Paste: Remove HTML Formatting Space

packages/blocks/src/api/raw-handling/html-formatting-remover.js

@@ -0,0 +1,92 @@
+/**
+ * Internal dependencies
+ */
+import { isPhrasingContent } from './phrasing-content';
+
+function getSibling( node, which ) {
+	const sibling = node[ `${ which }Sibling` ];
+
+	if ( sibling && isPhrasingContent( sibling ) ) {
+		return sibling;
+	}
+
+	const { parentNode } = node;
+
+	if ( ! parentNode || ! isPhrasingContent( parentNode ) ) {
+		return;
+	}
+
+	return getSibling( parentNode, which );
+}
+
+function isFormattingSpace( character ) {
+	return (
+		character === ' ' ||
+		character === '\r' ||
+		character === '\n' ||
+		character === '\t'
+	);
+}
+
+/**
+ * Removes spacing that formats HTML.
+ *
+ * @see https://www.w3.org/TR/css-text-3/#white-space-processing
+ *
+ * @param {Node} node The node to be processed.
+ * @return {void}
+ */
+export default function( node ) {
+	if ( node.nodeType !== node.TEXT_NODE ) {
+		return;
+	}
+
+	// Ignore pre content.
+	if ( node.parentElement.closest( 'pre' ) ) {
+		return;
+	}
+
+	// First, replace any sequence of HTML formatting space with a single space.
+	let newData = node.data.replace( /[ \r\n\t]+/g, ' ' );
+
+	// Remove the leading space if the text element is at the start of a block,
+	// is preceded by a line break element, or has a space in the previous
+	// node.
+	if ( newData[ 0 ] === ' ' ) {
+		const previousSibling = getSibling( node, 'previous' );
+
+		if (
+			! previousSibling ||
+			previousSibling.nodeName === 'BR' ||
+			previousSibling.textContent.slice( -1 ) === ' '
+		) {
+			newData = newData.slice( 1 );
+		}
+	}
+
+	// Remove the trailing space if the text element is at the end of a block,
+	// is succeded by a line break element, or has a space in the next text
+	// node.
+	if ( newData[ newData.length - 1 ] === ' ' ) {
+		const nextSibling = getSibling( node, 'next' );
+
+		if (
+			! nextSibling ||
+			nextSibling.nodeName === 'BR' ||
+			(
+				nextSibling.nodeType === nextSibling.TEXT_NODE &&
+				isFormattingSpace( nextSibling.textContent[ 0 ] )
+			)
+		) {
+			newData = newData.slice( 0, -1 );
+		}
+	}
+
+	// If there's no data left, remove the node, so `previousSibling` stays
+	// accurate. Otherwise, update the node data.
+	if ( ! newData ) {
+		node.parentNode.removeChild( node );
+	} else {
+		node.data = newData;
+	}
+}

packages/blocks/src/api/raw-handling/paste-handler.js

@@ -24,6 +24,7 @@ import shortcodeConverter from './shortcode-converter';
 import markdownConverter from './markdown-converter';
 import iframeRemover from './iframe-remover';
 import googleDocsUIDRemover from './google-docs-uid-remover';
+import htmlFormattingRemover from './html-formatting-remover';
 import { getPhrasingContentSchema } from './phrasing-content';
 import {
 	deepFilterHTML,
@@ -224,6 +225,7 @@ export function pasteHandler( { HTML = '', plainText = '', mode = 'AUTO', tagNam
 
 		piece = deepFilterHTML( piece, filters, blockContentSchema );
 		piece = removeInvalidHTML( piece, schema );
+		piece = deepFilterHTML( piece, [ htmlFormattingRemover ], blockContentSchema );
 		piece = normaliseBlocks( piece );
 
 		// Allows us to ask for this information when we get a report.

