Style engine: enqueue block supports styles in Gutenberg #42880

ramonjd · 2022-08-02T05:44:46Z

What?

A follow up to:

Style engine: enqueue block support styles #42452

What?

Let's enqueue stored block supports styles in Gutenberg, rather than in the style engine.

At the same time, we'll combine block style and layout styles into one style tag.

Why?

This PR follows up on a discussion in #42452, the substance of which is that it's probably better to give the consumer control over how styles are rendered via hooks / enqueues etc

How?

Adding a new function in script-loader.php - gutenberg_enqueue_stored_block_supports_styles() (name to be debated) that grabs block supports styles from the style engine store and enqueues the stylesheet.

Testing Instructions

Test in both classic and block themes.

Check that elements and layouts styles are enqueued.

Create a new post and insert some text links and group blocks (or any layout styles)

Here's some test HTML:

Example

<!-- wp:group {"style":{"color":{"background":"#ee6b6b"},"elements":{"link":{"color":{"text":"var:preset|color|vivid-cyan-blue"}}}},"layout":{"type":"flex","flexWrap":"nowrap","justifyContent":"space-between","orientation":"vertical"}} -->
<div class="wp-block-group has-background has-link-color" style="background-color:#ee6b6b"><!-- wp:paragraph {"style":{"elements":{"link":{"color":{"text":"var:preset|color|white"}}}}} -->
<p class="has-link-color"><a href="https://word">Test</a></p>
<!-- /wp:paragraph -->

<!-- wp:paragraph {"style":{"elements":{"link":{"color":{"text":"#eaf1f6"}}}}} -->
<p class="has-link-color"><a href="https://word">Test</a></p>
<!-- /wp:paragraph --></div>
<!-- /wp:group -->

<!-- wp:group {"style":{"color":{"background":"#c4cce6"}},"layout":{"contentSize":"342px"}} -->
<div class="wp-block-group has-background" style="background-color:#c4cce6"><!-- wp:paragraph {"style":{"elements":{"link":{"color":{"text":"var:preset|color|vivid-red"}}}}} -->
<p class="has-link-color"><a href="https://word">Test</a></p>
<!-- /wp:paragraph --></div>
<!-- /wp:group -->

<!-- wp:group {"style":{"color":{"background":"#9eebbc"}},"layout":{"contentSize":"342px"}} -->
<div class="wp-block-group has-background" style="background-color:#9eebbc"><!-- wp:paragraph {"style":{"elements":{"link":{"color":{"text":"#cd31d6"}}}}} -->
<p class="has-link-color"><a href="https://word">Test</a></p>
<!-- /wp:paragraph --></div>
<!-- /wp:group -->

Expected CSS output. For classic themes this style tag should appear at the bottom of the page. For block themes, in the head.

<style id='block-supports-styles-inline-css'>
.wp-elements-bfe89bcbd3fbf76f9bf93ac884528d9a a {color: var(--wp--preset--color--vivid-cyan-blue);}.wp-elements-85694c8ee6aa1fb95efca209b615f0b0 a {color: var(--wp--preset--color--white);}.wp-elements-6ac1c4acebd6a433ed64dc4e6b0eb185 a {color: #eaf1f6;}.wp-elements-469734e762a60395c96370480f3c9414 a {color: var(--wp--preset--color--vivid-red);}.wp-elements-9307c4f13980ac2b051b4fcf5f7b28fa a {color: #cd31d6;}.wp-block-group.wp-container-1 {flex-wrap: nowrap; align-items: flex-start;}.wp-block-post-content.wp-container-4 > :where(:not(.alignleft):not(.alignright):not(.alignfull)) {max-width: 840px; margin-left: auto !important; margin-right: auto !important;}.wp-block-post-content.wp-container-4 > .alignwide {max-width: 1100px;}.wp-block-group.wp-container-2 > :where(:not(.alignleft):not(.alignright):not(.alignfull)),.wp-block-group.wp-container-3 > :where(:not(.alignleft):not(.alignright):not(.alignfull)) {max-width: 342px; margin-left: auto !important; margin-right: auto !important;}.wp-block-group.wp-container-2 > .alignwide,.wp-block-group.wp-container-3 > .alignwide {max-width: 342px;}.wp-block-group.wp-container-2 .alignfull,.wp-block-group.wp-container-3 .alignfull,.wp-block-post-content.wp-container-4 .alignfull {max-width: none;}
</style>

