GitLab Backend Pairing, 23 Aug 2021

Previous Meeting

⏯

youtube image

►

From YouTube: BE Pairing - 20210823 - Skip job in JH fork

Description

In this session, we pair up on skipping a CI job in JH's fork of GitLab.

https://gitlab.com/gitlab-jh/gitlab/-/merge_requests/81

A

Too late now anyways, uh so I um the issue I was going to bring up is we have a startup css check job, which checks if we need to update some auto-generated files, and this works nicely uh with an asterisk next to nicely.

A

um It is not quite working at all for um the g who um who's forked git lab the the the get lab fork for china or something like that, um their pipelines, uh it's very likely we'll just want to disable this job for the jh fork, and um I'm not entirely sure how to do that. um But I imagine that there's some jobs that we've done that for so I imagine it's also really short to do like that's not going to take a lot of time.

A

um So that was one task I was bringing to the table.

B

I don't think we can hear you chad,.

C

No I'm muted. Okay. I was hard hardware.

B

C

Worked on one of the jehu related ones before like, if it's just a ci thing, probably the answer is just they need to disable that job in their fork or or delete it right.

C

There, I think, there's some environment variable flags that we could like wrap around it, but I don't think that's the way. We tend to do this because that colliders, you know the rest of the open source code base.

A

Are you saying that we don't really have any like jh references in our in our gitlab ci stuff.

C

No, like the only like we'll make, only make a change to facilitate them to be able to do something like. I think we may be slightly modified the way we override modules and stuff like that, to make it easier for them to do something, but as far as like, actually putting flags in it in our code and the main open source code base, I don't think we do that. I think this is going to be on them that they need to uh just not run this job in in their fork in their ci config.

A

Yes, and so I was gonna, I was going to push a contribution for like this is how they could do that because to their fork, because that's what I think they may need.

A

I- and I just don't know- and I don't know if we're.

C

Able to that's the way you want to do it yeah.

A

But I also don't know if we can just remove, because my understanding they're continuously getting updates from the their upstream, which is our main project, so the only things they really can deviate on are files inside of their jh subdirectory.

A

So even things under gitlab ci or something we, those files can't deviate from the main upstream, and this is very similar to how we used to have an eece split before now. We have just one project for both of it and marcus. I think you were here long enough to remember that as well. I think it's very similar uh to where uh their jh projects, unless it's under the jh folder only things on the jh folder can deviate from the main upstream one. So somehow I I don't know.

B

So I wonder what you're saying boys this is much much bigger and what we need to actually do is create a way if it doesn't already exist for someone to create a yammer which will override enough.

A

Yeah, maybe I really don't know, um I really just have no idea what.

C

To do just the ci config or like by not running it at all, are they gonna are other parts of the app that consume the startup css also gonna break? Can those have to be have logic in them?.

A

um So let me let me share. uh Let me let me share some more information on it. So since we started talking about it, I'll share my screen and we can start diving into this. Let me double check that I am allowed to share my screen and, yes, that should be okay, great, okay, I'm sure all right! This is the issue you were talking about lee, um but let me hop over to this issue and they are running into and I'd love to know your thoughts on this.

A

um They are running into an issue. Yes, so they have an mr! Yes, let me go, let's see if we can hop to that, mr and I think unfortunately, mr has been blocked on this for some time. Unfortunately,.

A

So they're kind of doing some they're doing something which isn't something we should really try to do um and they are throwing in these render if exists like they want to extend the sign up box.

A

So they're throwing in these render, if exists. They're gonna make an ee main our main project contribution to throw these in there, but these files won't exist in our code base at all. um To me, that's a little bit of like a yellow flag because knowing just how quickly get gitlab changes, um if it looks like something's not used, I'm no it'll, probably just get deleted.

A

B

A

Saying they want to push.

B

This upstream, to our main repo, but only create the file in their repo.

C

So I made exactly this point on the jhmr that I worked on, that this would get deleted uh because people didn't think it was used and I think the result was um we can't avoid that. I don't know I could try to find the link, but it was.

A

B

C

Just a module extend at the bottom; it wasn't something like this that actually pointed to a file that didn't exist, which is much more confusing and concerning, I think I I would put it more much more orange than yellow personally.

