Backdrop CMS Weekly, 14 Dec 2017

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Backdrop Weekly Dec 14th

Description

Today’s development agenda: http://bit.ly/2Bg3auD

A

All right we are on air, it is Thursday December 14th. This is a weekly check-in meeting on active development tasks for backdrop CMS before we get down to what we're working on in backdrop the product.

A

Let's talk about, what's been going on in contribs this week, so we had a couple of new project releases patch info, simple pic, quick block and wordpress import or address important I'm personally, very excited about in opiez, currently working on block reference module, but he's asking for people who want to use it to test it and leave feedback in the queue before the first official release comes out. So if you want to grab that one from github, it is ready for testing, though not officially released, um it looks like drush. Has some updates Jeff?

A

Do you want to do those.

B

Yeah, basically, last week, Jen Hopi tested PR for enabling and disabling modules via trush, and we found some more issues, but Hopi presented a fix and so we're ready for another round of testing, particularly on Drupal 7. When you have the backdrop crush extension installed and also particularly PHP 5.5 I.

A

Have a question about that is: is all of the work going on in that branch in your repository, called the ongoing set saga of n is that the branch I should be pulling or are these separate for requests? Houseware? That's.

B

The branch, okay.

B

We can't do PHP 5 3, because drush requires Phi, 4 or higher for a dot X.

A

Okay, I can just.

B

So that's a wrap with with and if we can get that in I'd like to cut a 1 dot, oh release. Okay, it's like I think that's a good MVP for some pretty good functionality. There. Okay.

A

I can test that after this meeting, great I think I did test it after Opie's last changed and it didn't change anything for me, but I wanted to make sure I was testing the right things because it seemed to like the fact that there's work going on nothing's changing for me. It seems like the problem might be me again today, because.

B

I, don't think he actually pushed his stuff. He just recommended the change and then I put and then I put it in at some point, but I'm not sure when that was so. Maybe there is a little crossover there. Okay, I'll.

A

B

A

And then we had been doing some discussion in the page list, nodes issue and I think decided to pursue the rabbit-hole approach, which basically we'll take notes and make their paths access tonight or otherwise based on settings, and so we decided to try that rabbit hole, module which had been ported to backdrop by herb dual while ago and then layer and apparently went in and sorted out some folks and so that module kinship module is currently ready for testing and so I think. Our plan for future integration would be to test the control.

A

For a backdrop, and if we like that approach, we can then try and figure out how we can take the settings from that. Can triple and merge them into core I, don't think rabbit hole. Module itself would be moved as a standalone module into core I. Think that feature would just become part of node or entity or wherever it is. We decided to put it so we'll just figure out that down the road, but the first step is getting it working per backdrop.

A

So, if anyone who has a site that needs to prevent access to its specific node type by its path, that's ready for testing how to you all right. So we have a ton of stuff going on and can trim. So thank you. Everybody who's doing all the amazing work on contrib, which probably means you're working on sites and backdrop, which is even more exciting. This is a bunch of great progress.

A

Lately we have a couple of things going on on backdrop: CMS websites, like our CMS backdrop, CMS org, main website, Jeff I, know you've been working on this try out backdrop, page I have intended to test that last week and failed I'll test it today.

A

I promise last week ahead, impromptu haircut for my dad that I'd forgotten about today and then also happening on our forum site forum that backed up CMS org is up. We now have adequate spam prevention on it. So, hopefully, most of our swarm forum, posts won't be posted by Russian Box anymore, which is great and I would encourage anyone who has any questions to jump on there.

A

The last few times I've jumped on there I've been really impressed by the number, who are asking questions and I tend to get distracted and want to spend time answering them for people, because people are great, so I would encourage anyone else to just poke around in there and see.

A

If there's any questions that you can answer to, I think it looks fantastic I would love to announce to the world that we have a forum in order to do that we need a blog post, written on a backdrop CMS to Org, announcing it and I had reached out to doc Lamont last week asking if he wanted to be the one to write the forum post since he's the one that did all of the work on getting the forum ready, I haven't heard back from him yet but I would it would be good I think if we got another person who was interested in writing that blog post in case he decides, he doesn't want to write it.

A

I would love this month's newsletter to highlight the forum. We also want to ask kind of a survey to our users to ask what other tools might benefit them in terms of helping them decide to adopt bactrim or helping them through the actual adoption process, and so I think being able to say we know you wanted a form. We built you a forum.

