.Net Foundation .Net Community Standups, 21 Aug 2018

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: ASP.NET Community Standup - August 21, 2018 - ASP.NET Core 2.2.0 preview 1 release party!

Description

ASP.NET Community Standup - Aug 21, 2018

Community links for this week: https://www.one-tab.com/page/Lbn9-De1RbOCEvxKyt4tSw

A

All right, we should be live. You.

B

Want it gonna hit your dual stream there. You.

C

B

I'm not gonna leave until you guys know.

C

B

Online right now, I'm gonna push.

B

D

B

So you got rid of the feedback, so you should mute that yeah cause.

E

B

To save bandwidth cool so.

C

What do we do now? Oh hello,.

B

I didn't see you there how many people are online Jon? We wait.

E

B

Usually wait till like 50 there.

E

Are 61 watching now already.

B

E

Now Hashem.

B

Hashem just said finally Scott in the new studio: hey everybody check this out.

B

We're gonna make dotnet core run on a 10 inch floppy disk. This is a new feature. It's coming soon. It's gonna fit in a hundred and sixty K, but we're gonna have to make it double sided so watch out for that floppy disk. What's that.

E

B

Fits in the new ultrabooks so watch for that cool, hi friends, I Scott.

E

B

It going I'm so excited to be in the new studio and I've actually not been on the show in a while and your remote still John in awesome and all of your all of your your wall. Hangings are straight, which makes me feel better because it's usually one of them's crooked I.

E

Learned it Scott when you said that you intentionally have some demos failed to keep people like on the edge of their. You know, I didn't.

D

A

Lies lies so just below the screen. There I'm sure his laundry is actually sitting in that.

B

It's true, that's true cool, so we're in the new studio dan. How are you I'm good doing? Well come on Kyle? How are you everything's great so, have you been on the you been on the show a couple times right, I actually,.

C

B

C

Who are you hi everyone? My name is Mikael Mengistu I'm, an engineer on the asp net core team I work specifically on signal, R and I'll be showing the signal Java client. Today, I've been on the team almost two years now and yeah things are going great cool and also from Portland. So we got to Portland people representing here.

B

In the stand up, which makes me feel much better and more comfortable, mm-hmm. Sorry with your piece, you can.

F

Come to my house in Portland sometime, we actually went to the same alma mater. Oh, that is true. Yeah.

B

Yeah I did so did he yeah I went to Portland Community College.

A

B

Little bit he did.

A

G

B

Once it was really cool cool all right, so how do we want to do this? You.

B

People know about our buttons.

E

This is our control.

B

Panel now to control streaming for OBS, because we're like official now, instead of using Google Hangouts we're using open broadcasting software and golnaz from the up, and you just left John well,.

E

I switched my thing over to okay.

B

Well, so we're using open broadcasting services now above broadcasting studio the way the twitch does it, and we've got this really cool macro button thing here, so I get to play director, which is very exciting. So if I push.

F

This little button says John gee.

B

Yeah, look at that.

A

B

A

It's amazing he's the only one that has his own, but he.

B

Does not special that way.

A

B

E

So I, don't I, have just a few super high quality links this week. This one.

D

I, don't know how deep.

G

E

Depth I should go into it. This is Steve Gordon and he's talking about endpoint, routing yeah, but I'm, not sure. If you're gonna talk about that I.

A

Was going to show a little.

E

A

E

So Steve goes through and he writes up. You know what is endpoint routing I think previously it was talked about his dispatcher he's been updating it based on conversations with with James and other people. So the name now in point routing and my understanding high-level- is that this allows you the goodness of like MVC routing, but it allows it earlier in the pipeline, so your middle, where your other middleware can take advantage and understand the routes as well. That's true, so yeah cool, so.

F

He goes, Turner shows a small.

E

Example: I'm not going to show much more of this, because I don't want to spoil your fun. That's.

G

Super awesome to.

E

G

People like trying out the CI, builds and keeping up with, what's happening in the repo like that I love.

E

It and it's it's, this great feedback, loop I've got some other. Some others will be showing too from people that regularly, post and and that we feature and it's cool, seen the interaction between them and so I'll kind of say: hey. Did you check this out or that kind of stuff? So here's a neat one from Philip, so he wrote about an AI file provider for Azure, blob storage. So the idea here is asp.net has had this I file provider thing forever and the ideas you can abstract away.

E

You can talk to something as if it was a file, but it's not a file. So the idea is I. People have used this for things like reading assets out of like a new get package, or something like that. So what he's doing here is he's actually set up a sure, blob storage as an AI file provider. So, what's neat with that is in one place you register your Azure blob, storage and then anywhere else in the site. You can just interact with it as if it's a file under slash files, so meaning you can.

E

You can refer to the code as if it's inside your app and boom magic, so I think this is yeah. This is really really cool, so um I'd like to I'm gonna, keep an eye on that one, this one's interesting, John Bristow just so this is a this is from a survey. This is survey data. This is definitely weighted towards the like they if you scroll down to the bottom here it talks about this is from an online pop-up at TELRIC calm.

E

So this is weighted towards people who are buying you know or on the telluric site, but it is interesting seeing over time the demographic shift, so here's here's the numbers and I'm just going to go to the actual. You know the the data stuff that they saw. They are seeing more people aware of SP net core, which is good and they're. Seeing you know, people continuing to look at an adopt asp net core I did think it was a little interesting that you know this is.

E

This is folks who are just kind of getting on board with asp net core and he's worried about their awareness of blazer, which is completely new and experimental. So I am not so worried about that personally, but but I thought this was really cool, seeing the kind of change over time as people are getting more involved with asp net core all.

A

E

So andrew lock is writing about segments and segments is something I was not aware of before. It is basically a an a layer on top of other analytics engines. It's they have an analytics jeaious library and the idea is you can feed in from one place to multiple things so say you want to be tracking. You want to use Google, Analytics and Mixpanel, and you know so whatever else they they have tons of other integrations. So what he's gone through and done here is kind of showed.

E

The plumbing of you know, you're setting up your segment account, but then plumbing this data into multiple different things. So, for instance, you know this could go into and feed your your analytics. Could feed Bing ads could feed all different kinds of services there.