A

Sure yeah and I've seen one other, mr and I've made that concern too. It was, um I saw the main gitlab project contribution, which was adding a reference to a file that didn't exist, and I said we need to at least wrap this with a bunch of comments of like this is actually referenced and something else that's.

C

Exactly what I said, what I would do, instead.

A

C

Like, rather than uh try to have the branching at the lower level here and to like, say, okay, we're going to conditionally include something that exists in one place, but not the other like at the top level say you know, there's a um one view that does nothing but include another partial right and that and that one like passes a flag down that is used in their fork or something like.

A

C

So it's less weird.

A

Yeah, I think I think that is because you, if you squint your eyes just right. This looks like dependency injection and you need to make these these base modules, extensible and they're. Just not we're in lining we're in lining files that don't exist so that they could be mixed in.

C

Which which one of the letters in solid we're violating I.

A

Think it's the oh, it's gonna be one or two of those. I think it's the open close where.

B

So how would you, how would you sort of make it, though, so they could squirt something in anywhere, so to speak?.

A

Yeah, so one this, that's such a good question, and I think this is outside the scope of our time together, but the more we can make um the presentational layer data driven the better, as opposed to where we've just inserted a whole bunch of fields and duplicated ourselves. If I could just say here's an array of fields, you can easily extend that data you're, providing to the presentation um we're just not there and have.

C

We done that anywhere else in the app paul. Oh.

A

That's a good question. uh Maybe I should do a search for a view model pattern see if that shows up anywhere. This is chad and I paired on doing the the top nav, uh and this is that was the approach we took for this is making it data driven.

B

I mean, I guess, that's, maybe not quite the same thing, but where we've done with, would you call it like docs and things like mvps and stuff? The the list of you know that pulls from data.yaml or whatever it is.

A

Yeah sort of what's new is very probably very similar to that too, like all of this is very similar to it, but it's it can get um the what's new bit because this all comes from a yaml file. um This is like basic data-driven presentation pres, because we're just rendering a list- and every list contains the same stuff. Things get interesting when, when the lists contain different stuff and but that's something, that's completely obvious.

B

Yeah yeah absolutely.

A

You could you could also throw different types of yields in here. If that's was something we were interested as well, but that also brings up the in the main gitlab project. The yield wouldn't be referenced at all, um which is uh which is interesting. um So the main problem at hand, though, what what what this is gonna do is we're updating a um main gitlab project file, which is only going to be different when jh is run and.

A

So if, if, if we're in a jh environment, I basically don't want startup css check to run because it's gonna include files, it's gonna include things that aren't in our um and our generated files. The other option is, we could create so give me one minute to go on this other path. The other option is, we create um startup css files specific for jh, so we have um ee startup css files and we have startup css files for ce, because these two things can be different.

A

We could do the same thing for jh, but there's a huge downside, meaning that whenever there's a change in here, um the uh whenever we're gonna push whenever they're gonna get updates from the upstream, they might need to like regenerate their files, and that's gonna be problematic, uh because right now they can just take whatever's the latest, because jh is assumed to be like a superset of ee and stuff, and as long as it's like that, we should be fine. um So the one option is: let's just make.

A

Startup css check work for jh by giving jh their own files for this, but I think that's going to mess up when they get upstream updates and it's going to be more of a headache than if we just disable it, because, because of the nature of that project, it's only needed. I think, because jh is just an extension on top of these things, it should only be needed if we're just an ee or ce.

C

Because the results of it are committed is the crisis of it.

A

C

Take that over every time they sync.

A

Yeah and- and I think that it's gonna, that would be more of a headache. um So that's why I landed on. I think we can just disable the job for that environment, but I really don't know how to do that um because, as I look into here, I see I'm looking, I did a search for jh inside of dot git lab and I see us every time we are doing an ee reference.

A

We also do a jh reference for whether we should do something or not, but we don't totally have like a is jh or like skip for jh or something.

C

I'm gonna try to find what we did for the the module extension library, because I know this came up with that.

B

So do we think at the moment these ian j h is saying only do this. If you are this or this so there's not one say: don't do this if you uh or not that we know immediately.

A

Yep yeah, I kind of- would want to.

A

I would kind of want to do something like I don't I don't this I this. I have no idea something like don't.