A

What else do you want is a great way to kind of open that door to taking feedback from people in terms of the tools that they want for working on backdrop, and so having the post-its like we built you, a forum is going to be really important and I know a doc. Well was really busy working law in the forum, and so I don't want to ask too much more of his time as well, but I would love to be able to announce all of the fantastic things that he's done.

A

So if anyone else is a nurse and working on the forum blog post, let us know we definitely could use some more help in that area.

A

Okay, that's it for backdrop, can't rib and our infrastructure websites. Nate. Do you want to talk a little bit about progress on backdrop, CMS itself, sure.

C

So, let's talk about the next bug, fix release which is back drop. One point eight point: one I know that we are closing in on one point: nine here and we're good to actually release any bug-fix releases for one point eight, but it looks like there's a good chance that may soon happen thanks a lot over the past week, actually or two weeks. I wasn't in the meeting last week on tackling bug, fixes and Jeff I saw you've been merging fixes into the 1.8 branch, which is very highly appreciated.

C

Thanks for encouraging a team collaboration there and getting some issues knocked out, let's see, there's a number of items that are currently in the RT bcq about 8 or so items that have been reviewed by the community. That's fantastic I started looking through those, and many of them are looking like they're in really great shape, so we'll probably knock out a couple of those today like cannot delete cloned views issue 2845 doing a node should not validate the forum.

C

2594 and I saw you knock out the add a redirect link shouldn't be shown if you're creating a new note, which of course, is obvious like.

C

Why would you create a redirect if you haven't even created the content, yet doesn't make any sense so 2046 that one was knocked out awesome so we'll be doing more issue review today, hopefully during sprinting and knocking out some of the items that have been piling up, I hate to just keep piling on the praise to one person, but hope he's really been like just knocking this out like doing tons of great code reviews and moving things forward.

C

So thanks again for all of your work there, let's see I, always call it these cross ports from Drupal that I'd really like to see finished. I've got four issues: twenty thirteen and then at 2774 2775 from seven seventy six that are various cross ports from Drupal, 7 top 50 and higher, and all of them have had poor requests to kind of finish them up but haven't gotten completed.

C

Although I know, with the time running out for code freeze on 1.9, those might not be the highest for any things, because we can always continue doing bug fixes after sorry after feature freeze, which is January 1st.

C

So at this point like yeah I would be great if we can get those merged in or get to get those fixed up so that we could include them in one point 8.1 sooner, but really they're, just not as high priorities as adding new functionality is for for the 1.9 release, so yeah, as always, would be nice to get those in, but I'm not sweating them at the moment, at least.

C

Let's see so that's the status of things generally for 1.8 minor release and bug fixes there anything else. You guys would like to add regarding fixes that have been happening recently.

C

No okay, okay! So let's talk about 1.9, 1.9 media library still is our highest priority item here and we're still kind of stuck in the same pattern. Holding pattern here that were hung up on issue 27 96, which is the ability to edit an existing file, still really close on that. But we just haven't seen activity there I know.

C

My poor course is the last version that was put up there a while back. If anybody wants to pick that up, that would be a huge help. Just push that issue. It's saying: hey I'm, gonna work on this.

C

Otherwise I'll assume that I'll still be the one to work on it, but yeah don't feel like it's not an opportunity.

C

After the after editing, existing file is complete, 827 96, it kind of unblocks the subsequent issues, field, bundle area file, entity bundles, which is 26 32 and uploading a file independently of other types of content. 26:33 we've only got two weeks left, I know before future. Freeze I, don't feel like these are super ambitious features.

C

I think we can still knock them out for 1.9 and we do need something like 1.9 right now is essentially the same thing as 1.8 and I mean we've been seeing a lot of really a lot of community activity in peripheral activities outside of court and can't rib and the forum's in the website the showcase kind of a lot of really exciting movement over the past couple of months, but most of that activity hasn't been in the core queue and that's fine, but it still would be good with us to get something as a banner feature for 1.9 to show progress in the actual product so hoping file management.

C

We can still put those things together. Let's see, we've got a new item in the agenda. Here. Translation has been a really hot topic, with the translation, extractor getting updated. The translation, server localized backdrop cm esta org kind of coming together and Jeff you've been doing a little bit of communicating working with Olaf these past couple of weeks. Do you wanna talk about this translation, section sure.

