Backdrop CMS Weekly meetings, 26 Aug 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: 2021/08/26 - Backdrop Weekly Dev

Description

Agenda: https://bit.ly/2UOEtke

A

All right we are live today is thursday august 26th, and this is our weekly developer check in before we get going on all the exciting things in the agenda. We're gonna do a quick round of introductions. I am jen lampton joining from oakland california. I'm really excited about everybody's enthusiasm and the huge flurry of work we've seen lately in the issue too. So thank you, everybody for all of that. uh Let's turn it over to olaf.

B

Hello, I'm olaf from hamburg in germany, and I don't know what to say so. I pass to somebody else. Let's take joseph.

B

Who doesn't hear me, I guess hi so.

C

Nope sorry hi, I'm joseph I'm from las vegas nevada. I'm a backhand web developer, yeah.

D

uh Luke hi I'm mccormick in san ramon. California, I've been doing some more coding recently, so so maybe I'll I'll try to chip in with something sometime soon. uh Let me throw it over to justin.

E

I'm justin christopherson, I'm in denver, colorado and volunteer with infrastructure things when I can and pass it over to tim.

F

Hi, I'm tim erickson saint paul tim, I'm in deerwood, minnesota and yeah. I'm gonna keep it short and sweet. Today, who's left great.

G

Greg yep did you say: greg yep, yep great um greg. I'm joining from greece uh really excited about working in the issues. I haven't been doing that for a long time and it's good to be back and I'll pass it to andy.

H

All right, andy chilling forward. uh I helped her. On the backdrop. um Sorry I wasn't listening, I'm not. I.

I

Don't remember what I'm supposed to say: that's nothing where you from that's it.

G

Maybe I think nate is the only left then.

J

Yeah and I'm nate, I'm a quick sketch uh on the internet from oakland, california. I think that's the first time we've made a full circuit uh without accidentally calling on someone twice so with nine people. It's great all right, um let's see actually jen I'll pass it back to you for a minute and before we get into some of the other items.

A

Sure so we have a couple of new contributed projects this week, field, visibility, condition and spam bot um andy. Thank you for working on spambot, I'm very excited to get that on backdrop cms.org um and then, uh let's see there are a couple of uh uh new applications in the contribute which I just wanted to point out. It's very exciting: we've had a bunch of people have been pretty enthusiastically moving stuff over. So all right back to you, nate.

J

Okay, well um a couple of meetings ago we started doing a new policy where we have uh a forum post each week uh that is posted to forum or sorry yeah forum, dot back forum.backdropcms.org and in there we solicit ideas for discussion for the next weekly meeting, and there was a couple of posts in there that listed a number of items for discussion.

J

Some of them got discussed in the previous meeting, so I'll just note what those ones were and if you are interested in that discussion, you can go back and watch the uh the outreach meeting that was prior to this one. First, one is better default positions for the breadcrumbs on the installation of new sites issue 41 13. that was discussed in the previous meeting.

J

Next up is config recipes issue, 3763.

J

And I think uh yeah, the question that we have here is: uh what can we do for the next release uh and how can we promote it uh as part of the next release, so config recipes?

J

The meta issue is issue 3763, but I think, as far as what we're talking about for uh 1.20 is this blocking issue issue 3224, which is uh the mod the module config needs to be synced on sorry, the config needs to be synced on module install uh and that issue we've been working on uh this past week, uh drawmont has filed a number of uh updated pull requests uh incorporated some feedback. We've got some reviews on the issue and let me check what the very latest of that issue is uh yeah.

J

It looks like uh most recently tim, uh you and andy have been going back and forth. Maybe uh one of you might actually be the better person to give the update on the status of that issue. uh I'll I'll put forward. You tim first uh as as a candidate sure.

F

uh Basically, I think the the problem was that there was a bug in core that, when you enabled a module that had config files, that the database tables weren't being made created, andy created a pull request which solves that, and I created some sample simple, like config recipes that basically just adds like a content, type and a view and have tested it and it seems to work so in theory, you know just by fixing that other bug which already existed, we've made something of a config recipe possible, but the next decision, but that, but we never really decided what we wanted.

F

Config recipes to look like so now we kind of got to decide. If, if this is really what config recipes should be and or you know, how do we manage them? How do you know do we have a special a new? uh Do we have a new project, type on backdrop.org and blah blah blah? And I don't I don't know what, if we'll be able to make those decisions in the next couple, you know two or three weeks before release.

F

C

H

H

No, I think that that sounds about right. I did.

J

Yeah there's a technical concern in there that uh that the latest pull request has made some comments on about what happens when you disable and uninstall the module like what happens to the config.

J

The field related config when a module is uninstalled and the current uh setup is that we have a an exception in the not maybe not an exception, but we're changing the way. The field sync config sync works that if the database tables are already there, it just ignores them rather than failing to install it just says: oh, I don't need to install the field config because or the field database tables because they're already there, and so that's kind of a keep work around to make it.

