Update wcscoll forgotten in previous patch
This commit is contained in:
parent
cbdca8b2ff
commit
a1df81e716
Notes:
svn2git
2020-12-20 02:59:44 +00:00
svn path=/projects/collation/; revision=286535
@ -1,4 +1,5 @@
|
||||
/*-
|
||||
* Copyright 2010 Nexenta Systems, Inc. All rights reserved.
|
||||
* Copyright (c) 2002 Tim J. Robbins
|
||||
* All rights reserved.
|
||||
*
|
||||
@ -38,50 +39,162 @@ __FBSDID("$FreeBSD$");
|
||||
#include <wchar.h>
|
||||
#include "collate.h"
|
||||
|
||||
static char *__mbsdup(const wchar_t *);
|
||||
|
||||
/*
|
||||
* Placeholder implementation of wcscoll(). Attempts to use the single-byte
|
||||
* collation ordering where possible, and falls back on wcscmp() in locales
|
||||
* with extended character sets.
|
||||
*/
|
||||
int
|
||||
wcscoll_l(const wchar_t *ws1, const wchar_t *ws2, locale_t locale)
|
||||
{
|
||||
char *mbs1, *mbs2;
|
||||
int diff, sverrno;
|
||||
int len1, len2, pri1, pri2, ret;
|
||||
wchar_t *tr1 = NULL, *tr2 = NULL;
|
||||
int direc, pass;
|
||||
|
||||
FIX_LOCALE(locale);
|
||||
struct xlocale_collate *table =
|
||||
(struct xlocale_collate*)locale->components[XLC_COLLATE];
|
||||
|
||||
if (table->__collate_load_error || MB_CUR_MAX > 1)
|
||||
if (table->__collate_load_error)
|
||||
/*
|
||||
* Locale has no special collating order, could not be
|
||||
* loaded, or has an extended character set; do a fast binary
|
||||
* comparison.
|
||||
* Locale has no special collating order or could not be
|
||||
* loaded, do a fast binary comparison.
|
||||
*/
|
||||
return (wcscmp(ws1, ws2));
|
||||
|
||||
if ((mbs1 = __mbsdup(ws1)) == NULL || (mbs2 = __mbsdup(ws2)) == NULL) {
|
||||
ret = 0;
|
||||
|
||||
/*
|
||||
* Out of memory or illegal wide chars; fall back to wcscmp()
|
||||
* but leave errno indicating the error. Callers that don't
|
||||
* check for error will get a reasonable but often slightly
|
||||
* incorrect result.
|
||||
* Once upon a time we had code to try to optimize this, but
|
||||
* it turns out that you really can't make many assumptions
|
||||
* safely. You absolutely have to run this pass by pass,
|
||||
* because some passes will be ignored for a given character,
|
||||
* while others will not. Simpler locales will benefit from
|
||||
* having fewer passes, and most comparisions should resolve
|
||||
* during the primary pass anyway.
|
||||
*
|
||||
* Note that we do one final extra pass at the end to pick
|
||||
* up UNDEFINED elements. There is special handling for them.
|
||||
*/
|
||||
sverrno = errno;
|
||||
free(mbs1);
|
||||
errno = sverrno;
|
||||
return (wcscmp(ws1, ws2));
|
||||
for (pass = 0; pass <= table->info->directive_count; pass++) {
|
||||
|
||||
const int32_t *st1 = NULL;
|
||||
const int32_t *st2 = NULL;
|
||||
const wchar_t *w1 = ws1;
|
||||
const wchar_t *w2 = ws2;
|
||||
|
||||
/* special pass for UNDEFINED */
|
||||
if (pass == table->info->directive_count) {
|
||||
direc = DIRECTIVE_FORWARD | DIRECTIVE_UNDEFINED;
|
||||
} else {
|
||||
direc = table->info->directive[pass];
|
||||
}
|
||||
|
||||
diff = strcoll_l(mbs1, mbs2, locale);
