| *** kkup has joined #storyboard | 00:02 | |
| *** jtomasek has joined #storyboard | 06:55 | |
| *** jtomasek has quit IRC | 07:18 | |
| *** jtomasek has joined #storyboard | 07:19 | |
| *** jtomasek has quit IRC | 07:58 | |
| *** jtomasek has joined #storyboard | 07:58 | |
| *** jtomasek has quit IRC | 08:06 | |
| *** jtomasek has joined #storyboard | 09:05 | |
| Zara | yeah, I'm thinking about task-tracking more generally, as a 'has anyone done this anywhere? (how) would this even work?' thing. | 09:22 |
|---|---|---|
| SotK | on merging duplicates, its a pretty interesting thought but idk how it would work properly in my head yet | 11:54 |
| SotK | it certainly seems like it'd be better than the normal "here is a link which duplicated this" thing | 11:54 |
| Zara | yeah, I think if someone wanted to do it, they'd need to start with a list of the things that could go wrong, then work out how to get around those | 11:55 |
| Zara | stunning insight of the day :P | 11:56 |
| Zara | hm, gerrit patch got stuck in gate (I think?) on Friday, have rechecked it | 12:18 |
| Zara | hm, also, just tried to update my db schema diagram so I could have the python client progress up in the open-- 'separating out tables in inkscape' step no longer works. just moves everything at once. maybe that hints at something funny in our db, or maybe it means I'm missing a step. | 12:44 |
| Zara | so yeah, not a big deal in itself, might hint at something more interesting, might not. | 12:45 |
| Zara | oh actually | 12:46 |
| Zara | svg, not png | 12:47 |
| Zara | it works fine | 12:47 |
| Zara | :P | 12:47 |
| Zara | hah, so there were no filter_criteria on march 15th :P comments history and story_permissions are also newer, but I was expecting those. | 13:35 |
| Zara | computer's being weird, am restarting, will be afk for a bit | 14:40 |
| anteaya | oh okay as a philosophical excercise for task-tracking it is an interesting conundrum | 14:53 |
| anteaya | I hadn't thought about it myself | 14:53 |
| persia | One of the tricky aspects is that there are many situations in which a set of apparent duplicates are non-duplicate, or a set of apparent non-duplicates are apparently duplicate. | 14:56 |
| persia | For example, if a hardware-dependent process dies at a certain point on several different pieces of hardware (but works on some others), it may be that these are not actually the same problem (in terms of code), although they may generate logs that are identical excepting the differing hardware. | 14:57 |
| persia | An example in the other direction is the use-before-set class of problem, where a program may behave oddly or crash in a number of places, but the core issue is that someone forgot to initialise a variable in some library. | 14:58 |
| anteaya | I'll take your word for it, as I said I haven't thought about it | 15:18 |
| anteaya | okay dhellmann is taking a sick day today: http://eavesdrop.openstack.org/irclogs/%23openstack-release/%23openstack-release.2016-07-25.log.html#t2016-07-25T11:46:36 I was hoping to get some feedback from him regarding the release team's storyboard needs | 15:22 |
| anteaya | Tuesday is a big day for the release team I understand, so I will try to talk to him later in the week | 15:22 |
| anteaya | fungi: ^^ so it appears I will not have anything new to add to markus_z's request for an agreement regarding how much buffer we assure folks to re-write scripts by tomorrow's infra meeting | 15:23 |
| anteaya | is there anything I can help the storyboard team with today? | 15:24 |
| fungi | anteaya: it seems like something we could just decide? | 15:24 |
| anteaya | fungi: I'm good with that | 15:24 |
| anteaya | since as you point out any minimum we set will be surpassed anyway | 15:24 |
| anteaya | with the time it takes to get agreement | 15:25 |
| fungi | i.e., if we want there to be a warning period between feature completeness (for the sake of migration needs) and a migration date, then we can set that to be however long we want | 15:25 |
| anteaya | agreed | 15:25 |
| anteaya | fungi: I'll let you word it how you like | 15:25 |
| anteaya | I'm not feeling tip top in languaging today | 15:25 |
| anteaya | but I think you have the spirit of what markus is looking for here | 15:25 |
| fungi | anteaya: sure, where did this wording need to be recorded? in the infra meeting? in the tracking story? in an ml thread? | 15:26 |
