.Net Foundation Design Reviews, 28 May 2019

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: .NET Design Review: GitHub Quick Reviews

Description

00:00:00 - Approved: SslStream.AuthenticateAsServer/ClientAsync methods should default CancellationToken to default(CancellationToken) https://github.com/dotnet/corefx/issues/35283#issuecomment-496610570
00:26:43 - Approved: Add Environment.LongTickCount https://github.com/dotnet/corefx/issues/35499#issuecomment-496612786
00:32:01 - Rejected: Consider moving RuntimeHelper.GetSubArray to Array https://github.com/dotnet/corefx/issues/36741#issuecomment-496613684
00:34:26 - Approved: Remove in and readonly ref from Activity apis https://github.com/dotnet/corefx/issues/37396#issuecomment-496615275
00:38:47 - Rejected: Comparing Utf8JsonReader to default instance with the == operator https://github.com/dotnet/corefx/issues/37534#issuecomment-496618927
00:49:33 - Approved: StreamPipeReader/Writer should allow to leave underlying stream opened https://github.com/dotnet/corefx/issues/37701#issuecomment-496622595

A

All right so know we are live, I believe so. This is basically list of issues that we have I to my knowledge, there's nothing that the super urgent, except the one that Tom pink me on. So let's maybe first okay, take a look at this one here, I believe they talked to us before, because it looks very familiar. I. Think, though, the only thing they want to do here is just add, display units which I, don't think changes anything besides that you know what unit this thing is to be interpreted a.

A

B

Those display units is just an arbitrary streamer, not something that is interpretable at.

C

During runtime really I believe that's correct, I mean they already give you the scale right.

D

Is this going to be parsing at some point or.

A

Me I, don't believe it is I mean I. Think there was the special man within last time that this data is just surfaced in the UI, so that, when you look at this thing that he actually know what this is supposed to be.

D

Sorry I was not in the other discussion.

E

What you saw there? Basically, they I'm actually, because that the Ganges Ababa does not know any your localization issues right right.

D

Yeah this is my concern was, if it is you I, so you have to support localization.

A

Yet this is an evolution of the diagnostic contrary. If I we ended, which is effectively.

A

A reprimand even for the API, the basic, gives you metadata and other things that is actually used. I'm actually dicta is a difficult idea. Old, API, yeah I believe that what you said.

A

D

I'm just wondering why we uses the word spirit: I mean like white, cannot be like name units rate time scale. That's a really healthy name, I! Think.

B

It's to clarify that it's for display purposes and not for other purposes, because you might have a name and a display name and a unit into display unit and they're, logically distinct things.

B

E

E

Maybe you can think the other I was watching.

C

A

Either way yeah but I think like this it last time you place it very similar things.

A

We'll just leave it, as is because they already have basically the last time will be approved to stay at a bunch of derived classes. So we have to be a tavern yet event counter, so they added polling, counter, incrementing event, counter and income at the pool encounter, and then we added a display name property, which is will be displayed in the payload for event counter the property a user can set it. The user wants a separate field that represents how this counter should be displayed on a viewer.

A

It is different from the name of this counter. For example, a counter may be instantiated with your name request counter, and it's explained everybody can be set to a number of requests which would be more friendly if you are friendly name. This is useful when trying to create counters interact with the CLI. We're having spaces in the name might cause trouble or confusion for the CLI parser and the user actually so framing doesn't provide.

E

The string kits the yes and then.

A

They have a generic metadata API, where you can obviously add key value pairs of strings, so they have, for example, priority. That's a few others that they already said, and this is basically another one. It's just I think most of these things ended up ending up in the property back, so they're, better displayed by you know a CLI or some sort of viewer, so yeah, there's no parsing.

D

Okay, so yeah, it looks good in just. We need to have a comment like we color crisis is our documentation. Yes, this is just.

D

D

F

F

F

Is the plan here to add this property to the base type diagnostic counter.

A

That's how I would read it? Yes, I think it's the event counter.

F

But if I couldn't find how I meant called related.

A

You couldn't find what I don't.

F

Qualify how event counters related like increment incrementing, encounter implements diagnostic comfort.

F

And the same is incrementing polling counter. Let's hear it, it's a memory.

A

Yeah I think they inherit from diagnosed account yeah we're probably on diagnostic counter, because that's where the display name for natives to.

B

Yeah, that's what he says at the bottom: puzzled added to diagnose fortune right, 400.

A

Damages provide this to make sure that it is actually.

