Custom Query

1. Set below in config.js

   config.entities = true;
   config.entities_processNumerical = true;
   

2. Go to editor, switch to source and paste below

   aaaaaa​​bbbbbbbbbbbbbb​​ccccccccccccccccccccc​​ddddddddddddddddd​​eeeeeeeeeeeeeeeeeeeeeeeeee​​fffffffffffffffffff​​ggggggggggggggggggggg
   

3. Switch to WYSIWYG and switch to source:

Entity gets doubled.

Problem can be reproduced from CKEditor 3.0 but I'm setting 4.0 as this is most likely going to be fixed in CKE 4.x.

1. Turn off ACF (allowedContent:true)
2. Load sample, switch to source and paste below:

   <audio controls>
     <source src="http://www.w3schools.com/tags/horse.ogg" type="audio/ogg">
     <source src="http://www.w3schools.com/tags/horse.mp3" type="audio/mpeg">
     Your browser does not support the audio element.
   </audio>
   

3. Switch to wysiwyg and right-click on audio element.

Result: JS error is thrown:
Error: Permission denied to access property 'nodeType'
Line : 998
URI: core/selection.js

Despite error I think whole plugin with placeholder (fake element) is missing here

Meta information can be found in dev/langtool/meta. It is used by CKEditor Translation Center (@ Transifex) to provide hints for translators ("Developer note").

When inserting a row/column/cell to the child table using the context menu, it will insert the row/column/cell on the parent table.

step to reproduce:

1. add a table (parent table) without header
2. add another table (child table) with header inside the parent table
3. right click on the first cell of the child table and insert a row/column/cell

Expected: Row/column/cell will insert in the child table

Actual: Row/column/cell added on the parent table

For more detail, video is attached.

Idea

Very often people don't realize that that our product is getting stable and richer. Users would benefit from updating CKEditor and, with a simple notification, we would let them know that "CKEditor lives".

The notification displayed in About dialog could look like that:

The newer version of CKEditor (<a>A.B.C</a>) is available (<a>Changelog</a>).

Technical background

When the About dialog is opened, a file is downloaded from ckeditor.com. It contains the latest version number and date, which is compared to the current version. If different, the dialog shows a message with a link to the download page.

Problem: If user explicitly disallows figcaption in ACF, figures are upcasted but the contents of (just stripped out) figcaption remain as a plain, non-editable paragraph within widget's body.

Expected: Once figcaption is disallowed, its contents should be removed from the widget. Such widget should be upcasted either to an inline widget or a block widget without caption (to be decided).

1. Open replacebycode.html sample
2. Clear page with New Page button and click on one of list buttons to insert list
3. Type "test1", Hit Enter, "test2", Hit Enter, "test3" and then stop
4. Press Ctrl+A to select all text
5. Select font-size 16px from Size dropdown
6. Click behind 3 in test3 to change selection (You need to press Enter in such way that selection changes but cursor is still in span)
7. Press Enter, press Enter and so on

Result: Items get created and you can't leave the list that way.

NOTES:

1. This TC won't show errors with list pasted into source
2. This is reproducible only in Firefox from CKEditor 4.0

1. Open replacebycode sample.
2. Using mouse select entire list ('Command module ... on the Moon').
3. Press backspace/delete.
4. Selection is placed between two paragraphs, directly in body (you can see that on elements path).

This is caused by poor implementation in ​editable.js. See also ​http://dev.ckeditor.com/ticket/10055#comment:11.

It's a part of bigger task - complete custom backspace/delete handling.

The definition of the function returned from dialog.validate.functions suffers from two bugs.

1. When obtaining the "value" from the arguments, the inner function should be referencing its own arguments, not those of the outer function.
2. When calling functions, it should be using "call" and passing "this" so that other validation functions passed as arguments will have the item as "this" as they expect.

If I set the editor content to a document with lots of images where the image size is not initially set, the autogrow plugin will not expand the editor as images are loaded. After all images are loaded, their size is known, and the document height are expanded without the autogrow plugin knowing. This results in a scrollbar being shown for the editor area.

It's hard to debug, but a simple fix should be enough:

With IE 11, load for example ​http://nightly.ckeditor.com/13-11-29-07-05/standard/samples/api.html

make sure to disable ACF for our tests, in the console: CKEDITOR.instances.editor1.filter.disabled=true

Then insert this "innocent" HTML <a style="display:inline-block">x</a> with the button available (using the console for this call seems to work fine, might be related to document focus)

IE11 throws an error in the scrollIntoView

to avoid it, it seems that it could be enough to check that type exists:

if ( this.type && this.type != CKEDITOR.SELECTION_NONE )

I just hope that you're able to reproduce it because I've spent most of this evening trying to find out what's wrong with IE11

Whenever I paste text with LINUX line breaks from clipboard into CKEDITOR with a toolbar that contains a 'Fotmat' button the text is processed correctly with line breaks, but if there's no 'Fotmat' button in toolbar - then the text is inserted in a single line.

Is it possible to configure CKEDITOR to process the text correctly in both cases without adding 'Fotmat' button in toolbar. (See attached files)

Reproducible in Version 4, and on CKEDTIOR demo page

​http://ckeditor.com/demo#standard - correct insert ​http://ckeditor.com/demo#widgets - inserted as single line

Edit:
In this case this so called text copied from browsers is in fact html where top tag is <pre> tag.
Because no pre is reported to ACF, filter removes pre tags but that is all it does.
ACF doesn't do anything to convert new line characters to e.g. <br /> or <p> .

When loading a CKFinder instance with readonly = true, the following javascript error is thrown:

// In IE and Chrome:
Uncaught TypeError: Cannot read property 'pW' of undefined ckfinder.js:7

// In Firefox:
TypeError: as is undefined ckfinder.js:7

The code in question seems to be this in IE/Chrome:

L.prototype.setUploadCommand=function(ar,as){var at,au,av,aw;if(!as.pW||as.pW.length<1)return;

Sorry, no unminimized example.

Finder was created thusly:

var finder = new CKFinder();
finder.basePath = '/Scripts/ckfinder_pdfmerge/';
finder.readOnly = true;
finder.create();

Everything seems to work, but the error is very annoying. Browsers tested: Firefox 24, IE 11 and Chrome 31.

Currently ignored:

	'/dt/core/style/editor.html#test apply inline style on non-editable inline element - at non-editable inline boundary':
		'env.ie && env.version == 8',
	'/dt/core/style/editor.html#test remove inline style from non-editable inline element - at non-editable inline boundary':
		'env.ie && env.version == 8',
	'/dt/plugins/indent/indent.html#test indent next to inline non-editable':
		'env.ie && env.version == 8',

We thought that they will be fixed by #11042, but nothing has changed.

Perhaps related issues: #7895 and #8989.

There is bad TC called test avoid having inline event handlers executed in dt\core\htmldataprocessor.html
The reason is that it attempts to execute editorBot#setData multiple times (using doTest function. setData does call wait and resume inside itself so any code after first wait() call will not be executed.

Problem can be reproduced in Firefox only from CKEditor 3.6.4 rev. [7527] in both CKE 3.x and 4.x

1. Create subdomain.example.com and example.com domains
2. Put ckeditor with attached sample under subdomain.example.com
3. This sample has document.domain='example.com’ set
4. Load attached sample from subdomain (​http://subdoamin.example.com/ckeditor/samples/replacebycode2.html)
5. Press preview

Result: Permission denied error is thrown.

