00:01:02 #startmeeting api wg 00:01:03 Meeting started Thu Jun 25 00:01:02 2015 UTC and is due to finish in 60 minutes. The chair is elmiko. Information about MeetBot at http://wiki.debian.org/MeetBot. 00:01:04 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 00:01:06 The meeting name has been set to 'api_wg' 00:01:15 anyone here? 00:02:45 \o 00:02:57 if it's just us, we can skip this week 00:03:06 yea 00:03:16 i do appreciate the re-review on the 5xx stuff, thank you =) 00:06:10 np, anything else dying for review? 00:06:21 I think I hit all the frozen ones. 00:06:23 hmm 00:06:37 i'm really curious to hear what people think about the 1xx advice, if we need it or not. 00:06:51 i had an extended conversation with notmyname and lifeless about it 00:07:15 and basically our advice is, use it if you need to, but be sure you read the rfc and the wsgi docs. 00:07:25 which kinda does seem like non-advice. i dunno 00:07:53 i think i could re-write the guidance, but i could use some.... guidance ;) 00:08:28 is it something we see people making mistakes on? 00:08:33 Yeah, I'd probably say it's not necessary guidance 00:08:33 HTTP is HUGE 00:08:42 writing a shadow RFC is going to consume you :) 00:08:44 since afaik nobody has asked about it 00:08:49 lifeless: lol 00:08:59 I wrote up a spreadsheet once 00:09:02 it was added in the spirit of completeness, so that we had something for each category of response 00:09:11 of all the MUST, MUST NOT, etc etc etc in HTTP/1.1 00:09:17 and? 00:09:22 hundreds of rows 00:09:27 lol ouch 00:10:17 o.O 00:10:30 my question for the group is, is it worth it to add a small paragraph or two describing that mostly the 1xx series doesn't need to be used manually, but that 100 should get special attention and to be carefule when using it (paraphrase)? 00:11:02 I think it got lost when the squid-cache server had a crash 00:11:41 fair, but no more than 2. 00:12:37 "two paragraphs only Mr. Vasily" ;) 00:12:46 lol 00:13:04 ahhhttp://web.archive.org/web/20041204194210/http://devel.squid-cache.org/squidhttp1.1.htm 00:13:07 http://web.archive.org/web/20041204194210/http://devel.squid-cache.org/squidhttp1.1.htm 00:13:17 gotta make sure we don't alert the americans. :) 00:13:27 hehe 00:13:38 so just under 500 distinct requirement 00:13:50 lifeless: impressive 00:13:55 lifeless: goodness... 00:14:00 yeah 00:14:11 so- I think we should provide advice on things people get wrong 00:14:13 not on the spec 00:14:22 cause - bottomless pit of despair 00:14:30 we've had 0 occurences of this being wrong :) 00:15:01 sigmavirus24: oh yeah - you'll likely find ^^^ entertaining, speaking of RFC-heads :) 00:15:16 so, 1 for yes, 1 for no... ಠ_ಠ 00:15:24 oh right 00:15:29 Forgot about this meeting 00:15:29 sorry 00:15:33 np 00:16:39 sigmavirus24: break the tie! 00:16:51 * sigmavirus24 is reading the scrollback 00:16:56 * sigmavirus24 wonders where miguelgrinberg and stevelle are 00:18:06 Yeah I was worried we were reiterating in shorter, less clear (to some) language what the HTTP specs say 00:18:45 I'm not sure how much of HTTP the protocol we need to cover. Error codes make sense, but I'm not certain we need guidance for every class 00:18:56 I doubt anyone is misusing 301/302/307/308 for redirects 00:19:17 And I doubt their misusing permanent redirects otherwise requests would have already been yelled at on the openstack-dev mailing list again 00:19:31 ok, fair. i'll abandon the 1xx review 00:19:45 Sorry for all that hassle elmiko 00:20:06 * sigmavirus24 goes back to hacking on cryptography 00:20:08 no problem, i learned a ton about 100-continue thatnks in no small part to lifeless 00:20:44 the only other question i had for the group, was about moving this guideline into freeze 00:20:48 #link https://review.openstack.org/#/c/180094 00:20:59 any thoughts? (seems like we have good approval so far) 00:21:32 +1, freeze it 00:22:13 F-f-f-f-f-f-freeze it 00:22:22 hehe, ok. will do 00:22:42 #action elmiko to freeze and notify about https://review.openstack.org/#/c/180094 00:23:22 well, if no one else has a topic i suppose we can close this out and take back the 35 minutes left =) 00:24:04 woot 00:24:10 more time to run cryptography's tests =P 00:24:15 all 140k+ of them 00:24:15 lol 00:24:20 thanks everybody! 00:24:24 lucky you 00:24:26 #endmeeting