A

A

A

F

So displays a rate display rate is on the concrete type, but display name and units would be on the base type.

A

I mean that international I think this was because the the types are not the same. I believe.

A

I mean this was the one even being pre-approved previously right. So I, don't think you need to reopen that one, but I think this one is just the screen lights. It would be similar to display name, I, guess: okay,.

A

You, but in the Canaries.

C

C

A

A

A

B

So, presumably when playing, they would have the display name as essentially the column, header and then they'd have their value and then post fix it with display.

C

I believe so what.

B

If it's a prefect.

B

Like what, if it's a dollar side,.

A

I, don't agree mother mommy but yeah, but it wouldn't work again.

D

But how are the eyes decided to put what they are I mean is a prefix though well.

B

I mean if this is for display purposes, yeah and you're, measuring something where it's a prefix display. You.

G

B

Rather than a postfix display unit, you might.

H

B

To know which side the up decides what to put there and let those work how it can be interpreted well, unless it's being configured by one app and then being interpreted by another, like you might have a performance profiler that sets up all of this, and then you might have something else with like say: benchmark net, which is the consuming that data. They.

E

Have a capability of adding metadata data into the table? Okay, dusty! Does the work work on there right, so I can do the measures.

D

Yeah I mean the guy doesn't provide any formatting.

D

Want to support this, you have to be white light on formatting for burial, whatever.

E

Also, the the and a unicast, a separate corn, completely selfish yeah.

D

I mean his: this scenario is valid, I mean the either. You have prefect said something like the column. You don't know if you couldn't get comes before or after anything,.

E

Is more like a thank.

A

E

Can also use a format format like there with the zero buckets unit. Technically, no, you.

D

E

D

E

D

Is like how they know that if it is predicted or not, you don't, but you can use a format as the display of the tried you.

A

In the format string, basically but I think that will not work necessarily depending on the beta's visualizer I.

B

Was like I think, given the existing API in adding this there's no way other than doing add metadata to relay how the ultimate string should be displayed when combining display units with value.

B

And so I was saying, if we're going at this, because it's common enough and it might be worth considering whether or not adding something that dictates how this string should be put together.

D

Yeah this also, this can be added later. If we.

B

Like maybe a display string, which it's.

D

Like they're, like currency format, today, you invite the format for the current thing like these are the currency comes before or after the numbers, three spaces or not.

F

As an aside, should we add the label annotation to the API, his laser golden display, name hazard?

F

So you see it again. They add, like question, mark strain question mark display units Oh.

F

Somebody favorite move, the chair is blocking the camera. You.

A

Don't you don't like looking at the back of a chair, yeah.

D

It's what you do cannot move it yourself.

G

Buddy units returned an empty string or no, if it's not set I'm.

B

Gonna hope an empty string in.

G

Which case it has the right annotations already right.

F

Except except display name is set to none nullable, so I would imagine or follow that no.

B

Right, but this is a new API rather than an existing one. No.

G

But that was new in three hours: I think.

F

I thought so do this well yeah.

F

Yes, it is display, name is also new. We just annotated to be nullable.

F

I mean in all the places where display name is used. If it's null the data empty, they said the empty empty string, but I don't know if you should be consistent or not.

A

I'll be happy with that.

G

Except for we probably want to revisit this play name as well. If this one is not knowable.

A

A

A

A

All right, so you looked at it, didn't approve it yet, because there are questions but good enough.

C

I believe they're currently sorted all this 10 years.

A

Bubbly will be good to have Levi our German here to talk to there's even the Atlantic magazine, yeah.

G

So we added some methods that take cancellation token and we didn't add default values for them. The question is, should we add equals default in these api's.

A

Positives and all the cider was it I.

G

Think it we just didn't take about it, so you know: do we kind of one a rule moving forward that cancellation tokens in this position end up as having a default we've been adding that for all the I think all the api's we've added with a cancellation tokens in the last year, we've added defaults for- and this was just in two one- there are other places as well, where we don't have them. So if we want to do it here, we could revisit those as well like WebSocket.

G

It's send and receive methods, take cancelation tokens, but they're, not they don't have default value.

A

As it seems like an excellent analyzer to have to an awesome.

A

Yeah under anybody objects, I've.

G

Asked this before but declare this is totally non breaking right. Just making an existing parameter have a default value. It doesn't cause any problems. It.

