Tuesday, 2015-06-09

« Monday, 2015-06-08 Index Wednesday, 2015-06-10 »
*** Apoorva has quit IRC00:00
*** Apoorva has joined #openstack-api00:27
*** sigmavirus24 is now known as sigmavirus24_awa01:06
*** sigmavirus24_awa is now known as sigmavirus2401:20
*** Apoorva has quit IRC01:54
*** Apoorva has joined #openstack-api02:51
*** Apoorva has quit IRC02:57
HenryGelmiko: the plan is indeed to use microversions, but new features cannot be blocked waiting for that to happen. So new feature APIs will need to start off as extensions until microversioning is established.02:59
*** ameade has quit IRC03:06
*** briancurtin has quit IRC03:06
*** blmartin has joined #openstack-api03:35
*** sigmavirus24 is now known as sigmavirus24_awa03:40
*** Apoorva has joined #openstack-api03:50
*** pballand has joined #openstack-api04:33
*** pballand has quit IRC04:52
*** blmartin has quit IRC04:54
*** e0ne has joined #openstack-api05:29
*** Apoorva has quit IRC05:38
*** e0ne has quit IRC05:47
*** alex_klimov has joined #openstack-api06:22
*** ameade has joined #openstack-api06:41
*** xylan_kong has quit IRC06:53
*** woodster_ has quit IRC06:53
*** briancurtin has joined #openstack-api07:09
*** woodster_ has joined #openstack-api07:34
*** subscope has joined #openstack-api07:36
*** fzdarsky has joined #openstack-api07:44
*** cdent has joined #openstack-api07:54
*** xylan_kong has joined #openstack-api08:05
*** lucasagomes has joined #openstack-api08:06
*** e0ne has joined #openstack-api08:46
*** woodster_ has quit IRC09:41
*** e0ne is now known as e0ne_10:02
*** e0ne_ has quit IRC10:12
*** e0ne has joined #openstack-api10:21
*** e0ne is now known as e0ne_11:34
*** e0ne_ has quit IRC11:40
*** woodster_ has joined #openstack-api11:48
*** lucasagomes is now known as lucas-hungry12:05
*** e0ne has joined #openstack-api12:14
*** e0ne is now known as e0ne_12:21
*** e0ne_ is now known as e0ne12:21
*** e0ne is now known as e0ne_12:30
*** e0ne_ has quit IRC12:35
*** sigmavirus24_awa is now known as sigmavirus2412:41
*** e0ne has joined #openstack-api12:55
elmikoHenryG: ack, thanks13:05
HenryGelmiko: also, if I understand salv-orlando correctly, it looks like neutron will still need to support extensions, even after getting microversioning implemented13:06
openstackgerritSean Dague proposed openstack/api-wg: http guideline expansion: background  https://review.openstack.org/18193113:07
openstackgerritSean Dague proposed openstack/api-wg: add subsection around caching behavior and http  https://review.openstack.org/18352313:07
openstackgerritSean Dague proposed openstack/api-wg: add section describing 501 common mistake  https://review.openstack.org/18345613:07
sdagueok, I think I cleaned up all the typos13:07
elmikosdague: cool13:08
elmikoHenryG: ok, i was just curious. i thought i had heard neutron was moving away from extensions.13:08
*** lucas-hungry is now known as lucasagomes13:08
sdaguehmmm... I can't seem to get tox -e docs to run locally13:47
sigmavirus24sdague: oh?13:47
sigmavirus24what do you see/13:47
sigmavirus24*?13:47
sdagueoh, there was a .# file hanging out13:47
sigmavirus24emacs?13:48
sdagueok, i need to redo these rfc markup bits, I didn't quite get out they render13:48
sdagueyeh13:48
sigmavirus24EMACSSSSS13:48
sigmavirus24just use vim13:49
sigmavirus24;)13:49
sdaguethen I wouldn't have awesome like this - https://dague.net/2015/03/24/openstack-emacs-tools/13:49
sigmavirus24flyhack == syntastic13:52
sigmavirus24pycscope == ctags13:53
sigmavirus24Yeah, multi-mode stuff in emacs is the best way of doing it13:53
sigmavirus24That said, with the way Vim parses documents, you can do that too, it just takes a little extra magic13:53
sigmavirus24Oh yeah, lambda-mode is something we can't get in vim13:53
sigmavirus24My friend had a problem writing clojure with emacs live because of indentation and the fact that with the λ the indentation was correct, but on disk it wasn't13:54
sigmavirus24so he had to turn that off because it was eating up too much of his time13:54
elmikosigmavirus24 you must have not gotten enough argumentation yesterday, starting a emacs->vim conversion this early in the morning... ;)13:55
sdagueheh13:56
sigmavirus24elmiko: I just like teasing people13:56
elmikook, well that's a given13:56
sigmavirus24elmiko: I am very close to saying that "I respect your opinion, but it doesn't make your argument right"13:56
elmikoha!13:56
