[00:02:29] (CR) Ejegg: [C: 2] "Good to have these, even if they don't all pass yet" [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/149515 (owner: Awight) [00:04:58] ejegg: yah I think another aspect of the fail is that I rebased all the test patches to come after the fixes I wanted to get out last week. That can't have been good for them. [00:05:17] and no fun for you, either [00:12:34] (CR) Ejegg: [C: 2] "Could add more assertions about refund results, but this is a great start." [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/149517 (owner: Awight) [00:14:49] is Tony Le actually our Paypal account holder? [00:19:40] ejegg: yep [00:20:20] just looking for ipn message digestion code. found a table on the web says EOT messages have a bunch of vars that aren't in that json [00:20:32] but I'm guessing the json has all the ones we use [00:21:27] ejegg: ah interesting. They may have added stuff--URL? [00:21:39] https://www.paypal.com/cgi-bin/webscr?cmd=p/acc/ipn-subscriptions-outside [00:21:45] I think what happened is I captured and sanitized a live message, which doesn't necessarily have all the fields [00:21:51] table of which message has what is at the bottom [00:22:02] nice one, thx. [00:22:02] oh, cool [00:22:27] so, is there actually code for Paypal in queue2civivrm? [00:22:53] yes, recurring.module normalizes into a format that wmf_civicrm can consume [00:22:57] nvm, found it [00:23:04] remember to bleach your eyeballs after reading that code [00:23:09] heh [00:23:11] I don't want that spreading... [00:24:11] oh, hello db query! [00:26:13] ejegg: eh? [00:26:29] (CR) Ejegg: [C: 2] "could come in handy" [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/149520 (owner: Awight) [00:27:06] thought it was just mapping array keys, then it got into some heavier logic [00:28:25] ys... We've been having interesting conversations about doing normalization through a generalized thing, to capture that part of the logic in a uniform way. [00:29:03] i guess if some gateways just don't give you as much info, you need to do more logic. [00:29:30] though a generic test for missing stuff and remapping after doing the array key normalization might work [00:31:22] Yeah there are many cases where we need code and not just a simple rule, but I'd like all the codebases that do that to present their transformations in the same way [00:31:29] Then we can actually see what's happening... [00:32:51] Basically, the same idea as in DonationInterface, but we would split the transformations out from the rest of the application logic [00:34:10] awight: I'll need production cluster access to view the HideBanners logs set up by that udp-filter that was added today, right [00:34:13] ? [00:34:30] ejegg: That's a good question, you should ask jgreen [00:34:51] ok, cool [00:34:55] I don't think it will be production cluster access, necessarily, probably he will replicate a syslog stream to frack [00:35:33] oh, interesting, didn't realize these filters were pointed at a syslog, but that makes sense [00:35:41] I think so [00:35:43] see pm [00:36:02] thank you [00:36:34] yuck. [00:37:13] gotta run. [18:19:57] (CR) Ejegg: [C: 2] Make prepared query less magical [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/151230 (owner: Awight)