A

Shouldn't I mean it usually depends on on the exact overloads of the. If the method isn't overloaded, it's obviously safe. If there are overloads, then the question is: can you construct cases where the compiler can't decide which ones to pick but for those kind of things I would assume? That's not the case. I mean if we already have novel all it takes only one argument, then that's probably bad right, but yeah.

G

We wouldn't add the tonight, but.

A

A

Like long T coming up.

G

So I think email email. You were on that's right with this one, but basically we have environment account which returns an end. That means that, after 24 days or 48 days, whatever it is, it rolls around and question is: should we have one that effectively? Doesn't you know that, along with Windows, there's get to account 64 I believe there's an equivalent on Linux or UNIX as well?

G

So should we have one that's long instead of it.

A

Yeah I don't see a problem with it. I think the only discussion about like how do we name is gonna rain. How.

E

Did you call the stream the bottom line, because there's also similar property I? Think a mistake is to write the language, no.

A

I think, on stream, it's its length and position above lungs I, don't.

B

Yeah for ready, we have one line and I think in other places. We'd use blob-like, but the framework design guidelines always say: don't use language, specific names like law, so.

A

Yeah, we usually like we usually differentiate by their by their names. Why? It's so like you know! We have right into or like reading, 32 verses, read in 64 and there's some ideas on process that then brought up ever just savages with 64 moved on that also works.

A

I think t count, 64 seems reasonable, because then.

D

It's like it's also nice.

A

To each other or.

D

Can we have it like fall tech can't work? What font, ful and alt account yeah.

A

Whenever it is long ticking follower.

H

Will not be alive.

A

Man, stuff runs on c-sharp I get concerned, though I see my after lie being coping with my reference exceptions.

A

All right so yeah, honestly t-con 64 makes the most sense to me because at least it looks similar. We don't process I think long would be weird yeah.

B

I think long is the most consistent with what we've done in other places, but it count 64 makes sense from discoverability and framework design guideline. So.

A

You see other names, but honestly the only thing I can think of is literally literally long wings, Manos.

I

Used it in a few plays yeah monitors. Not this.

E

I

A

To say that the most consistent will be in 64 I mean that's the one we usually use every single time he's review ain't but I, but I think.

B

Tick cap 64 still differentiate, since it's a that.

D

A

B

Integer and generally.

A

Speaking I would not add in 64 if the other one is not suffix with the type name like I think it makes more sense of everything, a suffix of the type name, but in case we just have to differentiate between 32 and 64-bit.

A

E

Like you said, 128 is gonna like expiring, like for millionaires, but what, if we change, the accuracy might be much.

B

Where I don't think we're ever going to change this from 100 nanosecond tix, because everything relies on a big 100 nanosecond and also the 64-bit overflowing is with like significantly more accuracy than 100 nanoseconds. Don't like won't expire ever it's enough to like it's 4 billion ^, 4 billion. So.

D

To bid all right, yeah or the time today is long.

A

All right any other concerns. Otherwise that's what we will approve.

G

We decided on two-count, 64 yeah.

A

C

D

A

I mean we're on fire today, all right.

A

It's talking about this one because hopefully we can just say it's good.

A

Does anybody care.

D

A

I would discuss the work like.

D

I didn't hear like a strong opinion. We have to move it, but it's my yeah.

B

I thought that the the feedback was whether David Wright thought there was a need for it, and I thought he had replied saying if we think there's more general use than already is fine.

B

B

Very bottomless: oh yeah, yeah, no objection removing it comes like other method.

A

I think if this one, the only sticking point, is to compile and to change right.

G

I think at this point it's probably worth leaving it where it is. The only reason I was questioning it I think I was the one who was questioning the original issue. Is you know, looking over our own code, there are lots of places where we do things like new array, source dot copy and then, like returns, basically implementing this manually in a bunch of places.

H

G

We can use the language syntax to do it so, and you know we're if you're allocating and copying an array, the perf cost of intent versus range is nothing.

G

It just felt a little weird to me that we had this helper that was buried and runtime helpers rather than every.

A

G

Related help, but that's on a ray degree.

A

I mean you could also just literally add the other API and I. Didn't just call this one insanity, you.

G

Always added in the future, and it became useful all.

D

Right and it's just saying it's a compiler syntax is enough.

D

D

D

All right, you.

A

So idea about these guys here.

A

A very context in this one bubbling all because Jeremy's the machine.

G

You should discuss this women charmiest here, yeah.

A

I think so let me just check whether he's actually because in some way, yeah he's bounced away. Alright, let's give this one.

