| *** gouthamr has joined #storyboard | 00:03 | |
| mordred | dhellmann, persia, fungi: I'm still catching up on scrollback, but hopefully everyone knows my feelings on database constraints and my wholehearted support for not having them encoded in the db layer | 00:07 |
|---|---|---|
| mordred | ok. dhellmann already said the words I would have said | 00:07 |
| persia | I'm slow to embrace new concepts, and still don't like the business logic being outside the DB layer, but must agree that when it is, it requires extra cognitive effort (and costs performance) to have them in the DB layer. | 00:19 |
| *** jamesmcarthur has joined #storyboard | 00:31 | |
| fatema_ | dhellmann, Thank you, I was checking that it is really to be done not a dead end. But anyways I'm truly interested in knowing how to implement it -it provokes me that it is strangely easy i concept and taking such long time, that's all- | 00:35 |
| fatema_ | persia, I guess I would discuss it again but anyways lovely to know this new information thank you fungi & persia | 00:36 |
| fatema_ | and dhellmann of course | 00:36 |
| persia | To me, it sounds like the hard part is the transition: figuring out how to make the DDL change without breaking sqlite again. | 00:37 |
| *** jamesmcarthur has quit IRC | 00:44 | |
| fatema_ | I guess this might start by dropping the column enum constraint | 00:52 |
| fatema_ | this requires copying the whole table with the column with no constraint | 00:53 |
| fatema_ | but this is hard to be done with SQL - can't really tell if it's possible or not- | 00:54 |
| fatema_ | basically yes persia this is the confusing part | 01:00 |
| *** fatema_ has quit IRC | 01:27 | |
| *** fatema_ has joined #storyboard | 01:28 | |
| *** fatema_ has quit IRC | 02:00 | |
| *** jamesmcarthur has joined #storyboard | 02:01 | |
| *** jamesmcarthur has quit IRC | 02:19 | |
| *** diablo_rojo has quit IRC | 02:20 | |
| *** persia_ has quit IRC | 02:29 | |
| *** diablo_rojo has joined #storyboard | 02:37 | |
| *** udesale has joined #storyboard | 03:47 | |
| *** jtomasek has joined #storyboard | 07:09 | |
| *** jpich has joined #storyboard | 07:29 | |
| *** tosky has joined #storyboard | 07:34 | |
| *** diablo_rojo has quit IRC | 07:38 | |
| SotK | that was an interesting backscroll discussion | 07:43 |
| SotK | yes, I'm fine with dropping the enum constraint now that its enforced by the API | 07:44 |
| SotK | though I think we'll need a plan for making it easy to deal with changes in acceptable values at the API level being reflected in the database; it would be bad if we changed `merged` to `resolved` but didn't have a simple way for all the tasks with state `merged` to have the state renamed | 07:47 |
| openstackgerrit | lei zhang proposed openstack-infra/storyboard master: Disable the badge from openstackdocstheme 1.23 https://review.openstack.org/595524 | 07:59 |
| *** jaosorior has quit IRC | 08:10 | |
| *** dtantsur|afk is now known as dtantsur | 08:33 | |
| *** jaosorior has joined #storyboard | 09:08 | |
| *** jaosorior has quit IRC | 09:54 | |
| *** chason has joined #storyboard | 10:37 | |
| *** chason has quit IRC | 10:39 | |
| *** chason has joined #storyboard | 10:41 | |
| *** udesale has quit IRC | 11:31 | |
| *** jaosorior has joined #storyboard | 11:54 | |
| *** udesale has joined #storyboard | 12:54 | |
| *** jaosorior has quit IRC | 13:09 | |
| dhellmann | SotK : sure, we still need migration scripts to convert existing values | 13:20 |
| dhellmann | the way I've supported sqlite in the past is to have 1 schema file that describes the full, final, database schema and then the separate migration files to deal with "real" databases like mysql | 13:21 |
| dhellmann | since we test both in the gate, we should be able to keep them in sync, but it is a bit of redundant information to manage | 13:21 |
| dhellmann | alternatively we could write each migration script to be database-aware, and do different things for different databases | 13:22 |
| dhellmann | that doesn't entirely solve the problem of how to write a migration script to change a column type, though | 13:22 |
| dhellmann | it would let us have a migration that dropped the column constraint for task status, though, since we could just skip that step for sqlite | 13:23 |
| dhellmann | and if we combine that with a change to the original migration to avoid adding the constraint to the column under sqlite in the first place, the resulting schema would be the same | 13:23 |
| SotK | I like the idea of database-aware migration scripts, I *think* we either already have some bits that are database-aware or have at least considered it in the past | 13:49 |
| *** ironfoot_ is now known as ironfoot | 14:53 | |
| *** jamesmcarthur has joined #storyboard | 14:57 | |
| *** diablo_rojo has joined #storyboard | 16:03 | |
| *** jamesmcarthur has quit IRC | 16:31 | |
| *** jamesmcarthur has joined #storyboard | 16:32 | |
| *** jpich has quit IRC | 16:32 | |
| *** udesale has quit IRC | 16:40 | |
| *** dtantsur is now known as dtantsur|afk | 17:46 | |
| *** tosky has quit IRC | 17:54 | |
| *** jamesmcarthur has quit IRC | 18:19 | |
| *** jamesmcarthur has joined #storyboard | 19:45 | |
| openstackgerrit | Kendall Nelson proposed openstack-infra/storyboard-webclient master: Show Email Addresses when Searching https://review.openstack.org/589713 | 19:51 |
| *** jamesmcarthur has quit IRC | 21:00 | |
| *** jamesmcarthur has joined #storyboard | 22:07 | |
| *** jamesmcarthur has quit IRC | 22:08 | |
| *** jamesmcarthur has joined #storyboard | 22:08 | |
| *** jamesmcarthur has quit IRC | 22:33 | |
| *** jamesmcarthur has joined #storyboard | 22:59 | |
| *** jamesmcarthur has quit IRC | 23:24 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!