J

So you can disable and even uninstall a module but leave the fields in place. But then, when you re-enable the module a second time, it will again import the config out of the config directory, but it will leave your current data alone um and that's kind of a weird behavior but um yeah.

J

It doesn't bother me too much in the way that.

G

I have, I have one concern.

J

G

So if if, if the time between disabling and re-enabling the module passes, and then this schema updates.

G

Happens there with the existing data, how do we I mean? We need to find a way to keep track of the uh schema version. That was there when the module was uninstalled so that we know which updates we need to run just to be safe right.

J

Yeah, so one thing is that the fields module still owns the like. Let's say you have a module that creates a new text field or a new date field. The text module and the date module respectively, are still owning that data and the field is still there and the content type is still there. Even though the feature module is uninstalled, so it's actually still maintained and still updated, and the schema is still kept track of because another module actually owns that content type.

J

So it kind of is maybe a little bit strange that the uninstall of the module doesn't clean up. The content type doesn't clean up the fields which is good from a maintenance perspective that can continue working on the site and it's good from a data preservation standpoint. But it's a little bit strange that uninstalling the module doesn't actually remove everything that it created so.

A

With other things, we're doing in backdrop, though, where like we require that you delete your fields manually rather than doing like a config, sync removes them. So I think there's some stuff that you know we tend to air in in the direction of like if you want to delete it, delete it yourself, and I think that lines up with this proposal.

C

So, for what it's worth, that is exactly the behavior of the features module in drupal 7.

J

Yeah, so I guess maybe this doesn't sound too bad, the the all of that actually will continue working exactly as it is right now without um this one fix that uh andy put in the fix makes it so that when you turn on the module a second time, it doesn't throw a fatal error or it doesn't fail to install because because otherwise it would try to import the field config and see that the table is already there and it would just bail out and say I can't create this field, because the table already exists.

C

Sorry I haven't looked at the latest version of this, but if I remember correctly, the behavior that I ran into when I started that when I made that issue was that what it does is it puts the config file in place and then checks if the config file exists and if it does doesn't create the table.

C

And so it was failing to create the table. um Well, actually, actually the original behavior before I started working on it was that it just didn't. Do it at all, because I needed it needs to run a hook to create the tables.

H

Yeah, I think, if you don't mind, if I just come in here, the the error that you were seeing joseph was because the module invoked when you're invoking config create you're doing that after the config save. So I just swapped those two lines and that fixed it.

H

So what what was happening is it was saving the config and then, when you ran config create, was looking and seeing that the field config already exists, so it was stopping so right now a switch swapped it so that you'll do the config create first and then save the field. I think that's what the config import uh does. So I just copy that from there. Okay, so that fixed the error that you had there before.

C

That's an easy change. I guess.

J

Yeah cool yeah and we do have um test coverage now, uh thanks for for writing that and adding that into the pull request, and it seems, like tim you've, confirmed that it all seems to be working as expected as well. So it seems like overall, really closing in on this fix um to make this possible. I think the only thing that uh we need to agree upon is you know if the compromises that we've put in place are acceptable for the time being, we can change.

J

We can probably change the way that this works in the future. This all is like edge case handling of uninstalling, a module completely and then reinstalling it like. If that's the scenario that is holding this whole thing up, it probably shouldn't be holding it up, because that's it's just a strange scenario.

J

In the first place, I think we'd all prefer to have this capability, um even if it has this edge case um scenario, because right now, like you, can't get into that education area, because the field config just doesn't do anything uh right now, so we're definitely removing a lot of the weirdness and getting it closer to where we want to be in this change.

J

So yeah, that's looking really good, um I think maybe we'll move on from that. I mean there's other things around config recipe still like what we do on backdrop cms.org, how we surface them, how we categorize them and all of that, but I think that um that may be a discussion for um after we get in this. This main underlying issue.

A

Question what's left on that issue, so the it's currently mark needs code, review and nate. You left a code review on there with a handful of questions. I think have been addressed.

A

Do you want to look at it again? I was going to do it a code review, but if you want to see it, maybe yeah well.

J

uh Yeah, but your your card review would be sufficient for me.

C

J

Already looked at it again and it seems, seems fine.

G

Before we move on to the other issues, can we discuss two issues that are config related? So we don't go off that topic. uh One of them is uh luke's. First attempt from what I can remember to to file a pull request, so the issue is 5146 uh and it solves a usability problem when people try to import config and then they get multiple errors, I think it's a great usability fix and it's been rtbc'd.

G

I just would like to have that on the radar.

J

That's a great fix this. This drives me nuts, honestly, like the you need to turn on this module, and then you turn on that module, and then you try to import config it's like now. You need to turn on this module yeah yeah. That would be. That would be a great enhancement. It looks like um I haven't code reviewed this, but it uh looks like someone else. Has it's marked rtbc already um yeah this? This is great. It's only been open for two weeks, so uh yeah great work on the collaboration here.