Error: Permission denied to access property '_cke_htmlToLoad'
Code: doc.write( window.opener._cke_htmlToLoad );
URI: ​http://www.subdomain.example.com/ckeditor/plugins/preview/preview.html

If widget editables are not direct descendants of widget root element, it is possible to edit/delete elements' editables with CreateDiv context menu options. Deleting results in loss of editable element and its content is moved into editables parent element.

In example below with CreateDiv context menu you could remove content-left and content-right columns. Moving content from those divs to their parent divs (col-xs-12).

Juicy parts of an widget definition:

allowedContent: 'div(!row);'+
            'div(!col-xs-12,col-sm-6,col-sm-3,col-sm-9);'+
            'div(!content-text,content-left,content-right);',

// Minimum HTML which is required by this widget to work.
requiredContent: 'div(!row);div(!col-xs-12);div(!content-left);div(!content-right);',


// Define two nested editable areas.
editables: {												
            contentLeft: {
              selector: '.content-left'
            },
            contentRight: {
              selector: '.content-right'
            }
},
// Define the template of a new widget.
template: '<div class="row">' +					
            '<div class="col-xs-12 col-sm-6">' +
              '<div class="content-text content-left">' +
                '<h2>Title 1</h2>' + 
                '<p>Some text</p>' +						
              '</div>' +
            '</div>' +
            '<div class="col-xs-12 col-sm-6">' +
              '<div class="content-text content-right">' +
                '<h2>Title 2</h2>' + 
                '<p>Some text</p>' +						
              '</div>' +
            '</div>' +					
          '</div>',

For example: ​http://ckeditor4.t/dt/core/focusManager/focus.html (#11153)

Mocked editor does not behave like the real one. It's cool to create light unit test, but most important is to write precise and realistic tests. All those which create CKEDITOR.editor instances manually are unsafe and should be rewritten.

Description (I attached a detailed image that shows every aspect of this issue well - please see it):
1) Create a cell and fill them with numbers - just for knowing where the cells moves at the end.
2) Merge down a cell at the right edge of the cell.
3) Split the merged cell vertically.

RESULT: The new cell comes to the first place of the affected row instead of the place where the split command was applied. Every other cells are shifted right. Note: If the merge/split action is applied to an intermediate cell (not on the last one in the row), there is no issue.

I can recreate it with the demo page with IE9, IE10.

I don't have a hosted example, but if you take

​http://nightly.ckeditor.com/13-12-04-07-05/standard/samples/ajax.html

and simply change the line where the config is defined to read:

var config = {startupFocus: true};

you will notice that creating the editor will result in a focused cursor on desktop but not in iPad

When the new language menu provided in CKEditor 4.3 contains a large number of language options, the menu is just displayed as one long list - see attached screen capture. This menu should be styled in a different way so that the language list can be displayed in a more user friendly way when a large number of language options are provided e.g. the menu could be scrollable or presented as a grid layout etc.

Otherwise there should be some way to customize the styling of this menu so that we could use CSS to change the appearance of the menu. Currently there is no easy way to customize this menu.

One idea which could help to make the styling of menus more flexible is adding a feature specific class to the cke_menu_panel div like cke_menu__<menuName> e.g. cke_menu__language, cke_menu__scayt etc. We realize that this div is reused for all menus though so this would need to be maintained as menus are opened and closed. Would this be possible?

until master

I feel that there is a case when down arrow with buttons does not work as expected. It works good with i.e. language button, color buttons - but still does not work as expected with richcombos, because it still does not focus first option.

The exception here is format menu, which checks currently active format.

1. open any sample with CKEditor (i.e. samples/replacebyclass.html)
2. navigate using keyboard to styles dropdown (focus it)
3. press down arrow

Expected result:
First entry in dropdown should be focused, in this case Italic Title.

Current result:
Only menu is focused.

additional info:

1. funny thing is that we don't block down arrow in this case, if scrollbar is visible (in document contianing editor), and its moveable towards bottom, then pressing down arrow will move screen

origin issue:
​t/11075

master

1. open any sample with CKEditor (i.e. samples/replacebyclass.html)
2. use following source code:

   <p>&nbsp;</p>
   
   <p><textarea name="foo"></textarea></p>
   
   <p><textarea name="foo"></textarea></p>
   
   <p>&nbsp;</p>
   

3. go back to wysiwyg mode, and put cursor in first line
4. press down arrow

Important note:
With firefox even more weird things happen, we are able to place cursor inside textarea, and some exception like:

Error: Permission denied to access property 'nodeType' @ http://presets.ckeditor.dev/4.3.1/full-all/ckeditor/ckeditor.js:353

Are thrown, however it has already been that way at least since 4.3 (#11262).

Expected result:
Caret should move to paragraph with textarea

Current result:
Caret skips paragraph with textarea, and moves to last line

additional info:

1. confirmed in chrome @ win8
2. it was even reproducable when i placed 2 textareas in paragraph
3. my assumption is that it's related to #11042 but it's only my gutts tell

Case 1: An error is thrown when clicking some tag in elementspath (see: screencast​).

Case 2: Confusing selection after clicking a tag in elementspath (see: screencast​).

Quite likely both cases have the same root: to be investigated.

Use this HTML:

<table border="1" cellpadding="1" cellspacing="1" style="width:500px">
	<tbody>
		<tr>
			<td>
			<p>&nbsp;</p>
			</td>
			<td>&nbsp;</td>
		</tr>
	</tbody>
</table>

1. Place caret in first table cell.
2. Move your mouse cursor to a place which will trigger magicline.
3. Press delete.
4. Error is thrown.

Uncaught TypeError: Cannot read property 'type' of null range.js:1526
CKEDITOR.dom.range.setStart range.js:1526
CKEDITOR.dom.range.setStartAfter range.js:1573
CKEDITOR.dom.range.setStartAt range.js:1618
CKEDITOR.dom.range.moveToPosition range.js:1489
CKEDITOR.dom.range.moveToElementEditablePosition range.js:2047
CKEDITOR.dom.range.moveToElementEditEnd range.js:2157
(anonymous function)

Starts in: ​https://github.com/ckeditor/ckeditor-dev/blob/c240965c7134620f8d2fe2c36f865197dc280aa9/core/editable.js#L653

Reproduced it on every browser.

1. Open replacebycode sample.
2. Click somewhere in the text.
3. Open find dialog and type "the", click "find" and close dialog.
4. Repeat 2&3 until error is thrown.

I couldn't reproduce it on Chrome and I haven't checked other browsers.

Edit:

Click somewhere in the text.

The main problem I had with reproducing this was that you have to click on "the" word like in the screen cast and not just anywhere in the text.

1. Open Opera or Chrome.
2. Go to replacebyclass.html sample.
3. Paste in the sorce mode:

   <p>lorem</p>
   <hr />
   <p>ipsum</p>
   

4. Switch to the WYSIWYG mode.
5. Put cursor after "lorem".
6. Press right arrow button.
7. Press enter.

Result: Two paragraphs have been created instead of one.

