Difference between revisions of "What's new in TSmtpCli V7"
Line 30: | Line 30: | ||
==== New Property ConvertToCharset ==== | ==== New Property ConvertToCharset ==== | ||
− | If set to TRUE enforces charset conversion | + | If set to TRUE enforces charset conversion to the character set specified in property CharSet. This property is ignored by UNICODE compilers. |
Revision as of 09:54, 29 May 2009
Unicode Changes
Since Delphi/C++Builder 2009 and better "string" maps to a UnicodeString. That's why the component has to convert these UTF-16 strings to ANSI strings with a correct code page first before transmission. Also SMTP allows 7-bit ASCII chars in the range hex 00-7F only. So if the (converted) AnsiString contains characters in the 8-bit range (hex 80-FF) it has to be encoded with a proper transfer-encoding as well.
TSmtpCli V7 is capable to perform the right string conversion and encoding automatically in the background, provided the following conditions met:
- Property CharSet is properly set, it defaults to current system ANSI charset.
- Property Allow8BitChars is set to FALSE.
- Property DefaultEncoding is set to either smtpEnc7bit or smtpEnc8bit.
- Message text is assigned to property MailMessage and property OnGetData is unassigned.
Notes on property CharSet (breaking change in V7!)
- Assigning an unsupported value raises an exception and sets the default system charset.
- Assigning an empty string sets the default system charset silently.
Mapping of MIME charset names to Windows code page identifiers is done in new unit OverbyteIcsCharsetUtils. If a charset name is mapped to a code page identifier it's not guaranteed that this ID is available and supported by Windows as well. For example, the Chinese character sets are not supported by default on western Windows installations. After installation of the optional Far-East language pack those charsets are supported as well.
Avoid use of "utf-7", UTF-7 does not yet work very well, instead use "utf-8" if you want to send UNICODE.
New Property ConvertToCharset
If set to TRUE enforces charset conversion to the character set specified in property CharSet. This property is ignored by UNICODE compilers.