanKattouw [17:57:58] I'm looking at stuff like -- If it's unfeatured we can ignore it, that would be -50 after 50 for featured which would be 0 [17:58:02] Or the +50 stuff [17:58:08] no - that's fine [17:58:14] OK, changing that and rerunning [17:58:44] auroraeosrose: I don't understand why the comment I just featured on testing does not appear at the top of the default relevant list: http://test.wikipedia.org/wiki/Special:ArticleFeedbackv5/Golden-crowned_Sparrow/31 [17:58:55] fabriceflorin: just a moment please [17:59:01] until RoanKattouw has rerun stuff [17:59:16] Done [18:00:08] Thanks, RoanKattouw! I will clear my cache now and try again. [18:00:48] hmmm, looks like hte default sort direction is still wrong [18:02:36] auroraeosrose: Yes, I can confirm the default sort order is wrong on all 3 platforms I am testing on, in both logged out, reader and editor modes. [18:03:15] fabriceflorin: it's just flipped correct? [18:03:22] it sorts right, just the default is the wrong direction [18:03:59] fabriceflorin: Would it be cool if I gave mlitn https://bugzilla.wikimedia.org/show_bug.cgi?id=35744 to start with? [18:04:04] auroraeosrose: Yes, it's flipped. If I click on Sort by relevance, I can get it to display featured items first. But this should be done by default. [18:04:06] it's a simple fix in AFT [18:04:19] just a minute fabriceflorin [18:04:25] javascript fix... gimme a moment [18:08:39] rsterbin: I just completed an extensive series of tests on testwiki, all events are being logged as expected, now hoping that this fixes the prefix issue when we push to production [18:09:03] DarTar: great [18:09:19] just for the record, http://test.wikipedia.org/wiki/Special:ArticleFeedbackv5/Golden-crowned_Sparrow is completely non-functional in IE7 [18:09:35] not sure about other IEs [18:09:57] yoni_omniti: we are looking for a simple bug for Matthias ( mlitn ) to start working on, so he gets more familiar with our code. Kaldari proposes that we give him https://bugzilla.wikimedia.org/show_bug.cgi?id=35744 . Do you think that would be a good bug for mlitn to start with? Has it been addressed on your end or not? I see that there are no more flag counters on testing, is that normal? [18:10:14] chrismcmahon: aftv5 should be disabled for ie7 [18:10:28] Hi chrismcmahon: Thanks for the heads-up about IE7. [18:10:29] it should work in ie8/9/etc [18:11:03] rsterbin: it is not disabled in IE7, I see "0 feedback posts... 60% found what they..." [18:11:17] ah, interesting [18:11:20] and some other random bits of the page [18:11:28] chrismcmahon: We are only supporting these platforms at this time, for testing purposes: http://www.mediawiki.org/wiki/Article_feedback/Version_5/Feature_Requirements#Platforms [18:11:55] RoanKattouw: 5898 please? js fix and the -1 [18:12:16] But chrismcmahon is right that we should disable IE7 -- or at least show a message. [18:12:35] !g 5898 [18:12:35] https://gerrit.wikimedia.org/r/#q,5898,n,z [18:13:16] * chrismcmahon should really get some more Win VMS in place [18:15:34] chrismcmahon: we have some browserstack accounts as well [18:16:24] although they're quite slow compared to the VMS [18:17:09] er VMs :) [18:17:54] fabriceflorin: good idea [18:18:01] i mean the bug [18:18:57] and as far as I remember, we only block the AFTv5 front-end on IE7 [18:19:02] need to double-check though [18:19:49] auroraeosrose: OK JS fix pushed [18:20:09] Thanks, yoni_omniti -- glad you like the idea. But I don't understand why the flag counters are not working on testing -- they are working on prototype. Any ideas? [18:20:29] fabriceflorin: please clear your cache and refresh the page [18:21:07] sorry, lost connection - any news on kaldari's bug proposal already? :) [18:21:16] yes, it's been approved [18:21:22] sent you an email [18:21:30] aight, will look into it tomorrow; thx [18:21:51] awesome, feel free to hit me up on IRC while you're working on it [18:22:12] in case you have any questions [18:22:48] or just to get advice on best-practices, etc. [18:22:49] I assume you'll be sleeping at that moment - but will definitely contact you when in doubt [18:22:59] I might not be :) [18:23:33] Thanks kaldari and mlitn1 -- Matthias, please email us (including the OmniTI team) if you have questions about this bug. I haven't looked at it in-depth, and it seems to be partially working on testing already, but I don't see any flag counters, which may need a fix. To be continued by email after deployment. [18:24:57] okidoki! [18:25:04] fabriceflorin: How's testwiki looking? I have a lunch appointment so I'd like