1. Open any browser (tested with IE8 and Chrome).
2. Go to "New Image plugin" sample.
3. Insert flash with dimensions (ex. 500 width; 500 height).
4. Select placeholder cut it and paste somewhere.

Result: placeholder changes size.

Moreover if you do not set dimensions of flash after cut & paste you will get width = 1 and height = 1.

Also in "Replace by class" sample everything is ok.

In IE8:

1. Open "New Image plugin" sample.
2. Click next to widget handler.

Result: you can drag widget using IE native D&D.

Consider the usual workflow — typing along, and now you want to make a link. Until you have been re-trained by CKEditor, you are likely to press the link button, and now you have your URL— but no custom anchor text, and no way to create it, except by deleting the link, typing some text, selecting it, and now making pressing the link button (if you even know this is an option).

This could be alleviated simply by having a text box for the anchor text. If left blank we put in the URL as the anchor, as is done now. But if something is put in there (or what was there is edited) then that is used as the anchor text. So the proposed fix is to allow adding or editing anchor text when creating or editing a link, but the user's pain is mostly for new links i'm sure.

If this has a chance of going in i think i'd be able to make a patch. My end-users thank you for your consideration!

Checked on IE8 and IE9:

1. Open replacebycode sample.
2. Click in various places outside body, on the left side of the editor area.
3. See that selection in elements path does not reflect the place where caret is placed. Also, if you apply a style, it will be applied in different place (previous selection position).

This means that selectionChange wasn't correctly fired perhaps due to lack of native selectionchange.

Panels are too narrow, so even our default settings look bad if font is missing (FF@Fedora):

Or if someone wants a longer name:

A better solution would be to allow configure dropdowns width or, better, if editor chooses width automatically based on the content.

Steps:

1. instantiate ckeditor with mathjax plugin installed using https protocol.
2. clicking mathjax throws error in console:

Chrome:"[blocked] The page at '[site with https]' was loaded over     HTTPS, but ran insecure content from 'http://cdn.mathjax.org/mathjax/2.2-latest/MathJax.js?config=TeX-AMS_HTML': this content should also be loaded over HTTPS."

in Firefox: "[14:41:12.367] Blocked loading mixed active content "http://cdn.mathjax.org/mathjax/2.2-latest/MathJax.js?config=TeX-AMS_HTML" @ https://[site]"

os:pc/mac

browser:chrome

according to the Mathjax website: ​http://www.mathjax.org/resources/faqs/#problem-https

their CDN does not offer https support and instead, asks users to use ​https://c328740.ssl.cf1.rackcdn.com/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML as an alternative address. It'd be great if ckeditor can test the protocol of the browser accessing the script and change the path accordingly. even better would be leaving plugin.js in the mathjax plugin folder as it was in version 4.3 beta.

Based on: #11315.

There are couple of scriptLoader.load usage and none check if request completed. When request failed error should be thrown, so developer knows what happened.

1. Copy ckeditor_test.html and blink.html to samples directory.
2. Click at the region below the paragraph, e.g. bottom right corner

Problem: Cursor is not visible.

Tested in IE9-11. The similar issue was fixed with #1659 and #7932

Using the Link dialog's Upload tab to upload a file shows no indication that something is happening in IE<=11 until after the file is fully uploaded. Chrome and Firefox both show their activity spinners once you start the upload, but nothing happens with IE. Using CKFinder for the file management. Also happens on your demo page.

For certain block elements inside CKEditor, if an element is at the beginning of the content, it is impossible to delete it using the Backspace key.

Sample elements: <form>, <blockquote>, <div>

Note: deleting does work in one particular case: only if there is no content after the element, so make sure to add some text to the content as well.

Test Case 1

1. ​http://ckeditor.com/latest/samples/replacebyclass.html
2. Paste in source mode

<blockquote>
<p>&nbsp;</p>
</blockquote>

<p>Text</p>

3. Switch to wysiwyg, place cursor in the Blockqoute element, press Backspace
4. Result: Blockqoute is not deleted.

Note 2: Pressing enter deletes Blockqoute (?)

Test Case 2

Repeat steps 1-4 with:

<form action="a" name="n">&nbsp;</form>

<p>Text</p>

(Enter key works as expected)

Test Case 3

Repeat steps 1-4 with

<div class="class">
<p>&nbsp;</p>
</div>

<p>Text</p>

(Enter key works as expected)

It is technically possible to workaround this bug by using magic line to insert first an element above (or use the Source mode), but I don't think that's a real solution for this bug.

observed at master

FF does not display images in preview.

1. open any sample with CKEditor (i.e. samples/replacebyclass.html)
2. click on preview button

Expected result:
You should se our sample image at right hand side.

Current result:
Image is missing, and alt is displayed.

additional info:

1. It's due to missing base tag in head. Since we're opening preview in /plugins/preview/iepreview.html, browser (correctly) assume that all relative links should be resolved starting from preview directory. So assets/foo.png will be resolved as /plugins/preview/asstes/foo.png.

When ckeditor is inside div which is display: none and textarea is empty, then when you show div (change css to display: block) and checkDirty returns false.

Here is example:

​http://hosting.data365.eu/ckeditor_bug/index.html

My browser is firefox 26 and running on Ubuntu Linux 13.10.

If "enterKey" is enabled, hitting [Return] while in a simple block-level (div) widget will break the widget by creating a second "widget body" within the widget wrapper. This leads to odd formatting and an unexpected DOM.

See attached demo, the only plugins are enterKey and a custom widget plugin, put the cursor within the displayed word and hit the [Return] key. Instead of adding a line/paragraph break it will create a second widget within the container (with its own outline and everything).

Reported by a customer.

When using CKEDITOR.inline() to replace a textarea element, the width configuration setting does not work, the editor always takes the whole available width which can destroy the layout.

While investigating I've noticed that setting height do not work as well. This is in quite confusing and not documented (e.g. ​here).

Proposed solutions:

1. For width, apply the width inline style to the "main" <div class="cke_textarea_inline"> element

2. config.height should be set with min-height, to allow expanding the size of the editor while the content is added.

After call range.enlarge( CKEDITOR.ENLARGE_ELEMENT ); in some cases <\p> tag is inside enlarged range (dt/core/dom/enlarge.html 'test space 9', currently in ​git:t/10778)

<p>foo <i><b>[bar]</b></i> </p> ->
<p>foo [<i><b>bar</b></i> </p>]

and in some it is not (dt/core/dom/enlarge.html test_enlarge_element5):

<p>Test <b><i>[Enlarge</i> this]</b></p> ->
<p>Test [<b><i>Enlarge</i> this</b>]</p>

Reported through support channel.

Not sure if related to #11219.

When config.forcePasteAsPlainText is set, dragging&dropping inline widget results in widget being lost. Either drag&drop should be automatically disabled when config.forcePasteAsPlainText is set or widgets should not be lost in this very specific case.

A sample with inline widget is attached, drag & drop the widget and see that it is lost.

Hard to describe and hard to search for so my apologies if it's already been logged but I think even if it's been logged I've found more details I will explain below.

If I highlight a section of text and apply some sort of text formatting(bold for example). Then place the cursor at the end of one of those newly bolded lines and click bold again to remove it so I can continue typing in non-bold it will remove bold from the entire line.

