Node.js Diagnostics Working Group, 17 Apr 2019

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Diagnostics WG meeting - Apri 17 2019

Description

https://github.com/nodejs/bootstrap/issues/114

A

So welcome to the no Deus Diagnostics working group for April 17 2019, we'll follow the agenda, as mentioned in issue number 292. If everybody who's here, I'm just gonna paste the link into this- that, if everybody's here can add themselves to the present, that would be good first on the agenda is: does anybody have any announcements that they'd like to share.

B

The perfetto repo is it's great- has been created thanks to Michael, for that I posted on the perfetto issue, the link to it. Something wants to follow it. Yes, sir yeah.

A

So that's nodes trying to remember what it was called. It was perfetto I'm, just gonna paste the link to that into the minutes. I.

A

Think people will get automatically subscribed, so I think.

B

Yes, he did. I saw a number of all the nine to start us, but I.

A

B

A

B

A

See I'm subscribed anyway, so does.

B

A

And any next steps, or whatever you want to call out on that front, I'm.

B

I'm building well I've been able to build a federal budget, but I'm just already know a few details there and then oh I have some code, but my national review process would be if it's something we're like the entire reviews for code would happen when I opened up here for the main brain for the main, no repo or if anything happens there, it.

A

Really comes down to how many active participants you have I think you know so, for example, for an API, we did do reviews. You know we asked for like one of their PR review before we landed things, but we've had quite a few people working on it. There is like a full review when you make the PR into core anyway right yeah. So if you're the one doing most of the work I, don't think it necessarily needs to slow down waiting for other people's reviews.

A

But if you end up with two or three people contributing you might want to, you know ask for one review, just this personal coordination or whatever, but I, don't think, there's any hard and fast rules in that front. Okay,.

B

Well, either way I think I'm gonna have the perfetto and chromium people look at the code right before over here. So okay.

A

It might be good to just like on the read. My suggestion was to update the readme so that it doesn't, you know it's different, so.

B

You have no one and you.

A

Might put some information on that? That's like hey! If you want to become a contributor or if you you know, here's how we're doing reviews you know at this point just land, whatever you want or or I mean other people won't be able to land, because I think I only added you to the team right, but it could be like one sided to the team, feel free to land things and we'll do a final review at the end or whatever yeah.

B

B

A

I'll end up closing out the issue in the repo and, if there's anything else that needs tweaking or whatever just let me know.

B

A

I think I think the one needman repo still open and we'll probably Constance open news do.

B

You mean the one in that I'm in repo yeah.

A

Like I think I'll.

B

A

Back and say: hey it's created, let's, okay, I think I ready closed it. Okay,.

B

A

Already closed, even better yeah, perfect, okay, but.

B

Working a diagnosis window open right, the.

A

Diagnostics, we only made sure yeah, okay, okay, yeah, okay, any other announcements. I am.

C

Still working on the use cases and user journeys and I really wanted to share on this before this meeting, but I am little bit behind my schedule. I really hope I can share it tomorrow with the working group and we can discuss it in two weeks. Hopefully.

C

A

Okay, I'm, so moving on to items on the agenda, the first one is support: pausing the debugger on strip load. That's 2467,.

A

B

Think that was there last time. Okay um I took an action at him to check in on that see. Oh, that was a stagnated because there's no opended left google right got it. Okay, um but I've gotten to that. Yet! But I'm someone I'm.

A

Just going to add that come to the issue itself, so that's we go to the end. Maybe it's there, but not okay, okay! Next time. Next thing on the agenda is horse, more lore and debugging support inside of v8 no GS, which is to 270 I.

D

Don't have any updates on that I think we can remove from the agenda.

D

We had some alignments on the Diagnostics summit and I think now we have to wait for the work to be done, especially on v8 or migrating the objects to to torque. And after that we can start discussing this again.

A

Oh, he did okay.

D

A

Okay and then some moving on to the next one diagnostic best practices guide, Jewish, isn't here so I, don't know if he's made any progress, but I think that'll work that Peters working almost directly related to that.

A

E

A

We've had no I think it doesn't provided something feedback and so as to reach so I think we're progressing on that front.

A

Okay, diagnostic good guys, I, don't think.

A

Anybody else has not big; those are something so.

