来自 google 文档的 ckeditor 粘贴删除了样式
ckeditor pasting from google docs removes the styles
当我从 google 文档复制文本并使用 paste from word 按钮将其粘贴到 ckeditor 时，它删除了所有样式（粗体、斜体……）。
我该如何解决？
重现步骤：
从 google 文档中复制一个带下划线和斜体的单词 将其粘贴到编辑器中。
预期结果：
该词以斜体和下划线粘贴。
实际结果：
这个词是粗体的，没有下划线或斜体。
在config.js中添加以下行：
config.pasteFromWordRemoveFontStyles = false;
config.pasteFromWordRemoveStyles = false;
然后使用"Paste from Word"按钮（如下标记），不要直接使用CTRL+V或CMD+V粘贴。
我正在发布我的解决方案以防它对某人有帮助。
我在从 googleDocs 粘贴时遇到了同样的问题。
我发现有人在 github (13877) 中创建了 CKEditor 的一个分支来纠正这个问题。
因为我需要它与我的 4.5.6 版一起使用，所以我将它改编成一个插件。
我这样做是为了与 Processwire 一起工作，我无法更改 CKEditor 版本，并且只在这种情况下对其进行了测试。
- pasteFromGoogleDoc（文件夹）
- plugin.js
plugin.js的内容：
/**
* @license Copyright (c) 2003-2017, CKSource - Frederico Knabben. All rights reserved.
* For licensing, see LICENSE.md or http://ckeditor.com/license
* 2017-07-05
* the code originally from Frederico Knabben, written in ckeditor-dev-t-13877 branch
* (https://github.com/cksource/ckeditor-dev/tree/t/13877)
* has been adapted in a plugin by Rpapier for use in Processwire. It hasn't been tested elsewhere.
* DESCRIPTION
* Filter to paste from Google Doc and keep style (bold, italic, underline)
* Those style must be present in the toolbar for it to show, otherwise, it will bypass it
* For processwire :
* you must edit the field that has CKEditor and make sure that :
* - ACF is On
* - pasteFromGoogleDoc plugin is enabled
* - CKEditor toolbar configuration contains Bold, Italic and Underline
* - e.g : Format, Styles, -, Bold, Italic, Underline, -, RemoveFormat
* if Underline is not in the toolbar for example, it will be bypassed by the filter.
*/
( function() {
'use strict';
CKEDITOR.plugins.add( 'pasteFromGoogleDoc', {
requires: ['clipboard'],
init: function( editor ) {
// === arteractive hack for pasteFromGoogleDoc
var filterType,
filtersFactory = filtersFactoryFactory();
if ( editor.config.forcePasteAsPlainText ) {
filterType = 'plain-text';
} else if ( editor.config.pasteFilter ) {
filterType = editor.config.pasteFilter;
}
// On Webkit the pasteFilter defaults to 'webkit-default-filter' because pasted data is so terrible
// that it must be always filtered. (#13877)
else if ( CKEDITOR.env.webkit && !( 'pasteFilter' in editor.config ) ) {
filterType = 'webkit-default-filter';
}
//console.log(filterType);
editor.pasteFilter = filtersFactory.get( filterType );
editor.on( 'paste', function( evt ) {
// Init `dataTransfer` if `paste` event was fired without it, so it will be always available.
if ( !evt.data.dataTransfer ) {
evt.data.dataTransfer = new CKEDITOR.plugins.clipboard.dataTransfer();
}
// If dataValue is already set (manually or by paste bin), so do not override it.
if ( evt.data.dataValue ) {
return;
}
var dataTransfer = evt.data.dataTransfer,
// IE support only text data and throws exception if we try to get html data.