A

If we're jh, I want to do something like that, but I have no idea how to do that um are any of you all familiar more with the ci api or how this is maybe possible.

D

I was wondering if we could just redefine either the startup css check task or the check base above, like just redefine it in the g repository and then I guess it should replace this definition. So we could just have like an empty script block. Basically.

A

Oh, like re, redefine it and then do they have. I don't know.

D

I'm not sure how to actually do this too, but I think you can include yaml files and then maybe they will need to have their own top level. Gitlab cr yama file, which would just include our top level, see I will find it then they could have overrides there. Maybe.

A

Yeah, let me see if uh oh yeah I already have from there. Oh okay.

D

All right on the cheap.

A

Oh, so this is a special job, making sure that every merge request only contains jh changes. So that's that's one thing that they do is they're making sure that there's no deviation between the main project, so they can continue to extend that's a special job that they have.

C

Just for reference like what we ended up doing with the modules like the pre-pinned mod thing, that's at the end of everyone uh I could drop a link to the one I'm talking about uh is.

C

We we we added this pre-pin mod that isn't strictly necessary, and but this uh note I'm going to link to as a suggestion said, why don't we just add prepaid mod to every single, the bottom of every single module by by default programmatically, so that was sort of like a specific fix to that particular problem. But the class of solution is like okay, if there's something that they they nee only need to do sometimes, and we have to sort of opt into it. Is there any reason we can't just programmatically?

C

Do that all the time and.

A

C

They always have this hook would be the the general class of solution. That's.

A

C

No idea how that applies to this problem, but just throwing it out there.

B

A

Yeah, that's interesting.

A

So yeah um we've found there is a jh specific one, and I wonder if this is somehow referenced somewhere. I want to see if maybe we're referencing jh.getlab.

C

Yeah, I think what what we would it'd be something similar to what I said before there were for the startup css. There would be like one one included yaml that doesn't actually do anything but includes the other yaml so like inserting this level and that's the level at which you would override the the yaml and in the jh code base. That would say just don't. Do this basically to be a no op.

A

Yeah yeah yeah, you see that they are um including.

A

Wait, what is that? How are they, including, is that the top level gitlab ci ammo, I guess, do we have a dot.

C

Does local override declarations, I'm assuming that's how it works and that's the whole idea behind this thing.

A

D

Local just means it's in the same repository. You can also include something from a different repository.

C

That's like the second one wins if it's the same.

A

Are we saying that yeah is there an override? um Is there also.

D

Projects, I think to say, like the jh gitlab cr yaml is like the default ca configuration instead of the top level. One.

A

Yeah it's so if, if we can override things, is there a way for um is? Is there a way for me to.

A

A

Just remove a rule or maybe.

C

Do I need to like be a no op, it would do nothing. Can we look at what on their fork? What is what sorts of things they already have in the jh that get lab yeah.

A

Let me let me open up their, I have their fork, I have it. I have their fork. Forked.

C

Essentially like in the main repo, the job to execute would be like, whatever it really does, and theirs it could just be. The job would be like echo, we're not running startup css, because.

A

Yeah, I wonder if we could just throw skip there, though I wonder if like is it possible for us to just okay? um I guess I I lately. I always don't trust the author, uh even for our stuff, I'm just like. I don't trust anybody anymore.

C

You don't trust what so, what.

A

Vs code keeps asking if I trust the authors, um because it wants to know if it can run like fancy extensions or something. And I.

B

Trust yourself, I guess no.

A

Like that's, that's one of the answers, um okay, so it's! What do you think is a good path forward. The way I would test this is I'm just going to push up an mr and see what happens and what the pipeline decides.

A

uh Do you think adding some sort of do you think adding some sort of skip to here makes sense, or do you think, somehow overwriting something makes sense.

C

So this whole uh the startup css, is it in its own independent file under gitlab for for ci? If we.

B

C

That's what we need to do and then that's in the upstream and then that's what they will override to be a no up, because if it's mixed currently in the same file mixed in with other stuff that they still need to do.

B

Yeah yeah yeah. So so that's what it looks like right. Oh it's! In with all this other front end all these other front end jobs. So we need to remove it from here. First put it in its own file. This file then references that file um and that's the file that we overwrite.

