.Net Foundation .Net Community Standups, 14 Apr 2020

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: ASP.NET Community Standup - April 14th 2020 - SignalR with Brady Gaster and David Pine

Description

Join members from the ASP.NET teams for our community standup covering great community contributions for ASP.NET, ASP.NET Core, and more.

Community links for this week: https://www.theurlist.com/aspnet-standup-2020-04-14

A

A

We just press enough buttons little work right, I.

B

Think I can hear him. Can everybody hear now give us a thumbs up in the chat in there Stewart lying, cheap and now I? Like your name, it's a sound hello and now Scott addy, who we can trust this.

C

Is all wonderful?

C

B

C

Need so part of my deal is I. Do some stretch streaming on my Mac and I. Do this on my Windows machine and switching audio devices back and forth always loses them. So all right. Well great! So we're gonna do a signal. Our show. Let me start off with some community links, so let's dig right into those I'm gonna blaze through them. There's always some wonderful links, love to share them. Bleh bleh.

A

Through that there.

C

Are some yes well and on the on the chat make sure people can hear me I hear my I see my audio devices moving, so it looks like we're good fella, complaining, I guess if we can't Martin below he's talking about building asp net court egg helpers to show/hide who I elementary.

C

So this this is a cool thing he's doing this for a speaker, travel app that he's building. So that's I think that's just a really nice thing that you can build tag helpers throughout throughout your applications. It makes it really easy to reuse, reuse, UI and functionality without kind of cluttering up your front end feet. So here you crates that somebody's type them super loud and I bet it's Brady. So.

B

C

Don't know, though, how much it comes through there? Okay, so um so anyhow, going through registering it and- and you know some notes at the end, but for the most pretty most part, pretty straightforward- it's nice to that you can do this using um I when I end up doing this I'll look at the final new project templates, because there are some login logout things. Also there and well so crostini is continuing. His mobile blazer bindings series here he's looking at state management, looking at different things, storing state and components, app state, etc.

C

It's interesting to me on these to watch the know the model, because it's blazer and xamarin forms together. So it's you know it's interesting to kind of think through some of those things it's a, but so definitely with with application. You know mobile application. You need to understand things like your offline data and app sync and all that kind of stuff, so I'm, very cool, ok, Damian, Bowden I'm, partly just pointing this out to remind people that he's got this really cool library where and and he's been doing some stuff with it certificate management.

C

So there is some stuff for certificate management now built into you, asp, net core and he's built some nice helper libraries on top of it and then also blogging things like how to implement in your asp net core applications so certificates from as your key vault definitely important. You don't want to be managing your own certificates anywhere near your app. If you can help it.

C

So here he's you know walking through how to set that up how to configure that and as your key vault and then how to pull those in and use those in your application, stymie.

A

Ins, great, I we had him on our dotnet twitch stream for the doc show just recently he's amazing yeah.

C

Yeah he and one thing that I love about what he does is he actually very regularly will go back and update his previous blog post and I always feel like I'll. Look that good.

C

He actually doesn't it'll be really cool and it will post like update notes at the top. Alright, your head blogging up a storm, as always so he's talking about work as an organizational raise, peanut core 3-1- and you know talking about the different options available, different work, types and stuff beautiful diagram she had. So thank you for that.

C

You know the obvious kind of how you can configure it when you create your application, what you're doing there, but then some other things like app registration, how you set that up in Azure I've used this in some applications, and you know it does take a little bit of setup, but it's pretty cool once you get it configured, it's really nice to be able to UM to manage inside your application, and then so here's talking about you know how you manage that inside your app and some more advanced configuration I'm going a little bit fast, because I know we have a hard stop at the end and I want to see all your demos, so, okay, cool david brummel in visual studio magazine writing about three blazer extensions for visual studio code.

C

So one is the asp net core switcher and the the nice thing that this does is. It makes it easy to switch between your view code and your code code. You know your kind of code behind so in in MVC. This will make it easy to switch between a view and a controller, razor pages switches between a page and its page model and then in blazer, switching between the razor and the razor CS. So definitely a handy extension.

C

There, razor plus, is provide some advanced, some better syntax, colorization, so nice and then finally, blazer snippets so Andy to have have additional snippets, always I I, never mind that, and then here is also linking to this blazer snippet pack, which will pull those in stuff.

C

Okay, Jeremy Miller writing about using Alba, which is a library he writes for integration testing with ASP Annette core web services. So here this is really cool. He's had the FUBU MVC project for years and and a lot of cool things that were built as part of that have been kind of moved over.