Browser: FF 25.0, IE9.0.8112.16421(config.enterMode = CKEDITOR.ENTER_BR only for IE)

OS: Win 7

Reproducible location: Demo site

I tested with Chrome 31.0.1650.63 m as well and it worked properly. Also ctrl+a vs using the mouse can have slightly different results which I will explain after the steps to reproduce.

Steps to reproduce:

1. Download the standard ckeditor from the website.(Only necessary if you want to test CKEDITOR.ENTER_BR. Otherwise just go to the demo)
2. Open replacebycode.html sample and clear the content
3. Type aaaa aaaa, hit enter. Type bbbb bbbb, hit enter. Type cccc cccc, hit enter.
4. Press Ctrl+A and click Bold
5. Place the cursor at the end of any of the three lines
6. Press Bold again

If Using IE or FF and using CKEDITOR.ENTER_BR you will notice that the entire highlighted content will lose the bold formatting. If using FF and CKEDITOR.ENTER_P (Default) you will notice that just the line you placed the cursor on will lose bold. I did NOT test CKEDITOR.ENTER_DIV.

The slight variation I mentioned between ctrl+a and using the mouse can be tested by repeating all the steps above but step 4 use the mouse to highlight everything instead and step 5 place the cursor at the end of the "cccc cccc" line. In FF bold is already de-selected from the toolbar, in IE it's still selected but when you click it formatting isn't lost. If you test lines 1 or 2 you will see the same results as if you tested with ctrl+a.

EDIT:

Reproducible in Firefox and IE11

To reproduce:

1. Type three lines of text
2. Ctrl+A, hit Bold Button
3. Click at the end of any line and Hti bold Button again

Results:
In Enter Div and P Whole line gets unbolded
In Enter BR whole text gets unbolded

If you import the textcomibination "cke:" the text will be deleted until the next HTML-Tag.

Sample: [try to import] "Sie gehen über die Brücke: Thomas, Peter und Susanne.

I tried this with the words "Stücke:" and "Lücke:"

this includes " Br&uuml;cke: Thom......" I think "cke:" will be used from the editor :-)

Tests only feature request
Since 8abc140af4644f165b9c7d227919b0a2cdfdb8fd

Since PJ created pretty convenient function to test all 3 editor modes, which was commited by PK in 8abc140af4644f165b9c7d227919b0a2cdfdb8fd. We should update our old tests which do that with custom functions (each time), since it will also simplify it.

Other candidates for such refact

ckeditor-tests-v4/dt/core/editable/events.html
ckeditor-tests-v4/dt/core/editable/readonly.html

Set allowed content to true and then insert an image map in the content:

<img alt="" src="assets/sample.jpg" usemap="#imgmap201414175321zz" />
<map id="imgmap201414175321zz" name="imgmap201414175321zz"><area coords="5,5,155,195" href="page.html" shape="rect" /></map>

Now right click on the map with the developer tools open and the tab will crash in Chrome. Adding a check in the tabletools plugin seems to be enough:

From

editor.contextMenu.addListener( function( element, selection, path ) {
	var cell = path.contains( { 'td': 1, 'th': 1 }, 1 );
	if ( cell && !cell.isReadOnly() ) {

To

editor.contextMenu.addListener( function( element, selection, path ) {
	var cell = path && path.contains( { 'td': 1, 'th': 1 }, 1 );
	if ( cell && !cell.isReadOnly() ) {

Despite this patch, I've already filed a ticket in Chrome, but this patch will avoid getting the error. ​https://code.google.com/p/chromium/issues/detail?id=331664

Related issue: #11308.

After call range.enlange there should be both opening and closing tags in range or none of them, but tests created for #10778 shown some incorrect situations:

	// <div>x<p>[foo] <b></b></p>x</div> ->
	// <div>x[<p>foo] <b></b></p>x</div>

	//Should be:
	// <div>x<p>[foo] <b></b></p>x</div>

Or:

	// <div>x<p>foo <i><b>[bar]</b></i> </p>x</div> ->
	// <div>x<p>foo [<i><b>bar</b></i> </p>]x</div>

	//Should be:
	// <div>x<p>foo [<i><b>bar</b></i> ]</p>x</div>

Note that changes introduced in #10778 does not cause this behavior, the bug existed before. To see more incorrect situations go to enlarge.html tests.

To reproduce on iPad: Click the Table button to display the Table Properties dialog box. Click the OK button. Verify that a cursor is displayed in the first cell of the inserted table. Verify that the virtual keyboard is displayed. Type some characters on the virtual keyboard. Verify that the typed text is not inserted at the cursor.

We have a questionnarie answering page where multiple ckeditors are shown. Whenever we have 3 or more, every toolbar from the 3rd editor has remnants from the black/white styles left i plugins/icons.png.

This is a custom built ckeditor with only the moonocolor theme, but still this icon set is part of the plugin folder.

When adding a select box with a multiple attribute, Chrome crashes when ctrl-clicking on an option within it, instead of a context menu the browser tab just crashes, with no exception thrown or no crash report in Chrome.

This bug occurs on Mac OS X Mountain Lion (10.8.6), tested with CKEditor 4.0.1 and 4.3.1 on the site demo. Both have the same result.

It's a Blink bug - ​https://code.google.com/p/chromium/issues/detail?id=332719

Related issue: #11365.

On Ipad 2 (IOS 6.1.3) and Ipad 3 (IOS 7.0.3) elements are always placed at the very first position of the editor window rather than at the insertion point (current position of the cursor). This happens when using toolbar buttons such as special links and emoticons. On the other hand if you select a piece of text a link will be positioned correctly on top of it and an emoticon will replace it.

At the moment there's no way to specify advanced upcasting/match because enterMode in unknown for those methods (as well as other options).

Originated from #11283.

1. Open image2 sample.
2. Focus image.
3. Press ordered list button.
4. Press unordered list button.

Image's structure explodes.

<:figure class="caption cke_widget_element" data-widget="image" data-cke-widget-keep-attr="0" data-cke-widget-data='{"hasCaption":true,"src":"assets/image1.jpg","alt":"Saturn V","width":"200","height":"","lock":true,"align":"right"}'><SPAN class=cke_image_resizer_wrapper><IMG alt="Saturn V" src="assets/image1.jpg" width=200 data-cke-saved-src="assets/image1.jpg"><SPAN class="cke_image_resizer cke_image_resizer_left" title="Click and drag to resize" data-cke-expando="138">​</SPAN></SPAN><:figcaption class=cke_widget_editable contentEditable=true data-cke-expando="false" data-cke-display-name="caption" data-cke-filter="107" data-cke-widget-editable="caption" data-cke-enter-mode="2">Roll out of Saturn V on launch pad</:figcaption></:figure>

FF's implementation of focus/blur handling on nested editables seems to be very fragile - tests in ​http://ckeditor4.t/dt/plugins/widget/nestededitables.html randomly fail from time to time depending on how they were ran. Additionally, I had to add one of the tests to regressions, because it started to fail after b671945e@tests.

None of these instabilities occur when testing manually.