packages/blocks/src/api/raw-handling/test/html-formatting-remover.js

@@ -0,0 +1,110 @@
+/**
+ * Internal dependencies
+ */
+import filter from '../html-formatting-remover';
+import { deepFilterHTML } from '../utils';
+
+describe( 'HTMLFormattingRemover', () => {
+	it( 'should trim text node without parent', () => {
+		const input = 'a';
+		expect( deepFilterHTML( input, [ filter ] ) ).toEqual( input );
+	} );
+
+	it( 'should remove formatting space', () => {
+		const input = `
+			<div>
+				a
+				b
+			</div>
+		`;
+		const output = '<div>a b</div>';
+		expect( deepFilterHTML( input, [ filter ] ) ).toEqual( output );
+	} );
+
+	it( 'should remove nested formatting space', () => {
+		const input = `
+			<div>
+				<strong>
+					a
+					b
+				</strong>
+			</div>
+		`;
+		const output = '<div><strong>a b</strong></div>';
+		expect( deepFilterHTML( input, [ filter ] ) ).toEqual( output );
+	} );
+
+	it( 'should not remove leading or trailing space if previous or next element has no space', () => {
+		const input = `
+			<div>
+				a
+				<strong>b</strong>
+				c
+			</div>
+		`;
+		const output = '<div>a <strong>b</strong> c</div>';
+		expect( deepFilterHTML( input, [ filter ] ) ).toEqual( output );
+	} );
+
+	it( 'should remove formatting space (empty)', () => {
+		const input = `
+			<div>
+			</div>
+		`;
+		const output = '<div></div>';
+		expect( deepFilterHTML( input, [ filter ] ) ).toEqual( output );
+	} );
+
+	it( 'should remove block level formatting space', () => {
+		const input = `
+			<div>
+				<div>
+					a
+				</div>
+				<div>
+					b
+				</div>
+			</div>
+		`;
+		const output = '<div><div>a</div><div>b</div></div>';
+		expect( deepFilterHTML( input, [ filter ] ) ).toEqual( output );
+	} );
+
+	it( 'should remove formatting space around br', () => {
+		const input = `
+			<div>
+				a
+				<br>
+				b
+			</div>
+		`;
+		const output = '<div>a<br>b</div>';
+		expect( deepFilterHTML( input, [ filter ] ) ).toEqual( output );
+	} );
+
+	it( 'should remove formatting space around phasing content elements', () => {
+		const input = `
+			<div>
+				<strong>
+					a
+				</strong>
+				<strong>
+					b
+				</strong>
+			</div>
+		`;
+		const output = '<div><strong>a</strong> <strong>b</strong></div>';
+		expect( deepFilterHTML( input, [ filter ] ) ).toEqual( output );
+	} );
+
+	it( 'should ignore pre', () => {
+		const input = `<pre> a\n b\n</pre>`;
+		expect( deepFilterHTML( input, [ filter ] ) ).toEqual( input );
+	} );
+
+	it( 'should not remove white space if next elemnt has none', () => {
+		const input = `<div><strong>a </strong>b</div>`;
+		const output = '<div><strong>a </strong>b</div>';
+		expect( deepFilterHTML( input, [ filter ] ) ).toEqual( output );
+	} );
+} );

test/integration/fixtures/apple-out.html

@@ -19,27 +19,9 @@
 <!-- /wp:list -->
 
 <!-- wp:table -->
-<figure class="wp-block-table"><table class=""><tbody><tr><td>
-One
-</td><td>
-Two
-</td><td>
-Three
-</td></tr><tr><td>
-1
-</td><td>
-2
-</td><td>
-3
-</td></tr><tr><td>
-I
-</td><td>
-II
-</td><td>
-III
-</td></tr></tbody></table></figure>
+<figure class="wp-block-table"><table class=""><tbody><tr><td>One</td><td>Two</td><td>Three</td></tr><tr><td>1</td><td>2</td><td>3</td></tr><tr><td>I</td><td>II</td><td>III</td></tr></tbody></table></figure>
 <!-- /wp:table -->
 
 <!-- wp:paragraph -->
-<p>An image: </p>
+<p>An image:</p>
 <!-- /wp:paragraph -->

test/integration/fixtures/classic-out.html

@@ -15,7 +15,7 @@
 <!-- /wp:paragraph -->
 
 <!-- wp:paragraph -->
-<p>Fourth  paragraph</p>
+<p>Fourth paragraph</p>
 <!-- /wp:paragraph -->
 
