ckeditor inserts hidden characters into HTML when using soft return (shift-return). Hex codes: E2 80 8B

[Vern Jensen]

When using shift-return to add a soft return to ckeditor (tested on both versions 4.5.3 and the latest 4.5.10), there are hidden characters that are inserted. Their hex codes are:

E2 80 8B

If you insert many soft returns, these hidden characters tend to 'bunch up' all on one line. For instance, if I type the following all with soft returns between each:

asdf sdfg dfgh fghj

then the following hidden hex codes will be inserted after the *last* line (after "fghj"):

E2 80 8B E2 80 8B E2 80 8B E2 80 8B

(i.e. E2 80 8B repeated 4x, one for each soft return)

To duplicate:

1) Go to the online ckeditor demo: ​http://ckeditor.com/demo

2) Type a few characters (like "asdf") on several lines, using shift-return to do a soft return between each line.

3) Click on the "Source" button to view the HTML. Select and copy an area large enough to more than fully contain the area you typed (make sure to copy a little extra *below* what you typed). Then paste it into a hex editor. There is one online at:

​https://hexed.it/

You will see the same hex codes:

E2 80 8B

have been added as strange characters, one set for every soft return.

OS: Windows 7 Browser: Chromium (also same behavior duplicated in Qt's WebEngine, which uses the same source as Chrome)

[Jakub Ś]

Please see this link - ​http://docs.ckeditor.com/#!/guide/dev_browsers

We support Chrome not Chromium. Anyway I have tried reproducing this issue in Chromium on Linux and Chrome on Windows and everything seems to be fine. These codes were not entered.

1. Can this issue be reproduced in Chrome browser?
2. Could you try doing the same in plain contenteditable element? Perhaps there is an issue in contenteditable?
3. The E2 80 8B is a Unicode Character for 'ZERO WIDTH SPACE' which is inserted so that it was possible to write inside the editor. Please see ​https://dev.ckeditor.com/ticket/10031#comment:18
   How are you getting data from editor? Are you using editor.getData() or something else?

[Vern Jensen]

3. To get data I am using

editor.document.getBody().getHtml();

[Vern Jensen]

1. I tried in Chrome and couldn't reproduce, so I tried again in Chromium... and now can't reproduce there either!! This issue IS real... but I'm not sure what's different about conditions now vs. a couple days ago. I'm doing the exact same steps.

2. Do you mean a plain <textarea> element? (contenteditable is an attribute... and can be applied to any element... but perhaps that's what you meant?)

[Vern Jensen]

Also, based on the link to ticket 10031, this IS a bug -- because apparently the ZWS character is supposed to NOT be included when you get output from ckeditor using

editor.document.getBody().getHtml();

(or should I be using getData() instead?)

[Piotrek Koszuliński]

(or should I be using getData() instead?)

Yes, innerHTML !== data. There are more difference between them – especially when you use widgets. So the only way to get a clean data is through editor.getData().