[00:03:43] cwd config is definitely still persisting :S [00:03:55] and with it the jobQueue [00:04:15] awight: looks like that puts it at 5pm for you - is that too late? [00:05:41] ejegg: and this is even before the create tables [00:05:59] somehow your blanking the config fixed that error...i don't understand [00:06:50] me neither. there's a bunch of lingering statics and I need to track 'em down [00:08:49] eileen: That works! [00:08:55] cool [00:09:45] ejegg: you think it has to do with the Configuration class? [00:09:46] cwd hmm, the configuration constructor isn't being called for each setUp in the paypal test [00:09:59] maybe specifically the Paypal testing config? [00:10:43] hmmm [00:12:22] (PS10) Eileen: [deploy with tools] Update preferred language based on contribution_tracking. [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/310219 (https://phabricator.wikimedia.org/T96410) [00:12:37] (CR) Eileen: "Dang I thought I had killed those implicit joins - gone now" (4 comments) [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/310219 (https://phabricator.wikimedia.org/T96410) (owner: Eileen) [00:12:57] ejegg: I just fixed those implicit joins [00:17:48] eileen: cool, looking again! [00:17:54] thanks [00:18:09] cwd never mind about the lingering config, I was confused by that foreach loop [00:18:53] cwd also I noticed your $message_locations has a duplicate 'recurring' key [00:19:03] ohhh, but that's actually OK [00:19:06] Just looking at dedupe efforts on recent imports and I'm seeing a tonne coming in right now where the name is throwing conflicts because one is in English & another in Japanese [00:19:09] freaking php [00:19:39] cwd anyway, I run phpunit locally and the pop throws an exception like it should [00:19:44] set, vector, map, same thing right? [00:19:54] * awight abuses airplane lavatory [00:22:24] (CR) Ejegg: "column is just 'id' not 'contact_id', right?" (1 comment) [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/310219 (https://phabricator.wikimedia.org/T96410) (owner: Eileen) [00:23:37] ejegg: arg I copied & pasted - fixing! [00:25:03] ejegg: oh really? [00:25:08] * cwd tries again [00:25:08] (PS11) Eileen: [deploy with tools] Update preferred language based on contribution_tracking. [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/310219 (https://phabricator.wikimedia.org/T96410) [00:26:23] cwd oh man, I don't think I even needed to null that out in tearDown - it's always overwritten in the base test case's setUp [00:26:37] which makes it more confusing that your tests are passing [00:27:14] in CI [00:28:00] ejegg: i'm not getting that exception. are you on the latest rebase? [00:28:17] i thought i was. lemme see... [00:32:21] (CR) Ejegg: "converging on a solution... just found one more typo" (1 comment) [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/310219 (https://phabricator.wikimedia.org/T96410) (owner: Eileen) [00:32:24] (CR) Eileen: "Opps! fixed again ejegg" [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/310219 (https://phabricator.wikimedia.org/T96410) (owner: Eileen) [00:34:33] (PS12) Eileen: [deploy with tools] Update preferred language based on contribution_tracking. [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/310219 (https://phabricator.wikimedia.org/T96410) [00:35:08] (CR) Eileen: "Ah yeah - note to self - last minute adds are bad" (1 comment) [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/310219 (https://phabricator.wikimedia.org/T96410) (owner: Eileen) [00:35:35] Fundraising Sprint Rocket Surgery 2016, Fundraising-Backlog, MediaWiki-extensions-CentralNotice, Operations, and 3 others: Banner not showing up on site - https://phabricator.wikimedia.org/T144952#2672632 (AndyRussG) @awight it takes a while for webrequest logs to land into Hive.... Also, regard... [00:37:22] (CR) Ejegg: [C: 2] "Looks like this patch has reached its final form!" [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/310219 (https://phabricator.wikimedia.org/T96410) (owner: Eileen) [00:37:56] cwd ok, re-downloading [00:39:17] cwd yeah, PS4: BadMethodCallException: Call to a member function execute() on a non-object (boolean) [00:39:42] (Merged) jenkins-bot: [deploy with tools] Update preferred language based on contribution_tracking. [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/310219 (https://phabricator.wikimedia.org/T96410) (owner: Eileen) [00:40:29] is CI running with the latest PHPQueue somehow? [00:40:50] ejegg: ok it's merged - I shall try to deploy! [00:41:00] logging in... [00:42:19] ejegg: ah ha [00:42:20] eileen: gogogo! [00:42:23] good thinking [00:42:30] cwd oh it is! [00:42:34] it does composer install doesn't it? [00:42:51] yeah... I dunno. I guess not! [00:43:04] or it wipes out the .lock file [00:43:10] that's ungood [00:43:27] but i'm not running latest php-queue [00:43:27] Yeah, I see right in the jenkins test log it cloned 7834573f5dc2ab8935 [00:43:43] or am i [00:43:47] which is your 'return null' patch [00:44:12] ejegg: just trying to cherry-pick [00:44:19] soanyway, let's just get those tests set up right with the persistent queue object [00:44:26] yeah nm that was it [00:44:35] eileen: sorry, is it really thorny again? [00:44:50] should i undo the change to php-queue? it addresses the same thing [00:45:00] cwd nope, it doesn't [00:45:29] your test isn't looking at the same instance, so the null assertion will always pass [00:45:37] even if the listener DID enqueue something [00:45:38] ejegg: both modified: sites/all/modules/globalcollect_audit/globalcollect_audit.module [00:46:09] eileen: ah, you want to keep the version that writes to 2 queues [00:46:25] ugh, but there's a name change. [00:46:25] ejegg: yeah there are quite a few changes in there [00:46:35] including a conflict on wmf_civicrm.install :-( [00:46:37] wait, why is that coming up with a cherry-pick? [00:46:51] how much are you cherry-picking? [00:47:25] phooey, I have a merge up in review that resolves all sorts of these things, at least up as far as the donation queue consumer switch [00:47:49] ejegg: ah gotcha. should i roll back php-queue? [00:47:51] (PS2) Eileen: Fix previous update query. [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/312951 (https://phabricator.wikimedia.org/T96410) [00:48:09] cwd no, the php-queue thing is fine for making push and pop act the same [00:48:27] ejegg: so I should look at your merge in review or does that need to be someone else? [00:48:30] just gotta fix up the test so you're looking at the same instance in both places [00:48:35] ok [00:48:49] BTW I realise this needs merging https://gerrit.wikimedia.org/r/#/c/312951/ - a fix to my not-yet-deployed upgrade command [00:48:57] eileen: I guess I'll have to re-do it on top of your language cherry-pick [00:49:07] I was really hoping to get that out today [00:49:30] ejegg: is it just reviewing the merge resolution? [00:49:40] (CR) Ejegg: [C: 2] Fix previous update query. [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/312951 (https://phabricator.wikimedia.org/T96410) (owner: Eileen) [00:50:20] eileen: no, it depends on us getting the paypal queue switched over and the donation producers mirroring [00:50:56] ejegg: ah ok [00:50:59] ejegg: do i have to add those persistent flags in the listener code too? [00:51:12] cwd nope, no need there, just for the tests [00:51:31] does it reconnect to the db whenever object is called? [00:51:39] have the test call ->object first with the persistent flag [00:51:59] and then when the listener calls ->object, it will see the instance already exists and use that one [00:52:06] (Merged) jenkins-bot: Fix previous update query. [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/312951 (https://phabricator.wikimedia.org/T96410) (owner: Eileen) [00:52:08] cwd yeah, PDO constructor connects to the db [00:52:21] but what if the listener calls it before the test? [00:52:34] then you see different instances [00:52:41] so just make sure the test gets in there first [00:53:04] so i need to call object from the test for any time when the calling code might call it? [00:53:22] that seems like confusing default behavior [00:53:53] yeah, it's just fixture setup [00:54:53] i'm sorry, i'm confused [00:55:10] testCapture calls execute before object [00:55:22] but when it pops off the queue, the thing is there [00:55:45] (PS1) Eileen: Update English languages to only keep a small number of recognised variants [wikimedia/fundraising/crm] (deployment) - https://gerrit.wikimedia.org/r/313159 (https://phabricator.wikimedia.org/T96410) [00:55:47] (PS1) Eileen: [deploy with tools] Update