A

This one seems like I need somebody for me like no 16 and there's. Somebody here has super strong opinions on.

B

It looks like this is already approved and we're just and then wants to change it from refereed only to not because he was able to shrink the size.

B

B

Not remove it, he just wants to change how it's being returned and past yeah.

A

Basically, once I, maybe instead of record only just copy this yeah.

A

Yeah I, don't have any concerns in that particular thing like I. Have no idea honestly with the applications. Have an operator is in I supposed to shop? Does it transparently, so the reverse would also be transparent, good right.

B

It's c-sharp lets you annotate the operators and if you only have operators which are in it, we'll choose those. But if you have both ones that are in and not in, then it will only ever choose the ones that are not in the.

A

Base it, if you leave them, they just become dead code. You kind of called it mystically right.

B

Except for a through reflection, yeah, which is there or if.

C

They ever expose a way to differentiate.

A

Yes suppose that's what we generally drew. Why wasn't there an email conversation recently about how in is also expensive at times right.

B

Because you can have hidden copies even within when you end up invoking an instance method which might you take let's see and then, if you're smaller than the register size, it's going to be more expensive, because you have to pass by reference then, rather than in register.

B

A

Do you have any opinions on this I think.

G

The if the rest read only in the in our unnecessary or worse hurting its slurred for the API to not have them so I'm all forward. If it's not going to negatively impact anything.

A

Who's the owner for these api's is that a Pollock butcher, his name, but is a strong Jonghyun Wang like I, think he was the person that submitted the event color API proposal. So I guess he is I. Don't.

G

I'm, not sure fans had been doing a decent amount of the work.

G

I would probably start with Noah and.

G

Yeah I started No, maybe Brian.

B

Ya know uh um Saddam and anything.

B

Yeah and a root and.

D

Maybe I think Tom.

C

People digging for edits.

A

A

I think it was shows he may not get an email and it will show up in the dashboard. I know whatever that.

B

A

Get too many so I see it works. Well, it's as well. You cannot on no chick.

A

A

Yeah, what is this.

A

G

J

Know what this.

G

Is so this came about from some feedback that someone was basically, they won't even want to try pattern. They wanted to create a UTI, Jason reader and return it, but then they wanted their API to potentially just like an try basically be able to return kind of equivalents of null so that the caller could check and say like. Did you create one or not I, um as.

A

Opposed to comparing the default explicitly.

G

Which? There's no way to do well.

A

What can you say, Jason reader everybody, since it's double equals equal, doesn't.

G

Have operator equals bust, it's a straight destruct and it doesn't have operator equal to find this API proposal is potentially is requesting adding an operator equal.

A

How do you always have a W equals M stress? There's no.

G

You have equals. You have like the equals method, which just defaults from the base implementation to doing in a comparison of each of the fields. Basically very.

B

G

I think they're I mean my I understand why it's being requested the whole proposal feels a little business gets to me. Certainly, the operators I'm not sure, make a whole lot of sense because, as was pointed out like what does it mean for what is the? What is the meaning of equals? How do you define it on these readers, yeah.

A

We don't know I would say, like you, I mean in stuff all I've used about myself, because this convenient even on things that are super easy. It's usually nice to have by.

C

We're gonna do.

G

It we should I think we should do it as the is default. Yeah.

C

I, don't think.

G

We should write.

B

It seems like the Equality wants. One only make sense that these were like reference types.

A

The other hand for refs trucks that you're not very likely to have them as default guide like it. You can only have there was somebody literally explicitly create service default right right.

B

Because you can't just.

A

Put them as a field, whatever the weather usually would get them on it. Wise.

B

You can, if it's a field of another restaurant, sure.

A

In which case you are in the 0.03% video.

B

All right, but it isn't that the one of the reasons why this is a.

A

Ref struct yeah I guess you could argue the whole fiber than the point of three percent. So I go booth. In that said, maybe they usage is higher. The.

G

Reason this was a ref struct is so that it can store span.

A

Yeah I don't have a problem with this per se. I mean that gives us all. Is the user pattern that you have for things that we think it's coming to the offset in specially code like this seems reasonable right.

G

I mean yes and no I mean yeah sure, but you could also just try and read from it like I would imagine the majority of cases just try and read and they're told and there's nothing more to read.

A

Yeah I guess it depends on whom you see this like? Is it against the reader? It's not something we would be very likely passing along between public API. It's probably more implementation that will tie me. Yeah I mean right.