This is more reminder to create tests for issue #11331

There is an issue with the Editor being used on the iPad (not saying it is limited to the iPad, but that's what I've been testing in) where the Editor will stop accepting any input. I can get this to happen in general when I'm switching focus between other INPUT elements on the screen and then try to focus on the Editor, but it is most easily repeated by using the styles/normal/font/size menus on the Editor itself.

All test cases were performed on ​http://ckeditor.com/demo#full

TC1

1. Press "New Page" button.
2. Select "Marker" style from the Styles dropdown.

Note: at this stage the editor looks like it has focus, the cursor is blinking.

3. Start typing.

Result: typed text is not entered at all (and the cursor is still blinking).

4a) Touch the wysiwygarea to set focus, wait 2 seconds for the native context menu to appear and then start typing.

Result: typed text is entered and style is correctly set.

4b) Touch the wysiwygarea to set focus and quickly start typing.

Result: typed text is entered but selected style is ignored.

Touching the wysiwygarea to set focus before or after pressing the "New Page" button did not change anything, the editor behaves in the same way.

The problem with user not being able to type after using select dropdown happens with Styles/Format/Font/Size and also when setting Font/Background color.

TC2

Related to TC1, important to notice the difference between issues in TC1, points 3 and 4. Other styles like Bold also do not work in certain cases, I reported it in a separate ticket #12044

blackey: I'm experiencing this issue on an iPad 3, running iOS 7.0.3 and using the Safari browser.

wwalc: confirmed on iOS 7.1.1.

since: 4.0 (didn't check earlier) until master

1. open any sample with CKEditor (i.e. samples/replacebyclass.html)
2. focus editor
3. press alt+f10 to focus toolbar
4. navigate to preview button
5. press button using space

Expected result:
Preview should appear

Current result:
firefofx: Preview is blocked by popup block
ie11: Exception is being thrown SCRIPT5007: Unable to set property 'location' of undefined or null reference, and popup blocker appears. Error is reported in #11597.

additional info:

1. preview is not blocked when we do this using mouse
2. tested with FF 26.0 at Win7, IE11 at Win8.1

edit: it appears that popup blocker also triggers in IE10 (@Win8), IE9 (@Win7)

since: 4.2 until master

1. open any sample with CKEditor (i.e. samples/jquery.html)
2. click maximize button in framed editor
3. click it once again to minimize editor
4. you can now edit everything in whole page

Expected result:
Body should remain untouched.

Current result:
Body element (in main window) gains contenteditable=true.

additional info:

1. I assume it's not a FF bug, coz when i checked IE11 + 4.2 (back then ie11 was recognized as gecko) it shown exaclty the same deffect

1. Open editor (replacebyclass.html) and paste following html as a source:

   <ol>
   	<li>aa
   	<blockquote>
   	<ol>
   		<li>bb</li>
   		<li>cc</li>
   	</ol>
   	</blockquote>
   	</li>
   	<li>dd</li>
   </ol>
   

2. Go to WYSIWYG mode and select [bb ... cc].
3. Press "Insert/Remove Bulleted List" button.

Result:

• nested list type is unchange,
• error in console: "Uncaught TypeError: Cannot read property 'parent' of undefined."
• when you do to the source mode you will se bookmarks:

  <li><span id="cke_bm_152S" style="display:none">&nbsp;</span>bb</li>
  <li>cc<span id="cke_bm_152E" style="display:none">&nbsp;</span></li>
  

Tested with Chrome and FF.

1. Open any page with CKEditor in Safari.
2. Maximize browser window (an arrow in the upper right-hand corner).
3. Open a dialog (link, image).
4. ESC.

Expected: Dialog is closed. Actual: Dialog is closed but window is no longer fullscreen.

My first guess is that some preventDefault is needed. It is not reproducible in Chrome though, so additional research is needed to figure out if the problem is CKEditor or Safari itself.

1. Open plugins/mathjax/samples/mathjax.html
2. Wait for formulas to load.
3. Double click a formula, don't change or touch anything.
4. Press ESC.

Expected: Dialog should close

Actual: Editor says that some dialog contents has been altered (but really hasn't).

5. Press ESC to get rid of the window.

Expected: Dialog should remain open (even though the prompt is invalid), cancel should prevent dialog from closing.

Actual: Sometimes another prompt is displayed which, when ESCed, sends us back to the editor. Sometimes no additional prompt is displayed but this is still invalid because dialog is closed.

I'm unable to reproduce in other browsers/dialogs.

I managed to reproduce it with sourcedialog, randomly (see another screencast).

1. Open placeholder plugin.
2. Select "is a sample placeholder. You".
3. Drag it and drop at the end of paragraph.
4. Drag&Drop widget using drag handler.

Result: It is not possible to d&d widget.

Tested on Chrome, FF and IE 11. On each I get different results but on any of them I was not able to drop widget.

1. Go to placeholder plugin.
2. Select whole content of the editor.
3. Copy (Ctrl+C).
4. Paste into notepad (or similar).

Result on Chrome:

This is a [[sample placeholder]]Click and drag to move. You are using CKEditor.

On FF:

This is a [[sample placeholder]] [Click and drag to move] . You are using CKEditor.

"Click and drag to move." should not be there.

1. Open IE8 and go to replacebyclass.html.
2. Put cursor after Apollo 11 header.
3. Press "Replace" button.
4. Set Find what: Apollo and leave Replace with: empty.
5. Double click on Replace button so the second Apollo (beginning of the first paragraph) will be replaced by empty string.
6. Press "Cancel" button.

Result: You cannot close the dialog and error in the console: 'type' is null or not an object.

Since 4.0.

I've tested this with Chrome and IE 10 and everything is fine there.

1. Open placeholder.html sample in IE 8.
2. Copy sample placeholder and paste it at the end of paragraph.
3. Drag the second placeholder (using drag handler) and drop in directly after the first one.

Result: Placeholder disappears.

Sometimes it is possible to drop it correctly.

Sometimes I get "Unspecific error" doing the same with image inline widgets.

1. Open image2 plugin in IE 8.
2. Click Image button and open "Image properties" dialog.
3. Set http://upload.wikimedia.org/wikipedia/commons/6/63/Wikipedia-logo.png as a URL (wikipedia logo with width 200 and height 200).
4. Change the width of the image to 20.

Result: height is 21 instead of 20.

5. Change the width of the image back to 200.

Result: height is 214 instead of 200.

Clearly IE 8 do not know how to math.

In Chrome everything is ok.

In 4.3-beta everything was fine.

1. Open replacebyclass.html sample in IE 8.
2. Select content: from strike through three and up to the begging of the document.
3. Press Ctrl+C or Delete.

Result: Error: Invalid argument and Browser crash.

The same in inline.

IE9 and Chrome works fine.

Since 4.0.

Buttons in elementspath should not be draggable. Now you can drag it into editable. What's worse in IE10, IE11 it will cause following exception:

SCRIPT5009: 'CKEDITOR' is undefined 
replacebyclass.html, line 355 character 1

1. open any sample with CKEditor (i.e. samples/replacebyclass.html)
2. click somwehere inside editable
3. click move body elements path part into editable

additional info:

1. it should be easily fixable with setting onmousedown="return false" on anchors

1. open any sample with CKEditor (i.e. samples/replacebyclass.html)
2. use "Find" button from toolbar
3. (you'll have Find what input focused) start pressing space

Expected result:
Spaces should be placed at the beginning of input

Current result:
No spaces are placed (they may be placed at the end of the input though)

1. open any sample with CKEditor (i.e. samples/replacebyclass.html)
2. clear editor contents
3. insert new image
4. set following properties

• URL: ​http://i.imgur.com/Gj8iBon.png
• alternative text: altText
• border: 2
• hspace: 2
• vspace: 2

5. double click at image (in order to open properties dialo)

Expected result:
Image properties dialog should be shown.