C

So here he's talking about HTTP on track testing, and so the idea here he kind of walks her an example doing a simple math, API and just kind of talks about how he creates some reusable infrastructure, so that his tests are a lot more lightweight, so setting up a fixture and then he's using the e you know built in HTTP.

C

What is it? The factory stuff? That's in SP net core, but kind of wrapping it in this integration context, and so the end result of that is test can be pretty lightweight because all they need to do is you know just inherit from this integration context and then the test itself is super lightweight done. I need stuff, open telemetry, so, on last week's show we talked to one of the links.

C

I shared had some open, telemetry stuff in it, and so Rob pointed out that the team the.net team itself is working with open, telemetry and integrating that in so that is pretty cool, so the open, telemetry dotnet SDK some of the work. That's in progress includes work in the framework and then there's also this open, telemetry SDK NuGet package to make it easier. So the end result here is that library, authors simply write platform, standard API and it just lights up in open telemetry. That is really neat.

C

So the the context I was mentioning it in last time was Jimmy. Bogart was writing it. You know a decent amount of code to light up telemetry throughout a pretty complex, microservice app- and in this case here, this is hopefully making it simpler to do that, all right congrats to the project, Thai team and they are tearing it up. So this is release 0.1, and so here, they're showing you know, release now. It's they updated Doc's.

C

It's kind of funny I went through at rich stream and failed horribly on some of the steps, and so they added some idiot-proof notes on that. We so I'm glad to have added some value there, and so you know some some great I love their Docs and I love their just. It continues here in their release, notes all the stuff they're writing up.

C

So things like you know and I'm gonna have to dig into this more, but some things like yeah so Brady feel free to jump in and I'm sorry for shaming you for typing loud feel, free to you know it's a party I.

B

Actually worked with Glenn the other day, I haven't done much with with tie, but the other day I worked with Glenn a little bit and we were able to get tie the tie repository configured in a fork that I have so that you could have the open in Visual Studio online button on the repo. So you can actually click it, and not only could you build in the cloud, but you could code in the cloud as well, so you can actually use tie from within a BSO.

C

B

There's a there's a bit of setup now, because you have to kind of customize the container that the the service that starts up to do some darker stuff that you can't do without it, but um I think that was a pretty pretty fun experiment. So people are interested in that reach out. You can definitely get you the fork and maybe see if you can make it even better for us it was cool cool idea. Alright,.

C

Cool one other thing, I wanted to point out in here toward Z and oh and I love this like the dashboard. The dashboard is looking good. One other thing that I'm really happy about is look at all these community contributions. I want to especially call out Simon crop man he's just tearing it up, look at all these contributions and and it across the rest of the community. So this is really cool good to see people jumping in the last thing I had to share.

C

Is this def around the Sun, and this is just kind of an upcoming thing that the community is putting on I just want to let people know about it. So this is coming up. May 12th and this is being put on the dotnet foundations, helping out a lot of community leaders ready, Jeff Fritz. You know so it's just excited to see this kind of shaping up and all the links are here and I will share this out in the chat and in the show notes, but first I'm gonna pass it back over to you.

C

Over to you, okay, so I think you're, muted,.

B

C

Cool, what do you got.

B

C

B

Well, I just wanted to talk about a couple of quick things and I wanted to introduce our good friend mr. pine here David was working in the I. Think I'm going to say this right, the cognitive services offering team summary and he this week is joining the dotnet authoring team. So, let's welcome David to our lovely family of awesome, writers, David and I have long been tinkering around with Sigmar and blazer ideas.

B

Some of you may be familiar with this chuck norris experience about that john pinged me and let me do a quick screen share and make sure I'm on the right browser window over here um before I jump to it. John painting the other day, and he asked me if there was anything I wanted to show around single arm I couldn't remember what I showed the last time so I'm gonna go ahead and share my screen and I want to highlight a couple of really important features that we shipped in our last page I'll.

B

Wait until one of y'all confirm that you can see my screen. Okay, I got it cool, I! Think I've talked about this last time. Do you want to point out? You can see here that I'm on the dotnet client and if I were to flip over here, you can actually see on the JavaScript client, but we recently added automatic reconnection. So in the in the old days of sigmar, we would do a lot of crazy things like try to replay messages and whatnot trying to reconnect your clients.