preferred language based on contribution_tracking. [wikimedia/fundraising/crm] (deployment) - https://gerrit.wikimedia.org/r/313160 (https://phabricator.wikimedia.org/T96410) [00:56:42] (PS1) Eileen: Fix previous update query. [wikimedia/fundraising/crm] (deployment) - https://gerrit.wikimedia.org/r/313161 (https://phabricator.wikimedia.org/T96410) [00:56:57] ejegg: like this https://github.com/wikimedia/wikimedia-fundraising-SmashPig/blob/master/PaymentProviders/PayPal/Tests/phpunit/CaptureIncomingMessageTest.php#L57 [00:57:14] (CR) Eileen: [C: 2] Update English languages to only keep a small number of recognised variants [wikimedia/fundraising/crm] (deployment) - https://gerrit.wikimedia.org/r/313159 (https://phabricator.wikimedia.org/T96410) (owner: Eileen) [00:57:19] sorry to bombard you but i really want to unblock you [00:57:20] (Merged) jenkins-bot: Update English languages to only keep a small number of recognised variants [wikimedia/fundraising/crm] (deployment) - https://gerrit.wikimedia.org/r/313159 (https://phabricator.wikimedia.org/T96410) (owner: Eileen) [00:57:22] oh huh [00:57:35] (CR) Eileen: "Cherry-pick to deployment - self merging" [wikimedia/fundraising/crm] (deployment) - https://gerrit.wikimedia.org/r/313160 (https://phabricator.wikimedia.org/T96410) (owner: Eileen) [00:58:00] (CR) Eileen: [C: 2] [deploy with tools] Update preferred language based on contribution_tracking. [wikimedia/fundraising/crm] (deployment) - https://gerrit.wikimedia.org/r/313160 (https://phabricator.wikimedia.org/T96410) (owner: Eileen) [00:58:24] (Merged) jenkins-bot: [deploy with tools] Update preferred language based on contribution_tracking. [wikimedia/fundraising/crm] (deployment) - https://gerrit.wikimedia.org/r/313160 (https://phabricator.wikimedia.org/T96410) (owner: Eileen) [00:58:28] (CR) Eileen: [C: 2] "Cherry-pick to deployment - self merging" [wikimedia/fundraising/crm] (deployment) - https://gerrit.wikimedia.org/r/313161 (https://phabricator.wikimedia.org/T96410) (owner: Eileen) [00:58:35] (Merged) jenkins-bot: Fix previous update query. [wikimedia/fundraising/crm] (deployment) - https://gerrit.wikimedia.org/r/313161 (https://phabricator.wikimedia.org/T96410) (owner: Eileen) [01:00:26] cwd weird, maybe I've been wrong this whole time [01:00:31] ejegg: ccccccebfefrtkfttdfhbckrnnikniiefrdfhjffblut [01:00:33] opps [01:01:35] just wondering if I should take a back up like CREATE TABLE backup_pre_language_update_7260 SELECT id, preferred_language FROM civicrm_contact; [01:01:58] could it be that the sqlite persists during the php session? [01:04:04] eileen would the log tables not be good enough? [01:04:47] ejegg: yeah they should really - it would be a little harder to mass revert - but probably not that hard since the connection_id would be the same for all [01:04:48] cwd yikes, and all of our queues and dbs are sharing one in-memory db across all tests? [01:05:01] !log civicrm upgrade from d30a5e454e6614214faef8354a3a23d71d573c7f to 26e5214e66f2e4b5de89c27dd5121dfde43269e2 [01:05:06] Logged the message at https://wikitech.wikimedia.org/wiki/Server_Admin_Log, Master [01:05:06] cwd I hope that's not happening [01:05:25] well it shells out a new php for each test [01:05:36] so i don't think it is probably that bad [01:05:54] wait, really? But we have to clear out static vars or they stick around for the next test [01:06:15] oh...huh [01:06:28] i certainly thought that's what it was doing [01:06:34] dang, I think I gotta take off for the evening [01:06:38] sorry, I'll puzzle over this more later [01:06:48] so long! [01:07:00] yeah me too, have a good one! [01:07:09] OK - I've disabled debug & donations & thank you mail - lets try that update [01:09:28] (PS1) Eileen: Add CiviCRM initialize to upgrade function. [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/313163 [01:10:09] Is anyone still about - I need to get this one line merged https://gerrit.wikimedia.org/r/#/c/313163/1/sites/all/modules/wmf_civicrm/wmf_civicrm.install [01:10:25] I think it's probably trivial enough I could self-review but would prefer not to [01:10:48] XenoRyet: still here? [01:13:18] awight: are you still around? [01:15:16] Fundraising Sprint Rocket Surgery 2016, Fundraising-Backlog, MediaWiki-extensions-CentralNotice, Operations, and 3 others: Banner not showing up on site - https://phabricator.wikimedia.org/T144952#2672707 (awight) > @awight it takes a while for webrequest logs to land into Hive.... Yes, come to... [01:15:23] OK - I'm gonna self merge this on the basis of triviality since the main commit has merged [01:15:30] eileen: yes, hi! [01:15:34] (CR) Eileen: [C: 2] Add CiviCRM initialize to upgrade function. [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/313163 (owner: Eileen) [01:15:36] just diving into Hive data [01:15:53] awight: ah well I self- merged - but so trivial I feel OK [01:15:57] https://gerrit.wikimedia.org/r/#/c/313163/1/sites/all/modules/wmf_civicrm/wmf_civicrm.install [01:16:01] (CR) Awight: [C: 2] "Posthumous approval!" [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/313163 (owner: Eileen) [01:16:17] :-) [01:16:21] thanks awight [01:16:34] hehe that civicrm_initialize thing gets me nearly every time, too [01:17:16] yep - I think the upgrade script could still fail on a trigger thing that I just thought of - but if it does there is no real impact except it won't happen today [01:17:55] (Merged) jenkins-bot: Add CiviCRM initialize to upgrade function. [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/313163 (owner: Eileen) [01:18:16] Does drupal have anything like Civi's "interrupted update" mechanism? [01:18:30] (PS1) Eileen: Add CiviCRM initialize to upgrade function. [wikimedia/fundraising/crm] (deployment) - https://gerrit.wikimedia.org/r/313164 [01:18:38] I seem to remember needing to update civicrm_domain.version manually... [01:19:13] (CR) Eileen: [C: 2] "merge to deploy" [wikimedia/fundraising/crm] (deployment) - https://gerrit.wikimedia.org/r/313164 (owner: Eileen) [01:19:20] (Merged) jenkins-bot: Add CiviCRM initialize to upgrade function. [wikimedia/fundraising/crm] (deployment) - https://gerrit.wikimedia.org/r/313164 (owner: Eileen) [01:19:28] the system.schema_version won't update if the sql fails [01:19:37] so it will just run again next time [01:20:38] !log updating CiviCRM from 26e5214e66f2e4b5de89c27dd5121dfde43269e2 to 88000acc826592e535491bde6a47f8976deaf07a [01:20:43] Logged the message at https://wikitech.wikimedia.org/wiki/Server_Admin_Log, Master [01:21:06] well it's running now - it's take a while to build the temp tables so we'll see [01:21:27] actually they are not temp tables - I'll have to drop them [01:22:12] btw awight I think the dedupe jobs fall over less if we run the temp table cleanup job occassionally [01:34:22] eileen: Fine by me--is there any advantage in not clearing them at the end of each run? [01:34:29] or um beginning of the run? [01:35:26] OMG, Hive can't parse \t in queries. That doesn't mean good things about their SQL parser. [01:35:37] awight: it actually is a bit too aggressive [01:36:00] I think it might clear out search results & possibly some in progress dedupe finds [01:36:09] the latter don't matter in the scheme of things [01:37:20] I forget when we acquired such a job--does it just run a Civi/drush command, or is it a custom script? [01:40:45] it's just an api call - so can be enabled in civicrm sched jjob list or scheduled using drush [01:42:24] but by the same token, something we probably don't want to hack? [01:44:13] well I just looked & the expiry_date field doesn't seem to be being used on the civicrm-cache table as I think it probably was intended - but I should test that [01:44:30] also I was looking into those global collect charges [01:44:35] Ooh [01:44:55] I have a query for finding anything dodgey in the next scheduled_date field [01:45:22] yess, thanks [01:45:31] I'm inclined to either a) run that query as a job with fail_mail if it returns anything or b) run the query before the gc charges run & exit the job if not empty [01:45:36] Does it work as soon as September ends? [01:45:59] all of the above :) [01:46:15] well the query basically looks at the gap between the last contribution & the next scheduled one [01:46:37] & says if it is not one month +- 2 days then it looks wrong [01:46:56] ah hrm, remind me, when the bug happened people got charged once per day for two days? [01:47:01] yes [01:47:42] so we'd have to run this safety check in-between the last job to make charges on the 30th and the first job of the 31st, I guess? [01:48:08] well I was thinking it might