/* This function is equivalent to strncasecmp() for multibyte
* strings. */
int mbstrncasecmp(const char *s1, const char *s2, size_t n)
{
#ifdef ENABLE_UTF8
if (use_utf8) {
char *s1_mb, *s2_mb;
wchar_t ws1, ws2;
if (s1 == s2)
return 0;
assert(s1 != NULL && s2 != NULL);
s1_mb = charalloc(MB_CUR_MAX);
s2_mb = charalloc(MB_CUR_MAX);
for (; *s1 != '\0' && *s2 != '\0' && n > 0; s1 +=
move_mbright(s1, 0), s2 += move_mbright(s2, 0), n--) {
bool bad_s1_mb = FALSE, bad_s2_mb = FALSE;
int s1_mb_len, s2_mb_len;
s1_mb_len = parse_mbchar(s1, s1_mb, NULL);
if (mbtowc(&ws1, s1_mb, s1_mb_len) < 0) {
mbtowc_reset();
ws1 = (unsigned char)*s1_mb;
bad_s1_mb = TRUE;
}
s2_mb_len = parse_mbchar(s2, s2_mb, NULL);
if (mbtowc(&ws2, s2_mb, s2_mb_len) < 0) {
mbtowc_reset();
ws2 = (unsigned char)*s2_mb;
bad_s2_mb = TRUE;
}
if (bad_s1_mb != bad_s2_mb || towlower(ws1) !=
towlower(ws2))
break;
}
free(s1_mb);
free(s2_mb);
return (n > 0) ? towlower(ws1) - towlower(ws2) : 0;
} else
#endif
return strncasecmp(s1, s2, n);
}
/* This function is equivalent to wcwidth() for multibyte characters. */
int mbwidth(const char *c)
{
assert(c != NULL);
#ifdef ENABLE_UTF8
if (use_utf8) {
wchar_t wc;
int width;
if (mbtowc(&wc, c, MB_CUR_MAX) < 0) {
mbtowc_reset();
wc = bad_wchar;
}
width = wcwidth(wc);
if (width == -1) {
wc = bad_wchar;
width = wcwidth(wc);
}
return width;
} else
#endif
return 1;
}
/* c is a multibyte non-control character. We return that multibyte
* character. If crep is an invalid multibyte sequence, it will be
* replaced with Unicode 0xFFFD (Replacement Character). */
char *mbrep(const char *c, char *crep, int *crep_len)
{
assert(c != NULL && crep != NULL && crep_len != NULL);
#ifdef ENABLE_UTF8
if (use_utf8) {
wchar_t wc;
/* Reject invalid Unicode characters. */
if (mbtowc(&wc, c, MB_CUR_MAX) < 0 || !is_valid_unicode(wc)) {
mbtowc_reset();
*crep_len = bad_mbchar_len;
strncpy(crep, bad_mbchar, *crep_len);
} else {
*crep_len = wctomb(crep, wc);
if (*crep_len < 0) {
wctomb_reset();
*crep_len = 0;
}
}
} else {
#endif
*crep_len = 1;
*crep = *c;
#ifdef ENABLE_UTF8
}
#endif
return crep;
}
/* c is a multibyte control character. It displays as ^@, ^?, or ^[ch],
* where ch is (c + 64). We return that multibyte character. If crep
* is an invalid multibyte sequence, it will be replaced with Unicode
* 0xFFFD (Replacement Character). */
char *control_mbrep(const char *c, char *crep, int *crep_len)
{
assert(c != NULL && crep != NULL && crep_len != NULL);
#ifdef ENABLE_UTF8
if (use_utf8) {
wchar_t wc;
if (mbtowc(&wc, c, MB_CUR_MAX) < 0) {
mbtowc_reset();
*crep_len = bad_mbchar_len;
strncpy(crep, bad_mbchar, *crep_len);
} else {
*crep_len = wctomb(crep, control_wrep(wc));
if (*crep_len < 0) {
wctomb_reset();
*crep_len = 0;
}
}
} else {
#endif
*crep_len = 1;
*crep = control_rep(*c);
#ifdef ENABLE_UTF8
}
#endif
return crep;
}
/* Assess how many bytes the given (multibyte) character occupies. Return -1