B

We don't do that stuff in Signal Corps anymore, but what we have added in recent update is the ability for you to add automatic reconnection and if you look in the docs you'll see that we have a couple nice ways that we can do this. Essentially, if you just called it with automatic reconnect method, whenever your client disconnects from the hub, it will automatically start trying to reconnect. You can customize that we have information in the docs on how to do that.

B

We've also added on reconnecting and on reconnected so that he could handle those methods to kind of change your UI to so people. You can't take action now.

B

You're disconnected wait a second whatever you want to do, but you can actually use that to kind of sugar up the UI and make sure they won't do something while they're disconnected you can also scroll down lower in the code and you'll see here that we have a couple different ways that you can customize, that with automatic reconnect behavior if you're on the dotnet client side, you can actually pass in time span and objects here, instead of just passing an insta, you know describe the number of milliseconds that you can wait.

B

This is one of my favorite new features in safe mode or I know. A lot of customers have been asking for this I as well as many other customers always had interesting ways of accommodating this type of what we like to call the tunnel effect. John I think we've talked about this in some of the mobile demos. We've done before you know in and out you know you do so in this case, we've kind of given you that capability and like I, mentioned it's also in the.net client.

B

So you can do things in the.net continent as well, and a.

C

Demo I did and it's really fun like it's. You know you have to write the code for your on reconnected and all that kind of stuff, but it's just when it like works. It's like. Yes, it's so cool. It's.

B

Cool it's cool and I, bring up the dotnet client specific reason. Recently, our hero, mr. Damien Roth, PM extraordinaire, a blazer blazer creator, du jour. Let everybody know that we had preview one update and since then I think, we've done pretty pretty sturdy and will continue to do updates on it. But one thing that we did in this update foscar over here make everybody dizzy, was that we added support for the dotnet signal, art, client now.

A

B

Like myself, who've been doing signal are, for years, the idea of having a signal come in from the server and hit my back-end and then I could pop up message up to the client was just always. You know second nature to me, but when I started doing things with blazer, was it as easy as I thought it was going to be to kind of tie those two items together: I'm actually had a label in our github issue. Tracker called signal are loves blazer. Just so we made sure that everything we did around there was.

B

You know we are trying to create a better relationship between the signal, our clients and blazer itself. We've actually been able to D do that here, so you can see that in this case the Dan kind of walks through an example that David's going to do an interesting job up here in a minute. I've got a hub that runs on the server with my serger side post from a blazer webassembly app I'll wire up in my chat hub, just like I always have in the past.

B

But then, when you get into the actual c-sharp code in my client, you can see here that I'm mooing up a hub connection and I'm in the uninitialized async method, I'm, actually creating that connection building it and then wiring up all my handlers for my events and calling start async. So this is where I get to actually used doc, Nick I'd, rather than JavaScript code directly inside my blazer pages, so that I can use blazer and signal our client together.

B

The dots are fun but I know demos are a lot more fun and I know my my good buddy did a pond has a pretty awesome demo, so I'm gonna stop. My screen share here turn it over to mr. pine and let him blow our minds for a minute. Yeah.

A

So I'll share my screen in a second I want to add a little bit more context. So last year in October, I was getting ready to board a plane to Bulgaria. To give a talk on signal. R and I was in communication with Brady and he was trying to convince me to switch over my app that was using blazer extensions at the time right. It was the only dotnet capability to use like the dotnet client and that's when we actually uncovered a bug.

A

Do you remember that, like there was an issue where webassembly and like the Blazer implementation actually was using signal are under the covers right and I? Don't I, don't recall correctly what the issue was if it was like a singleton or something, but you know we couldn't have multiple signal art channels from the same asp, net core host was pretty much the end game yeah, so I ended up using javascript and you you're like so you you reach out yesterday and you're like hey I, have an idea and I was like yeah. What's your idea?

A

Do you want to come on the show? And it's like? Oh, my god. Yes, let's do so I'll share my screen. Now. Let me know when you can see my screen.

A

Okay, so the demonstration that I had prepared was a demo that I gave in dev reach in Bulgaria last year and the fall. But that was the JavaScript version. So Brady says: can you do that in dotnet? Now, with the.net client and I said sure, so, I wrote this last night, borrowing a bunch of the bits from before, but basically it's following Dan Roth's blog, where you do like file new and you get the signal R bits. So you have.

A

You know a webassembly blazer application, so you get the goodness of like a shared class library. So you have just poco sitting here and those can be shared on the server and the client which is nice and then the server as you'd imagine is pretty straightforward.

