=== almaisan-away is now known as al-maisan === al-maisan is now known as almaisan-away [20:32] wgrant: in support of heat updates being wonky: [20:32] https://bugs.launchpad.net/launchpad/+bug/816870 [20:32] <_mup_> Bug #816870: Distribution:+search (package search) timeouts < https://launchpad.net/bugs/816870 > [20:32] 36 heat, but AFAICT it should have 40 [20:32] 2*6 * 6*4 + 2*2 [20:33] 44 if the subs from dups are really honoured [20:33] wgrant: I just added a dup, and its off by 8 => 2 for the subscriber, 6 for it being a dupe -> I do think there is a glitch somewhere [20:34] or perhaps memcache [21:36] hi all [21:37] lifeless, if i was to put up a patch that re-added g+ (and maybe other) 'share' features, implemented just as a button linking to a url on the remote site, do you think that would be accepted? [21:39] that would have less (but still some) disclosure implications; it would have no security implications. It may have taste/design/usabilility aspects which I'm not the best person to address (such as it wouldn't show the counts -> wouldn't grow LP's capabilities at all). [21:39] the remaining disclosure implication would be the invitation to share private objects, definitely addressable. [21:50] poolie: speaking of patches [21:50] poolie: you've not removed the duplication in the show-timeline-to-developers code you landed [21:50] poolie: I'm feeling a little let down, as you committed to do so as part of landing it without delay [21:54] i was just thinking about that too [21:54] i will do it soon [21:55] my january was a bit fuller than expected, what with two hospital trips [21:55] i know we said 'january' [22:01] I understand, thank you [22:02] january FSVO january will be fine :) [22:02] this is a good example of why not to make decisions based on planned future work :) [22:02] hah, indeed [22:06] so i think it seems interesting enough that it's worth fixing the bugs and unifying the code [22:06] the bugs being mostly cross-browser js things [22:06] it got one or two upvotes [22:09] its nice [22:09] I wouldn't want it removed [22:09] I just don't want it to be costly in a costly area [22:17] flacoste: I'd kindof like to blog about our slack set-based approach; that ok with you? [given that you're orchestrating it, I don't want to steal your thunder...] [22:51] poolie: I fixed most of the bugs last week [22:51] poolie: Because I needed it [22:51] poolie: It works in Firefox now, toggles, and jumps to the right place when it shows. [22:54] wgrant, oh hooray [22:54] thanks very much! [22:55] and i'm so glad you found it worth doing [22:58] wgrant, so the only remaining significant thing is to unify the rendering with the oops code? [22:58] I think so [22:58] oops/profile [23:14] lifeless: Ah, duh [23:15] lifeless: The count at the top of the page is users affected including dupes. [23:15] that's a feature :) [23:15] possibly showing the details in a tooltip or something would be good though [23:16] That bug's users_affected_count == 4, so 2*6 + 4*4 + (2 + 2 on dupes)*2 == 36 [23:16] lifeless: Not a regression, and technically correct by the docs, but we should possibly fix it. [23:17] wgrant, ? [23:17] poolie: Hm? [23:18] i wondered if you considered the inclusion of dupe-affected users a regression or something [23:18] Ah, no, it's just that it's displayed including dupes, but heat calculation uses the raw value without dupes. [23:19] oh ok [23:19] Since heat calculation is now more transparent (because the hilarious aging stuff is gone), this has only become obvious recently. [23:19] don't talk to me about heat :) [23:20] Last time someone talked to me about heat, I deleted 7/8 of it :) [23:25] wallyworld_: Morning [23:25] wgrant: hello [23:26] wallyworld_: Can you (no-)qa your picker thing? [23:27] ok, give me a few minutes [23:27] Thanks :) [23:30] wgrant, :) [23:35] wgrant: can you create an appropriate bug please [23:36] wgrant: I doubt this will be the last we hear of it [23:36] wgrant: we may want to keep the updateheat garbo job but change it to use a featureflag to determine the oldest date of heat which is *invalid* [23:37] Possibly, yeah. [23:37] wgrant: then we can just land new algorithm, set the flag to the right date, and watch it run [23:37] (and then stop automatically) [23:37] That sort of thing is probably going to be useful in several places. [23:37] yes [23:37] eg. the bug denorm tables [23:37] Once I work how HTF we can recalculate BugSummary. [23:42] lifeless: bug #936607 [23:42] <_mup_> Bug #936607: Bug heat calculation's affected users count doesn't match the displayed value < https://launchpad.net/bugs/936607 > [23:42] thanks; uhm, I might tweak the title for search [23:43] Search is hopeless anyway, but feel free :) [23:43] bug 936607 [23:43] <_mup_> Bug #936607: Bug heat appears wrong due to affected users from duplicates not being included < https://launchpad.net/bugs/936607 > [23:44] I figure folk will try bug heat wrong / bug heat broken etc; fingers crossed [23:44] Yeah [23:46] * wallyworld_ takes car to mechanic - stupid fuel pump is broken :-( [23:46] wallyworld_: needs a software upgrade? [23:47] lifeless: i wish that's all it was. i used a rubber mallet to get it started [23:47] but that's impractical [23:48] hopefully won't cost toooo much [23:51] wgrant: so, sso detangle. You're all set on that ? [23:52] lifeless: Yeah. Hopefully make some progress on it soon. [23:53] wgrant: I think its more urgent than continued db efficiency in the medium term [23:54] wgrant: because it is a one off cost with a fairly long deploy tail [23:55] Yeah, but it's also a bit more work and harder to do in small chunks of downtime throughout my days. [23:56] yah [23:56] I hear a rumour you have mondays to self-direct [23:57] I've got stuff to finish off this week, but I expect so. [23:57] :P