G

Yeah great great uh thanks for holding there, the fourth uh look and persisting through the multiple code, reviews and requests for changes all the time but yeah and then uh the other issue is 5163, which you have added to the the bottom of the list, which is it solves another similar issue and we actually have a few other issues in the queue that might be solved. Because of that.

G

So another thing is that if, if you try to move config from one environment to another and the target environment doesn't have a module enabled, uh it will start complaining and throwing errors. So um I thought that this this issue doesn't exist in drupal, drupal, 8 and beyond, because they actually have um they track, whether modules are enabled or not in their config, and they have this core extensions yaml file, which finds modules there and basically enables them for you.

G

um So the current pull request. What it does. It basically audits it's a there's, an install, an update, hook that basically audits the site and creates a json file with every module that is installed, uh enabled or disabled sorry.

G

And then the bit that is left to be done is on the target system when uh a comparison is done between what's in the system table and what's on that file, and then, if there's modules that have not been enabled, it enables them for you, the question there remains. When do I do that?

G

Do I do it on chrome, my ideally, my expectation would be to be done automatically and not require a user action, at least for enabling not disabling modules, which is not what I'm after in that issue, but yeah I don't know like do I do it, do it during bootstrap? Would that be expensive.

C

So this is for doing a full import of the config.

G

Yeah so um yeah before we go there, because you mentioned that there's two workflows. I know that most of you, like the most of the advanced developers, they have a um workflow where they sync the staging with the active directory this. This basically deals with the simplest situation, where there's only uh a person that uploads their their config files into active and expect things to work.

G

So they start they go through the sync process, but then it starts complaining about missing modules.

G

If the configuration included the list of modules that were there in the source uh environment and enable them, then you wouldn't have those arrows and that's what I'm trying to achieve.

G

J

Code doesn't do.

G

J

Go ahead, joseph.

C

So that would that would fix it. Basically, if you were doing a full config import, but it wouldn't fix it if you were only moving one file.

G

Yeah, I guess yeah partial import. I haven't thought about it yeah, unless we have, we have it so that uh doing a partial import, always checks for that specific file.

G

Sort of like a special case.

J

Yeah kind of the only protection we have in here is that um when you import config, some module has to claim that config, and so, if a no module claims the config at all, then the config import completely fails.

J

So if that module has a settings page, you know it saves one value, then the config import would fail entirely because there's no module available to validate that that config is okay, so order of operations greg you'd have to um definitely turn on all of the needed modules in like a pre-phase and then do another step to actually run the validation on everything and turning on a module. I think um that would it would need to be done as a separate step like when you run the config import.

J

It would need to like make one request, turn on the modules, and then we need to start over a second request which we already have like. You know it's imported via batch api, so that would work fine. It would just need a preliminary step to turn on the modules first and then in the second and subsequent requests start start importing. The config.

G

Yeah and if the modules are missing, if the modules are missing from the code base, you could complain and say: hey you're, trying to enable a module because it will know it which is not in the code base.

G

If you so choose it could download it if it's contrib and enable it so there's a lot of possibilities to do not throw errors. Rather than throwing errors, throwing messages and saying hey. This module is sort of like enabling config, but it doesn't exist in the cloud base. Would you like to download it downloads it, for you enables it inputs, config, so basically yeah? That's the functionality that I want to implement. Now it seems simple, but I don't know the entry point like. Where would I be doing that only during config config input or.

C

Probably yeah, so um we were talking about having some way of keeping track, of which config files depend on which modules and jen said that she actually puts on underscore modules in her config files already.

C

So we could just do that and then say: okay, this and then, and then that would work, even if you only imported one config file, because that one config file would know which module created. It.

G

Okay, how do we do that? Should we open another issue then and make sure that all core core config files, those those that exist and those that are generated- also have that underscore module? Should we do that, then.

A

Yeah there is an issue for that already. um I don't think it includes a change to all, but I'm not even sure, there's a pull request, but it doesn't. I know it doesn't include all of the core config stuff, but this goes even crazier, because all of the can this would apply to all of config and contrib too, and so we can do it in core.

A

But I think we just need to be careful that we check to see if it exists, like don't assume it's there, because there is going to take a while to get updated. Yeah.

C

Right, uh if you have, if you have, if you create a new config file through code and then save it, it will automatically fill in for you that can underscore config name parameter and put it in your config file. So we can do the same thing with module, which is yeah just basically, when you save a new config file at any time um check what module owns it and put that in there for them.

A

That's great because then we can update any number if it's a backdrop site running a module, even if the module didn't include it. Initially, we cleaning up the mess. That's a great.

J

Idea, yeah. That would make the preliminary like what modules need to be turned on a lot cleaner.

A

J

Now, the only way to do it is to run the validation and see what validation errors come back and say. Oh, these modules are missing and it would be much better if we could determine that information without asking without running any php code, like from individual modules, we could just look at the files and say what what are the modules needed here? Yeah.

