On 2005-10-18 23:31, Elias Mårtenson said:
GCC isn't mangling the string. @"" is designed to take ASCII
encoding, not UTF8 or MacRoman or anything else. What you'll get
for any character above 127 is just undefined. Why Apple/NeXT chose
to make @"" work this way, I don't know, but that's the situation.
No, it's not GCC that is mangling the string, it's whatever code runs
when the @ is seen before a string. GCC itself delivers UTF-8 when it
encounters a \u inside a string. Therefore the following works:
(remember that you need the -std=c99 compiler option for this to work)
NSString *s = [NSString stringWithUTF8String:"Mac\u00A0OS\u00A0X"];
And yes, I just tried it myself. :-)
How lame it is that
NSString *s = [NSString stringWithUTF8String:"Mac\u00A0OS\u00A0X"];
is ok, but
NSString *s = @"Mac\u00A0OS\u00A0X";
is not. But it's not gcc's fault. The Objective-C documentation says:
"@"string" - Defines a constant NSString object in the current module
and initializes the object with the specified 7-bit ASCII-encoded string."
So the language is defined that way, the compiler must follow. The
compiler can warn you though, with this option:
-fconstant-cfstrings
It will say "warning: non-ASCII character in CFString literal".
Of course our beloved xCode does not turn this on by default, nor does
it have any GUI option for it. In fact, from the gcc man page, it
doesn't even seem to me that -fconstant-cfstrings is a warning option,
nevertheless it works if you put in 'other warning flags'. I hope it
doesn't break something else! :)