// This html data object may also be empty if we drag content of the textarea.
value = dataTransfer.getData( 'text/html' );
if ( value ) {
evt.data.dataValue = value;
evt.data.type = 'html';
} else {
// Try to get text data otherwise.
value = dataTransfer.getData( 'text/plain' );
if ( value ) {
evt.data.dataValue = editor.editable().transformPlainTextToHtml( value );
evt.data.type = 'text';
}
}
}, null, null, 1 );
editor.on( 'paste', function( evt ) {
var dataObj = evt.data,
type = dataObj.type,
data = dataObj.dataValue,
trueType,
// Default is 'html'.
defaultType = editor.config.clipboard_defaultContentType || 'html',
transferType = dataObj.dataTransfer.getTransferType( editor );
// If forced type is 'html' we don't need to know true data type.
if ( type == 'html' || dataObj.preSniffing == 'html' ) {
trueType = 'html';
} else {
trueType = recogniseContentType( data );
}
// Unify text markup.
if ( trueType == 'htmlifiedtext' ) {
data = htmlifiedTextHtmlification( editor.config, data );
}
// Strip presentational markup & unify text markup.
// Forced plain text (dialog or forcePAPT).
// Note: we do not check dontFilter option in this case, because forcePAPT was implemented
// before pasteFilter and pasteFilter is automatically used on Webkit&Blink since 4.5, so
// forcePAPT should have priority as it had before 4.5.
if ( type == 'text' && trueType == 'html' ) {
data = filterContent( editor, data, filtersFactory.get( 'plain-text' ) );
}
// External paste and pasteFilter exists and filtering isn't disabled.
else if ( transferType == CKEDITOR.DATA_TRANSFER_EXTERNAL && editor.pasteFilter && !dataObj.dontFilter ) {
// 2017-07-05 comment out this filter because it is already processsed somewhere...
//data = filterContent( editor, data, editor.pasteFilter );
//console.log(data);
}
if ( dataObj.startsWithEOL ) {
data = '<br data-cke-eol="1">' + data;
}
if ( dataObj.endsWithEOL ) {
data += '<br data-cke-eol="1">';
}
if ( type == 'auto' ) {
type = ( trueType == 'html' || defaultType == 'html' ) ? 'html' : 'text';
}
dataObj.type = type;
dataObj.dataValue = data;
delete dataObj.preSniffing;
delete dataObj.startsWithEOL;
delete dataObj.endsWithEOL;
/* evt.data.dataValue = data;
evt.data.dataValue = evt.data.dataValue
.replace( /zooterkins/gi, 'z********s' )
.replace( /gadzooks/gi, 'g******s' );*/
}, null, null, 6 );
}
} );
function filterContent( editor, data, filter ) {
var fragment = CKEDITOR.htmlParser.fragment.fromHtml( data ),
writer = new CKEDITOR.htmlParser.basicWriter();
filter.applyTo( fragment, true, false, editor.activeEnterMode );
fragment.writeHtml( writer );
return writer.getHtml();
}
// Returns:
// * 'htmlifiedtext' if content looks like transformed by browser from plain text.
// See clipboard/paste.html TCs for more info.
// * 'html' if it is not 'htmlifiedtext'.
function recogniseContentType( data ) {
if ( CKEDITOR.env.webkit ) {
// Plain text or ( <div><br></div> and text inside <div> ).
if ( !data.match( /^[^<]*$/g ) && !data.match( /^(<div><br( ?\/)?><\/div>|<div>[^<]*<\/div>)*$/gi ) )
return 'html';
} else if ( CKEDITOR.env.ie ) {
// Text and <br> or ( text and <br> in <p> - paragraphs can be separated by new \r\n ).
if ( !data.match( /^([^<]|<br( ?\/)?>)*$/gi ) && !data.match( /^(<p>([^<]|<br( ?\/)?>)*<\/p>|(\r\n))*$/gi ) )
return 'html';
} else if ( CKEDITOR.env.gecko ) {
// Text or <br>.
if ( !data.match( /^([^<]|<br( ?\/)?>)*$/gi ) )
return 'html';
} else {
return 'html';
}
return 'htmlifiedtext';
}
// This function transforms what browsers produce when
// pasting plain text into editable element (see clipboard/paste.html TCs
// for more info) into correct HTML (similar to that produced by text2Html).
function htmlifiedTextHtmlification( config, data ) {
function repeatParagraphs( repeats ) {
// Repeat blocks floor((n+1)/2) times.