E

So he's he's shown that and then, of course, wrapping up with you've got to write this analytic Scot out in so he's showing writing that out and razor pretty cool alright, so here's one from Mohammed, Rohan, Sahid and he's showing about optimally configuring asp, net, core HTTP client factory, so these are kind of his opinionated- looks at here's. How I would like to see HTTP client factory set up some some interesting things he pulls out in here. He talks about setting up poly for retries. He talks about gzip, compression and and setting up compression.

E

He talks about as well and I think this is a really interesting one. The XCore X correlation ID HTTP header. So the idea is having this correlation ID to track all the way through the logs. You know where, where things are going so so you know, there's several different things in here: setting up the policies, etc, etc.

E

Here are some where he talks about the the correlation IDs, so kind of plumbing. That through and of course, as always, he's got a github project, we can dig into the code. It almost seems.

G

Like a bunch of stuff that could be in like a like a nudey package, like you know his apply these conventions or policies to my HP client Factory is like a just a broad convention across the board. I, don't know if it's! If, if he feels like those are good practices just for any app, maybe that should be something that could be pre-built and just added to an existing application. You're.

E

Right yeah that sounds good I'll. Ask him about that: okay, here's David pine, doing a walkthrough of blazer, so here he's just kind of over viewed. He previously did an interview with Steve Sanderson about blazer. Here he's just kind of pointing people to the getting started stuff and then he went through and wrote an internet, Chuck Norris database client, so he's pulling data from the the Chuck Norris FAQs, and so here's showing how to do that going through and setting it all up.

E

So it's nice as a walk through and and at the very end you end up with with the Chuck Norris. So here we can get information about Chuck Norris. What.

D

E

D

E

Chuck Norris needs protein shakes there. You go all right, big, exciting post. This one is about bing.com and they're, moved to you dotnet core to one so yeah.

E

Excuse me bing.com very large service. They also have a lot of internal services calling each other so ton of stuff going on. So this is, this is cool to see a you know, large real-world production application moving to dotnet core and talking about why they did it and what they got out of it, what they learned so as they talk about in here. This is things like their their view. Logic is written in Razer so that you know I mean this is kind of all the way up. The stack.

E

It's not just some dotnet core stuff. In the background, so some neat things, that's ready to run images thing is exciting. So the idea there is, you know when you deploy your code out to a server, there's still some JIT that goes on and with ready to run that kind of engine step happens at Build time and then you're pushing that out. It's already pop it on a server and it's ready to go so that so that's exciting! You see here the performance, so this has been. This is showing their their server latency dropping 34%.

E

So that's neat then take some time on this. It's easy to kind of scroll. You know to.

B

See him the most control.

E

Plus that, for me, sure.

B

Thank you and go back to that chart for a second, because that's the real, that's a beautiful charge. Let that drop in latency is what 34 percent mm-hmm yeah. Thank. That's.

E

Just huge, so this next part here is, is really interesting and I recommend kind of digging through a clicking through on these links here. What they're, showing is here's why they got that 34% job, so they dug in specifically to their. You know their performance data and showed what was going faster. So these are the kinds of things like as I've been doing dotnet for talks lately and I'll. Tell people like this is why being on dotnet course, so great every release gets you all these.

E

These performance benefits that are just magic behind the scenes right. So, if you're, using, if you're checking string equality and you move to dotnet core, your code gets faster, you know, I mean stuff, that's just happening behind the scenes.

B

Master is a, and one thing if I may that's worth pointing out- and we talked to the Bing team is even though Bing is Microsoft and we are dev, dev and Microsoft. You would think that we like to talk to each other all the time, but it's really different companies right Microsoft is a hundred thousand people. It's basically, you know a thousand little hundred person companies if being wanted. Some modification to like the CL are fundamentally or they wanted to change c-sharp.

B

Fundamentally, it's not quite as easy as just ringing us up demanding a feature and fitting it into the thing. The fact that dotnet core is open source literally allowed them to modify it change it without talking to us and then then their negotiations change. They can show us well here's the thing that we want. Here's a feature: we want, here's a drop that we did and.

G

B

G

Get some good good contributions back from the the Bing team as I seem it's.

B

Also, there's gonna be features in c-sharp 8.0 that are modified specifically because being found them to be useful, and if we weren't open-source, it.

E

B

Have made the friction between organizations greater and we might not have seen that cool feature.

E

And some of the things on this list actually I think it was this. This call I this an internal call of optimization and that that was something where they, the Bing team, said hey. We could use this or you know people. There was an open issue and the dotnet core team was like yeah, not sure. If this is going to be useful and then the Bing team, like prototyped it and said, look here are the performance improvements we get and because of that, that's that's why that feature shipped there, so so this is.

E

This is some really cool stuff like call I I wasn't aware of that? That's there's something that's been around for a while, but wasn't inlinable, and this is call like a or or whatever is, is a it's like P invoke, but a lot faster. So in certain cases things can go, you can make internal calls much faster. So one other really cool thing here. Is this runtime agility, so the the ability to have X copy? You know side by side, CI, all that kind of stuff being able to roll things out so much quicker.

E

This is pretty big that there you know this is a huge application and two days after dotnet core two one was released. They they were live running on it. So neat, stuff cool. This is just a quick one giraffe. This is you.

B

Gotta make that bigger my friend. Thank you, sir.

E

So trap is an f-sharp based web framework right. You know, Scott, you've done some posts on these in the past. I think you covered giraffe as part of that right.

B

And it's really cool the way that they've plugged it into asp.net core, like you can see right there, how the giraffe add giraffe extension method, our registers, dependency, they're, basically using parts of asp.net core that they like and the parts that they don't want. They just plug into the pipeline and they take over. So it's really the best of both worlds and it makes a really comfortable environment for our people, cool.

E

So it's neat to see the continued like they're they're continuing to build out on this big feature here is: this is up updates for authorization, policy and and policy based authorization, so yeah a lot of good stuff last link that I always end up with, as we get closer to Netcom, just reminding people. So one thing that a favor I would ask of you, even if you're totally aware of dotnet comp go ahead and and save the date, so we can kind of track and get an idea of.

E

You know how many people are planning on showing up that kind of thing. So so yeah we've got a cool agenda. The agenda has been published. We had last year we had 70 submissions this year we had over 200 can.

B

E

B

E

I'd, rather, people just go and read this on their own time. There's there's a lot of stuff right so, but there's a lot of there's.

