util_unistr.c File Reference

#include <arpa/inet.h>
#include <atalk/logger.h>
#include <errno.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/param.h>
#include <sys/stat.h>
#include <atalk/unicode.h>
#include <atalk/byteorder.h>
#include "precompose.h"

Functions
int	strlower_w (ucs2_t *s)
	Convert a string to lower case.
int	strupper_w (ucs2_t *s)
	Convert a string to upper case.
size_t	strlen_w (const ucs2_t *src)
	wide strlen()
size_t	strnlen_w (const ucs2_t *src, size_t max)
	wide strnlen()
ucs2_t *	strchr_w (const ucs2_t *s, ucs2_t c)
	wide strchr()
ucs2_t *	strcasechr_w (const ucs2_t *s, ucs2_t c)
	wide strcasechr()
ucs2_t *	strcasechr_sp (const ucs2_t *s, uint32_t c_sp)
	sp strcasechr()
int	strcmp_w (const ucs2_t *a, const ucs2_t *b)
	wide strcmp()
int	strncmp_w (const ucs2_t *a, const ucs2_t *b, size_t len)
	wide strncmp()
ucs2_t *	strstr_w (const ucs2_t *s, const ucs2_t *ins)
	wide strstr()
ucs2_t *	strcasestr_w (const ucs2_t *s, const ucs2_t *ins)
	wide strcasestr()
int	strcasecmp_w (const ucs2_t *a, const ucs2_t *b)
	wide strcasecmp()
int	strncasecmp_w (const ucs2_t *a, const ucs2_t *b, size_t len)
	wide strncasecmp()
ucs2_t *	strndup_w (const ucs2_t *src, size_t len)
	wide strndup()
ucs2_t *	strdup_w (const ucs2_t *src)
	wide strdup()
static ucs2_t	do_precomposition (unsigned int base, unsigned int comb)
static uint32_t	do_precomposition_sp (unsigned int base_sp, unsigned int comb_sp)
static uint32_t	do_decomposition (ucs2_t base)
static uint64_t	do_decomposition_sp (unsigned int base_sp)
size_t	precompose_w (ucs2_t *name, size_t inplen, ucs2_t *comp, size_t *outlen)
	pre|decomposition
size_t	decompose_w (ucs2_t *name, size_t inplen, ucs2_t *comp, size_t *outlen)

Function Documentation

decompose_w()

size_t decompose_w	(	ucs2_t *	name,
		size_t	inplen,
		ucs2_t *	comp,
		size_t *	outlen
	)

do_decomposition()

static uint32_t do_decomposition ( ucs2_t  base )

static

do_decomposition_sp()

static uint64_t do_decomposition_sp ( unsigned int  base_sp )

static

do_precomposition()

static ucs2_t do_precomposition ( unsigned int  base, unsigned int  comb  )

static

binary search for pre|decomposition

do_precomposition_sp()

static uint32_t do_precomposition_sp ( unsigned int  base_sp, unsigned int  comb_sp  )

static

precompose_w()

size_t precompose_w	(	ucs2_t *	name,
		size_t	inplen,
		ucs2_t *	comp,
		size_t *	outlen
	)

pre|decomposition

we can't use static, this stuff needs to be reentrant static char comp[MAXPATHLEN +1];

We don't implement Singleton and Canonical Ordering. We ignore CompositionExclusions.txt because they cause the problem of the roundtrip such as Dancing Icon.

exclude U2000-U2FFF, UFE30-UFE4F and U2F800-U2FA1F ranges in precompose.h from composition according to AFP 3.x spec

strcasechr_sp()

ucs2_t * strcasechr_sp	(	const ucs2_t *	s,
		uint32_t	c_sp
	)

sp strcasechr()

Note

separately process BMP and surrogate pair

strcasechr_w()

ucs2_t * strcasechr_w	(	const ucs2_t *	s,
		ucs2_t	c
	)

wide strcasechr()

Note

separately process BMP and surrogate pair

strcasecmp_w()

int strcasecmp_w	(	const ucs2_t *	a,
		const ucs2_t *	b
	)

wide strcasecmp()

case insensitive string comparison

Note

surrogate pair support

strcasestr_w()

ucs2_t * strcasestr_w	(	const ucs2_t *	s,
		const ucs2_t *	ins
	)

wide strcasestr()

Note

surrogate pair support

strchr_w()

ucs2_t * strchr_w	(	const ucs2_t *	s,
		ucs2_t	c
	)

wide strchr()

Note

hi and lo of surrogate pair are separately processed.

strcmp_w()

int strcmp_w	(	const ucs2_t *	a,
		const ucs2_t *	b
	)

wide strcmp()

Note

no problem of surrogate pair

strdup_w()

ucs2_t * strdup_w

(

const ucs2_t *

src

)

wide strdup()

duplicate string

Note

no problem of surrogate pair

strlen_w()

size_t strlen_w

(

const ucs2_t *

src

)

wide strlen()

Count the number of characters in a UTF-16 string.

Note

one surrogate pair is two characters.

strlower_w()

int strlower_w

(

ucs2_t *

s

)

Convert a string to lower case.

Returns

True if any char is converted

Note

surrogate pair support

strncasecmp_w()

int strncasecmp_w	(	const ucs2_t *	a,
		const ucs2_t *	b,
		size_t	len
	)

wide strncasecmp()

case insensitive string comparison, length limited

Note

compare up to 'len+1' if 'len' isolate surrogate pair

strncmp_w()

int strncmp_w	(	const ucs2_t *	a,
		const ucs2_t *	b,
		size_t	len
	)

wide strncmp()

Note

no problem of surrogate pair

strndup_w()

ucs2_t * strndup_w	(	const ucs2_t *	src,
		size_t	len
	)

wide strndup()

duplicate string

Note

not check isolation of surrogate pair

if len == 0 then duplicate the whole string

strnlen_w()

size_t strnlen_w	(	const ucs2_t *	src,
		size_t	max
	)

wide strnlen()

Count up to max number of characters in a UTF-16 string.

Note

one surrogate pair is two characters.

strstr_w()

ucs2_t * strstr_w	(	const ucs2_t *	s,
		const ucs2_t *	ins
	)

wide strstr()

Note

no problem of surrogate pair

strupper_w()

int strupper_w

(

ucs2_t *

s

)

Convert a string to upper case.

Returns

True if any char is converted

Note

surrogate pair support