// Even number of repeats - add <br> at the beginning of last <p>.
return CKEDITOR.tools.repeat( '</p><p>', ~~( repeats / 2 ) ) + ( repeats % 2 == 1 ? '<br>' : '' );
}
// Replace adjacent white-spaces (EOLs too - Fx sometimes keeps them) with one space.
data = data.replace( /\s+/g, ' ' )
// Remove spaces from between tags.
.replace( /> +</g, '><' )
// Normalize XHTML syntax and upper cased <br> tags.
.replace( /<br ?\/>/gi, '<br>' );
// IE - lower cased tags.
data = data.replace( /<\/?[A-Z]+>/g, function( match ) {
return match.toLowerCase();
} );
// Don't touch single lines (no <br|p|div>) - nothing to do here.
if ( data.match( /^[^<]$/ ) )
return data;
// Webkit.
if ( CKEDITOR.env.webkit && data.indexOf( '<div>' ) > -1 ) {
// One line break at the beginning - insert <br>
data = data.replace( /^(<div>(<br>|)<\/div>)(?!$|(<div>(<br>|)<\/div>))/g, '<br>' )
// Two or more - reduce number of new lines by one.
.replace( /^(<div>(<br>|)<\/div>){2}(?!$)/g, '<div></div>' );
// Two line breaks create one paragraph in Webkit.
if ( data.match( /<div>(<br>|)<\/div>/ ) ) {
data = '<p>' + data.replace( /(<div>(<br>|)<\/div>)+/g, function( match ) {
return repeatParagraphs( match.split( '</div><div>' ).length + 1 );
} ) + '</p>';
}
// One line break create br.
data = data.replace( /<\/div><div>/g, '<br>' );
// Remove remaining divs.
data = data.replace( /<\/?div>/g, '' );
}
// Opera and Firefox and enterMode != BR.
if ( CKEDITOR.env.gecko && config.enterMode != CKEDITOR.ENTER_BR ) {
// Remove bogus <br> - Fx generates two <brs> for one line break.
// For two line breaks it still produces two <brs>, but it's better to ignore this case than the first one.
if ( CKEDITOR.env.gecko )
data = data.replace( /^<br><br>$/, '<br>' );
// This line satisfy edge case when for Opera we have two line breaks
//data = data.replace( /)
if ( data.indexOf( '<br><br>' ) > -1 ) {
// Two line breaks create one paragraph, three - 2, four - 3, etc.
data = '<p>' + data.replace( /(<br>){2,}/g, function( match ) {
return repeatParagraphs( match.length / 4 );
} ) + '</p>';
}
}
return switchEnterMode( config, data );
}
function filtersFactoryFactory() {
var filters = {};
// GDocs generates many spans and divs, therefore `all` parameter is used
// to create default filter in Webkit/Blink. (#13877)
function setUpTags( all ) {
var tags = {};
for ( var tag in CKEDITOR.dtd ) {
if ( tag.charAt( 0 ) != '$' && ( all || tag != 'div' && tag != 'span') ) {
tags[ tag ] = 1;
}
}
return tags;
}
// Checks if content is pasted from Google Docs.
