On Wed, April 21, 2010 12:03, Graeme Geldenhuys wrote: > Tomas Hajny het geskryf: >> >> What do you mean by "messing with the -Fcutf8 parameter"? If you include >> some character constants outside the us-ascii set, you should always >> tell >> the compiler how to evaluate them, because the evaluation depends on the > > fpGUI uses UTF-8 internally on all platforms. Specifying some Unicode > characters as a sequence of bytes stored inside an TfpgString type (alias > for AnsiString) causes no problems with FPC. To the compiler it's simply a > sequence of bytes. fpGUI has it's one UTF-8 file and string manipulation > functions.
This is the important part which was missing in your original message - you don't use Unicode specific types (UnicodeChar, UnicodeString, UTF8String, etc.). As mentioned in my e-mail, my comment was primarily related to using these types. Obviously, if you use a type which is "transparent" from this point of view, everything is up to you. > The usage of #xxxx format also overcomes some problems with editors not > fully supporting Unicode or when you use a non-Unicode font in your > editor. > For example, I use a non-Unicode font (Raize) under Linux with Lazarus > IDE. > If I try and insert true Unicode characters inside the editor, things > screw-up and text is garbled. Yes, the alternative notation solves that problem indeed. Tomas _______________________________________________ fpc-pascal maillist - fpc-pascal@lists.freepascal.org http://lists.freepascal.org/mailman/listinfo/fpc-pascal