G

That's for ambivalent about this.

A

Yeah I don't think it would be horrible, but I remember it seems even.

A

B

B

B

Do we really want a field between default and a new instance? That's created that is effectively the same as default. So if.

G

You scroll up. That was my question to him right.

B

And he responded with, he was relying on having some field to do that right.

G

So that's I mean there. There is one field today that which I mentioned that can be used for that. But you know it's implementation detail. So if, in the future that were to change, we might lose that and we'd have that something I mean.

A

Generally speaking, that's well. How is default works anywhere else right. You basically just compared to the old zero pattern, whatever it is right instruct right.

G

The problem is that the all-0 pattern is a valid result of an actual construction sure.

B

Right because default state is still valid. It's just you can't do anything useful with it. If you're trying to read in some scenarios.

A

Yeah I mean the property itself makes no statement just says: I'm the default value. It's like, if that's valid, for me, then great. If it's not, then I mean take. For example, the text span that has to in so I starred in the length is default is completely reasonable right and it just is a single. You know: zero span with on the first character and that's a valid radio, but still there are some cases where you just want to differentiate them from actually write.

G

The question that was why you care about is default if the scenario that caused this was wanting to know whether one was constructed or not like, are you getting back because the person didn't like the try pattern, so they wanted to use something where they could just return the reader and in the case where they returned, would have returned true, while the reader is valid in the case where they would have returned false to just returning the default.

G

But in the text span example you gave it would be perfectly valid scenario to be returning true and a text ban of offset zero length to zero, but.

B

I educate, so my comment was basically that I don't think. That's necessarily a useful scenario.

G

Which is an evil scenario being.

B

Able to differentiate between someone saying return, default and return new constructor that is effectively default.

E

G

I, don't disagree with you I, don't think this is all thing is valuable in general, but if, if you're trying to differentiate, if you're trying to do that, try thing, then by definition you care about that.

B

Well, I'm not sure because the its it's basically a question of the is default question is basically: is there any data I can read or not it's a quick check for that.

G

I, don't think that's true, you can be. You can be a non-default reader and have zero data to read.

E

Should be acceptable.

E

It seems like I, don't.

A

Understand that the original problem like here, there was another issue where I think you responded with. That seems like the problem of you on making like, if you don't want to use to try better. No, you have to leave more cases like that. I, don't understand what the problem is with using the try better, but that's what it's for right. They should say, give me data and the answer might be yes or no, and then you just move on with your life I.

J

A

A

All right, I just called it, and let me let me take a stronger stance here, because otherwise you're not making forward progress.

A

B

It looks like asp.net right now is just doing if not read or not read, bro knew invalid data, except.

G

Which may expense maybe actually perform the operation you really care about, and if you don't get back the results, you expect anything yeah.

B

They're, basically just using and are we at the end of the payload check.

B

This would be the mincing of yours right, in which case, maybe the correct property to expose. If you were to is a is and is end of stream or isn't the reader or something like that which effectively says do I? Have any data like to read right.

G

Probably already have that, is it and they're like a try peak or something.

B

Right but it's a it becomes a question of like I, guess, usability and friendliness like on stream. We've got a is end of stream property, but you could also peek into a bunch of in other ways, but we've got that dedicated property for the carbon case.

G

Yeah, although that's problematic in and of itself, for other reasons,.

A

Yeah I mean we don't have trip' as it seems it is not.

G

I'm, not thinking of a sequence reader, think.

A

You think I can sequence yeah but yeah. If somebody wants to expose a different property that basically indicates that they have data left or whatever. That seems more reasonable to me than just default anyway, so we I posted my comments. I closed it so that get some urgency behind it. Nobody feels strong about it. I want to come all right, so this one. Let's refresh this page.

A

/ look at this one.

A

Look at this one.

J

E

E

It makes sense if you didn't ship it yet with a great sense. When you start just not, was it ever.

H

E

Hey hey: do we have reshipped pipelines, yeah.

G

We should pipelines, but I, don't think we should this API I, don't think we should we ship stream right breeder options. Well,.

A

Because it's literally new in this release, do we to be able to wrap a pipe over stream. What.

B

Am I advice, did we ship.

A

B

Pipe rider options as well: I think those are both no okay, because at the same time, if we're changing this, that buffer size shouldn't be 4096 screen previously said, it should be some non explicit value that way we can change it without and allow any app to using it to update some pointless one or something right.