lib/compat/wordpress-6.1/script-loader.php

aristath · 2022-08-02T07:28:21Z

lib/compat/wordpress-6.1/script-loader.php

+/**
+ * Fetches, processes and compiles stored block supports styles, then renders them to the page.
+ * Styles are stored via the style engine API. See: packages/style-engine/README.md
+ */
+function gutenberg_enqueue_stored_block_supports_styles() {
+	$styles  = gutenberg_style_engine_get_stylesheet_from_store( 'block-supports' );
+	$styles .= gutenberg_style_engine_get_stylesheet_from_store( 'layout-block-supports' );
+	if ( ! empty( $styles ) ) {
+		$key = 'block-supports-styles';
+		wp_register_style( $key, false, array(), true, true );
+		wp_add_inline_style( $key, $styles );
+		wp_enqueue_style( $key );
+	}
+}
+


I think it would be better if we abstract this function a bit... Instead of having a function specifically for the block-supports, we could have it for all stores. In another PR I'm converting global-styles to use the style-engine, so if we make this function specifically for block-supports, then we'd have to add a separate function for global-styles, then another one for theme-styles and any other implementation we may come up with in the future.

This works fine:

Suggested change

/**

* Fetches, processes and compiles stored block supports styles, then renders them to the page.

* Styles are stored via the style engine API. See: packages/style-engine/README.md

*/

function gutenberg_enqueue_stored_block_supports_styles() {

$styles = gutenberg_style_engine_get_stylesheet_from_store( 'block-supports' );

$styles .= gutenberg_style_engine_get_stylesheet_from_store( 'layout-block-supports' );

if ( ! empty( $styles ) ) {

$key = 'block-supports-styles';

wp_register_style( $key, false, array(), true, true );

wp_add_inline_style( $key, $styles );

wp_enqueue_style( $key );

}

}

/**

* Fetches, processes and compiles stored styles, then renders them to the page.

* Styles are stored via the style engine API. See: packages/style-engine/README.md

*/

function gutenberg_enqueue_stored_styles() {

$stores = WP_Style_Engine_CSS_Rules_Store_Gutenberg::get_stores();

foreach ( array_keys( $stores ) as $store_name ) {

$styles = gutenberg_style_engine_get_stylesheet_from_store( $store_name );

if ( ! empty( $styles ) ) {

$key = "wp-style-engine-$store_name";

wp_register_style( $key, false, array(), true, true );

wp_add_inline_style( $key, $styles );

wp_enqueue_style( $key );

}

}

}

Fair call on the name change. Much better.

I was also wondering about how we'd cater for global styles and make it backwards compatible.

I left it out of the equation so we could decide later how to handle them. Mainly because of what's going on in gutenberg_enqueue_global_styles() and the ever-changing nature of global styles.

Also, this would print out every stored store. Is that the intention?

Would there be any issue against being explicit about which styles Gutenberg prints, i.e., adding them as we need them?

Aside from the certainty/control aspect, we'll need to determine the order eventually, e.g., we definitely want block supports styles to be printed after global styles.

I can imagine it would be handy for theme authors to be able to queue styles via the style engine, but I'm unaware if it's something that they should control themselves (?)

Oh, great discussion here, I think I agree with both of you 😄 — it'd be good to come up with a combined function that works for those future use cases, rather than needing it to be tied specifically to block supports 👍

Would there be any issue against being explicit about which styles Gutenberg prints, i.e., adding them as we need them?

I agree that it'd be better to be explicit about which particular stores we're enqueueing, so that we can both determine the order, and have the function be clear about what's being enqueued (which should help with debugging).

Ideally, if something somewhere (a plugin or theme) adds another store, then maybe it shouldn't be automatically enqueued. Enqueueing styles should probably still be the responsibility of the theme to do explicitly, rather than happening automatically?

