base_rune.c

来自「SumatraPDF是一款小型开源的pdf阅读工具。虽然玲珑小巧（只有800多K」· C语言 代码 · 共 175 行

enum
{
	UTFmax = 3, /* maximum bytes per rune */
	Runesync = 0x80, /* cannot represent part of a UTF sequence (<) */
	Runeself = 0x80, /* rune and UTF sequences are the same (<) */
	Runeerror = 0x80 /* decoding error in UTF */
};

enum
{
	Bit1	= 7,
	Bitx	= 6,
	Bit2	= 5,
	Bit3	= 4,
	Bit4	= 3,

	T1	= ((1<<(Bit1+1))-1) ^ 0xFF,	/* 0000 0000 */
	Tx	= ((1<<(Bitx+1))-1) ^ 0xFF,	/* 1000 0000 */
	T2	= ((1<<(Bit2+1))-1) ^ 0xFF,	/* 1100 0000 */
	T3	= ((1<<(Bit3+1))-1) ^ 0xFF,	/* 1110 0000 */
	T4	= ((1<<(Bit4+1))-1) ^ 0xFF,	/* 1111 0000 */

	Rune1	= (1<<(Bit1+0*Bitx))-1,		/* 0000 0000 0111 1111 */
	Rune2	= (1<<(Bit2+1*Bitx))-1,		/* 0000 0111 1111 1111 */
	Rune3	= (1<<(Bit3+2*Bitx))-1,		/* 1111 1111 1111 1111 */

	Maskx	= (1<<Bitx)-1,			/* 0011 1111 */
	Testx	= Maskx ^ 0xFF,			/* 1100 0000 */

	Bad	= Runeerror
};

int
chartorune(int *rune, char *str)
{
	int c, c1, c2;
	int l;

	/*
	 * one character sequence
	 *	00000-0007F => T1
	 */
	c = *(unsigned char*)str;
	if (c < Tx)
	{
		*rune = c;
		return 1;
	}

	/*
	 * two character sequence
	 *	0080-07FF => T2 Tx
	 */
	c1 = *(unsigned char*)(str+1) ^ Tx;
	if (c1 & Testx)
		goto bad;
	if (c < T3)
	{
		if (c < T2)
			goto bad;
		l = ((c << Bitx) | c1) & Rune2;
		if (l <= Rune1)
			goto bad;
		*rune = l;
		return 2;
	}

	/*
	 * three character sequence
	 *	0800-FFFF => T3 Tx Tx
	 */
	c2 = *(unsigned char*)(str+2) ^ Tx;
	if (c2 & Testx)
		goto bad;
	if (c < T4)
	{
		l = ((((c << Bitx) | c1) << Bitx) | c2) & Rune3;
		if (l <= Rune2)
			goto bad;
		*rune = l;
		return 3;
	}

	/*
	 * bad decoding
	 */
bad:
	*rune = Bad;
	return 1;
}

int
runetochar(char *str, int *rune)
{
	int c;

	/*
	 * one character sequence
	 *	00000-0007F => 00-7F
	 */
	c = *rune;
	if (c <= Rune1)
	{
		str[0] = c;
		return 1;
	}

	/*
	 * two character sequence
	 *	0080-07FF => T2 Tx
	 */
	if (c <= Rune2)
	{
		str[0] = T2 | (c >> 1*Bitx);
		str[1] = Tx | (c & Maskx);
		return 2;
	}

	/*
	 * three character sequence
	 *	0800-FFFF => T3 Tx Tx
	 */
	str[0] = T3 |  (c >> 2*Bitx);
	str[1] = Tx | ((c >> 1*Bitx) & Maskx);
	str[2] = Tx |  (c & Maskx);
	return 3;
}

int
runelen(int c)
{
	int rune;
	char str[10];

	rune = c;
	return runetochar(str, &rune);
}

int
runenlen(int *r, int nrune)
{
	int nb, c;

	nb = 0;
	while (nrune--)
	{
		c = *r++;
		if (c <= Rune1)
			nb++;
		else if (c <= Rune2)
			nb += 2;
		else
			nb += 3;
	}
	return nb;
}

int
fullrune(char *str, int n)
{
	int c;

	if (n > 0)
	{
		c = *(unsigned char*)str;
		if (c < Tx)
			return 1;
		if (n > 1)
			if (c < T3 || n > 2)
				return 1;
	}
	return 0;
}