G

B

Is what we've done on other API? Is that way we just change it under the covers.

G

To answer I figured who asked but I think there are scenarios for both leaving open and not just like. There are, with all of our streams that a couple will leave open.

A

B

So the other api's that do this take in encoding a buffer size and they believe open and sometimes they detected coming for Bible.

C

Evening news news.

G

Mother I mean we did we do this elsewhere, like on the plate. Stream takes a pool about whether it check out the underlying street gzipstream right.

B

I was just giving the list of what the other eight guys are taking yeah.

A

I mean the problem is, most of them are not defaulted, so calling them.

B

An open is that we change them all to be defaulted well,.

A

Some of them we did one.

E

Thing we checked them on distance if this isn't, if we just remove all of the arguments and the work and then basically default or there are you we want, and then, if someone wants to check out, you can just explicitly set that right. But that's what these are I'm. Sorry, they just haven't. The gutters I think that.

A

They're, the the type was designed to be read only.

A

Anyway, so I think this looks fine to me. I mean model the other change here, but that I regarded I think they believe openSUSE problematic. So now.

G

Can you scroll down I think there's more to it? Yeah.

G

To that Andrews proposal, because we had commented that the previous proposal was only covering one direction, it was only covering the I was in the pipe wrapping. The stream and there's also the direction of the stream wrapping the pipe and so I think Andrews proposal incorporates both directions.

F

F

It's incorporating both, which is awesome, I only see going one way right: one pie, pie, pie, brighter Lorraine.

F

Yeah, a stream hazard- oh you mean the constructor of team. Okay, bye-bye officers has it the.

G

Stream pipe rider options and stream pipe reader options are this proposal adds to that, so that covers the creating a pipe around the stream and then, as stream, covers creating a stream around the pipe.

F

G

You're right call.

F

A

I mean it looks very reasonable to me, but with its more symmetric but against that wasn't part of the original design so situated.

G

Do you want to address Tanner's question about the yeah.

B

Because if we hard-code it, then we change it in the future because of better improvements than anyone doing roll forward both pick it up. They have to recompile to get it.

A

So hopefully we should all separate issue for that, though, because that's kind of like orthogonal to this one here right, if you discuss them all in the same one and which is tougher than the woman I, think.

B

I can open one right now. Look.

A

Over the easiest one and then assign the two walls pipes.

G

Decided to enter, he seems to have an interesting.

C

C

C

F

It's a guideline for all api's that extreme to have these buffer size encoding boot leave open parameters. Why.

B

F

B

All of them have them I think it's mostly a case-by-case basis. It looks like like a binary reader writer doesn't look to have an API that takes off for size, but stream readers rewriter do.

B

D

A

H

C

Honestly, a service-based seems like a really weird me who's, the area over for once that her engine.

A

Or equal dignity.

E

You does it have to have windows in the name like a flag. We ever have a way of checking that on.

A

And so mentally sure would.

D

What shampoos Chris? That's? What guys do it.

A

Service basis, the thing you're extending when you actually expose yourself as a Windows service. It basically has um you're basically deriving from the sky. Then you can override that things that gonna shut down, for example, and do whatever you want.

D

It looks weird place to check yeah.

A

D

Environment yeah, maybe environment, will be more reasonable. I was.

A

About to say the same thing because quite frankly,.

B

What is full runtime information? What if you wanted to check if it was like a eunuch statement instead, which is I'm in the same thing,.

A

Yeah, the question is whether you can detect that.

A

Yeah I think this one doesn't have enough information to decide what's useful.

A

E

The service basis window specific right across, but from aspirin on my neck. So it looks like it returned fault.

B

Their implementation.

C

A

Like white frenzy, I, don't care for the detection. I mean the detection is whatever it is, but.

D

A

D

Today, to detect it without exposing gaining your property Michigan, we have a way to detect if you are running in wonders and in service in same time today. Well.

A

I think the way they implemented it here was where they suggest checking the parent process right. The bassist said is the parent process.

A

D

They check it services, services, the name of the process itself and.

G

D

And it's easy to check if and so why this is not good enough anyway, we will no can do.

A

Well, I guess they expose the API because they don't want to spread this thing through all the codebase right.

A

The way he described it was that easy to hook me right.

A

A

But yeah I didn't appreciate, discussed this more me. I think you need to move to the other room, we're losing this one here. Alright, let me disconnect this one and then we can reconnect the other one.

D

We should collect the question mark.