Open Telemetry Uncategorized, 15 Jun 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: 2022-06-15 meeting

Description

Open Telemetry Meeting 1's Personal Meeting Room

A

A

I think we start we already couple of minutes.

B

Okay, yeah: your volume is a little low.

A

Who is it okay, let me see just a second.

A

A

A

There is some problem in my microphone: I'm just checking that.

B

Okay, uh yeah no problem yeah same a little better, then okay,.

A

Okay, I'll probably speak bit loud. I think, if that helps.

B

Okay, are you in the office or.

A

B

A

You know I'm at home I'll be going into office after the meeting. Is it better.

B

A

Think I was just trying to increase the pattern. Is it? Is it better? Now.

B

uh Yeah a little better yeah, okay,.

A

Okay, let's try it. Let me know if I'm not audible, just let.

A

Let me share the screen first and then.

B

Yeah, I can see your screen yeah.

A

Okay, so I think I've just added one of the items with mark is sent.

B

A

B

I mentioned yes.

A

What do you suggest should we.

B

B

I I'm fine to to disable the outer clothes, even I think then, the same rule applied to other open telemetry report. Right. I think I I copied the rule from the spec repo so and.

A

B

There's uh seven days alert um alert about.

A

B

And then, after the seven days it will be automatically closed but yeah. I think it is fine to to remove the auto close.

A

Yeah, I mean the only the uh I think in the specs repo. Probably they would be getting lots of issues and I think they probably they wanted to just automate it. Okay,.

B

A

That that that is a problem, I think it's, I think, even if we are manually, uh probably checking all those tail issues, I think it should be still. Okay.

B

Yeah, but uh that's right: yes, back, get more issues than us, so once if we can like, then this requires us to charge all like all issues right like regularly in a month or so. Yes,.

A

B

If we don't close them,.

A

Yes, we should start doing it.

B

Yeah, but we still need to charge them and see whether it deserves to be open and like don't stay or like some issue, still need to be closed right after some time. Nobody.

A

B

About yeah, okay,.

A

B

Let me, if respear, to disable the auto close for now.

B

Oh, do you think so? I think yeah.

A

That's really good.

B

A

A

I think, let's, let's monitor unless we are getting lots of issues. Probably we can always change this. I mean I don't know, I think it's okay, even if you just mark it as a scale and let's try it.

B

Yeah, that's right and for the same, I think we have similar concern because for maybe uh we still will have more issues than prs right. So we don't apply the same same policy on pr's because from prs we are, I think,.

A

We would better money trash even for clothes.

A

We don't have this action for pr right. It's only yeah.

B

We don't yeah, I think, that's our purpose. Yeah. We have to manage rgbr.

A

B

We don't want to out of close uh even mark stale of them.

A

Okay, yeah thanks.

A

I was thinking if we can start creating uh release milestones.

B

B

Start creating.

A

B

A

We should have have a milestone for each release with the date each future releases, which we are planning to create.

A

And add I mean like it's something like this probably.

A

I mean I just wanted to discuss and what do you think about this? So we already have project milestones here so like something like.

A

Okay and then probably, we can also start to.

A

Add the issues like if we feel that this grpc exporter for matrix should be part of this release.

A

We can okay.

B

A

Have the future releases, we can define couple of co, two three future releases for each of the ones.

B

A

Based on the priority of the issue, what we feel these I mean like matrix is: okay matrix would be part of a release candidate, but follow for all the other issues. Something like this one, which probably we can decide that if this should be part of uh sorry, which.

B

Is? Which milestone.

A

If you feel it's not higher priority, we can move it to the base of the amount of work and the priority. You can move it if we have release milestone defined for next two three months.

B

A

Probably move it to one of those microphones, so we will. We will have idea. Where do we for each of those releases? Where do we I mean? What's the current state of that? Is it good to.

A

Issues have been closed, which was decided for to be part of this release.

