These are chat archives for akkadotnet/akka.net

12th
Aug 2015
Ryan Davis
@rdavisau
Aug 12 2015 00:30
@rogeralsing with regards to the above, what is the magic that allows you to Ask outside of an actor and still receive the response? For example here, if you change L94 to an actor that echos back to the sender, and L102 to an <?, you (happily) will get the response, despite being outside of an actor. I inspected Sender() and saw that it was [akka://cluster/deadLetters]. I guess because of the synchronous processing/ordering guarantee you are able to do something clever there?
Aaron Stannard
@Aaronontheweb
Aug 12 2015 00:50
@bruinbrown is it marked as internal today?
Jonathan Ohlrich
@johlrich
Aug 12 2015 01:16
Thanks for the clarification @rogeralsing, using tell directly and adding the sender argument explicitly made it behave like the others. makes sense
Sean Gilliam
@sean-gilliam
Aug 12 2015 03:25
that case study was awesome https://petabridge.com/blog/akkadotnet-goes-to-wall-street/ :thumbsup:
Aaron Stannard
@Aaronontheweb
Aug 12 2015 05:48
glad you liked it @sean-gilliam
we have some others we're working on too
Aaron Stannard
@Aaronontheweb
Aug 12 2015 06:03
thing I really like about what he wrote
was how he mentioned that he could easily configure all of his session state actors to live somewhere else on the network either
through Akka.Remote or Akka.Cluster in the future
and it wouldn't require a big re-architecture job - it's mostly a configuration detail
Christian Sparre
@christiansparre
Aug 12 2015 06:06
I really like the case studies, have read a few of the Akka JVM ones too.
Aaron Stannard
@Aaronontheweb
Aug 12 2015 06:07
@christiansparre yeah, me too - made it easy for me to see how I might be able to use the actor model to solve real world problems
Christian Sparre
@christiansparre
Aug 12 2015 06:07
I'm wondering if there are any examples of doing "service discovery"?
Aaron Stannard
@Aaronontheweb
Aug 12 2015 06:09
code sample wise, that's what Akka.Cluster can do
and I think @rogeralsing's team is working on an integrated Akka.NET + Service Fabric example that uses some of ServiceFab's discovery tools
Christian Sparre
@christiansparre
Aug 12 2015 06:10
Ok, haven't looked at Akka.Cluster too much yet, will have to do that
Aaron Stannard
@Aaronontheweb
Aug 12 2015 06:10
we just published some new doc pages about it on Friday: http://getakka.net/docs/clustering/cluster-overview
that's a good place to start
Marc Piechura
@marcpiechura
Aug 12 2015 06:12
@Aaronontheweb could you maybe take a short look at my mspec PR? I'm afraid that I messed up the PR with the rebase from upstream :(
Aaron Stannard
@Aaronontheweb
Aug 12 2015 06:12
haha, sure thing @Silv3rcircl3
Marc Piechura
@marcpiechura
Aug 12 2015 06:12
Thx!
Aaron Stannard
@Aaronontheweb
Aug 12 2015 06:12
would it be helpful if I did a video showing how to do all of that Git workflow for rebasing / squashing PRs?
you're not the first person who's asked
Marc Piechura
@marcpiechura
Aug 12 2015 06:13
Yeah that would be awesome, I have only used the build in VS stuff because I never needed more, until now ;)
Christian Sparre
@christiansparre
Aug 12 2015 06:14
I think that would be great @Aaronontheweb. I'm pretty new at git, but love it. Read your contrib guidelines, but a video about it would be awesome!
Aaron Stannard
@Aaronontheweb
Aug 12 2015 06:14
no worries man - I remember when I had to learn this workflow when I sent PRs to someone else's project a few years ago :p
Marc Piechura
@marcpiechura
Aug 12 2015 06:18
@christiansparre do you mean this http://getakka.net/docs/Contributor%20guidelines or are there more?
shows up at the top of each PR when you're in the process of sending it in, which makes it become sort of whitenoise after a while :p
Ryan Davis
@rdavisau
Aug 12 2015 06:20
I have honed my use of rebase to a simple n-step sequence
step 1: copy "git rebase --abort" to clipboard. so you're ready.
step 2: git rebase -i
step 3: ?????????????
ste!?!?!?!?!??!?!?!
step n: paste from clipboard, go to step 1
Marc Piechura
@marcpiechura
Aug 12 2015 06:22
Lol maybe I was to fixed on the green "send pr" button ;)
Aaron Stannard
@Aaronontheweb
Aug 12 2015 06:27
@Silv3rcircl3 see, even I screw it up: akkadotnet/akka.net#1239
accidentally pulled in a bunch of commits to the master branch that never made it back into dev
I guess that's just technical debt for not pulling in those PR commits to master earlier :p
I was able to squash all of the commits for my PR, but not get rid of those ones that exist in the timeline for master
@Silv3rcircl3 do you have some time to Skype tomorrow? it's just about bed time here in the US, but I'll be on in about 8 hours or so. might be easiest if I talk you through how do to the workflow
I'll ping you my Skype handle
Aaron Stannard
@Aaronontheweb
Aug 12 2015 06:40
.... man, why does this StructureMap test keep failing sometimes? akkadotnet/akka.net#1216
none of the other DI implementations have that problem
Marc Piechura
@marcpiechura
Aug 12 2015 06:42
In 8 Hours I'm at home, maybe tomorrow the same time the virtual meetup is today?
It 8:43 in the morning here ;)
Aaron Stannard
@Aaronontheweb
Aug 12 2015 06:43
sure thing
Marc Piechura
@marcpiechura
Aug 12 2015 06:44
Cool! I will ping you on skype tomorrow
Doesn't need to be the computer were I created the PR or?
Ryan Davis
@rdavisau
Aug 12 2015 07:52
@rogeralsing @Horusiath I'm playing with the (unpublished, potentially incomplete) migrant fixes in issue88 branch for the FSI support, all appears to be working so far which is very exciting :+1:
Roger Johansson
@rogeralsing
Aug 12 2015 08:21
nice!
Anthony Brown
@bruinbrown
Aug 12 2015 10:31
@Aaronontheweb Yeah marked as internal, it's exposed by the Scheduler but only as a TimeSpan, which doesn't give us nanos
Steffen Forkmann
@forki
Aug 12 2015 12:06
I created an experimental PR which converts to Paket: akkadotnet/akka.net#1243
Bartosz Sypytkowski
@Horusiath
Aug 12 2015 12:32
@forki great for me! But you've noticed some problems with different versions of xUnit? I thought that paket doesn't have any problems with that
also I'm curious what others will say about that :)
Steffen Forkmann
@forki
Aug 12 2015 13:01
paket usually installs only one version (we don't have version in path).
from here it would only be a small step to allow packages with two versions
but in general I don't think this is a good idea.
welshtom
@welshtom
Aug 12 2015 14:42
Does akka work with .NET micro?
Bartosz Sypytkowski
@Horusiath
Aug 12 2015 16:43
@welshtom probably not
but it still works on embedded devices, we seen it running on Raspberry Pi under mono
Aaron Stannard
@Aaronontheweb
Aug 12 2015 18:14
@/all doing our virtual Akka.NET meetup live via Google Hangouts and YouTube in 15 minutes! https://petabridge.com/blog/akkadotnet-virtual-meetup/ - the broadcast will be recorded and published on our YouTube channel once the livestream is over: https://www.youtube.com/c/PetabridgeAcademy
Arjen Smits
@Danthar
Aug 12 2015 18:31
have we started ? I opened the hangout, but nothing is showing yet. Not sure if its google hangout who's giving me the finger, or something else...
jweimann
@jweimann
Aug 12 2015 18:31
i'm in there too, still silent from what I can tell
Arjen Smits
@Danthar
Aug 12 2015 18:32
yeah you cant really tell how many people are in there. its not even showing a counter or something
the live stream is working
jweimann
@jweimann
Aug 12 2015 18:34
questions are popping up at least :)
Arjen Smits
@Danthar
Aug 12 2015 18:34
ah, its working now. A good ol' ctrl-f5 did the trick
Andrew Skotzko
@skotzko
Aug 12 2015 18:35
@/all we’re live in the meetup everyone! https://www.youtube.com/watch?v=YuY1ziEqifU
Natan Vivo
@nvivo
Aug 12 2015 19:43
FSMs are a good example for sagas
Andrew Skotzko
@skotzko
Aug 12 2015 19:57
sorry for some of the issue / cutouts everyone. Hangouts decided to throw a hissy fit
(we're back btw)
Aaron Stannard
@Aaronontheweb
Aug 12 2015 20:43
@nvivo as for your question about Akka.NET 2.0
we're going to do whatever we need to do to make .NET developers productive as .NET evolves
given how much is changing right now, I imagine the way we write and deploy code is going to be very different in 2016/2017 from 2015
we're not limited to what the JVM does
but we try to follow it because it has years of being battle-tested in really large systems
and we have the benefit of not making the same mistakes they made
but as you heard today, we're starting to get those battle-tested stories too!
the fact that @annymsMthd was able to eliminate 85% of his legacy code in the areas where he used actors is pretty amazing
Natan Vivo
@nvivo
Aug 12 2015 20:49
@Aaronontheweb thanks for the response.
Aaron Stannard
@Aaronontheweb
Aug 12 2015 20:56
anytime man
btw, want to give props for @annymsMthd's presentation... even though he's told me a bunch of that stuff before, I got a lot of new ideas about how to do deployments / design based on what he said
great job Josh!
there was something about the format today that made it easier for me to "get" that than before
in particular the way to do logging with Elastic search and logstash
Bartosz Sypytkowski
@Horusiath
Aug 12 2015 20:57