// Google Docs wraps everything in element with [id^=docs-internal-guid-],
// so that function just checks if such element exists. (#13877)
function isPastedFromGDocs( element ) {
if ( element.attributes.id && element.attributes.id.match( /^docs\-internal\-guid\-/ ) ) {
return true;
} else if ( element.parent && element.parent.name ) {
return isPastedFromGDocs( element.parent );
}
return false;
}
// Process data from Google Docs:
// * turns `*[id^=docs-internal-guid-]` into `span`;
// * turns `span(text-decoration=underline)` into `u`;
// * turns `span(font-style=italic)` into `em`
// * turns `span(font-style=italic)(text-decoration=underline)` into `u > em`. (#13877)
//
function processDataFromGDocs( element ) {
var styles = element.attributes.style && CKEDITOR.tools.parseCssText( element.attributes.style );
if ( element.attributes.id && element.attributes.id.match( /^docs\-internal\-guid\-/ ) ) {
return element.name = 'span';
}
if ( !styles ) {
return;
}
if ( styles[ 'font-style' ] == 'italic' && styles[ 'text-decoration' ] == 'underline' ) {
element.name = 'em';
element.wrapWith( new CKEDITOR.htmlParser.element( 'u' ) );
if (styles[ 'font-weight' ] > 400) {
element.wrapWith( new CKEDITOR.htmlParser.element( 'strong' ) );
}
} else if ( styles[ 'text-decoration' ] == 'underline' ) {
element.name = 'u';
if (styles[ 'font-weight' ] > 400) {
element.wrapWith( new CKEDITOR.htmlParser.element( 'strong' ) );
}
} else if ( styles[ 'font-style' ] == 'italic' ) {
element.name = 'em';
if (styles[ 'font-weight' ] > 400) {
element.wrapWith( new CKEDITOR.htmlParser.element( 'strong' ) );
}
}
}
function createSemanticContentFilter() {
var filter = new CKEDITOR.filter();
filter.allow( {
: {
elements: setUpTags(),
attributes: true,
styles: false,
classes: false
}
} );
return filter;
}
function createWebkitDefaultFilter() {
var filter = createSemanticContentFilter();
// Preserves formatting while pasting from Google Docs in Webkit/Blink
// with default paste filter. (#13877)
filter.allow( {
: {
elements: setUpTags( true ),
attributes: true,
styles: true,
match: function( element ) {
return isPastedFromGDocs( element );
}
}
} );
filter.addElementCallback( processDataFromGDocs );
return filter;
}
return {
get: function( type ) {
if ( type == 'plain-text' ) {
// Does this look confusing to you? Did we forget about enter mode?
// It is a trick that let's us creating one filter for edidtor, regardless of its
// activeEnterMode (which as the name indicates can change during runtime).
//
// How does it work?
// The active enter mode is passed to the filter.applyTo method.
// The filter first marks all elements except <br> as disallowed and then tries to remove
// them. However, it cannot remove e.g. a <p> element completely, because it's a basic structural element,
// so it tries to replace it with an element created based on the active enter mode, eventually doing nothing.
//
// Now you can sleep well.
return filters.plainText || ( filters.plainText = new CKEDITOR.filter( 'br' ) );
} else if ( type == 'semantic-content' ) {
return filters.semanticContent || ( filters.semanticContent = createSemanticContentFilter() );
} else if ( type == 'webkit-default-filter' ) {
// Webkit based browsers need semantic filter, because they produce terrible HTML without it.
// However original `'semantic-content'` filer is too strict and prevents pasting styled contents
// from many sources (e.g. Google Docs). Therefore that type extends original `'semantic-content'` filter. (#13877)
return filters.webkitDefaultFilter || ( filters.webkitDefaultFilter = createWebkitDefaultFilter() );
} else if ( type ) {
// Create filter based on rules (string or object).