A

We have our startup and I have some blazer Twitter services I'll register. So in here we add signal R and then the specific application is actually going to rely on Twitter and the intent of the application is to have essentially a a filtered Twitter stream. So I have some handles and hashtags that we're concerned with and I have a hosted service. That runs here, especially that the Twitter service and the hosted service basically sits here in acts as a background service right.

A

So what's an asp net core background services are in process services that run outside of the context of the request response pipeline.

A

So I'm relying on a filtered stream- and this is coming from tweet NB, I believe, is the pronunciation of that and it's just a Twitter SDK and we'll will instantiate the stream we'll wire up to some events and when those filtered tweets come back in, we can broadcast those back out over signal our to the client that will be listening to them. So there is the server bits right and then the client is full-on.

A

Webassembly blazer, which is awesome, they've, made some improvements to actually enable the async main, which I think that was while it was introduced. It was like c-sharp 7.1, or something like that. I, don't think you could do that before with blazer. Now you can so then we have our index, and this is where things get into like the fun the fun parts right.

A

So one of the interesting things is you'll see a lot of people using like the little code like at code and then that's where you kind of ingest your your c-sharp instead I have just my markup and I'm, inheriting from an index components that index components is essentially like the code behind for the application right. So this is where you get first class representation of c-sharp, and so we have a component that inherits from component base.

A

We have a couple properties right, so we have a status message whether or not we're streaming, whether or not we're actually connected to our connection, and this is in the client app. This is where we're actually consuming that the dotnet signal our SDK, so we have those hashtags and handles. We have some tweets again the the tweet result here. This is coming from our shared library. So the server has that same interaction, which is nice it can, they can use those same objects.

A

We have a navigation manager, a JavaScript runtime, and this is where things get a little bit disheartening, so don't be too upset with the the the Twitter when the tweets come through and they render we just get a blob of HTML. So in order for them to actually update and look like an actual tweets, we have a bit of JavaScript and this is where I have to nudge the Twitter card. So this is the only JavaScript there's, no other JavaScript hidden up my sleeves here. So if we go back over here so if.

C

That's not very much JavaScript. That's.

A

Not no, that's that's five lines right, so it's not super upsetting and then just go ahead. I just.

B

Want I just want to make sure you're romancing the fact that you have two files open they're. Not one yes, makes me very happy. I love to hear in the chat is the separation of the razor and the scene. Yes, my folks happy it definitely does me. I was excited to see that implement. That I was excited to see that feature shut up. I guess I should say yeah.

A

Yeah I for one I, like the separation, I'm a I'm, not afraid to say I'm an angular fan, so a lot of people are either. You know angular or react or insert the other thing. But I like the separation and react is kind of a mixed bag. So that's one thing: I was trying to avoid so I like to have in just the c-sharp class, so.

C

A

Uninitialized async, we instantiate our connection builder and then we use the navigation manager to get our stream hub right. So this is our hub that we have and then we can wire up to the events, just as you would on any hub connection. So we have our tweet results when it's received, we'll add it to our collection will state what will invoke.

A

The state has changed to let the webassembly app update its representation and then, after that, we fire and forget and again this is where I apologize for the bit of JavaScript, where we edged the twitter card. You.

C

Know, I have to say I like I, get it because you're you're working to do it all kind of you know beautiful and architecturally right, but I think actually I love that you're not stuck with it's, not an either/or. You can use JavaScript as much as you want. You know what I mean so you're, not you're, never kind of like locked in because I've worked on things, and so like hey, here's, a nice, javascript library, I'll go ahead and use it. You know it's like yeah.

B

Yeah I couldn't agree more, like I actually hit David recently and I said dude I. Have this? Have this audio project you both know about it. I think the internet knows I've been working with audio project for a while and I said I want to in the sounds in the client, but I want to do it from within blazer, and he sends me how the Jas and there's apparently a blazer wrapper for it as well, but it's great like if you've got great JavaScript libraries to party no yeah.

A

And that's the nice thing about this, like the companion repository actually exemplifies the same verbatim demo, but using the JavaScript equivalent right so yeah. So we have our status here and then we start it and then I'll add the tracks. These are the tracks that we have here on the clients and then it's just basically the the romance of interacting with the hub and updating things.

A

So, let's just run it real, quick and I will admit that earlier this morning, when I was running, it I hit rate limiting because I was doing some final tweaks, so hopefully, I've laid off I've been off of it for like four or five hours now so, hopefully, hopefully it'll work. Oh.

B

A

Don't you love something? Yes,.

B

So what are the.