E

I don't I don't know greg if that's if.

J

That's absolutely needed for this phase or not, it seems like it would be beneficial. It might be easier than to include it. Then.

D

Would people be be confused if installing a module caused the module's code to change but like it would show up, and if you get it would say well.

C

We would just be changing. We would just be changing what gets put into the active configuration directory. Oh, it wouldn't uh yeah.

J

D

C

Change their code.

J

Yeah, well that that is considered acceptable too, um because we have to do this all the time with module updates that sometimes the config needs to change in the files, and we have update hooks that change what's stored in the config, and so it is expected actually that after you run update.php, you probably will have config changes on production um so that that's the way that we have had to handle this in the past already. So it would be totally acceptable if we wanted to accelerate that process to have an update hook.

J

That retroactively goes through all of the active config and populates all of the module lines. That would be completely within the scope of things that we do normally.

G

I was about to say that, were you sorry, because I was adjusted away for a moment. Were you discussing about contrib config as well or just poor? Now.

J

We could populate it for everything already, because.

G

J

Every module already has to claim a config file if it's in the active directory, it's already claimed by a module, and we can do that by using uh hook, config info and like finding like what module claims each config file and then populating that value.

G

But then I think that joseph also mentioned that we also have config that gets it's. It's not an actual file that we copy over and install it or enable it it's generated on the fly right.

A

J

He's talking about the properties inside of the config files are generated on the fly.

C

Right, so if you, if you go, look at a config file on a on a live backdrop site the first one to four properties will have an underscore in front of them and we generate those four properties on the fly.

G

A

Also, while we're talking about this, um we just talked about config recipes and now we're talking about issues with config syncing. There was an issue that was talking about like how do we handle um enabling a recipe that imports config for a module? That's not enabled and greg. I think you said you thought that was okay, but um because it wouldn't do anything until the module is enabled.

A

But now we're gonna run to this problem, where, if you go to do a config sync on that site, the sync will fail, because the sync won't work unless the module's enabled. So I think, whatever solution we end up arriving on for doing a config import enables modules. We need to make sure that also um we can handle that for recipes.

A

I don't know that needs to be like the recipes need to be blocked on that, but I think we need to figure out, um like maybe don't allow imports if the modules not enabled or enable the module automatically uh for the recipes.

C

Well, you could, you could have your recipe just depend on the other module yeah, and then that would fix that problem. But if you so, if you have a config file and in your module and you enable that module, uh it will check if there's already a config file with the same name and if there is it will just not import that file.

G

So if you, the assumption there is that recipes are indeed another project type and they have a dot info file, so we can specify uh dependencies right. Yes,.

C

Well, I think they should just be modules, but I don't know if we want to treat them differently in the ui or not.

A

And right now they are just modules, so I think we should move forward assuming they're, just modules and then later, if we want to make them into some fancy sub-type of module. We can't.

F

Meet you, I think you're muted yeah. I know um I I it's why I raised this discussion that we have laryn is also is working on sort of a side project which is specifically done designed around paragraphs, but it's a kind of a recipes thing that doesn't treat them as modules.

F

There are some pros to that and cons to that right and um in my fear, you know, I feel like we can't announce a config feature that assumes that configs are module modules until we've decided for sure that's the route we want to go, because I don't think we can just change what configs are. Config recipes are six months from now right that, whatever that decision we make now is.

A

I think like because they're rep modules now six months from now we're not going to not have modules right, it would still work in the future, but if we then define a new type, I think that would also be okay. So I think okay.

F

Right config recipes work. We have them, you know as soon as this patch goes in. The question is whether or not we tell everybody hey. We have this thing called config recipes, and this is what they look like right. If we do that, then it might be hard to change that later. That doesn't mean we can't go ahead and use them.

F

Let's hold off.

A

On the press release but still do the feature.

F

That's that's my point.

D

That's my technical from uh from a usability perspective. Just kind of you know throwing out my impressions. I, the the functionality of config recipes is, is so distinct from the general functionality of modules that I think it's a mistake to lump it with that, like. I think I think that was a misfeature of triple seven's modules, that of drupal 7's features rather module that it dumped in with modules and which is confusing and kind of unnecessary.

D

In there I mean they're, not modules, so like they're, not they're, not really doing the things that others, so so, even even if for implementation purposes, if it happens to be they actually, you know really are modules. I I think from the perspective of the users, it would be good for us to to actually hide that and keep them as a separate thing and and not not let anyone have to worry about them being modules like unless they're.

D

Actually, you know with us, I think a general user doesn't need to know that and that's just confusing.

G

You mean that the the perception that modules means code and people are not quarters, whereas recipes is a simpler thing, yeah. Well, we can market it. The other way around that this is the simplest form of a module and modules are not difficult in backdrop anymore.

D

The module lobby will be in favor of that, but you know the other constituencies might have a problem with it.