*** blmartin has joined #openstack-api14:03
sigmavirus24<Movie trailer voice> 'In a world where "I disagree" invalidates all other arguments, yaks get a new shed every minute' </Movie trailer voice>14:03
elmikovery nice, i would actually go see that movie14:04
sdagueheh14:04
* sigmavirus24 is rather glib today. I shouldn't be doing any reviews14:09
sigmavirus24If only I was glibc, then I'd be timeless =P14:10
elmikobut what about eglibc?14:12
openstackgerritSean Dague proposed openstack/api-wg: http guideline expansion: background  https://review.openstack.org/18193114:18
openstackgerritSean Dague proposed openstack/api-wg: add subsection around caching behavior and http  https://review.openstack.org/18352314:18
openstackgerritSean Dague proposed openstack/api-wg: add section describing 501 common mistake  https://review.openstack.org/18345614:18
sdagueok, that should fix the rfc tags, I think14:18
elmikocool, thanks14:19
elmikoi need some advice on https://review.openstack.org/#/c/183698/14:19
sigmavirus24:rfc:`rfc`14:19
elmikodstanek has some valid complaints about the language, but i'm not sure what i could say instead of "HTTP framework", any suggestions?14:20
openstackgerritSean Dague proposed openstack/api-wg: add section describing method use in http  https://review.openstack.org/18973814:22
sigmavirus24elmiko: responded14:22
elmikothanks sigmavirus2414:22
sigmavirus24elmiko: we probably need an appendix14:22
sigmavirus24I guess Glossary is more appropriate14:22
elmikoyea, maybe lol14:22
sdaguecdent: https://review.openstack.org/189738 is an attempt to sort out that back and forth you were having with kenichi via a table, as I thought it might explain things a bit better14:23
sigmavirus24"Request" the HTTP message sent to the server by the client14:23
sigmavirus24"Client" Person who is making an HTTP Request to the server14:23
sigmavirus24"Person" ...14:23
sigmavirus24=P14:23
elmikovery nice ;)14:23
elmikowould be kickass if the whole glossary was basically a linked-list14:24
sigmavirus24Sphinx's refs can give you links to every other thing =P14:26
elmikoi didn't realize you could use the rfc directive to link to sections, that's cool14:28
sigmavirus24yeah but it produces ugly output14:28
sigmavirus24=/14:28
sigmavirus24bitprophet and I were talking about overriding that directive to produce nicer output14:29
sigmavirus24But we never did14:29
elmikonow, i'm curious14:29
sigmavirus24Pretty sure the syntax is :rfc:`7231#3.3` or something, right? That then renders to RFC 7231#3.314:29
sigmavirus24And you can link to pages with :rfc:`7231p33` which renders to RFC 7231p3314:30
sigmavirus24like I said, ugly =P14:30
elmikoyea, hmm14:30
elmikothe way that sdague did it in https://review.openstack.org/#/c/183523 looks reasonable on output14:31
elmiko:rfc:`7231#section-6.1` just produces RFC 723114:32
*** subscope has quit IRC14:34
sdaguesigmavirus24: yeh, the output doesn't tell you it's linking to a section, so I added it in parens after for clarity14:34
elmikosdague: +1, lgtm14:34
sdagueoh... wait, so upstream it is rendering different then on my box14:35
sdagueok, so I'm going to say just grr, we'll typo scrub this stuff later14:36
sdaguebut that's annoying14:36
elmikowait, how does it render on your machine?14:36
sdagueso... on my machine14:37
sdague:rfc:`7231#section-6.1` => RFC 723114:37
elmikoi get the same14:37
sdaguehowever, apparently in the docs upstream14:37
sdague:rfc:`7231#section-6.1` => RFC 7231 Section 6.114:37
elmikouh-oh...14:37
* elmiko facepalm14:37
sdagueno idea why14:37
sigmavirus24heh14:38
sigmavirus24let me look14:38
sdagueanyway, we should just address that later I think14:38
sigmavirus24maybe they did make that directive look better14:38
* sigmavirus24 shrugs14:38
sigmavirus24sdague: what version of sphinx are you running?14:38
sdaguehttp://paste.openstack.org/show/277916/14:40
sdagueanyway, we should probably focus on content, then just do a format scrub later, or we're going to be running around in circles on format for a bit I think14:41
elmikoyea, makes sense14:43
*** terrylhowe has joined #openstack-api14:44
cdentthanks for 189738 sdague, especially the point about it not being 1:1 with CRUD14:44
cdentI'll give it a proper review later, my stack is all confused at the moment14:45
sdagueyeh, no problem. It could probably use better language, but that was one way I found in trying to describe it.14:46
sigmavirus24sdague: http://docs-draft.openstack.org/23/183523/8/check/gate-api-wg-docs/f8aa626//doc/build/html/guidelines/http.html#http-caching-and-proxy-behavior14:47