If we want the function to help themes or other use cases with the task of enqueueing, perhaps one way to do it might be to have a param that defaults to default or something like that. If it's default then we enqueue the Gutenberg stores, and if it isn't default then it looks up the store name to be enqueued? 🤷

Also, this might just be a personal preference, but it can be useful to have explicit strings for style keys where we can, so that if you're inspecting a rendered page and need to look up where that style gets generated, you can search the codebase for the string in order to find it. Not very DRY, but can sometimes help with debugging.

it'd be good to come up with a combined function that works for those future use cases, rather than needing it to be tied specifically to block supports

Yeah, it'd be great to be able satisfy both causes:

having control over the stores that Gutenberg requires to be enqueued

giving theme authors the opportunity to opt-in to this behaviour

A naive version might be to filter out and enqueue Gutenberg's stores in gutenberg_enqueue_block_support_styles(), then loop over the remaining (presumably added elsewhere by themes or whatever).

It'd be better to explicitly define "equeuability" so maybe we could utilize the context and enqueue parameters so that we store all styles that specify a context, but only enqueue them if enqueue === true 🤔

The consequence is that we'd have to filter out rules from the store that shouldn't be enqueued. 🤔 🤔🤔🤔🤔🤔

return gutenberg_style_engine_get_stylesheet( $layout_styles, array( 'context' => 'layout-block-supports', // It will be stored in the 'layout-block-supports' store. 'enqueue' => true, // It will be enqueued. ) );

A naive version might be to filter out and enqueue Gutenberg's stores in gutenberg_enqueue_block_support_styles(), then loop over the remaining (presumably added elsewhere by themes or whatever).

I mean something like this:

/** * Fetches, processes and compiles stored core styles, then combines and renders them to the page. * Any other registered stores are enqueued separately after the core styles. * Styles are stored via the style engine API. See: packages/style-engine/README.md */ function gutenberg_enqueue_stored_styles() { $core_styles_keys = array( 'block-supports', 'layout-block-supports' ); $compiled_stylesheet = ''; foreach ( $core_styles_keys as $style_key ) { $compiled_stylesheet .= gutenberg_style_engine_get_stylesheet_from_store( $style_key ); } // Combine Core styles. // @TODO check for `defined( 'SCRIPT_DEBUG' ) && SCRIPT_DEBUG` and split them for ease of debugging. if ( ! empty( $styles ) ) { $key = 'block-supports-styles'; wp_register_style( $key, false, array(), true, true ); wp_add_inline_style( $key, $compiled_stylesheet ); wp_enqueue_style( $key ); } $additional_stores = WP_Style_Engine_CSS_Rules_Store_Gutenberg::get_stores(); foreach ( array_keys( $additional_stores ) as $store_name ) { if ( in_array( $store_name, $core_styles_keys, true ) ) { continue; } $styles = gutenberg_style_engine_get_stylesheet_from_store( $store_name ); if ( ! empty( $styles ) ) { $key = "wp-style-engine-$store_name"; wp_register_style( $key, false, array(), true, true ); wp_add_inline_style( $key, $styles ); wp_enqueue_style( $key ); } } }

Also, this would print out every stored store. Is that the intention?

Yes

Would there be any issue against being explicit about which styles Gutenberg prints, i.e., adding them as we need them?

No issue with that, but I don't think that having a separate function per-store would be the right way to approach this 😅

Aside from the certainty/control aspect, we'll need to determine the order eventually, e.g., we definitely want block supports styles to be printed after global styles.

Well... not necessarily! I agree with the "we'll need to determine the order eventually" part of the sentence, but I don't believe they should actually be enqueued separately!
Ideally, block supports and global styles would be in the same store (let's say a core store).
Block supports would be added to the store after global styles, so any CSS that needs to override global-styles will be merged and override things properly. Having these 2 in the same store would allow for further optimization of the resulting CSS by combining things where appropriate 😉

Maybe instead of a priority on enqueue it would be more appropriate to have a priority at the time we add styles to the store... This way we'd be able to sort the array based on priorities, and merge styles properly, ensuring overrides are correctly applied?

I can imagine it would be handy for theme authors to be able to queue styles via the style engine, but I'm unaware if it's something that they should control themselves (?)

