How to display UTF-8 in a HITextView (MLTE) control?

Question

I have a UTF-8 string which I want to display in an HITextView (MLTE) control. Theoretically, HITextView requires either "Text" or UTF-16, so I'm converting:

UniChar uniput[STRSIZE];
ByteCount converted,unilen;

err = ConvertFromTextToUnicode(C2UInfo, len, output,
                               kUnicodeUseFallbacksMask, 0, NULL, 0, NULL,
                               sizeof(UniChar)*STRSIZE,
                               &converted, &unilen, uniput);
err=TXNSetData(MessageObject, kTXNUnicodeTextData, uniput, unilen, kTXNEndOffset,
               kTXNEndOffset);

I have defined the converter C2UInfo as follows:

UnicodeMapping    uMapping;
uMapping.unicodeEncoding = CreateTextEncoding(kTextEncodingUnicodeV2_0,
                              kUnicodeCanonicalDecompVariant,
                              kUnicode16BitFormat);
uMapping.otherEncoding = GetTextEncodingBase(kUnicodeUTF8Format);
uMapping.mappingVersion = kUnicodeUseLatestMapping;
err = CreateTextToUnicodeInfo(&uMapping, &C2UInfo);

It works fine for plain old ASCII characters, but multi-byte UTF-8 is being mapped to the wrong characters. For example, æ (LATIN SMALL LETTER AE) is being mapped to 疆 (CJK UNIFIED IDEOGRAPH-7586).

I've tried checking and unchecking "Output Text in Unicode" in Interface Builder, and I've tried varying some of the conversion constants, with no effect.

This is being built with Xcode 3.2.6 using the MacOSX10.5.sdk and tested under 10.6.

Answer 1

The “Text” that ConvertFromTextToUnicode expects is probably the same “Text” that is one of your two options for MLTE. If you had the sort of “Text” that ConvertFromTextToUnicode converts from, you could just pass it to MLTE directly.

(For the record, “Text” is almost certainly either MacRoman or whatever is dictated by the user's locale-determined current script.)

Instead, you should use a Text Encoding Converter. Create one, use it, finish using it, and dispose of it when you're done.

There are two other ways.

One is to create a CFString from the UTF-8, then Get its characters. You would do this instead of using a TEC. It's functionally equivalent and possibly a little bit easier. On the other hand, you don't get to reuse the converter, for whatever that's worth.

The other, since you have an HITextView, would be to create a CFString from the UTF-8 and just use that. Like Cocoa objects, HIToolbox objects have an inheritance hierarchy; since an HITextView is a kind of HIView, HIViewSetText should just work. (And if not, try HIViewSetValue.)

The last method also gets you that much closer to your eventual move away from MLTE/HITextView, since it's essentially what you'll do with an NSTextView. (HITextView and MLTE are deprecated.)