| *** nikhil has quit IRC | 02:58 | |
| *** david-lyle has joined #openstack-searchlight | 03:13 | |
| *** lakshmiS has joined #openstack-searchlight | 03:23 | |
| *** nikhil has joined #openstack-searchlight | 04:00 | |
| *** nikhil has quit IRC | 04:14 | |
| *** GB21 has joined #openstack-searchlight | 06:31 | |
| *** lakshmiS has quit IRC | 08:33 | |
| *** sballe has joined #openstack-searchlight | 09:15 | |
| *** lakshmiS has joined #openstack-searchlight | 10:21 | |
| *** lakshmiS has quit IRC | 10:37 | |
| *** lakshmiS has joined #openstack-searchlight | 11:11 | |
| *** GB21 has quit IRC | 13:07 | |
| *** lakshmiS has quit IRC | 13:20 | |
| *** nikhil has joined #openstack-searchlight | 14:06 | |
| openstackgerrit | Steve McLellan proposed openstack/searchlight: Add faceting https://review.openstack.org/222388 | 14:53 |
|---|---|---|
| sjmc7 | TravT - about the created_at, updated_at, lakshmi's comment made me wonder. should this be more a documentation thing, rather than enforced in code? can we legitimately say that everything must always have a created and updated time? | 14:55 |
| sjmc7 | or should we say that if your resource does, you should provide those fields? i can't think of anything that doesn't right now (although metadefs only have created) | 14:56 |
| TravT | sjmc7 let me go read his comment | 14:58 |
| sjmc7 | https://review.openstack.org/#/c/226524/ | 14:58 |
| sjmc7 | he suggested moving it to a base class, which is reasonable given that we're enforcing the rule, but that made me wonder if we should be enforcing it at this level | 15:01 |
| sjmc7 | the other option is to have e-s ignore missing sort fields | 15:01 |
| TravT | well, I don't think we should auto-add the fields if they don't exist at all. | 15:03 |
| TravT | but if they do exist, we should map them to the same field. | 15:04 |
| sjmc7 | "we" can't do anything; the plugin author has to do it | 15:04 |
| sjmc7 | that being the case, there's not really anything to do in this patch other than adjusting the metadef plugin's behavior | 15:04 |
| TravT | maybe we could provide a helper function (normalize fields): | 15:04 |
| TravT | nova is actually wrong at the moment. | 15:04 |
| TravT | the plugin has a mapping for created_at / updated_at | 15:04 |
| TravT | but that isn't what is provided in data | 15:04 |
| TravT | it provides created / updated | 15:04 |
| sjmc7 | ah yes, i forgot that. right - so that's something that should be done in the nova serialization | 15:04 |
| sjmc7 | ok. i'm leaning more towards correcting the behavior of these plugins and making this a documentation issue | 15:05 |
| TravT | what if in base or in utils, we provided a helper function | 15:05 |
| sjmc7 | what would it do? | 15:05 |
| sjmc7 | take a list of before/after pairs? | 15:05 |
| TravT | well, right now in my horizon plugin, i'm doing something similar | 15:06 |
| TravT | function tempNormalizeCreateUpdateDateField(item) { | 15:06 |
| TravT | item.created = item.created || item.created_at || null; | 15:06 |
| TravT | item.created = $filter('date')(item.created, 'short'); | 15:06 |
| TravT | delete item.created_at; | 15:06 |
| TravT | item.updated = item.updated || item.updated_at || null; | 15:06 |
| TravT | item.updated = $filter('date')(item.updated, 'short'); | 15:06 |
| TravT | delete item.updated_at; | 15:06 |
| TravT | } | 15:06 |
| TravT | only inverse | 15:06 |
| sjmc7 | yeah | 15:06 |
| TravT | a similar function would just take an object | 15:06 |
| TravT | see if it has created and add created_at | 15:06 |
| TravT | same for updated | 15:07 |
| sjmc7 | i'm not sure about a base function that does a load of guessing | 15:07 |
| sjmc7 | i guess if it's just for these fields it's ok | 15:07 |
| sjmc7 | ok, i'll have another crack at it | 15:07 |
| TravT | normalizing the date format would be kinda nice too | 15:08 |
| sjmc7 | e-s does that | 15:08 |
| TravT | ok | 15:08 |
| sjmc7 | although if we have this normalize function do we let it implicitly map those? | 15:09 |
