Strings in VCL

There are four different string types in VCL:

In addition of these there is String alias that is mapped to one of the above strings. In Delphi 1 String was ShortString. After that it was AnsiString until Delphi 2009/C++Builder 2009 where the default string was changed to UnicodeString.

When Delphi or C++Builder compiler compiles an application and saves form data into the application resource the compiler encodes all strings. There are four different string encoding methods:

In general Sisulizer uses the same encoding in localized strings as in original strings. For example if the original form is in English most strings are encoded as short string using code page 1252. Strings that are longer than 255 characters are encoded as long strings using code page 1252. If a string contains non-ASCII characters (e.g. ©, Ä or Ö) it is encoded as UTF-8 or UTF-16, whichever takes less space.

When Sisulizer writes localized forms the string value are different to the original values. Their character sets are also different. This might make Sisulizer to use Unicode much more than Ansi. Especially when you are localizing to a non-Latin based language such as Greek or Japanese. If you use Delphi/C++Builder 6 or later there are no problem. All strings encoding are valid. However if you use Delphi/C++Builder 5 or older Unicode encoding are not available for properties that has String type. There is not easy and 100% reliable way to detect what was the Delphi/C++Builder compiler that was used to compile the EXE. This is why you can configure if and how Sisulizer writes Unicode strings.

Use the Options sheet of VCL source dialog. Right click your application on the project tree and choose Properties. Options sheet contains Widestring (Unicode) usage part.

Widestring (Unicode) usage

Specifies if Ansi strings are converted to Unicode strings when creating localized forms. Possible values are:

Unicode VCL

Delphi 2009 and C++Builder 2009 and later have a Unicode enabled VCL. It means that the default string is now UnicodeString instead of AnsiString. If you localize an Unicode VCL application Sisulizer writes all strings as Unicode using either UTF-8 or UTF-16.

Unicode Components

TNT and ElPack components provide Unicode support even when used with non-Unicode VCL. Sisulizer automatically detects that you use Unicode components and writes all strings as Unicode. Sometimes usage of Unicode components can not be detected. In that case turn Widestring (Unicode) usage to Yes.

If your application is written in Delphi 5, C++Builder 5 or older and it contains Unicode enabled components such ad TNT or ElPack and you have not turned off Unicode writing you will get the following error when you try to run your localized application:

To make you localized application to work you have to turn Widestring (Unicode) usage to No.