Differences

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

Link to this comparison view

irc:1434405600 [2017/05/27 13:44] (current)
Line 1: Line 1:
 +[04:38:35] <Jin> .
 +
 +[04:39:09] <Jin> hi
 +
 +[08:41:28] <​quazimodo>​ hullo
 +
 +[08:46:59] <​cescoffier>​ purplefox pmlopes temporalfox : Hi guys, I've fixed the web site build on cloudbees. It was not building the right branch. Looks like metrics and RX has reproducible test failures.
 +
 +[08:47:23] <​temporalfox>​ cescoffier good, I will investigate these :-)
 +
 +[08:47:40] <​cescoffier>​ temporalfox Plane is made for that right ?
 +
 +[08:47:53] <​temporalfox>​ plane is at 7pm
 +
 +[08:48:38] <​cescoffier>​ oh plenty of time :-)
 +
 +[08:48:49] <​temporalfox>​ these errors are reproducible locally ?
 +
 +[08:49:04] <​cescoffier>​ I've seen one on my machine, so I would say yes
 +
 +[08:49:16] <​cescoffier>​ the RX one seems to be a timing issue
 +
 +[08:49:22] <​cescoffier>​ (unfortunately it's in the TCK)
 +
 +[08:50:05] <​cescoffier>​ BTW, yesterday I've made a full build of the aggregator from an empty maven repository
 +
 +[08:50:48] <​temporalfox>​ ah you mean reactive-streams
 +
 +[08:51:00] <​cescoffier>​ hum yes sorry reactive streams
 +
 +[08:51:00] <​temporalfox>​ and not vertx-rx
 +
 +[08:51:04] <​temporalfox>​ not the same :-)
 +
 +[08:53:36] <​cescoffier>​ and ovbiously, I can't reproduce them on my machine right now (both succeed)
 +
 +[08:54:10] <​temporalfox>​ can you show me the metrics failure ?
 +
 +[08:54:36] <​cescoffier>​ ok commuting, back online in 15 minutes
 +
 +[08:54:47] <​temporalfox>​ sure
 +
 +[08:55:13] <​cescoffier>​ https://​vertx.ci.cloudbees.com/​view/​vert.x-3/​job/​vert.x3-dropwizard-metrics/​391/​
 +
 +[09:12:45] <​temporalfox>​ ok I see
 +
 +[09:12:47] <​temporalfox>​ java.io.FileNotFoundException:​ /​i/​am/​not/​here/​missingfile.json (No such file or directory)
 +
 +[09:12:47] <​temporalfox>​ at java.io.FileInputStream.open0(Native Method)
 +
 +[09:12:47] <​temporalfox>​ at java.io.FileInputStream.open(FileInputStream.java:​195)
 +
 +[09:12:47] <​temporalfox>​ at java.io.FileInputStream.<​init>​(FileInputStream.java:​138)
 +
 +[09:12:47] <​temporalfox>​ at java.util.Scanner.<​init>​(Scanner.java:​611)
 +
 +[09:12:47] <​temporalfox>​ at io.vertx.ext.dropwizard.impl.VertxMetricsFactoryImpl.loadOptionsFile(VertxMetricsFactoryImpl.java:​87)
 +
 +[09:12:47] <​temporalfox>​ at io.vertx.ext.dropwizard.impl.VertxMetricsFactoryImpl.metrics(VertxMetricsFactoryImpl.java:​65)
 +
 +[09:12:48] <​temporalfox>​ at io.vertx.ext.dropwizard.impl.VertxMetricFactoryImplTest.testLoadingWithMissingFile(VertxMetricFactoryImplTest.java:​128)
 +
 +[09:13:01] <​temporalfox>​ the test must somehow be OS dependant
 +
 +[09:13:13] <​temporalfox>​ usually when it works on OSX and not on jenkins
 +
 +[09:13:18] <​temporalfox>​ it can be a case problem
 +
 +[09:13:33] <​temporalfox>​ well this file is supposed not to exist
 +
 +[09:13:44] <​temporalfox>​ perhaps the reason is that the failure is not the same in this os
 +
 +[09:16:04] <​temporalfox>​ ah actually it's another test
 +
 +[12:34:02] <​pmlopes>​ purplefox there was a bug report about integrating stormpath with shiro that uncovered a couple of bugs with the shiro implementation,​ I've a PR ready for review that demonstrates that we can now work with shiro and that plugins should also work
 +
 +[14:49:25] <​cescoffier>​ pmlopes are you around ?
 +
 +[14:49:32] <​pmlopes>​ yes
 +
 +[14:49:44] <​cescoffier>​ I've a couple of questions for you
 +
 +[14:50:02] <​cescoffier>​ I'm trying to provide as an example a custom implementation of cluster manager
 +
 +[14:50:09] <​cescoffier>​ I was thinking about using redis
 +
 +[14:50:25] <​cescoffier>​ do you think it would work (we don't need something bullet proof)
 +
 +[14:51:13] <​pmlopes>​ you mean using the pub-sub mode?
 +
 +[14:51:32] <​cescoffier>​ I will definitely need this
 +
 +[14:51:39] <​cescoffier>​ to listen node arrival and departure
 +
 +[14:51:53] <​cescoffier>​ I also need a way to register the current instance in the "​cluster"​
 +
 +[14:52:01] <​pmlopes>​ i think it can work but i'm not familiar with the cluster API
 +
 +[14:52:08] <​cescoffier>​ was thinking mapping this to a simple entry addition to a list
 +
 +[14:54:41] <​cescoffier>​ ok, well, gonna try will tell you
 +
 +[14:55:35] <​pmlopes>​ ok, i am looking at a bug on jdbc client
 +
 +[14:56:41] <​pmlopes>​ all tests here implemented using h2/​hypersonic however other DBs like Oracle/​Postgres/​Derby use more types that h2 such as BigDecimal, RowId etc and all those types are not handled when converting to JSON
 +
 +[14:57:17] <​pmlopes>​ in Derby/​Oracle you can reproduce it just by inserting an autoincrement row, since it returns a BigDecimal for whatever reason
 +
 +[14:57:25] <​pmlopes>​ even if you declare the column as INT
 +
 +[15:00:59] <​LaSombra>​ Hi aesteve
 +
 +[15:01:38] <​LaSombra>​ aesteve: I started investigating that bug and it seems it would require a change on Codegen, which I'm looking into
 +
 +[15:54:13] <​Jotschi>​ Hi
 +
 +[17:21:52] <​Jotschi>​ The AuthService is completely gone in M6?