* if the byte sequence is invalid, and return the number of bytes minus 8
* when it encodes an invalid codepoint. Also, in the second parameter,
* return the number of columns that the character occupies. */
int length_of_char(const char *c, int *width)
{
assert(c != NULL);
#ifdef ENABLE_UTF8
if (use_utf8) {
wchar_t wc;
int charlen = mbtowc(&wc, c, MB_CUR_MAX);
/* If the sequence is invalid... */
if (charlen < 0) {
mbtowc_reset();
return -1;
}
/* If the codepoint is invalid... */
if (!is_valid_unicode(wc))
return charlen - 8;
else {
*width = wcwidth(wc);
/* If the codepoint is unassigned, assume a width of one. */
if (*width < 0)
*width = 1;
return charlen;
}
} else
#endif
return 1;
}
/* This function is equivalent to strchr() for multibyte strings. */
char *mbstrchr(const char *s, const char *c)
{
assert(s != NULL && c != NULL);
#ifdef ENABLE_UTF8
if (use_utf8) {
bool bad_s_mb = FALSE, bad_c_mb = FALSE;
char *s_mb = charalloc(MB_CUR_MAX);
const char *q = s;
wchar_t ws, wc;
int c_mb_len = mbtowc(&wc, c, MB_CUR_MAX);
if (c_mb_len < 0) {
mbtowc_reset();
wc = (unsigned char)*c;
bad_c_mb = TRUE;
}
while (*s != '\0') {
int s_mb_len = parse_mbchar(s, s_mb, NULL);
if (mbtowc(&ws, s_mb, s_mb_len) < 0) {
mbtowc_reset();
ws = (unsigned char)*s;
bad_s_mb = TRUE;
}
if (bad_s_mb == bad_c_mb && ws == wc)
break;
s += s_mb_len;
q += s_mb_len;
}
free(s_mb);
if (*s == '\0')
q = NULL;
return (char *)q;
} else
#endif
return strchr(s, *c);
}
/* This function is equivalent to strncasecmp() for multibyte strings. */
int mbstrncasecmp(const char *s1, const char *s2, size_t n)
{
#ifdef ENABLE_UTF8
if (use_utf8) {
wchar_t wc1, wc2;
while (*s1 != '\0' && *s2 != '\0' && n > 0) {
bool bad1 = FALSE, bad2 = FALSE;
if (mbtowc(&wc1, s1, MB_CUR_MAX) < 0) {
mbtowc_reset();
bad1 = TRUE;
}
if (mbtowc(&wc2, s2, MB_CUR_MAX) < 0) {
mbtowc_reset();
bad2 = TRUE;
}
if (bad1 || bad2) {
if (*s1 != *s2)
return (unsigned char)*s1 - (unsigned char)*s2;
if (bad1 != bad2)
return (bad1 ? 1 : -1);
} else {
int difference = towlower(wc1) - towlower(wc2);
if (difference != 0)
return difference;
}
s1 += move_mbright(s1, 0);
s2 += move_mbright(s2, 0);
n--;
}
return (n > 0) ? ((unsigned char)*s1 - (unsigned char)*s2) : 0;
} else
#endif
return strncasecmp(s1, s2, n);
}
/* This function is equivalent to isalnum() for multibyte characters. */
bool is_alnum_mbchar(const char *c)
{
assert(c != NULL);
#ifdef ENABLE_UTF8
if (use_utf8) {
wchar_t wc;
if (mbtowc(&wc, c, MB_CUR_MAX) < 0) {
mbtowc_reset();
wc = bad_wchar;
}
return iswalnum(wc);
} else
#endif
return isalnum((unsigned char)*c);
}
/* This function is equivalent to ispunct() for multibyte characters. */
bool is_punct_mbchar(const char *c)
{
assert(c != NULL);
#ifdef ENABLE_UTF8
if (use_utf8) {
wchar_t wc;
if (mbtowc(&wc, c, MB_CUR_MAX) < 0) {
mbtowc_reset();
return 0;
}
return iswpunct(wc);
} else
#endif
return ispunct((unsigned char)*c);
}