B

A lot of local events aren't there well.

E

Yeah get into that in a second yeah. So if you go through and look at the speakers, there are tons of speakers from around the world, and this is the thing we've got local events, but we also have day three. Actually, let me go back to this agenda. Bart's.

F

Gonna, be speaking, that's true. That's.

E

Exciting so today three of this we're going all around the world. There are twitch livestream things from all around the world. It's it's going round the clock. So this is great. We've got people, I mean we've got every continent speakers presenting from so.

C

That's crazy and.

E

Then local events- this is neat, so we have over a hundred local events. So this includes live viewing parties during the event that twelve through the fourteenth, but then we also have live, or we have follow-on events that you can go register for. So these are all around the world. These are places you know near you, so number one find one! That's near you! If there is not one near you, you can register a new one, so you can go and create a new event and tell us about it even know.

E

What's that they can do that even now, even now, yeah they can go. So we have. We have some swag. We have sent out to events that have already registered, can't promise we'll get you swag if you, if you register one late, but still if you want to have an in-person event and tell us about it and get it get on the map. We're happy to add you awesome, yeah, all right, I'm done, shiny I will share out the links in the chat and put it on the description, cool.

B

Thanks all right cool, so we're gonna share your screen, sir sure.

G

B

Right, we click.

G

So we thought we would spend some time talking about what's coming into to specifically in preview, one we've been harder work on to to I'm. We've already saw someone to the community, apparently as he's on the ball and already looking at the vets and trying things out and writing content, which is fabulous but preview on is pretty much done.

G

It's not public yet, but maybe this week like, we hope to have it out this week the the bits that you can install and start using and start playing around the features it is early in the Tootoo milestones, so things are still being being worked on. Things are being polished, so there there certainly be bugs.

G

This is I would say this is more like an alpha quality preview for for this release, but we expect to have a couple more previews over the course of the remainder of the year before we then ship, it and I think the goal is to have it shipped before the the end of calendar year of 2018.

G

That's what we're hoping we'll see how the the releases go, but this is all the new stuff. That's going to be available to start playing around with in in preview, 1 and I thought it might be fun just to take an early look literally a preview preview of the the new features that are coming in this release. So new stuff, the new templates are being updated with getting a facelift.

G

We were moving from bootstrap 3 to boot, strap for and I think we're on four one three now so we're trying to keep up with the latest releases and our third-party dependencies, and so as part of doing the bootstrap 4 update. We also did a bit of a design clean-up on the templates, simplify, make them a little cleaner. Looking, we also updated the the spa templates to update their dependencies and we move to angular 6, specifically for the for the angular template and I.

G

Think angular 7 actually is slated to come out somewhere later in the fall, so we may have to do another update I, don't want to have to look to see how their their release goes. Http 2 is coming in this release. We've been actually working on the HB 2 for yeah. It's been it's. It's I mean HB. 2 is a pretty involved spec. Actually, when I talk to the the devs that are working on the that protocol implementation, they they make it clear that this is. This is not as simple I mean actually be.

G

Hp1 itself was not not a simple effort that took honestly like uh I would say even a couple releases in the a core before the kestrel implementation of HB. One really became came of age like for a while. We were doing that thing where we're saying we really need to run kestrel behind some sort of proxying server like IAS or nginx, and which took a little while before we be more comfortable with putting it as an edge server.

G

So HB 2 is a even more complicated protocol and we've been working on it, and we now have bits that you can try out with with with freebie one and we'll see a little bit of that is in process hosting. This is the ace net core module like historically that's been an out of process model where you all your requests, get forward forwarded to a dotnet that eggsy process effectively and that has performance implications.

G

We've been looking at for again for a little while trying to improve the performance of that setup by moving the a spec or logic into the EIS process itself. So you don't have to have that orbit of having an IPC channel communication from the is process to the dotnet exit process and we have the the Bitzer we're planning to ship them with 2.2, which I think we tried to get them out with with 2.1.

G

It wasn't ready quite yet again one of those features that has some amount of complexity involved with it, and so we need a little more time to bake, but in which it's available, and it's now enabled by default in the releases. So we'll take a look at that as well.

G

Big theme for this release has been services and building web api's, and so we're really trying to improve the experience of building api's that are descriptive and some sense self-documenting. So if you have an API- and you want to generate an open API specification from your API we'd, like the to have a really good experience in doing that, if you build an API with previous releases of a spec or there was a lot of sort of flexibility, I would say and how those your action methods could be called like. You could have a parameter.

G

That could be it's just as equally been a query string, value or a form data value or a header value like it was a fairly flexible model, binding system and that's great for in terms of functionality. But it means that, in terms of providing a description of what your API supports, it meant the matrix was pretty large. So we're I've been trying to add conventions that make things a little bit more opinionated they give.

G

Your API is a better description that you can then use for things like code generation, documentation, testing tools and the like, so another round of Web API conventions in this release, endpoint routing we saw the blog post earlier about what what we've been doing their thumbs in the past. You may, for the the term dispatcher. This is an effort to improve our routing systems so that it is available throughout the stack, as opposed to being kind of an NBC.

G

Specific thing like you should be able to link to a NBC action or a page from your from middleware or you should actually should be able to link to like a signal our hub from from from it aware, though, enabling those types of scenarios is what the end point routing is is all about, and performance honestly gets gets better across the board, with the new, the new routing system, health checks making it so that your M, you can stand up endpoints so that you know your services are live and healthy and chugging along.

G

We are now shipping the health checks API with a spinet core into two and then Sigma, R Java client, a new new Java client. So you can write all your Android apps and talk to your signal, our hubs and get real time real time data. That's that's the the list of stuff. That's coming in preview! One and some things are more. Are you know more feature? Complete than others in this release, but you know how many views will be. We expect I think what we published in there was a roadmap announcement.

G

Actually one just bring that up so that everyone knows where to to see this such.

B

A comforting dialogue, oh.

B

G

So if you go to the announcements repo on github, so hast net announcements and if you look at the issues, there was a a spec or two to roadmap that we published that long ago. So, if you click on that, that has a sort of a outline of all the features that we were planning to do with this release.

