Are there any compiler directives or preprocessor commands that need to be set in a particular way to make Indy return Utf-8 strings rather than truncating them into Ansi strings? The project I'm working on has all kinds of Delphi-mode flags all over it if that matters.
If I directly set the subject line to a UTF-8 string (like below) it displays correctly on the GUI, so utf-8 support is set up correctly and I'm using an appropriate font and all of that good stuff. Subject is declared as Utf8String for clarity in this code.
MailItem.Subject := 'îņŢëŕŃïóЙǟŁ ŜũƥĵεϿד'; //Displays correctly
However, if I pull the same subject line from the header, using Indy to decode it, I get every international character replaced with exactly one question-mark, exactly the number of should-be characters. Looks like it's converting UTF-8 to ANSI, which is not what I want.
MailItem.Subject := IdCoderHeader.DecodeHeader('=?utf-8?B?w67FhsWiw6vFlcWDw6/Ds9CZx5/FgSDFnMWpxqXEtc61z7/Xkw==?='); //Displays '?????????? ???????'
So what things could be going wrong and/or how can I fix it?
I am using the latest version of Indy10 from Indy's website, and Lazarus 1.0 on Windows, so I don't think this is a "my software needs updating" bug, I think it's probably some sort of configuration issue.
No, there are no compiler flags you can set for that.
Indy uses AnsiString in Delphi prior to D2009, and in FreePascal prior to 3.0. Indy uses UnicodeString in Delphi 2009+ and FreePascal 3.0+. There is no option to change that.
However, in non-Unicode versions of Delphi and FreePascal, in some places of Indy, you can instruct Indy to interpret AnsiString input values as UTF-8, and return UTF-8 encoded AnsiString output. However, DecodeHeader() is not one of those places.