C

A

I don't know if it works like that. Okay, I don't know if it works like like, um like hamill views, do where you can have files where overwriting files and marcus correct me. If I'm wrong, uh I think it I think, looking at this bit.

A

This is that our ee entry point and that's gonna, include everything, and then this is just going to be. An additive include everything. On top of it, uh I think.

B

So yeah did you so were you just pulling down the the repo to have a quick look? What they've currently got in there in case, I think that's what chad was suggesting. They probably already got something that does something I mean but yeah. Let's do.

A

That, oh gosh, what am I out? I think I need to fetch something.

D

It says in the docs, the include files are deep merged, so I guess every level like every key on each level will be merged. Somehow wow.

C

Which is fine as long as we're overriding, if there's like one branching point that in the g, who is overwritten to do nothing.

A

Why am I having such a hard time with this? I want to? um Oh, it's not maine. It's like maine jh, that's right!.

C

Here I'm pasting the browse link to it: okay, on the web on the web interface, okay,.

A

Oh, so, okay, so these are all these should be exactly the same as the ones, and this is actually our.

D

Gitlab ci directory, not the yes.

A

A

Yes, they do have. um Yes, they do have something like this. They have jh get lab, yes, um okay, so these are the ones that they have and.

A

I guess maybe I could do this and, if they're deep merged, maybe we could just do this.

A

Oh gosh, oh gosh, something like this and use something like what's the other startup well, if.

B

It's got the same name, do you think it'll just override it.

A

Well, it sounds like you're, deep merged, that's what the dogs say yeah. I guess we have to try it out yeah and we have a pretty good docs team.

D

But they they get all the input from us from the engineers.

A

Phil, maybe I could just do something like this: it is actually.

D

Like a van, never when colin, never, which should just not run it like that.

A

You want to just try that out. I'm not.

D

Sure if it works on the dot, startup css check base or if you need to do it on the actual job, like the startup css check, let's.

A

Let's try this one out and see, but I do now that are talking about it.

A

I'm hoping I want to make sure it should run at some point anyways like do you know what I'm saying I don't know.

A

I want to make sure that if it was gonna run, it should have run except for art change,.

B

So if you want to squirt something else into your, mr, that will force it to.

A

Yeah and I could try- um I could try throwing in just these changes, because this is where part of it's failing. So I could to like almost base my mr off of this one yeah you wanna. Maybe I should do that. Instead, yeah yeah check out that branch yeah. Okay, let me commit what I have and and uh never run startup css and jh.

C

You're going to base nanimar is in their fork right, so you're going to base it off of their fork.

A

Yes, are you going to? Are you going to, I know, there's a rebase onto there somewhere.

A

So I'm gonna do this, I'm going to fetch this branch- and I am going to oh man I'm about to do something really crazy. I know there's a rebase onto here, but uh I'm not sure what it is.

C

A

You want to help me with the rebase on too sure, so I made my commit off of their latest main, but I actually now want to just move it on top of this. Okay.

C

Do you get log dash 2? What's the last commit that you were on so the 7c grab that one.

A

C

A

C

So get rebased on two, whatever that other branch is that you're wanting to rebase onto.

C

And then, what you're rebasing relative to is the thing that you caught.

A

C

Yeah so you're, taking that one commit throwing that on top of this other branch, but not worrying about any other differences in history.

A

Okay, can I go, can I go dash, I can I do interactive. No.

C

There's no need to. I do not allow you to pause. Okay.

A

All right, I'm doing it.

A

Wow freebasing one it did it man, okay, so that was cool. So I made I've never had on onto really work for me, but this is the. This is the one.

C

Case it saved me like there's the two approaches: there's the the combined one and then there's the big chain of ones. If you're doing the big chain of ones onto is your friend, it will.

A

C

A

That was awesome, so I'm gonna just push this to um he slaughtered. I don't know what it was never run.

A

Oh gosh, why can't I do that? What is wrong? Why is that work working? It probably doesn't like um uh I'm. Probably under my um I said I didn't have permission to.

C

Yeah, you can't push to their fork. That's what I was. What.

A

I have, but I have my fort I'm trying to push to my fork.

C

Okay, you want to push here.

D

Is the url correct like if you do get remote? Does it show the correct url.

C

Yeah just explicitly specify the remote yeah, so your remote is so.

A

I was doing pslotter.