the fact that @annymsMthd was able to eliminate 85% of his legacy code in the areas where he used actors is pretty amazing

I've quite often reach 100% ration. It's called "rewrite that shit" development :P

it's natural reaction, when a new team comes as the replacement of exiting one and is ordered to maintain their application
Joshua Benjamin
@annymsMthd
Aug 12 2015 21:14
Thanks @Aaronontheweb and thank you guys. It was my first time doing a presentation like this and I was pretty nervous. Thanks for sticking it out
Andrew Skotzko
@skotzko
Aug 12 2015 21:16
you did great @annymsMthd !
Aaron Stannard
@Aaronontheweb
Aug 12 2015 21:17
:+1:
I think you were great - you should do this more often!
Joshua Benjamin
@annymsMthd
Aug 12 2015 21:18
We actually had a vm host go down at work and it made the HA stuff more important than ever. We are now looking at all aspecs of our system to ensure we get rid of all single points of failure
Joshua Benjamin
@annymsMthd
Aug 12 2015 21:27
@Aaronontheweb with what @rogeralsing was talking about with Serilog you could cut out logstash altogether and use Elasticsearch and Kibana
Sean Gilliam
@sean-gilliam
Aug 12 2015 21:28
great job guys. google hiccups and all :)
Joshua Benjamin
@annymsMthd
Aug 12 2015 21:28
But having the elk has been a lifesaver with our entire stack.
Andrew Skotzko
@skotzko
Aug 12 2015 21:45