| elmiko | etoews: have a fun vaca! | 01:02 |
|---|---|---|
| etoews | elmiko: thx! i'll be around tomorrow but after that gonzo | 01:03 |
| elmiko | k, cool | 01:03 |
| elmiko | win 13 | 01:04 |
| *** elmiko is now known as _elmiko | 01:04 | |
| *** bitblt has quit IRC | 01:45 | |
| *** sigmavirus24_awa is now known as sigmavirus24 | 01:49 | |
| *** sigmavirus24 is now known as sigmavirus24_awa | 02:17 | |
| *** woodster_ has quit IRC | 04:41 | |
| *** ig0r_ has joined #openstack-api | 05:52 | |
| *** ig0r__ has quit IRC | 05:55 | |
| *** cdent has joined #openstack-api | 06:39 | |
| *** fzdarsky has joined #openstack-api | 06:42 | |
| *** fzdarsky_ has joined #openstack-api | 06:45 | |
| *** fzdarsky_ has quit IRC | 06:53 | |
| *** fzdarsky has quit IRC | 06:56 | |
| *** fzdarsky has joined #openstack-api | 07:01 | |
| *** lucas-dinner has joined #openstack-api | 08:02 | |
| *** lucas-dinner has quit IRC | 08:03 | |
| *** lucasagomes has joined #openstack-api | 08:03 | |
| *** park_heijlong has quit IRC | 09:14 | |
| *** e0ne has joined #openstack-api | 09:28 | |
| *** fzdarsky has quit IRC | 09:32 | |
| *** e0ne is now known as e0ne_ | 10:24 | |
| *** e0ne_ is now known as e0ne | 10:31 | |
| *** e0ne is now known as e0ne_ | 10:51 | |
| openstackgerrit | Alex Xu proposed openstack/api-wg: Clarify the return code when conflicting input parameters https://review.openstack.org/197871 | 11:39 |
| *** e0ne_ is now known as e0ne | 11:50 | |
| *** lucasagomes is now known as lucas-hungry | 11:51 | |
| *** e0ne is now known as e0ne_ | 12:02 | |
| *** e0ne_ is now known as e0ne | 12:09 | |
| *** annegentle has joined #openstack-api | 12:15 | |
| *** e0ne has quit IRC | 12:32 | |
| *** annegentle has quit IRC | 12:38 | |
| *** lucas-hungry is now known as lucasagomes | 12:46 | |
| *** pserebryakov has joined #openstack-api | 12:47 | |
| *** e0ne has joined #openstack-api | 13:00 | |
| *** _elmiko is now known as elmiko | 13:15 | |
| *** fzdarsky has joined #openstack-api | 13:18 | |
| *** terrylhowe has joined #openstack-api | 13:27 | |
| *** sigmavirus24_awa is now known as sigmavirus24 | 13:47 | |
| elmiko | sigmavirus24: hey, do you happen to know which rfc talks about query parameters for http? | 13:50 |
| sigmavirus24 | elmiko: well rfc3986 discusses query parameters for URIs | 13:50 |
| sigmavirus24 | 3987 is IRIs | 13:50 |
| sigmavirus24 | URLs are URIs but a restricted subset | 13:50 |
| elmiko | excellent, thanks | 13:51 |
| sigmavirus24 | I would presume you're looking for 3986 because I don't recall if the 7230-7235 series talk about query params | 13:51 |
| elmiko | couple questions came up recently about using query params on POST and PUT, just curious to read more | 13:51 |
| sigmavirus24 | 2616 may have but having standardized that elsewhere, it makes sense they would have omitted it | 13:51 |
| sigmavirus24 | elmiko: ah, if you're looking for semantics then you are looking for something in the 7230-7235 series | 13:51 |
| sigmavirus24 | I forget which one though | 13:51 |
| elmiko | ok, still, you've reduced my search parameters =) | 13:52 |
| sigmavirus24 | no problem | 13:53 |
| sigmavirus24 | I need to finish writing up that list for the API wg | 13:53 |
| sigmavirus24 | with summaries | 13:53 |
| sigmavirus24 | so its easier to look at the topics covered and say "I think it should be in this one" | 13:53 |
| elmiko | ooh, that would be nice | 13:53 |
| elmiko | yea | 13:53 |
