The following accounts have interacted with this PR and/or linked issues. I will continue to update these lists as activity occurs. You can also manually ask me to refresh this list by adding the props-bot label.

Core Committers: Use this line as a base for the props when committing in SVN:

Props jonsurrell, westonruter, peterwilsoncc, johnbillion.

To understand the WordPress project's expectations around crediting contributors, please review the Contributor Attribution page in the Core Handbook.

Test using WordPress Playground

The changes in this pull request can previewed and tested using a WordPress Playground instance.

WordPress Playground is an experimental project that creates a full WordPress instance entirely within the browser.

Some things to be aware of

• The Plugin and Theme Directories cannot be accessed within Playground.
• All changes will be lost when closing a tab with a Playground instance.
• All changes will be lost when refreshing the page.
• A fresh instance is created each time the link below is clicked.
• Every time this pull request is updated, a new ZIP file containing all changes is created. If changes are not reflected in the Playground instance,
  it's possible that the most recent build failed, or has not completed. Check the list of workflow runs to be sure.

For more details about these limitations and more, check out the Limitations page in the WordPress Playground documentation.

Test this pull request with WordPress Playground.

#10641 includes some protection to ensure bad style tag contents are not saved (including </style> or ending in something that could become </style> after concatenation).

This change should be fine because of the work in #10656. Is it worth including that additional safety here?

On one hand, this change should be perfectly safe now.

On the other hand, there's not a very good reason to allow possible style tag closers inside of style tags. It will be safely escaped by the HTML API, but I wonder whether it's more helpful to prevent it from ever being stored.

In other words, to restore the validate method but to change this:

if ( preg_match( '#</?\w+#', $css ) ) {

To be rather:

if ( false !== stripos( $css, '</style' ) ) {

Right? That seems fine to me. To be more conservative, and only make it even looser if we find this is actually needed.

The check I've proposed is a bit more complicated. It also covers text that ends in partial style close tags <, </, … </style. This should ensure that even when concatenated with other text in a style tag, it cannot produce a style close tag. I do believe that these customizer styles are concatenated with other styles by the global styles system, so this check may be worth having.

The HTML API should still protect against broken style tag contents, so I'm not sure whether this protection is worth bringing here as well.

	protected function validate_custom_css( $css ) {
		$length = strlen( $css );
		for (
			$at = strcspn( $css, '<' );
			$at < $length;
			$at += strcspn( $css, '<', ++$at )
		) {
			$remaining_strlen         = $length - $at;
			$possible_style_close_tag = 0 === substr_compare( $css, '</style', $at, min( 7, $remaining_strlen ), true );
			if ( $possible_style_close_tag ) {
				if ( $remaining_strlen < 8 ) {
					return new WP_Error(
						'rest_custom_css_illegal_markup',
						sprintf(
							/* translators: %s is the CSS that was provided. */
							__( 'The CSS must not end in "%s".' ),
							esc_html( substr( $css, $at ) )
						),
						array( 'status' => 400 )
					);
				}


				if ( 1 === strspn( $css, " \t\f\r\n/>", $at + 7, 1 ) ) {
					return new WP_Error(
						'rest_custom_css_illegal_markup',
						sprintf(
							/* translators: %s is the CSS that was provided. */
							__( 'The CSS must not contain "%s".' ),
							esc_html( substr( $css, $at, 8 ) )
						),
						array( 'status' => 400 )
					);
				}
			}
		}


		return true;
	}

I'm going to rework this based on what landed in r61486.

No, I'm not. This is ready to go, I'll land it.

Can we get test coverage added for the example CSS in the PR description?

Some users may have the customize capability but not the unfiltered_html capability.

This may not be entirely relevant here because the Custom CSS requires the edit_css capability:

And this is treated the same as unfiltered_html:

Some users may have the customize capability but not the unfiltered_html capability.

I believe this is covered by edit_css capability as Weston mentioned #10667 (comment).

For those users I think it would be good to run urls through the kses protocol check, see this section of code in the safe css portion of kses.

I'm reluctant to add this if avoidable. If there are specific concerns I'm happy to work on addressing them. Do you know whether this filter was already being run on the content?

One concern I have is that these functions are working with HTML entities, indicating they're not designed to operate on CSS or the RAWTEXT content of STYLE tags. This type of content-type processing mismatch is why this change is necessary.

Thanks everyone! I believe all feedback is addressed.