C

G

The plan so August preview is what we're doing right now. We plan to do another preview in September and October and then try and ship it before the the end of the year. That's the the plans schedule and you can get a feel for all the features that we're trying to do in this release through the roadmap. So definitely.

B

Check that out.

G

Some of these things are done. Some of them are not, for example, you saw at the bottom. There was the the C++ client for Lauren, like that's, not that's not.

E

Something else yes,.

G

E

Actually, already.

G

Been pushed out to so we so you know the roadmaps. We do our best to try and predict the future. Looking our crystal ball about how much we can actually get done in and release and something's making it something stoat. So that's a an example of a feature that actually got pushed out, but a lot of this stuff is is certainly in alright cool. So shall we shall we take a look? Shall we play around with the bits? So what I did is I?

G

Basically, the latest CI build and I think it's actually still the candidate build for what we're planning to ship and I started playing around with the features to see you know what the state of things are and I thought I would do a few few demos, let's try and take a look, a few things. So let's go to visual studio and if you do file new project new web application, my 23rd web application- apparently, let's make it 20-7 for good measure.

G

Since that's my favorite number and what we should see, I've installed the latest SDK and you can see I have the a Smith core to showing up as an option in the dialog. The SDK version that I currently have. If we take a look, it's a dotnet version I'm on 202, 103, B, 193 49. Is this the current build that that I'm on? Can you see that so there that's the current build of this of this SDK, and so you can create an ASIC or a two application yeah.

G

We can always go ahead and do that and let's take a look and see what we now get in these in these templates.

G

No interesting, that's interesting! Dial! I can see that one before that bad. So this interesting question about what version of Visual Studio you should be using with with two bits: the the plan is that then we're going to be recommending. 15 959 is the version of Visual Studio that you're going to want to use right now. Yes, we're currently on a the 59, is in the preview Channel.

E

G

Don't have, if you know in preview installed yet probably should be doing that you.

B

Want to do that and switch right.

G

Now, no, no, as this will be fine, so the there's a currently an issue with the SDK, where we're with Visual Studio, actually where it's saying that you need fifteen seven or newer, really should be saying: fifteen nine or newer. Oh, that.

E

G

Will be been updated, but fifty-nine is the plans targeted, updated for for a snack or two two and dotnet cor 2:2, but it still works with the older versions of Visual Studio. There are just newer features that you may run to issues with so, for example, the new asp net core module that supports in process hosting. That's something that to wire that up with ice express. We need extra Visual, Studio changes, so that's gonna be one of the features that's going to require fifty nine. It's just a new template. This.

B

G

New template so I think we showed this a little bit before in the the last community stand up, but this is bootstrap for base. We can see that if we go into our the blip root folder and look at the actual, let's look at the the index- the layout. Let's look at a couple things here and there's a big enough. Yeah.

B

Maybe bigger or not big enough, you.

G

B

That I, add to the people and to the community, is to just sit and tell everyone to make their fonts bigger until they become either irritated. With.

G

This dick stick, though, that's the newbie.

B

F

Okay, so things you'll notice.

G

The bootstrap link is a bootstrap for one three links, so we have bootstrap for one three in the templates. You'll also notice that we're not pulling it off of cDNAs, so instead of using our own CDN we're now just using the the public existing ones that are popular in the community. This aligns us with Lib Man, which is the new vs tooling, for managing client-side dependencies. They also by default, will pull from CDN j/s we'd love to hear what people think about this. Like. Is this just a good idea?

G

Should we be using a different CDN than this one? There's a there's, a few options out there we looked at them and this one seemed to have the most usage behind it. So that's that's currently what we're trying out and it's definitely an area we'd love to get feedback on and.

E

One nice feature: there is it's more likely, someone might have those files already cached on their machine right. Yes, it's.

G

A fairly commonly used CDN, so you get that additional benefit. The other things you'll notice about the template is that it's much more minimalistic than previous versions of the template. The the Jumbotron I call it like the both of the billboards that we used to have in the old template with a.

E

Carousel the carousel is now.

G

Is now gone, we've simplified all the links on there that kind of cluttered up the page we, the idea here, was to try and reduce the amount of code that you have to deletes when you actually create an app and once and then use it for development again love to know what people think about the new look and the sort of minimalistic approach to the to the project templates that.

B

You add a couple more bullets with more links to stuff. You, like you like bullets and links somewhere between where we had before, which was too many carousels and graphics, and what we have now, which is the word, welcome yeah.

F

B

Would probably add a bulleted I'd.

C

Be interested in what is one of people in the consensus joan was personally a fan of little more minimalistic style.

B

Welcome to asp.net core yeah.

E

Like I think other ones, when you create a new app, it says, like you, just created an app with few or are you created? You know what I mean it says so.

F

At least advertise that it's with which technology you just used to create the project, people are saying, keep.

B

It simple there's less to delete.

E

G

E

Got page two is interesting is but like I'm, you stories the about in contact and all that I mean I think.

B

The idea is just to.

E

Show that you can have n number of pages right, yeah.

G

That's cool, so this is also the new cookie consent. People.

B

Are saying they want totally blank templates completely blind.

B

G

Yeah used to be that the cookie consent dialogue would are you I would like comma.

F

The nav bar and people didn't.

E

Seem to like that so now it's a dismissable.

G

Thing, that's cool in that bar the what's one. Other interesting is page 1 and page 2. Actually don't go anywhere, they're, just showing you how to use the nav bar. We didn't actually even add in Solon.

E

In page two, that.

G

Was based on, like the bootstrap theme that we based the template on they had that similar pattern where they just had some links in the nav bar and we're like it.

B

Is how you do not bar links and the thing I don't like about that is you're brand new to asp.net you've just shown up. You see this nice empty template. You click on page one, and you see this goofy looking URL and you think that that's the way to do things. No, no.

E

Yeah, but if you actually had.

B

A page one, it would say, slash page one right now, as a new person I'm like wow now.

G

You were very like this goofy looking URL. This is actually. This is actually an example of like why this is an alpha alpha release like this this. This is actually a bug in the template.

E

The page links.

G

On this, like this is a racer pages based template, but they're. Currently the page 1 and page 2 links that are generated are using the.

B

Specifying an action and controller wouldn't.

G

They have index well because there is no route for home and index.

E

To fill it in and you end up with the routing pattern.

G

Where you end up with the action control end up in the query string so.