They definitely should be able to! I can think of many scenarios where theme-authors would be able to take advantage of it and append their styles to the Core styles, or use the style-engine to add block-specific styles! It would certainly be a lot more efficient (for performance and sustainability) if they added their opinionated block-styles this way instead of enqueueing monolithic stylesheets.
Additionally, it's not just themes that can use it... Think block.json and styles we can add in there - both in Core blocks and 3rd-party blocks.

Ideally, if something somewhere (a plugin or theme) adds another store, then maybe it shouldn't be automatically enqueued. Enqueueing styles should probably still be the responsibility of the theme to do explicitly, rather than happening automatically?

Hmmm what if we add a prop to stores that can define if the store should be automatically enqueued or not?

Also, this might just be a personal preference, but it can be useful to have explicit strings for style keys where we can, so that if you're inspecting a rendered page and need to look up where that style gets generated, you can search the codebase for the string in order to find it. Not very DRY, but can sometimes help with debugging.

Intersting. I suppose we could do something like that if SCRIPT_DEBUG is set to true... but I don't think it should be the default behavior. The default output should always be optimized for consumers, not developers 😉

A naive version might be to filter out and enqueue Gutenberg's stores in gutenberg_enqueue_block_support_styles(), then loop over the remaining (presumably added elsewhere by themes or whatever).

Seems reasonable!

No issue with that, but I don't think that having a separate function per-store would be the right way to approach this

Totally agree 👍

I agree with the "we'll need to determine the order eventually" part of the sentence, but I don't believe they should actually be enqueued separately!

I think we're on the same page. I was only referring to having the ability, however we do it, to respect the styles hierarchy and preserve the cascade.

So I mean Global styles -> Elements -> Block styles -> Block supports.

Ideally, block supports and global styles would be in the same store (let's say a core store).

This makes sense! I like!

Block supports would be added to the store after global styles, so any CSS that needs to override global-styles will be merged and override things properly. Having these 2 in the same store would allow for further optimization of the resulting CSS by combining things where appropriate

Oooh now I see what you mean. Sweet! Thanks for the explainer. 🍺

I'm off for a couple of days but will finish this off and do some more testing on Friday.

Thanks for all the great feedback!

A naive version might be to filter out and enqueue Gutenberg's stores in gutenberg_enqueue_block_support_styles(), then loop over the remaining (presumably added elsewhere by themes or whatever).

I've done this for now and have added test coverage for it.

Happy to refine things in follow ups!

aristath · 2022-08-02T07:32:28Z

This looks nice!

I pushed a small commit, removing the constructor and the get_instance() method from the main class. All methods in that class are static now, so we don't need to get any instances etc, and the constructor no longer makes any sense if we remove the get_instance() method.

I only have one thing that I would like to see changed in this PR, and I left a comment above 👍

packages/style-engine/class-wp-style-engine.php

ramonjd · 2022-08-05T01:14:02Z

packages/style-engine/class-wp-style-engine.php

@@ -646,10 +570,10 @@ function wp_style_engine_get_styles( $block_styles, $options = array() ) {
 	}

 	if ( ! empty( $parsed_styles['declarations'] ) ) {
-		$styles_output['css']          = $style_engine::compile_css( $parsed_styles['declarations'], $options['selector'] );
+		$styles_output['css']          = WP_Style_Engine::compile_css( $parsed_styles['declarations'], $options['selector'] );
 		$styles_output['declarations'] = $parsed_styles['declarations'];
 		if ( true === $options['enqueue'] ) {


enqueue doesn't really work as an option anymore since all the style engine is doing is "storing"

Maybe we can default the "context" to null and store wherever a non-empty context string value is passed (?)

Then we could remove enqueue altogether for now.

I'll do this in a follow up if folks think its legit.

I'll do this in a follow up if folks think its legit.

Sounds legit for a follow-up to me! Makes for a slightly simpler call to store styles, and means that the calling code doesn't need to think about what the style engine is doing.

andrewserong

This is testing really nicely for me @ramonjd! Thanks for all the follow-ups. Just left a few nits — the one item that we should probably address before merge is the Gallery block's usage of the gutenberg_enqueue_block_support_styles function. I think the choices are probably between updating the call in that block, or deferring introducing the _deprecated_function call to a follow-up, once we've refactored the usage.

