mirror/openldap
2
0
Fork 0
mirror of https://git.openldap.org/openldap/openldap.git synced 2025-01-30 13:30:57 +08:00
Code Issues Packages Projects Releases Wiki Activity

Use <lber.h> ber_memalloc, not malloc. Shortcuts for pure-ASCII strings.

Browse Source 
Still need to add some kind of isNormalized() check to avoid rework on
UTF8 strings that are already properly normalized.
This commit is contained in:
Howard Chu 2002-01-03 16:22:53 +00:00
parent 33e6f6e65c
commit 8c04993cdf
1 changed files with 51 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

55
libraries/liblunicode/ucstr.c
View File

@ -4,6 +4,12 @@
#include <ac/string.h> #include <ac/string.h>
#include <ac/stdlib.h> #include <ac/stdlib.h>
#include <lber.h>
#define malloc(x) ber_memalloc(x)
#define realloc(x,y) ber_memrealloc(x,y)
#define free(x) ber_memfree(x)
#include <ldap_utf8.h> #include <ldap_utf8.h>
#include <ldap_pvt_uc.h> #include <ldap_pvt_uc.h>
@ -99,18 +105,35 @@ char * UTF8normalize(
s = bv->bv_val; s = bv->bv_val;
len = bv->bv_len; len = bv->bv_len;
if ( len == 0 ) { /* See if the string is pure ASCII so we can shortcut */
out = (char *) malloc( 1 ); for ( i=0; i<len; i++ ) {
*out = '\0'; if ( s[i] & 0x80 ) /* non-ASCII */
break;
}
/* It's pure ASCII or zero-len */
if ( i == len ) {
out = malloc( len + 1 );
if ( i && !casefold ) {
strncpy( out, bv->bv_val, len );
} else {
for ( j=0; j<i; j++ )
out[j] = TOUPPER( s[j] );
}
out[len] = '\0';
return out; return out;
} }
outsize = len + 7; outsize = len + 7;
out = (char *) malloc( outsize ); out = (char *) malloc( outsize );
if ( out == NULL ) { if ( out == NULL ) {
return NULL; return NULL;
} }
/* FIXME: Should first check to see if string is already in
* proper normalized form.
*/
outpos = 0; outpos = 0;
/* finish off everything up to character before first non-ascii */ /* finish off everything up to character before first non-ascii */
@ -234,6 +257,30 @@ int UTF8normcmp(
return *s1 - *s2 > 0 ? 1 : -1; return *s1 - *s2 > 0 ? 1 : -1;
} }
/* See if we can get away with a straight ASCII compare */
len = (l1 < l2) ? l1 : l2;
for ( i = 0; i<len; i++ ) {
/* Is either char non-ASCII? */
if ((s1[i] & 0x80) || (s2[i] & 0x80))
break;
if (casefold) {
char c1 = TOUPPER(s1[i]);
char c2 = TOUPPER(s2[i]);
res = c1 - c2;
} else {
res = s1[i] - s2[i];
}
if (res)
return res;
}
/* Strings were ASCII, equal up to minlen */
if (i == len)
return l1 - l2;
/* FIXME: Should first check to see if strings are already in
* proper normalized form.
*/
ucs = (long *) malloc( ( l1 > l2 ? l1 : l2 ) * sizeof(*ucs) ); ucs = (long *) malloc( ( l1 > l2 ? l1 : l2 ) * sizeof(*ucs) );
if ( ucs == NULL ) { if ( ucs == NULL ) {
return l1 > l2 ? 1 : -1; /* what to do??? */ return l1 > l2 ? 1 : -1; /* what to do??? */