to go to prod soon if everything looks good [18:25:08] fabriceflorin: flag counters won't show up if you don't have permissions... are you logged in? because it's an ajax page sometime you can get logged out but not see it [18:25:26] RoanKattouw, rsterbin: what do you think: can we push 5863, there's no pending issue AFAIK on testwiki that holds us off [18:25:48] looks good to me - everything appears to be working [18:26:14] DarTar: agree [18:29:12] cool, RoanKattouw ? I'd like to check if it fixes the weird issue which apparently didn't affect testwiki [18:30:11] Thanks, RoanKattouw, rsterbin, auroraeosrose -- it's working for me as well. I think we're ready to push this to production. Is there a way to test DarTar's 5863 code fix? Or can we only do this on production? [18:30:21] I don't know, that's a question for DarTar [18:30:26] But it sounds like he wants to go to prod too [18:30:38] RoanKattouw: it works as expected on testwiki [18:30:51] OK sweet [18:30:53] Let's go to prod [18:31:06] This'll take a while, so please don't complain about broken things until I say it's done :) [18:31:15] np [18:31:17] RoanKattouw: OK, sounds good. I recommend we push everything to production now, unless someone has any final issues. [18:31:25] Hmm, schema stuff first probably [18:32:38] DarTar: I apologize for being a bit rude earlier, that was uncalled for. Please let us know if there is anything we can do to help test the new code on production. [18:33:59] OK, RoanKattouw, we'll be patient ;o) -- we'll wait to hear from you. [18:36:01] fabriceflorin: np, I'll just wait to hear from RoanKattouw, make a couple of tests and see if I can reproduce the issue [18:36:16] RoanKattouw: Do you have any time estimates of when you think it might be safe to test the code on production? [18:36:36] 15 mins? [18:36:42] This is sort of hard to estimate [18:36:52] The only answer that's really correct is "when I say so" :S [18:37:30] RoanKattouw: Sounds good. I am just trying to plan our time. Will take a quick bathroom break now. ;o) [18:39:20] DarTar: Thanks, glad we're OK. Hopefully, all will go well on production ;o) [18:51:56] 10 mins before I disappear (hosting the brown bag today), RoanKattouw are we good to go yet? [18:52:04] Yes, it's done [18:52:09] sweet [18:52:59] RoanKattouw: Cool. Is that all of the AFT code, or just the fix that DarTar is reviewing? [18:53:10] Everything [18:54:33] RoanKattouw: Wonderful. Is is safe to test, then? [18:55:35] Yes [18:55:37] Sorry about that [18:55:45] It was taking a long time and I was distracted [18:57:45] this is weird, I am editing an AFT5-enabled article and I don't see any event show up in the log [18:59:20] oh hang on, here they are [18:59:25] strange latency [18:59:57] rsterbin: prefixes for section edit / edit tab events look correct now [19:00:05] Yeah udp2log has latency sometimes, seems to be related to it batching things up [19:00:06] making one more test before I go [19:00:22] RoanKattouw: did the localsettings stuff get applied? [19:00:35] RoanKattouw, auroraeosrose, rsterbin: It appears that not all the code may have deployed yet. I am seeing strange artifacts on this page: http://en.wikipedia.org/wiki/Special:ArticleFeedbackv5/Golden-crowned_Sparrow For example, the Most Relevant drop down menu now says there is only 1 post, but I am seeing lots of posts in the relevance filter list. [19:01:06] and was the filter count script run after the relevance score sql? [19:01:09] or before? [19:01:48] Oh, crap [19:01:58] I haven't run the relevace & filter scripts yet [19:02:07] I deliberately deferred that until after the code was live [19:02:09] Running no [19:02:10] w [19:02:31] ah - that's why fabriceflorin [19:02:40] he needs to run a couple scripts here ;) [19:03:01] rsterbin, RoanKattouw: log looks kosher [19:03:07] thanks for the fix! [19:03:20] taking off for the brown bag, bbl [19:03:37] OK scripts run [19:03:43] Counts & relevance should be correct now [19:03:56] much better [19:04:02] work right for you now fabriceflorin [19:04:10] auroraeosrose: Thanks! That explains it. I am also noticing that the -1 value is still including a lot of crap on a large page like this. How easy would it be to change this config setting back to -5 after all? [19:04:41] fabriceflorin: um... RoanKattouw would have to change the number in the script and rerun both scripts again [19:04:43] It's not a config setting [19:05:00] it's hard coded in sql [19:05:14] RoanKattouw: did the permissions stuff for localsettings.php go in? [19:05:18] Oh, right [19:06:11] Pushing that now, sorry [19:06:18] * RoanKattouw was doing 4 things at the same time over the past hour [19:06:22] OK, done [19:06:34] hehe [19:06:38] as long as it all gets in [19:06:50] RoanKattouw: if it's hard-coded in SQL, would it be possible for you to assign a value of -5 to all posts up to today? or up to a certain date, like April 1? [19:07:16] fabriceflorin: I would have to alter the scripts and they'd have to be rerun [19:07:31] fabriceflorin: That would disqualify all of them automatically, because the threshold is also -5 [19:08:44] Anyway -- I'm heading upstairs for lunch [19:08:48] RoanKattouw: Yes, I realized that the minute I posted this. Maybe we should just stick with what we have now, then. The Obama page looks pretty good with the current settings: http://en.wikipedia.org/wiki/Special:ArticleFeedbackv5/Barack_Obama [19:09:55] So I think we're good to go on your end, RoanKattouw_away -- auroraeosrose, rsterbin, yoni_omniti and chrismcmahon, how is this feedback page working for you all? [19:10:04] hmm - fabriceflorin do you have featured stuff available in your toolbox? [19:11:30] auroraeosrose: Yes, I just featured a bunch of items on the Obama page, and it works as intended. http://en.wikipedia.org/wiki/Special:ArticleFeedbackv5/Barack_Obama [19:12:10] hmmm, wondering why I don't have it [19:12:13] weird [19:12:35] must just not be in the right group ;) [19:12:47] the aft5 group should get those permissions [19:13:07] i know i'm in it, but i don't have the feature tool, either [19:13:32] Strangely, the aft hide account we used before to get access to hide functions doesn't show the new tools (Feature this post, Mark as resolved). I wonder why that is. http://en.wikipedia.org/wiki/User:Aft5hide [19:14:01] http://en.wikipedia.org/wiki/Special:ArticleFeedbackv5/Barack_Obama/32831 - doesnt look right - i think we need to increase right margin for the comment block [19:14:09] fabriceflorin: those permissions are controlled in localsettings [19:14:20] if more groups need "feature" permissions that needs to be added in localsettings [19:14:58] aftv5hide and aft5 groups then? [19:15:03] auroraeosrose, rsterbin: It may be that the user rights for aft-hide need to be updated? Maybe we can ask RoanKattouw_away after his lunch. [19:15:15] fabriceflorin: sure [19:15:23] otherwise looks like it works great [19:15:27] yep [19:16:37] yoni_omniti: I agree that we need a bit more padding for http://en.wikipedia.org/wiki/Special:ArticleFeedbackv5/Barack_Obama/3283, but it's not horrible for me. Which platform are you testing on? [19:16:49] chrome/mac [19:17:10] on a smaller screen the toolbox sits on top of the comments too [19:17:12] probably the result of the tool panel width bugfix. can be fixed for the next release [19:18:22] fabriceflorin: if you resize the window (both ways) you will se that sometimes the tool panel partially overlaps the comment text [19:18:31] auroraeosrose: Yes, I believe that aftv5hide and aft5 groups both need to be updated in local settings. Is this something you can do? Or does RoanKattouw_away need to do this? [19:19:08] RoanKattouw_away: needs to do this [19:20:32] yoni_omniti: Yes, I think we will need to make a number of presentational tweaks to the feedback page, and have asked Pau to propose simple design tweaks that could allow us to make things more readable without radically changing the basic structure of the page. Maybe we can discuss his proposals on our Friday 2pm ET call? [19:21:59] auroraeosrose: Thanks for the clarification. Would your yoni_omniti be willing to write RoanKattouw_away an email explaining what needs to be done exactly, to make sure both user groups get access to this tool? Please Cc: everyone, so we are on the same page. [19:22:51] For now, I think we are good to go, based on your feedback. Now is the time to experiment with all this, and start filing bugs or feature requests to make this page work better -- hopefully without requiring major fixes ;o) [19:23:41] auroraeosrose, rsterbin and yoni_omniti: Thanks so much for all your good work today! I am really happy we have these new features on production and appreciate all that you have done to make this possible! [19:23:59] I am now going to join the lunch in progress, if that is OK with you. [19:25:19] (In the previous post, I meant to ask auroraeosrose OR yoni_omniti to email RoanKattouw_away (not "your yoni", sorry ;o) [19:36:53] thanks everyone [21:06:11] Hey RoanKattouw_away: When you get back, I have a Gerrit permissions thing to run by you. [21:25:58] rsterbin: quick update on the logs [21:26:26]