These are chat archives for akkadotnet/

Jun 2016
Jun 25 2016 00:27
@Aaronontheweb I am new and am wondering why Newtonsoft was dropped and HOCON adopted? Is it faster?
Jun 25 2016 00:33
You mean Wire in place of JSON.NET? (HOCON isn't directly related to that)
Jun 25 2016 00:36
@qwoz Ah, yes. Wire replacing Newtonsoft. Do you know the reason why it was switched?
Aaron Stannard
Jun 25 2016 00:39
JSON.NET is too stupid to serialize certain types
specifically some F# types and immutable collections
plus Wire is 20-25x faster
and compresses larger messages more efficiently
Jun 25 2016 00:41
So you get pretty much the same as what provided plus much more (including speed)? You have piqued my interest.
Wire was created by akkadotnet for this purpose? Is it primarily for or do you expect usage outside of the project?
Aaron Stannard
Jun 25 2016 01:17
it's primarily for Akka.NET
developed by @rogeralsing to solve some persistent issues we've had serializing F# types
specifically discriminated unions
which has been one of those chronic problems we tried to solve by hacking on top of JSON.NET many times over
over the past couple of years
@alexvaluyskiy lol crap - turns out I never got the CHR specs to pass locally; forgot to update my configuration for the MNTR
working on those now
ok, weird
they did just pass for me locally
but not on the build server
Aaron Stannard
Jun 25 2016 01:27
@alexvaluyskiy looks like the node sorting for the CHR is off
[Node1][FAIL] Akka.Cluster.Tests.MultiNode.Routing.ClusterConsistentHashingRouterMultiNode1.ClusterConsistentHashingRouterSpecs
[Node1][FAIL-EXCEPTION] Type: Xunit.Sdk.EqualException
--> [Node1][FAIL-EXCEPTION] Message: Assert.Equal() Failure
Expected: List<Address> [akka.tcp://MultiNodeClusterSpec@localhost:49447, akka.tcp://MultiNodeClusterSpec@localhost:49444, akka.tcp://MultiNodeClusterSpec@localhost:49443]
Actual:   List<Address> [akka.tcp://MultiNodeClusterSpec@localhost:49447, akka.tcp://MultiNodeClusterSpec@localhost:49443, akka.tcp://MultiNodeClusterSpec@localhost:49444]
either that or the test function is off
let me check
ah, I think I figured it out
this assertion isn't supposed to be sensitive to orderings
but it is
fixed that