How to save greek letters to a variable?

[YiannisKam]

Code: Pascal  [Select][+][-]

1. program Hello;
2.  
3. {$mode objfpc}{$H+}
4.  
5. uses
6.   windows, sysutils;
7.  
8. var
9.   name: utf8String;
10.  
11. begin
12.   SetConsoleOutputCP(CP_UTF8);
13.  
14.   writeln('Πως σε λένε φίλε μου;');
15.   readln(name);
16.   writeln('Καλημέρα φίλε ', name);
17.   readln;
18. end.

I'm not sure if utf8String is the appropriate type in this case but neither the String type works. Everything else is displayed just fine. Is there a solution?

[jamie]

Maybe you should include also "SetConsoleCP" , too ? that sets the input.

[YiannisKam]

I've tried that also. No progress.

[paweld]

With LConvEncoding - you must add LCLBase to project requiments: 

Code: Pascal  [Select][+][-]

1. program Hello;
2.  
3. {$mode objfpc}{$H+}
4.  
5. uses
6.   windows, sysutils, LConvEncoding;
7.  
8. var
9.   name: String;
10.   b: Boolean;
11.  
12. begin
13.   SetConsoleOutputCP(CP_UTF8);
14.   writeln('Πως σε λένε φίλε μου;');
15.   readln(name);
16.   name := ConvertEncodingToUTF8(name, GetConsoleTextEncoding, b);
17.   writeln('Καλημέρα φίλε ', name);
18.   readln;
19. end.

[wp]

Yes, this is not very intuitive... The only way I could get the 2nd string right was doing an explicit conversion by means of the routines in unit LConvEncoding (requires package LazUtils in case of console application) because the input codepage is 850 (setting it to UTF8 results in #0 strings after the ReadLn...). And of course - since the codepage is 850 - I cannot enter Greek characters here...

Code: Pascal  [Select][+][-]

1. program hello;
2.  
3. {$mode objfpc}{$H+}
4.  
5. uses
6.   windows, SysUtils,
7.   LConvEncoding;  // Add "LazUtils" to project requirements
8.  
9. var
10.   name: String;
11.   cp: Integer;
12.   cpStr: String;
13.  
14. begin
15.   SetConsoleOutputCP(CP_UTF8);
16.   cp := GetConsoleCP;
17.   cpStr := 'cp' + IntToStr(cp);
18.   writeln('Πως σε λένε φίλε μου;');
19.   Write('Enter some text: ');
20.   readln(name);
21.   name := ConvertEncoding(name, cpStr, 'utf-8');
22.   writeln('Καλημέρα φίλε ', name);
23.   Write('Press ENTER to quit...');
24.   readln;
25. end.

[EDIT]
paweld was faster ...  But don't add LCLBase to the requirements because this pulls in the LCL. Better to use LazUtils which does not depend on the LCL.

[paweld]

But don't add LCLBase to the requirements because this pulls in the LCL. Better to use LazUtils which does not depend on the LCL.

Thank you for the information. I didn't know about that.

[marcov]

Doesn't enabling utf8 in the application manifest screen work for you?

[YiannisKam]

I've tried everything but nothing works so far

[wp]

Delphi has the same behaviour. This is my test project usable in both Lazarus and Delphi:

Code: Pascal  [Select][+][-]

1. program hello;
2.  
3. {$IFDEF FPC}
4.  {$mode objfpc}{$H+}
5. {$ENDIF}
6.  
7. uses
8.   windows, SysUtils;
9.  
10. var
11.   name: String;
12.   i: Integer;
13. begin
14.   SetConsoleOutputCP(CP_UTF8);
15.   SetConsoleCP(CP_UTF8);
16.   writeln('Πως σε λένε φίλε μου;');
17.   Write('Enter some text: ');
18.   readln(name);
19.  
20.   Write('Bytes of "name":');
21.   for i := 1 to Length(name) do Write(Format(' $%.2x', [ord(name[i])]));
22.   WriteLn('; Length(name)=', Length(name));
23.  
24.   writeln('Καλημέρα φίλε ', name);
25.   Write('Press ENTER to quit...');
26.   readln;
27. end.

It looks as if Windows accepts only ASCII character (< #128) for input when SetOutputCP(CP_UTF8) has been called; other characters are replaced by zero.

[YiannisKam]

Yes, as it seems the string is zeroed out

[marcov]

I also experimented (with non-greek results), and it seems that this is a problem of readfile(), for unicode one need to use readconsolew().

[Warfley]

It's one of the problems I've found a few times now. Basically windows does not allow to change the input encoding to UTF-8. It's actually a really big problem for my TermUtils, which simply does not work correctly with Unicode on Windows. It's just something that Windows does not support. The "correct" way of doing it under windows is to read input as utf-16 (widestring/unicode string) and use this or convert it to utf-8 internally.

[pascal111]

It's not understood yet to me although the clarity of the code but my compiler still having no sensible reactions messages. May you provide us some details?