C

After I reassured the user, Chinese I think we should definitely do deep. Dives on a specific use case is to agree on recommended tools, best practices, what the how the future could look like. So this is just a heads up that I think there will be more topic in the future. I.

A

Think that would be a great way to push them forward to sort of get people together, bash it this sort of outline and what the thoughts are and then yet baby.

C

A

Next, one expectation about tear support, really the focus is on defining the best practices, so I think we're just gonna drive the expectation on the tear support I think. Maybe even we could take the agenda item off of this one as well, until we get that done and then we'll sort of drive tagging or pushing those things based on that. Does that make sense to people?

A

Yes, okay, so let me remove the dye.

A

Okay, so the stable epi, a saqqaq, stable, API tracking, issued I guess we don't have.

A

Strength to think whose that was Ali who's tagged as the champion I, don't think I know of any progress on that front.

B

Yeah I haven't talked to Italy much about the sin they there is that new, you think, storage like yeah, it's unopened after that. That's active I'm, not sure what this day. That is, it's.

A

Still being I like I, think, Vladimir is still trying to figure out a problem in one of the test cases where it doesn't like it loses contacts, okay,.

A

Okay, but nothing else really new in terms of getting out of experimental on that front. That might be something that, if we I don't know at the summit, if we have the right set of people, we might talk to about it as well. You think context, formalization and diagnostic support. You don't have my.

A

And then the last on this shoe is trace event, number 84, so I.

A

Guess you've already given us a bit of an update on that oh yeah.

B

I'm not aware of any other changes.

A

There was one person who might be in my trial he's had a couple of the friend rent them up, but I haven't seen anything happened from that. Yet so, okay, no, no progress on that front.

A

Okay, so that's the end of our agenda. Were there any other agenda items people want to put on or talk about.

E

mmm It's maybe interesting here, because now we have the flag for unhandled objections in note 4 and the question is: should we and decide on a new default for it and I personally, so we might talk about it at the summit, but just like briefly, because I'm not sure if we want to change it right away or not, but we could, for example, also have a survey from our users about this topic. What's the current, so the default is then nothing changed. Its.

A

E

And you can opt into having the strict mode, which means on uh no rejection. It's Rowan arrow or you can opt out certain of listening.

B

To any of those like.

E

Which is not notice here anymore or you can choose warrant, which is almost a default, but and without a duplication, warning.

A

A

E

If you use warn you opt in to explicitly, you want to be warned, but a bit way. We know. Okay, you don't need to play notice anymore, because you already know what you're viewing so.

B

E

And then there is no duplication warning anymore. In that case,.

F

This is separate mode, so.

E

It's a noise, I know it's it's on. You know.

E

No regular moaning would still be. There was born.

A

Yes, okay, got it; okay, so I.

E

Would like to have a survey involved us and that would be pretty cool, because I would like to ask our users what they would prefer in general, and also we should write up as something about it.

E

How am I no rejections in general work and what does it normally means and then ask our user base and what would you prefer to have as a general default, because we are not a browser and therefore we have different requirements for most people and also it would be interesting to see if there is a difference between different user bases, let's say: you're a clean user or in the company or I. Don't know what you know: I.

C

Personally, like the idea of being user driven here so I, like the survey and I stink, getting rid of the duplicated would be also a good step forward.

E

Yeah well, I mean at the moment it's still deprecated, because we didn't decide on any default for the future. I mean we did decide at some point that we actually want.

B

E

Shut down the application so similar to throwing as it is right now we stopped in, but then people were concerned about it again. I mean it's it's a long ongoing issue, yeah.

A

There was not enough consensus, I think yeah.

E

A

E

Consensus to to add the duplication warning in the first place, but really doing it.

B

E

Another thing and I guess: if we would ever do that, we would have to have a basis from the users right.

A

Yeah no I agree that the both like I, like Peter, says having it being user driven, would be a good idea, so we can do a user survey through with support of the user feedback team. We still pretty much need to to write up the survey, any questions that we want. So that's really the first step. The other thing we can do is in the user feedback team, we're working on a approach where we can actually get like in person back at meetups.

A

The first one is going to be about adoption of new releases and stuff, but it's a case where you know we can. We can also have you know, ask at some point that one of the questions that happens in those meetups is like. What's your feeling on this and then we would point them back to somewhere where they could do the result. You know it could comment and it might be through the survey or through something else or but I think writing up.

