[00:00:19] * James_F nods. [00:05:37] James_F: https://gerrit.wikimedia.org/r/#/c/338026/ [00:05:52] bah [00:05:53] Yeah, pulling locally now. [00:06:30] lol. unrelated lint failure [00:06:43] 00:05:45 Vagrantfile:77:4: C: Use 2 (not 1) spaces for indentation. [00:06:43] 00:05:45 override.vm.box = 'ira/leap' [00:06:43] * bd808 hates rubocop [00:06:46] How did that get merged? [00:06:55] ie, why didn't jenkins say no? [00:07:27] the place it's pointing to is wrong too [00:07:35] stupid rubocop [01:23:11] https://www.tindie.com/products/pons/wplight-iot-wordpress-light/ [01:23:15] We should do a MediaWiki one [01:27:48] Reedy: In your rôle as one of our security experts, you think we should associate MediaWiki with something called "Internet of Things"? :-) [01:28:28] Make IoT great again? [01:28:32] * Reedy hides [01:36:52] Wikipedia is not a thing, it's THE THING!!1 [01:37:15] therefore, there's no IoT, there's only Wikipedia [18:19:03] legoktm: could you take a look on https://gerrit.wikimedia.org/r/#/c/338023/? [18:30:19] uh [18:30:38] I think if a content handler isn't able to create empty content objects stuff will break [18:32:57] bb8fff662a1745c88c68beac45749bfd25755187 is what made me write the test [18:33:23] Daniel was the one who suggested I do it for all content models :P [18:46:00] legoktm: so how do you think it should be fixed? [18:46:42] because we're kind of stuck here - we can't enable wikidata content models properly since they fail this test [18:46:48] SMalyshev: making Wikibase's content handler return some dummy yet valid content for makeEmptyContent() [18:46:50] and wikidata can't really create valid empty objects [18:46:57] why not? [18:47:05] can't you have items that are entirely empty? [18:47:14] legoktm: nope [18:47:21] items should have an ID at least [18:47:54] as I understand, item without an ID is not valid [18:47:58] https://test.wikidata.org/wiki/Q45764 hmm [18:48:13] it has ID of Q45764 ;) [18:48:33] having no statements doesn't mean it's empty. There's still ID [18:48:39] "{\"type\":\"item\",\"id\":\"Q45764\",\"labels\":[],\"descriptions\":[],\"aliases\":[],\"claims\":[],\"sitelinks\":[]}" [18:48:46] exactly [18:48:57] to create it, you need an ID [18:49:01] Q45764 [18:49:17] but makeEmpty doesn't have an argument to supply an ID [18:49:32] moreover, the caller probably doesn't even know how an ID looks like.... [18:51:34] isValid() on entities essentially is: return $this->getEntityId() !== null; [18:51:58] right [18:52:02] so we can create entity without an ID, but it would be invalid [18:52:09] and that's intentional in the code [18:52:24] I'll just leave it up to Daniel then? He's the one who asked for the test and knows the nuances of Wikibase much better than I :) [18:52:32] so we have a contradiction here - that test says all empty entities are valid, and wikidata says empty entities are invalid [18:52:47] Daniel was the one who told me to add that check :) [18:52:52] heh [18:52:56] okay then [18:53:00] (except he's on vacation till 27 so not online :) [18:53:19] 15:09 the idea is that isValid should only be tested if supportsDirectEditing returns true. [18:53:29] I'll +2 then [18:53:30] so that's what I did [18:53:36] legoktm: ok, thanks :) [20:07:42] AaronSchulz: I thought you said doQuickOperation took arrays... [20:07:48] it doesn't for a delete op... [20:37:24] AaronSchulz: https://phabricator.wikimedia.org/T158351 [21:11:41] meh, seemingly transient [22:19:34] are you using doQuickOperation or doQuickOperations? [22:23:24] Plural [22:23:31] It seems it does [22:23:37] I was just going about it the wrong way [22:23:49] I was passing an array of files to delete with one op [22:24:05] when it should be an array of arrays, with operations in each, and the file defined there too