15 #define F 0   /* character never appears in text */
22 	F, F, F, F, F, F, F, T, T, T, T, F, T, T, F, F,  /* 0x0X */
24 	F, F, F, F, F, F, F, F, F, F, F, T, F, F, F, F,  /* 0x1X */
25 	T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T,  /* 0x2X */
26 	T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T,  /* 0x3X */
27 	T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T,  /* 0x4X */
28 	T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T,  /* 0x5X */
29 	T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T,  /* 0x6X */
30 	T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, F,  /* 0x7X */
32 	X, X, X, X, X, T, X, X, X, X, X, X, X, X, X, X,  /* 0x8X */
33 	X, X, X, X, X, X, X, X, X, X, X, X, X, X, X, X,  /* 0x9X */
34 	I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I,  /* 0xaX */
35 	I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I,  /* 0xbX */
36 	I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I,  /* 0xcX */
37 	I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I,  /* 0xdX */
38 	I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I,  /* 0xeX */
39 	I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I   /* 0xfX */
47  *      0: uses odd control characters, so doesn't look like text
61 	int gotone = 0, ctrl = 0;  in looks_utf8()
63 	for (i = 0; i < nbytes; i++) {  in looks_utf8()
64 		if ((buf[i] & 0x80) == 0) {	   /* 0xxxxxxx is plain ASCII */  in looks_utf8()
72 		} else if ((buf[i] & 0x40) == 0) { /* 10xxxxxx never 1st byte */  in looks_utf8()
77 			if ((buf[i] & 0x20) == 0)	/* 110xxxxx */  in looks_utf8()
78 				if (buf[i] > 0xC1)	/* C0, C1 */  in looks_utf8()
81 			else if ((buf[i] & 0x10) == 0)	/* 1110xxxx */  in looks_utf8()
83 			else if ((buf[i] & 0x08) == 0)	/* 11110xxx */  in looks_utf8()
84 				if (buf[i] < 0xF5)  in looks_utf8()
90 			for (n = 0; n < following; n++) {  in looks_utf8()
95 				if ((buf[i] & 0xc0) != 0x80)	/* 10xxxxxx */  in looks_utf8()
103 	return ctrl ? 0 : (gotone ? 2 : 1);  in looks_utf8()
110  *      0: invalid UTF-16
124 	int following = 0;  in looks_utf16()
127 		return 0;  in looks_utf16()
129 	bom = buf[0] << 8 ^ buf[1];  in looks_utf16()
130 	if (bom == 0xFFFE)  in looks_utf16()
131 		bigend = 0;  in looks_utf16()
132 	else if (bom == 0xFEFF)  in looks_utf16()
135 		return 0;  in looks_utf16()
144 			if (c < 0xD800 || c > 0xDFFF)  in looks_utf16()
146 					return 0;  in looks_utf16()
148 					following = 0;  in looks_utf16()
149 			else if (c > 0xDBFF)  in looks_utf16()
150 				return 0;  in looks_utf16()
155 		else if (c < 0xDC00 || c > 0xDFFF)  in looks_utf16()
156 			return 0;  in looks_utf16()
182 	if ((buf[0] & 0x80) == 0)  in decode_utf8()
183 		u = buf[0];  in decode_utf8()
184 	else if ((buf[0] & 0x40) == 0);  in decode_utf8()
186 		if ((buf[0] & 0x20) == 0)  in decode_utf8()
187 			u = (buf[0] ^ 0xC0) <<  6 ^ (buf[1] ^ 0x80);  in decode_utf8()
188 		else if ((buf[0] & 0x10) == 0)  in decode_utf8()
189 			u = (buf[0] ^ 0xE0) << 12 ^ (buf[1] ^ 0x80) <<  6  in decode_utf8()
190 			  ^ (buf[2] ^ 0x80);  in decode_utf8()
191 		else if (((buf[0] & 0x08) == 0))  in decode_utf8()
192 			u = (buf[0] ^ 0xF0) << 18 ^ (buf[1] ^ 0x80) << 12  in decode_utf8()
193 			  ^ (buf[2] ^ 0x80) <<  6 ^ (buf[3] ^ 0x80);  in decode_utf8()
216 		w1 = buf[0] << 8 ^ buf[1];  in decode_utf16()
218 		w1 = buf[0] ^ buf[1] << 8;  in decode_utf16()
220 	if (w1 < 0xD800 || w1 > 0xDFFF)  in decode_utf16()
222 	else if (w1 > 0xDBFF);  in decode_utf16()
228 		u = ((w1 ^ 0xD800) << 10 ^ (w2 ^ 0xDC00)) + 0x10000;  in decode_utf16()