just be part of the normal gc job [01:48:16] ie. don't proceed if this is not clean [01:48:50] I'll write in more detail on the ticket [01:49:34] Yeah that sounds like a great solution--exit(1)ing the job will trigger a failmail [01:50:21] We can eventually add more assertions there to prevent other regressions... [01:50:42] right - it's not a full fix - just a safety mechanism. I've looked into the cause too [01:57:01] Fundraising Sprint Rocket Surgery 2016, Fundraising-Backlog, MediaWiki-extensions-CentralNotice, Operations, and 3 others: Banner not showing up on site - https://phabricator.wikimedia.org/T144952#2672822 (awight) I'm seeing something even stranger when I sort by timestamp. There are all 87-byte... [02:01:13] Fundraising Sprint Rocket Surgery 2016, Fundraising-Backlog, MediaWiki-extensions-CentralNotice, Operations, and 3 others: Banner not showing up on site - https://phabricator.wikimedia.org/T144952#2672851 (awight) I'm still keeping the priority down to High, now that I've convinced myself that th... [02:01:31] Cache bugs are not fun. The CentralNotice thing seems to be something along those lines, too. [02:06:10] awight: there is a patch to deploy in tools [02:06:18] Flesh out required modules for all projects in this repo [02:06:24] Assume that is OK to go out? [02:06:33] oops--yeah that's safe [02:06:50] we don't run pip install -r on production... [02:12:17] ? [02:12:32] I don't know about that - is there a special way to deploy tools? [02:13:11] (PS1) Eileen: Merge branch 'master' of https://gerrit.wikimedia.org/r/wikimedia/fundraising/tools into deploy [wikimedia/fundraising/tools] (deploy) - https://gerrit.wikimedia.org/r/313169 [02:13:38] eileen: no, it's the same fundraising_code_update + rsync_blaster world [02:14:19] ok cool [02:18:04] (CR) Eileen: [C: 2] "Merge to deploy - self-approving" [wikimedia/fundraising/tools] (deploy) - https://gerrit.wikimedia.org/r/313169 (owner: Eileen) [02:18:10] (Merged) jenkins-bot: Merge branch 'master' of https://gerrit.wikimedia.org/r/wikimedia/fundraising/tools into deploy [wikimedia/fundraising/tools] (deploy) - https://gerrit.wikimedia.org/r/313169 (owner: Eileen) [02:20:30] (PS8) Eileen: Clean up legacy junk data from preferred_language field. [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/310218 (https://phabricator.wikimedia.org/T96410) [02:20:42] (CR) jenkins-bot: [V: -1] Clean up legacy junk data from preferred_language field. [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/310218 (https://phabricator.wikimedia.org/T96410) (owner: Eileen) [02:24:07] (CR) Awight: "recheck" [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/310218 (https://phabricator.wikimedia.org/T96410) (owner: Eileen) [02:24:47] Strange CI glitch... 00:00:02.029 warning: failed to remove sites/default/settings.php' [02:26:52] odd [02:27:32] Did I not comment with the right magic btw? Zuul should be picking that up and rebuilding... [02:27:44] oh ha V+2 [02:27:55] An unexpected success ;) [02:28:27] INFO : 2016-09-28 02:26:52,354 Stage-1 map = 45%, reduce = 15%, Cumulative CPU 284411.26 sec [02:28:35] * awight cringes at carbon footprint [02:48:15] awight: it sure seems like the sqlite store persists through the session [02:48:23] does this prove that? https://github.com/wikimedia/wikimedia-fundraising-SmashPig/blob/master/PaymentProviders/PayPal/Tests/phpunit/CaptureIncomingMessageTest.php#L57 [02:49:14] that test passes even though it sets the message in Listener&execute() [02:58:19] Aaargh [02:58:19] INFO : 2016-09-28 02:43:33,049 Stage-1 map = 100%, reduce = 100%, Cumulative CPU 540632.94 sec [02:58:27] ... org.apache.thrift.TException: Error in calling method FetchResults [02:58:31] Caused by: java.lang.OutOfMemoryError: Java heap space [02:59:20] How can a big data program possibly fail to allocate itself memory or work in appropriately sized chunks. And after finishing an extremely expensive computation... [03:02:24] cwd: I don't get it [03:02:44] Why should the test fail? [03:03:33] PHPQueue\PDO::push attempts to create the queue table if it doesn't exist--or you're beyond the table thing? [03:03:45] awight: listener&execute sets the message using config->object [03:03:53] and the test finds it using config->object [03:04:02] shouldn't the 2nd be an empty db? [03:04:22] you call listener.execute, right? [03:04:32] in test::capture [03:04:38] yeah [03:04:39] so I don't get why the test would not find it [03:05:00] if every call to config->object returns a new in memory db [03:05:02] what's the change-id again? [03:05:22] oh I'm starting to see [03:06:24] * cwd suspiciously eyes the & before the object() function signature [03:07:43] awight: no specific change id, this test has been passing for awhile [03:09:14] cwd: hey http://php.net/manual/en/ref.pdo-sqlite.connection.php#105350 [03:09:24] Undocumented flag... [03:09:50] We're shooting for something more granular than per-quasi-request, though [03:10:54] Nasty--so part of our base test tearDown should be to delete all tables or otherwise purge that bugger? [03:11:18] & I agree that the PHP "reference" could also be to blame [03:12:34] horror: ".tables" [03:13:03] awight: well i mean more persistence within a single php thread. i thought you were saying there is no such thing, like every call to config->object returns a new in memory db, so we need to memoize it using the configuration class [03:13:29] I did think that was necessary [03:13:45] (and still might be for other data stores...) [03:13:58] this person thinks so too: http://stackoverflow.com/questions/18814644/pdo-sqlitememory-sharing-and-lifetime [03:14:02] yeah I think we need to tearDown all the tables [03:14:07] but that is not the behavior i am observing [03:14:47] if it was how would that test pass? the db should be blank here https://github.com/wikimedia/wikimedia-fundraising-SmashPig/blob/master/PaymentProviders/PayPal/Tests/phpunit/CaptureIncomingMessageTest.php#L60 [03:14:57] While we're looking at it... Seems like the pass-by-reference and "persistent" param are incredibly misleading and we should remove if there's nothing depending on it. [03:15:10] cwd: I think we're agreeding! [03:15:13] hehe [03:15:44] yes i think so too :) [03:16:27] but i'm not sure if something is depending on it... [03:16:43] OK just reading the SO comment, yeah the 2013 answer could be bogus [03:17:38] (PS3) Eileen: Update backoffice message template to allow receipt resends. [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/312952 (https://phabricator.wikimedia.org/T137496) [03:17:40] (PS5) Eileen: Remove handling for bad data (language = '_' as we are cleaning that up [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/310724 (https://phabricator.wikimedia.org/T96410) [03:17:51] I need to walk my poor dog, catch you tomorrow! [03:34:14] !log tools upgrade from b0be0f9ca04191c4bab869bb81191c5c77c432ca to 5427a601ed64ecfe01ef84eda13f28c0dd08a3af [03:34:19] Logged the message at https://wikitech.wikimedia.org/wiki/Server_Admin_Log, Master [03:36:01] Fundraising Sprint Octopus Untangling, Fundraising Sprint Pretending This Isn't Happening, Fundraising Sprint Qwerty Thwacking, Fundraising Sprint Rocket Surgery 2016, and 3 others: Silverpop export not capturing language changes in Civi (not sure if... - https://phabricator.wikimedia.org/T96410#2673139 [03:37:20] Fundraising Sprint Octopus Untangling, Fundraising Sprint Pretending This Isn't Happening, Fundraising Sprint Qwerty Thwacking, Fundraising Sprint Rocket Surgery 2016, and 3 others: Silverpop export not capturing language changes in Civi (not sure if... - https://phabricator.wikimedia.org/T96410#2673144 [04:10:41] Fundraising Sprint Rocket Surgery 2016, Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM: Longterm fix + regression test for T144489 - https://phabricator.wikimedia.org/T144557#2673161 (Eileenmcnaughton) I tried the query below on live and locally and established that it successfully identifies rec... [09:33:11] Fundraising-Backlog, MediaWiki-extensions-CentralNotice: Fatal exception when cloning/saving banners - https://phabricator.wikimedia.org/T146880#2673576 (Pcoombe) [09:35:02] Fundraising-Backlog, MediaWiki-extensions-CentralNotice: Fatal exception when cloning/saving banners - https://phabricator.wikimedia.org/T146880#2673589 (Pcoombe) p:Triage>High [10:06:31] Fundraising-Analysis, Fundraising-Backlog, Analytics, MediaWiki-extensions-CentralNotice: Provide performant query access to banner show/hide numbers - https://phabricator.wikimedia.org/T90649#1064105 (Milimetric) +1 to the streaming way. We have to get