How to get a single Arabic letter in a string with its Unicode transformation value in DELPHI?

Question

Considering this Arabic word(جبل) made of 3 letters .

-the first letter is جـ, -name is (ǧīm), -its Unicode value is FE9F when its in the beginning, -its basic value is 062C and -its isolated value is FE9D but the last two values return the same shape drawing ج .

Now, Whenever I try to get it as a single character -trying many different ways-, Delphi returns the basic Unicode value. well,that makes sense,but what happens to the char with transformation? It is a single char too..Looks like it takes the transformed value only when it is within a string, but where? how to extract it?When and which process decides these values? Again the MAIN QUESTION: How can I get the Arabic letter or its Unicode value as it is within a string?

just for information: Unlike English which has tow cases for its letters(Capital and Small), Arabic has four cases(Isolated, Beginning,Middle And End) with different rules as well.

Answer 1

I'm not sure I understand the question. If you want to know how to write U+FE9F in Delphi source code, in a modern Unicode version of Delphi. Do that simply like so:

Char($FE9F)

If you want to read individual characters from جبل then do it like this:

const
  MyWord = 'جبل';
var
  c: Char;
....
c := MyWord[1];//this is U+062C

Note that the code above is fine for your particular word because each code point can be encoded with a single UTF-16 WideChar character element. If the code point required multiple elements, then it would be best to transform to UTF-32 for code point level processing.

---

Now, let's look at the string that you included in the question. I downloaded this question using wget and the file that came down the wires was UTF-8 encoded. I used Notepad++ to convert to UTF16-LE and then picked out the three UTF-16 characters of your string. They are:

U+062C
U+0628
U+0644

You stated:

The first letter is جـ, name is (ǧīm), its Unicode value is U+FE9F.

But that is simply incorrect. As can be seen from the above, the actual character you posted was U+062C. So the reason why your attempts to read the first character yield U+062C is that U+062C really is the first character of your string.

---

The bottom line is that nothing in your Delphi code is transforming your character. When you do:

S[1] := Char($FE9F);

the compiler performs a simple two byte copy. There is no context aware transformation that occurs. And likewise when reading S[1].

---

Let's look at how these characters are displayed, using this simple code on a VCL forms application that contains a memo control:

Memo1.Clear;
Memo1.Lines.Add(StringOfChar(Char($FE9F), 2));
Memo1.Lines.Add(StringOfChar(Char($062C), 2));

The output looks like this:

As you can see, the rendering layer knows what to do with a U+062C character that appears at the beginning of the string.