return new CKEDITOR.filter( type );
}
return null;
}
};
}
function switchEnterMode( config, data ) {
if ( config.enterMode == CKEDITOR.ENTER_BR ) {
data = data.replace( /(<\/p><p>)+/g, function( match ) {
return CKEDITOR.tools.repeat( '<br>', match.length / 7 * 2 );
} ).replace( /<\/?p>/g, '' );
} else if ( config.enterMode == CKEDITOR.ENTER_DIV ) {
data = data.replace( /<(\/)?p>/g, '<div>' );
}
return data;
}
} )();
当我从 google 文档复制文本并使用 paste from word 按钮将其粘贴到 ckeditor 时，它删除了所有样式（粗体、斜体……）。
我该如何解决？
重现步骤：
从 google 文档中复制一个带下划线和斜体的单词 将其粘贴到编辑器中。
预期结果：
该词以斜体和下划线粘贴。
实际结果：
这个词是粗体的，没有下划线或斜体。
在config.js中添加以下行：
config.pasteFromWordRemoveFontStyles = false;
config.pasteFromWordRemoveStyles = false;
然后使用"Paste from Word"按钮（如下标记），不要直接使用CTRL+V或CMD+V粘贴。
我正在发布我的解决方案以防它对某人有帮助。 我在从 googleDocs 粘贴时遇到了同样的问题。 我发现有人在 github (13877) 中创建了 CKEditor 的一个分支来纠正这个问题。
因为我需要它与我的 4.5.6 版一起使用，所以我将它改编成一个插件。 我这样做是为了与 Processwire 一起工作，我无法更改 CKEditor 版本，并且只在这种情况下对其进行了测试。
- pasteFromGoogleDoc（文件夹）
- plugin.js
plugin.js的内容：
/**
* @license Copyright (c) 2003-2017, CKSource - Frederico Knabben. All rights reserved.
* For licensing, see LICENSE.md or http://ckeditor.com/license
* 2017-07-05
* the code originally from Frederico Knabben, written in ckeditor-dev-t-13877 branch
* (https://github.com/cksource/ckeditor-dev/tree/t/13877)
* has been adapted in a plugin by Rpapier for use in Processwire. It hasn't been tested elsewhere.
* DESCRIPTION
* Filter to paste from Google Doc and keep style (bold, italic, underline)
* Those style must be present in the toolbar for it to show, otherwise, it will bypass it
* For processwire :
* you must edit the field that has CKEditor and make sure that :
* - ACF is On
* - pasteFromGoogleDoc plugin is enabled
* - CKEditor toolbar configuration contains Bold, Italic and Underline
* - e.g : Format, Styles, -, Bold, Italic, Underline, -, RemoveFormat
* if Underline is not in the toolbar for example, it will be bypassed by the filter.
*/
( function() {
'use strict';
CKEDITOR.plugins.add( 'pasteFromGoogleDoc', {
requires: ['clipboard'],
init: function( editor ) {
// === arteractive hack for pasteFromGoogleDoc
var filterType,
filtersFactory = filtersFactoryFactory();
if ( editor.config.forcePasteAsPlainText ) {
filterType = 'plain-text';
} else if ( editor.config.pasteFilter ) {
filterType = editor.config.pasteFilter;
}
// On Webkit the pasteFilter defaults to 'webkit-default-filter' because pasted data is so terrible
// that it must be always filtered. (#13877)
else if ( CKEDITOR.env.webkit && !( 'pasteFilter' in editor.config ) ) {
filterType = 'webkit-default-filter';
}
//console.log(filterType);
editor.pasteFilter = filtersFactory.get( filterType );
editor.on( 'paste', function( evt ) {
// Init `dataTransfer` if `paste` event was fired without it, so it will be always available.
if ( !evt.data.dataTransfer ) {
evt.data.dataTransfer = new CKEDITOR.plugins.clipboard.dataTransfer();
}
// If dataValue is already set (manually or by paste bin), so do not override it.
if ( evt.data.dataValue ) {
return;
}
var dataTransfer = evt.data.dataTransfer,
// IE support only text data and throws exception if we try to get html data.