A

The survey is probably the first thing is like we wrote up a set of questions agreed on those questions, and then you know we need a champion to take that to the user feedback team and it should be pretty easy to get the survey out for some period of time. So.

E

It would definitely like to be determine it. Okay sounds.

A

A

And work on more questions, yeah.

E

So in that case, the question is: should we try to push it out before the summit I? Don't think it really makes sense to to push it out before the summit I.

A

Yeah I, don't think, there's probably enough time to get some input on the question. Really. Yes, I.

E

A

It up to get people's opinions at the summit. That's.

E

What I have just to say? That was exactly my point: that I could gather enough information, put everything together and where I believe with some good stages and and then we can finalize it on the summit. Yeah.

A

That makes more sense to me.

A

Good any other thoughts comments.

C

C

Have another neat picking topic if we are done with this topic, while I was reading the the diagnostic working group read me. Basically, what is our mission? I was really surprised to be honest and let me copy paste it into the into the text. This. This mission was kind of stated by someone three years ago, who is not part of the working group anymore and I'm really surprised that what we are focusing these building documents and api's for the vendors basically and I'm, not sure this is really the status anymore. I see.

C

Now we are more into the tooling and providing more complete solutions. So what would be the way to improve it? Or do you agree or disagree with me? I know it really needs wicking, and this was the first time that I was reading it, but I find it little bit weird I.

A

Think the thinking at the time was we didn't necessarily want to compete with anybody like who's providing tooling. So we wouldn't necessarily write a debugger, but the API which people could connect to with the debugger was not an issue, but you know and I don't have any strong feelings like if we basically wanted to trim it after the NDP is or something or the way to move forward, just create a PR that suggests what you think it might be updated to, and we can have a good conversation in there.

F

C

A

You're right, probably after three years, we have a bit of a different viewpoint.

A

Before we call it a meeting.

C

Kind of discusses at the beginning of the meeting, but Ruben you were on here, for example, that who is coming to Berlin.

C

C

Do you know my code of Ganesh is coming for my being model uh I.

A

Haven't heard I was definitely telling him to try and get approval and everything lined up, but I think the other I know he is traveling the next two weeks. Sorry I, don't know I'm, certainly encouraging him sense, Sam Roberts, who isn't actually part of the working group. It does have some interest on the diagnostic side. I've heard we'll be there as well. Okay,.

C

Cool you know Ruben, that the need from people in short I would expect this.

E

Should be there? Yes, okay,.

C

Then science I have some people from the working group there yeah.

G

Okay, cool, okay,.

A

G

A

C

A

I'm gonna take a quick question to see quick, look to see at all. There's a couple people watching one minute you have any questions on you can put them on the YouTube channel, see if we can answer them.

H

A

See somebody's got their hand, so Wentz yeah.

H

I wish to ask I wonder one question.

G

H

Larry there is a foot West not about us in context and since it can be connected with this point, 100 South, synchronization, Diagnostics, port and I have posted to link.

A

Right so that's 107, sorry or a tracer to let me open that one I think that is there.

H

Is it near the church and I think that yeah.

A

I'm just clicking on it right, yes, that that is the one we mentioned in terms of the wounded Vladimir trying to dissolve the test case. For so we have had a discussion in terms of the stable API for a Seahawks and one of the thoughts we had.

A

To sum it was that if we define an API for context, propagation, which is kind of like at a different higher level than async hooks, that may then mean that we don't need to focus as much on getting Easton hooks out of experimental, because we could, you know, change them and have this higher level. Api stay consistent and I. Think that discussion is what triggers vladimir to open that that issue. So it is related in that sense,.

A

Does that answer your question or.

H

Yeah, okay, okay, so.

A

Yeah I think we're we're all keeping an eye on that one, and you know I think that it'll be interesting to see if we can get something like that into core. Although I think.

B

We should be.

A

You know we don't need to rush. We should be a bit careful because you know people start using it and everything we don't have to change it afterwards.

A

Any other questions I.

A

Don't see any questions on the YouTube channel, then so there's nothing else, we'll call it a week and I will heads up. I am off on vacation two weeks from now, but I'm hoping Mike will be Basco he'll share the meeting and I'll catch up with you all through github between now and then, but after that, probably at the next meet.

A

G

You bye thanks.