there eventually and this task is already... [15:51:27] hi awight [15:58:05] bonjour! [15:59:42] i was thinking about the sqlite::memory thing. if the temp db didn't persist in memory what advantage would it have over a stdClass? [16:09:38] Hi from Mexico once again, fr-tech! [16:11:01] Thx 4 being there and working on CN bugs while I was sick, traveling and kid-caring-while-spouse-was-at-funeral... [16:11:35] wb AndyRussG [16:11:41] sounds like a harrowing adventure [16:12:27] Heh thx it had some harrowing bits and some good bits [16:13:00] hey all — I’m trying to use civi’s import template and am getting an error, but not sure what the error is! can anyone help me diagnose the problem? :) [16:14:36] cwd: you could do a quick test, create two sqlite::memory stores and put something into one hat, see if a rabbit comes out of the other. [16:14:56] Either way, I think the punchline is that we need to purge the db during tearDown [16:15:03] Good find! [16:15:17] ccogdill: Sure thing! Lemme browse to the Civi logs [16:15:25] I ran it yesterday afternoon [16:15:29] let me run again really quick [16:15:31] thank you! [16:15:45] k done awight [16:15:46] ccogdill: it's fine, the logs will still be there [16:15:48] ah ok [16:17:04] huh. "error in line 2" [16:17:16] That's... less info than we need :) [16:17:19] haha yeah, that’s the whole row! I only tried 1 row of data as a test [16:17:21] :p [16:17:56] maybe worth noting that I tried the "Original Amount" line as both "30" and "USD 30" [16:18:03] Leanne thought that might have been the problem [16:18:22] There really should have been more debugging... [16:18:43] I think there is a phab task about this error messaging [16:19:41] Yeah, that would be a good place to comment about this--T88460 [16:19:55] sure I can throw a comment in there [16:21:21] ack, I'm looking at the source code and that's in fact all we log. [16:21:39] ugh [16:22:07] any chance I could screenshare and show you a row of data? or maybe is there a way to export a previously imported row that was correct? [16:23:04] I do see the row and could import it locally, but I have a quick fix for the logging... [16:23:24] well I need to import 700 rows of data [16:23:26] this was just one [16:23:34] should I try to import the whole file? [16:23:39] (PS1) Awight: Log the exception message for failed import lines [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/313218 (https://phabricator.wikimedia.org/T88460) [16:23:50] I’ll have to do this one more time… [16:25:07] ccogdill: It should be harmless to attempt the whole file, but we have that glitch where you might see 700 errors... Maybe try two rows, to see if it's a formatting issue? [16:25:14] And is this file on the fileserver somewhere? [16:25:46] awight: any thoughts on this? https://gerrit.wikimedia.org/r/#/c/313124/ it makes pop return null instead of erroring when the table is missing. you had mentioned doing https://gerrit.wikimedia.org/r/#/c/313127/ instead but ejegg said the former was fine (either way fixes the CI error) [16:25:54] fr-tech: quick CR? //gerrit.wikimedia.org/r/313218 [16:26:07] argh, debian-testing. https://gerrit.wikimedia.org/r/313218 [16:26:19] cwd: Sure lemme see... [16:26:34] (CR) Cdentinger: [C: 2] Log the exception message for failed import lines [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/313218 (https://phabricator.wikimedia.org/T88460) (owner: Awight) [16:27:07] I’ll add to the fileserver now and try the import [16:27:10] cwd: great--feel free to self-merge the library updates [16:27:22] awight: you aren't worried about the incidentals? [16:27:34] eh? The other changes to php-queue? [16:27:37] yeah [16:27:48] reading... [16:27:57] not sure why they were sitting undeployed [16:28:25] I'm really surprised about popAtomic. [16:29:18] (Merged) jenkins-bot: Log the exception message for failed import lines [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/313218 (https://phabricator.wikimedia.org/T88460) (owner: Awight) [16:29:18] That gets used all over the place, how were we not crashing and burning? [16:29:57] good question [16:30:36] it's unnerving that CI composer installs sometimes, cause the new code can make it up before the library updates [16:31:00] ohhhh awight! the error message got much better when I ran more rows [16:31:11] INVALID_MESSAGE No payment instrument specified [16:31:22] wat. [16:31:31] this is a new payment method for us — does it need to be added somewhere in the backed before I can import it? [16:31:31] I'm so crept out right now [16:31:41] ccogdill: for sure [16:31:46] well there you have it [16:31:52] okay should I create a task for that? [16:32:18] haha okay I see--the "repeated error" message actually did dump the exact error. [16:32:43] With that patch cwd just approved, we should also get the error message for single rows. [16:33:09] oh cool! [16:35:29] (PS1) Awight: Log the exception message for failed import lines [wikimedia/fundraising/crm] (deployment) - https://gerrit.wikimedia.org/r/313219 (https://phabricator.wikimedia.org/T88460) [16:35:47] (CR) Awight: [C: 2] Log the exception message for failed import lines [wikimedia/fundraising/crm] (deployment) - https://gerrit.wikimedia.org/r/313219 (https://phabricator.wikimedia.org/T88460) (owner: Awight) [16:40:09] awight: https://github.com/wikimedia/wikimedia-fundraising-SmashPig-vendor/blob/master/coderkungfu/php-queue/src/PHPQueue/Backend/Predis.php#L205 [16:40:14] yeah i'm not sure what is going on [16:40:26] (Merged) jenkins-bot: Log the exception message for failed import lines [wikimedia/fundraising/crm] (deployment) - https://gerrit.wikimedia.org/r/313219 (https://phabricator.wikimedia.org/T88460) (owner: Awight) [16:41:39] awight: betcha we've been updating the vendor repo but not the php-queue fork [16:42:03] !log update CRM from 88000acc826592e535491bde6a47f8976deaf07a to 6b2bd98fce5006030423bccc4f4b7fd9b5d14821 [16:42:08] Logged the message at https://wikitech.wikimedia.org/wiki/Server_Admin_Log, Master [16:42:19] cwd: That doesn't sound right, cos composer pulls directly from our fork#master [16:42:43] ccogdill: You should be able to get an error for single lines now [16:42:56] (or I broke everything, we'll see :) [16:42:59] I’ll try just for fun :p [16:43:17] oh! it was successful! [16:43:25] mmm I’ll force a failure [16:44:10] haha foiled again [16:44:37] oh weird it imported even when I misspelled the payment instrument [16:44:59] What does the contribution look like? [16:45:02] awight: right but it's possible to update vendor and deploy without touching composer [16:45:16] sorry i know you're popular this morning [16:45:33] cwd: uh, yeah on that note you probably want to start with the SmashPig#master composer.lock update, eh? [16:47:02] huh I guess i’m not finding the data when I search in Civi [16:47:09] maybe it didn’t import.... [16:47:34] oh nevermind. I misread the message. but it does tell me the error! [16:48:09] it just says "successful import!" at the top which is confusing [16:48:13] argh. [16:48:22] OK sorry I need to go for 30 min [16:48:39] awight|afk: it's just dev-master [16:48:39] thanks for the help anyway! [16:49:21] ccogdill: what's your timeline? [16:49:42] awight I’d like to be able to send the last test email for this PSP this week. today or tomorrow preferably [16:49:55] ok. code changes if any will be very simple [16:49:59] okay [16:50:03] I can put it in phab? [16:50:40] sure, for posterity i s'pose [16:51:35] sounds good [16:51:36] thanks! [17:01:48] (PS1) Ejegg: DO NOT MERGE investigate SQLite persistence [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/313224 [17:02:39] (CR) jenkins-bot: [V: -1] DO NOT MERGE investigate SQLite persistence [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/313224 (owner: Ejegg) [17:08:32] (PS2) Ejegg: DO NOT MERGE investigate SQLite persistence [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/313224 [17:09:06] awight|afk: cwd|brb ^^^ OK, so it's our config class that's doing the persisting (even when we don't ask for it), not sqlite or PDO [17:09:23] (CR) jenkins-bot: [V: -1] DO NOT MERGE investigate SQLite persistence [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/313224 (owner: Ejegg) [17:12:26] ejegg: will check that out in a minute--also, I realized we should do a test where we create two PDO sqlite::memory instances directly and see if they are isolated or not [17:12:44] ejegg: hmm? 17:09:23 db1 foo record: Array [17:12:46] 17:09:23 ( [17:12:48] 17:09:23 [bar] => 5 [17:12:50] 17:09:23 [0] => 5 [17:12:52] 17:09:23 ) [17:12:54] ejegg: hehe great, that's whatchu just did [17:17:29] yah, trying to see why the config class is caching instances always [17:18:26] ejegg: but that one doesn't use config [17:18:31] PDO directly [17:18:32] cwd awight|afk oh derp, persistent defaults to true! [17:18:52] cwd PDO directly acts isolated [17:19:04] we get an exception trying to exec the 2nd statement [17:19:17] cause it failed preparing, since the table wasn't in the 2nd db [17:19:27] I just didn [17:19:38] 't realize the default was $persistent=true [17:20:11] so your test /is/ getting the same instance as the listener, even though it asks for it later [17:20:50] ah you are right [17:20:57] that is so weird to me [17:21:09] cwd sorry to waste your time - your initial assumption that you were looking at the same thing was correct! [17:21:17] what is the point of an in memory store that has the same lifetime as a stdClass? [17:21:47] heh but not for the reasons i thought [17:21:49] I guess it CAN be made to persist across objects with that extra param [17:22:14] from the docs it looks like the persistent arg makes it write to a file for the lifetime of the session [17:22:23] but otherwise, just for testing, or doing sql-y things with a dataset that you don't want to do procedurally, I guess [17:22:49] http://php.net/manual/en/ref.pdo-sqlite.connection.php#105350 [17:23:20] yah, that ATTR::PERSISTENT thing [17:23:28] s/::/_/ [17:23:42] but... that's actually not what we want [17:24:01] request and session are annoyingly ambiguous terms here [17:24:05] we want to be able to nuke the config and get a whole new set of objects for our next test [17:24:25] yeah... saying session in PHP implies a much longer life [17:25:14] oh hey, Scrum of Scrums time [17:25:27] AndyRussG: we still need help with the CN bug, yes? [17:26:02] Did you and awight|afk get an idea of which code is the problem? I can ask about ownership during SoS [17:26:24] ejegg: yes, sounds like a great idea [17:26:34] awight|afk: has dones some really significant digging [17:26:58] all of this info is in T144952 ? [17:26:58] T144952: Banner not showing up on site - https://phabricator.wikimedia.org/T144952 [17:27:02] SoS is like nowish right? [17:27:10] AndyRussG: yes [17:28:09] ejegg: cool! yeah that's it. It's pretty clear it's an issue with MessageCache [17:28:19] I mean, yeah that's the bug [17:28:38] ok, cool, I'll ask for MessageCache whizzes [17:28:53] Pretty sure it's backed by ObjectCache [17:30:19] Yeah that's it ObjectCache... $this->wanCache = ObjectCache::getMainWANInstance(); on Line 162 of MessageCache [17:31:10] Might also be some weird interaction w/ TranslationExtension? not sure... [17:31:16] ejegg: ^ thx!! [17:31:59] awight|afk, ejegg: here's what i think about Configuration&object() - 1) persistence should not be optional, it should either be the behavior or not, and 2) it should not be returned as a reference, it should be explicitly changed if persistent [17:33:17] i tend to think it should *not* be persistent and that should be handled by something else, but that would probably break a lot of behavior we rely on [17:50:23] baahaha > 10:18 < ejegg> cwd awight|afk oh derp, persistent defaults to true! [17:50:48] awight: derp is an understatement... [17:51:23] AndyRussG: ejegg: The really freakish thing about the CN bug is that the cache is populated with correct stuff at a certain point, then *exactly* 10 minutes later to the second, all the caches are missing the string again. [17:51:52] awight: hey! whoa that''s an amazing discovery [17:51:55] AndyRussG: / awight SoS ppl suggested Brad Jorsch might be helpful [17:51:56] that 10 minutes clue should cause cache-saavy colleagues to know where the bug is, I'd hope [17:52:00] ejegg: i stared at that function for way too long without that registering [17:52:08] ah cool [17:57:31] AndyRussG: I guess another noteworthy detail, but probably unrelated to MessageCache, is that the banner didn't seem to get any desktop impressions even when set to 10% of enwiki, at emergency priority. [18:01:23] (CR) Ejegg: [C: 2] Explicitly create test tables in sqlite [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/313127 (owner: Cdentinger) [18:01:38] awight: yeah I saw that. Weird [18:02:06] Sure the banner was configured, no odd campaign settings, etc? [18:02:21] I mean config'd for desktop, logged-ins or anons, etc [18:02:21] (CR) Ejegg: "The test is slightly busted, but fixed in the next patch. The Listener code is perfectly fine. I'mma un-WIP this and merge it if nobody st" [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/312941 (owner: Cdentinger) [18:03:13] (PS5) Cdentinger: Just respond 200 to blank requests [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/312941 [18:03:17] I just mean, often I've thought I've found a targeting bug just to realize there was some config detail I'd forgotten... Though that happens more often locally.... [18:03:55] (CR) Ejegg: [C: 2] "Let's get this show on the road!" [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/312941 (owner: Cdentinger) [18:04:36] (Merged) jenkins-bot: Just respond 200 to blank requests [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/312941 (owner: Cdentinger) [18:05:53] (Abandoned) Ejegg: DO NOT MERGE investigate SQLite persistence [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/313224 (owner: Ejegg) [18:06:25] (PS3) Ejegg: Explicitly create test tables in sqlite [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/313127 (owner: Cdentinger) [18:08:56] (PS1) Ejegg: Merge branch 'master' into deployment [wikimedia/fundraising/SmashPig] (deployment) - https://gerrit.wikimedia.org/r/313229 [18:09:09] AndyRussG: oops, missed your messages due to VPN--hehe yeah that was silly. I'd forgotten that we were using a live banner, which did indeed target only mobile [18:09:13] I'll comment in the bug... [18:09:25] (CR) Ejegg: [C: 2] Merge branch 'master' into deployment [wikimedia/fundraising/SmashPig] (deployment) - https://gerrit.wikimedia.org/r/313229 (owner: Ejegg) [18:09:37] (Merged) jenkins-bot: Merge branch 'master' into deployment [wikimedia/fundraising/SmashPig] (deployment) - https://gerrit.wikimedia.org/r/313229 (owner: Ejegg) [18:10:39] awight: ah OK glad it checked out! (totally sure now?) thx [18:10:43] awight: fr-tech anytime u'd like to exploding CN stuff, I'm quite happy to do so [18:13:35] (PS2) Ejegg: Get all queues with QueueFactory::getQueue [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/313117 [18:13:37] (PS2) Ejegg: WIP use redis for Adyen jobs [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/313130 [18:14:38] (PS10) Ejegg: Move Unsubscribe Queue Consumer off ActiveMQ [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/312084 (https://phabricator.wikimedia.org/T145419) (owner: XenoRyet) [18:16:33] XenoRyet: want to help with a couple deploys? [18:16:51] Sure. What've we got? [18:17:19] SmashPig, for starters [18:17:28] then the old PaymentListeners [18:17:39] Ok cool, I've not done a SmashPig deploy before. [18:17:50] Cool. Pretty much like the rest [18:18:12] I've just merged master->deploy, so the rest is all up on deploy servers [18:18:42] want to drive this one? I'll watch a screenshare if you want confirmation on anything [18:19:16] Yea, let me get setup a little bit. Then we'll jump on the tech-talk link? [18:19:26] sure, ping me when you're all set [18:22:32] ejegg, alright jumping on the call. [18:23:04] Or at least I am if hangouts will let me. [18:23:05] XenoRyet: be right there... [18:24:28] k, i'm in [18:41:24] (PS1) Awight: Provide a workaround for confusing payment instrument/method mapping [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/313234 (https://phabricator.wikimedia.org/T146922) [18:42:15] ccogdill: OK that should fix the issue ^ -- change your column name to "Raw Payment Instrument"... [18:42:35] okay! [18:42:42] will try now... [18:42:53] ccogdill: uh one minute, please [18:42:58] oh oaky [18:43:06] *okay. day of typos, truly... [18:44:05] ccogdill: you can retry the import on staging civi, at your convenience [18:48:54] awight "Checks import complete. 