// This html data object may also be empty if we drag content of the textarea.
value = dataTransfer.getData( 'text/html' );
if ( value ) {
evt.data.dataValue = value;
evt.data.type = 'html';
} else {
// Try to get text data otherwise.
value = dataTransfer.getData( 'text/plain' );
if ( value ) {
evt.data.dataValue = editor.editable().transformPlainTextToHtml( value );
evt.data.type = 'text';
}
}
}, null, null, 1 );
editor.on( 'paste', function( evt ) {
var dataObj = evt.data,
type = dataObj.type,
data = dataObj.dataValue,
trueType,
// Default is 'html'.
defaultType = editor.config.clipboard_defaultContentType || 'html',
transferType = dataObj.dataTransfer.getTransferType( editor );
// If forced type is 'html' we don't need to know true data type.
if ( type == 'html' || dataObj.preSniffing == 'html' ) {
trueType = 'html';
} else {
trueType = recogniseContentType( data );
}
// Unify text markup.
if ( trueType == 'htmlifiedtext' ) {
data = htmlifiedTextHtmlification( editor.config, data );
}
// Strip presentational markup & unify text markup.
// Forced plain text (dialog or forcePAPT).
// Note: we do not check dontFilter option in this case, because forcePAPT was implemented
// before pasteFilter and pasteFilter is automatically used on Webkit&Blink since 4.5, so
// forcePAPT should have priority as it had before 4.5.
if ( type == 'text' && trueType == 'html' ) {
data = filterContent( editor, data, filtersFactory.get( 'plain-text' ) );
}
// External paste and pasteFilter exists and filtering isn't disabled.
else if ( transferType == CKEDITOR.DATA_TRANSFER_EXTERNAL && editor.pasteFilter && !dataObj.dontFilter ) {
// 2017-07-05 comment out this filter because it is already processsed somewhere...
//data = filterContent( editor, data, editor.pasteFilter );
//console.log(data);
}
if ( dataObj.startsWithEOL ) {
data = '<br data-cke-eol="1">' + data;
}
if ( dataObj.endsWithEOL ) {
data += '<br data-cke-eol="1">';
}
if ( type == 'auto' ) {
type = ( trueType == 'html' || defaultType == 'html' ) ? 'html' : 'text';
}
dataObj.type = type;
dataObj.dataValue = data;
delete dataObj.preSniffing;
delete dataObj.startsWithEOL;
delete dataObj.endsWithEOL;
/* evt.data.dataValue = data;
evt.data.dataValue = evt.data.dataValue
.replace( /zooterkins/gi, 'z********s' )
.replace( /gadzooks/gi, 'g******s' );*/
}, null, null, 6 );
}
} );
function filterContent( editor, data, filter ) {
var fragment = CKEDITOR.htmlParser.fragment.fromHtml( data ),
writer = new CKEDITOR.htmlParser.basicWriter();
filter.applyTo( fragment, true, false, editor.activeEnterMode );
fragment.writeHtml( writer );
return writer.getHtml();
}
// Returns:
// * 'htmlifiedtext' if content looks like transformed by browser from plain text.
// See clipboard/paste.html TCs for more info.
// * 'html' if it is not 'htmlifiedtext'.
function recogniseContentType( data ) {
if ( CKEDITOR.env.webkit ) {
// Plain text or ( <div><br></div> and text inside <div> ).
if ( !data.match( /^[^<]*$/g ) && !data.match( /^(<div><br( ?\/)?><\/div>|<div>[^<]*<\/div>)*$/gi ) )
return 'html';
} else if ( CKEDITOR.env.ie ) {
// Text and <br> or ( text and <br> in <p> - paragraphs can be separated by new \r\n ).
if ( !data.match( /^([^<]|<br( ?\/)?>)*$/gi ) && !data.match( /^(<p>([^<]|<br( ?\/)?>)*<\/p>|(\r\n))*$/gi ) )
return 'html';
} else if ( CKEDITOR.env.gecko ) {
// Text or <br>.
if ( !data.match( /^([^<]|<br( ?\/)?>)*$/gi ) )
return 'html';
} else {
return 'html';
}
return 'htmlifiedtext';
}
// This function transforms what browsers produce when
// pasting plain text into editable element (see clipboard/paste.html TCs
// for more info) into correct HTML (similar to that produced by text2Html).
function htmlifiedTextHtmlification( config, data ) {
function repeatParagraphs( repeats ) {
// Repeat blocks floor((n+1)/2) times.
