greptilian logo

IRC log for #rest, 2015-10-22

https://trygvis.io/rest-wiki/

| Channels | #rest index | Today | | Search | Google Search | Plain-Text | plain, newest first | summary

All times shown according to UTC.

Time S Nick Message
00:36 fuzzyhorns joined #rest
00:56 zama joined #rest
01:46 fuzzyhorns joined #rest
02:56 baweaver joined #rest
04:02 fuzzyhorns joined #rest
05:21 fuzzyhorns joined #rest
05:31 _ollie joined #rest
06:37 Davey joined #rest
06:37 Davey joined #rest
06:46 fuzzyhorns joined #rest
07:10 interop_madness joined #rest
07:49 graste joined #rest
07:54 trygvis joined #rest
08:11 fuzzyhorns joined #rest
08:34 trygvis joined #rest
09:37 fuzzyhorns joined #rest
10:14 chthon joined #rest
10:21 rosstuck joined #rest
10:25 trygvis http://elektronikknett.no/Artikkelarkiv/2015/Oktober/Mikrokontrollere-for-romfart
10:25 trygvis oops
11:03 fuzzyhorns joined #rest
11:35 mezod joined #rest
12:28 fuzzyhorns joined #rest
13:04 rosstuck joined #rest
13:54 fuzzyhorns joined #rest
14:28 bluezone joined #rest
15:02 trygvis joined #rest
15:04 dkm_ joined #rest
15:05 impl joined #rest
15:10 _longines_ joined #rest
15:11 sfisque joined #rest
15:12 Davey joined #rest
15:12 Davey joined #rest
15:12 locks joined #rest
15:21 trygvis_ joined #rest
15:44 ralphschindler joined #rest
15:50 pith joined #rest
16:54 ralphschindler joined #rest
17:31 simonv3 joined #rest
17:41 anth0ny If I POST to /cameras, a modern browser knows to invalidate the cache on a GET of /cameras, correct?
17:41 asdf sure, lots of stuff do that, including browsers
17:42 anth0ny great, and how about if I POST to /cameras/1234, will the browser be aware that /cameras should have an invalidated cache?
17:42 fumanchu_ any cache should invalidate on POST
17:42 fumanchu_ no. POST <url> tells a cache to invalidate that URL only
17:43 anth0ny obviously, this isn't to speak of EVERY browser in every situation, I just don't know much about how browsers interact with REST apis (things like invalidating the cache)
17:43 asdf this is explicitly said in http spec i think, it's not just a rest guideline
17:44 anth0ny so is there a best practice for saying "hey, I'm updating Camera record 1234, invalidate the cache on list all Camera records"?
17:45 asdf hmm, i don't think there is, i think the browser needs to check with the server next time it fires a get (eg. last-modified, or etag)
17:45 fumanchu_ no, there isn't
17:46 anth0ny great, thanks
17:47 fumanchu_ some folks make the client smart enough to do both. some folks use ETags. some folks use server-side caches which have group invalidation policies. some folks (like me) design the resources to minimize such issues. they all have different limitations.
17:48 asdf how do you design resources to minimize that?
17:48 asdf do you mean eg. not including the /cameras/1234 attributes in the /cameras list, so that on update, the list usually doesnt need to be changed?
17:49 fumanchu_ allow that what your OO code wants to call "an object" actually has data from several different scopes, which should be split up among the container resource (cameras/), each entity (cameras/1234/), and "views" which cross-cut multiple entities
17:50 fumanchu_ don't duplicate data between the container and the entity, which means if you want to alter the data in the container, you POST/PUT/PATCH the container instead
17:51 interop_home joined #rest
17:51 asdf right
17:51 asdf how common are the 'views' in your design? this sounds like something that's only used in uncommon cases?
17:52 fumanchu_ https://bitbucket.org/fumanchu/shoji/src/tip/spec.txt?#spec.txt-658
17:52 asdf oh right, i meant to read that; thanks
17:52 fumanchu_ yes, a "view" is uncommon, and explicitly read-only
18:06 rosstuck joined #rest
19:10 ralphschindler joined #rest
20:27 Coldblackice joined #rest
21:55 fuzzyhorns joined #rest
22:22 ralphschindler joined #rest
23:42 fumanchu joined #rest
23:57 fuzzyhorns joined #rest

| Channels | #rest index | Today | | Search | Google Search | Plain-Text | plain, newest first | summary

https://trygvis.io/rest-wiki/