Differences

This shows you the differences between two versions of the page.

Link to this comparison view

irc:1466632800 [2017/05/27 13:44] (current)
Line 1: Line 1:
 +[00:00:52] <​temporalfox>​ AlexLehm do you have wireshark ?
 +
 +[00:00:54] <​AlexLehm>​ ah yes sorry
 +
 +[00:00:57] <​AlexLehm>​ yes
 +
 +[00:01:03] <​temporalfox>​ let's focus on
 +
 +[00:01:04] <​temporalfox>​ testSearchDomain
 +
 +[00:01:13] <​temporalfox>​ can you run wireshark on port 53
 +
 +[00:01:24] <​temporalfox>​ actually no
 +
 +[00:01:32] <​AlexLehm>​ https://​gist.github.com/​alexlehm/​b080e3e61e78e011717e89270c0d23b6
 +
 +[00:01:52] <​temporalfox>​ on port 53530
 +
 +[00:02:03] <​temporalfox>​ and tell me the DNS query and answer you see
 +
 +[00:02:24] <​temporalfox>​ so basically you do in wireshark
 +
 +[00:02:26] <​temporalfox>​ filter
 +
 +[00:02:30] <​temporalfox>​ "udp port 53530"
 +
 +[00:02:34] <​temporalfox>​ on localhost
 +
 +[00:02:38] <​temporalfox>​ and in vertx
 +
 +[00:02:50] <​temporalfox>​ mvn test -Dtest=HostnameResolutionTest#​testSearchDomain
 +
 +[00:04:48] <​AlexLehm>​ i think it doesn'​t work to sniff localhost in windows
 +
 +[00:04:50] <​AlexLehm>​ let me try
 +
 +[00:05:32] <​temporalfox>​ ok
 +
 +[00:05:40] <​temporalfox>​ hard to find out then
 +
 +[00:05:52] <​temporalfox>​ so we need to try println
 +
 +[00:05:56] <​temporalfox>​ and / or debug
 +
 +[00:06:28] <​temporalfox>​ we are going to do the println way I think
 +
 +[00:06:32] <​temporalfox>​ if that does not work
 +
 +[00:09:06] <​AlexLehm>​ when i run just one test, it works
 +
 +[00:09:22] <​temporalfox>​ ok
 +
 +[00:09:33] <​temporalfox>​ what if you run two tests ?
 +
 +[00:09:38] <​temporalfox>​ so it's not dramatic
 +
 +[00:09:42] <​temporalfox>​ more like a timing issue
 +
 +[00:09:45] <​temporalfox>​ in the tests
 +
 +[00:10:19] <​temporalfox>​ maybe for tests we should run the DNS server with different ports
 +
 +[00:10:25] <​temporalfox>​ can you try that ?
 +
 +[00:10:34] <​temporalfox>​ in FakeDNSServer
 +
 +[00:10:36] <​temporalfox>​ you change
 +
 +[00:10:40] <​temporalfox> ​  ​public static final int PORT = 53530;
 +
 +[00:10:52] <​temporalfox>​ by something that incrments
 +
 +[00:10:52] <​temporalfox>​ for each test
 +
 +[00:11:02] <​temporalfox>​ like you remove the final keyword
 +
 +[00:11:13] <​temporalfox>​ and in setup() of HostnameResolutionTest
 +
 +[00:11:16] <​temporalfox>​ you increment it
 +
 +[00:11:53] <​AlexLehm>​ when I run  mvn test -Dtest=HostnameResolutionTest#​test*SearchDomain* , it has 2 tests failing
 +
 +[00:12:01] <​AlexLehm>​ let me the port number change
 +
 +[00:12:52] <​temporalfox>​ it may be due to port recycling
 +
 +[00:12:59] <​temporalfox>​ I don't know how windows work
 +
 +[00:13:01] <​temporalfox>​ :-)
 +
 +[00:13:10] <​temporalfox>​ but thing is that this DNS server is not a Netty one
 +
 +[00:20:03] <​AlexLehm>​ with increasing port numbers it works i think
 +
 +[00:20:54] <​AlexLehm>​ running all tests to make sure
 +
 +[00:22:58] <​temporalfox>​ ok
 +
 +[00:35:37] <​AlexLehm>​ ok, the tests all work except for 2 that were failing before (which I haven'​t figured out yet)
 +
 +[00:38:47] <​temporalfox>​ but if you run them individually they pass ?
 +
 +[00:43:46] <​AlexLehm>​ no, they always fail
 +
 +[00:44:31] <​AlexLehm>​ these are in RedeployTest,​ I didn't pay attention to these since they failed for me since time ago
 +
 +[00:44:47] <​AlexLehm>​ so they failure is not related to the dns changes
 +
 +[00:44:54] <​temporalfox>​ ah ok
 +
 +[00:45:06] <​temporalfox>​ but I'm wondering about dns test
 +
 +[00:45:21] <​AlexLehm>​ the dns test works when running all tests in the test class together now
 +
 +[00:45:30] <​temporalfox>​ with the port change
 +
 +[00:45:34] <​AlexLehm>​ yes
 +
 +[00:45:56] <​temporalfox>​ ok
 +
 +[00:46:17] <​temporalfox>​ I've deployed the artifacts in nexus
 +
 +[00:46:21] <​temporalfox>​ and I will close the staging now
 +
 +[00:46:35] <​temporalfox>​ and tomorrow there will be tests with this repo to validate the artifacts
 +
 +[00:46:37] <​temporalfox>​ for 3.3.0
 +
 +[00:49:06] <​AlexLehm>​ ok
 +
 +[00:50:22] <​temporalfox>​ thanks for your precious help tonight
 +
 +[00:50:36] <​temporalfox>​ that helps for sure :-)
 +
 +[00:52:17] <​AlexLehm>​ no problem
 +
 +[00:58:54] <​AlexLehm>​ i wonder if it would make sense to set up a CI running on Windows, there is a hosted CI servers that does that, but I haven'​t used that yet
 +
 +[01:06:33] <​temporalfox>​ yes it would certainly be helpful
 +
 +[01:06:52] <​temporalfox>​ we do already run the examples on windows before a release
 +
 +[01:06:53] <​temporalfox>​ with various langs
 +
 +[01:09:24] <​temporalfox>​ AlexLehm good night
 +
 +[01:09:37] <​AlexLehm>​ good night
 +
 +[01:12:39] *** ChanServ sets mode: +o temporal_
 +
 +[09:42:19] <​noam_>​ Hello, any news on 3.3?
 +
 +[10:00:50] <​cescoffier>​ noam_ : we have stage the final version, we are running the checks
 +
 +[10:02:30] <​noam_>​ Thanks! looking forward to the release
 +
 +[15:14:34] <​pouyanster>​ I have a really interesting problem with HttpClient: when using getAbs with the absolute URL, nothing seems to happen (handler is not invoked) but if I use getNow with explicit port number, host, and path, it seems to be working fine. Tried with both 3.0.0 and 3.2.1. anyone else having this problem or am I using the API in a wrong way?
 +
 +[15:29:27] <​pouyanster>​ my bad! `end()`should be placed after `getAbs` whereas getNow automatically does that! #​readthesource #​addthistodocs
 +
 +[21:09:00] *** ChanServ sets mode: +o temporalfox
 +
 +[21:22:18] *** ChanServ sets mode: +o temporal_