E

That that will change that.

G

Would be just paid one okay, that would be so. Yes,.

G

Go ahead! Sorry, if.

E

I did have identity, that'll just go page 1, page 2 log never register. Yes,.

G

So if you add an edit, you get the right register and login links and you get the normal identity, UI and those appear on the upper right-hand side as well. I realize.

B

That everyone who is on chat here is saying that they desperately want us to have nothing, but there's got to be at least one additional.

G

Page I'd be curious for the people that are saying they want. Nothing is the empty template exactly what you want, or is even that not enough of nothing. People.

B

Are saying links that do nothing or misleading okay, you need to be able to show routing in the template. You.

E

B

A new template for these scenarios, a couple of people that desperately ask them for a left side, nav bar I, think nav bars are up to bootstrap, so we're probably not going to do that, but I think the point is there are two kinds of people there are people who understand that empty is all they want. Mmm-Hmm.

F

B

Then there are the rest of us that understand that a new person needs to at least see something yeah.

G

What's right, what's.

B

My ten-year-old gonna do at this point the get.

G

Started product in with that, you do the first time you use a spin accord on a core I think needs to have a little bit of stuff in it needs to have some pages needs to have. Some links needs to show you how to do a few things for the people that are doing like day-to-day development, that I've already done the getting started experience and what a starting point for building their next app for their for a new project, I'm. Guessing that that's what the people that are asking for empty that! That's! That's! What writer.

B

So I just think that in the refreshing of this standard or default template, let us put extra attention on empty. Okay,.

G

Okay and I would love to hear from people what they think about the existing empty template like is it? Is it what you want? A car I think empty right now has just like a single middleware that returns HelloWorld right. Is it too empty yeah? Is it to him like something some people have asked? Can you take the empty template and just add a single page or add a single.

E

Control or something.

G

Like like at least have MVC concepts in the empty template, so that you know how to use that, but that makes it less empty. Is that too much like? We would love to hear what people have thoughts on that? Obviously we're not gonna be able to. Maybe we won't be able to make everyone happy, but we'd love to how make the majority as happy as possibly our empty.

C

Template has like the middleware, which just writes that response hello world, but like if you look at like node and Express, they have like basic routing, and it just sounds like it gives you a page. It's like welcome to Express right.

B

I almost feel like the one that we just saw. There was almost an appropriate empty template. Right I mean that's a pretty.

A

G

A

Show me the other one, so this.

G

Is the new spot template, so the spot template still does have some some content in it um curious to hear what people think about that for people who want to know.

E

More contact little.

G

Bit more content, we did shift the spot templates to like we decided that the template should be consistent in terms of their nav bars like before. The spot templates had a left nav. So now it has a top nav. It still has the same functionalities before you know a counter and a fetch data, page yeah for consistency here instead of a left nav. But you know, let us know what you see: I.

B

You know: I know that people hate deleting stuff but I tell you. I would I'd put in a link to the asp.net community, stand up online.

E

Page, you got to tell people about.

F

Community with the.

G

G

Marketing, so just to show the interest of time what else we got in the angular template just to show that this. This also now has angular 6. So if we, if we look in the package, so JSON file for this guy, you can see it's using angular 6 across the board. I think angular 6 is now just 6.1. So we will update that as the angler.

E

Well, maybe 7.

G

Like there's 7 supposed to come out in September, you.

E

Search over yeah so.

G

They move pretty fast, like this.

E

G

This is actually one of the reasons why we're thinking about taking the spot templates and not having them baked in the SDK, because they they need to update. Whenever the client libraries update right.

E

And we can't ship an SDK.

G

Release every single time.

E

G

Size update so we're thinking that maybe all these templates actually just need to be just new, get path. Template pack set on new get. That's that's the current thing. We're not gonna do that in 2 2, but the future thinking. That's what we're looking at. Ok, what else should we look at? Let's look at the wire a little bit for this project and notice that something that's happening on the network.

G

So if I refresh you notice, there's a protocol column down here and it's showing h2, which is a preheated version of HTTP 2, so this application out of the box has HTTP to turned on talking to kestrel yeah. This is talking directly to to to well actually, in this case this, this app is going to your.

B

Address you ready, yeah, you ran this in visual studio. Should we use it so we have expressed in the corner there. Yeah.

G

B

Your running this on battery good job, no.

G

It's it's gutsy time.

G

Should we try it on Castro like I, haven't, actually do a test line.

B

Go for it, let's do it.

G

B

F

No done hold on hold.

B

On yeah wow, this is like a daily build. So don't judge us, that's good.

G

B

E

B

Turned on it's.

E

Because your homepage is too empty here,.

B

E

Yeah, okay, you had.

B

More links in the home.

G

Castrol now has native support for hb2, so things that we allow. We support here now, like HP, who has HTP to has the multiplexing of the the connection.

B

G

B

G

It doesn't feel like.

B

It should um okay I'd like to know that that would be interesting.

F

No I mean I'm kind.

B

Of an you know, I'm a little og and that I like when I learn about HTTP and I teach classes on HTTP I, open up, telnet and I, make the students telnet to port 80 and then type HTTP I need to refresh my knowledge of HTTP 2 to know if such a thing is possible anymore, yeah, it's value.

G

And understand their.

B

Like experience up to the top.

G

Here, in addition to just URLs, by the way, you're binding to not just 81 but also HP, to just.

B

Like we tell them about SSL, you know we should tell them about that. So.

G

Hp 2 is working with Castro things. We do support the multiplexing of the streams. You know the compressed headers all that stuff things that we do not support. We don't support server, push, that's not something we currently plan to add I, believe we don't currently have prioritization of the streams in this release either. So those are two things that are not enabled that.

B

G

Priorities. Well you, if you have multiple streams going over the same channel, you can specify I believe you can specify probably.

B

Super-Important.

G

The stock ticker is more important than the chat I believe. That's the way that works yes, I use.

B

Signaler well.

G

Actually, that's a great question like I may be you know, is there any thinking or thoughts around doing signal are over an HTTP 2 channel? Is that a thing is that something that we should do have you guys talked about that? That's.

C

Not something we really talked about, yeah we're still sort of in the middle of like working on to to stuff that we've got slated and figuring out like 3 Oh planning. We've got a bunch of features that hasn't really come into conversation, yet I hope you bring it up. I'm, probably asking a question.