// Even number of repeats - add <br> at the beginning of last <p>.
return CKEDITOR.tools.repeat( '</p><p>', ~~( repeats / 2 ) ) + ( repeats % 2 == 1 ? '<br>' : '' );
}
// Replace adjacent white-spaces (EOLs too - Fx sometimes keeps them) with one space.
data = data.replace( /\s+/g, ' ' )
// Remove spaces from between tags.
.replace( /> +</g, '><' )
// Normalize XHTML syntax and upper cased <br> tags.
.replace( /<br ?\/>/gi, '<br>' );
// IE - lower cased tags.
data = data.replace( /<\/?[A-Z]+>/g, function( match ) {
return match.toLowerCase();
} );
// Don't touch single lines (no <br|p|div>) - nothing to do here.
if ( data.match( /^[^<]$/ ) )
return data;
// Webkit.
if ( CKEDITOR.env.webkit && data.indexOf( '<div>' ) > -1 ) {
// One line break at the beginning - insert <br>
data = data.replace( /^(<div>(<br>|)<\/div>)(?!$|(<div>(<br>|)<\/div>))/g, '<br>' )
// Two or more - reduce number of new lines by one.
.replace( /^(<div>(<br>|)<\/div>){2}(?!$)/g, '<div></div>' );
// Two line breaks create one paragraph in Webkit.
if ( data.match( /<div>(<br>|)<\/div>/ ) ) {
data = '<p>' + data.replace( /(<div>(<br>|)<\/div>)+/g, function( match ) {
return repeatParagraphs( match.split( '</div><div>' ).length + 1 );
} ) + '</p>';
}
// One line break create br.
data = data.replace( /<\/div><div>/g, '<br>' );
// Remove remaining divs.
data = data.replace( /<\/?div>/g, '' );
}
// Opera and Firefox and enterMode != BR.
if ( CKEDITOR.env.gecko && config.enterMode != CKEDITOR.ENTER_BR ) {
// Remove bogus <br> - Fx generates two <brs> for one line break.
// For two line breaks it still produces two <brs>, but it's better to ignore this case than the first one.
if ( CKEDITOR.env.gecko )
data = data.replace( /^<br><br>$/, '<br>' );
// This line satisfy edge case when for Opera we have two line breaks
//data = data.replace( /)
if ( data.indexOf( '<br><br>' ) > -1 ) {
// Two line breaks create one paragraph, three - 2, four - 3, etc.
data = '<p>' + data.replace( /(<br>){2,}/g, function( match ) {
return repeatParagraphs( match.length / 4 );
} ) + '</p>';
}
}
return switchEnterMode( config, data );
}
function filtersFactoryFactory() {
var filters = {};
// GDocs generates many spans and divs, therefore `all` parameter is used
// to create default filter in Webkit/Blink. (#13877)
function setUpTags( all ) {
var tags = {};
for ( var tag in CKEDITOR.dtd ) {
if ( tag.charAt( 0 ) != '$' && ( all || tag != 'div' && tag != 'span') ) {
tags[ tag ] = 1;
}
}
return tags;
}
// Checks if content is pasted from Google Docs.
