Differences

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

Link to this comparison view

irc:1464040800 [2017/05/27 13:44] (current)
Line 1: Line 1:
 +[00:00:49] <​AlexLehm>​ for some reason proxy support is a bad thing in most programming languages, I had to change a perl script recently to use curl or wget since the proxy support for perl was broken in the linux version we had installed
 +
 +[00:07:56] <​AlexLehm>​ ok, squashing the commits gives me a chance to hide such commit messages like "​forgot to add class file xxx"
 +
 +[00:13:04] <​temporalfox>​ we all do such kind of mistakes!
 +
 +[00:19:09] <​AlexLehm>​ i think the guy who wrote Jenkins said his motivation was that he usually forgot to add some files when committing something, so he started to write a tool to check that automatically
 +
 +[00:19:27] <​melvinross>​ if you use setCertPath in HttpSever options, is it expecting a bundle?
 +
 +[00:50:06] <​AlexLehm>​ temporalfox:​ commited the squashed version now https://​github.com/​alexlehm/​vert.x/​tree/​issues/​%23101-socks45-squashed
 +
 +[00:50:25] <​temporalfox>​ ok will have a look tomorrow
 +
 +[00:50:35] <​temporalfox>​ can you make a PR ?
 +
 +[00:50:48] <​temporalfox>​ going to bed :-)
 +
 +[00:51:53] <​AlexLehm>​ ok sure will do
 +
 +[14:28:57] <​kurt-o-sys>​ hey all... design questions, always difficult, but here's one, just fishing for opinions. I've got a few modules now, and one of them is '​validators',​ containing standard validating rules.
 +
 +[14:29:45] <​kurt-o-sys>​ However, that module contains validors for social security numbers, IBAN/BIC (bank account) numbers, and others. I'm thinking of splitting that module...
 +
 +[14:30:19] <​kurt-o-sys>​ but that would result rather a lot of very small modules... Any pros/cons? Or experiences?​ Or opinions?
 +
 +[14:43:54] <​kurt-o-sys>​ vert.x is unopiniated?​ :p
 +
 +[15:36:54] <​melvinross>​ so I've been trying to get SSL working with HttpServer, using the PemKeyCertOptions. But when it gets a connection all I get are exceptions
 +
 +[15:37:01] <​melvinross>​ like this SEVERE: java.lang.IllegalArgumentException:​ invalid version format: <​￯VTG￞;​X￯`,​
 +
 +[15:37:26] <​melvinross>​ and that's using real crts, self signed, etc
 +
 +[15:59:50] <​melvinross>​ hmm no one eh? I've got a valid cert and a pkcs8 key. I'm actually at a loss for what I'm missing
 +
 +[16:39:15] <​LostCoder>​ Is there anything similar to the ReplayingDecoder from netty in vertx? ​ (to parse data on a net socket)
 +
 +[17:21:37] <​ChicagoJohn>​ so i believe your codegen tool needs  a bit more work
 +
 +[17:23:34] <​ChicagoJohn>​ i have found it to be a tad infelxible on its class generation. ​ for example: ​ the Rxified HttpClientRequest contains only one implementation of HttpClientRequest putHeader() ​ which only accepts strings whereas the core HttpClientRequest contains multiple implementations accepting CharSequences
 +
 +[17:24:30] <​ChicagoJohn>​ and considering io.vertx.core.http.HttpHeaders are all CharSequences,​ a .toString is needed. ​ its not the end of the world obviously but again, minor inconveniences.
 +
 +[17:26:02] <​ChicagoJohn>​ Ive also experienced difficulties with the object converters handling to/from Json.  they dont handle super calls to convert parent object attributes
 +
 +[17:46:05] *** tepper.freenode.net sets mode: +o temporalfox
 +
 +[18:06:29] *** tepper.freenode.net sets mode: +o temporalfox
 +
 +[19:34:50] <​ChicagoJohn>​ anyone ever chain async calls using rx flatmap?
 +
 +[20:13:01] *** ChanServ sets mode: +o tempora__