B

That exposes my extreme ignorant, but.

G

No, it's a good question, because what is the.

B

Relationship between HTTP 1, 2.

G

And WebSockets, so WebSockets is an upgrade from HP. 1Hp 2 is a separate protocol that you negotiate.

B

It's partly it's an upgrade as I've seen the upgrade requested, the red WebSockets runs on top of HTTP 1 or to the side of it's more like kind.

G

Of to the side, because you I mean the the upgrade handshake is done as part of the HP one protocol, but what port is it over? It's still over at the same port, it's.

B

G

The same the same.

B

Well, there's WebSockets have headers for HTTP, or is it no? No.

G

Once you jump right into a WebSocket, it's like having a TC raw TCP connection, so.

B

How would that? How would that relate to HTTP 2, so.

G

It's the thing about HB 2 is it also gives you this effectively a duplex channel, so you could. A lot of people have actually asked a question like do we even need WebSockets anymore, given that you can have an HTTP 2 channel, because you can just have like these long-standing responses.

B

On the server, hence the question: how would signal our over HTTP to work yeah.

G

Is that thing I actually hadn't I know if there was plans, sounds like we're still still thinking about that I mean it's worth, noting that the HTTP to implementation in a snit core is still pruning new, even though we've been working on it for a while there's a lot of work still to be done with it, and we've turned it on by default deliberately in the previews, because we want to hear feedback from people, we want them to be trying it.

G

So we can, you know, flush out those bugs like there's, there's known issues right now with the performance of the hb2 implementation in kestrel, that we still need to optimize and work on like one of the whole points of hb2 is supposed to be perfect and we're not quite there yet. So it may be that the signal our folks are just sort of like well, let's, let's wait to.

B

F

G

But it's there so please try it and let us know how it goes if you hit by this, please file them performance issues. Please let us know so: that's HB 2, a next one I want to talk about is the a spec core module yeah go ahead, so I, don't think I'm gonna, actually there's not much to demo on this one, but this is a moving moving.

G

A spec core apps to run in is in process is something I've been working on for a little while the the diagram on the left is how the a spec core apps work in is today, where the process forwards, requests to the don''t eggsy process, where your app actually runs in in the in process model, which we are previewing with preview one, instead a spec or actually then, and the.net core itself runs in process in the eye worker process and that it gives us a much better performance.

G

We've also have some done a bunch of work to give us some release agility. As part of the of this work, where a lot of ace of the ACE met core model can be updated. As part of like bringing new packages from your application, so remind.

B

Me it's been years: I is worker process, that's w3, w p exe yeah.

G

E

B

In the context on the left there, if dotnet exe poof's, if it disappears, if it, if an exception, rolls all the way up and that thing dies mm-hmm this a NCM in the auto process, context fires up another yeah.

G

And, or or like one common problem that we hear about a lot is: if the don''t exit process fails to start okay, and then you just get an ACM can try to restart it, but it doesn't it's. It keeps failing to start, and then you get like a 502 touch style error where, like your application, failed to start and that that's kind of tricky to diagnose, because you need to go figure out. Why did that?

G

Why did it on an exit process have an issue with startup like you get through a second exception at some point, something you go digging through logs to figure that out one of the nice things about the in process model. Is that diagnosing any issue with your startup logic becomes a lot easier because you're debugging it right within the iOS worker process itself? Okay and performance is like almost four times faster like this. Is that our latest numbers from our internal benchmarks?

G

This is the returning a chunk of Jason benchmark from my blue from Tekken power. The green line at the bottom is the performance of is hosting with the 2-1 bits and the sort of pinkish line at the top. That's with the new in process model and in a snake or two too so quite a bit better in terms of performance. Now, with the preview one bits getting actually using the the new imprecise ma in proc model does require a little work. You do have to set an MS build property in your CS proj file.

G

You have to then publish your app and then and then you should be good to go on yes, but it's not something that you'll get out of the box with the f5 experience in visual studio until a I think like 59 preview, to is, when we plan to actually have that integration done. So the dev experience will take a little longer, but in four deployments you can at least try it in with PB 1 bits, and we are getting a little long on time and I want to make sure.

G

We give count the time to show the signal our client. So why don't we then move to sitting all.

B

Right so let's switch over here and then I also have to go film as your Friday. Soon to that is HTTP HTML I'm gonna upgrade to WebSockets right now. Here we go cool.

C

So yeah here, I'm just gonna, go over really briefly. Some of the signal, our Java client and.

B

This is out or.

C

Coming what's the status of this.

B

C

Technically there is a like staged public buildable, but our official preview on BET's will be available this week. Hopefully we.

G

Have a candidate.

C

Build cool so I just want.

G

To show you, how, like you.

C

G

Out with this particular sample on the signal, our samples we build yeah.

E

G

Poking around in the signal, our samples, repo, you can see the code that he's gonna show today and.

B

I'm gonna give you an opportunity to make your fonts bigger before. Oh, that's right before I, publicly, shame you into making the font you can do that, but it's not a live view. So as long as you're, not changing any good, that's using zoom it yeah yeah. You can't.

C

Like type sure, okay cool, so basically, this is just a java application. What.

B

C

Idea are you in right here? This is IntelliJ IDEA, so, basically without here what we have to do for now, because it's not a public build. Yet you have to define our staging repository and then from there we can add the signal, our dependency mm-hmm. So we just add another sort of node within our dependent. These tags, and this.

B

Is like maven yeah? This is maven okay, so this is the Java nougat. This is like evil, evil Spock with the beard and.

C

Whole thing the.

B

Dark side, the dark side, okay, cool. So in a parallel universe, you say: comm dot, Microsoft at asp.net, yeah, artifact, 86,.

C

R and then the version number and then, if we go over to chat I, just wrote a simple chat. Application mm-hmm, which news up a hub connection, takes input from the scanner for defining, which signal our server. You want to connect to. We define our on handlers, so when we receive an indication from the server with the target of send, we basically just take the arguments from that invocation from the server and we can do whatever logic we want here. We just print it. It.

B

Looks it looks and feels surprisingly like c-sharp like yeah.

C

That kind of like Spanish.

B

And Italian kind of thing like these languages.