B

I'll see so, I think, maybe then we'll need to charge the release status regularly right to see what's a gap for their next release and yeah yeah. I think this makes it easy for us to track release and before release. We need to check whether all the associated items or issues are being resolved right or completed.

A

Yes, probably we have to spend some time in that case to really decide that this issue could go to which release yeah. We can take our probably we can in case I mean I I really it should be. I think it should be very straightforward to do that in case in case, if we have some gap or we don't really have clear understanding that, whether it's high, how how much priority it is, we can always discuss that. Yes, in the community meeting.

B

Okay, so yeah. That sounds good. I think um when to create the new release the like that, when to create that entry, what do you or what mouse in advance or right.

A

Yeah, I think we can have- I mean we can have entries, probably for next three months.

B

A

And then let me just delete this like this created together.

B

Yeah, I think once we we decided, like we have a rough date about the next release. We can create one or even more.

A

uh Okay, I think I mean I was thinking that so three milestones for next three days: okay,.

B

A

It because we we already have the cadence of one circular release once a month, so we can always have it. Let's say release one say we'll define it probably for if 24th of june is for the first release, then for the next release would be say july of 24 and.

B

A

Can have defined like that, based on that cadence and then we can, based on the priority and the amount of report required for these. We will move them to each of those milestones and.

B

A

Then probably I think we can see like right now. If I see this.

A

For the project, we will have an idea like if I go now here for the release candidate for sdk, I see it for june 30 and it's still 33 complete, so whether it's the right time to really have a release or net not. We can have an idea whether we should postpone this release or it's good to have a release. Now.

B

A

We're going to have a release all those pending issues. We can move it to the next release.

B

A

That looks good.

A

It's not something which I probably haven't had a thought we, this was something which was done already in microsoft, cpp telemetry.

B

A

It's not being followed, but I think I was this like. If you see here there used to be all these release, milestones defined and then probably they they used to have all those releases already part of those milestones.

B

A

Mean I just just probably got it inspired from here: if we can do something we'll.

B

Go back to look at the history of the open kind of recipe, uh people they're really in the milestone.

A

B

A

Microsoft would be very less that we don't have much.

B

A

Created the one.

B

For alpha release right, I think I remember you yeah.

A

B

A

We we only had it for probably alpha in the release for traces, and now we have it for matrix, but not the regular releases, which we have.

B

Okay, this sounds good yeah. I think this is a good tool I thought for us to to manage in their.

A

Progress yeah we'll have a good good, probably good, to really uh track the issues also- and somebody who comes to this and just want to understand that whenever you're going to fix this issue, they can just go to the milestones and see.

A

What's the plan for that.

A

A

B

Yeah, I think, for the for the release where we were targeting these this week, I'm going to create a milestone and yeah. Try to add. I think we almost even almost maybe not still a few atoms need to be added to the milestone, like the async right. That should.

A

B

Part of the milestone.

A

Probably I cannot say that you can yeah when you can create probably two three releases, let others be empty and let's so that I think we can start. uh At least I I can also go through the issues and based on the priority. I think I could start moving it to the particular a milestone.

B

A

And for the release which we planned for this week, we have a plan for this week or what is.

A

We had 28 days back, we can. We can have uh release this week, but probably good to see if basic changes are going. I think async changes. I just want to discuss once more before really planning for that.

A

A

Okay, I think probably we can start doing it. I mean if you agree with this yeah.

B

I shouldn't agree that I think this is a good tool.

A

A

You are supporting dln building.

A

I mean I don't know I've. I know people asking for this they're having some discussion both internally and externally, also yeah, but I'm.

B

Sure they just asked about dynamic safety runtime right, I'm just curious, whether that's the same as dl, and I think for static linking our library. We can still use the dynamics runtime for for windows, and this is only for windows right.