B

A

So hashtag signal our hashtag blazer hashtag developer community at nets and at my handle, so wait for it to start and then, while it's loading, it's gonna, open up, chrome I have a tweet here that I've prepared ever so elegantly that I will tweet. But once our our web assembly app launches we'll we'll get a better kind of contextual setting of what this app does.

B

When you're sharing on teams or any of this training program, when you try to run your app, it always runs a little slower I.

A

Definitely agree with that yeah this morning it.

B

A

Much faster yeah we just run those over here. Thank that, like that, wait for it blazing dotnet, tweets loading. Please wait.

A

All right so we're connecting.

A

All right, Twitter stream resumed. So now, if anyone around the world is the tweet any of these hashtags or handles our application will pick it up. So that's where I encourage. You know viewers that do that. I've got one here, so I'll tweet it right now you saw on Twitter. It is now live. Tada, we're gonna minimize that go back over here boom right. So now it's already picking these up.

A

You can see that the tweets are coming in live they're coming in hot people are tweeting them and as they come in so on, the backend the hosted service has access to the Twitter API s, we're filtering streams specifically to those hashtags and handles, and as people tweet them live, our our service will push them through and signal our will pick them up and then the webassembly app here will display them. You can see lots of loved this.

C

B

We need multiple columns and I'll just start using this. Instead of what I use.

A

I always like to tell people like okay, I I would dare you to like, say something offensive or upsetting or potentially whatever, but I I. Don't like doing that, and one of the reasons I don't is cuz. It's not great for demos, but I actually do have as an added bonus here in the application. When we go back over to the Twitter service.

A

One of the neat things is if the tweet like, if Twitter, believes, that it's potentially sensitive I'll, take it a step further and I actually have the sentiment analysis from ml net running here as well. We're all took a look at the the text of the tweet and predict whether or not it has negative sentiment, and if it does, we don't we don't let it through. So.

C

I usually have a breakpoint here.

A

So if it's running and people are sending offensive things over we'll hit, this and I'll show them hey we're not like having that it's.

B

Not as down so I just had a great conversation with Bree before our show she goes would with. Would there ever be room for an ml net demo on an all web show I was like.

C

Her did the idea I've always thought with dotnet. That's so cool is you can do so many things with it and it's great to see the things kind of crossover like we've had laser mobile bindings and we've had like you know: okay, you know we're some of the stuff we're doing with microservices is kind of crossing over and did it you know I mean so absolutely yeah. Work with Bree will definitely get around to show that so.

A

I don't know if there's any questions if anyone was curious about anything. If you want me to die I've been to any more of the code, I'd be happy to show that off otherwise I can stop sharing my screen to.

B

Answer any questions that you have all this stuff already yeah, you.

C

Know I didn't have any extra questions if you have any that you've answered that might be good to highlight.

C

There was one question on the automatic reconnection going back. Aways was somebody said. If you don't actually connect, then reconnection doesn't work for them like they're. Expecting that reconnect would work regardless. Does that make sense.

B

I'm curious about why it wouldn't work I mean at this point. You would do, though this is where you would do with automatic reconnect whenever you're doing your connection would say that this is only available in the asp.net core client, so I do want to call that out. This is not available and done at frameworks ignore David I know they want your repo URL, so you might want to pull that up as well. For mayor.

B

But the other thing to call out is with automatic reconnect will connect back. Do the same sir. So if, for whatever reason, the reason that you've you know disconnected is because that server is gone bye-bye and like perma bought by you will create a new connection back so I do want to do want to call that uh with automatically connect. Is that the one caveat is that server has to be there? The server from which you got disconnected needs to be there for you to reconnect to it.

B

If you get a different server, we can I shouldn't set the new connection. I know that's the semantic thing, but.

B

There's any other questions and.

A

For an added bonus, while we were all chatting about Auto, reconnect actually added that to the source code and push that up to the repository, so people will have it.

B

And Blaser web RTC api Jil Sander ah David. Have you ever done anything there? I'm not.

A

What was it again, sorry, bad.

B

Party C API J. Is it our up I.

A

Have not is that um that sounds interesting, video stuff or no, that's that's different I.

C

B

A

Done a bunch with actual like native video stuff, with angularjs and stuff like that, but I've never done it with blazer I haven't got into that yet I'd be curious. How that would work.

B

Louis has a question here: if you could change screen one more time, babe, if you have it up, not sure if I'm mrs. to the beginning with a signal, our ruling on the server or webassembly I wanted to call that out in David's project. He has the server like in the server project. He's got a normal blazer web assembly project. You can see that in the server project he's actually got his hub, but then in his client project.