// Google Docs wraps everything in element with [id^=docs-internal-guid-],
// so that function just checks if such element exists. (#13877)
function isPastedFromGDocs( element ) {
if ( element.attributes.id && element.attributes.id.match( /^docs\-internal\-guid\-/ ) ) {
return true;
} else if ( element.parent && element.parent.name ) {
return isPastedFromGDocs( element.parent );
}
return false;
}
// Process data from Google Docs:
// * turns `*[id^=docs-internal-guid-]` into `span`;
// * turns `span(text-decoration=underline)` into `u`;
// * turns `span(font-style=italic)` into `em`
// * turns `span(font-style=italic)(text-decoration=underline)` into `u > em`. (#13877)
//
function processDataFromGDocs( element ) {
var styles = element.attributes.style && CKEDITOR.tools.parseCssText( element.attributes.style );
if ( element.attributes.id && element.attributes.id.match( /^docs\-internal\-guid\-/ ) ) {
return element.name = 'span';
}
if ( !styles ) {
return;
}
if ( styles[ 'font-style' ] == 'italic' && styles[ 'text-decoration' ] == 'underline' ) {
element.name = 'em';
element.wrapWith( new CKEDITOR.htmlParser.element( 'u' ) );
if (styles[ 'font-weight' ] > 400) {
element.wrapWith( new CKEDITOR.htmlParser.element( 'strong' ) );
}
} else if ( styles[ 'text-decoration' ] == 'underline' ) {
element.name = 'u';
if (styles[ 'font-weight' ] > 400) {
element.wrapWith( new CKEDITOR.htmlParser.element( 'strong' ) );
}
} else if ( styles[ 'font-style' ] == 'italic' ) {
element.name = 'em';
if (styles[ 'font-weight' ] > 400) {
element.wrapWith( new CKEDITOR.htmlParser.element( 'strong' ) );
}
}
}
function createSemanticContentFilter() {
var filter = new CKEDITOR.filter();
filter.allow( {
: {
elements: setUpTags(),
attributes: true,
styles: false,
classes: false
}
} );
return filter;
}
function createWebkitDefaultFilter() {
var filter = createSemanticContentFilter();
// Preserves formatting while pasting from Google Docs in Webkit/Blink
// with default paste filter. (#13877)
filter.allow( {
: {
elements: setUpTags( true ),
attributes: true,
styles: true,
match: function( element ) {
return isPastedFromGDocs( element );
}
}
} );
filter.addElementCallback( processDataFromGDocs );
return filter;
}
return {
get: function( type ) {
if ( type == 'plain-text' ) {
// Does this look confusing to you? Did we forget about enter mode?
// It is a trick that let's us creating one filter for edidtor, regardless of its
// activeEnterMode (which as the name indicates can change during runtime).
//
// How does it work?
// The active enter mode is passed to the filter.applyTo method.
// The filter first marks all elements except <br> as disallowed and then tries to remove
// them. However, it cannot remove e.g. a <p> element completely, because it's a basic structural element,
// so it tries to replace it with an element created based on the active enter mode, eventually doing nothing.
//
// Now you can sleep well.
return filters.plainText || ( filters.plainText = new CKEDITOR.filter( 'br' ) );
} else if ( type == 'semantic-content' ) {
return filters.semanticContent || ( filters.semanticContent = createSemanticContentFilter() );
} else if ( type == 'webkit-default-filter' ) {
// Webkit based browsers need semantic filter, because they produce terrible HTML without it.
// However original `'semantic-content'` filer is too strict and prevents pasting styled contents
// from many sources (e.g. Google Docs). Therefore that type extends original `'semantic-content'` filter. (#13877)
return filters.webkitDefaultFilter || ( filters.webkitDefaultFilter = createWebkitDefaultFilter() );
} else if ( type ) {
// Create filter based on rules (string or object).
return new CKEDITOR.filter( type );
}
return null;
}
};
}
function switchEnterMode( config, data ) {
if ( config.enterMode == CKEDITOR.ENTER_BR ) {
data = data.replace( /(<\/p><p>)+/g, function( match ) {
return CKEDITOR.tools.repeat( '<br>', match.length / 7 * 2 );
} ).replace( /<\/?p>/g, '' );
} else if ( config.enterMode == CKEDITOR.ENTER_DIV ) {
data = data.replace( /<(\/)?p>/g, '<div>' );
}
return data;
}
} )();