UTF8UpperCase
lazarus / 31-LCL

----------------------------------------------------------------------
Ocye · 2012-12-03 22:02

When I call UTF8UpperCase('ɑbɑri'); (pchar #201#145'b'#201#145'ri')
any further action leads to external SIGSEV. I guess some pointer are
faulty; e.g. OutCounter for "Final correction of the buffer size" is
9.
If someone can confirm I'll post a bug report.

Code: [Select]unit Unit1;

{$mode objfpc}{$H+}

interface

uses
  Classes, SysUtils, FileUtil, Forms, Controls, Graphics, Dialogs,
StdCtrls,
  LCLProc;

type

  { TForm1 }

  TForm1 = class(TForm)
    Button1: TButton;
    Memo1: TMemo;
    procedure Button1Click(Sender: TObject);
  private
    { private declarations }
  public
    { public declarations }
  end;

var
  Form1: TForm1;

implementation

{$R *.lfm}

{ TForm1 }

procedure TForm1.Button1Click(Sender: TObject);
begin
  Caption:=UTF8UpperCase('ɑbɑri');
  Memo1.Lines.Add(Caption);
end;

end.
Lazarus 1.1 r39360M FPC 2.6.0 x86_64-linux-qt

----------------------------------------------------------------------
Bart · 2012-12-03 23:09

I can confirm.
It seems that this happens when a char has an uppercase equivalent
that is longer:
#$C9#$91 -> #$E2#$B#$AD and apparently you need at least 2 of those:
UTf8UpperCase(#$C9#$91) will not crash,
Utf8UpperCase(#$C9#$91#$C9#$91) will.

Bart

----------------------------------------------------------------------
Bart · 2012-12-03 23:33

Reported as issue #23428.

Bart

----------------------------------------------------------------------
Ocye · 2012-12-04 00:12

Quote from: Bart on December 04, 2012, 12:33:27 amReported as issue
#23428.Thank you! But is the resulting character still UTF8 with more
than two bytes?
Anyway, at least no sigsev should be thrown.

----------------------------------------------------------------------
Bart · 2012-12-04 10:25

Quote from: Ocye on December 04, 2012, 01:12:14 amThank you! But is
the resulting character still UTF8 with more than two bytes?

Yes, the upper of #201#145 (2 bytes) is #$E2#$B#$AD (3 bytes).

Quote from: Ocye on December 04, 2012, 01:12:14 amAnyway, at least no
sigsev should be thrown.

I don't get a sigsev, it just crashes (the app vanishes) with the
follwing message:

Marked memory at $0023DBE8 invalid
Wrong signature $7D64F58E instead of 4C796538

Also, this only happens when HeapTrace is enabled (option -gh).
(Tested on Win7-64, Lazarus 1.1/Fpc 2.6.2r1 (32-bit))

Bart

----------------------------------------------------------------------
Bart · 2012-12-04 11:23

I posted a workaround in the bugtracker.

Bart

----------------------------------------------------------------------
Ocye · 2012-12-06 19:09

Unfortunately the bug tracker didn't sent me an email. So I have to
thank Felipe here. Great work!
And thanks to you, Bart, for the management .