0.2.4 commit breaks highlighting of first and last incorrect words in rich text editors / HTML

[alang]

Start a new line and add 3 misspelled words, only the 2nd word will be highlighted.

Issue seems to be related to wrapping p tags, as redactor demo does not wrap lines in p tags. Happens each paragraph.

WebService correctly responds with incorrect words.

[markose]

The problem is in method HtmlParser.prototype.highlightWords:

The author forgot the ^| at the beginning and the |$ at the end of the expression.

Just replace lines 587-590:

var regExp = '';
regExp += '([^' + letterChars + '])';
regExp += '(' + incorrectWords.join('|') + ')';
regExp += '(?=[^' + letterChars + '])';

with:

var regExp = '((^|[^' + letterChars + ']))(' + incorrectWords.join('|') + ')(?=([^' + letterChars + ']|$))';

and line 592:

this.replaceText(new RegExp(regExp, 'g'), element[0], this.highlightWordsHandler(incorrectWords), 2);

with:

this.replaceText(new RegExp(regExp, 'g'), element[0], this.highlightWordsHandler(incorrectWords), 3);

and it will work properly

Regards,
Marko

[johan-ludvigsson-stratsys]

That did indeed fix the problem. Thanks!

[markose]

You're welcome :-)

On Thu, Feb 21, 2013 at 11:43 AM, johan-ludvigsson-stratsys <
notifications@github.com> wrote:

That did indeed fix the problem. Thanks!

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/26#issuecomment-13882877.

[vladotg]

@markose - I tried to use your fix and it works except in one case: for example, if you have two lines of text and mispelled word at the end of first line, the word wont be highlited.
for example:

This is the first sentensce
This is the second sentence.

I purposely didn't write dot in first sentence. If I write dot, it will work.

[johan-ludvigsson-stratsys]

@vladotg @markose I can also confirm this, but as we are looking into it, it seems to be more related to _getText in the replaceTextInDOM method, which just appends text as it goes. This works as long as there only are inline-elements such as span in the text, but for p-tags it does not work very well. So this is a two-part bug I would say

[vladotg]

If you manage to solve this, please write here.

[alang]

@johan-ludvigsson-stratsys I also got as far as that method in trying to fix this bug. From my (quick) tests 0.2.3 does not have this problem. One of these two commits (or both) seem to of caused this regression:
2454c12
14a1d95

[peter-klaesson-stratsys]

I added some tests for the cases where using p and li elements. I located the problem for when stripping the tags. The p and li elements should break the words, that was not the case.

[willy-michel]

We are using your plugin with CKEditor 4 and have the same problems with first and last words. Is it already planned to fix this issue? Or is there a workaround for this?

[badsyntax]

Apologies for the belated response and thanks for your patience. I have been on holiday for the past 4 weeks. I believe this issue is related to the regex as @markose pointed out. I will work on this issue over the weekend.

[willy-michel]

Thanks for your reply. I tested the fix from @markose and it works fine. We can use it until the issue is fixed in the next release. Thank you.

[badsyntax]

Please can you help me test the latest changes which should accomodate these issues.

You can either test by pulling in the latest code from the develop branch, or just download this version: https://github.com/badsyntax/jquery-spellchecker/blob/develop/dist/js/jquery.spellchecker.js