[04:53:29] (PS3) Legoktm: Add hook to support UserMerge [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/147323 (https://bugzilla.wikimedia.org/67761) (owner: Ejegg) [04:53:40] (CR) Legoktm: [C: 1] Add hook to support UserMerge [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/147323 (https://bugzilla.wikimedia.org/67761) (owner: Ejegg) [13:06:15] (CR) Siebrand: "i18n/L10n reviewed." (2 comments) [extensions/CentralNotice] - https://gerrit.wikimedia.org/r/146362 (owner: Mwalker) [16:59:25] (CR) Ejegg: [C: 2] missing static signature [wikimedia/fundraising/crm/civicrm] - https://gerrit.wikimedia.org/r/146967 (owner: Awight) [17:06:48] (CR) Ejegg: [C: 2] typo causing whitescreen [wikimedia/fundraising/crm/civicrm] - https://gerrit.wikimedia.org/r/142168 (owner: Awight) [17:13:30] (CR) Mwalker: [C: 2] typo [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/145233 (owner: Awight) [17:21:19] (CR) Ejegg: "Should update the call to the fn in this same commit, no?" (1 comment) [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/146978 (owner: Awight) [17:23:15] (CR) Mwalker: [C: 2 V: 2] parameterize postback url [wikimedia/fundraising/PaymentsListeners] - https://gerrit.wikimedia.org/r/146961 (owner: Awight) [17:27:42] ejegg... https://bugzilla.wikimedia.org/show_bug.cgi?id=37790 [17:28:09] :( [17:28:21] reopened, huh? [17:28:56] They're going to have to build a self-hosted shop or find someone else to host it then [17:29:11] I'll put some explanation on the bug [17:30:28] ejegg, atgo; I would recommend closing WONTFIX [17:30:37] yeah that's sort of what i'm thinking [17:30:44] oh right, that's an option [17:31:05] HTTPS works just fine; the rest of our site doesn't redirect to HTTPS automatically; and shopify does push you to HTTPS when you checkout [17:32:09] how upset will folks be if we close as wontfix? [17:33:18] hell if I know; jeremyb should've known better -- his comments should've been in a new bug [17:33:39] but we dont control shopify; the only thing we can control is the links we push to the shop [17:33:56] yeah... i mean the only thing we can do is redirect from http to https [17:34:33] we can put in js to bounce the user to https, but it turns out that breaks logins [17:35:32] hmm [17:35:37] well.. let's table it for now at least [18:48:41] (CR) Qgil: "This is one of the many open reviews that is making DonationInterface look bad at http://korma.wmflabs.org/browser/gerrit_review_queue.htm" [extensions/DonationInterface] - https://gerrit.wikimedia.org/r/64442 (owner: MarkTraceur) [18:54:31] awight that's 130 actually :) [18:57:14] Oh wow my testing patch is still there [18:57:23] I...cannot even say that's still applicable [19:02:16] ejegg: so the unsub error... I would start by looking at Extension:FundraisingEmailUnsubscribe [19:03:43] yeah, I see it tries to send messages on the stomp queue with process, email, and contribution-id params [19:04:45] then the failmail just shows a single array element 'redacted contents of message ID: manual-removal-12312312' [19:05:12] guessing the manual-removal is the process, and the digits are the contribution id? [19:05:32] Not sure if email is just redacted for the failmail or if it was the missing required field [19:05:36] ejegg: we should change that failmail to link directly to the "damaged queue" like the other failmails do [19:05:47] ejegg: but, look at the other failmail and it has the details [19:05:57] the "build failed" [19:06:03] oh, thanks [19:06:19] ah, much more info [19:06:38] oh, can we see the damaged queue in civi? [19:07:09] yep, here it is [19:08:50] ejegg: the dequeue loops are supposed to go through code in crm/modules/wmf_common/Queue.php, which includes the item_url() in the failmail body [20:47:47] (PS1) Ssmith: Add popover for SQL query access [wikimedia/fundraising/dash] - https://gerrit.wikimedia.org/r/147648 [20:48:19] (CR) jenkins-bot: [V: -1] Add popover for SQL query access [wikimedia/fundraising/dash] - https://gerrit.wikimedia.org/r/147648 (owner: Ssmith) [21:03:12] awight at some point today... could you show me the 2 spreadsheets for CB that need mashing for mispayments? i'd like to send instructions to rosie for dealing with that [21:04:08] atgo: sure [21:04:14] atgo: maybe 3:30 or 4 [21:04:18] ok! [21:07:13] awight want to join this GR thing? [21:07:15] we should get a room [21:07:17] if so [21:07:27] awight mushroom? [21:07:34] atgo: I'm finishing a really rocky deployment, I'll try to be there in < 10min [21:15:22] ejegg: fwiw, if you can figure out the URL to that damaged message, you'll be able to download the contents as json, then drop the message so it stops failing [21:15:45] normally this is not necessary, so again we have to make that queue consumer use the common dequeue_loop and reject routines. [21:16:22] awight: I got a look at the filing messages [21:16:34] there are a dozen or so with process='wmf-manual' [21:16:38] and email = '' [21:16:49] But for the life of me I can't figure out how they got there [21:16:54] ejegg: are u sure they aren't all the same message? I see the ID is the same in the failmails [21:17:03] ejegg: yeah that is an important thing to chase down if possible, thx [21:17:15] oh, a bunch have different contribution-id [21:17:29] the only code with that process is the drupal wmf_unsubscribe module [21:17:46] and it does a civi search for that email, and rejects if zero users are found [21:17:48] oooh. [21:17:54] this must be a backend unsubscribe? [21:18:05] so, we should just TODO add validation to that form [21:18:14] much less serious than a frontend failure to validate [21:18:22] I created some email-less users on my own box, and the civi api search for email '' still comes back with zero [21:19:08] so even without client-side validation on the form, the api search/user check should have exited before putting a message on the stomp queue [21:20:18] ejegg: but is this in the drupal module form for admins, or the wiki form? [21:20:29] the drupal module for admins [21:21:05] so yeah, not as bad as if this was external [21:21:08] just puzzling [21:22:33] awight think you'll be able to join here? [21:22:53] atgo: yeah I'm telling ops to clean their damn room and let me out [21:22:58] ;P [21:23:13] ha alright [21:30:13] ejegg - awight and i are in a meeting for a bit. can we postpone standup? [21:30:29] ok, works for me [21:31:48] ok. will ping when we're ready [21:45:18] hey ccogdill [21:45:31] hey atgo [21:45:33] what’s up? [21:45:41] so... i hear this doesn't work [21:45:41] https://wikimedia.mingle.thoughtworks.com/projects/online_fundraiser/cards/1655 [21:46:01] from sisolak... did you create another card for it? [21:46:02] hah I literally just opened this card [21:46:11] well great :) [21:46:20] I think what doesn’t work is the ? issue [21:46:28] oooooh [21:46:29] it was determined that was a result of the automated upload [21:46:32] that makes sense [21:46:38] ok... is that the only thing about it that is broken? [21:46:57] I think so! [21:50:58] ejegg awight standup? [21:52:50] I deployed a fix to the Translate workflow and am waiting for QA now. I'm going to try to continue work on #1623 this afternoon, and maybe take a look at basic things like Civi scheduled reminders. [21:53:17] cool thanks. ejegg do you want to do with camera or IRC? [21:54:04] because hangouts doesn't seem to want to work [21:54:13] ok, in that case IRC is cool [21:54:29] ok :) [21:54:33] I have been confusing myself most of the day... [21:55:01] trying to track down the cause of the unsubscribe failmail [21:55:26] is that related to the bounces? [21:55:28] but the code looks like it shouldn't be allowing the sorts of queue messages that we're getting [21:55:58] nope, just some emails we were getting and Jeff was curious about [21:56:05] oh ok [21:56:34] also with Jeff, trying to find where the Paypal listener actually uses a db connection. I think K4 needs to explain that to us, b/c we don't see it anywhere [21:57:27] then trying to get a mailserver going on my local machine that can actually get some bounce messages back for testing civi config [21:58:03] that's about it [21:59:10] ejegg: I don't think the listener connects to the db. [21:59:13] cool. [21:59:24] yeah she said something abotu listeners yesterday that i can't totally remember [21:59:29] here's pizzza's update: working on the fraud widget, came into contact with some event bugs but working through those. I may switch gears for a bit to make some changes to Megan's prototype [22:02:32] pizzzacat!!! [22:02:43] meowreeer [22:03:24] ejegg: sorry for redundancy, but did you say that the mystery thing came from the admin backend? [22:04:26] yes, from the wmf_unsubscribe feature in civi [22:05:36] all the messages were queued