 <!-- wp:more -->

test/integration/fixtures/evernote-out.html

@@ -1,7 +1,5 @@
 <!-- wp:paragraph -->
-<p>This is a <em>paragraph</em>.
-<br>This is a <a href="https://w.org">link</a>.
-<br></p>
+<p>This is a <em>paragraph</em>.<br>This is a <a href="https://w.org">link</a>.<br></p>
 <!-- /wp:paragraph -->
 
 <!-- wp:list -->
@@ -17,13 +15,7 @@
 <!-- /wp:separator -->
 
 <!-- wp:table -->
-<figure class="wp-block-table"><table class=""><tbody><tr><td>One
-</td><td>Two
-</td><td>Three
-</td></tr><tr><td>Four
-</td><td>Five
-</td><td>Six
-</td></tr></tbody></table></figure>
+<figure class="wp-block-table"><table class=""><tbody><tr><td>One</td><td>Two</td><td>Three</td></tr><tr><td>Four</td><td>Five</td><td>Six</td></tr></tbody></table></figure>
 <!-- /wp:table -->
 
 <!-- wp:image -->

test/integration/fixtures/markdown-out.html

@@ -7,8 +7,7 @@ <h1>This is a heading with <em>italic</em></h1>
 <!-- /wp:paragraph -->
 
 <!-- wp:paragraph -->
-<p>Preserve<br>
-line breaks please.</p>
+<p>Preserve<br>line breaks please.</p>
 <!-- /wp:paragraph -->
 
 <!-- wp:heading -->

test/integration/fixtures/ms-word-out.html

@@ -1,11 +1,9 @@
 <!-- wp:paragraph -->
-<p>This is a
-title</p>
+<p>This is a title</p>
 <!-- /wp:paragraph -->
 
 <!-- wp:paragraph -->
-<p>This is a
-subtitle</p>
+<p>This is a subtitle</p>
 <!-- /wp:paragraph -->
 
 <!-- wp:heading {"level":1} -->
@@ -29,25 +27,7 @@ <h2>This is a heading level 2</h2>
 <!-- /wp:list -->
 
 <!-- wp:table -->
-<figure class="wp-block-table"><table class=""><tbody><tr><td>
-  One
-  </td><td>
-  Two
-  </td><td>
-  Three
-  </td></tr><tr><td>
-  1
-  </td><td>
-  2
-  </td><td>
-  3
-  </td></tr><tr><td>
-  I
-  </td><td>
-  II
-  </td><td>
-  III
-  </td></tr></tbody></table></figure>
+<figure class="wp-block-table"><table class=""><tbody><tr><td>One</td><td>Two</td><td>Three</td></tr><tr><td>1</td><td>2</td><td>3</td></tr><tr><td>I</td><td>II</td><td>III</td></tr></tbody></table></figure>
 <!-- /wp:table -->
 
 <!-- wp:paragraph -->

test/integration/fixtures/ms-word-styled-out.html

@@ -1,15 +1,7 @@
 <!-- wp:paragraph -->
-<p>
-<strong>Lorem
-ipsum dolor sit amet, consectetur adipiscing elit&nbsp; </strong>
-</p>
+<p><strong>Lorem ipsum dolor sit amet, consectetur adipiscing elit&nbsp;</strong></p>
 <!-- /wp:paragraph -->
 
 <!-- wp:paragraph -->
-<p>
-Lorem
-ipsum dolor sit amet, consectetur adipiscing elit. Pellentesque
-aliquet hendrerit auctor. Nam lobortis, est vel lacinia tincidunt,
-purus tellus vehicula ex, nec pharetra justo dui sed lorem. Nam
-congue laoreet massa, quis varius est tincidunt ut.</p>
+<p>Lorem ipsum dolor sit amet, consectetur adipiscing elit. Pellentesque aliquet hendrerit auctor. Nam lobortis, est vel lacinia tincidunt, purus tellus vehicula ex, nec pharetra justo dui sed lorem. Nam congue laoreet massa, quis varius est tincidunt ut.</p>
 <!-- /wp:paragraph -->