A

Yeah I mean I don't really understand that. Actually, what do what does it mean? What what was what what's the dynamic c runtime, and what exactly is that I I mean my understanding is that we can have a static library and we can have a shared libraries in windows. The shared libraries are dlls and uh the problem which we have right now is that uh uh I mean if we have api just a second.

A

If we create an implementation library as a dnl, uh it cannot, uh I mean the instrumentation library which is using our ap, open, telemetry api and it is created as a bl. It cannot be integrated with an application.

A

uh It cannot be used with an application because of the boundary of tracer provider and all the global static boundaries. But this is a different issue. You are saying the dynamics run time which you are seeing.

B

Yeah, I think there could be a different issue like even for for windows. We we can still build our open, telemetry cpp as a static library. We just link to the application, but um as we were using c runtime we linked to provided by visual studio, we can still link to a dynamic c runtime, which is a dll. I mean we for if we like, like um call any c runtime function like critique or anything like for io. We just record into that dail system provided not not a link that statically to our code.

A

Okay, yeah, probably you can understand it better.

B

Yeah, I think the there could be some issue, like um user dynamic serial time, um we can share one single runtime with other components, even very for non-open telemetry. If we currently, we use a static one. We can't we can't share the run time with others, and if we do hip allocation, like our memory allocation, we actually created our own like hip structure, so.

B

The pointer can't be freed or can't be passed to other component or right because they have to be freed in in our component. I think this is one thing.

A

B

Yeah and if this is uh overlocker, I think maybe the um this can be supported like to make some minor changes or say make script or stuff, because this stands for require a major reflector like provided the dll export declaration.

A

B

A

Yeah, I think I probably I think I still need to understand it better. So this is a different issue. You are seeing from that supporting dll building, that's a different.

B

Yes, I think slightly.

A

B

A

B

Component as a dll, because this is means- we call the c runtime in another drill, instead of link that into our in ripple.

B

Yeah, I think that's one ask.

A

B

From the the adaptive team from azure, okay.

A

Okay, but but the ln issue is still there, that's this yeah.

B

So you're still there yeah.

A

B

A

Okay, then probably I think I have for azure. I think I that's what I was I I didn't. I didn't wanted to reply. Probably I think I have taken that discussion in a different direction.

A

uh Probably you may want to deploy. On top of that I mean.

B

A

B

Yeah I'm trying.

A

To understand more, I'm a.

B

A

A

B

A

News on that yet yeah.

A

Yeah I'll keep it with you, like you, probably you can discuss whenever you want. I think it's bad.

A

Okay- and apart from that, I think this yeah, let's discuss for async the problem is right. Now there is a. uh I think. I've already commented on on that issue. There is a change, spec specification.

A

So there is a change in the specification.

A

This this so right now this casing changes are done for the purpose that if we can have simultaneous, I mean if it can, if we can have simultaneous export export method, calls uh triggered from processor uh to exporter.

B

A

That that was a that was the motive behind this acing changes and uh as for the specs change, the reason why those that was done was because uh the specs as for the stacks as of now I export method, cannot be called concurrently. I mean we cannot call make export method, call uh concurrently and export. The first export should finish before we can initiate the second export. That was how specs was okay.

A

That was something which was mentioned in the text, but now it is going to change until the wordings have been changed, so it says that uh what export does before returning language and exporter dependent. That means that, even though export method can return without doing the actual export, I mean it's something like the export. Even though the export is being done, exporter may be using a separate thread to do the actual export, but it can still return and which will allow the processor to initiate the second export.

B

A

So the which means that the exporter may have a separate thread to handle this, and uh even though we are calling export um export is not called confidently, but still exporter may be doing multiple uploads on a separate thread so which so the exporter result. Success will mean that the batch has been successfully passed to the export earlier. The success was meant that the batch has been successfully exported.

A

If you see the difference earlier, the exporter result. Success meant that that has been separately exported, but now it means that the batch has been successfully passed, so this has changed now. That means that, whenever processor uh initiate the export, it is not wait. It is not expected that actual export has been happened. It just means that the exporter has got the events and it is going to eventually export it. So that means that we can simplify our changes right now and which is which is.