B

That's where he's using the the signal, our client, so the signal or client part is running in webassembly with the hub Runnels yeah.

A

And my hub is, admittedly a lot more complex than probably the standard thing. I was trying to flex a bunch of different muscles for some of the capabilities that aren't normally demonstrated with with signal are so anti Brady's point. This is the server we have a hub on the server, so you get all your native asp net core single our greatness from the server aspect, but then over in the client. The client then also has a dependency on the signal, our client right. So this is the dotnet signal.

A

Our client, so signal are both ways once client, one server and getting back to like flexing the muscles for the like the strong typing here. So hubs can be when you inherit from hub, you can inherit from just hub plane or can do it from an actual like of tea, so you can add a type parameter.

A

So in this case, I have a bit of complexity with how I've architected this, and it's somewhat intentional, because I want to access the my my twitter, client and service inside my hosted bits, but then also from my stream hub. But then also I from my it's almost like a conflict if you will, but my twitter service also takes on the the hub context. So there's a bit of airing out some dirty laundry here in this, and I should probably clean it up, but it's an interesting thing that you're able to do so.

A

My twitter service takes down the hub context and, as you may know, the hub context is a way to actually have like a an abstraction over the actual hub. So I have my I hub context of tea where it has the twitter clients and then, with that context, I can still, from this hosted service pump messages out to the hub. That is part of my my web app, which is kinda cool.

A

B

Is cool? Okay, well,.

C

B

Just gonna say the the the hub of tea is also kind of a new thing for folks who might have done so more a lot in framework days and then came around in core. Essentially, you know the hub itself, just pub with no inheritance or no generic argument. It's gonna basically give you the ability to do die, not dynamic, because now we've actually got a dummy in the cup and the dynamic up.

B

We use dynamic methods where, as this hub, you could actually in this case david has an I Twitter client, if you think about that, that tea, that he's passing in there that he is essentially this interface, and that is the interface of the clients not on the interface of his top. So if you think about it, this is the interface that is hub knows how to call so kind of inside-out thinking.

B

But if you've done, if you never I hate to date myself here, but if you've done something like full-duplex WCF in the past, where you had to have a service contract and a client contract for that callback similar to that. But what it does is it gives you in the dotnet side it gives you that nice type specificity on the client side, because you've got the interface that you can kind of Bank off yeah.

A

Exactly so from the hub context, I can walk up to all the clients and I can broadcast out on a strongly typed interface right, so this is tweet received, so this this isn't normally available right this this method here. This is something that is specific, because my hub context is specifying that type parameter of the Twitter client you'd.

B

Either have to call sender and book or you'd, have to use dynamic and then not get the arguments.

C

Okay, nice, nice.

C

One thing: there was a question on on the retry policy or reconnect, and you talked about this a little bit before Brady with it like you can specify times times, and things is there, but it's is that is that kind of the main there are there ways of doing kind of more custom policy sorts of things like an exponential back-off or things like that? Yes,.

B

So he's actually gonna pass in like an array of time spans here, but we do have the capability for you to provide kind of your own logic. Their exponential back-off is the default and it goes it's in the docs goes 30, 15 and 5. You can customize that, like he's doing here, if you wanted to provide I, think Steven wrote in another way to do it, but I can't remember what that way was he's. Gonna.

A

B

In here no, you just need to go into the code. Yet we talked about yeah, so we actually talked about giving you a different type of retry policy for each one. So you could write, write your own way to do it, so we do have abstraction there. So if you wanted to write your own with automatically connect handler, you could do that. But we've given you a you, know automatic back off in the beginning as well.

B

A

The way that would work is you give them an implementation of this interface and they'll. Call next retry delay, giving you a context and from the context, you kind of determine how you want to handle it.

C

A

C

There's a question on your experience on maximum number of open connections. Signal are, can use, doesn't.

A

Depend on the underlying protocol.

B

If you want to scale off signor I, don't have the slide up and unfortunately I can find it. But essentially, if you want to go from, you know, a hub is server round so cool thanks talking about here. You know this hub that David has lives on the server and if you wanted to make a call to that server and then, for whatever reason, you love balance your app now. You're actually gonna have your hub over here your hub over here.

B

So if your clients connect to this hub and some clients connect to this hub, these two clients sets aren't going to be able to talk to one another unless you do what's called a backplane put it back behind them now right here they just put up the outer signal or service the address signal or services and service inside a browser that gives you the capability to be able to scale out to pretty much infinite numbers of connections. By default. You can enlist in terms of units and I, think you can go.