B

Yeah, we made a conscious effort to start making backdrop for good website a use case for making a translated website for multiple languages and that's going great and we're revealing issues and sorting about which ones are kind of core issues and which ones are state specific issues. So that's kind of what's driving a lot of this in Olaf is, is doing great, with coordinating us and helping us, especially folks like myself, that don't really understand the core concepts of translating so I'm getting better because he's helping so from that came in the internationalization module.

B

There's a the ability to hide the navigation links that are displayed per node, so ported that into core, and there were some updates on that I'm based on some feedback from Nate last week or two weeks ago. Maybe- and that has an upgrade path, not just for that one feature. So that's I think that's great for a few and we'd like to get that in so we could get that onto our working websites and then the CMI translation. So this is basically being brought up because of blocks.

B

So we want to be able to translate blocks and blocks or configurations, so there's no reason to not translate everything so I made a start on that to make this get translated, function, I'm, not sure how hook config info and peix play into it. So if maybe, if we could get, if get our heads together get out the video a little better outline I'd love to continue working there and then block translation came up because a backdrop for good.

B

Yeah, that's the fact. I already spoke about yeah, so some exciting work there. It's been great working with folks on this stuff.

C

Awesome I've got some ideas regarding how config hope, config info might play into things and and I've been brainstorming, some other alternatives there, because it I think we might have some better options. I'm gonna write that up right now, I get that pushed it by the end of the meeting, probably great, or we can do it after talk afterwards.

C

So, okay, awesome, yeah, all that's very exciting. Yeah and translation has been an item that we had slated for 1.8, but when things don't happen, sometimes we'll give a cycle to kind of let things kind of incubate for a little while and see if we can hit it for the next release.

C

But yea 1.9 is still in play, we're just not targeting it as like a. We absolutely want this to happen, but it might, and that would be really fantastic.

C

Ok, I think that's it for the product updates for 1.9.

C

That's it tada. We go into general discussion and Q&A.

C

And it's funny, it looks like the only idea that we've actually got here in the agenda is actually talking about configuration management translation. So maybe we should roll into that.

C

Okay, I'll try to summarize the issue here. So this is we're talking about issue 704, which is ad config translation, in which.

C

Let's Jeff, can you talk about how the port requests currently works? You have to call the config and then you said, get translated, and then you specify the key. Oh.

C

I see and then you pass in args, like as you would for, for the T function. Oh, that makes sense. So when you say get translated, you specify the key of the config item that you want to translate, but then you also need to specify the substitutions, like you would in the T function. That makes sense.

C

Okay, yeah. That makes sense. So the question then comes in to how we make the string extractor, the P otx module, aware of what strings are translatable by looking at a module. So if a module had some config in I had some translated strings or strings that need to be translated in config, how would the translation extractor identify those strengths of saying these strings need to be translated and put them into a p OT file?

C

The translation template files for the p OT file is so early on. I had suggested that so ok, let's first of all, compare ourselves Drupal, 8 here and hunter plate managed. This problem is that they they created schema files for their config files, so all of the config also has a separate schema file. So every config file has two files, one for the actual values and then one for the schema and the schema specifies data types like.

C

Is this a string or is this an integer, but it also has more complex things other than just doing an integer like this thing is a translated string where that gets. Messy, though, is that schemas can be like groups of things can be repeated like if you have a view and you specify hey. These are the filters that are applied to the view. Those filters can be repeating.

C

So it's not quite just so simple as saying you know, this static string will always be this particular type of data, because you end up with dynamic data structures that can be expanded like into really big things. So this game is get really complicated.

C

I was thinking about initially co-opting, the idea of schema, but instead of using a separate schema file, we'd use, hope, config info as say, these are the keys that are translatable and that solves the basic problem of which strings within the config file are translated but doesn't solve the. How do we deal with complex data structures that, like we don't know what the keys are that need to be translated? Quite yet so.

C

That was my initial idea and a new idea that I'm thinking about is instead of doing it through hook, config info or schema file, because dealing with those complex data structures is a problem in both of those solutions. It's complicated and it's hard to to describe that data format.

C

A new idea instead is at the time that we write the config file like let's save use, is saving the view config file, it's responsible for adding all of the keys, adding all the values and then at the very end it says, config arrow save and it saves the config file. I was thinking what, if, at that time before, we call the save method on a configure item.

C

If we also had the ability in views to say you know, config arrow set translatable and then the key and then true or false, and that would save into the config file itself an array of all of the translatable keys.

