Jonathan Lahijani

As far as I can tell, that's the way to do it. Assuming your repeater field is called 'my_repeater_field' and it contains a field called 'foo' that you want to add a make-shift markup field above, here's what you'd do: $wire->addHookAfter('InputfieldWrapper::render', function(HookEvent $event) { if(!$event->object->children->count) return; if($event->object->children[0]->name!="my_repeater_field") return; $event->return = str_replace( "<li class='Inputfield InputfieldText Inputfield_foo_repeater", " <li class='Inputfield'> <label class='InputfieldHeader'> my custom repeater markup field </label> <div class='InputfieldContent'> content goes here </div> </li> <li class='Inputfield InputfieldText Inputfield_foo_repeater", $event->return ); });

@Sebi Right now, it is possible to view the OpenAPI json by going to /admin/setup/appapi/endpoints/action-get-openapi/ I want to automate building html-based documentation with Redocly CLI based on that JSON data, however because how it is currently programmed, I can't use a script to cleanly grab that JSON nor is there a method that easily gets it in that format given how the code is structured (using the executeEndpoints method doesn't get the JSON in the same way and it relies on urlSegments). Is it possible to refactor the code / create a new method that would allow getting the OpenAPI JSON directly?

Just curious, does anyone know the approach @ryan uses to minify module JavaScript files? For example, how is he minifying this: https://github.com/processwire/processwire/blob/dev/wire/modules/AdminTheme/AdminThemeUikit/scripts/main.js to this: https://github.com/processwire/processwire/blob/dev/wire/modules/AdminTheme/AdminThemeUikit/scripts/main.min.js Ideally I want to use a CLI-based tool (Linux-based), but I want to avoid using a Node-based solution (such as https://github.com/terser/terser).

If you truly want to make an admin theme from scratch like I'm doing, it's best to just take AdminThemeUikit, since that is the "official" and most supported theme and rip out UIkit and start replacing it with your own approach and just hack away at it. Keep in mind that ProcessWire makes heavy use of jQuery UI and a few other libraries so you'll have to play nicely with them unless you want to replace them too, but that takes it to another level. With Bootstrap, it's straight-forward enough given the similarities with UIkit, although this is turning out to be more work than I anticipated. But that was the point since I want it to force me into looking at how everything is interconnected. One idea for an admin theme is to do it with pure, modern CSS and as little JS as possible and as accessible as possible (good reason why here).

@Mustafa-Online I made nice update today on this module after not having touched it in several weeks. It's now basically complete but before I release it I still need to clean up some old code that's left over, make it work more nicely with SelectizeAll and provide similar overriding capabilities that the UIkit theme has. This module includes Bootstrap Icons and it substitutes Font Awesome icons accordingly (I went through each one and found the equivalents!). It also uses the Bootstrap navbar and dropdowns. A side-effect of this is ajax-loaded dropdown content won't work in the first release. I may backtrack on using Bootstrap dropdowns depending on if that becomes difficult. Anyway, it's looking good overall. If anyone is wondering why use this as opposed to UIkit, it may be beneficial if you are doing a lot of custom admin stuff and want to use pre-made Bootstrap styled components since the Bootstrap ecosystem is much much larger than UIkit. Also, it feels a little more fresh, although my actual reason for creating this was to for a way for me to get a deeper understanding of how admin themes work.

@ryan @Pete Can we get this fixed? Or if you'd like, I can help maintain this site.

I wanted to start a conversation about other systems out there that are "ProcessWire-like", meaning they also go all in on everything being a custom field, similar database architecture and other things that make ProcessWire seem unique. What makes ProcessWire unique to you that you just haven't seen elsewhere and how deep does ProcessWire go with certain decisions or features that other systems don't from your experience? One very recent system in the JS world that's similar (at least on the surface, I've never worked with it) is Payload CMS: https://payloadcms.com/

If I'm understanding this correctly, this would mean all of ProcessWire would have to be open-source, meaning the commercial modules, correct? Related.

Thanks for releasing this Ryan. Although I've been using ProcessWire daily for over a decade at this point, it's nice to compare my version of "The ProcessWire Way" with that of the creator himself and learn any tips and tricks along the way. This is probably going to be very helpful for beginners. When I was researching Tailwind a while ago, the creator (Adam Wathan) spoke about how he made many demo videos of himself replicating websites with his CSS framework so developers could get an idea of how the creator of the tool himself would approach using it. I feel like these site profiles provide a similar and more realistic learning experience in addition to ProcessWire's great documentation.

