[00:00:48] (PS1) Adamw: disable Contacts and Contributions export and search stuff. [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/90458 [00:00:49] (CR) jenkins-bot: [V: -1] disable Contacts and Contributions export and search stuff. [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/90458 (owner: Adamw) [00:01:16] K4-7131: ^^ [00:01:48] (PS2) Adamw: disable Contacts and Contributions export and search stuff. [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/90458 [00:01:49] (CR) jenkins-bot: [V: -1] disable Contacts and Contributions export and search stuff. [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/90458 (owner: Adamw) [00:01:51] hehe. [00:02:07] ah, cool. I have to get out of here to not miss my flight and whatever... but I might have a minute in the airport. [00:02:20] I'll sic mwalker on it, no worries [00:02:27] mwalker: ^^ ! [00:02:39] heh [00:11:29] mwalker: https://gerrit.wikimedia.org/r/#/c/90459/ [00:15:58] (CR) Mwalker: [C: 2 V: 2] disable Contacts and Contributions export and search stuff. [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/90458 (owner: Adamw) [01:06:48] (PS1) Adamw: eliminate old special case handling that was messing with throttling. [extensions/CentralNotice] (wmf_deploy) - https://gerrit.wikimedia.org/r/90469 [01:07:06] (Abandoned) Adamw: eliminate old special case handling that was messing with throttling. [extensions/CentralNotice] (wmf_deploy) - https://gerrit.wikimedia.org/r/90469 (owner: Adamw) [01:07:35] (PS1) Adamw: eliminate old special case handling that was messing with throttling. [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/90470 [01:07:55] mwalker: ^^ if u would [01:10:39] (CR) Mwalker: [C: 2] eliminate old special case handling that was messing with throttling. [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/90470 (owner: Adamw) [02:20:45] (PS1) Adamw: Found a bug with priority-based banner hiding [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/90475 [02:22:06] mwalker: well, that was annoying. There was another bug, minor this time. banners from lower priority campaigns were incorrectly assigned a minimum of one slot, when it should have been zero. https://gerrit.wikimedia.org/r/#/c/90475/ [16:45:06] (CR) Adamw: [C: 1] "(6 comments)" [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/90261 (owner: Katie Horn) [17:31:21] can someone please fix the banner for the fundraiser, I have clicked it away for about 5 times already, seems to be vanished for some time but then returns [18:22:10] Romaine: are you logged-in? banners shouldn't be showing to logged in users [18:22:20] yes logged in [18:22:27] happens on multiple projects [18:22:37] * Romaine is located in NL [18:22:53] thanks for the info, Romaine, let me go take a look [18:23:20] normally a notice should be shown only once and be hidden after clicked away [18:23:49] right, banners are supposed to hide in several ways [18:24:51] Romaine: you're referencing fundraising banners, not the funds dissemination committee, correct? [18:25:05] i see FDC banners when logged in [18:25:22] sorry [18:25:27] funds dissemination committee, [18:25:31] * jvandavier phew [18:25:43] i mean, not great, but i was very concerned for a second [18:27:02] I got the fundraiser only once earlier today [18:28:38] but getting the FDC at least four times on several projects is really annoying [18:29:15] seems like a bug to me [18:29:58] (PS1) Adamw: Look in /etc and other reasonable paths for config [wikimedia/fundraising/tools] - https://gerrit.wikimedia.org/r/90577 [18:30:19] Jeff_Green: ^^ [18:33:01] looks ok to me but I'm not a python wiz [18:33:38] Jeff_Green: or do u prefer a "-f/--config=" argument? [18:34:41] might be nice to have that as an option, having a default location in /etc seems good to me [18:36:03] (PS2) Adamw: Look in /etc and other reasonable paths for config [wikimedia/fundraising/tools] - https://gerrit.wikimedia.org/r/90577 [18:36:13] (CR) Adamw: "Tested. Merged with tacit nod from Jgreen." [wikimedia/fundraising/tools] - https://gerrit.wikimedia.org/r/90577 (owner: Adamw) [18:36:22] (CR) Adamw: [C: 2 V: 2] "Tested. Merged with tacit nod from Jgreen." [wikimedia/fundraising/tools] - https://gerrit.wikimedia.org/r/90577 (owner: Adamw) [18:36:23] Romaine: i'm asking around to see if that was intentional or not [18:36:59] gerrit needs a tacit nod button... [18:37:14] :) [18:37:33] Please talk to my consigliere, Michael [18:37:51] (+2) [18:38:22] i had to look that up: http://upload.wikimedia.org/wikipedia/commons/1/1c/Mafia_family_structure_tree.en.svg [18:38:50] that... is exactly it [18:39:05] * awight drifts off into articles using this images [18:39:26] jesus, it's latek or