Greek characters are stored in two different formats?

Question

I think there is a problem with the encoding of Greek characters on this Q2A site, look at this post:

learn-greek-online.com/ask-greek/1128/answering-the-telephone-in-greek

In the Question you will see a string in Greek: "Ιάκοβος" (right after "Thank you").  It is stored in UTF-8 (I think) as it appears exactly like this in the source code of the page: "Ιάκοβος".

Now go to the Answer of this post.  You will notice a string: "Hello Ιάκοβος".  This is stored with "HTML character codes", if you go to the source code you will see:

Hello Ιάκοβος

So the Greek name Ιάκοβος is stored in two different encodings in the same page.  How can I change this?  I want the simple utf-8 used everywhere.

BTW, in this q2a site, it works fine, look at this thread:

http://www.question2answer.org/qa/67670/greek-characters-displayed-properly-notification-emails

The letter "Λ" has the correct encoding (utf-8) in the question, answer, and comment section:

test (question): Λ
test (comment): Λ
test (answer): Λ

Comments

Some test in Greek (Ελληνικά). From word doc.

---

Some test in Greek (Ελληνικά).  From word text editor.

---

Some test in Greek (Ελληνικά). Bolded.

Answer 1

Hi

After some more tests I found this: 

Only posts with formating (eg BOLD text) have a problem.  When I remove the bold, the Greek chars appear normal in the source code.  However this only happens on my website.  When I tested it here, the source code is always correct. 

So there must be a configuration problem on my website.

Answer 2

I read this related bug report/fix about CK-Editor here:
https://www.imediainc.com/blog/february-2017/prevent-html-encoding-of-greek-characters-in-ck

So I added this line in the file qa-plugin/wysiwyg-editor/ckeditor/config.js :
config.entities_greek = false;

Unfortunatelly it did not fix the problem.