openldap/libraries/liblunicode/ucstr.c

419 lines
8.7 KiB
C
Raw Normal View History

2002-01-04 00:36:46 +08:00
/*
2003-01-04 04:20:47 +08:00
* Copyright 2000-2003 The OpenLDAP Foundation
2002-01-04 00:36:46 +08:00
* COPYING RESTRICTIONS APPLY. See COPYRIGHT File in top level directory
* of this package for details.
*/
2000-09-08 07:19:10 +08:00
#include "portable.h"
#include <ac/ctype.h>
2001-06-15 08:02:12 +08:00
#include <ac/string.h>
#include <ac/stdlib.h>
2000-09-08 07:19:10 +08:00
#include <lber.h>
#include <ldap_utf8.h>
#include <ldap_pvt_uc.h>
2003-04-11 11:57:10 +08:00
#define malloc(x) ber_memalloc_x(x,ctx)
#define realloc(x,y) ber_memrealloc_x(x,y,ctx)
#define free(x) ber_memfree_x(x,ctx)
2000-09-08 07:19:10 +08:00
int ucstrncmp(
const ldap_unicode_t *u1,
const ldap_unicode_t *u2,
ber_len_t n )
{
for(; 0 < n; ++u1, ++u2, --n ) {
if( *u1 != *u2 ) {
return *u1 < *u2 ? -1 : +1;
}
if ( *u1 == 0 ) {
return 0;
}
}
return 0;
}
int ucstrncasecmp(
const ldap_unicode_t *u1,
const ldap_unicode_t *u2,
ber_len_t n )
{
for(; 0 < n; ++u1, ++u2, --n ) {
ldap_unicode_t uu1 = uctolower( *u1 );
ldap_unicode_t uu2 = uctolower( *u2 );
2000-09-08 07:19:10 +08:00
if( uu1 != uu2 ) {
return uu1 < uu2 ? -1 : +1;
}
if ( uu1 == 0 ) {
return 0;
}
}
return 0;
}
ldap_unicode_t * ucstrnchr(
const ldap_unicode_t *u,
ber_len_t n,
ldap_unicode_t c )
{
for(; 0 < n; ++u, --n ) {
if( *u == c ) {
return (ldap_unicode_t *) u;
}
}
return NULL;
}
ldap_unicode_t * ucstrncasechr(
const ldap_unicode_t *u,
ber_len_t n,
ldap_unicode_t c )
{
c = uctolower( c );
2000-09-08 07:19:10 +08:00
for(; 0 < n; ++u, --n ) {
if( uctolower( *u ) == c ) {
2000-09-08 07:19:10 +08:00
return (ldap_unicode_t *) u;
}
}
return NULL;
}
void ucstr2upper(
ldap_unicode_t *u,
ber_len_t n )
{
for(; 0 < n; ++u, --n ) {
*u = uctoupper( *u );
}
}
struct berval * UTF8bvnormalize(
struct berval *bv,
struct berval *newbv,
2003-04-11 11:57:10 +08:00
unsigned flags,
void *ctx )
{
int i, j, len, clen, outpos, ucsoutlen, outsize, last;
2002-10-09 03:45:01 +08:00
char *out, *outtmp, *s;
unsigned long *ucs, *p, *ucsout;
unsigned casefold = flags & LDAP_UTF8_CASEFOLD;
unsigned approx = flags & LDAP_UTF8_APPROX;
static unsigned char mask[] = {
0, 0x7f, 0x1f, 0x0f, 0x07, 0x03, 0x01 };
if ( bv == NULL ) {
return NULL;
}
s = bv->bv_val;
len = bv->bv_len;
if ( len == 0 ) {
2003-04-11 11:57:10 +08:00
return ber_dupbv_x( newbv, bv, ctx );
}
/* FIXME: Should first check to see if string is already in
* proper normalized form. This is almost as time consuming
* as the normalization though.