Current result:
Browser crashes.

additional info:

1. sometimes you're able to repro it without setting alt text

TC 1 (general, IE9)

1. Open samples/replacebyclass.html
2. Insert page break.
3. Select inserted page break.
4. C-c
5. Change selection.
6. C-v

SCRIPT5007: Unable to get value of the property 'isBlock': object is null or undefined 
editable.js, line 1708 character 5

Works in 3.x.

TC 2 (IE11)

1. open any sample with CKEditor table plugin (i.e. samples/replacebyclass.html)
2. switch to source mode, and use following source code:

   <div style="page-break-after: always;"><span style="display:none">&nbsp;</span></div>
   <table border="1" cellpadding="1" cellspacing="1" style="width:500px">
   	<tbody>
   		<tr>
   			<td>
   			<p>aa</p>
   
   			<p>bb</p>
   			</td>
   			<td>&nbsp;</td>
   		</tr>
   		<tr>
   			<td>&nbsp;</td>
   			<td>&nbsp;</td>
   		</tr>
   	</tbody>
   </table>
   

3. switch back to wysiwyg mode
4. copy page break into first (top left) table cell

Expected result:
Operation should not cause exception.

Current result:
Exception is being thrown:

SCRIPT5007: Unable to get property 'isBlock' of undefined or null reference
File: ckeditor.js, Line: 323, Column: 45

Using current CKEditor standard editor demo on an iOS device (tested on iPad 2 with iOS 7.0.4):

1. Position the cursor on the first word in the first paragraph ("Apollo") which is formatted bold.
2. Observe the 'B' toolbar button is correctly show depressed.
3. Move the cursor somewhere else in the paragraph that is not bold.
4. Observe the 'B' toolbar button still shows bold, event though the cursor is not on bold text.

This problem seems to be consistent across all tools on the toolbar; they are not updating when the selection changes.

The problem seems to be that CKEditor is attaching to the DOM selectionchange event on the editable element (around selection.js:387) but according to ​http://help.dottoro.com/ljixpxji.php, the selectionchange event is only valid when attached to the document object (not any element).

I've worked around this problem outside CKEditor using something like this, which fixes the problem:

this.editor.document.on('selectionchange', function () {
  var selection = this.editor.getSelection(),
    path = selection && new CKEDITOR.dom.elementPath(
      selection.getStartElement(),
      selection.root
    );

    this.editor.fire('selectionChange', {
      selection: selection,
      path: path
    });
  }, this);

There are several things concerning the dialog that should be investigated/fixed/re-factorized:

1. Keyboard button labels are not consistent

   Editor Context Menu Press Shift+Ctrl+F10 or APPLICATION KEY to open context-menu. Then move to next menu option with TAB or DOWN ARROW. Move to previous option with SHIFT+TAB or UP ARROW. Press SPACE or ENTER to select the menu option. [...]

This is because some labels are hard-coded (upper case) and some are generated from langfiles (lower case). #11196 allowed translation of keys and we should take advantage of that feature. All hard-coded labels must be converted into dynamic ones.

2. We must verify whether help strings still match editor features. Some of them may be outdated/irrelevant/misleading. If necessary, some new ones should be written.

3. Some instructions are too complicated, some are unclear, e.g. lots of translators do not understand the MagicLine-related texts so we should revise the contents overall.

4. We must investigate a11y aspect of the dialog, e.g. compatibility with screen readers (JAWS). Perhaps <kbd> tags would be suitable here and there.

Related tickets we could also take into consideration:

• #8490: Plugins should extend a11yhelp to make the dialog dynamic and (build/editor)-dependent. At the moment a11yhelp creates a delusion that some features are enabled while, in fact, the plugins to which they belong are not even loaded.
• #9638: "Press ALT 0 for help" iframe voice label makes no sense when a11yhelp is not loaded (i.e. basic build).
• #10929: Spelling Mistakes on Accessibility Instructions dialog.

Hi, when you open SharedSpace demo

​http://ckeditor.com/ckeditor_4.3_beta/samples/plugins/sharedspace/sharedspace.html

and you add those styles

style="position: fixed; top: 0px; left: 0px; right: 0px; bottom: auto; z-index: 10000; width: auto;"

to div with id "top" it will work well for editable areas in div, but for textareas it's broken. Dropdown is moving along with page scroll.

​http://i.imgur.com/2cOkbBT.png

env.ie6Compat ... env.ie9Compat are deprecated since 4.0. After clean-up (#11422) these variables are rarely used:

env.ie9Compat -> 1 time
env.ie8Compat -> 2 times
env.ie7Compat -> 0 times
env.ie6Compat -> 0 times

They should be finally removed.

Steps to reproduce:

1. Add a text input to the editor, it’s most notable when the text input is placed beyond the size of the editor window
2. Click somewhere in the page outside of the editor so it doesn’t have focus
3. Right or ctrl + click on the text input element

When the input element is clicked, the focus jumps to the top of the editor window, and when adding a contextMenu listener this actually passes the top most element as the one being clicked.

Browser & OS: Google Chrome 32.0.1700.77 Mac OS X 10.9.1

The problem with quirks appeared after removing IE 6 & IE 7 support (#11422), because most of hacks now apply only to QM (instead of IE 7, IE 6 and IE QM). Even worse it is implemented in multipile ways (unsing "env.ie && env.quirks", "env.ie6Compat" or "document.documentMode").

I think that it would be misleading if we keep if ( env.ie6Compat ) in our code only because of QM (env.ie6Compat is true in QM) and it is better to replace it with if ( env.ieQuirks ).

Also after clean up we can save some bites (compressed ckeditor is 322 byts smaller with env.ieQuirks).

I'm not sure if there is a point in keeping env.quirks == true for non-IE browsers. Maybe env.quirks should be true only in IE 8 and IE 9 QM (so we will use just env.quirks instead of env.ieQuirks).

Steps to reproduce:

• Insert 3 links on the page, which URL "www.google.com" and Id: 1,3,3.
• Double click on the inserted links to check the Ids, only the first link Id is retained, all other links after that have empty Id.

Browser and OS: Window Firefox 26.0 & Chrome 32.0.1

In Simple Plugin (Part 1) guide, in Dialog Window Tabs part, after first code segment, there is a phrase, that tells reader, that result of changes in code segment can be seen immediately and that it will have two tabs in place.

Expected: As told Instead: There are no tabs, assuming because the tabs have no elements inside at that moment.

​http://docs.ckeditor.com/#!/guide/plugin_sdk_sample_1

If you create a blockquote, enter some text, highlight the text and then press the Delete key. The paragraph(s) inside the blockquote disappear and you can type directly into it.

For FF you need more than one line of text highlighted.