What do you think?

andrewserong · 2022-08-08T05:52:16Z

packages/style-engine/class-wp-style-engine.php

@@ -646,10 +570,10 @@ function wp_style_engine_get_styles( $block_styles, $options = array() ) {
 	}

 	if ( ! empty( $parsed_styles['declarations'] ) ) {
-		$styles_output['css']          = $style_engine::compile_css( $parsed_styles['declarations'], $options['selector'] );
+		$styles_output['css']          = WP_Style_Engine::compile_css( $parsed_styles['declarations'], $options['selector'] );
 		$styles_output['declarations'] = $parsed_styles['declarations'];
 		if ( true === $options['enqueue'] ) {


I'll do this in a follow up if folks think its legit.

Sounds legit for a follow-up to me! Makes for a slightly simpler call to store styles, and means that the calling code doesn't need to think about what the style engine is doing.

andrewserong · 2022-08-08T06:11:53Z

lib/compat/wordpress-6.1/script-loader.php

+			$compiled_core_stylesheet .= "/**\n * Core styles: $style_key\n */\n";
+		}
+		// Chain core store ids to signify what the styles contain.
+		$style_tag_id             .= $style_tag_id ? '-' . $style_key : $style_key;


Since $style_tag_id is defined as 'core' by default it looks like this ternary will always be true?

Indeed! T'was an artefact from something... somewhere.

andrewserong · 2022-08-08T06:13:06Z

lib/compat/wordpress-6.1/script-loader.php

+	foreach ( $core_styles_keys as $style_key ) {
+		// Add comment to identify core styles sections in debugging.
+		if ( defined( 'SCRIPT_DEBUG' ) && SCRIPT_DEBUG ) {
+			$compiled_core_stylesheet .= "/**\n * Core styles: $style_key\n */\n";


Great idea to aid debugging!

packages/style-engine/README.md

andrewserong · 2022-08-08T07:02:07Z

lib/compat/wordpress-6.1/script-loader.php

 * @param string $style String containing the CSS styles to be added.
 * @param int    $priority To set the priority for the add_action.
 */
 function gutenberg_enqueue_block_support_styles( $style, $priority = 10 ) {
+	_deprecated_function( __FUNCTION__, '6.1' );


I notice that with this PR applied, this is being logged once per page request:

It looks like it's being called from the Gallery block. I imagine we should update that call, too, otherwise we'll potentially risk flooding logs of sites running the GB plugin?

Alternately, we could leave the _deprected_function call to a follow-up if you'd prefer.

Will do, thanks!

deferring introducing the _deprecated_function call to a follow-u

Actually, I think this you had the best idea for now. We can deprecate when we deal with the final Gutenberg usage in the gallery.

Just in case the gallery gets hairy 😄

I've reverted this change.

ramonjd · 2022-08-08T07:33:45Z

the one item that we should probably address before merge is the Gallery block's usage of the gutenberg_enqueue_block_support_styles function

Good call. I had that buried deep on my list of TODOs. Thanks for reminding me. I'll get a PR up for that ASAP. 🙇

andrewserong

Thanks for the follow-up @ramonjd — this is testing well for me manually. I just noticed that I think the output is being processed twice, I've left a comment around the add_action lines. I wasn't too sure which way would be best for handling it, but happy to do more testing / exploration!

lib/compat/wordpress-6.1/script-loader.php

andrewserong · 2022-08-09T03:02:47Z

lib/compat/wordpress-6.1/script-loader.php

+ * Fetches, processes and compiles stored core styles, then combines and renders them to the page.
+ * Styles are stored via the style engine API. See: packages/style-engine/README.md
+ */
+function gutenberg_enqueue_stored_styles() {


Just checking, the idea with using a generic name like wp_enqueue_stored_styles rather than being tied to the style engine, is that we're not exposing any style engine behaviour here? I think that sounds good to me, in that this is about enqueuing stored styles in general, (and deals with essential styles for WordPress) rather than being specific to the style engine.

Another tricky naming issue 😄

"Stored" indirectly implies the style engine, but yeah the motivation was to expand this function later to apply to all stores. See comment above.

If that's true, it will probably require a strategy to subsume/integrate gutenberg_enqueue_global_styles() when we get to global styles. ?

is that we're not exposing any style engine behaviour here?

