This version (2017/05/27 13:44) is a draft.
Approvals: 0/1

[00:16:16] <jeremy_prime> cool, I'll take a look at some point tomorrow to see if I can identify what's going on (someone else might spot it before me though)

[09:45:10] <aesteve> cescoffier: thanks for the explanations on Docker, I had to go yesterday but I've read them :)

[09:45:53] <cescoffier> aesteve: no problem, I've discovered it too

[09:48:09] <aesteve> I guess we probably have this question in the Google Group some day. When you're deploying your app to Docker the first question you'll ask yourself is : “How many instances of my verticle should I deploy”

[09:48:19] <aesteve> we'll*

[09:49:38] <tsegismont> hi everyone

[09:49:40] <tsegismont>

[09:49:42] <tsegismont> 404

[09:49:48] <tsegismont> is the doc still alive somewhere?

[09:51:01] <tsegismont> Ha, found it

[09:51:02] <tsegismont>

[09:51:06] <tsegismont> sorry for the noise

[10:11:23] <temporalfox> need to finish this

[10:11:23] <temporalfox> api changed a bit since

[10:11:24] <temporalfox> new stuff

[18:44:44] <bytor99999> Does this sentence in the unit docs make sense? It seems both are saying the same thing to me….

[18:44:46] <bytor99999> The previous examples supposed that test cases were terminated after their respective callbacks, this is the default behavior of a test case callback. Often it is desirable to terminate the test after the test case callback, for instance:

[18:45:03] <bytor99999> Both say after callbacks.

[18:46:13] <bytor99999> I think it is trying to say. In the first sentence that this is the callback..

[18:46:14] <bytor99999> suite.test(“my_test_case”, { context →

[18:46:33] <bytor99999> And the second for the code after

[18:46:36] <bytor99999> { context →

[18:46:36] <bytor99999> def async = context.async()

[18:46:36] <bytor99999> eventBus.consumer(“the-address”, { msg →

[18:46:56] <bytor99999> Being the callback in the eventBus.consumer call, the one starting with { msg →

[18:46:58] <bytor99999> Correct?

[18:48:40] <bytor99999> And my other question regarding async. Is with multiple async calls, are they nested. Meaning you can start them all without the previous one having to be complete() first? Or does the first one have to be completed before you can start a second one?

[18:49:17] <bytor99999> Reason I ask, is I finally got a couple of integration tests to actually pass with real messages through the eventBus to the handler and the response to the test and it have data. YAY!

[18:50:07] <bytor99999> But, following the exact same pattern in other tests fail, mostly saying that a handler isn't registered for the address I am sending to, but I know that there is one.

[18:54:47] <bytor99999> OK, I think I figured out a pattern where I only need to create one async()

[18:55:20] <bytor99999> So next question. Does automatically complete() the current async() for you?

[18:59:42] <AlexLehm> i think the first sentense should say “terminated before their respective callbacks”

[19:02:23] <AlexLehm> you can create more than one async and have them completed in any order, the test will finish when all are completed

[19:02:54] <AlexLehm> fails the complete test, the async objects are not checked then anymore

[19:10:30] <bytor99999> Thanks Alex

[19:20:55] <AlexLehm> if you are getting failures where the handler is not registered it may be a timing problem, this can either be resolved by waiting a few ms before starting the test or having a callback in the setup of the receiver and start the client test in this callback

[19:21:15] <AlexLehm> or set up the receiver in a before message with async

[19:21:20] <AlexLehm> before method

[19:22:26] <bytor99999> Yeah. I am having to wiggle around setting async correctly in all the right places. The first is in deploying the Verticle, to wait till it is deployed. Then create another one that the actual test method sending a message and receiving the reply to complete the second.

[19:30:21] <bytor99999> I don't think putting it in the before method is working. When I have an async started as the last line of the @Before method, it never calls the @Test methods.

[19:41:32] <bytor99999> This is getting a little bit frustrating because it should work. In my base test class I have an static boolean to hold whether the verticle has been deployed before or not then in the @Before check the boolean and if false then deploy the verticle where i first start async() then deploy and when it is finished deploying to then call complete() then in the @Test method I start a new sync() for the test and call complete. But I am getting the sa

[19:41:32] <bytor99999> me test fail because no handler. Even though the Verticle with the handler has been deployed for sure.

[19:51:25] <AlexLehm> does this work if you wait in the test for e.g. 1 second and then run it?

[19:52:01] <AlexLehm> most of the concurrently problems I had resolved with that only then the tests take longs, which is not good

[19:55:07] <bytor99999> I haven't tried with wait, just trying to coordinate async() calls. I am posting to the Google Group, just because I think it has gotten to a bigger discussion that to fill in IRC.

[19:55:13] <bytor99999> But, you know what I miss…

[19:56:02] <bytor99999> I think we need to make our Google Group and IRC channel legit. And to do that, we need to get some of us, (me) needs to get Brocked. Then we can say it is officially a real forum and IRC channel. :D

[20:08:13] <Elisha> yay :D

[20:22:58] <bytor99999> I think I might have gotten my tests to finally work. It has to do with the fact that Each test method runs as a separate instance and therefore state held gets overwritten. So making that state as static. State is the vertx instance as well as a boolean that the verticle is deployed.

[20:40:41] <zerkz> Hmm, odd question , but has anyone ever tried to get Vertx running on Android?

[21:12:19] <diega> hi! Is there some plan or WIP for supporting OAuth in vert-x3/vertx-auth?

[22:53:22] <kontoo> quick question about vert.x: imagine an app that let's clients connect via web sockets and proxies their messages to another server via a custom protocol. In netty you can achieve that quite easily, is that also possible with vert.x?