B

A

I mentioned here and even also said that.

A

We can remove all those batch log, processor, batch fan, processor and even we can remove all those async export functions which we have written.

B

A

That means that this pr actually will get simplified quite a lot, and uh so what do you think I mean, but.

B

This packaging has not been merged already.

A

B

Is still open right.

A

Yeah this this, this vr has all those changes, batch, lock, processor, batch pen, processor, which we don't need anymore now.

B

A

B

This back and the proposed spec has not been merged, which means yes, it is okay. Okay, I got it, it is not merged.

A

But I think it's already approved if we see it's a tool by the uh rally and uh this guy- and I don't see that it will not be immersed.

A

I was thinking that, instead of so there are two options. One one option is that that go ahead and review these changes. Yeah.

B

A

Then have it which, which overtakes the string that first of all, let's go ahead, review it and merge these changes and then.

B

A

Separate pr to remove it.

B

A

I feel that that would be a lot of effort both in reviewing and also for movement, also to really create another pr after that, for the cleanup and so.

B

A

You you suggest.

B

We do the clean cleaning up for now.

A

Yeah, I am suggesting that let's wait uh for this to get merged.

B

A

And I think it should be much better to do a cleanup. I think this is already. I don't think this could not be merged.

A

I mean I don't know. I just don't want that this. If we do all the reviews for something which we don't know, which we know that we may have to remove it in future,.

B

I think I think yeah it makes sense, because for now we're almost to sure this is not going to be needed right.

A

Yeah, it's mostly sure because I think yesterday it got because.

A

By jack, I think one term.

B

B

If we want to wait this to be merged and then um another, I make the changing that I think should we are, which means we may need to delay the merge of async like for one week. That will be good right.

A

Yeah, I mean that's what I am thinking I mean. The problem here is that even owent has to probably do a merge conflict every time, but it's good to wait and let's get.

A

Let's see what what what's being decided for this.

B

I'll see yeah also.

A

B

Want to take this spec update, which means- maybe I think, for the async change. I think it will be. It looks, looks hard to be merged in this this week.

A

Yes, yeah yeah, it cannot be done if, even if you want to do, you won't want to take this change. It will not be merged. This.

B

A

A

So I mean I was thinking even though I said agree, but I think I gave it a thought. I thought I would discuss with you in the meeting before really putting any other comment.

A

I just thought that it would be good if we have a pr which uh I mean with. If we do have this pr only contains the change based on this uh specs changes. I think if we, the pr can be modified and we can remove lots of changes that are not required.

B

I see okay yeah that starts promising. If we can remove a lot of unnecessary changes. There.

A

Is no point in having these are very big uh code, I mean, and then lots of macros are being created which we don't need it. uh uh Well, I mean if we have to go through this, if we have to really incorporate the spec changes and lots of lots of things which we can clean up in this, and I think it will be very simple.

A

So that's what I'm thinking probably.

B

I agree so yeah if we can to make the change and wait for the spare catalyst. I think we can do that in parallel, but we need to merge the change after the spec is merged right.

B

We can merge this pr, we'll just say if we we're going to make the change in the async pr, but we can only merge the async pr after the spec changes version.

A

Yes, that's that's something I think which is uh which we have today. We have to take a decision. Whether we should do that.

B

Yeah, I think um if we want to do that, then we will. We need to postpone this merge of this one to the release candidate of metrics right is it? Do you think that will work.

A

B

That's also, maybe one or two weeks later.

A

Yes, yes, probably probably I don't want both of them to go together, I mean so, let's take a call.

A

I think, let's take a call, probably next week or something I think, but should we should we communicate to owen that he can start doing those changes, yeah.

B

A

Should come, communicate with him.