Only pulling from style engine stores. Do you think we should be more specific in the naming? 🤔

Do you think we should be more specific in the naming? 🤔

Naming's hard! For anything that's specifically style engine related (rather than the style engine being an implementation detail), I like the idea of including style_engine in the public function name so that it's all grouped together. So I'd probably lean toward wp_style_engine_enqueue_stored_styles as the function name, but I don't feel at all strongly about it if you prefer the shorter name 🙂

This sounds infinitely reasonable.

I don't really have a preference either 🙉

I suppose a weak argument to keep it generic is that wp_style_engine_enqueue_stored_styles might communicate to some that it's a public function of the style engine package.

I suppose a weak argument to keep it generic is that wp_style_engine_enqueue_stored_styles might communicate to some that it's a public function of the style engine package.

Good point! Since neither of us feels strongly about it, I think that nudges us more toward keeping the shorter title — after all, it's a function defined outside of the style engine package.

Updating README.md

Hard code a list of core styles that we want to fetch from the store and the order in which we fetch them.

For classic themes, in the footer.

andrewserong

Thanks for all the follow-up @ramonjd! This is testing nicely for me now, with the callback only firing once. I smoke tested that the output still appears to be correct, and is only being enqueued once in wp_enqueue_scripts in blocks-based themes, and once in wp_footer in Classic themes.

This LGTM, and we can continue to refine and adjust the logic within gutenberg_enqueue_stored_styles in follow-ups as we explore global styles, etc, I reckon.

Nice work! 🎉

packages/style-engine/README.md

Co-authored-by: Sören Wrede <soerenwrede@gmail.com>

oandregal · 2022-08-10T08:33:18Z

For the changelog, as this is a follow-up to #42452 (marked as experimental) I've labeled this as experimental as well.

ramonjd · 2022-08-10T21:48:50Z

For the changelog, as this is a follow-up to #42452 (marked as experimental) I've labeled this as experimental as well.

@oandregal Is there anything you need me to do here or is it still automated?

oandregal · 2022-08-11T07:17:02Z

Nah, everything is fine. I was just editing titles and labels and wanted to tell people why I did.

enodekciw · 2022-08-19T21:32:42Z

Is there a specific reason why this is being enqueued in wp_footer when using classic theme?

The website I'm currently working on doesn't use a FSE theme, but uses more like a hybrid approach - header and footer are loaded from header.php and footer.php but everything else is managed with Gutenberg. I suppose this is considered "a classic theme", right?

Let's say I build a hero section now:

Group block with specific layout sizes
Inside of that Group block there's a center aligned Heading block
center aligned Buttons block comes after the Heading block

Now what happens is that all of the layout / alignment CSS from Gutenberg is enqueued in footer and this creates a huge flash of unstyled content issue. Also, Core Web Vitals (especially, CLS) isn't happy. Basically, I have to load custom CSS in the head section just to get rid of this, atleast for a hero section.

Maybe it would be a good idea to atleast have a filter to be able to enqueue this in the head section instead of the footer for classic (more like hybrid) themes?

ramonjd · 2022-08-21T22:31:26Z

Is there a specific reason why this is being enqueued in wp_footer when using classic theme?

This PR doesn't change the functionality that's been in place all this time.

I started looking in the Git history for a specific reason. It goes way back and I believe it's for backwards compatibility with Core. This PR provides some clues:

The dynamic block styles for layout and elements should be loaded in the for block themes. While that should also be the case for classic themes, the current mechanism we use (render_block) does not allow us to do that

If you manage to go deeper and find out more it'd be great to know.

Maybe it would be a good idea to at least have a filter to be able to enqueue this in the head section instead of the footer for classic (more like hybrid) themes?

This is an interesting suggestion, thanks. Adding filters is definitely something that folks could look at, and I know, once the style engine is in Core and is stable, there'll be some in the style engine package at least.

