TStringList Delimiter Char

[jcmontherock]

Could we use any characters as string delimiter ? I try '§' without success.

[jamie]

That is unicode, its more than a single character.

I believe there is a UTF8StringList somewhere in the collection, that may do it.?

[jamie]

Just looking and found TstringlistUtf8GFast, but that does not seem to allow me to set a Utf8Char, instead its still char.

You could use   yourString.Slip, that should work.

[Bart]

Just looking and found TstringlistUtf8GFast, but that does not seem to allow me to set a Utf8Char, instead its still char.

You could use   yourString.Slip, that should work.

Problems with your keyboard, or with your coordination 

Bart

[jcmontherock]

What is: String.Slip ?

[TRon]

I think that jamie wanted to write yourString.Split  (but somehow managed to flip his keyboard during the process)

[Bart]

What is: String.Slip ?

String.Split

Code: Pascal  [Select][+][-]

1. uses
2.   SysUtils;
3. var
4.   Arr: TStringArray;
5.   S: String;
6.   i: Integer;
7. begin
8.   S := 'abc§def§ghi§jkl§';
9.   Arr := S.Split('§', TStringSplitOptions.ExcludeLastEmpty);
10.   for i := Low(Arr) to High(Arr) do writeln(Arr[i]);
11. end.

This will split the string S, using '§' as the separator.
Split accepts both chars and strings as separator.
TStringSplitOptions.ExcludeLastEmpty ensures that the last value will not be an empty string.

Bart

[Zvoni]

Errrr..... someone will have to explain to me, how "§" (=ASCII 167) is Unicode/UTF8?!?!?!

[dseligo]

Errrr..... someone will have to explain to me, how "§" (=ASCII 167) is Unicode/UTF8?!?!?!

It is U+00A7:

Code: Pascal  [Select][+][-]

1. var
2.   s: String = '§';
3.  
4. begin
5.   WriteLn(Length(s)); // 2
6.   WriteLn(Ord(s[1])); // 194
7.   WriteLn(Ord(s[2])); // 167
8. end.

P.S.: I tried this in Lazarus 3.2.

[wp]

someone will have to explain to me, how "§" (=ASCII 167) is Unicode/UTF8?!?!?!

Only characters below #128 are 1 byte. Anything above may depend on the code page. #167 is `§`on the CP125x code pages, but on CP437 (original IBM) or 850 (Latin 1) it is 'º', and on CP852 (Latin 2) it is 'ž'. Therefore, it consists of at least two bytes in UTF8. A nice tool is the Lazarus character map in menu Edit.

[Zoran]

Errrr..... someone will have to explain to me, how "§" (=ASCII 167) is Unicode/UTF8?!?!?!

I'll try.

In short:

It is not ASCII 167.
This character is encoded as (in hex):
  - a7 (167 decimal) in cp1250 (not ASCII!),
  - 00 a7 in UTF-16
  - c2 a7 in UTF-8
  - has no representation in ASCII


Longer explanation:

This is not ASCII 167. There is no ASCII 167.
ASCII is 7-bit encoding. There are no values above 127 in ASCII!
There are several so called ANSI 8-bit encodings (code pages), also known as Windows code pages. You probably mixed up one of these. This character is indeed encoded as 167 in cp1250 (sometimes called win-1250) which is the ANSI page used for Croat alphabet, as well as for several other east European latin languages -- Check, Slovak, Hungarian, Polish... For other languages there are other ANSI code pages, such as cp1252 for west european latin languages, cp1251 for cyrilic languages, cp1253 for Greek, etc.
It is not ASCII!

All these 8-bit ANSI encodings are compatible with ASCII in the first 128 characters, which have values below 127. Every ANSI page, such as cp1250 mentioned above, have the same first 128 characters (0-127); these are taken from ASCII. However, they differ in values above 128 (in these position above 128 -- cp1250 has letters used in east european latin languages, such as č, ć, đ ..., cp1252 has letters used in west european languages, such as ü, ö, ç, ..., cp1251 has cyrilic letters there, etc.).

Furthermore, these ASCII values up to 127 are also encoded the same in utf-8, but these are only characters which have one-byte encodings in utf-8.

Any character which appears in some ANSI (not ASCII!) encoding with a value 128 or above (that is, which has the bit 7 set), such as § which is encoded as 167 in cp1250, is represented in utf-8 with at least two bytes.

[dseligo]

It is not ASCII 167.
This character is encoded as (in hex):
  - a7 (167 decimal) in cp1250 (not ASCII!),
  - 00 a7 in UTF-16
  - c2 a7 in UTF-8
  - has no representation in ASCII

And in command prompt of my Windows 11, § sign has value 245 (I have code page 852).

[jcmontherock]

Thanks to everybody.
I finally found that a character like '§' or '°', in windows 11 running in utf-8 are not a really a char, using 1 byte.