Yes, I can post something that shows the behavior.
But I have found that my changing only the English font doesn't resolve things completely. Just yesterday, and after about twenty reloads, MSFTEDIT kicked in and changed about 30% of the English characters.
I am going to go to symbol fonts for everything... and I can only hope that it will cause RichEdit to leave things alone. It is my idea that using Unicode causes MSFTEDIT to do Font Binding. I have found numerous complaints about Font Binding on the Internet, and I have assumed that it is all Unicode related.
RichEdit does this with out respect to a font's quality. It does it with the standard fonts that are packaged with the Microsoft system. When I enter English text I don't do any manipulations to the keyboard, and when I do so for the other languages, I only change the key code... no RTF inserts.
Font Binding behaves very much like a computer virus. It monitors your behavior, and at some point it activates. That my document was fine until about the 20th time it was loaded is evidence that Font Binding is unpredictable (on my part) and very difficult circumvent. Once it even switched fonts on a complete paragraph that was in English. Yet sometimes it attacks Greek as well. Oddly as well, it has never messed with the Hebrew or Syriac (and the same developer had built both the Hebrew and Greek fonts).
In my opinion, RichEdit is unmarketable for Unicode applications with mixed languages. By unmarketable I mean... my application couldn't be given away at no cost. No one could rely on its product. Even if their document looked good, they wouldn't know that someone on another computer would see the same fonts (even if they used your program). You can't even be confident that they will see the same language. Nevertheless, I hadn't intended to sell it. I have always planned to give it away to everyone.
RichEdit with Unicode is only reliable if your entire document is one language and one font... and it is only on account of Font Binding.
The following is a typical post that I found on an Internet blog.
Andrew West on 11 Jul 2007 7:45 AM:
There are a couple of reasons why I personally think that an option to show characters with fallback/substitution/linking applied would not be such a cool idea.
Firstly, I'm sure that I am not the only person who loathes the way Microsoft's font substitution works. When I select a specific font to render my text with then what I really want is the application to respect my choice of font and not act as if it knows better then me and use some different font.
For example, in Notepad or Word if I select Code2000 to display some Syriac text it will be displayed with Estrangelo Edessa and there is no way I can get it to use Code2000 even though Code2000 supports Syriac just as well as Estrangelo Edessa does (and even worse in Word the font dropdown box shows Code2000 as the selected font when you click on the text, deceiving users into thinking that they have successfully
applied the font to the text when they haven't). If there was a registry setting where the user could define the font to use as a fallback for each script (or if there was an option to disable font substitution at the application level) then it wouldn't be so bad, but as it is, from a user's perspective the font substitution behaviour is unpredictable and often unwanted.
Secondly, font substitution behaviour is application-specific. Microsoft apps such as Word and Notepad do apply font substitution and linking, but many non-Microsoft apps such as OpenOffice.org and BabelPad do not. Thus if charmap showed characters with fallback/substitution/linking applied this would not necessarily be what you will get if you select the same font in some other application.
Incidentally, when you told us in September (http://blogs.msdn.com/michkap/archive/2006/09/03/737553.aspx) that charmap on Vista would still not display anything beyond the BMP and would still only display the
very limited list of Unicode blocks (nothing beyond Unicode 2.0 for @%#'$ sake !!!) I was stunned. How hard could it possibly be to update charmap to show everything up to and including Unicode 5.0 ?
So I think that a far more useful feature for charmap would be the ability to display all currently defined Unicode characters.
His remark about the font changing, but the font box still shows it by the original name, is also something that I have experienced. It's a sanctioned virus.
Rick