As for a filter for enqueuing itself, it could be worth a try so long as we take into account the historical compatibility issues during testing.

aristath · 2022-08-22T09:47:44Z

Regarding loading in header vs footer in classic themes vs block themes: it all comes down to the timing and when we parse blocks.
In block themes, we get the contents of a page and render the blocks before wp_head when rendering a page. In classic themes however, rendering happens differently and when wp_head runs, we don't actually know what blocks exist on the whole page. Blocks get rendered as we discover them in the content and the only viable option is to print their styles in the footer because the head has already been rendered.
So the main difference is that block themes get the whole template/content/etc before the head, while classic themes have "progressive" rendering.

ramonjd · 2022-08-22T22:24:01Z

Thanks for the explanation @aristath!! 🙇

It sounds like something I could add to the docs or in a code comment for posterity.

Edit: oh, there's a note related this already here:

gutenberg/lib/compat/wordpress-6.1/script-loader.php

Line 16 in 090171d

* For classic ones, it's loaded in the body

…sts to 6.1. This changeset backports the following changes: - Implement [WordPress/gutenberg#42880 gutenberg#42880]: Backport script loader: enqueue stored block supports styles - Allow a way to bypass `SCRIPT_DEBUG` in tests. See [WordPress/wordpress-develop#3259 (comment) comment] and the related [WordPress/gutenberg#44248 Gutenberg pull request] Props ramonopoly, gziolo, bernhard-reiter, audrasjb, costdev. See #56467. Built from https://develop.svn.wordpress.org/trunk@54214 git-svn-id: http://core.svn.wordpress.org/trunk@53773 1a063a9b-81f0-0310-95a4-ce76da25c4cd

…sts to 6.1. This changeset backports the following changes: - Implement [WordPress/gutenberg#42880 gutenberg#42880]: Backport script loader: enqueue stored block supports styles - Allow a way to bypass `SCRIPT_DEBUG` in tests. See [WordPress/wordpress-develop#3259 (comment) comment] and the related [WordPress/gutenberg#44248 Gutenberg pull request] Props ramonopoly, gziolo, bernhard-reiter, audrasjb, costdev. See #56467. Built from https://develop.svn.wordpress.org/trunk@54214 git-svn-id: https://core.svn.wordpress.org/trunk@53773 1a063a9b-81f0-0310-95a4-ce76da25c4cd

…sts to 6.1. This changeset backports the following changes: - Implement [WordPress/gutenberg#42880 gutenberg#42880]: Backport script loader: enqueue stored block supports styles - Allow a way to bypass `SCRIPT_DEBUG` in tests. See [WordPress#3259 (comment) comment] and the related [WordPress/gutenberg#44248 Gutenberg pull request] Props ramonopoly, gziolo, bernhard-reiter, audrasjb, costdev. See #56467. git-svn-id: https://develop.svn.wordpress.org/trunk@54214 602fd350-edb4-49c9-b593-d223f7449a82

ramonjd requested a review from spacedmonkey as a code owner August 2, 2022 05:44

ramonjd self-assigned this Aug 2, 2022

ramonjd added [Package] Style Engine /packages/style-engine [Type] New API New API to be used by plugin developers or package users. [Status] In Progress Tracking issues with work in progress labels Aug 2, 2022

ramonjd mentioned this pull request Aug 2, 2022

Style engine: enqueue block support styles #42452

Merged

ramonjd requested review from andrewserong and aristath August 2, 2022 05:48

ramonjd commented Aug 2, 2022

View reviewed changes

lib/compat/wordpress-6.1/script-loader.php Outdated Show resolved Hide resolved

aristath reviewed Aug 2, 2022

View reviewed changes

andrewserong reviewed Aug 3, 2022

View reviewed changes

packages/style-engine/class-wp-style-engine.php Show resolved Hide resolved

ramonjd force-pushed the update/style-engine-enqueue-block-supports-scripts-in-gutenberg branch 3 times, most recently from 8917291 to ba6f124 Compare August 3, 2022 04:11

ramonjd mentioned this pull request Aug 3, 2022

Tracking: Add a Style Engine to manage rendering block styles #38167

Closed

25 tasks

ramonjd force-pushed the update/style-engine-enqueue-block-supports-scripts-in-gutenberg branch 2 times, most recently from 2e2abce to 8d227f6 Compare August 5, 2022 00:01