C

So up at the top of the config file we currently stored. Some really basic metadata, like the name of the config, is in the config file itself, as well as whether or not the config file uses static. In-Memory caching can underscore config underscore static. Equals true, I think is what it's called and I was thinking that, using that same, like config metadata concept, we would use this method.

C

That says like set translatable and then take a key, and if you set something as translatable, it would add to an array up at the top of the file to be like underscore, config underscore translations or underscore config underscore config. Actually, config underscore translatable or something like that, and then it's just a big array of all of keys and then what the the VOT extractor module would do. Is it wouldn't have to do anything crazy with like parsing schemas and then finding all the keys?

C

It would literally just read the config files and say what keys are translated in here and pull that list out of the config file and then find the Associated keys and say these are the keys that need to be translated in this file. So wouldn't do any logic or expansion of like reading a schema, and these are the keys. It would literally just say, give me the keys and then get the associate keys and say these drinks need to be translated.

C

So basically it it's just the responsibility of this being like a magic thing that happens like with one thing over here and then on save it combines the two things together and then magically creates like a list of translated strings and instead just requires that modules. If they want translated strings before they call configure, are safe, they say which keys are translatable beforehand.

C

And I'm writing that up in a way that you know, and it doesn't need to be understood via hangout in the middle of the week, but I think that that might work. What do you think Jeff that.

B

Sounds great and in the case of blocks, so that would be the block module, so we could like set in the translatable x' array in the config file, the body value or whatever. We want really so like with nodes. You know that in the node table, there's like a canonical language, node and then the other ones have references to that. Node ID and that's how they. If your URL is like mysite.com slash yen, then it knows to figure out which one is the English node or the French node and vice and so on.

B

So how can we make this kind of reference system with blocks? Can we like save a UUID to the block config file, or something like that? So.

C

Because blocks are config and pipe design, you know they. They are going to use like a different translation system than content, because content like nodes and taxonomy terms have content translation, but config is using a separate system because things that are in the config files are considered to be well, not content.

C

So what that would ultimately mean for this would be I'm, not sure this would work well for body or for node type. What I'm saying for block body content, but just because it wouldn't work well, doesn't mean it wouldn't work, because what you'd end up with is you'd end up with a long string of text. That is the body content of the block and then that string would be registered into the I.

C

Don't know the name of the table, but at the locale module table for like this, is it translates or translatable string, and then someone would go into the locale module interface and find untranslated strings than they would be able to translate that entire string. So basically the block content would be run through T. Ultimately, so it wouldn't be managed in a way that there would be.

C

You know, different blocks and like a different interface, saying, like edit, the french version of this block at the Spanish version of this block. It would literally be go over to the interface translation, UI and then find the block string and then translate it via locale module, I.

A

Think that makes the most sense for config translation stuff so like blocks that are config. And if we have blocks that are content like ageless node style, then it would make more sense to have the like edit, the French edit. The Spanish version of the block yeah.

C

You know what some I mean the suggested approach here, that we ultimately running-config translation through T we're doing it this way, because this is basically the way can fake is translated right now, it's just that it's a manual process and we're running the T function on variables. So this basically just evolves the cept to to make it officially supported.

C

But once during this approach we basically can real Everage the locale module to do config translation, and we don't explicitly need a new user interface to translate config based strings, which is nice, but you could make an interface that manages config based strings. That says like on a settings form, you know you save it and then maybe the settings form could also have another tab. That is like translate and translate all the settings that were in the form that you were just in and they could do that by taking a look at the config file.

C

Seeing what keys are in the config file that need to be translated and then making a page. That is just the strings for that particular config file and you could translate them independently, the rest of the user interface. So that could be like an optional like improvement. If you wanted an easier way to find your strings, rather than the need to use the little Cal module to go and independently find those strings, then translate them as part of the rest of your website.

D

Yep I think that makes sense what you said made because, for example, look at custom blocks which may be editable by content. Editors and.

D

Firstly, it's very complicated to go to the user interface translation page to translate a body also, their interface is not adapted for this use case. I think you have very small, it's text, widgets and so on, and maybe they have not the permission to translate the interface and so on. So it would be nice if they could.

D

The person who does the configuration has also the possibility to translate it right at the place that would solve the permission problem of the interface problem, so I think that makes perfectly sense and using the user interface translation is also sort of misleading, because it's not the user interface. It's graduation, yeah.

