greptilian logo

IRC log for #rest, 2016-04-23

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
00:17 quimrstorres joined #rest
02:51 fuzzyhorns joined #rest
05:17 quimrstorres joined #rest
07:02 fuzzyhorns joined #rest
08:02 fuzzyhorns joined #rest
08:25 quimrsto_ joined #rest
08:45 gigo1980________ joined #rest
09:03 fuzzyhorns joined #rest
09:14 quimrstorres joined #rest
09:46 gigo1980________ joined #rest
10:04 fuzzyhorns joined #rest
11:17 quimrsto_ joined #rest
11:47 gigo1980________ joined #rest
12:00 gigo1980________ joined #rest
12:05 fuzzyhorns joined #rest
12:51 gigo1980________ joined #rest
13:06 fuzzyhorns joined #rest
14:35 wsiqueir joined #rest
15:05 huckleberry78 joined #rest
15:35 fuzzyhorns joined #rest
16:17 quimrstorres joined #rest
17:19 huckleberry78 joined #rest
18:47 fuzzyhorns joined #rest
18:54 huckleberry78 joined #rest
18:56 Coldblackice_ joined #rest
19:31 quimrsto_ joined #rest
19:54 huckleberry78 joined #rest
20:20 fuzzyhorns joined #rest
20:22 ozette joined #rest
20:23 ozette anyone using postman to test their web rest api by any chance?
20:23 fuzzyhorns ozette: i have done a bit of that
20:26 ozette fuzzyhorns: do you know if the form-data encodes files as "multipart"?
20:27 fuzzyhorns i do not
20:27 ozette i'm trying to do a POST here, but end up with no confirmation at the backen
20:28 ozette i'm not sure if the problem is with my route, or if it's postman..
20:28 fuzzyhorns have you tried anything besides postman? like a curl?
20:49 ozette not yet, that's an idea
21:02 pdurbin ozette: eventually, you'll probably want to write a test suite for your API. I kind of doubt you can use postman for that but who knows.
21:12 whartung I do all my testing with curl
21:13 pdurbin whartung: no test suite?
21:14 whartung I've never been much of a tester outside of simple tests during development.
21:15 pdurbin let QA write the API tests, I guess
21:16 whartung I asked the develpoer of our last project to "test first".
21:16 whartung talk about pulling teeth from a chicken
21:16 whartung I wrote up the api spec. I bundled in sample payloads.
21:16 whartung I asked him to not write  line of code until he had tests that failed given the API that I've documented.
21:17 whartung that was a real struggle to get done, and eventually gave up.
21:17 pdurbin bummer
21:17 whartung but in the end, you shouldn't be testing this stuff trhough HTTP anyway. Just test the code
21:17 whartung "something" accepts https packets, and "something else" marshals up the payloads (json in our case).
21:17 whartung so, assumng that works, you have some high level object to pass to you api
21:17 whartung so just test that.
21:18 whartung 99.999% of the time we can assume the HTTP and Marshalling parts work
21:18 pdurbin yeah, I guess
21:18 whartung so why bother standing up the all the infrastructure and such when you can just make some method calls
21:19 pdurbin yeah
21:19 whartung yea, sometime println will fail -- but, it's pretty rare.
21:19 whartung so, lets not bother with the transport part
21:19 whartung because that is really integration testing
21:20 whartung "what if someone sends you a badly formatted payload?" Then the system explodes. Don't Do That. Iintegration problem. Find out who did it, grab the wiffle ball bat and go beat them.
21:20 pdurbin I guess I'm a fan of integration testing. End to end testing. With the database running and all that.
21:20 quimrsto_ joined #rest
21:20 whartung yea -- and that's a pain to do
21:20 whartung especially in a really complicated system
21:21 whartung so, my "integration test" is posting a payload to "http://localhost/app/TestThing"
21:21 pdurbin oh yeah, it kind of sucks writing integration tests. but I'm happy to have them
21:21 whartung and that kicks the whole kit off
21:21 whartung but I'm not auotmating that. "Let's embed teh EJB container. Don't forget to get JMS running. Are you sure you have the timers working?"
21:21 whartung yea. no.
21:22 whartung "no thanks"
21:22 pdurbin whartung: should we be worried about JAX-RS? https://java.net/projects/jax-rs-spec/lists/users/archive/2016-04/message/11
21:24 whartung not saying Burke's wrong, but Burke's a whiner.
21:24 pdurbin heh
21:25 whartung I mean about the JSR treatment.
21:25 pdurbin I don't know him.
21:25 whartung I can't say about his proxy thing
21:25 whartung tacitly I agree it doesn't belong in Jax RS
21:25 whartung I know Jersey has had some Client framework, dunno if the RS guys are trying to standardize that or not.
21:25 whartung I find the idea of a Proxy makes me ithc
21:26 whartung I mean, yay…welcome to 1988 and Sun-RPC…oh, no, sorry, Welcome to 1999 and Remote EJB -- shit, wait no, that's not it…Welcome to 2001 and SOAP-XML…wait…have we learned anything yet?
21:26 whartung damn, I skipped CORBA in 1994
21:27 * whartung is a slacker
21:27 pdurbin :)
21:28 whartung stuff like his proxy just takes "REST" and stomps the boot down harder on its head.
21:29 whartung "Would be incredibly useful, especially as a replacement for remote EJBs.
21:29 whartung Now if there only was some support for security context propagation and
21:29 whartung distributed transactions, but I digress ;)"
21:29 whartung omg
21:31 pdurbin sounds like this proxy thing is a feature in http://resteasy.jboss.org
21:36 whartung gotta run, ttyl pdurbin
21:37 pdurbin o/
22:01 quimrsto_ joined #rest
22:05 huckleberry78 joined #rest
22:09 Coldblackice_ joined #rest
23:42 ozette joined #rest
23:52 quimrstorres joined #rest

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

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