Hi,

I created new project, then choose "File Settings -> Encoding -> UTF8 with BOM -> Change file"

Added new listbox1 to the form, in source code (button ) i added some items with my local language, this language in run time apear like "??????".

Is that a bug or i missed something?

The project is attached.

I am using Lazarus 0.9.31 r FPC 2.6.0 i386-win32-win32/win64

Thanks in advance.

Best Regards
Zaher Dirkey
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lazarus.freepascal.org/pipermail/lazarus/attachments/20120125/d003decc/attachment.html> -------------- next part --------------
A non-text attachment was scrubbed...
Name: utfbom.zip
Type: application/zip
Size: 2687 bytes
Desc: not available
URL: <http://lists.lazarus.freepascal.org/pipermail/lazarus/attachments/20120125/d003decc/attachment.zip>

When the compiler sees the UTF8-BOM it translates all strings to your system encoding, which is not UTF-8. The LCL expects UTF-8 strings. Result: you see only garbage.

Mattias

On Wed, Jan 25, 2012 at 1:33 AM, Mattias Gaertner <

How can i resolve this problem, because i share this unit with Delphi 2007, Delphi 2007 not understand UTF8 *without *BOM

Thanks

Zaher Dirkey
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lazarus.freepascal.org/pipermail/lazarus/attachments/20120125/4a6dc04e/attachment.html>

2012/1/25 Zaher Dirkey :

How can i resolve this problem, because i share this unit with Delphi 2007, Delphi 2007 not understand UTF8 without BOM

For FPC up to 2.6 you can use all text fields as ansistring variables and use ifdefs to call a procedure to convert all UTF-8 strings to Ansi only in Delphi, see this unit from the Magnifier which does exactly this:

http://magnifier.svn.sourceforge.net/viewvc/magnifier/trunk/translationsvmg.pas?revision=210&view=markup
See there procedure TTranslations.UpdateTranslations;

If you are using FPC 2.7 try keeping the BOM and add this call before Application.Initialize:

SetMultiByteConversionCodePage(CP_UTF8);

2012/1/24 Zaher Dirkey :

I created new project, then choose "File Settings -> Encoding -> UTF8 with BOM -> Change file"

Although BOM is valid in UTF8, its use is not recommended (http://en.wikipedia.org/wiki/Byte_order_mark#cite_note-1). I'm surprised Lazarus supports this feature should be removed and UTF-8 files with BOM should be rejected by FPC and Lazarus as invalid.

Flame war in 3, 2, ... ;-)

Hello Lazarus-List,

Wednesday, January 25, 2012, 11:13:16 AM, you wrote:

H> Although BOM is valid in UTF8, its use is not recommended H> (http://en.wikipedia.org/wiki/Byte_order_mark#cite_note-1). I'm H> surprised Lazarus supports this feature should be H> removed and UTF-8 files with BOM should be rejected by FPC and Lazarus H> as invalid.

Please, re-read the cite. It should be used in undetermined encoding protocol media, like files, in fact it is quite mandatory in many of them. The cite is about setting a BOM in things like string variables where the protocol is expected (programatically) or database fields where the field definition already states the encoding used.

A pascal program file without BOM is ansi as you can say that a file could not be UTF8 but you can not say that it is not ansi unless BOM is present.

25.01.2012 2:02, Zaher Dirkey ?????:

When BOM or {$codepage utf8} directive is present compiler converts your strings to widestrings. So you have two options

1. Add {$codepage 8859-1} directive at the start of the .pas file.

2. Use UTF8Encode() function, e.g.

ListBox1.Items.Add(UTF8Encode('?????? ?????'));

Maxim Ganetsky wrote

{$codepage 8859-1}

Thanks, Until now this sloved my problem in both Delphi and Lazarus

{$ifdef FPC}
{$mode delphi}
{$codepage 8859-1}
{$endif}

But strange, it is not as expected!.

Best Regards

Zaher Dirkey
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lazarus.freepascal.org/pipermail/lazarus/attachments/20120125/138e9712/attachment-0001.html>

26.01.2012 1:57, Zaher Dirkey ?????:

No, it works exactly as documented:

http://wiki.lazarus.freepascal.org/LCL_Unicode_Support#Free_Pascal_Particularities

http://wiki.freepascal.org/Language_related_articles#UTF-8_support