C

Yeah, it doesn't work real great for blocks, it would work, but it's not definitely not a good solution. Yeah.

D

But so we talked a lot about the custom blocks and I, don't know if you had the ability to follow it. We also talked about the node s, block approach, which is more workaround, which also the reason why we pushed the activities for rabbit hole where you can use nodes and hide them and use them instead of custom blocks, which sauce I think a lot of problems, because custom blocks are also problematic, permission, wise and and a top of that.

D

If you use nodes, you can also add other fields like an image field and so on so I think that's a very good solution, but it's yeah I, think more, a solution for experience, site builders and not for first time, users of backdrop and I had actually I had the idea. If we could rebuild custom blocks, maybe not at the moment, but maybe for the release, don't know one time, 1.10 or 1.11 in a way that it's internally uses these hidden nodes. I.

D

C

That black slave yeah yeah I what.

D

We do now as a workaround, which is I bet. I can do. We could arrange it internally so that this certain kind of nodes would be used for custom blocks and it would solve a lot of problems, permissions field ability and translation. At the same time, just an idea, yeah.

C

I wonder if we could also yeah there's problems with this, but maybe you know there are problems, but maybe we just acknowledge them and say yeah. That's a problem like that. The one I'm thinking about in particular is that any type of feel double entity. Putting it into a layout is mixing configuration and content because the content is stored in the database and the layout is stored in configuration and as soon as you have anything, that's filled. Full content place into a layout.

C

You've found the two together which makes moving configuration between environments a lot Messier, because sometimes the content won't exist on the destination.

E

Location they're already doing it with fields and layouts or existing.

D

C

But yeah we I mean we explicitly allowed that when we said it's okay for you to just place a node into a layout, and then you know that's a design decision that you've made for your website that you're mixing together your content and your configuration and when we built out blocks, we specifically configuration to make it so that blocks could be moved across environments without binding them to content.

C

So that's intentional, but I think your suggestion is still a good one. That's you know we need a way to make some translatable block a lot easier and if it really turns out to be like maybe just embedding nodes needs to be made an easier process, and that would also solve the problem of translation and I. Think we we still I mean that issue 100, the you know make a system of URL list nodes, slash for making entity that just doesn't have a doesn't have a path on it.

C

You know something that could be used as a content block. Basically, you know those are open issues that either one that we implement might solve. This translation problem kind of as a as a side effect, so yeah I think, let's continue pursuing. You know options there, and maybe this isn't a specifically translation problem. Maybe it's just a problem of how do we make it easier to embed Tran? How do we make it easier to embed pieces of content into a layout?

C

That might be what we're working with here or finding a way to make it so that we can embed content within a layout that somehow doesn't find it so tightly to the configuration and I have no idea what that would look like, but something like I I, don't know. I saw some of the things I've seen before been like.

C

You know, let's place a view in the in the layout and the view is specifically filtered to be like one or two pieces of content, but not based on node ID based on some other selection criteria, and then that makes it so that the view is pulling out the content that you want and the view it does the query to select the content but doesn't explicitly include any IDs or anything like that. It's all dynamically figured out at the time, but I mean those are complex.

C

Solutions and I think that if we can find ways to streamline that with a lot simpler, premade solutions to solve common use cases, then that would be a great thing.

D

Yes, however, tricky I think because configuration content will overlap, we can't separate them totally, think of menus and so on so and I think also do great and they still have issues with it because yeah, it's not totally separate separate bill. Yeah.

C

It's it's really not, but the real I mean, but the benefits that we get from having config attempted to be separate and explicitly separate is still just a huge game, even though it isn't perfect, I see yeah. Just getting all of that stuff out of the database is just a huge, huge boon. Yeah.

C

All right so basically problems or working on them.

D

Yeah but I found it nice last the last two weeks because we identified a lot of issues and there was sort of focus and motivation to solve exactly these problems which, for the use case in mind from backdrop for good, so it was good decision decision to to trial. It I think yeah.

C

And this is awesome that this can pick translation. Is it moving forward? That's really exciting.

C

All right, I'm gonna, continue writing up kind of my summary that I talked about in this config translation issue just so that we can talk about it offline as well and yeah. Let's, let's keep cracking on that stuff.

C

Okay, I think that's it. Thank you. Everybody for joining great me today lots of really exciting we're capping across the spectrum the world of backdrop. So that's it. Thank you guys. All for coming and we'll see you guys next week, bye.