 +
 +[20:23:09] <​ChicagoJohn>​ anyone on?
 +
 +[21:33:28] <​tempora__>​ me
 +
 +[21:36:10] <​tempora__>​ hi AlexLehm
 +
 +[21:36:24] <​AlexLehm>​ tempora__: there is already a package io.vertx.example.core.http.proxy (for a http proxy example impl in vertx) should I put the https proxy example into ...http.https?​
 +
 +[21:36:27] <​AlexLehm>​ hello Juline
 +
 +[21:36:30] <​AlexLehm>​ hello Julien
 +
 +[21:36:40] <​tempora__>​ it looks overally good
 +
 +[21:36:48] <​tempora__>​ however I'm not a fan of ChannelProviderAdditionalOperations
 +
 +[21:36:54] <​tempora__>​ the channel pipeline becomes very complex
 +
 +[21:37:01] <​AlexLehm>​ yes, i was not sure how to do that properly
 +
 +[21:37:49] <​tempora__>​ need to think about it
 +
 +[21:38:42] <​AlexLehm>​ i tried to refactor the two connect classes and wasn't quite sure how to do it (at that point I hadn't quite understood how the pipeline works though)
 +
 +[21:39:40] <​tempora__>​ yep
 +
 +[21:40:57] <​tempora__>​ given that wit hHTTP pipeline setup is quite complex
 +
 +[21:41:00] <​tempora__>​ I don't mind complexity
 +
 +[21:41:04] <​tempora__>​ when it is a single class :-)
 +
 +[21:41:13] <​tempora__>​ but when it start to span several classes
 +
 +[21:41:19] <​tempora__>​ it becomes hard to reason about
 +
 +[21:41:29] <​tempora__>​ (but perhaps it's just me)
 +
 +[21:42:17] <​tempora__>​ one thing I was thinking is that maybe ChannelProvider should be a Netty handler
 +
 +[21:42:26] <​tempora__>​ (but that's not related to the topic)
 +
 +[21:44:02] <​tempora__>​ I will try to spend time on this soon
 +
 +[21:44:08] <​tempora__>​ and work on the branch
 +
 +[21:46:18] <​AlexLehm>​ ok great
 +
 +[21:49:01] <​tempora__>​ mayeb the whole thing should be improved
 +
 +[21:49:16] <​tempora__>​ (I mean also the ConnectionManager setup)
 +
 +[21:49:32] <​tempora__>​ the main issue I thin kis that we mix various cases
 +
 +[21:49:40] <​tempora__>​ and http1 / http2 with fallback
 +
 +[21:49:45] <​tempora__>​ complexifies it
 +
 +[22:20:02] <​AlexLehm>​ tempora__: did you test the class not found check for the netty proxy package? It looks like the exception is happening in the constructor
 +
 +[22:20:20] <​tempora__>​ I don't understand
 +
 +[22:21:58] <​AlexLehm>​ when I use the NetClient with the netty-handler-proxy package missing, I get an exception in the new ProxyChannelProvider call, not on the connect call
 +
 +[22:22:37] <​AlexLehm>​ didn't test that
 +
 +[22:28:02] <​tempora__>​ ok
 +
 +[22:28:13] <​tempora__>​ it should not fail like this
 +
 +[22:37:09] <​AlexLehm>​ i noticed that when writing an example, i will take a look and change the check if necessary later
 +
 +[22:37:56] <​tempora__>​ thanks
 +
 +[23:03:21] *** ChanServ sets mode: +o temporalfox
 +
 +[23:07:17] *** ChanServ sets mode: +o temporal_
 +
 +[23:27:34] <​AlexLehm>​ temporal_: does it make sense if i implement this issue https://​github.com/​vert-x3/​vertx-mail-client/​issues/​67 or is that too short before the release
 +
 +[23:27:50] <​AlexLehm>​ i got somewhat sidetracked by the proxy stuff and forgot about it
 +
 +[23:28:40] <​temporal_>​ if it's a trival thing with 1 or 2 days go for it
 +
 +[23:30:02] <​AlexLehm>​ ok, its a very small thing (one property)
 +
 +[23:58:53] <​ChicagoJohn>​ temporal_ if you are still on i have a quick question
 +
 +[23:59:03] <​temporal_>​ ChicagoJohn yes
 +
 +[23:59:36] <​ChicagoJohn>​ https://​groups.google.com/​forum/#​!topic/​vertx/​ew64orj-Zpo
 +
 +[23:59:40] <​ChicagoJohn>​ here is my write up