For those who want to block the vendor folder using ProcessWire's main htaccess file, here's a snippet: https://github.com/processwire/processwire-requests/issues/191#issuecomment-1969214984

With Geffen Playhouse for example, which was launched in 2019 right before custom fields for files/images was introduced, we needed custom fields for images. At the time, the best approach was to use the RepeaterImages module, which uses repeaters and all the functionality that comes with it. This includes the ability enable/disable a repeater item. In a recent update, I wanted to remove that dependency and switch to just a normal images field with custom fields, but the client still wanted enable/disable capability on images, hence my approach to it described in this post. I think about it just like repeaters. There are times when you want a piece of data to exist but not be visible. Without being able to disable an image, you would have to delete it (or do some other weird hack like perhaps add an image tag), which is less than ideal. With enable/disable capability, it brings it more in line with how multi-item fields, like repeaters, work.

As of today, this is now a native feature on the dev branch: https://github.com/processwire/processwire-requests/issues/520#issuecomment-1961794624

Lets say you have an multi-image field named 'images' and you want to be able to mark one or more images as disabled as well as make it visually appear as disabled in the admin. Currently, ProcessWire does not support this natively (requested here), however we can still easily achieve this with using custom fields for files/images, introduced in PW 3.0.142, as well as a hook to achieve the visual effect. Follow these steps (modify as needed): create a checkbox field called 'disabled' create a template called 'field-images' which will ProcessWire will detect as a custom template for the 'images' field add the 'disabled' field to that template add the following code to /site/templates/admin.php $wire->addHookAfter('InputfieldImage::renderItem', function(HookEvent $event) { if($event->object->name!='images') return; if(!$event->arguments('pagefile')->disabled) return; $event->return = "<div style='opacity:.2'>{$event->return}</div>"; }); Of course, if you don't want to display disabled images on your frontend, make sure to update your selector on your 'images' field, like so: // before (will still select disabled images) foreach($page->images as $image) { ... } // after (will ignore disabled images) foreach($page->images->find("disabled=0") as $image) { ... }

As of today on the dev branch, this is now possible with a new hookable method InputfieldPage::renderPageLabel() https://github.com/processwire/processwire-requests/issues/460

Hey there fellow Angeleno, Let's assume you have a bunch of 'video' pages with a 'file' field where the single video gets uploaded and another field called 'file_transcoded' that contains the transcoded/sized-down version. You could create an external script that bootstraps into ProcessWire, loops through each page where 'file_transcoded' doesn't exist (meaning it hasn't been acted upon yet), run ffmpeg on it, take the transcoded file and add it to file_transcoded ($page->file_transcoded->add($transcoded_file)), save and repeat for all the other non transcoded pages. That external script could be run using a cronjob. I believe you can do a background process or queue however I don't have experience with that in the context of ProcessWire yet although it's been discussed. The approach I mentioned keeps all the files in the natural /site/assets/files/(page-id)/ directory. I think the main point here is that where the original video files precisely exist is irrelevant as long as you use the $pages API which will take care of that for you.

Because I don't write JS often, hold my nose when I do, and since it's usually for little tweaks or UI adjustments, my knowledge of the syntax and weirdness of the language overall doesn't seem to stick (how do I add a class to a div again in vanilla JS?). Google is not so great because oftentimes the information is outdated on Stack Overflow results and requires a lot of sifting. However ChatGPT is phenomenal with JavaScript (and jQuery) if you want develop AND learn at the same time. Seriously a game changer for me. Unfortunately it's not so great for AutoHotkey, or at least it seems to keep giving me v1 code as opposed to v2 despite telling it to give me v2, but that's OK because it still provides some guidance which I supplement with AHKs documentation.

Indeed. Pico CSS, which is a minimal CSS framework, takes that approach: https://v2.picocss.com/docs https://v2.picocss.com/docs/accordion I wish they supported tabs however.

I would think that in ProcessWire, you would be able to do this natively: $users->get("name=bob")->logout(); Guess I'll write my own method in /site/classes/UserPage.php

Actually I just did a find for 'uk-' and replaced it with '', then replaced 'Uikit' with 'Bootstrap' and I was done! Just kidding, it's a little more involved but I got very far and it's usable now. I still need to clean up a few things. Here's a screenshot (I need to style buttons among other things): To be more clear, while I won't use it immediately, another reason why I personally am making a Bootstrap based admin theme is because it could theoretically take advantage of the many Bootstrap-based admin themes and components (examples), or at least some of the screens and layouts they provide. That would be very convenient as I do see myself making more web-app type sites where it's purely using PW's backend. @bernhardYou're right, it's not deprecated, but in maintenance only mode. I don't know the inner-workings of jQuery UI but I would assume even though it's functional, browser capabilities are now much more robust and perhaps much of its capabilities are not using JavaScript's newer, native features? Don't quote me on that, but it feels very old. I see that CraftCMS uses jQuery UI as well.