|
||||
sverrno = errno;
|
||||
free(mbs1);
|
||||
free(mbs2);
|
||||
errno = sverrno;
|
||||
if (direc & DIRECTIVE_BACKWARD) {
|
||||
wchar_t *bp, *fp, c;
|
||||
if ((tr1 = wcsdup(w1)) == NULL)
|
||||
goto fail;
|
||||
bp = tr1;
|
||||
fp = tr1 + wcslen(tr1) - 1;
|
||||
while (bp < fp) {
|
||||
c = *bp;
|
||||
*bp++ = *fp;
|
||||
*fp-- = c;
|
||||
}
|
||||
if ((tr2 = wcsdup(w2)) == NULL)
|
||||
goto fail;
|
||||
bp = tr2;
|
||||
fp = tr2 + wcslen(tr2) - 1;
|
||||
while (bp < fp) {
|
||||
c = *bp;
|
||||
*bp++ = *fp;
|
||||
*fp-- = c;
|
||||
}
|
||||
w1 = tr1;
|
||||
w2 = tr2;
|
||||
}
|
||||
|
||||
return (diff);
|
||||
if (direc & DIRECTIVE_POSITION) {
|
||||
while ((*w1 || st1) && (*w2 || st2)) {
|
||||
pri1 = pri2 = 0;
|
||||
_collate_lookup(table, w1, &len1, &pri1, pass,
|
||||
&st1);
|
||||
if (pri1 <= 0) {
|
||||
if (pri1 < 0) {
|
||||
errno = EINVAL;
|
||||
goto fail;
|
||||
}
|
||||
pri1 = COLLATE_MAX_PRIORITY;
|
||||
}
|
||||
_collate_lookup(table, w2, &len2, &pri2, pass,
|
||||
&st2);
|
||||
if (pri2 <= 0) {
|
||||
if (pri2 < 0) {
|
||||
errno = EINVAL;
|
||||
goto fail;
|
||||
}
|
||||
pri2 = COLLATE_MAX_PRIORITY;
|
||||
}
|
||||
if (pri1 != pri2) {
|
||||
ret = pri1 - pri2;
|
||||
goto end;
|
||||
}
|
||||
w1 += len1;
|
||||
w2 += len2;
|
||||
}
|
||||
} else {
|
||||
while ((*w1 || st1) && (*w2 || st2)) {
|
||||
pri1 = pri2 = 0;
|
||||
while (*w1) {
|
||||
_collate_lookup(table, w1, &len1,
|
||||
&pri1, pass, &st1);
|
||||
if (pri1 > 0)
|
||||
break;
|
||||
if (pri1 < 0) {
|
||||
errno = EINVAL;
|
||||
goto fail;
|
||||
}
|
||||
w1 += len1;
|
||||
}
|
||||
while (*w2) {
|
||||
_collate_lookup(table, w2, &len2,
|
||||
&pri2, pass, &st2);
|
||||
if (pri2 > 0)
|
||||
break;
|
||||
if (pri2 < 0) {
|
||||
errno = EINVAL;
|
||||
goto fail;
|
||||
}
|
||||
w2 += len2;
|
||||
}
|
||||
if (!pri1 || !pri2)
|
||||
break;
|
||||
if (pri1 != pri2) {
|
||||
ret = pri1 - pri2;
|
||||
goto end;
|
||||
}
|
||||
w1 += len1;
|
||||
w2 += len2;
|
||||
}
|
||||
}
|
||||
if (!*w1) {
|
||||
if (*w2) {
|
||||
ret = -(int)*w2;
|
||||
goto end;
|
||||
}
|
||||
} else {
|
||||
ret = *w1;
|
||||
goto end;
|
||||
}
|
||||
}
|
||||
ret = 0;
|
||||
|
||||
end:
|
||||
if (tr1)
|
||||
free(tr1);
|
||||
if (tr2)
|
||||
free(tr2);
|
||||
|
||||
return (ret);
|
||||
|
||||
fail:
|
||||
ret = wcscmp(ws1, ws2);
|
||||
goto end;
|
||||
}
|
||||
|
||||
int
|
||||
@ -89,24 +202,3 @@ wcscoll(const wchar_t *ws1, const wchar_t *ws2)
|
||||
{
|
||||
return wcscoll_l(ws1, ws2, __get_locale());
|
||||
}
|
||||
|
||||
static char *
|
||||
__mbsdup(const wchar_t *ws)
|
||||
{
|
||||
static const mbstate_t initial;
|
||||
mbstate_t st;
|
||||
const wchar_t *wcp;
|
||||
size_t len;
|
||||
char *mbs;
|
||||
|
||||
wcp = ws;
|
||||
st = initial;
|
||||
if ((len = wcsrtombs(NULL, &wcp, 0, &st)) == (size_t)-1)
|
||||
return (NULL);
|
||||
if ((mbs = malloc(len + 1)) == NULL)
|
||||
return (NULL);
|
||||
st = initial;
|
||||
wcsrtombs(mbs, &ws, len + 1, &st);
|
||||
|
||||
return (mbs);
|
||||
}
|
||||
|
Loading…
Reference in New Issue
Block a user