CKEDITOR.ui.dialog.select size option doesn't work due to css.

select.cke_dialog_ui_input_select {
	height: 24px;
	line-height: 24px;
	background-color: #FFF;
	border: 1px solid #C9CCCF;
	border-top-color: #AEB3B9;
	padding: 2px 6px;
	outline: none;
	-moz-border-radius: 3px;
	-webkit-border-radius: 3px;
	border-radius: 3px;
	-moz-box-shadow: 0 1px 2px rgba(0,0,0,.15) inset;
	-webkit-box-shadow: 0 1px 2px rgba(0, 0, 0, 0.15) inset;
	box-shadow: 0 1px 2px rgba(0, 0, 0, 0.15) inset;
}

height parameter is limiting select area.

Dear Team,

In CKEditor Inline when you make text selection and then right-click on it, context manu appears and selection is lost.

See e.g. #11295. Calling editor.destroy() right after (or before it's completed) CKEDITOR.replace() or editor.setData() causes errors.

That's because developer has to care about calling methods when previous finished. There are callbacks and events which notify developer when action is completed, but it's his duty to find which method can be called when.

However, not every developer understand problems with asynchronous methods and not in every case it's easy to handle this. For example when editor is loaded in one tab of some UI component and user switches between tabs too quickly, destroy() may be called before editor fully initialises.

On the other hand, if editor's methods would take care of that (e.g. destroy() would wait until initialisation finished), then API would start to work unpredictably. Developer would never know if destroy() will be done immediately or if it's going to wait until something (setData, initialisation, etc.) ends. This may be even worse situation than the current one.

We could make even longer step and make all editor's methods asynchronous and e.g. based on promises. Then everything would be fine... if you understand all of this :D.

Therefore, instead of forcing developers to understand some not trivial concepts to do basic stuff we can simply clarify this in docs. But there still will be cases in which it will be hard to handle editor state (like the tabs case).

Opinions needed.

More details about the current situation in ​http://dev.ckeditor.com/ticket/11502#comment:8

Cases we need to solve

1. Integration with styles applicable to widgets. See #11297.
2. Disabling ACF in body of some widgets - e.g. mediaembed. See #11737 and #10925.
3. Filtering pasted widgets. See #11115.

Cause of problems

1. ACF filters data before widgets are upcasted. Therefore for example mediaembed cannot notify ACF that children of some elements should not be filtered and it's not possible to create allowed content rule based on a widget name and style properties allowed on that widget element.
2. In pasted or dropped content we have the upcasted and initialized HTML of a widget. Therefore we cannot filter it, because ACF is applied to the data format.

Solution

ACF after upcasting

The advantage of this solution is that ACF would know everything about widgets, so it could make precise decisions. Additionally, while filtering pasted content there would be no problem at all, because it would be the same case.

However, there are two problems which makes this idea incorrect:

• Assuming that we've got some complex widget like image2, which makes a lot of transformations while upcasting, how would ACF filters the upcasted image? How could it strip, without breaking widget, figcaption and figure if they are not allowed? It's simply impossible, without some special actions from the upcast method. It could theoretically ask ACF if specific components of widget are allowed, but it's tricky and still doesn't solve this issue completely.
• Moreover, when ACF would filter upcasted version of widgets it would be impossible or a lot harder to define ACF rules. One would have to know internal widget structure to define those rules. In my opinion this makes the idea totally wrong. ACF filters data format and that's a foundation we have to keep.

ACF filters data

Previous section proved that ACF needs to be applied to data, not to inner HTML. This means that the current way of processing is the only correct one, but on the other hand we still have those three cases, which are listed at the beginning, that have to be solved.

1. Integration with styles applicable to widgets - a solution has been proposed in ​http://dev.ckeditor.com/ticket/11297#comment:8. It may not be 100% precise in some cases, but there's another ACF's characteristic which we have to keep in mind. It's not meant to be a security filter. It's meant to ensure good UX in as many cases as possible. Therefore when uncommon constellation of element and its attributes slips through its hands it's not a blocker.

2. Disabling ACF without a possibility to communicate through element's attributes. In this case the only possible solution is a callback executed by ACF which will notify it about how it should proceed. For example mediaembed will add a callback which will tell ACF that all descendants of element should be skipped if element has mediaembed's class. Callback will return binary flags, because we may need more options in the future. Example of flags that we may need to implement:

• DISABLE_ON_ELEMENT - skips only this element (continues on descendants),
• DISABLE_ON_DESCENDANTS - skips descendants,
• CONTINUE_ITERATION - continue iteration - used with DISABLE_ON_DESCENDANTS could inform ACF that it should look for a descendant that reenables filtering,
• ENABLE - reenable filtering.

But of course for now we'll need just one of these flags, so only one should be implemented.

Ticket: #11737.

3. Filtering pasted widgets. It's not possible to filter widget's upcasted version - only its source version can be filtered. If we had full control over clipboard we could downcast selected data on cut/copy and then filtering on paste would work like filtering on data load. Unfortunately, we don't have full control over clipboard and won't have even after #11437, because of IEs.

Therefore, the only solution might be to look for upcasted widgets in pasted data, extract their HTML, find out from which editors they came from, use those editors to downcast them, and replace their upcasted version with downcasted one in pasted data. This could be based on unique editors' ids and data-cke-widget-editor-id attributes marking widget wrappers. This won't work when copying between window tabs or frames or when editor instance has been destroyed in the meantime and I'm not sure what should happen in such situation.

Similar solution, maybe more straightforward and more reliable would be if editor, on cut/copy/dragstart if it cannot override the content, could retrieve widgets' HTML from selection, downcast them and store in the CKEDITOR object, so all other editors will be able to replace pasted widgets with their downcasted HTML stored in that object. Of course widgets have to be marked with unique ids to make this possible. This option is better than the previous one because widget's HTML will be kept even after editor is destroyed. We can also reuse the same mechanism in drag and drop for IEs in which we will need to store the data for a while. As the mechanism will be public 3rd party code will be able to integrate with it.

And the third solution, nearly perfect, is to control copy and cut and override data with HTML extracted by us and processed by htmlDP.toDataFormat. We will need to do the same for drag&drop, so it won't be a big problem to eventually make the same with copy/cut. Additionally, we would automatically solve the problem with Blink putting awful HTML into clipboard.

Descriptive summary: When you move the focus out of the image area and then do a right clic on the image, Image Properties option is not displayed in the context menu

• I just erased a very detailed steps to reproduce and other information because akismet rejected my submit lots of times -

This would help in debugging,

• when a user says "Feature X does not work as expected".
• you updated a plugin, but are not sure if it is in the browser yet or still cached somewhere server or locally.

I guess this is also the reason for any software's version string in about-dialogues.

PS: In my initial ​ request in the forum i also asked is there a way to get plugin versions on the JS console of the browser inspector? which is currently unanswered...

Using current standard CKEditor demo:

1. Select all the content in the demo editor and delete it.
2. Enter some text (e.g. "original").
3. Press Ctrl-A (or Cmd-A on Mac) to select all the text.
4. Hit the Bold button to bold the text.
5. Press Ctrl-A again to re-select all (even if it is already selected).
6. Type in some new text (e.g. "new"), which will replace the original text.