788 imported, 0 ignored, not including 0 duplicates and 0 errors." [18:48:55] !!! [18:49:34] right on. [18:49:46] fr-tech: a little UBN review when you can, //gerrit.wikimedia.org/r/313234 [18:49:49] * awight dies [18:49:54] https://gerrit.wikimedia.org/r/313234 [18:50:17] thank you! so you’ll let me know when I can run it on the live version? [18:50:33] (omg gnome http://unix.stackexchange.com/questions/290544/double-click-selection-in-gnome-terminal ) [18:50:42] ccogdill: will do! [19:15:40] awight: what's with the wall of smashpig failmail in the last hour? Anything to worry about? [19:16:31] oh, hey. That's a lot of failmail. ejegg and I just did a smashpig deploy, but it looked like it went well. [19:16:57] Yeah that seems to be an active leak in the hull [19:17:34] We hadn't gotten to the part where he showed me how to revert if we needed to. [19:17:54] XenoRyet: fundraising_code_update -p SmashPig=rollback [19:18:00] then rsync_blaster it out [19:18:05] k, doing that now. [19:18:34] Also, try to !log here as you do deployments [19:18:52] Yea, forgot that step. My bad. [19:18:53] thanks for owning it! [19:19:24] um... I guess the stuff in the damaged queue will need to be flopped back into the real queue but only once the code has been fixed [19:20:23] argh, sorry [19:21:05] ejegg: XenoRyet awight anything we need to tell fr-online or MBeat ? [19:21:16] Ok, rolled back [19:21:52] dstrine: I don't think so... [19:21:57] ok coo [19:22:00] *cool [19:22:48] !log reverted SmashPig from 4b930ada57d99070713be18e235238078e1c1e48 to 372cd4008fee3fd02ad2eae9163cf7b28d2ef7c8 [19:22:54] Logged the message at https://wikitech.wikimedia.org/wiki/Server_Admin_Log, Master [19:24:02] (CR) Ejegg: [C: 2] Provide a workaround for confusing payment instrument/method mapping [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/313234 (https://phabricator.wikimedia.org/T146922) (owner: Awight) [19:25:25] ejegg: XenoRyet: cwd|afk: I'm curious what the status of the PP listener is. Are we still trying to deploy the new code and rolling back to the old code? Or making tweaks to the new code? [19:26:00] awight: tweaks to the new code to avoid spamming the queue on monitor hits [19:26:30] Fundraising Sprint Rocket Surgery 2016, Fundraising-Backlog, MediaWiki-extensions-CentralNotice, Operations, and 3 others: Banner not showing up on site - https://phabricator.wikimedia.org/T144952#2675165 (awight) Just a silly follow-up to my meta vs meta.m remarks: the banner we were investigati... [19:26:43] then once that works we can re-deploy your redirect over the old code [19:26:46] (Merged) jenkins-bot: Provide a workaround for confusing payment instrument/method mapping [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/313234 (https://phabricator.wikimedia.org/T146922) (owner: Awight) [19:27:05] ejegg: That's confusing. But the old code is still deployed at /paypal? [19:27:13] awight yep [19:27:24] How did we get such a high volume of failmail, then? [19:27:37] ah I see, this is unrelated--pending queue [19:27:37] the new db error checking [19:27:49] yeah, those payments_initial selects were failing all along [19:27:55] we just weren't logging anything [19:27:58] ouch. [19:28:07] yeah, checking on the credentials now [19:28:32] ... I do fear we've suddenly introduced a whole bunch of tight coupling and circular dependencies with that stuff [19:28:58] Guess we'll cross that bridge later tho [19:34:52] (PS1) Ejegg: Temporary workaround: don't check payments_initial [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/313241 [19:35:00] awight: have you peeked at the code that actually writes the CN banner i18n message to the cache? (Was thinking of debugging in to find it, but if know its hiding place already....) [19:35:20] I guess it must be in MessageCache [19:35:30] awight: what do you think of that temp workaround? ^^ [19:35:52] since payments_initial checks weren't actually doing anything so far [19:37:24] ejegg: That doesn't sound right--but if they're not doing anything at the moment, yeah [19:37:55] There are no loglines for "Skipping finalized" ? [19:38:00] awight: nope [19:38:14] since those selects were failing silently till we deployed the error checking [19:46:36] (PS1) Awight: Provide a workaround for confusing payment instrument/method mapping [wikimedia/fundraising/crm] (deployment) - https://gerrit.wikimedia.org/r/313246 (https://phabricator.wikimedia.org/T146922) [19:46:49] (CR) Awight: [C: 2] Provide a workaround for confusing payment instrument/method mapping [wikimedia/fundraising/crm] (deployment) - https://gerrit.wikimedia.org/r/313246 (https://phabricator.wikimedia.org/T146922) (owner: Awight) [19:47:14] (Merged) jenkins-bot: Provide a workaround for confusing payment instrument/method mapping [wikimedia/fundraising/crm] (deployment) - https://gerrit.wikimedia.org/r/313246 (https://phabricator.wikimedia.org/T146922) (owner: Awight) [19:48:19] !log update fundraising CRM from 6b2bd98fce5006030423bccc4f4b7fd9b5d14821 to 86371233a9218a526f85cd76d7008f966042fde7 [19:48:24] Logged the message at https://wikitech.wikimedia.org/wiki/Server_Admin_Log, Master [19:48:57] ccogdill: You can try importing the Raw Payment Instrument file on production whenever you have an opportunity. [19:49:06] cool! [19:49:42] awight 1 question — will the civi import automatically send a TY? [19:49:49] it didn’t on staging but that makes sense... [19:49:55] I just want to make sure it will [19:50:28] Yes, it will unless you put something into the Thank You Date or No Thank You Reason columns [19:50:33] got it [19:50:34] thanks [19:50:57] ccogdill: a side question--you're using the Generic Import Template.xls, right? [19:51:35] The Excel version has some cell comments that may or may not be useful... [19:51:56] feedback and edits welcome :) [19:51:59] Leanne sent me her version, no comments on it [19:52:05] where does that file live? [19:53:07] d'oh. [19:54:47] so, I'm going to re-queue all those damaged pending messages [19:55:07] imported successfully! thanks, awight [19:56:09] ccogdill: Whew--the templates live in a very strange place, https://github.com/wikimedia/wikimedia-fundraising-crm/tree/master/sites/all/modules/offline2civicrm/templates [19:57:20] (PS2) Ejegg: Temporary workaround: don't check payments_initial [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/313241 [19:57:29] (CR) Awight: [C: 2] "PS 2: explanatory comment" [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/313241 (owner: Ejegg) [19:58:20] haha never would’ve thought to check there [19:58:33] yeeeah... that'll need a wiki page [20:02:40] ccogdill: I've heard some good reports about your talk [20:03:00] oh awesome! it was a lot of fun [20:03:22] yeah, congratulations! [20:03:43] (Merged) jenkins-bot: Temporary workaround: don't check payments_initial [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/313241 (owner: Ejegg) [20:04:20] they are all quoting you 'the boring email is the best' [20:04:57] haha uh oh [20:06:53] * awight eyes spam filters [20:35:35] fr-tech this patch is suddenly kinda needful: https://gerrit.wikimedia.org/r/313117 [20:41:23] (PS3) Awight: Get all queues with QueueFactory::getQueue [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/313117 (owner: Ejegg) [20:41:43] sure, will review. [20:41:48] thank you! [20:42:23] mostly ingenico in pending, so it's only orphan txns that should be affected [20:46:08] (CR) Awight: Get all queues with QueueFactory::getQueue (3 comments) [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/313117 (owner: Ejegg) [20:46:24] not necessarily -1'ing about that stuff, just irking [20:47:04] awight: cool, I'll fix it up right quick [20:47:26] only if you agree [20:47:42] yeah, seems to make sense [20:48:19] (CR) Awight: [C: 1] "Fool-resistant! Holding off until tweaks." [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/313117 (owner: Ejegg) [20:49:07] I've been battling AbuseFilter, just trying to write a short wiki page of documentation. Why does anybody use our software? [20:49:17] https://phabricator.wikimedia.org/T146938 [20:52:52] ejegg: Are you already amending, or can I just +2 the original patch? It's not the fault of that patch... [20:53:09] awight: about to push up a change [20:53:18] kk [20:53:25] but it's a kinda deeply nested array to override with... [20:54:56] (PS4) Ejegg: Get all queues with QueueFactory::getQueue [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/313117 [20:56:23] (CR) Awight: [C: 2] Get all queues with QueueFactory::getQueue [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/313117 (owner: Ejegg) [20:56:28] thanks! [20:56:44] That was indeed funky. But the QueueFactory is much more usable than bare config stuff [20:56:56] XenoRyet: as soon as that merges to master, you want to prep another deploy branch merge? [20:57:20] (Merged) jenkins-bot: Get all queues with QueueFactory::getQueue [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/313117 (owner: Ejegg) [20:57:28] sure [20:57:34] thanks! [21:00:35] ejegg: the purpose there is just to insinuate a default queue into the config? [21:00:49] cwd yep :) [21:01:00] and