C

You're a fork: okay, yep.

A

And then I was uh doing my branch name, I think that's how it works, because normally I'll do like origin of the branch name, but this isn't working. I think because of some ssh thing on.

A

All right there's, I think that there's I was doing something and I did messed up something with my ssh thing locally. uh I can.

C

Remote for http and type your password.

A

I could yeah I could do that. I think I'm going to use the web id. uh I know I know yeah look! Oh look at that button, edit fork and web ide.

A

That was money that was pretty cool now, if only I could like copy and paste the patch into here. That would be amazing. uh That would be really cool. uh So I'm going to jhlan gitlab ci, I created a new file and I called it.

B

But then has this got. That is this off of the branch that you're working or are you working off of the main now.

A

Oh gosh you're right yeah. This is not on the right branch that I was really hoping to be working on. Oh forget it.

A

I gotta fix my. I gotta fix my ssh thing, uh um but I'm probably gonna stop sharing my screen when I do that. So does someone wanna? Maybe um I'm gonna push this up and we can see if that change would have fixed it or not, and but it may take a minute or two before we get some feedback on it. um Does uh someone wanna drive on um on lee's issue of the confidential um notes showing up in activities.

B

I'm happy to do it, I'm not sure that I'm prepared like I'd like to be, but I'm happy to do it. Okay, um let me I've got the issue.

A

Then I'm gonna, gracefully pivot, look at how look at how graceful this transition is. It's like we're on live television, not a pirouette, a pirouette.

A

uh That sounds like a ballet turn.

A

The act of spinning on one foot, typically with the raised foot touching the knee of the supporting leg, a pirouette that is graceful.

B

It's just gonna, probably unnecessary, but create a new branch.

C

You could call marcus on noticing that that was uh actually a deep merge that made it a lot simpler.

D

Okay, I think the race won't get merged, which can be surprising.

C

About making all this stuff easier to do in yammel, that's sort of been languishing or an issue at least, and there were some proposals and then people started uh bike shedding and they never went anywhere.

D

It's a common theme in github issues.

B

So I'm guessing the first thing we should try and do for this is recreate the issue right, which isn't always going to be the easiest thing in the world.

B

But we'll try.

B

If jdk wants to.

D

I'm interested in this feature for service desk as well like to have internal comments.

B

Yeah so again it's one that's dragged on for far too long. Unfortunately, like I don't know when um I want to say that the git lab team, if I've got my terminology right, started work on this maybe nine months ago, and then I started picking some of it up about six months ago, because I was like oh awesome, been waiting for this, um but I think both the gitlab team and I have lost enthusiasm and um we really need to find some. I don't know some carrots to dangle to get it over the line.

B

So we want a public project. I guess.

B

Oh an issue: I've probably already got some um confidential notes on here. Let's pick a nice new one.

B

I've already got the feature flag switched on, I'm almost certain of it.

C

I dropped the mr about merging hammer gamma rays in the chat.

A

Oh, it's not so it's not that deep it does it it'll just overwrite the array.

D

I think the deep watching only happens for hashes and not for race, probably but.

C

Hopefully, there's no arrays involved here that there's a job, it's just a hash.

B

Of being array, but if you're just telling it to never run, then I think you're, all good yeah.

A

That's true, um yeah, that's interesting, deep merging is always a it's always a problem. Sometimes you don't want to deep merge.

A

It's always an it's an interesting problem.

A

It's one of those extensibility modification ones.

C

Too far into shallows now.

A

I want to merge these. I want to merge these two hashes. We need to go deeper.

B

A

B

Handstands, do I have to do after enabling a feature flag for it to uh sometimes hey there? It is.

D

I think it gets crashed up to a minute also sometimes.

A

Yeah, I usually have to do like a like a check. I have to fetch it, and my rails console will almost kind of force the cache to update.

B

So this is a bit frustrating because I can't tell what I'm sure I feel like I've already fixed um the user's activity fade. um So let me just like that and go like that.

A

Yeah, I don't see the comment there is that expected.

B

Well, yes, and no it's kind of good, but at the same time I'm logged in as the administrator who has access to that comment. So I'm kind of feel like it didn't need to hide it from there yeah um and it's not appearing in here. So we might be in super luck and find that it's fixed um and just the issue hasn't been updated or I.