On Chrome and Safari, the new text is bold (i.e. it retains the styling of the original text). On Firefox and IE11, the new text does not retain the original styling; it just gets the default styling.

I think the Chrome and Safari behaviour is correct, so Firefox and IE11 should work the same way.

Related issues: #11162, #8439, #8440.

Tested on the demo using the BBcode plugin to confirm the bug. ( ckeditor.com/demo#bbcode )

When adding some line breaks between inline text and a block element, each time the user switch to source code, then switch back to preview. One line break disappear.

To reproduce the bug, use the demo (link above). Switch to source code and paste this code :

This is some [b]sample BBCode text[/b]. You are using CKEditor.





[list=1]
[*]test
[*]test
[*]test
[/list]



test test test



[quote]Test content[/quote]




[quote]Test content[/quote]

Switch back to preview, then switch back and forth to see that the line breaks between the first line of text and the list are disappearing and also the line breaks between the line "test test test" and the first quote, and finally also the line breaks between the two quotes.

It doesn't only happen when switching back to source. It has the same behavior when saving the content. If you save the content in a database, each time you save you lose some of the line breaks exactly as if you were switching to source then back to preview before saving.

I would vote to extend templates and make them dynamic through the usage of popup dialog helper. The dialog would:

• contain user-configurable form
• go through validation & error display
• implement ui dependencies
• return html from a function, given formState hash.

This would allow a web site builder to provide client with smart templates that would react to web editor input.

Enclosed please find a very rudimentary example of what it would do. Validation and error handling aspects are not polished at this moment - subject to further development upon feature approval.

Steps to reproduce:

1. Visit a url where it ends with a # and you have a ckeditor instance (e.g. www.example.com#)
2. Click on a menu option that has a flyout menu (e.g. fonts)

Result: You'll see that the flyout gets anchored to the top of the page Note: This only happens the first time you click on the option for flyout. All subsequent clicks anchor the flyout correctly.

This happens with barebones install of ckeditor 4.0+ & usage of CKEDITOR.replace

My environment:

• OSX 10.8.5
• Chrome 32.0.1700.102

Dear Team,

In Find and Replace, "Match Whole Word" is not working.
when typing the whole word and searching it shows

The Specified Text was not found

It works fine for "Match Case" & "Match Cyclic"

Kindly check with the demo site as mentioned below[[BR]]
http://ckeditor.com/demo#full

Please guide me to resolve this problem.

Thanks in Advance.

When config.allowedContent = true; is set, the editor throws a JS error when inserting content (such as an image) into any unknown element. A simple test case is this:

<nobr>a</nobr>

Steps to reproduce: With the above content entered into the editor, set the cursor focus to be inside the unknown tag, then insert an image.

Inside core/editable.js, the function extractNodesData throws an error trying to look-up the allowed names for this element.

This is invalid HTML, so it is a user error, but the config.allowedContent = true; prevents it from being filtered out (as it should). We've seen it pop-up when users are copy and pasting content from other sources. So at a minimum, I think the editor should handle this case in some way.

My suggestion is when allowedContent is set to true, the editor should treat all unknown tags as a semantically meaning-less tag, perhaps inheriting the allowedNames from DIV elements.

I would be willing to submit a fix with the above suggestion, or implement another suggested fix if needed.

Thanks

Chrome error message:

Uncaught TypeError: Cannot read property 'img' of undefined ckeditor.js:319
c ckeditor.js:319
(anonymous function) ckeditor.js:326
CKEDITOR.editable.CKEDITOR.tools.createClass.proto.insertHtml ckeditor.js:300
(anonymous function) ckeditor.js:306
h ckeditor.js:10
CKEDITOR.event.CKEDITOR.event.fire ckeditor.js:11
CKEDITOR.editor.CKEDITOR.editor.fire ckeditor.js:13
CKEDITOR.tools.extend.insertHtml ckeditor.js:225
$.extend.editor.paste

Edit: This issue is about CKEditor throwing error if DTD was not extended.

I have noticed that there is no plugin for inserting footnotes into text using CKEditor although I found a Drupal module at ​https://drupal.org/project/footnotes bundled with CKEditor which has a footnote plugin.

I also found a DokuWiki module called ckgedit at ​https://www.dokuwiki.org/plugin:ckgedit used for editing bundled with CKEditor including a footnote plugin.

So one of these footnote plugins could be merged into the main release of CKEditor.

Setup:

• OS-X 10.9.1
• Chrome 32.0.1700.107 or Safari 7.0.1 (9537.73.11)

Steps te reproduce:

1. Include the image2 plugin in your CKEditor build.
2. Open /ckeditor/samples/replacebyclass.html
3. Create a table with 2 rows and 2 columns (2x2).
4. Put an image widget in the first cell, and some dummy text in the other three cells.
5. Open a context menu on the first cell (containing the image) and select 'Merge Right'.

The following javascript error occurs:

Uncaught TypeError: Object #<error> has no method 'getDocument' plugin.js:402 placeCursorInCell plugin.js:402 createDef.exec plugin.js:767 exec command.js:52 CKEDITOR.tools.extend.execCommand editor.js:822 CKEDITOR.menu.CKEDITOR.tools.createClass._.onClick plugin.js:193 (anonymous function) tools.js:573 (anonymous function) tools.js:573 (anonymous function) plugin.js:361 (anonymous function) tools.js:659 CKEDITOR.tools.callFunction tools.js:687 onclick replacebyclass.html:1

Tested with FF25 and 27.

Reproducalbe in demo

Type "AAA"
Enter
Backspace
Shift+Enter

--> expected: new line, but it make p

Backspace

--> cursor is on line under "AAA"

Type "B" -> In Line 1 is "AAAB"

With "down" u can reproduce it.

Extracted from ​http://dev.ckeditor.com/ticket/11439#comment:16

Due to invalid focus in anchor dialog we can experience few issues. You're not able to:

• focus other fields by pressing tab
• close anchor dialog with esc

1. open any sample with CKEditor (i.e. samples/replacebyclass.html)
2. using Source button set html to:

   <p><a href="#foo">bar</a></p>
   

3. return to wysiwyg mode
4. double click link bar
5. press esc key

Expected result:
Dialog should be closed (as it is for typical links)

Current result:
It's not closed unless you'll focus any input control inside of it.

additional info:

1. Tested in Opera (blink), IE10, FF 27.0.1 @ Win8
2. Issue was also present in 4.0 release (didn't check earlier releases)

TC1:

1. Open image2 sample.
2. Doubleclick image.
3. Change the width to 300.
4. See that height was set too although it was empty at the beginning. I'd expect that it was left empty.

Notice that in step 2 "keep ratio" is on. I think that it's ok, because it tells user that ratio will be kept if he changes size, which is true because the missing dimension is calculated automatically.

TC2:

1. Open image2 sample.
2. Doubleclick image to verify that only width is set. Close the dialog.
3. Resize image using handler.
4. Doubleclick image.
5. See that both dimensions are set now.

• Edit: TC2 is not about reading dimensions by dialog but about editor calculating both dimensions. In HTML when only one dimension is set the other one is automatically calculated by browser so we should not interfere it.

See attached image.

It's a Firefox bug: ​https://bugzilla.mozilla.org/show_bug.cgi?id=965966