J

Yeah, I think I think luke as long as uh I think we could do that as long as if, if recipes became their own thing, they couldn't execute php code and I think that you could have a module that included.

J

You know a lot of functionality and it would have custom hooks and things like that, and then that's really a module that has a bunch of pre-baked functionality and that's just something you can do in backdrop more easily now. But if we take this other route, then recipes would really be. There's no module file. There's no php code in it, because if it is executing php code, then it's a module like. I think that that would be kind of the defining line. Is it php code if it adds functionality? It's a module.

A

So just your module in if you had a module, could you include a recipe in it or do they have to be different?

A

J

It gets into the fuzzy area yeah. What do we call that.

G

F

That I see it, I'm not.

G

Sure, today, the way that I see it is that recipes are the simplest form of a module that can involve evolved to become modules by adding php code and features eventually.

D

Right so well, I mean I mean writ large I'll I'll go like philosophical on everybody for a second, but this. This is what that's. What computers are like, ultimately, we're we're just throwing ones and zeros and like like it all. You know it's all just ones and zeros anyway.

D

So so I mean I mean all this other stuff that we do is trying to simplify it and conceptualize it and and make it be reasonable metaphors that people can use to understand some important piece without having to go one one, one: zero zeros or one one, one: zero, zero, zero, like like we're, simplifying things all the time, so it comes up.

D

For example, it's the same reason why you want meaningful variable names right, like it'll still function, even if you give them terrible variable names, but yet it matters like it matters how people perceive it and how what they think it is and and whether whether the first guess turns out to be correct.

G

If I'm not mistaken, the way that you have tested recipes is by basically making a blank module. That's what you did right.

F

Yes, but I I want, I want to talk about things right now that can get into the next core, and I I don't see this discussion, helping us get anything into.

F

This is a much bigger discussion. I think we have to come back to it later as fun as it is. Yes, oh, I really want to have it.

D

This is an incredibly important feature like like it's groundbreaking once this gets in. So so, let's, let's shepherd it in bigger small steps, but let's, let's get it there somehow.

J

Yeah uh all right, let's see next up jen we've got uh this. Oh boy, we're gonna have a hard time fitting everything user. Compare visibility, rule issue 4829, uh I don't know anything about this. Can you give us an intro.

A

uh Yeah, so this is a visibility rule that was in panels that didn't make it into backdrop. It allows you to basically it's something different like when I'm looking at my own user page versus when I'm looking at someone else's user page, so it compares the person who's looking versus the person you're.

A

Looking at, um though that use case is rather simple, the user interface for it is very convoluted because it's like user one user two and it doesn't really help anyone know what these entities are, that you're comparing uh and so I thought um tim was testing it he's like. I don't really understand, what's going on here, which is fair, um but I thought that now that we have relationships and layouts, I might be able to come up with a more useful um example, because we can use the author.

A

You know note, author of node relationship, that's in core to have like a third user, and then you could compare user one and user three.

D

A

And user three, rather than you know, just the two that my um your own youtube profile. Other users profile example- and I thought maybe if we had more examples of users to compare, we could maybe come up with some better terminology around how to explain to people what the things that they are they're choosing. um So it's uh I think it works as it is now, but the user interface is not intuitive, and so I'm hoping to come up with a better example. That'll help us change the text, but I think functionality wise it works.

A

um I don't know if it has a test with it or not. If it had one in panels, it might have one in the pull request. But I'll have to remind myself.

G

Would it would it help if, if you were being destructive, descriptive about that, instead, instead of calling them user one and user two and you actually used more words, I know it's.

A

Yeah, it's just.

G

A

Hard because, without knowing the use case, that people are building for how to describe.

D

A

Is which is hard um so yeah? I thought, having maybe more examples of use cases might help us figure out how to name them more appropriately in an abstract way. That still makes sense, which is very hard.

A

um Tim also recommended, um rather than asking user one user two, but we could provide, is a very limited feature. That's like compare currently logged in user with profile being viewed, um but that would limit the functionality to only that one use case, whereas right now, when you have the ability to use user one user two, you can also compare to like a node author or any of the other users that we can have in core like via relationships.

A

So maybe it's worth um asking everyone in this zoom. If you think it would be better to limit the functionality in order to make the user interface make sense to people or if we're gonna expect that people will trudge through the interface and trying to figure out what this feature is.

G

So you mentioned you mentioned this one use case, but there's more is the the number of used cases a finite number or can be limited to imagination. Okay, yeah, it's limited.

C

It's limited by the number of relationships you have on your website.

G

A

But the one use case that we are trying to put in court is like the most common one, so it might be even worth like building us doing the same thing twice, but doing it once with an interface that people understand and once with this like convoluted interface, that you could use for a more sophisticated scenario.

G

Yeah under under the hood, they would both be the same thing yeah. I think that's why yeah that's a good compromise. I think okay.