B

It's a slider I can't what the baseline is, which you can go up to a hundred thousand concurrent connections on one signal, our service instance and the documentation talks about. If you think, you're going to need above 100 thousand concurrent connections, you can do what's called charting and basically create a series of the different just ignore service instances. You can wire them all up and you can use the documentation spring slate and actually goes into how you can use signal groups to say people in the east group are gonna be on.

B

This instance of people in the West group are going to be on this instance and they round-robin everything for you automatically. Well. That also does it's a really great side car. If you want to do if you want to host your asp met after your blades or apps up in has your app services, you can actually add the tools will go for you recommend as your signal our service.

B

If you're using signal are and what ends up happening is all of your WebSocket traffic, instead of it going directly to your app service and showing up whatever bandwidth and whatever request you need pulls you want there. You basically take all that WebSocket traffic and move it over to the address signal service so that your application server is only serving the HTTP traffic to be able to serve the patients.

B

All your fast-moving WebSocket connections, those get offloaded over to the address signal or service and the connections that are managed for you on the server side rather than on the client side and in a secure manner. So.

C

Easy easy to to set it up to it's like one line of code and a configuration setting for this.

B

It is even less than that now and if you, if you have the connection string in your app, we figure out that you need to go there and we go there. Oh you don't even have to change the it's. It's the. There is a way that you can customize that behavior, which is all documented, but you don't even need a server anymore with a just ignore service. You can actually go a certain of this mode and you can do things with like.

B

Have your functions and as your signal or service bindings, which we've had Anthony chew on the show and talk about a few chunks which you can you know you don't even need the server or the hub anneal or you can just back stuff off the server sync are crazy, and since we have clients for JavaScript typescript, Java and.net, you can kind of run the gamut. There's a.

C

C++ one isn't there.

B

We want to ship that in the five wave it's been professionally in development quite some time. Brennan Conroy is just doing his best on that project, but we want to ship it in fact, and we hope that the C++ client would enable more lower-level development efforts.

A

B

Have an open-source iOS client which you can find, we do not have a plan at this time to ship a you know: Microsoft licensed iOS client for signaler. If you will, if you have input on that, please get it to me. The big question we've asked back and forth with folks is: would you want to have it in Swift or objective-c? It was like a 50/50 also because, like 50 people that wouldn't use the other, so you know we don't have to build and maintain both. So we were, you know, enter bets wisely there.

B

So you need an iOS client. I am the call you want to reach out to and I'll yell at Fowler and we'll figure out how to do it. One day.

C

You know this makes me think a little bit of like pwace. Is there any kind of anything special you would do or whatever, with a PWA with the.

B

There there is a little special stuff that you have to do inside the PWA. Now, there's some interesting check boxes in the latest, 3d release on the blazer side, or you can actually check a box and it becomes a PW way and that just works flawlessly. So you can build a blazer app today check the box and get a PW a right out of the gate. Since blazer uses say more than that's gonna work, you know for you.

B

I have talked to some customers who had some issues with some stuff that I do not know anything about in the JavaScript client side, which is was a service service workers. Would we want to do service workers I? Think with react.js? Is it reactance trying to learn? I'm, probably shown that I don't know what this frame works anymore, but we want to. We want to figure out that one if there was one bug that I got on, that we want to kind of investigate.

B

I've talked to Fowler about it, and we just tend that the time I really yeah.

C

Somebody was asking where to find the iOS client, so I in the docs I see a link there. Oh.

C

Yeah, a lot of good questions and excitement from the from chat people are excited about this.

C

How pretty cheap the signal? Our services as your signal, our services, some I- was surprised. You know.

B

I will say this if you start to look at that. You know. One thing to factor: is your build by connections? Talk to customers who have you know tiny little devices and their business might consist of lots of customers to which they shipped many many many many many tiny little devices that stay on all the time. You can imagine that bill would begin to pick up a little bit so so do have that in your mind, you know think about capacity planning when you're thinking about your your you know, consistent connections isn't.

A

There, like an Asscher signal, our service calculator, I thought addre had like all these different calculators: okay, peers and.

B

John I just dropped late for the Iowas client teams. So let you pop that back out to folks, okay.

C

Let me see, there's a question about how how signal are compares to gr PC streaming.