A

B

I might be way off can.

A

We can we try, it.

B

Let's add a normal note: yeah.

B

Get rid of this ah there you go. That's better! Okay, let's add a normal one.

B

C

I didn't even know you can make individual comments confidential.

B

That's uh yeah because it's behind the feature flag, so it's pretty! Okay! So again, is this working off of any kind of psychic background job or what well.

A

It might be working off of a psychic background job.

A

Do you know if all the background jobs are running successfully? Do you know how to check that in the yeah.

B

uh Not exactly but uh monitoring background job, I think that's it yeah.

B

So what we're looking for, broken somewhere good busy.

A

To test if the activity is just working in general, do you want to try leaving a comment on a non-confidential issue and we can see.

B

It's yeah, it is it's public issue.

A

Oh, that is but that's a that's a confidential issue, though right. Oh, it's not confidential. No, no.

B

No, no, no, no, no! It's public! So I've. Definitely I've seen this behavior before in my gdk and it's um just where I've given up and come back later and I'm sure things will start. Okay. That might help well, so sorry, restart, yeah yep sounds like a worthwhile.

B

Always works right.

A

A

We have invalid yaml over here on the original problem.

B

So what did you reckon this compulsory? Some of the notes are compulsory or something.

A

I I don't understand what what this error would be.

B

um Let me stop sharing, while I'm beating my screen. If you want to.

A

We are parallelizing our work. um Can you see this screen.

B

A

So I am seeing uh startup css check. Config key may not be used with rules when.

C

Oh, I think what that's saying is: when is the newer syntax for picking when something is supposed to be run.

A

I can't hear you anymore, hello, hey, you know.

C

There's maybe something that you're, I think it's the win. Never that's! The problem like try to just put uh look at the original one if it's just doing a script, just fake it with a script that does nothing or echo, it says a workaround.

A

So this thing extends uh it is doing a script. It is extending this thing um and this thing is extending.

A

Oh wait, wait, wait, wait! Oh, wait. Wait, wait, wait! Wait! Let me see this one.

A

Oh, I need to use the rules thing. I think yeah yeah yeah yeah.

C

Yeah, that needs to be nested in rules. It was just a syntax here.

A

Okay, I have no idea what I'm doing here with rules, though.

C

uh I think you just put the rules nesting level in there and then rules.

B

C

A

What is what is this? Do I need to worry about that.

C

I think you can just put the win uh straight after the dash, so okay.

A

So I could do something like rules.

C

And then make the win in array. I think.

B

C

uh I can try it well look at the docs real, quick.

A

Just try it uh should I try to edit this using the web ide.

C

A

C

You may need an if.

A

So we're talking about rules.

A

um Do you all mind if I cheat real fast and look at source graph and see if there's anything I can copy from there.

C

Well, just just kick this off first and see if okay.

A

So we're saying it's worth saying something like this.

C

Actually, it should work. I've seen it.

B

It's it's not when, when it's not when it's when an array, it's not is.

C

B

C

Yeah, I I'll see you in the darks. I think this will work I'll, give you a link to it.

B

So you do need the dash next to when you can. I should.

A

B

C

B

C

Still work because then this the deep merge of the rules will get overwritten with this array, but we're not trying to override the contents of the array itself.

C

But right there in the docs, you can see there's a bearing with nothing else.

A

Oh, is it really okay cool all right, so where's the merch? Oh here's, the cool, merge requests it's down here.

A

Oh, it looks yes, yes, yes, there's no pipeline errors, that's a that's a good one and.

A

A

It's not trying to run the other jobs. I don't know if it's like pick them up or not, but that's kind of cool.

C

I think this may work, but.

D

C

D

Other default shops like all these backdrops and all that.

A

Yeah, that's that's one, I'm not entirely sure about.

D

A

D

Do we not really have any changes in this much requested triggers.

A

Maybe I should have been targeting master. I thought that we would run it yeah, because this, mr I was targeting that target branch. That's.

D

Hard okay, so we only have to see I changes. I.

A

I guess that's what's happening so so. Let me then target master.

A

All right, let me target maine jh and um let me cancel this one. Oh, hopefully it'll. Allow me to start a new one.

C

This needs to be rebased well, this needs to be rebased on two main jh again, because now it's got all of those other other commits right. Well,.

