Native UI and language support

Yeah, its a funny one because the native labels obviously do support multi-language settings. You certainly don't need to change the native font, but the MoSync methods for setting properties take a const char* which will still be single byte characters.  I don't know of a way to set the native code page or anything, so I'd expect the same restrictions, unless someone knows any better.?

Hi is there any solution for this problem now?

For my NativeUI App I need german umlauts too.

Regards,
David

There are some people in MoSync looking at this problem for a long-term fix. I've had issues again with it recently (it was working, and then it stopped). I found out that my problem was caused by reading in the text from XML.

MoSync MTXml parser does support UTF-8. If you called the method 'feedProcess' instead of 'feed' it will convert UTF-8 characters into ANSI characters, using (I think) the 1252 codepage. It may even support UTF-16 now, with wide character support, but I've not tested it myself. with 'feedWide'.

This is the code from MTXml which does the conversion. You can use this to your strings into compatible ones with the right characters. This works for me in AMUI/MAUI and I'll try it later with NativeUI.

//processes UTF-8 of the data if the CONTEXT says the document uses that encoding.
//also processes standard entities if(ent).
//return <0 on error.
//if any data was successfully processed, return the length of it.
//remainsLen = 0;
//if any data remains, set remainsLen to the length of it and set sCurPtr to the start of it.
#endif
//for some functions, remaining data will mean error.
//remainsLen is partially used as "refLen" and "utf8Len" inside this function.

//todo, maybe: improve performance at the cost of code size
//by making this function a template on ent, sUtf8 and sWideBuf.
static int _proc(char* data, int* remainsLen, bool ent) {
	char* dst = data;
	wchar_t* wdst = sWideBuf;
	char* src = data;
	*remainsLen = 0;
	while(*src != 0 ) {
		int res = 0;
		if(ent && *src == '&') {	//reference
			src++;
			res = parseRef(src, remainsLen);
			if(res == PEC_INCOMPLETE) {
				sCurPtr = src - 1;
				break;
			} else if(res < 0 )
				return res;
			if(res > 0xFF && !sWideBuf) {
				res = sContext->unicodeCharacter(sContext, res);
				if(res == 0 )
					return PEC_OUTSIDE;
			}
		} else if(sUtf8) {	//normal character
			if(sWideBuf) {
				res = convertUtf8ToUnicode(src, remainsLen);
			} else {
				res = convertUtf8ToLatin1(src, remainsLen);
			}
			if(res == PEC_INCOMPLETE) {
				sCurPtr = src;
				break;
			} else if(res < 0 )
				return res;
		}
		if(res != 0 ) {
			if(sWideBuf) {
				*(wdst++) = (wchar_t)res;
			} else {
				*(dst++) = (char)res;
			}
			src += *remainsLen;
			*remainsLen = 0;
		} else {
			if(sWideBuf) {
				*(wdst++) = *(src++);
			} else {
				*(dst++) = *(src++);
			}
		}
	}
	if(sWideBuf) {
		*wdst = 0;
		return wdst - sWideBuf;
	} else {
		*dst = 0;
		return dst - data;
	}
}

//todo: optimize. maybe use mbtowc().
static int convertUtf8ToUnicode(const char* utf8, int* pnBytes) {
	char b = utf8[0];
	if(b & 0x80 ) {
		int nBytes = 0, unicode, i;
		do {
			nBytes++;
			b <<= 1;
		} while(b & 0x80);
		if(nBytes > 4) {
			//printf("CU8b: %i", nBytes);
			return PEC_INVALID;
		}
		unicode = (b >> nBytes) << ((nBytes-1) * 6);
		for(i=0; i<nBytes-1; i++) {
			if(utf8[i+1] == 0 ) {
				*pnBytes = i+1;
				return PEC_INCOMPLETE;
			}
			unicode |= (utf8[i+1] & 0x3F) << ((nBytes-2-i) * 6);
		}
		*pnBytes = nBytes;
		return unicode;
	} else {
		*pnBytes = 1;
		return b;
	}
}

static int convertUtf8ToLatin1(const char* utf8, int* pnBytes) {
	int unicode = convertUtf8ToUnicode(utf8, pnBytes);
	if(unicode > 0xFF) {
		//0xFEFF is ZERO WIDTH NO-BREAK SPACE, an encoding signature.
		if(unicode == 0xFEFF) {
			unicode = ' ';	//not a proper translation, but it shouldn't cause any problems.
		} else {
			//printf("CU8u: %i", unicode);
			unicode = sContext->unicodeCharacter(sContext, unicode);
			if(unicode == 0 )
				return PEC_OUTSIDE;
		}
	}
	return unicode;
}

It works with this workaround...

http://jira.mosync.com/browse/MOSYNC-1735

I'm pleased you've got a workaround, but its still not working for me. My text is in a UTF-8 XML document. It is being processed correctly, and everything tells me that the text readable, but it still crashes on character codes > 127.