C

May be related, that was a design goal, I believe yeah yeah. That was something that I, like was one of my main goals is like making the API sort of feel the same right.

E

C

So when you're writing the Java, it you're not like oh man, how do I do this Java specific, fitting? It's like okay, this is pretty much very similar to what we have in c-sharp. There will be some differences. Mm-Hmm I wanted it to really sort of feel the same right.

B

But you also are balancing the it should feel like it's been a net signal. R2 signal are familiar people, but it should also be idiomatic Java and using Java isms. It's.

C

Like one thing, it's like everything is like starting with lowercase, which is super annoying, but that's just.

B

E

C

Is super noon? It is if I zoom out. If we run this, you can see it'll ask us the URL to our chat. We want to join, and so this is an application. I have running the same.

B

Asp JSP dotnet.

C

Application with Signori reconfigured, we specified the hub URL. So that's a.

G

C

G

Yes, for that court hardened man excuse.

C

Me that's confuse people, and this is the application here running live so if I have a different client here, this is using the JavaScript client. If I hit broadcast with the message browser we go back, we can see that we have this printed. We receive that invitation, so we can say from here hello console and if we hit son we should be able to go back and see that a little family goes. Oh yeah.

C

Virtual desktops yeah, we can see very similar, so this uses Android studio in general uses great or rather than maven. So we can see how we would bring in the signal our Java client here instead of or we still basically have to define this repo, the staging where you go. So it's very similar. So this is in our project level built that Gradle and then for the app level build build our cradle.

C

We have to in our dependencies similar to maven, add another node and we just define it the same way just isn't a different shape to online instead of like seven. So we have calm, Microsoft ASP, not bah, blah blah, and there will be information in the 2.2 preview, one blog post, with the updated like build members for, like the actual, like release preview, one build mm-hmm.

B

Yeah, how did you deep bug this stuff is just like a Wireshark kind of situation where you're like watching bits on the wire or you think how much of this are you able to abstract away, or are you looking at the protocol and watching things goodbye like byte by byte, like when.

C

You were writing when you're, like.

B

Oh I guess I got to write the Java client for Sigler now, like are you drawing on whiteboards, like byte alignment, trying to figure out how things work like? How low-level was this well I.

C

Mean essentially, if we have our hub protocol, expect our totally and.

B

C

Kind of just starts from okay. What is like the quickest way? I can just get something like written in Java to connect to a signal. Our server, like literally just find a WebSocket library plug in a URL I swarm at the handshake and then send it, and just.

F

Get anything get some.

B

C

Of ACK yeah yeah like just connect and like get a positive response mmm and then from there like making it look more and more like the client. So.

E

C

Writing out all this stuff and figuring out like making it look professional from the first go. It's like a literally get start to work and then build the client around that and then sort of make it look more real and have it have. It feel like the c-sharp in JavaScript client did.

B

You bump into any problems with Java any robustness issues or does it work pretty solid.

C

So one thing that you can see here is like for, like the you can probably go: oh I see the.

F

Android outrun.

B

The Android app in the iPhone skin, what is it it's.

C

Pixel, actually it's a.

B

Pixel yeah so yeah just like the way it goes yeah. It was correct.

C

You running hack, C or which I mean later using. This is just like the built-in, like out of the box and.

E

C

Which is like really impressive, because I've been doing Android development like back in school and it used to when you hit run, you could like go run to the library wait.

E

A kiss now like.

C

No, it's it's really working well in the house snappier. um So if we say hello.

C

And enjoy then we hit the side.

B

I'm afraid I need to run as your Friday they're waiting for me is that.

C

Sweets got Android running yeah, and this is just a simple app that I have. It will be posted in this signal, our samples repo until then I just have it like on my personal github, but you'll be able to see it there in the signal, our samples, repo, but yeah. This is signal are the Android or the Java client right on Android any.

B

Call it 0.1 right now, yeah.

C

B

Will you how you decide to run? I 1.0.

C

Release, I'm not super sure, I'm not really in charge of like the version numbers I believe.

G

The planets are shipping, in line with the tutorial.

C

G

C

G

Change, of course, as always, because these are previews and we make decisions based on what we hear from customers and so forth.

B

G

B

Go do is as your Friday.

G

B

A

F

G

A couple quick, more things.

D

G

Talk a little bit about routing, so this are. These are the numbers that we're seeing now with the MBC route, the new endpoint routing system in a snit cor 2:2 we're seeing about a 10 percent performance bump with the new routing annotation. So when you think about endpoint routing, you should think about is basically a rewrite of routing that I think. That's honestly, the the right way to think about it. There's there's the existing routing system.

G

That's in a state court today that you're using if you're, using MVC your razor pages or building web api's we're doing another turn of the crank and though there's a couple reasons for that, one is to get you know. Performance benefits like this. Another is because we really would like the routing concepts to apply broadly across a snit core. You should be able to use link generation from middleware to route, to a hub or to an API action or to a to a controller action on in your in your application.

G

So that's that's the goal, as we.

E

Changed already, we kind of did some stuff with that with like way back in the day, routing came out with MVC and then we're like. Oh, this should be plumbed in through everything like vacuuming with web worms right, we kind of rolled it back into routing being kind of a more global concept, I hadn't. So it's.

G

True, it's true that the like routing the other, where is, is a decoupled piece, but each route is basically it's. um Each route is kind of like its own world of implementation, like you have to execute the route code in order to figure out what does it do and all of that logic really ends up being MBC logic like there's. No, there's no notion of that being something that's sort of independent that you can.

G

You could use from like a middleware like it's, so it's there's a lot of stuff that happens in routing that actually happens at the MVC layer. That's built upon on top of the core routing abstractions, but it ends up being really only useful the MVC. As a result, we're gonna really like to make that functionality more broadly applicable. It's really what this is all about. So.

E

And I hope you don't mind thrown in a few random questions here. We've taught I'll keep I'm sure it because I know we're time wise and all that, but we've talked about like hey, it would be nice to have them, have them available in middleware. What are some actual like uses of that yeah.

G

I would I actually.

E

Make use of that.

G

So like, if you want to have middleware that generates anything with links like generates a either a page or errors or whatever that needs to say and by the way, this is where the like: here's, my privacy page or here's, my trying to think would it be a great example of this.