ramonjd commented Aug 5, 2022

View reviewed changes

ramonjd added Needs User Documentation Needs new user documentation Needs Dev Note Requires a developer note for a major WordPress release cycle labels Aug 5, 2022

ramonjd force-pushed the update/style-engine-enqueue-block-supports-scripts-in-gutenberg branch 2 times, most recently from 0a41d80 to ed847e3 Compare August 8, 2022 00:54

andrewserong reviewed Aug 8, 2022

View reviewed changes

andrewserong reviewed Aug 9, 2022

View reviewed changes

ramonjd and others added 4 commits August 9, 2022 14:31

Enqueuing block supports styles in gutenberg

5e782cf

Updating README.md

get_instance() and constructor no longer needed

6196b91

remove $instance declaration

d6e3552

Rename hook callback to gutenberg_enqueue_stored_styles

255c66d

Hard code a list of core styles that we want to fetch from the store and the order in which we fetch them.

For block themes, print stored styles in the header.

13785f1

For classic themes, in the footer.

ramonjd force-pushed the update/style-engine-enqueue-block-supports-scripts-in-gutenberg branch from 1de71a1 to 13785f1 Compare August 9, 2022 04:39

andrewserong approved these changes Aug 9, 2022

View reviewed changes

Soean reviewed Aug 9, 2022

View reviewed changes

packages/style-engine/README.md Outdated Show resolved Hide resolved

Update packages/style-engine/README.md

93b49bf

Co-authored-by: Sören Wrede <soerenwrede@gmail.com>

ramonjd merged commit 090171d into trunk Aug 9, 2022

ramonjd deleted the update/style-engine-enqueue-block-supports-scripts-in-gutenberg branch August 9, 2022 22:16

github-actions bot added this to the Gutenberg 13.9 milestone Aug 9, 2022

ramonjd mentioned this pull request Aug 9, 2022

Style engine: remove enqueue flag #43103

Merged

oandregal added [Type] Experimental Experimental feature or API. and removed [Type] New API New API to be used by plugin developers or package users. [Status] In Progress Tracking issues with work in progress labels Aug 10, 2022

ryelle mentioned this pull request Aug 18, 2022

Theme: Fix layout style bleed from previews into main theme WordPress/pattern-directory#514

Merged

ockham mentioned this pull request Aug 26, 2022

Plugin: Backport PHP changes for WordPress 6.1 release #43440

Closed

37 tasks

ramonjd mentioned this pull request Sep 9, 2022

Backport script loader: enqueue stored block supports styles WordPress/wordpress-develop#3218

Closed

ockham mentioned this pull request Sep 15, 2022

Backport script loader: enqueue stored block supports styles (refresh) WordPress/wordpress-develop#3259

Closed

andrewserong mentioned this pull request Sep 16, 2022

(Alternative) Backport script loader: enqueue stored block supports styles (3rd attempt) WordPress/wordpress-develop#3262

Closed

This was referenced Sep 19, 2022

Add/enqueue stored block supports styles ramonjd/wordpress-develop#1

Closed

Backporting block supports filter callback and registrations and tests WordPress/wordpress-develop#3273

Closed

ramonjd removed the Needs Dev Note Requires a developer note for a major WordPress release cycle label Jun 30, 2023

Style engine: enqueue block supports styles in Gutenberg #42880

Style engine: enqueue block supports styles in Gutenberg #42880

Conversation

ramonjd commented Aug 2, 2022 • edited Loading

What?

What?

Why?

How?

Testing Instructions

aristath Aug 2, 2022 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

aristath commented Aug 2, 2022

Choose a reason for hiding this comment

Choose a reason for hiding this comment

andrewserong left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

ramonjd commented Aug 8, 2022

andrewserong left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

andrewserong left a comment

Choose a reason for hiding this comment

oandregal commented Aug 10, 2022

ramonjd commented Aug 10, 2022

oandregal commented Aug 11, 2022

enodekciw commented Aug 19, 2022

ramonjd commented Aug 21, 2022

aristath commented Aug 22, 2022

ramonjd commented Aug 22, 2022 • edited Loading

ramonjd commented Aug 2, 2022 •

edited

Loading

aristath Aug 2, 2022 •

edited

Loading

ramonjd commented Aug 22, 2022 •

edited

Loading