within a couple seconds, so it's possible the api call there is returning a bunch of users even though it doesn't on my machine. [22:06:00] Could just try submitting an empty form and see if we get a new batch [22:06:10] but I have no idea what it strenuous in that db [22:06:14] *what is [22:06:56] aha ok that's interesting [23:00:54] (PS1) Awight: Unsubscribe errors reject the message and are not fatal [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/147675 [23:07:03] (CR) Ejegg: [C: 2] Unsubscribe errors reject the message and are not fatal [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/147675 (owner: Awight) [23:07:32] something's wrong with our gate-and-submit workflow... [23:31:05] awight is there a card for getting the refunds and such to work for BTC? [23:31:19] atgo: No [23:31:24] .. [23:31:35] atgo: it will not take tech work [23:31:46] :D [23:31:59] too sleepy for sarcasm, awight :P [23:32:14] want me to make one or do you have it? [23:32:56] atgo: i would like to not [23:33:04] ha ok. i've got it [23:33:11] * awight drags backpack through the mud [23:34:37] awight https://wikimedia.mingle.thoughtworks.com/projects/online_fundraiser/cards/1815 [23:34:48] listen if you're not going to use that backpack, i need one :P [23:35:19] atgo: oh. did u want to encourage me to work on the "mispayments" [23:35:22] process? [23:35:47] at some point we need to send rosie some info about what she needs to do [23:36:19] just bc i don't have both spreadsheets so.. i don't know what will need correcting in the MO from the mispayments one [23:36:21] but we can do it monday if you want [23:36:58] I also need to rememeber to exclude mispayments from the main import. [23:39:48] ha [23:40:02] also is https://wikimedia.mingle.thoughtworks.com/projects/online_fundraiser/cards/1782 a necessary part of BTC? [23:45:07] atgo: no, i worked around for now. this will just cause original_amount to be unhelpful until fixed. [23:45:18] ok [23:46:10] and.. maybe i'll stop bothering you guys after this [23:46:19] but when is that cookie change goign out? i know it's pending deployment [23:46:23] awight ejegg ^ [23:46:55] should we put that up on beta first? [23:47:21] ejegg: goot idea! [23:47:56] can we put it on beta Monday? [23:48:12] what does this mean? [23:48:22] is beta not in prod? this is the first i've heard beta at the foundatin? [23:48:27] foundation. [23:48:35] i don't only speak in questions [23:48:44] (PS1) Awight: Merge remote-tracking branch 'origin/master' into deployment [wikimedia/fundraising/crm] (deployment) - https://gerrit.wikimedia.org/r/147687 [23:48:55] I think it's new. It's supposed to be set up almost identically to production [23:49:01] (CR) Awight: [C: 2] Merge remote-tracking branch 'origin/master' into deployment [wikimedia/fundraising/crm] (deployment) - https://gerrit.wikimedia.org/r/147687 (owner: Awight) [23:49:36] ejegg: okay, that unsub error should be rejected next time the job runs. [23:49:56] that will push all those messages into the damaged queue, right? [23:50:15] ejegg: I... think so. [23:50:16] So we don't have to manually zap 'em from the real unsub queue? [23:50:20] Cool [23:50:33] At least it should process the real unsub messages [23:51:45] so wait. what does beta do for us? [23:51:53] basically a staging environment? [23:52:57] atgo: yeah it's very much like production [23:53:17] atgo: but doesn't cause us public humiliation [23:53:22] oh ok [23:53:30] so we leave it there for what, a couple of days? [23:53:41] megan's been asking about the cookie change for a while :) [23:54:26] trying to get an estimated delivery date for her since they're ramping up testing [23:57:22] awight ejegg ^ [23:58:43] I have no idea how much traffic beta gets though. Or even how to access it. Let me see [23:58:53] ejegg: ah yeah it will only get our test traffic [23:59:05] I just went through these motions and can help whenever u want [23:59:30] yeah, let's at least put it up Monday for a couple of days and try to hit it from a bunch of clients [23:59:36] Apparently it deploys off of master branches, so there are only special steps if you want to set up configuration. [23:59:41] cool. [23:59:54] ok... so perhaps i can tell her she'll have it by end of next week? [23:59:57] then if nothing looks broken I'd be fine with deploying in the next slot after that [23:59:59] (barring issues)