B

Before I think before, we decided right because he need to do this anyway, once he decided yeah, it's okay, to make the change and we merged. I think, like two weeks or one or two weeks later, that would be good fine. I think this is a good way, but yeah needs his confirmation.

B

Yeah yeah also yeah. Maybe give us more more time to review the big pr. I think that the mark should show the interest to review the pr. Maybe we can discuss how maybe we can.

A

B

Split the focus, so we can review the pr more.

A

Yeah exactly right, because, if not there's a review, even markus have said that we definitely we know. Even I was thinking that we have to test it. Also, I mean instead of this. That's.

B

A

At least the existing functionality should not get affected. Yes, we cannot just rely on the ci test to do it. That's.

B

Right yeah, I totally agree, maybe more people to review. I think.

A

B

We will come back for next week, so yeah.

A

I'll I'll probably add it here that I think that's it's. We discuss it in the community meeting and I think it's probably better to incorporate the specs changes in this yeah. Hopefully we should be merging it on once. The specs is merged. Okay,.

B

B

Yeah and I just started to look at the pr.

A

Okay, I would get that that's something I think you'll probably be um the.

A

We really want to review it. In that case, uh there won't be major changes in the export expo, which.

B

Part, maybe we can no no.

A

At least you can we can ignore these all the changes which are there in the processor process.

B

A

We can all we can ignore them and all the okay. Let me just try to open this yeah like async, back, lock, processor. This this class will go away. Okay,.

B

A

Lock processor, this will go away and we'll just have batch band processor and batch lock, processor, which we already have.

B

A

Yeah so even async batch fan processor, this will go away and in batch span processor we will not have.

A

Okay in the export also will in the exporter. Also, if we see this one, we will not have where.

A

Basically sent the result as a contact.

B

A

B

Means there's this api change right. It means if we take the change like the current change for now,.

A

B

Take the update for the spec later and we will have api breaking and like we'll, remove there. No.

A

No, no, no, we so it won't be api breaking because we we already have the existing method there in place. Yeah.

B

We're adding a new method.

A

And that's in this is an exporter on the sdk, which is fine.

B

The people can use this new export right and if they are starting to use this one after.

A

Taking this release.

B

A

B

Yes, for the next releases.

A

Yes, if we remove it yeah once we, if we remove this, that would be a breaking thing.

B

Yeah- and this is not marked as preview red- this is just I mean it is stable, so people should, if we release it this month, people can take dependence on it.

A

Yeah they can, they can do it. Only thing is that they have to enable this flag explicitly yeah. This.

B

Is a feature flag and then not not paying them. This is preview and there may be working later, but there's no such sign, so they are.

A

Free exactly to depend.

B

Don't introduce this from because we know in the next month we need to for sure be deleted.

A

It in this changes.

A

And the building.

A

A

A

A

A

So I'll update that pr, I think that I know owen will not be happy with that, but I think probably it's good to have a cleaner than doing something and then removing it.

B

A

Okay and matic progress.

A

Yeah, so I mean I I'll be raising uh mostly, I have done the changes for otlp uh grpc exporter, and it's mostly done. I think, probably this. This.

A

I'm hoping that in next, probably by tomorrow or something.

B

A

How changes are yeah? I.

B

Remember some people who is asking this otrp support for metrics.

A

A

I'm just saying this week, but I think I'm more optimistic. It should be done before that.

B

I see- and this will be the release candidate right- should be part of the release.

A

B

A

Part of it otlp, hdtv exporter, ovitch, wants to work on that. I think it is.

A

This will have dependency on async changes which won't be doing so. We we don't. We don't need to have it as part of release candidate.

B

A

A

Yeah this these these- this is the so the the major change right now is so this one is uh otp grpc exporter, I'll assign it to myself.

A

I've already done this, and this is part of the microphone and http exporter.

B

So this, I think, will not be part of the release candidate right because okay.

A

A

I mean one of the major issue is major or critical issue in this is supporting the registered multiple acing callbacks.