| anteaya | I think infra meeting | 15:26 |
| anteaya | if you put it there I can copy/paste or point around as needed | 15:27 |
| anteaya | I don't think we are at ml thread yet | 15:27 |
| fungi | cool, i'll put something under the task tracker effort in the agenda | 15:28 |
| anteaya | thank you | 15:28 |
| *** kkup has quit IRC | 15:30 | |
| fungi | i think it should probably be a warning period between tc acceptance and migration date, as that's probably a suitable proxy for having the identified features implemented, and we'll want to do a pretty big and far-reaching advance notice announcement anyway | 15:32 |
| anteaya | agreed | 15:37 |
| anteaya | also here is a gerrit topic for ipsilon: https://review.openstack.org/#/q/topic:ipsilon | 15:37 |
| anteaya | ipsilon is the openid provider we are moving to, and which the task-tracking spec says the task-tracker has to be using | 15:38 |
| anteaya | so for interest purposes | 15:38 |
| anteaya | also for you to review as you see fit | 15:38 |
| anteaya | fungi: did I say that correctly? | 15:40 |
| fungi | sort of | 15:43 |
| fungi | ipsilon is an openid provider we are evaluating, and the task-tracking spec _actually_ says, "Before migrating new users, StoryBoard should be plugged into an OpenStack identity provider (OpenStackID, ipsilon...) and therefore this work should be completed first." http://specs.openstack.org/openstack-infra/infra-specs/specs/task-tracker.html#dependencies | 15:44 |
| anteaya | thank you | 15:45 |
| fungi | we _could_ point it at openstackid's openid implementation first and then switch it to use ipsilon later. there are already several systems (zanata, groups, et cetera) which will need similar treatment and the backend accounts will be consistent between the two openid implementations anyway so shouldn't be a hard transition | 15:46 |
| anteaya | oh okay so if it becomes a blocker, we can change the spec? | 15:47 |
| fungi | the spec doesn't need changing i don't think. it says we can use openstackid or ipsilon or "..." | 15:49 |
| fungi | whatever's available when the time comes | 15:49 |
| fungi | just specifies "an OpenStack identity provider" (meaning some openid backed by the foundation member database) | 15:49 |
| fungi | it was intentionally open-ended so that we wouldn't get tripped up by having to pick a specific openid implementation that wasn't available yet | 15:50 |
| fungi | as a prerequisite for the task tracker work | 15:51 |
| fungi | this way the transition from openstackid to ipsilon (if it comes to pass) goes on in parallel with storyboard migration or even afterward | 15:52 |
| fungi | i should specify, transition from the openid api implementation in openstackid, ipsilon wouldn't be a replacement for all of the identity management in openstackid anyway, just for the openid interface itself | 15:52 |
| fungi | like with pretty much everything else, we need to plan large projects like this so that they can go on asynchronously in parallel, and avoid blocking on one another even if it means slightly more work on one project or another to make that possible. otherwise we end up entirely serialized by a spec dependency graph that can cause us to deadlock progress on everything | 15:55 |
| anteaya | thank you for the clarification, I had missed that this was open ended | 16:18 |
| anteaya | and I agree with asynchronous parallel efforts | 16:19 |
| *** davidlenwell has quit IRC | 16:28 | |
| *** davidlenwell has joined #storyboard | 16:40 | |
| Zara | arghhhhhhhhhhhhhh just spent the last ages trying to get worklists via boards when I wanted a board_id filter on worklists (so a python clienty equivalent of: boards/2/worklists when I wanted worklists/?board_id=2 | 16:48 |
| Zara | grr grr grr, I'm off to sulk, I don't think I have anything to add to backscroll. | 16:49 |
| persia | Could use a URL rewrite function, so that boards/2/worklists would redirect to worklists/?board_id=2 | 16:50 |
| persia | Mind you, that should probably be a convenience function, rather than put in the client by default. | 16:51 |
| persia | (and may your sulk be satisfying and productive) | 16:51 |
| Zara | :) | 16:53 |
| Zara | I hope so too, especially since my day hasn't been. :D | 16:53 |
| *** dmsimard is now known as dmsimard|afk | 17:09 | |
| *** dmsimard|afk is now known as dmsimard | 19:26 | |
| *** kkup has joined #storyboard | 23:54 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!