G

If you wanted to generate documentation, sir, for your API honestly, you could do this at this layer now by scan, you can inspect and generate links for anything in your API from from a middleware layer as and have it be a reusable component. So that's something you could do as opposed to having that having to have that live in the MVC layers. Okay,.

E

Yeah- and maybe this fits in Eugene, just ask you a question: does: does endpoint Rowdy make it easier to support adding new routes to MVC controllers dynamically, not just at startup, but I'd? Remove routes like that map to our controller.

G

It's not something that I far as I know that we've been taught it may be, but it's I don't know honestly the answer to that one. If this is a really as a dynamic support, it's more than more about making like the the route table, inspectable, making it making the notion of what the endpoints are known up front, making them things that you can route to up front without having to get to some point in the pipeline, where you didn't have to execute a bunch of code to figure out.

G

Well, what are all the endpoints that I could potentially link to. That's the that's. The high level idea, there's a quick app that.

E

Does have any changes to attribute routing, or does that tribute routing just continue to do what it does at.

G

The the programming model for HP browning works the same way the this one, the so like, you still add the same attributes to your controllers and actions. The underlying infrastructure, that's actually resolving the routes and doing the linking has been, is being replaced and so you're getting the new implementation of the of the routing system under the covers.

G

Let's just to show an example of how what you can do today, it's worth calling out that also with the endpoint routing it's a bit of a journey like we, we have the the vision of it being this. You know generic dispatcher, that you can route to any endpoint in your application and then there's what's actually implemented in preview, one.

G

The things that you'll be able to do in preview, one is you'll, be able to get a new service, which is this link generator service which is available through di, and you can then use that from pretty much anywhere in the application to generate links to MVC endpoints. Ok, so you can generate links to your MVC actions, your your razor pages, that's.

E

What I'm doing.

G

Here, like I'm in this application, I'm using the link generator to generate a link to my to my privacy page and if we just run this real quick I. Think I've put it on a map to the slash routing path. So we should be able to see that working.

G

In the future, of course, you'd like to be able to link to other things as well, and that's one of the items that we still need to go and implement. But with what we've got, we go slash Browdy, you know, I can see. I get a correct route to my to my privacy, privacy page based on the the path to the page on disk. So that's that's kind of cool is to start playing around with that.

G

If we're interested in particular to see how this affects existing applications, because there are some some you know anytime, you draw sort of reimplemented something there are some subtle, behavior changes that if that occur like what happens when you link to something that doesn't exist, the behavior right now is a little bit different like it. It will generate a link with an empty string as opposed to a link with values that just goes to a for or, for you know, we'd love to hear people what people think about that.

G

So try upgrading your applications with this new routing system and and see. If how much it we actually curious to see how much is a break like? How do is there anything? That's a overly disruptive that we actually need to align closer with the old routing system. The way it's controlled like how you get the new routing system is purely by saying that your compatibility version is to two so that by having that that in there you will get the new routing system.

G

If you need to roll it back, you can go into the embassy options and flip a switch and say use the old routing system instead of the new one. If you hit too many breaking changes and you need to to get something working, so that's that's how you do that? That's the new routing system- and the last thing I just wanted to show is health checks, real quick, so health checks, health checks is a way to stand up. A simple endpoint that infrastructure can hit to tell that your services and your application are actually alive.

G

It's a library we've actually been playing around with for for a while. We just have haven't, had a chance to completely bake it and get it shipping, but it's now shipping with two two: it's in the box in this application, I've added health checks to the application, and you can see I've added a single check. It has a just. A name is what this config name is, and then this check is saying my app is healthy.

G

If this particular config, this --tz-- in the application you can think of this is sort of like being a knowledge. Astute ekang, like is my database connection. Life is what this is trying to show. You could do something very similar, in fact in it, we hope to in the future have like sugar methods where it's like, you know, like add, DB context, check or or something like that. Where, with that, you would actually would check to make sure that your DB context is there and it's live and running.

G

But for now this is just sort of a simple mock version of a health check where I'm just checking to see, if do, I have a particular required config value, and this is how you can sort of set your definition of what being healthy means. Yeah.

C

G

And you can add as many checks as you want and then down below tada stand up the health check. End point you say just use health.

D

G

If you run this version of the application.

G

And if we go to the slash health, let's see if it's okay, so currently I'm healthy. So that's good I go back to my application and mess around with the config. Let's say like: let's get rid of this required config value that I need and run to be healthy, just save that, and if I refresh you can see now the be the endpoint is saying that you're on healthy. So you can then integrate these endpoints with things like like kubernetes, which Cooper News has infrastructure for checking like house health endpoints of your application.

G

You can wire it up with like the asthma Traffic Manager. It has capabilities of pointing it at the the health endpoints for your app and it will then be able to detect, is the app online or not, and if it's not online, like the traffic manager, could then roll traffic over to two other nodes in your application to make sure that your customers are still happy. You.

C

Can configure this so that, like the the on call engineers, they then get like text that say that oh, the whatever service is not running properly.

G

Yeah they could do that or even like, if you're like in the middle of a deployments like you deployed something and the deployment you know broke something you didn't play something correctly, you could stop the deployment and then roll back and correct things in that way. So that's what that does so, just in summary, here's all the stuff that you'll see in preview, one we show most of these things I think we showed the web api conventions that the last community stand up.

G

So if you missed that, you should go back and check the last recording burn off showed some of the new conventions that we're adding and the analyzer that allows you to, you know, add attributes to indicate which HB status codes, your actions returns, and then things like that, and we should have bits for you to actually use hopefully later this week.

G

E

For people that are watching want to keep up with that, where do they get the bit so they watch for that.

G

Yeah we will publish blog blog posts on the ACE minute blog the web, dev blog. As soon as the release goes, live, you'll, see the the releases go, live and the repos and yeah they'll be also be I believe on the.net site, the the courts I done it so you'll see a banner advertising. The preview there as well great okay,.

E

This is exciting, so I'm, actually relatively new on the dzo, also known as dramatic zoom out technology that we can support in that room exist. Well, it's not that I know of I know. Damien did some weird thing with OBS, so may I I, don't know what we do. Do we just dramatically just stop and also the recordings running from your side. Yeah.

G

Yeah, should we do the we're gonna, we're gonna fall backwards or anyone I get two three.