| sjmc7 | guess that can be left up to the plugin | 15:12 |
| TravT | well, i've been thinking that we could just add a general template mapping that all mapping get | 15:19 |
| TravT | like what I had ekarlso do for _id fields | 15:19 |
| sjmc7 | yeah, that's not a bad idea | 15:19 |
| TravT | or we just have a "common mappings" utility that contains them and plugins just grab them and add to their mapping | 15:20 |
| sjmc7 | yeah. ok, i'll fix the nova issue and put up another patch. this might be somethng we revisit when there are more plugins | 15:21 |
| TravT | sure, sounds good. | 15:21 |
| TravT | i'll get back to my refactoring of the horizon plugin | 15:21 |
| sjmc7 | :) | 15:21 |
| ekarlso | TravT: whaaaa | 15:30 |
| sjmc7 | :D | 15:34 |
| TravT | ekarlso, yeah, i'll put some patch out by end of day no matter what. | 15:35 |
| *** GB21 has joined #openstack-searchlight | 17:09 | |
| sjmc7 | TravT - doc_values might be useful for some fields (like the metadef descriptions) but for most of our stuff the volume we're looking at is so low i'm not sure it'll matter a huge amount | 17:20 |
| *** sballe has quit IRC | 18:26 | |
| *** GB21 has quit IRC | 18:37 | |
| TravT | ekarlso: https://review.openstack.org/#/c/219271/ | 18:43 |
| TravT | couple comments on there. if you can take a look. | 18:43 |
| TravT | would like to get this one through | 18:43 |
| david-lyle | TravT: is the partial fix for https://bugs.launchpad.net/horizon/+bug/1496627 good enough for Liberty? | 20:18 |
| openstack | Launchpad bug 1496627 in OpenStack Dashboard (Horizon) "hz-magic-search-bar does not honor isServer property" [Medium,In progress] - Assigned to Rajat Vig (rajatv) | 20:18 |
| david-lyle | the one that already merged? | 20:18 |
| TravT | i'm actually about 30 minutes away from having my code updated to try that out again. | 20:19 |
| david-lyle | ok, I'll leave it in RC-1 for now, if it fixed it, can you remove the milestone? | 20:20 |
| TravT | battling a slightly different issues with magic search facets that I'm trying to figure out if it is my issue or not. | 20:20 |
| TravT | ok | 20:20 |
| david-lyle | trying to get down to final RC bugs | 20:20 |
| david-lyle | thanks! | 20:20 |
| TravT | makes sense | 20:20 |
| TravT | thanks | 20:20 |
| sjmc7 | let me know if i can help, TravT | 20:20 |
| TravT | can you read spaghetti? | 20:20 |
| TravT | sjmc7: i just incorporated your per type facet patch | 20:21 |
| TravT | that part seems to be working fine | 20:21 |
| TravT | on the searchlight side | 20:21 |
| sjmc7 | i am fluent! | 20:22 |
| TravT | and horizon side... the issue i'm having at the moment is with figuring out magic search events | 20:22 |
| sjmc7 | i don't speak magic search | 20:22 |
| sjmc7 | kelly might be able to help, depending on what she's up to her ears in | 20:23 |
| TravT | i'll ping her if needed... this just happens to be what i'm working with at this moment in time. | 20:23 |
| TravT | smc7: if you have some time, maybe you can look back over the functional tests patch to see if lakshmiS addressed your last comments | 20:24 |
| sjmc7 | yep, planning to | 20:25 |
| sjmc7 | many, many fires today | 20:25 |
| TravT | Fire Danger: Critical | 20:25 |
| sjmc7 | you have no idea | 20:25 |
| TravT | let's keep it that way | 20:25 |
| TravT | :P | 20:25 |
| sjmc7 | :) | 20:25 |
| openstackgerrit | Steve McLellan proposed openstack/searchlight: Standardize created_at, updated_at https://review.openstack.org/226524 | 20:27 |
| *** sigmavirus24_awa has quit IRC | 20:54 | |
| *** sigmavirus24_awa has joined #openstack-searchlight | 21:00 | |
| *** openstackgerrit has quit IRC | 22:16 | |
| *** openstackgerrit has joined #openstack-searchlight | 22:16 | |
| TravT | ekarlso: here's the initial code for the horizon panel: https://review.openstack.org/#/c/227036/ | 22:38 |
| TravT | review -d from horizon | 22:38 |
| ekarlso | TravT: i'll have to look at it tmrw :| | 23:33 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!