| sigmavirus24 | it's a work-in-progress review | 13:53 |
| *** pserebryakov has quit IRC | 14:00 | |
| *** annegentle has joined #openstack-api | 14:12 | |
| *** notmars has joined #openstack-api | 14:53 | |
| *** annegentle has quit IRC | 14:58 | |
| *** e0ne is now known as e0ne_ | 15:32 | |
| *** e0ne_ is now known as e0ne | 15:48 | |
| *** notmars has quit IRC | 16:03 | |
| *** bitblt has joined #openstack-api | 16:03 | |
| *** notmars has joined #openstack-api | 16:08 | |
| *** notmars_ has joined #openstack-api | 16:11 | |
| cdent | does the working group have a position on clients that do not configure themselves by asking keystone for an endpoint? | 16:12 |
| cdent | that is, do we assume that that is the only legitimate way to find an endpoint? | 16:13 |
| *** notmars has quit IRC | 16:13 | |
| *** notmars_ is now known as notmars | 16:13 | |
| *** terrylhowe has left #openstack-api | 16:13 | |
| ryansb | I don't know that we have an official stance as a working group | 16:20 |
| ryansb | but I'm pretty sure the general OpenStack concensus is "always use keystone, for it is the Source of Discovery Truth" | 16:21 |
| cdent | yeah, that's my understanding too | 16:21 |
| cdent | but I also want a world where I can use curl :) | 16:21 |
| cdent | (easily) | 16:21 |
| ryansb | yeah, you use curl to hit up the service catalog | 16:22 |
| ryansb | then use curl to hit up the endpoint you got back | 16:22 |
| cdent | mleh | 16:22 |
| cdent | :) | 16:23 |
| ryansb | yeah, I know | 16:23 |
| stevelle | "easy" | 16:24 |
| cdent | Are browser-based clients presumed to be doing the same kinds of lookups? | 16:27 |
| cdent | Or is there an expectation that the surrounding code which delivers the javascript to the browser wil have already done endpoint lookups? | 16:27 |
| cdent | Do other people think about these things or have I been at the koolaid for too long? | 16:28 |
| ryansb | I'd expect the JS code to do that lookup | 16:31 |
| *** notmars has quit IRC | 16:31 | |
| ryansb | but I don't see why you couldn't send the lookup results along with the JS code | 16:31 |
| stevelle | ditto. service catalog should be cached client-side | 16:31 |
| *** annegentle has joined #openstack-api | 16:35 | |
| *** e0ne has quit IRC | 16:42 | |
| *** bitblt has quit IRC | 16:45 | |
| elmiko | advising the use of keystone's service catalog seems a little out of scope for the wg, imo | 16:46 |
| elmiko | i mean, sure, it's good practice | 16:46 |
| *** annegentle has quit IRC | 16:46 | |
| *** Apoorva_ has joined #openstack-api | 16:50 | |
| *** annegentle has joined #openstack-api | 16:50 | |
| cdent | elmiko: having a position on and making a guideline about is not quite the same thing :) | 16:51 |
| cdent | we're in IRC, it's possible to have a position on everything! | 16:51 |
| elmiko | lol, good point | 16:51 |
| elmiko | my position would be use the service catalog, unless you have a compelling reason not to | 16:52 |
| *** lucasagomes is now known as lucas-beer | 16:53 | |
| sigmavirus24 | so | 16:54 |
| sigmavirus24 | I agree with elmiko that client work is a bit out of our scope | 16:54 |
| sigmavirus24 | We care about the APIs not the client best practices | 16:54 |
| sigmavirus24 | That said, keystone and other services behind an F5 are still a bit ... broken | 16:54 |
| sigmavirus24 | So relying solely on that has shown to be problematic without other ways of dealing with it. (See also recent cinderclient discussion on the ML and subsequent patch revert) | 16:55 |