F

That's where I'm leaning, doc wilmot, oh jack, what about go ahead? You talk.

H

Yeah, I was um sorry I'm in and out of the conversation, sorry, but um why isn't this good enough for contrib? uh Is this something that should be in core? I don't imagine that that would be a very common um something that would use commonly well. Why did we think that it's something that should be incorporated.

A

It is pretty common that you want to show like, especially if you take into scenario like the dashboard right like um a lot of times. People have their own dashboard, but you might want to see like okay. These are the eight things that I'm responsible for is editing, content or whatever. But then I want to go see what's on like bob's dashboard, and I don't want to see my tasks on bob dashboard.

A

I want to see box tests on bob's dashboard, and so you have to have some way to tell those pages apart, so it works for dashboards. It works for like actual profile pages. It works for anything where you have a website with a lot of authenticated uh users with higher levels of access, it becomes pretty common, um the it was in panels and, I think, just overlooked, like a lot of our visibility. Conditions are missing.

A

There's like a taxonomy term, though it's missing a user, which is all the entities that aren't nodes, just didn't make it into layouts when layouts went in, um and so I think I think it's a valid use case, the edge casey ones where you need the really complicated interface aren't uh so that, I think, is another thing where it's like it's.

A

Maybe we just wanted another thing, good question, for everyone, too, is: um where do we put all the stuff that was in panels that didn't make it into like? Should we just have a drip module called panels that has all of this stuff that was left out like because we and then we could have the like? Oh did you need the extra features from panels install the panels module for background layouts.

G

Bundles leftovers.

A

J

I I'd really prefer that we rename space that so that it makes sense in today's context, rather than looking backwards, you know say, like you know, lay out extra conditions or something like that or lay out extras. You know.

A

D

It strikes me that that the dashboard is kind of a subset of a fairly large area of usability that opens up for just customizing administrative interfaces.

D

I mean people, don't customize user interface, administrative user interfaces, but I mean they should like that. That's actually uh an important thing, so it'd be nice to have that as uh as an option.

C

I think I'll. Currently, a lot of things in our administrative interfaces are hard-coded, so we can't really move them around that much.

J

C

B

J

Maybe we need to move on from from this one. We've got a couple more things to get through and.

A

J

A couple minutes to get through it um if it's okay, I'll skip the next one, uh but I'll make a note of it that there's uh there's an issue. That's garnering some discussion about changing the default homepage from slash user to slash node. uh If the user removes the default home page, which is home, uh that issue is 33.43.

A

One question on it: not nate, do you remember when it was changed to user and if that was done intentionally from node? I don't remember.

J

um I don't remember quite when that was okay, it might have been done when node module was optional and so like the testing profile needs a home page that isn't slash node, because slash node isn't turned on in the testing profile.

J

That might be it. It might have also been that maybe we were hoping to get rid of slash node. You know at some point entirely, but I checked the code base and it's not deprecated, so I it doesn't seem that if we did decide that it wasn't reflected in the code base.

J

Okay, uh let's see oh boy, there's so many things, uh a details.

A

J

Element issue: 50 90.: uh oh that one was discussed in the previous meeting.

A

Yeah, I think the one question I don't know greg. If you know the answer to this already was, do we need to probably tell you sure we need to probably go for that. Did you want that answered or.

G

Well, it would be good to have it since we say we still support. Do we do we still say that we support older browsers or best efforts.

A

We don't support internet floor, eight.

G

No, I mean I mean nine, nine and above that's what I meant.

J

Yeah, I don't think we have an official policy on that, although we should, because it's probably time to update things considering or microsoft- doesn't even support internet explorer at this point, so yeah yeah, our compatibility- uh has stretched further than that is reasonable. Apparently,.

G

Yeah, I think that the stats now are like 98 of the internet, uses chrome or something like that. A ridiculous amount, so we're talking about a fraction of the two percent of the internet.

A

Okay, I will open the pmc issue and we can talk about it. There.

G

Yes, let's do that. Having said that, there's there's developing countries that are still stuck on windows, xp and uh government agencies that are stuck in windows, xp so or schools, or things like that. So just putting it out there.

C

Yeah, a lot of health care companies as well like I used to work at a health care company here in las vegas, and I did I.t and we had to maintain a. We had to maintain an exact point: release of internet explorer 9 in order to access records at one of the hospitals.

C

And yep it was a disaster, but we had to do it.

H

Behalf of all healthcare professionals, we don't change. Okay,.

C

Because it wasn't even the healthcare professionals who wanted to do it, it was like the iet department at this. One hospital just didn't want to update their infrastructure and they were running like an old is asp server for their medical records.

G

How do we, how do we load polyfills like do? We have an example in core of something like that, or would it be something like the way that jquery is being loaded or does it need to be conditionally loaded when, in pages that use that or always.

J

It it depends, I think we might have the html5 shim is still in there.

A

Specifically for internet.