A

This is a major change and I think this is something which is I've tried it a couple of times and somehow the change did not work and then I have to probably uh pause it after a while and start working on otlp, but I think I'll take it up uh this week once the otp is uh otlp. Stuff is done. Okay, I think this is probably there is any, uh and this can have an effect on the delay.

A

If uh the changes are, you know not working as expected, but yeah, let's just just giving things across that. I think there won't be any issues coming with this. Okay.

B

A

But but yeah it just needs some.

A

Some, uh I think complicated changes in the existing food.

B

Yeah, so maybe you can assign this to to yourself.

A

Yeah I like should be with me. I don't know why.

A

I'm just not keeping everything just in case somebody comes and they wanted to take it. So I just kept it. Okay,.

B

Sure yeah for this one, maybe mark.

A

B

A

Yeah I'll probably give a liver for them. That's a good point, just in case somebody interested to contribute.

A

I think it's a small issue. First.

A

Okay and yeah, even this is a small issue and I already done it.

A

This also is a small one from not a major issue. Probably somebody.

A

And yeah this I'll take it. I think it would be a need some changes in this room: okay, yeah! That's what we've assigned okay.

A

Guys also, I think.

A

Yeah and okay, I think we should be done with all of them either assigned or either.

A

Okay, and apart from that.

A

No discussion on interrupt repeating the other languages. Log sdk still log meter changes are there and.

A

B

Yeah and for the locker name, I think it does. It relies on some asynchronous rather there's just some confidence.

A

A

It's okay, I think probably he wanted to take it up uh over and wanted to work on that once yeah, uh I think kanji got immersed.

B

Okay, yeah: that's when yeah.

B

That looks good to me.

A

Okay, anything else we had, we wanted to discuss, take it all on the new next release, probably, but I think you may be creative, you will create a website for that, and uh probably you keep a date based on what you think would be a good based on the correct scenario.

A

uh Anyway, we can have it for 30th if we plan to do it for matrix daily standard.

B

Okay, matrix that's 30, and for this week, are we still going to do a monthly release? I think right, even without the async change, shall we I think, do you does it deserve to do our muslim release.

A

I don't know whether we have anything substantial.

B

If not, isn't substantial or nothing broken.

A

Yeah we can have, we have an upgrade to jso json. There are some changes for matrix.

A

You did change it for current time, storm with time stream and.

B

Then there are some ecw exporter fakes. I think.

B

Yeah, quite a few, I think I.

A

Think maybe for.

B

A

Yeah, you can have one small release for this, um even but there are few pr's which I wanted to be used. Let me just see uh put requests it's already merged. What.

A

Okay, this is not.

A

This one we can, we can close it. Sorry, I think I'll close it now. It's a this also should be part of that.

B

A

B

C-R-Ah yeah, we can merge it. It's.

A

Done already, okay,.

A

Yeah, so I think probably we can do that. We can have a patch release.

B

Okay, that sounds cool yeah.

B

Okay, um yeah for patch release. I think that's when yeah less work than when released with the async change. The follow-up will be. Communicator was always on this thing.

A

B

Before we decided to make relief, we said the icing change.

A

Okay, I think if we don't have anything new in the agenda, I think we can quickly go to the.

A

Issues- one of the thing I before I forget there was a issue raised by.

A

Yes, this one, I think, uh mark wanted to take this issue. I think in that mail he sent a mail, and he just mentioned that this was created by him and I think he has better idea. Okay,.

B

A

To fix it, so we can continue.

B

A

Okay, just again.

A

A

A

A

This is something I think we probably fix is required for this, and no man gets on fails to install and I'd say it's something important issue, I'm just keeping it as health wanted.

B

A

But I think probably once assahan is back, I think he can take this over. He has because the change was done by him and he had better idea how to fix it.

B

A

So mention it what I think once he's there.

A

Tagging me up now.

A

Okay, this one is something which would be fixed in uh async changes, so.

