shithub: femtolisp

ref: 212e0968cc531399fb96edef89f34c4cac29078b
dir: /utf8.c/

View raw version
/*
  Basic UTF-8 manipulation routines
  by Jeff Bezanson
  placed in the public domain Fall 2005

  This code is designed to provide the utilities you need to manipulate
  UTF-8 as an internal string encoding. These functions do not perform the
  error checking normally needed when handling UTF-8 data, so if you happen
  to be from the Unicode Consortium you will want to flay me alive.
  I do this because error checking can be performed at the boundaries (I/O),
  with these routines reserved for higher performance on data known to be
  valid.
  A UTF-8 validation routine is included.
*/

#include "llt.h"

static const uint32_t offsetsFromUTF8[6] = {
	0x00000000UL, 0x00003080UL, 0x000E2080UL,
	0x03C82080UL, 0xFA082080UL, 0x82082080UL
};

static const char trailingBytesForUTF8[256] = {
	0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
	0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
	0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
	0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
	0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
	0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
	1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
	2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2, 3,3,3,3,3,3,3,3,4,4,4,4,5,5,5,5
};

// straight from musl
int
u8_iswprint(Rune c)
{
	if(c < 0xff)
		return ((c+1) & 0x7f) >= 0x21;
	if(c < 0x2028 || c-0x202a < 0xd800-0x202a || c-0xe000 < 0xfff9-0xe000)
		return 1;
	return !(c-0xfffc > Runemax-0xfffc || (c&0xfffe) == 0xfffe);
}

/* returns length of next utf-8 sequence */
size_t
u8_seqlen(const char *s)
{
	return trailingBytesForUTF8[(unsigned int)(uint8_t)s[0]] + 1;
}

/* byte offset => charnum */
size_t
u8_charnum(const char *s, size_t offset)
{
	size_t charnum = 0, i = 0;

	while(i < offset){
		if((s[i++] & 0x80) != 0 && !isutf(s[++i]) && !isutf(s[++i]))
			i++;
		charnum++;
	}
	return charnum;
}

size_t
u8_strwidth(const char *s)
{
	size_t i, w;
	Rune r;

	for(i = w = 0; s[i];){
		i += chartorune(&r, s+i);
		w += wcwidth(r);
	}
	return w;
}

/* next character without NUL character terminator */
Rune
u8_nextmemchar(const char *s, size_t *i)
{
	Rune ch = 0;
	size_t sz = 0;

	do{
		ch <<= 6;
		ch += (uint8_t)s[(*i)++];
		sz++;
	}while(!isutf(s[*i]));
	return ch - offsetsFromUTF8[sz-1];
}

int
octal_digit(char c)
{
	return (c >= '0' && c <= '7');
}

int
hex_digit(char c)
{
	return (c >= '0' && c <= '9') || (c >= 'a' && c <= 'f') || (c >= 'A' && c <= 'F');
}

char
read_escape_control_char(char c)
{
	switch(c){
	case 'n': return '\n';
	case 't': return '\t';
	case 'a': return '\a';
	case 'b': return '\b';
	case 'e': return 0x1b;
	case 'f': return '\f';
	case 'r': return '\r';
	case 'v': return '\v';
	}
	return c;
}

static inline int
buf_put2c(char *buf, const char *src)
{
	buf[0] = src[0];
	buf[1] = src[1];
	buf[2] = '\0';
	return 2;
}

int
u8_escape_rune(char *buf, size_t sz, Rune ch)
{
	assert(sz > 2);
	if(ch >= 0x20 && ch < 0x7f){
		buf[0] = ch;
		buf[1] = '\0';
		return 1;
	}
	if(ch > 0xffff)
		return snprintf(buf, sz, "\\U%.8x", ch);
	if(ch >= Runeself)
		return snprintf(buf, sz, "\\u%04x", ch);
	switch(ch){
	case '\n': return buf_put2c(buf, "\\n");
	case '\t': return buf_put2c(buf, "\\t");
	case '\\': return buf_put2c(buf, "\\\\");
	case '\a': return buf_put2c(buf, "\\a");
	case '\b': return buf_put2c(buf, "\\b");
	case 0x1b: return buf_put2c(buf, "\\e");
	case '\f': return buf_put2c(buf, "\\f");
	case '\r': return buf_put2c(buf, "\\r");
	case '\v': return buf_put2c(buf, "\\v");
	}
	return snprintf(buf, sz, "\\x%02x", ch);
}