J

Explorer and actually depending on the outcome of that internet explorer discussion like that would mean that we might remove that shim, because that only exists for internet explorer.

G

Yeah, very nice, okay I'll have a look at that thanks.

J

um Let's see on the category of things that I just will mention in passing this uh there's a long-standing sim link problem that if you sim link the backdrop web route, then uh the module updater will fail to update modules issued 3008 3008.

J

um This issue has been coming along nicely: uh indigo zella and the author of the pull requests uh and then it's brad, uh brad bulger um have been not knocking back and forth a lot of uh suggested approaches, and I think that they've landed on one. That seems to fix everything.

J

So the latest pull request looks pretty promising um from brad there, and so I'm hoping that that will reach conclusion here shortly.

B

um Nate, have you seen that there's also a new pr of indigues indigo sailor.

J

J

uh Can you elaborate on that olaf.

B

um It's a bit hard for me because it's a very technical issue, but it's a much more smaller pr for a three-liner, and um so I can aside from the comment of indicator, she says it only covers the problem. Backdrop root is a sibling and doesn't bother any other siblings, and this might not fix all possible problems with siblings. But fix is the one we're struggling with so long now.

J

Oh yeah, how about that? I had not seen this this version interesting.

B

And yeah, I think it's also interesting because you the time ago, said that it would be good to.

B

Keep the current behavior in addition to the new behavior, where either the destination directory is assembling all the entire web route, and so I'm not sure if I understand it correctly, but I guess uh indicator shoes chose another approach here.

J

Yeah, it looks like yeah, I think. Actually I could be wrong about this, but I think brad's latest pull request also is the old behavior and the new behavior both so.

A

J

Looks like we have two approaches that both may provide the new behavior. That is actually the current behavior. That's there and the the previous triple seven behavior, which is what we're trying to solve so.

A

J

It's interesting that indigosalis is so much simpler and solves the primary use case just as well. um I think maybe I'll need to read this issue and see if brad uh and indigozella can reach a compromise or conclusion on their own, because I would love it if one or the other would say, let's go with your pull request if they both.

A

J

The problem in different ways, because I don't really have a preference for implementation, one versus the other- I don't think um as long as they both can be confirmed and tested that they work. So so I think what I'd like to see is one of the other say: let's use your pull request.

F

I'll just uh indigozella asked to put this on the agenda and I think her hope was that she would get direct core committer feedback.

F

So it could be that that's what they're looking that they they're reluctant to decide, because they want to know from a court committer- and maybe you just gave it- maybe that's all the feedback she needed was.

J

Yeah well, I need to well the indigo. Zealous is a lot easier to understand, and so I can provide the feedback on that right away. Saying oh yeah that looks like it would be adequate um and if brad is happy with that, then let's go with indigozales, because it's really easy to understand.

J

However, brad's I think is more comprehensive and that it deals with sim links everywhere, nested anywhere within the backdrop route, and so that may be preferable because it's a comprehensive fix, but I haven't, haven't, tested and read through his implementation, because it's a little bit more gnarly to wrap your head around. So I think brad's would be preferable if we can get it confirmed and working uh confirmed and tested if it's a comprehensive fix and if we just want something that fixes the immediate problem, then let's go with indigozales.

J

So that's kind of my my take on that uh yeah. So if brad's happy with indigos ella's, that's the fastest way forward and if ina gozala wants complete fix, reviewing pads would be a different option.

B

And also, I think manual testing is needed as a next step. So if anyone wants to help with it, because both prs and the current state aren't confirmed that they really fix the problem.

J

Yeah and we can have um zen ci- confirm the problem by updating the does nci script. That runs the test. I think brad's actually does change this nci script into indigo. Zealous does not, but we could temporarily change the script either way that is set up in in brad's pull request to check both ways uh that that are confirmed or you can set it up locally, but it definitely is confusing.

J

Setting up all these sim links all over the place to actually test the problem.

J

uh Well, anyway, I hope that that answers things. um Let's see, we still have more, um let's see, let's actually skip down.

J

uh We have github actions, that's continuing to come along uh we'll skip over that, though, for today, um one issue that I still would love to see in 1.20 is the media library dialogue we're still trying to get that file. Name filter, exposed issue 3293..

J

um I think we have the time we have the approaches that uh are we. We have gotten really close uh to working in this one. So I think that 3293 is a matter of uh checking the approaches. We have resolving any last issues with it um and then deciding like you know which approach we want to do. I think uh andy, you posted a comment, saying hey. We had one that was working a long time ago.

J

Maybe we should just go with that, um even though it's not like a comprehensive fix- and I think that's you know that that's an option. That's on the table too. um I'd love to see all views filters working if we can get there because it seems like we're really close, um but if we needed to compromise for 120 and use the kind of halfway fix be kind of okay. With that.