A

This is experimental three options. um This is something which I think tom, I think I've just assigned to you and so yeah.

B

I just took a quick look. I think this one is low priority for mine, yeah yeah.

A

Yeah, I don't see it's a higher priorities for tracking purposes with you.

B

Yeah may need some work to support this in sdk.

A

Static variables from sdk header files to this one yeah it's low priority. I think it's it's fine. You can keep it if somebody want to pick it up. Okay,.

A

This also, it's probably libraries delivered document the libraries which we are delivering and ensure that both bazel and cma should have consistent library names. Okay, it's a good issue yet, but probably keep it. Okay, I just created these two remove log name: sdk version is an empty string. This is this should be fixed now, right.

B

Then we can mark it as fixed.

A

Yeah yeah, okay, yeah.

B

A

B

I'm going to create a new uh new milestone and then include this.

A

B

Even if you're not running a.

A

Script actually just ensure that even if you're manually modifying these, uh which is also still okay but just ensure that this is this, this value is not empty.

B

Yeah sure yeah, I will run the.

A

Script and yeah make.

B

Sure the barriers are there.

A

Okay, well, I'm 64. I think you created this topic.

B

uh Yeah, no, no news on this yet.

A

Yeah allocation of attribute value from knowledgeably span. He told this.

B

A

Creator of this issue he wants to take it. He didn't reply right owen had some comment. Do you have plans to add the support? I think he didn't have reply on this dinner.

B

It's a breaking change.

A

It's a ebi breaking chain, so I think probably.

A

We have to think when to do it.

A

There is another.

B

Abi breaking.

A

Chain which is which which will come and then probably I'm just good- we can take it if required. We can take this change along with.

B

I think the locker name design is one api is still.

A

Okay, I think I won't say wait say we.

A

Instrumentation scope.

B

Is this a bridger change? Can this add.

A

Yeah, it will break it. It's a api breaking king right yeah. We are adding a new attribute in our existing method or if we may create a new method, new virtual method, with a with this extra either we over a new virtual method or we modify the existing method either way it will break the api.

B

I see but abi we don't. We don't provide api compatibility right.

A

We provide abi compatibility for api right api level. We do provide it.

B

A

So now the problem would be that, if, uh at api level, if any avi waking change is going to happen, I mean that that is not going to work.

A

So what it means is that if, if somebody builds it this with a new api, it may not be useful with the existing lt case, existing sdk, which is not compiled, built using this uh instrument. This new attribute.

A

So I I built up my instrumentation library, which, with the new version, but the sdk is still the old version, so it is going to break that the new instrumentation library is going to break or it won't work. With the uh I mean it's going to crash so.

B

A

That that's, that would be a vibration, change and baby 11.

A

A

Yeah, probably that's what we have to decide. I mean I think that the specs is all already going to modify. I mean they are already going to incorporate this change.

B

A

And I think I I already mentioned it this one. You see this thing in there I already mentioned it.

A

We already closed it or what yeah this one.

A

And drive some new interface of the old one.

A

I'm not sure what he's talking about.

A

But here is this yeah if you prefer compatibility over the conventionally useful interface, you can go on convoy like that like here, and derive some new interface from the old ones. Okay, okay, you can go to my.

A

A

And we don't have any new pr's.

A

So in this pr, I don't know why this is the doc effects is feeling the changes are valid, but the doc doc effects fails.

A

A

It should be, I think it is.

A

But I didn't find any solution. I tried, but then I didn't spend much time on this.

B

A

Little thing yeah interesting.

A

Yeah, I think that's what we have from 10 30. I think anything anything else you want to discuss.

B

Nothing more from my side, thanks yeah, since we're also running a lot of time today,.

A

Yeah thanks uh you'll be in office.

B

Probably today, for today, I probably will not because my child is at home today, they didn't go through.

A

Something yeah: okay, I'm planning to go to the office now again.

B

A

Have a good day thank.

B