sigmavirus24am I missing something?14:47
sdaguehuh14:48
sigmavirus24the way RFC is formatted14:48
sigmavirus24er :rfc:14:48
sigmavirus24seems to be working as you expected14:48
sdagueso look at the previous revision - http://docs-draft.openstack.org/23/183523/7/check/gate-api-wg-docs/f736a39//doc/build/html/guidelines/http.html14:48
sigmavirus24Ah14:49
sdagueoh, never mind, that was the manual links14:49
elmikohehe14:49
sdagueI got myself confused on which was where14:49
sdagueok, sorry about that. Too many patch revisions this morning14:49
sigmavirus24that's fine14:49
*** salv-orl_ has joined #openstack-api14:51
*** salv-orlando has quit IRC14:52
salv-orl_elmiko: the bottom line of the story between neutron and versioning is as follows: neutron will introduce versioning in a phased fashion, because the developer community wants to have a proof that microversioning can be successful. Even then, plugin-specific extensions would be loaded outside of any versioning mechanism and won't be considered part of the official neutron API14:59
salv-orl_elmiko: there are plenty of commercial and open source backends integrating with neutron15:00
elmikosalv-orl_: gotcha, it makes sense. there is a great deal of supported code in neutron, no reason to throw out the baby with the bath water ;)15:01
sdaguesalv-orl_: long term, have you thought about moving from plugins to separate end points?15:02
sdagueI was thinking about if/when a scheduler api exposes in nova, starting with a second endpoint might give us the most options to eventually making it independent code15:02
sdagueand I wonder if that's a pattern which might make sense in neutron (it might not, but worth a quick thought)15:03
salv-orl_sdague: it is my understanding that additional services (load balancing, firewall, vpn, etc.) are moving in this direction. There are pro and cons, but I can definetely seem more the pros than the cons15:05
salv-orl_sdague: however, for plugins providing L2/L3 networking functionality, they are conceptually similar to nova drivers, so they can hardly have their own endpoint15:06
*** salv-orl_ has quit IRC15:06
*** e0ne is now known as e0ne_15:08
*** salv-orlando has joined #openstack-api15:08
*** Apoorva has joined #openstack-api15:12
sdaguesalv-orlando: yeh, makes total sense15:16
sdaguewill be great once the advanced services are their own endpoints15:16
*** annegentle has joined #openstack-api15:16
*** annegentle has quit IRC15:18
*** e0ne_ has quit IRC15:18
*** e0ne has joined #openstack-api15:23
sigmavirus24Unrelated, but I'm still super mad about the datetime format used in HTTP (https://tools.ietf.org/html/rfc7231#section-7.1.1.2)15:24
*** cdent has quit IRC16:01
*** e0ne is now known as e0ne_16:03
*** e0ne_ is now known as e0ne16:09
*** alex_klimov has quit IRC16:12
*** annegentle has joined #openstack-api16:25
*** salv-orlando has quit IRC16:33
*** e0ne has quit IRC16:38
*** elmiko is now known as _elmiko16:41
*** pballand has joined #openstack-api16:44
sdaguesigmavirus24: yeh, it's such a weird freaking format16:47
sigmavirus24sdague: thank netscape16:47
sigmavirus24if memory serves me well, it's netscapes fault16:47
sigmavirus24it became a defacto standard for cookies and utterly breaks the ability to join all headers with ','s16:47
* sigmavirus24 gnashes teeth16:47
*** salv-orlando has joined #openstack-api17:09
*** lucasagomes is now known as lucas-dinner17:10
*** Apoorva has quit IRC17:14
*** e0ne has joined #openstack-api17:24
*** Apoorva has joined #openstack-api17:33
*** _elmiko is now known as elmiko17:52
*** annegentle has quit IRC18:30
*** fzdarsky has quit IRC18:35
*** blmartin has quit IRC18:39
*** openstackgerrit has quit IRC18:48
*** openstackgerrit has joined #openstack-api18:48
*** annegentle has joined #openstack-api19:09
*** annegentle has quit IRC19:14
*** e0ne has quit IRC19:15
*** annegentle has joined #openstack-api19:30
*** alex_klimov has joined #openstack-api20:12
*** annegentle has quit IRC21:00
*** annegentle has joined #openstack-api21:10
*** annegentle has quit IRC21:21
*** annegentle has joined #openstack-api21:22
*** pballand has quit IRC21:28
*** pballand has joined #openstack-api21:31
*** annegentle has quit IRC22:08
*** pballand has quit IRC22:17
*** pballand has joined #openstack-api22:27
*** pballand has quit IRC23:26
*** lucas-dinner has quit IRC23:54
« Monday, 2015-06-08 Index Wednesday, 2015-06-10 »

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!