B

That's a great question signal or streaming have quite a few differences I'm not too deep in the gr piece of the area we want to have. Maybe James Newton on I just came on to talk a little bit more about that. Obviously, gr PC is going to use HTTP 3 signal or does not sit on top of a CT p3. Yet I'd love to do that. One day, don't have any cards for it's still experimental, but with Dr PC streaming you get to bidirectionality with signal, or you also get to bidirectionality the streaming.

B

The client to server and server to client streaming API is in signal our we have some opportunities to make them a little bit more elegant in the future, especially when it comes to doing things like juggling multiple streams. At the same time, if you were to remember the demo that I did on the show about a year ago, we announced client-side streaming called what was the thing called streamer streamer, as your website's thought net that has some code in. It is interesting code where we're literally managing the incoming and outgoing streams.

B

We want to make that more elegant for folks in the future. I think that experience and the G RPC client is a little bit more elegant, not a lot more elegant, we'd like to do the same thing in the signaler area, but in terms of turf I can't give you perfect numbers on it again. This probably in it depends what kind of server or micro service you're running in the back. Ok, it's also the protobuf description, which is nice yeah,.

C

B

It's experimentation with describing signal hubs: I, love, Rico I love the work he does, but we just people weren't into it. Customers didn't like it. So customers told us they didn't really like signal or description. It wasn't something that they.

C

C

Any other questions I mean there's. People are just generally like they love it. Let me see this is a little different. I mean this is a question on G RPC. Instead of graph QL between backends I. Don't think this is necessarily the.

C

We did a show week ago, I believe it was with. It was very recently we did G RPC, so I'm, probably some questions that are coming up on G RPC. Those would be good, ok, very cool yeah. It's interesting like all that I'm just thinking about the things you're doing because part of what's amazing about signal art. Is you abstract over the different transports and so as a developer, building across different languages and browsers and now devices? You know what I mean I don't have to think about that I use a hundred.

B

C

Out and then this magic happens, yeah.

B

Yeah exactly right, but.

A

You still get that specificity right where you can configure it, and it's finite. You can say I want it to be this. This transport or this protocol right.

C

But you don't have to do that. You know.

B

C

A WebSocket guru or whatever it is the next thing you know or when so so, that's that's cool I! You know, I was wondering a bit like with the you were talking about like the C++ and then also you mentioned some of the things with like small devices. When you have like lots of IOT things connecting. Is there any way to throttle down the connection like how chatty it is? I?

C

Guess, that's more of a I need to not put stuff on the hub that I don't want sent because you're just you're you're, just transferring the information we're.

B

Really just transferring you can control both the both the way, you're feeding data into the hub, and you can feed the the way you're getting it out on the client as well. You do have the capability to customize that now, if you were to get into the streaming world, then you can get access to the channel and then you can get access to back pressure.

B

You know what I mean, so you could say: I only want to have a I only want to process hundred at a time when we process the housing at a time, or you can do something really dangerous and say unbounded, which, if you do that in toddler size, your code.

A

B

Could your server at that point, but you do have that you know that control, but when you're getting into the streaming stuff, you're gonna need to do a little more coding. Gonna be a little more work. You start messing around I. Think a new normal t. You really gotta think about how you're gonna deal with that yeah for me, starting to work with I think innumerable was fun because I would say well. I want to get back now, run my code and another.

B

Your turn the thing on and turn the thing on it all sudden. They start. You know this is that a weight just keeps hanging out every time. I get one, that's a weird way to think for a dotnet programmer, because I think we always think we'd have to go back down the sockets, even though those want to well.

C

I know you've got a big meeting here coming up in just a few minutes, so how about we wrap up and if you want to share I, can add the link for source code and anything else into the community links for this week. Anything else you want to point people to just for keeping up to date with what's new in blazer land.

A

Probably share that one blog from Dan Ross that day introduces the the dotnet client yeah.

B

I'm saying ASP nut, you know our site whenever we do even the most incremental of an update, we'll put a blog post out there and I have an email in my inbox right now from Dan. We want to work together to make sure that the documentation is better, so we're gonna be doing some stuff there. If there are any needs that you have in the wiser documentation area or anything specific people have to hear about it. We want to make sure that that experiences I'm giving for you is.

C

There, like a public blazer roadmap, that people can keep an eye on I.

B

Don't know that would be okay, alright I, don't have a sequined jacket, so I'm not.

C

Yes, you need a signal. Our sequin jacket, t-shirt.

C

Well, I'm conscious, you got a big meeting so and thanks this was awesome and thank you for welcome to David thank.

A

You thank you swell.

C

Have you on awesome.

A

C