smthing [18:45:05] Jeff_Green: k, that change is pushed, so config can begin its life as cp /opt/fundraising/tools/dedupe/config.yaml.example /etc/fundraising/dedupe.yaml [18:48:28] (PS1) Adamw: more manual queries [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/90580 [18:50:20] awight: so wmf_civicrm vs civicrm? [18:50:28] (CR) jenkins-bot: [V: -1] more manual queries [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/90580 (owner: Adamw) [18:50:44] should I s/wmf_//g the file? [18:50:52] uh [18:50:55] oh yes [18:51:02] and with the new db creds obviously [18:51:26] but the schema lines too, right? [18:51:44] yeah, sorry for the pointless duplication [18:51:48] ha [18:52:18] * Jeff_Green tries to remember the service hostname we created for the master db... [18:52:52] ah ha [18:59:44] (Abandoned) Mwalker: Ugh, this bug was fixed in 2008! [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/45509 (owner: Mwalker) [19:03:04] (Abandoned) Mwalker: Producing files for consumption by Limn [wikimedia/fundraising/tools] - https://gerrit.wikimedia.org/r/51973 (owner: Mwalker) [19:07:07] btw, it would be nice to have g+rx on /etc/fundraising, for debugging [19:27:02] (PS1) Mwalker: Sane .gitignore for csv2sql [wikimedia/fundraising/tools] - https://gerrit.wikimedia.org/r/90630 [19:27:26] (PS2) Mwalker: Sane .gitignore for csv2sql [wikimedia/fundraising/tools] - https://gerrit.wikimedia.org/r/90630 [19:27:42] (CR) Mwalker: [C: 2] Sane .gitignore for csv2sql [wikimedia/fundraising/tools] - https://gerrit.wikimedia.org/r/90630 (owner: Mwalker) [19:28:57] mwalker: Did I not give you the link to the upstream bug I filed about that Zuul issue? [19:29:16] it's actually in the commit message of the patch I linked to [19:29:22] i meant to add it to the email; but I forgot [19:29:30] Oh cool [19:29:34] * marktraceur just making sure [19:29:48] yep; you were intelligent in that regard -- I would have long lost the link otherwise [19:54:47] (PS1) Mwalker: Serve Mobile URLs to Mobile Users [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/90639 [19:57:01] (PS2) Mwalker: Serve Mobile URLs to Mobile Users [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/90639 [19:59:41] (PS1) Adamw: Optimize tagging by performing bulk insert [wikimedia/fundraising/tools] - https://gerrit.wikimedia.org/r/90640 [20:00:02] (PS6) Hashar: Dummy commit [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/87457 (owner: Mwalker) [20:00:08] (CR) jenkins-bot: [V: -1] Dummy commit [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/87457 (owner: Mwalker) [20:02:23] mwalker: you have a minute for CR? ^^ [20:02:37] (PS7) Hashar: Dummy commit [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/87457 (owner: Mwalker) [20:02:56] awight: sure; one moment [20:02:57] (CR) jenkins-bot: [V: -1] Dummy commit [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/87457 (owner: Mwalker) [20:03:21] mwalker: thanks, no rush [20:03:46] how long is this going operation going to take? [20:03:57] I'm worried that with our DB mode you'll be eating locks [20:03:59] mwalker: which op? [20:04:09] oh. there is a 1.5 minute select [20:04:15] which i don't believe eats a lock [20:04:25] the bulk insert should be pretty much instantaneous [20:04:50] it will be 10,000 inserts on the entity_tag table [20:04:58] so all the processing between start of transaction to commit will be fast? [20:05:23] the default mode right now is REPEATABLE_READ which is crazy strict about locking [20:05:35] Jeff_Green: you know anything about this? ^^ [20:05:42] mwalker: there is no transaction :( [20:06:10] if there were, it would be up to 15 min. I'm trying to get actual timing now. [20:06:12] I don't [20:06:29] this is a question for Mr. DBA [20:07:10] yeah i think it's fine at the moment, cos of not wrapping in a transaction. [20:07:11] heh; so; basically; because we use STATEMENT logging, MySQL defaults to using REPEATABLE_READ locking so that changes are atomic and ordered [20:08:07] but ok; if you;re not using transactions; why do you need the explicit commit on line 68 of quick_autoreview.py? [20:09:24] (CR) Adamw: [C: 2 V: 2] "Nice. And I assume there are no references in the PHP which should also be fixed?" [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/90639 (owner: Mwalker) [20:09:52] mwalker: err that is a function named "commit" :) [20:10:14] which doesn't do a DB commit? crazy [20:10:15] but ok [20:10:52] hehe, neither do my functions named "begin" and "end" [20:13:49] (CR) Mwalker: [C: 2] Optimize tagging by performing bulk insert [wikimedia/fundraising/tools] - https://gerrit.wikimedia.org/r/90640 (owner: Adamw) [20:14:21] anyways; regarding the whole repeatable read thing; it