For fun, I've been working on what you would call AdminThemeBootstrap to get a deeper understanding of ProcessWire's admin framework (maybe I'll release it one day but it's more of a pet project). As you may know, the current admin system has a deep dependency on jQuery and jQuery UI (which is now deprecated in maintenance mode). Things like the navbar dropdowns, tabs and accordions are jQuery UI based components, not (for example) UIkit's dropdown or Bootstrap's dropdown. This has gotten me thinking about if there was (not that it's necessary) a comprehensive re-thinking of the admin, what libraries in today's world would be the ideal fit for a project like ProcessWire?

For Flowbite's Javascript file, I just load that file manually. For CSS, I didn't realize they have a ready-to-go CSS file as well, but that seems kind of pointless since Tailwind (similar to UIkit and Bootstrap) would require a build process for it to be customizable. I simply use Tailwind CLI directly (without NPM). But as you mentioned, given the nature of how Tailwind works whereby it will only include classes if they are used, it makes using it for a page builder more complicated unless you safelist a bunch of classes. As far as not being able to use Flowbite for end products, yea that's a downside. But I don't think there's a problem configuring a page builder / RPB with blocks for a specific client and not general use. I'm preferring it over UIkit these days because doing CSS using pure utility classes has won me over as opposed to a mix of utility classes/separation of concerns (UIkit, Bootstrap) vs. pure separation of concerns. For example, I like that I can copy a block from Flowbite without any thought of it not working or having other styles in the cascade affect it, but at the same time having full flexibility to change it easily. Tailwind also just makes more sense for designs where you have to "color outside the lines" of a CSS framework like UIkit. It's horrific to look at all those classes when using it initially, but then you get used to it.

If you dabble in Tailwind some more, I strongly recommend using Flowbite, which is still pure Tailwind but with an added JS library that makes it more batteries included like UIkit and Bootstrap. Also there are free and premium blocks and layouts which I've purchased that speeds things up considerably while maintaining the flexibility of Tailwind. To me, it's the winner of the all the various Tailwind competing user interface kits (ie compared to Preline, Tailwind UI, Daisy UI, etc.).

Hi @ryan, Thanks for this update. One easy to overlook but important missing part of the repeater update that was made is what happens if you edit a repeater item directly. Assume we have a repeater field named 'books' and you edit the repeater item page by going to /admin/repeaters/books/(for-page)/repeater-item-page When editing it directly, the page to be edited simply gets the RepeaterPage class instead of BooksRepeaterPage class. I managed to hack-fix this by doing this in /wire/core/Templates.php: // determine if custom class available (3.0.152+) if($usePageClasses) { // generate a CamelCase name + 'Page' from template name, i.e. 'blog-post' => 'BlogPostPage' $className = ucwords(str_replace(array('-', '_', '.'), ' ', $template->name)); // *** *** hack fix: if we are editing a page with a template that starts with 'repeater_', assign the correct page class! *** *** if(str_starts_with($template->name, 'repeater_')) { $className = __NAMESPACE__ . "\\" . str_replace(' ', '', $className) . 'RepeaterPage'; $className = str_replace('ProcessWire\Repeater', 'ProcessWire\\', $className); } else { $className = __NAMESPACE__ . "\\" . str_replace(' ', '', $className) . 'Page'; } if(class_exists($className) && wireInstanceOf($className, $corePageClass)) { $pageClass = $className; } } Can what I described be supported as well?

This is a known issue. I actually made a big report in the PW issues repository on GitHub.

After over 10 years, I didn't know you can simply give a page's name when assigning a value to a page field. I discovered this by accident. I used to do this: $page->setAndSave('order_status', $pages->get('/path/to/order-statuses/pending/')); But I realized you can simply do this in ProcessWire making it much easier on the eyes, with the added benefit of not having to update the code if the path to the page changes: $page->setAndSave('order_status', 'pending'); My mind is blown. It makes me wonder what other little things I haven't realized yet.