These are chat archives for akkadotnet/akka.net

20th
Jun 2015
Chris Martin
@trbngr
Jun 20 2015 00:31
@Aaronontheweb Can you look this over for me and tell me how to proceed? https://gist.github.com/trbngr/46dea593e2e180d3bf8e#file-gistfile1-txt-L29
Aaron Stannard
@Aaronontheweb
Jun 20 2015 00:32
@trbngr there's no IActorContext available there?
Chris Martin
@trbngr
Jun 20 2015 00:32
nope
Aaron Stannard
@Aaronontheweb
Jun 20 2015 00:32
which version of Akka.Persistence is this plugin using?
Chris Martin
@trbngr
Jun 20 2015 00:33
103 I believe
1.0.3.0-beta
Aaron Stannard
@Aaronontheweb
Jun 20 2015 00:34
ok, so it's the latest then
Chris Martin
@trbngr
Jun 20 2015 00:34
Not a super huge deal. I'm just trying to avoid the constant polling of the persistence journal.
and use the ES subscirptions
Aaron Stannard
@Aaronontheweb
Jun 20 2015 00:34
the PersistentView is getting replaced is it not cc @Horusiath
so how does EventAppeard get called?
Aaron Stannard
@Aaronontheweb
Jun 20 2015 00:36
do you really care about the Sender in this case?
because that's what is throwing
you could try passing in ActorRefs.NoSender
for the sender on that tell
and that should work
Chris Martin
@trbngr
Jun 20 2015 00:36
oh yeah. lemme try that
HO LEE FUK! It lives! :P
Thanks so much, dude
Aaron Stannard
@Aaronontheweb
Jun 20 2015 00:38
anytime
Sender, Self, Parent all use Context
Chris Martin
@trbngr
Jun 20 2015 00:38
Right on
Aaron Stannard
@Aaronontheweb
Jun 20 2015 00:38
and Context is bound the to ExecutionContext of the actor
so you can't really access Context in an async call or Continuation
can only be called inside the Receive methods
Chris Martin
@trbngr
Jun 20 2015 00:39
So I don't need to capture _self? I can just use Self. Noted
Aaron Stannard
@Aaronontheweb
Jun 20 2015 00:39
no
you do need to capture Self
because that's part of the context too :p
Chris Martin
@trbngr
Jun 20 2015 00:39
well crap. haha
Am I right to capture it in PreStart?
Aaron Stannard
@Aaronontheweb
Jun 20 2015 00:40
yeah
the Context is active there
Chris Martin
@trbngr
Jun 20 2015 00:40
nice. This is pretty bad ass
Aaron Stannard
@Aaronontheweb
Jun 20 2015 00:40
basically anything inside the actor is safe
it's when you start trying to use the Context outside of the actor's running scope where things get weird
Chris Martin
@trbngr
Jun 20 2015 00:41
cool
Chris Martin
@trbngr
Jun 20 2015 01:25
Optimized Akka.Persistence.EventStore. It will now use only one IEventStoreConnection per the ES docs. I also add an implementation of PersistentView called EventStoreSubscriptionView that uses an EventStoreSubscription to stay up to date instead of an PersistentView.AutoUpdateInterval
Still no tests, however ;)
Chris Martin
@trbngr
Jun 20 2015 01:30
Is there a way to grab a ILoggingAdapter from an ActorSystem? I need one pretty early
:blush: system.Log
Thomas Lazar
@thomaslazar
Jun 20 2015 08:28
@Aaronontheweb: i know apache thrift employs or better recommends an extend-only design of their interfaces. to make it backwards compatible to an extend
they have numbered fields in their definitions that you're not supposed to change on edit. and just extend with new numbers
varghesep
@varghesep
Jun 20 2015 16:18
I don't normally get any response for the questions that I post here. They get lost in this chat. So I posted some code on Stack Overflow at http://stackoverflow.com/questions/30956153/service-fabric-akka-net-and-persistent-actor-integration
Please share your comments, improvements needed so everyone can learn from it. The example has the integration of Service Fabric, AKKA.NET and Persistence.
Roger Johansson
@rogeralsing
Jun 20 2015 18:40
@Horusiath under what conditons does akka.persistence need to serialize journal/snapshot with special transport information?
Weston
@ronnyek
Jun 20 2015 23:10
So i've been lookng a bit into project orleans, and it seems like its much much more targetting the idea of basically short lifecycle fire up a grain, change state persist state dismiss
not so much having a distributed pool of workers
so I'm wondering if akka.net is more for that. Seems like its a good way to manage a distributed pool of workers, with managers at certain levels
Error handling etc