*/
/* finish off everything up to character before first non-ascii */
if ( LDAP_UTF8_ISASCII( s ) ) {
if ( casefold ) {
outsize = len + 7;
out = (char *) malloc( outsize );
if ( out == NULL ) {
return NULL;
}
outpos = 0;
for ( i = 1; (i < len) && LDAP_UTF8_ISASCII(s + i); i++ ) {
out[outpos++] = TOLOWER( s[i-1] );
}
if ( i == len ) {
out[outpos++] = TOLOWER( s[len - 1] );
out[outpos] = '\0';
return ber_str2bv( out, outpos, 0, newbv);
}
} else {
for ( i = 1; (i < len) && LDAP_UTF8_ISASCII(s + i); i++ ) {
/* empty */
}
if ( i == len ) {
2003-04-11 11:57:10 +08:00
return ber_str2bv_x( s, len, 1, newbv, ctx );
}
outsize = len + 7;
out = (char *) malloc( outsize );
if ( out == NULL ) {
return NULL;
}
outpos = i - 1;
memcpy(out, s, outpos);
}
} else {
outsize = len + 7;
out = (char *) malloc( outsize );
if ( out == NULL ) {
return NULL;
}
outpos = 0;
i = 0;
}
p = ucs = malloc( len * sizeof(*ucs) );
if ( ucs == NULL ) {
free(out);
return NULL;
}
/* convert character before first non-ascii to ucs-4 */
if ( i > 0 ) {
*p = casefold ? TOLOWER( s[i - 1] ) : s[i - 1];
p++;
}
/* s[i] is now first non-ascii character */
for (;;) {
/* s[i] is non-ascii */
/* convert everything up to next ascii to ucs-4 */
while ( i < len ) {
clen = LDAP_UTF8_CHARLEN2( s + i, clen );
if ( clen == 0 ) {
free( ucs );
free( out );
return NULL;
}
if ( clen == 1 ) {
/* ascii */
break;
}
*p = s[i] & mask[clen];
i++;
for( j = 1; j < clen; j++ ) {
if ( (s[i] & 0xc0) != 0x80 ) {
free( ucs );
free( out );
return NULL;
}
*p <<= 6;
*p |= s[i] & 0x3f;
i++;
}
if ( casefold ) {
*p = uctolower( *p );
}
p++;
}
/* normalize ucs of length p - ucs */
2003-04-11 11:57:10 +08:00
uccompatdecomp( ucs, p - ucs, &ucsout, &ucsoutlen, ctx );
if ( approx ) {
for ( j = 0; j < ucsoutlen; j++ ) {
if ( ucsout[j] < 0x80 ) {
out[outpos++] = ucsout[j];
}
}
} else {
ucsoutlen = uccanoncomp( ucsout, ucsoutlen );
/* convert ucs to utf-8 and store in out */
for ( j = 0; j < ucsoutlen; j++ ) {
/* allocate more space if not enough room for
6 bytes and terminator */
if ( outsize - outpos < 7 ) {
outsize = ucsoutlen - j + outpos + 6;
2002-10-09 03:45:01 +08:00
outtmp = (char *) realloc( out, outsize );
if ( outtmp == NULL ) {
free( out );
free( ucs );
2002-10-09 03:45:01 +08:00
free( ucsout );
return NULL;
}
2002-10-09 03:45:01 +08:00
out = outtmp;
}
outpos += ldap_x_ucs4_to_utf8( ucsout[j], &out[outpos] );
}
}
2002-10-09 03:45:01 +08:00
free( ucsout );
ucsout = NULL;
if ( i == len ) {
break;
}
last = i;
/* s[i] is ascii */
/* finish off everything up to char before next non-ascii */
for ( i++; (i < len) && LDAP_UTF8_ISASCII(s + i); i++ ) {
out[outpos++] = casefold ? TOLOWER( s[i-1] ) : s[i-1];
}
if ( i == len ) {
out[outpos++] = casefold ? TOLOWER( s[len - 1] ) : s[len - 1];
break;
}
/* convert character before next non-ascii to ucs-4 */
*ucs = casefold ? TOLOWER( s[i - 1] ) : s[i - 1];
p = ucs + 1;
}
free( ucs );
out[outpos] = '\0';
return ber_str2bv( out, outpos, 0, newbv );
}
/* compare UTF8-strings, optionally ignore casing */
/* slow, should be optimized */