H

Yeah that that comment was me giving up. Essentially I I spent a while looking at that and the jquery that jquery has never been my friend in the first place, but I think it's it's really hard to try and figure this thing out that I give up. So um looking back, I think, to solve this specific problem about being able to filter for files, then just start autocomplete um field that sounded brilliant now. Somebody else is gonna have to try and figure out how to get all of these fields to work.

H

uh I've tried it's hard.

J

Yeah well, thank you for your efforts there um and yeah. If it comes down to it, we we may do exactly that. You know: go go back to that partial fix.

J

uh Let's see- and lastly, I want to talk about telemetry issue. 285 is the big meta issue. There's been a lot of activity on this in the last week, big thanks uh to draw month for doing the reviewing on that uh checking it out. It looks like herb.

J

Duel also uh came through and re-reviewed a couple of times as well and this issue now it seems, like people are in agreement that, like this looks good uh for core inclusion, like the we've had it code, reviewed now uh by multiple people, which is awesome, um I had hoped to get the project module portions merged into project and deployed this past week, but I'm sorry that I didn't get that together, but I think that we're now even more confident than we were before that this looks all ready to go so next step here is to get project module changes merged in so that telemetry and core actually has something to communicate with.

J

H

Have a look at that as well, and it looks perfectly fine yeah, just just one I'll: do a prt project module just a couple little things that I found as well: okay,.

G

We're still we're still we're still adding it in core uh disabled. Does the plan right? That's the plan.

J

Yeah- and I want to talk about that- a little bit too- that.

J

This could be a like fuzzy zone that uh that, if we put telemetry in core and have it disabled, um we could use this uh period in between feature freeze and release to finesse. What it looks like um when it's enabled or when on new installs or something like that.

J

I don't think that's a big ask, and so I think that we could possibly use that window to do any final finessing as far as like announcing that it's there or maybe even leave it off by default, but add a checkbox during the installer, something you know to better expose this telemetry module to end users does.

A

This mean we're not worrying about putting it in as hidden first before it's disabled.

G

That's what I meant I actually meant hidden.

J

Oh yeah, what do you think jen? I.

A

Don't know I got excited.

J

Yeah, I don't, I personally don't think it needs to be hidden. Okay,.

A

J

Because it doesn't do a whole lot, obviously, um and um yeah and.

A

Everything's really.

E

J

Think that it exposes like adequately like here's, the information you're, giving up it's pretty clear about all of its consequences. I don't think that any of us.

G

Any of us actually has any uh reservations like we're happy to put it there. It was. It was because we thought that we had reservations that we sort of, like said, let's add it as hidden we're more than glad to yeah yeah, okay,.

A

J

I don't have any reservations.

A

Recommending it under an overabundance of caution or it's like well, if we have to put it in court, we could put it in as hidden if anyone's worried about it and then, but if no one's worried about it, I recommend disabled yeah that'd be great.

F

Maybe nate are you suggesting that that by feature freeze we have it uh committed but disabled, but that we work on the ui and could potentially enable it before the actual.

G

J

F

At least give the.

J

Option to enable it in the installer yeah, maybe we don't enable it by default, and but maybe we expose it somehow surface it somehow to to the end you're saying you know, could you please turn this on or but the nice thing is for existing sites. The existing sites problem, I think maybe we should hold off on.

A

J

Hey, if you have an existing site turn on this module, we'd like it. If you did that, I think that that shouldn't be at this phase. I think that we should do that. Probably in the next release.

J

Yeah sounds like everybody likes that idea. So, okay, uh well we're super. I mean we're not super overtime but we're running over. uh Does anyone have any last thoughts or anything else, they'd like to bring up before we close.

F

I'll just say: olaf hang around for a minute. If you want to talk about nodeless content types after the meeting.

B

I

J

Well, let's uh close for today, then thank you all for for being here and thanks for all of your work, uh this past week and always um yeah. This is definitely the most exciting part of a release and looks like this. One is no exception that the flurry of activity- oh and it was mentioned uh before the recording that we were talking about code sprints um this weekend, uh depending on how things shake out, um there's a good chance that I'll be available most of tomorrow and hopefully most of monday.

J

So this weekend will be a good weekend for opportunities tim. You you had suggested just if somebody is doing actively like setting aside time that uh start up a uh a chat or sorry a a jitsi or a zoom, or something like that, put it into zulip, and then anybody that wants to join can- and I think that we'll probably be doing a little bit of that uh this weekend- uh code freeze itself isn't actually until wednesday. So we actually have. You know a little bit time after the weekend as well.

G

Yeah, we usually start.

J

G

Channel that says, uh 120 sort of like release sprints and then anyone can just jump in at any time and say: hey, I'm currently working for like an hour or two or whatever, and then others can jump in.

F

And we did designate uh 11 to 4 p.m. Pacific time uh this saturday, some of us will be online and so that's a designated time. But people can work anytime and open up other spaces.

J

Okay, all right! Well, thank you. All um I've got to drop and yeah. Thank thank you and we'll end in the broadcast. So thanks.

I