A

I wanted that I wanted those others.

A

Just to test it out, um hopefully, the pipeline starts and stuff.

A

um Thanks so much for helping me out with this, because this has been on my plate for a while.

C

I'm still concerned about the, if exists of things that never exist.

C

A

C

Address that in a particular way.

A

So I don't see startup css check here. Oh no, I do see it. Oh.

C

Yeah it'll be there but uh click on it and see what it's gonna do.

A

I I think it's not supposed to be there.

A

And here's now what I think is going on. We did this for this, like I, don't know what they call them, but it's like an abstract job or it's like a yeah. We did. We did it for this one, but the when is actually being set for this one.

A

C

It's a startup css check. Instead of that thing,.

A

Yeah, I think, that's, I think, that's what I need to do so. I'm yep yep yep. So I'm just going to do something like this.

C

Because those dots are like included at a different time that isn't subject to the deep merge or something like that.

A

I think it's yeah. I think this. I think this contains our when never, but this one is technically overwriting it with the right things. I think that's what's happening, so I'm just going to do this. That looks good. We feel good about it, um hey. Why don't? I just use the web ide for making quick, simple changes.

A

uh It's gonna work.

B

I guarantee I use it for dna to um about.getlab the docs, because the project's impossible to pull down, because it's so large.

A

Yeah, uh okay, it's gonna, take a it's it's freaking out, because when you open the merge request with the web id, it starts trying to open all of the files and unfortunately we still have this issue where the tabs wrap. So it looks like some good old some good old wind forms, um but I need to close these things because I definitely don't need to open them. I'd like to like close all of them. That would be a cool feature.

A

We're not there, though, all right, I'm going to go back to here, I'm going to try to just ignore that these things have been opened and I'm going to go to jh gidlab ci this one and we're talking about just this right.

C

Yep, that's the name of it.

A

That's the name of it. Okay, let me double check, that's the name of it.

A

Yeah, that's the name of it, there's an as if foss, but this one should be fine, because it's as if foss so that's actually, okay, um it really should be just that one. I think so, and maybe I'd like to add a comment here.

C

A

Because it's not going to run it in the jh context like if it runs an ee context or j or as a foss context, that's all great um technically in jh land there's an as of ee, uh which is strange, but um but it's there.

A

uh I'm gonna leave a comment to hear, though of um we don't need to run this for jh since um which should only extend what exists in the main project.

A

I'm gonna add something like that that that all looks valid. I guess I hit commit um fix it for reals this time. Okay,.

A

Now that would be a cool feature for the web ide um to do an automatic fix up commit of the last commit head. That would be really sweet, so I wouldn't have to worry about a message or anything I can just say I just want to fix up the last commit. Oh man that would be pretty cool.

B

A

Well doing a fix up commit, uh it's just changes the the commit message to oh fix up thing.

B

Oh, so you want to change the mess. Okay yeah! No, I wish you could um amend the commits all the time in the web idea, because you know what it's like. You try 10 different things and say.

A

That would be cool too, um so this is yeah fix it for reals this time and the test. We should not have a startup css.

A

Yes, it's not there. This is it. You didn't just scroll past.

C

It I'm pretty sure I'm at the bottom. You can go to the thing and do the jobs page.

A

Oh really like going.

C

A

Here yeah there you go, I'm just going to run 110 jobs, nope.

C

Did it all right, I have to hop off. It was fine yeah.

B

A

B

A

Well, that was good.

B

C

B

At least managed to fit it well turn it off and on again, the event didn't get created, but turn it off and then again and then add in some new comments. The events did get created even.

A

The confidential ones.

B

Yes, yes, so I've recreated the problem, um which is always the first step. I guess.

A

C

So I think yours paul was, it was a perfect fit for this form. It was not a like. This is an intractable problem that is going to take months to figure out. It was like what is the syntax and how is the existing plumbing work? This weird.

A

Yeah ci stuff is a good. Those are good things to throw in there, because it's always there, the rules behind it, I'm never quite sure about, and everyone has different experiences with it and seeing different patterns. So.

C

C

This is a good.

A

Point too yep uh yeah cool. Well thanks for hopping on to help me out. This is a humongous help for me, um so I will catch you all later next time.