int UTF8bvnormcmp(
struct berval *bv1,
struct berval *bv2,
2003-04-11 11:57:10 +08:00
unsigned flags,
void *ctx )
{
int i, l1, l2, len, ulen, res = 0;
char *s1, *s2, *done;
unsigned long *ucs, *ucsout1, *ucsout2;
unsigned casefold = flags & LDAP_UTF8_CASEFOLD;
unsigned norm1 = flags & LDAP_UTF8_ARG1NFC;
unsigned norm2 = flags & LDAP_UTF8_ARG2NFC;
if (bv1 == NULL) {
return bv2 == NULL ? 0 : -1;
} else if (bv2 == NULL) {
return 1;
}
l1 = bv1->bv_len;
l2 = bv2->bv_len;
len = (l1 < l2) ? l1 : l2;
if (len == 0) {
return l1 == 0 ? (l2 == 0 ? 0 : -1) : 1;
}
s1 = bv1->bv_val;
s2 = bv2->bv_val;
done = s1 + len;
while ( (s1 < done) && LDAP_UTF8_ISASCII(s1) && LDAP_UTF8_ISASCII(s2) ) {
if (casefold) {
char c1 = TOLOWER(*s1);
char c2 = TOLOWER(*s2);
res = c1 - c2;
} else {
res = *s1 - *s2;
}
s1++;
s2++;
if (res) {
/* done unless next character in s1 or s2 is non-ascii */
if (s1 < done) {
if (!LDAP_UTF8_ISASCII(s1) || !LDAP_UTF8_ISASCII(s2)) {
break;
}
} else if (((len < l1) && !LDAP_UTF8_ISASCII(s1)) ||
((len < l2) && !LDAP_UTF8_ISASCII(s2))) {
break;
}
return res;
}
}
/* We have encountered non-ascii or strings equal up to len */
/* set i to number of iterations */
i = s1 - done + len;
/* passed through loop at least once? */
if (i > 0) {
if (!res && (s1 == done) &&
((len == l1) || LDAP_UTF8_ISASCII(s1)) &&
((len == l2) || LDAP_UTF8_ISASCII(s2))) {
/* all ascii and equal up to len */
return l1 - l2;
}
/* rewind one char, and do normalized compare from there */
s1--;
s2--;
l1 -= i - 1;
l2 -= i - 1;
}
/* FIXME: Should first check to see if strings are already in
* proper normalized form.
*/
ucs = malloc( ( ( norm1 || l1 > l2 ) ? l1 : l2 ) * sizeof(*ucs) );
if ( ucs == NULL ) {
return l1 > l2 ? 1 : -1; /* what to do??? */
}
/*
* XXYYZ: we convert to ucs4 even though -llunicode
* expects ucs2 in an unsigned long
*/
/* convert and normalize 1st string */
for ( i = 0, ulen = 0; i < l1; i += len, ulen++ ) {
ucs[ulen] = ldap_x_utf8_to_ucs4( s1 + i );
if ( ucs[ulen] == LDAP_UCS4_INVALID ) {
free( ucs );
return -1; /* what to do??? */
}
len = LDAP_UTF8_CHARLEN( s1 + i );
}
if ( norm1 ) {
ucsout1 = ucs;
l1 = ulen;
ucs = malloc( l2 * sizeof(*ucs) );
if ( ucs == NULL ) {
return l1 > l2 ? 1 : -1; /* what to do??? */
}
} else {
2003-04-11 11:57:10 +08:00
uccompatdecomp( ucs, ulen, &ucsout1, &l1, ctx );
l1 = uccanoncomp( ucsout1, l1 );
}
/* convert and normalize 2nd string */
for ( i = 0, ulen = 0; i < l2; i += len, ulen++ ) {
ucs[ulen] = ldap_x_utf8_to_ucs4( s2 + i );
if ( ucs[ulen] == LDAP_UCS4_INVALID ) {
free( ucsout1 );
free( ucs );
return 1; /* what to do??? */
}
len = LDAP_UTF8_CHARLEN( s2 + i );
}
if ( norm2 ) {
ucsout2 = ucs;
l2 = ulen;
} else {
2003-04-11 11:57:10 +08:00
uccompatdecomp( ucs, ulen, &ucsout2, &l2, ctx );
l2 = uccanoncomp( ucsout2, l2 );
free( ucs );
}
res = casefold
? ucstrncasecmp( ucsout1, ucsout2, l1 < l2 ? l1 : l2 )
: ucstrncmp( ucsout1, ucsout2, l1 < l2 ? l1 : l2 );
free( ucsout1 );
free( ucsout2 );
if ( res != 0 ) {
return res;
}
if ( l1 == l2 ) {
return 0;
}
return l1 > l2 ? 1 : -1;
}