| sigmavirus24 | Also, I'm certain Swift would disagree given how much work they put into being able to be deployed sans *any* other OpenStack services | 16:55 |
| elmiko | sigmavirus24: imo, that would fall under "compelling reason not to" | 16:57 |
| sigmavirus24 | Which? | 16:57 |
| elmiko | like, you are only using swift and now for sure where the endpoint is | 16:57 |
| elmiko | *know | 16:57 |
| sigmavirus24 | Yeah sure | 16:57 |
| sigmavirus24 | Or you have a load balancer so you only ever need to know one endpoint ever | 16:58 |
| * notmyname lurks | 16:58 | |
| elmiko | right, that too | 16:59 |
| elmiko | but again, following cdent's lead, this is just opinion fodder and i don't think is worthy, or even appropriate, for guideline material | 17:00 |
| cdent | It's in my consciousness because I've been doing devstack work to make services run on the same host and port with prefixes | 17:02 |
| elmiko | interesting | 17:03 |
| cdent | which according to that gang is the desired default | 17:03 |
| cdent | api.openstack.example.com/{compute,network,identify} etc | 17:04 |
| elmiko | that's not the default installation expectation though, is it? | 17:05 |
| cdent | Is there such as a thing as a default installation? I don't know | 17:06 |
| cdent | but from a human usability standpoint a portless single endpoint rather makes sense (whether it is by way of a series of proxies or direct)? | 17:06 |
| elmiko | by default i meant more what is in the offical install/ops guide | 17:07 |
| elmiko | guides | 17:07 |
| cdent | I don't think it is now, because it hasn't been easy to do with the eventlet-based api | 17:07 |
| cdent | (unless you used a proxy) | 17:08 |
| cdent | but in the move to "prefer mod wsgi" then it is becomes more straightforward | 17:08 |
| elmiko | right | 17:08 |
| *** notmars has joined #openstack-api | 17:15 | |
| *** annegentle has quit IRC | 17:16 | |
| *** salv-orl_ has quit IRC | 17:20 | |
| *** salv-orlando has joined #openstack-api | 17:22 | |
| *** cdent has quit IRC | 17:39 | |
| *** annegentle has joined #openstack-api | 17:41 | |
| *** e0ne has joined #openstack-api | 17:51 | |
| *** salv-orlando has quit IRC | 17:55 | |
| *** annegentle has quit IRC | 18:00 | |
| *** notmars has quit IRC | 18:10 | |
| *** notmars has joined #openstack-api | 18:10 | |
| *** annegentle has joined #openstack-api | 18:19 | |
| *** annegentle has quit IRC | 18:20 | |
| *** notmars has quit IRC | 19:24 | |
| *** notmars has joined #openstack-api | 19:31 | |
| *** annegentle has joined #openstack-api | 20:04 | |
| *** fzdarsky has quit IRC | 20:16 | |
| *** notmars has quit IRC | 20:17 | |
| *** salv-orlando has joined #openstack-api | 20:19 | |
| *** fzdarsky has joined #openstack-api | 20:31 | |
| *** e0ne has quit IRC | 20:42 | |
| *** annegentle has quit IRC | 20:56 | |
| *** sigmavirus24 is now known as emacsraptor | 21:10 | |
| *** fzdarsky has quit IRC | 21:10 | |
| *** emacsraptor is now known as sigmavirus24 | 21:10 | |
| *** notmars has joined #openstack-api | 21:21 | |
| *** notmars has quit IRC | 21:55 | |
| *** sigmavirus24 is now known as sigmavirus24_awa | 21:59 | |
| *** lucas-beer has quit IRC | 22:20 | |
| *** salv-orlando has quit IRC | 23:18 | |
| *** openstack has joined #openstack-api | 23:39 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!