How do I prevent the warning "unicode constant cast with potential data loss" in Lazarus?

Question

I'm trying to assign the Tricolon Unicode character #$205D as the caption to a button in a Lazarus Windows program like this:

MyButton.Caption := #$205D;

It works, the button displays the Tricolon fine, but the compiler emits the warning "Warning: Unicode constant cast with potential data loss".

How do I correctly assign the Tricolon character to the caption of an LCL control to get rid of the warning?

What compiler options are active relating to character encoding. — David Heffernan, Jun 04 '17 at 09:51
Try `MyButton.Caption := #$E2#$81#$9D;` LCL using UTF8 encoding but `#$205D` is UTF16 character constant. — Abelisto, Jun 04 '17 at 10:04
@Abelisto You were right, that's a (the?) solution. Care to make this an answer so I can accept it? — dummzeuch, Jun 04 '17 at 10:23
@DavidHeffernan no idea, I installed the latest Lazarus release for Windows and haven't changed anything. I don't even know (yet) where to look for such compiler options. — dummzeuch, Jun 04 '17 at 10:25
http://wiki.freepascal.org/Character_and_string_types with such complexity I doubt trial and error is the way to go — David Heffernan, Jun 04 '17 at 11:36
@DavidHeffernan Yes, it seems it is your personal force to make everything more complicated then it is. However, just remember, string is just the sequence of bytes. The magic starts at the representation, not at the assignment. — Abelisto, Jun 04 '17 at 21:47
@Abelisto everything in the computer is just bytes, but that isn't so helpful. FPC has a lot of options and variations. It's trickier than you think. — David Heffernan, Jun 04 '17 at 21:50
@DavidHeffernan "It's trickier than you think." - ABSOLUTELY RIGHT!!! — Abelisto, Jun 04 '17 at 21:57

score 3 · Accepted Answer · answered Jun 04 '17 at 10:27

LCL uses UTF8 encoding but #$205D is UTF16 character constant. So use UTF8 encoded constants instead:

const
    CTricolon = #$E2#$81#$9D;
    //CTricolon = '⁝'; // Also works fine if using character(s) as is in the source

...

    MyButton.Caption := CTricolon;

score 1 · Answer 2 · answered Jun 11 '17 at 16:43

The problem is that the detection of a 2-byte -> (default) 1-byte conversion is compiletime, and the exact codepage of the default 1-byte type is runtime.

(either varying Windows encodings depending on locale or set to UTF8 on startup in Lazarus)

The compiler is warning you that this is dangerous. To fix it set the source encoding to utf8 and assign an utf8 string.

How do I prevent the warning "unicode constant cast with potential data loss" in Lazarus?

2 Answers2