any other queue defaults we might want to set later [21:01:05] a good behavior for sure [21:01:15] like putting the creds in one place instead of copying with yaml voodoo [21:01:17] but i think we should make it the only one [21:01:46] instead of having it wrap config, because that leaves this dangling interface that you are free to use but you will have a bad time [21:02:06] sorry, make what the only one? [21:02:33] make QueueFactory the only way to get the queue, if it's the only correct way [21:02:53] instead of wrapping config which is now a deprecated interface [21:03:39] that or change the behavior of config to Do The Right Thing [21:03:43] It seems fair to leave the raw access, and just document that QueueFactory is the only correct way [21:04:06] what's the point? seems like a trap [21:04:07] We've wrapped the databases as well, and IMO it's a good thing [21:04:38] Not much of a trap if we say, $config->object allows you to get raw stuff with no helpful overrides [21:05:13] but in this case it sounds like there is no correct behavior without the overrides [21:05:38] Which code would be allowed vs disallowed from using Configuration? Hmm e.g. PayPalAPI and its factory [21:05:42] cwd hard to enforce there [21:06:38] Fundraising Sprint Rocket Surgery 2016, Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM: Not able to save activity edits - https://phabricator.wikimedia.org/T137740#2675570 (Eileenmcnaughton) Open>Resolved [21:06:38] I agree that we should now only use the new interface, but how would we hide the raw one? And if we wanted to write a new class to wrap some config in niceties, that would be allowed to use config->object I assume [21:07:36] Fundraising Sprint Rocket Surgery 2016, Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM, Patch-For-Review: Issues manually merging records - https://phabricator.wikimedia.org/T144440#2675573 (Eileenmcnaughton) Open>Resolved Checked & this seems OK now [21:08:38] well my first thought is config&object is broken if it's not doing what you want [21:08:50] Fundraising-Backlog: Option to specify primary address on merge - https://phabricator.wikimedia.org/T146941#2675580 (Eileenmcnaughton) [21:08:55] (PS1) XenoRyet: Get all queues with QueueFactory::getQueue [wikimedia/fundraising/SmashPig] (deployment) - https://gerrit.wikimedia.org/r/313317 [21:08:58] AndyRussG|eats: Did you still want to discuss the BannerLoader bug? [21:09:20] XenoRyet: hmm, I think you can merge both that and the parent patch [21:09:35] just check out deployment and git merge master [21:09:53] Should I abandon this one first? [21:09:58] ejegg: I think there's a lot of stuff that's been cherry-picked around. [21:10:05] yah, might as well [21:10:09] awight: not in SmashPig [21:10:20] oops yeah sorry to butt in [21:10:30] (Abandoned) XenoRyet: Get all queues with QueueFactory::getQueue [wikimedia/fundraising/SmashPig] (deployment) - https://gerrit.wikimedia.org/r/313317 (owner: XenoRyet) [21:12:23] Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM: Civi: Contribution report: add column for preferred language - https://phabricator.wikimedia.org/T146942#2675606 (DStrine) [21:13:16] ejegg: Just to make sure I'm doing the right thing, the merge complains about all the tests, but I just take those out because we don't deploy tests. [21:13:31] XenoRyet: yep! [21:13:48] git mergetool should ask about 'em and you can just reply 'd' to all [21:14:52] (PS1) XenoRyet: Merge branch 'master' into HEAD [wikimedia/fundraising/SmashPig] (deployment) - https://gerrit.wikimedia.org/r/313319 [21:15:06] gonna relocate for the meetingthon [21:15:33] XenoRyet: cool, you can self-+2 that [21:15:44] (CR) XenoRyet: [C: 2] Merge branch 'master' into HEAD [wikimedia/fundraising/SmashPig] (deployment) - https://gerrit.wikimedia.org/r/313319 (owner: XenoRyet) [21:16:06] Anything I can do to help with the queue overhaul [21:16:08] ? [21:16:15] we often put a list of the merged patches in the commit message, which you can generate with 'git log --oneline --no-merges --reverse deployment..master' [21:16:31] Yea, just realized I forgot that step. [21:16:47] no worries. want to deploy again? [21:16:48] Or also "git cherry -v origin/deployment origin/master" [21:17:12] same as before, only in the _update step you have to say -P SmashPig=HEAD [21:17:27] because the rollback pinned the revision [21:18:17] Sounds good [21:22:31] (Merged) jenkins-bot: Merge branch 'master' into HEAD [wikimedia/fundraising/SmashPig] (deployment) - https://gerrit.wikimedia.org/r/313319 (owner: XenoRyet) [21:25:09] !log update SmashPig from 372cd4008fee3fd02ad2eae9163cf7b28d2ef7c8 to d2589271d43d8895fbb928f54f8feb43fcbc43c9 [21:25:15] Logged the message at https://wikitech.wikimedia.org/wiki/Server_Admin_Log, Master [21:26:01] XenoRyet: ejegg awight can anyone pass me a link to the Japanese adyen form? [21:27:40] dstrine: https://payments.wikimedia.org/index.php?title=Special:AdyenGateway&appeal=JimmyQuote&payment_method=cc&recurring=&uselang=ja&language=ja¤cy_code=JPY&amount=1000&country=JP [21:28:31] XenoRyet: ok, want to re-run that Requeue job? [21:28:42] Yea, will do. [21:28:44] Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM, Epic: Merge conflict on source - https://phabricator.wikimedia.org/T146946#2675687 (Eileenmcnaughton) [21:44:21] XenoRyet: want to deploy PaymentListeners ? [21:45:14] Give me like 10 mins? [21:45:17] that just for the redirect? [21:45:30] Or go ahead and go without me if you need to. [21:45:48] cwd yep, just the redirect [21:46:06] actually, shoot, i need to relocate still. cwd you want to deploy that? [21:46:21] sure thing [21:47:05] woot! back soon... [21:50:43] Fundraising-Backlog, fundraising-tech-ops, FR-Smashpig: SmashPig fredge database credentials are incorrect - https://phabricator.wikimedia.org/T146950#2675781 (awight) [21:54:37] !log PaymentListeners updated from b4d77a991e100f97d98fcd72eaf03940a4e1845d to 21647c8f4b781b74ae2dc4377334410b4eed7 [21:54:42] Logged the message at https://wikitech.wikimedia.org/wiki/Server_Admin_Log, Master [22:01:24] Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM, Epic: Merge conflict on source - https://phabricator.wikimedia.org/T146946#2675797 (Eileenmcnaughton) @CCogdill_WMF Do you use the field 'civicrm_contact.source' at all? Major Gifts & Donor services were happy just to keep the value associated with... [22:02:14] Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM, Epic: Merge conflict on source - https://phabricator.wikimedia.org/T146946#2675799 (CCogdill_WMF) No, I don't use that field! [22:05:29] Fundraising Sprint Octopus Untangling, Fundraising Sprint Pretending This Isn't Happening, Fundraising Sprint Qwerty Thwacking, Fundraising Sprint Rocket Surgery 2016, and 3 others: Silverpop export not capturing language changes in Civi (not sure if... - https://phabricator.wikimedia.org/T96410#2675803 [22:06:02] !log PaymentListeners rolled back [22:06:07] Logged the message at https://wikitech.wikimedia.org/wiki/Server_Admin_Log, Master [22:33:23] ______________ [22:33:25] < fr-tech talk > [22:33:27] -------------- [22:33:29] \ [22:33:31] \ [22:33:33] ^__^ [22:33:35] _______/(oo) [22:33:37] /\/( /(__) [22:33:39] | W----|| |~| [22:33:41] || || |~| ~~ [22:33:43] |~| ~ [22:33:45] |_| o [22:33:47] |#|/ [22:33:49] _+#+_ [22:34:18] yesss [22:40:11] Fundraising-Backlog, MediaWiki-extensions-CentralNotice, Tracking: [Tracking] CentralNotice: Unexplained banner display failures - https://phabricator.wikimedia.org/T145719#2675902 (DStrine) [22:40:14] Fundraising Sprint Qwerty Thwacking, Fundraising Sprint Rocket Surgery 2016, Fundraising-Backlog, MediaWiki-extensions-CentralNotice, and 2 others: Banners not allocated correctly - https://phabricator.wikimedia.org/T144393#2675901 (DStrine) Open>Resolved [22:43:05] (PS1) Eileen: WIP add check for stable data before running GC charges. [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/313329 [22:47:15] * AndyRussG switches to monospaced font [22:50:11] Fundraising-Analysis, Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM: Civi: Contribution report: add column for preferred language - https://phabricator.wikimedia.org/T146942#2675938 (Eileenmcnaughton) [22:51:05] (PS11) Ejegg: Move Unsubscribe Queue Consumer off ActiveMQ [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/312084 (https://phabricator.wikimedia.org/T145419) (owner: XenoRyet) [22:51:39] (CR) jenkins-bot: [V: -1] Move Unsubscribe Queue Consumer off ActiveMQ [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/312084 (https://phabricator.wikimedia.org/T145419) (owner: XenoRyet) [22:56:37] (CR) Ejegg: "recheck" [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/312084 (https://phabricator.wikimedia.org/T145419) (owner: XenoRyet) [23:00:27] (CR) Ejegg: "recheck" [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/312084 (https://phabricator.wikimedia.org/T145419) (owner: XenoRyet) [23:06:47] (PS1) Aaron Schulz: Use DatabaseBase::factory() over Database::factory() [extensions/ContributionTracking] - https://gerrit.wikimedia.org/r/313334 [23:07:03] Fundraising-Backlog, MediaWiki-extensions-CentralNotice: Fatal exception when cloning/saving banners - https://phabricator.wikimedia.org/T146880#2675993 (awight) @AndyRussG note that we can get the details of these exceptions from fluorine:/a/mw-log/ exception.log or fatal.log, by searching for those ex... [23:09:13] (PS1) Ejegg: Logging for JobQueueConsumer [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/313337 [23:09:14] Fundraising Sprint S 2016, Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM, Epic: Merge conflict on source - https://phabricator.wikimedia.org/T146946#2675998 (DStrine) [23:09:17] Fundraising Sprint S 2016, Fundraising-Backlog, MediaWiki-extensions-CentralNotice: Fatal exception when cloning/saving banners - https://phabricator.wikimedia.org/T146880#2675999 (DStrine) [23:09:19] Fundraising Sprint S 2016, Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM, Spike: SPIKE: investigate level of completeness of civi deduped DB - https://phabricator.wikimedia.org/T146384#2676000 (DStrine) [23:09:21] Fundraising Sprint S 2016, Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM, Epic, Patch-For-Review: Resolve conflicts on preferred language - https://phabricator.wikimedia.org/T146344#2676001 (DStrine) [23:09:25] Fundraising Sprint Rocket Surgery 2016, Fundraising Sprint S 2016, Fundraising-Backlog, FR-ActiveMQ, MediaWiki-extensions-DonationInterface: Orphan rectifier not sending to payments-initial nor payments-fraud queues - https://phabricator.wikimedia.org/T145848#2676003 (DStrine) [23:09:28] Fundraising Sprint Rocket Surgery 2016, Fundraising Sprint S 2016, Fundraising-Backlog, MediaWiki-extensions-CentralNotice, Patch-For-Review: CentralNotice: Support MariaDB Strict Mode - https://phabricator.wikimedia.org/T145591#2676004 (DStrine) [23:09:31] Fundraising Sprint Rocket Surgery 2016, Fundraising Sprint S 2016, Fundraising-Backlog, FR-ActiveMQ, MediaWiki-extensions-DonationInterface: [Tracking] Stop producing to ActiveMQ - https://phabricator.wikimedia.org/T145660#2676005 (DStrine) [23:09:34] Fundraising Sprint Rocket Surgery 2016, Fundraising Sprint S 2016, Fundraising-Backlog, MediaWiki-extensions-CentralNotice: Impressions slow to update on jaJP dsk campaign - https://phabricator.wikimedia.org/T145446#2676008 (DStrine) [23:09:37] Fundraising Sprint Qwerty Thwacking, Fundraising Sprint Rocket Surgery 2016, Fundraising Sprint S 2016, Fundraising-Backlog, and 3 others: Investigate the discrepancy between pageview counts and banner impressions for WLM 2016 - https://phabricator.wikimedia.org/T145565#2676006 (DStrine) [23:09:40] Fundraising Sprint Rocket Surgery 2016, Fundraising Sprint S 2016, Fundraising-Backlog, MediaWiki-extensions-CentralNotice, Patch-For-Review: CN Campaign Setting oddities - https://phabricator.wikimedia.org/T145447#2676007 (DStrine) [23:09:42] Fundraising Sprint Rocket Surgery 2016, Fundraising Sprint S 2016, Fundraising-Backlog, FR-ActiveMQ, Patch-For-Review: Migrate FundraisingEmailUnsubscribe to new queue backend - https://phabricator.wikimedia.org/T145419#2676010 (DStrine) [23:09:44] Fundraising Sprint Qwerty Thwacking, Fundraising Sprint Rocket Surgery 2016, Fundraising Sprint S 2016, Fundraising-Backlog, and 2 others: Migrate refunds to new queue - https://phabricator.wikimedia.org/T145234#2676011 (DStrine) [23:09:46] Fundraising Sprint Qwerty Thwacking, Fundraising Sprint Rocket Surgery 2016, Fundraising Sprint S 2016, Fundraising-Backlog, and 2 others: Move recurring queue consumer off ActiveMQ - https://phabricator.wikimedia.org/T145233#2676012 (DStrine) [23:09:50] Fundraising Sprint Rocket Surgery 2016, Fundraising Sprint S 2016, Fundraising-Backlog, MediaWiki-extensions-CentralNotice, and 4 others: Banner not showing up on site - https://phabricator.wikimedia.org/T144952#2676013 (DStrine) [23:09:52] Fundraising Sprint Rocket Surgery 2016, Fundraising Sprint S 2016, Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM: Longterm fix + regression test for T144489 - https://phabricator.wikimedia.org/T144557#2676014 (DStrine) [23:09:54] Fundraising Sprint Octopus Untangling, Fundraising Sprint Qwerty Thwacking, Fundraising Sprint Rocket Surgery 2016, Fundraising Sprint S 2016, and 2 others: Epic: Create frack vm cluster - https://phabricator.wikimedia.org/T142533#2676017 (DStrine) [23:09:56] Fundraising Sprint Octopus Untangling, Fundraising Sprint Pretending This Isn't Happening, Fundraising Sprint Qwerty Thwacking, Fundraising Sprint Rocket Surgery 2016, and 7 others: Move legacy PayPal listener to SmashPig - https://phabricator.wikimedia.org/T141654#2676018 (DStrine) [23:09:59] Fundraising Sprint Pretending This Isn't Happening, Fundraising Sprint Qwerty Thwacking, Fundraising Sprint Rocket Surgery 2016, Fundraising Sprint S 2016, and 5 others: Delete from the pending database when a transaction is failed or completed - https://phabricator.wikimedia.org/T143945#2676016 (... [23:10:01] Fundraising Sprint Pretending This Isn't Happening, Fundraising Sprint Qwerty Thwacking, Fundraising Sprint Rocket Surgery 2016, Fundraising Sprint S 2016, and 4 others: Donation import should merge and delete pending db info - https://phabricator.wikimedia.org/T141477#2676019 (DStrine) [23:10:03] Fundraising Sprint Octopus Untangling, Fundraising Sprint Pretending This Isn't Happening, Fundraising Sprint Qwerty Thwacking, Fundraising Sprint Rocket Surgery 2016, and 5 others: Decimal dates are breaking our queue consumer - https://phabricator.wikimedia.org/T140667#2676020 (DStrine) [23:10:05] Fundraising Sprint Rocket Surgery 2016, Fundraising Sprint S 2016, Fundraising-Backlog, Wikimedia-Fundraising-CiviCRM, Patch-For-Review: Receipt/Thank you letters from Civi - Fix template? - https://phabricator.wikimedia.org/T137496#2676021 (DStrine) [23:10:07] Fundraising Sprint S 2016, Fundraising-Backlog, FR-PayPal-ExpressCheckout, FR-Paypal: Renew PayPal API certificate. Due date: Aug 1st - https://phabricator.wikimedia.org/T136742#2676022 (DStrine) [23:10:10] Fundraising Sprint Pretending This Isn't Happening, Fundraising Sprint Qwerty Thwacking, Fundraising Sprint Rocket Surgery 2016, Fundraising Sprint S 2016, and 6 others: Migrate donations to new queue - https://phabricator.wikimedia.org/T131277#2676024 (DStrine) [23:10:13] Fundraising Sprint Hermit Crab Husbandry, Fundraising Sprint Internet Exploring, Fundraising Sprint Jabberwock Slaying, Fundraising Sprint Killing Time, and 16 others: [Epic] All pending producers should write to a single pending queue - https://phabricator.wikimedia.org/T133195#2676023 (DStrine) [23:10:15] Fundraising Sprint Octopus Untangling, Fundraising Sprint Pretending This Isn't Happening, Fundraising Sprint Qwerty Thwacking, Fundraising Sprint Rocket Surgery 2016, and 4 others: Silverpop export not capturing language changes in Civi (not sure if... - https://phabricator.wikimedia.org/T96410#2676025 [23:21:43] eileen: Hi, getting on the 4:30 call now [23:21:57] err, sorry to be so Pac-centric [23:23:01] be there in 5 [23:26:58] (CR) Awight: [C: 2] Logging for JobQueueConsumer [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/313337 (owner: Ejegg) [23:27:42] (Merged) jenkins-bot: Logging for JobQueueConsumer [wikimedia/fundraising/SmashPig] - https://gerrit.wikimedia.org/r/313337 (owner: Ejegg) [23:34:33] ejegg, awight - is that to track down the segfault? [23:35:36] cwd I was thinking it might help [23:36:23] good call. mysterious [23:37:37] (PS1) Ejegg: Merge branch 'master' into deployment [wikimedia/fundraising/SmashPig] (deployment) - https://gerrit.wikimedia.org/r/313339 [23:42:23] (CR) Ejegg: [C: 2] Merge branch 'master' into deployment [wikimedia/fundraising/SmashPig] (deployment) - https://gerrit.wikimedia.org/r/313339 (owner: Ejegg) [23:48:34] (CR) Ejegg: [V: 2] Merge branch 'master' into deployment [wikimedia/fundraising/SmashPig] (deployment) - https://gerrit.wikimedia.org/r/313339 (owner: Ejegg) [23:51:15] !log updated SmashPig from d2589271d43d8895fbb928f54f8feb43fcbc43c9 to 3458f93599084815da46a2540e9ed762c8b120ce [23:51:20] Logged the message at https://wikitech.wikimedia.org/wiki/Server_Admin_Log, Master