shouldn't actually matter unless you're on the master [20:14:36] but I ran into that problem when I was writing the export scripts for silverpop [20:14:59] i am on master [20:15:01] hmm [20:15:12] ok really good thing to fix in the future [20:15:25] it would be correct to wrap all this in a transaction [20:16:00] (PS8) Hashar: Dummy commit [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/87457 (owner: Mwalker) [20:16:01] (CR) Mwalker: [C: 2] Found a bug with priority-based banner hiding [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/90475 (owner: Adamw) [20:16:03] (CR) jenkins-bot: [V: -1] Dummy commit [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/87457 (owner: Mwalker) [20:19:21] (PS1) Adamw: quick_autoreview.py can be called from the shell [wikimedia/fundraising/tools] - https://gerrit.wikimedia.org/r/90644 [20:19:44] Jeff_Green: ^^ [20:20:49] what's the /usr/bin/env thing do? let you control by user which python runs? [20:21:39] it wipes out all env vars, i tink [20:21:44] stupid h key [20:21:52] oic [20:21:53] new ersey [20:22:12] (CR) Jgreen: [C: 1] quick_autoreview.py can be called from the shell [wikimedia/fundraising/tools] - https://gerrit.wikimedia.org/r/90644 (owner: Adamw) [20:22:25] there, I >tacit approved [20:22:44] not gonna get this guy to sign anything fishy... [20:22:54] (CR) Adamw: [C: 2 V: 2] quick_autoreview.py can be called from the shell [wikimedia/fundraising/tools] - https://gerrit.wikimedia.org/r/90644 (owner: Adamw) [20:23:18] I mean, i'm just an associate anyway... [20:23:25] ;) [20:24:27] running [20:29:40] (PS9) Hashar: Dummy commit [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/87457 (owner: Mwalker) [20:32:18] Jeff_Green: oh here is a good explanation of the /usr/bin/env thing: [20:32:30] https://en.wikipedia.org/wiki/Shebang_(Unix) - This mostly works because the path /usr/bin/env is commonly used for the env utility, and it invokes the first sh found in the user's $PATH, typically /bin/sh, if the user's path is correctly configured. [20:33:49] awight: anything that mostly works ;-) [20:34:17] Jeff_Green: sorry my hora de partir has arrived [20:34:51] alright. I'll fight a bit more with sudo and punt in 20 anyway too. have a good weekend! [20:35:03] if I accidentally start it, can I just ctrl-c ? [20:35:07] yep [20:35:09] k [20:35:17] thanks, i think you have it working hands-free anyway! [20:37:51] (PS2) Mwalker: more manual queries [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/90580 (owner: Adamw) [20:40:31] (CR) Mwalker: [C: 2] "horrifying; but ok" [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/90580 (owner: Adamw) [23:09:20] mwalker: i'm on 6 if u wanna chat about anything... [23:09:35] why the heck did you come in? [23:09:37] crazy person [23:09:55] but no; I'm good; I'm reading about how to write C extensions to Node.JS [23:10:25] baaahaha WTF [23:10:33] now you've gone too far [23:10:35] you might also find this interesting -- I think it might explain some of our problem https://developers.google.com/speed/articles/use-compression?csw=1 [23:10:53] I need to do the GeoIP lookup in Node and not on the varnish server -- which means I have to get GeoIP working in Node :p [23:11:06] initial reading indicates it shouldn [23:11:08] t be too bad [23:11:37] specifically on that page "The data from Google's web search logs show that the average page load time for users getting uncompressed content is 25% higher compared to the time for users getting compressed content. In a randomized experiment where we forced compression for some users who would otherwise not get compressed content, we measured a latency improvement of 300ms. " [23:12:04] oh damn [23:12:06] not that line [23:12:59] basically -- there are a lot of users out there with browsers that could accept compressed content; but don't because of intermediate proxies/software/bugs/pineapples [23:13:22] better page :http://www.stevesouders.com/blog/2009/11/11/whos-not-getting-gzip/ [23:13:58] yeah that first link did not say how many people this affects [23:14:49] so; in 2009; about 15% of users -- I've a patch in to udp2log that will tell us if it's still an issue [23:14:51] so, <15% of people seeing 25% speed decrease does not explain much... [23:15:01] but I'm curious, cool that you're pursuing [23:15:11] sure it does -- if they're the demographic that donates to us :) [23:15:24] hey. did u see jeff's directive, that we have a QA server? [23:15:29] holy cow [23:15:40] uh; yep; but still not sure what we can do with it [23:16:16] i'm running the deduper, for example [23:16:28] I think we can just hammer it with whatever we want. [23:16:43] eventually, we can get full integration tests of the pipeline.