Unicode languages and text selection...

[Paul Dupuis]

In moving a commercial application that allows researchers to perform a 
lot of textual analysis  from LC6.7.11 to LC9.0.x, we have been trying 
to make sure we take advantage of "Unicode Everywhere".

In do in so, we test with text in as many different languages 
(representative of different glyphs/alphabets and sets of Right-To-Left 
(RTL) and LTR languages). We certainly do NOT have staff that know even 
a fraction of the languages. We craft test content using Google 
Translate. In doing this, we have discovered and reported a few long 
standing Unicode related bugs in the current release of LiveCode, some 
going back to LC7.Some examples:

https://quality.livecode.com/show_bug.cgi?id=22249
https://quality.livecode.com/show_bug.cgi?id=22213
https://quality.livecode.com/show_bug.cgi?id=22202

  I trust that all will eventually get fixed, hopefully sooner rather 
than later. We have now seen another oddity that we can not figure out 
if it is an intentional behavior of certain languages in Unicode or a bug.

  In certain languages in fields, Arabic for example, when we select 
portions of the text in a field, certain glyphs and even words seem to 
change. Not knowing the language, we can't be sure. Has anyone else seen 
this odd behavior in any language (seems to appear more in RTL 
languages) and is it a characteristic of Unicode for some languages or 
some sort of bug?