size_t
u8_escape(char *buf, size_t sz, const char *src, size_t *pi, size_t end, int escape_quotes, int ascii)
{
	size_t i = *pi, i0;
	Rune ch;
	char *start = buf;
	char *blim = start + sz-11;
	assert(sz > 11);

	while(i < end && buf < blim){
		// sz-11: leaves room for longest escape sequence
		if(escape_quotes && src[i] == '"'){
			buf += buf_put2c(buf, "\\\"");
			i++;
		}else if(src[i] == '\\'){
			buf += buf_put2c(buf, "\\\\");
			i++;
		}else{
			i0 = i;
			ch = u8_nextmemchar(src, &i);
			if(ascii || !u8_iswprint(ch)){
				buf += u8_escape_rune(buf, sz - (buf-start), ch);
			}else{
				i = i0;
				do{
					*buf++ = src[i++];
				}while(!isutf(src[i]));
			}
		}
	}
	*buf++ = '\0';
	*pi = i;
	return (buf-start);
}

char *
u8_memchr(const char *s, Rune ch, size_t sz, size_t *charn)
{
	size_t i = 0, lasti = 0;
	Rune c;
	int csz;

	*charn = 0;
	while(i < sz){
		c = csz = 0;
		do{
			c <<= 6;
			c += (uint8_t)s[i++];
			csz++;
		}while(i < sz && !isutf(s[i]));
		c -= offsetsFromUTF8[csz-1];

		if(c == ch)
			return (char*)&s[lasti];
		lasti = i;
		(*charn)++;
	}
	return nil;
}

/* based on the valid_utf8 routine from the PCRE library by Philip Hazel

   length is in bytes, since without knowing whether the string is valid
   it's hard to know how many characters there are! */
int
u8_isvalid(const char *str, int length)
{
	const uint8_t *p, *pend = (uint8_t*)str + length;
	uint8_t c;
	int ab;

	for(p = (uint8_t*)str; p < pend; p++){
		c = *p;
		if(c < 128)
			continue;
		if((c & 0xc0) != 0xc0)
			return 0;
		ab = trailingBytesForUTF8[c];
		if(length < ab)
			return 0;
		length -= ab;

		p++;
		/* Check top bits in the second byte */
		if((*p & 0xc0) != 0x80)
			return 0;

		/* Check for overlong sequences for each different length */
		switch(ab){
			/* Check for xx00 000x */
		case 1:
			if((c & 0x3e) == 0)
				return 0;
			continue; /* We know there aren't any more bytes to check */

			/* Check for 1110 0000, xx0x xxxx */
		case 2:
			if(c == 0xe0 && (*p & 0x20) == 0)
				return 0;
			break;

			/* Check for 1111 0000, xx00 xxxx */
		case 3:
			if(c == 0xf0 && (*p & 0x30) == 0)
				return 0;
			break;

			/* Check for 1111 1000, xx00 0xxx */
		case 4:
			if(c == 0xf8 && (*p & 0x38) == 0)
				return 0;
			break;

			/* Check for leading 0xfe or 0xff and then for 1111 1100, xx00 00xx */
		case 5:
			if(c == 0xfe || c == 0xff || (c == 0xfc && (*p & 0x3c) == 0))
				return 0;
			break;
		}

		/* Check for valid bytes after the 2nd, if any; all must start 10 */
		while(--ab > 0)
			if((*(++p) & 0xc0) != 0x80)
				return 0;
	}

	return 1;
}

int
u8_reverse(char *dest, char *src, size_t len)
{
	size_t si, di;

	dest[di = len] = '\0';
	for(si = 0; si < len;){
		switch((uint8_t)src[si]>>4){
		case 0xf:
			di -= 4;
			dest[di+0] = src[si++];
			dest[di+1] = src[si++];
			dest[di+2] = src[si++];
			dest[di+3] = src[si++];
			break;
		case 0xe:
			di -= 3;
			dest[di+0] = src[si++];
			dest[di+1] = src[si++];
			dest[di+2] = src[si++];
			break;
		case 0xd:
		case 0xc:
			di -= 2;
			dest[di+0] = src[si++];
			dest[di+1] = src[si++];
			break;
		default:
			di--;
			dest[di+0] = src[si++];
			break;
		}
	}
	return 0;
}