return error_quark;
}
+static gunichar*
+utf8_case_conv (const gchar *str, gssize len, gboolean upper)
+{
+ glong i, u16len, u32len;
+ gunichar2 *u16str;
+ gunichar *u32str;
+ gchar *u8str;
+ GError **err = NULL;
+
+ u16str = g_utf8_to_utf16 (str, (glong)len, NULL, &u16len, err);
+ u32str = g_utf16_to_ucs4 (u16str, u16len, NULL, &u32len, err);
+ for (i = 0; i < u32len; i++) {
+ u32str [i] = upper ? g_unichar_toupper (u32str [i]) : g_unichar_tolower (u32str [i]);
+ }
+ g_free (u16str);
+ u16str = g_ucs4_to_utf16 (u32str, u32len, NULL, &u16len, err);
+ u8str = g_utf16_to_utf8 (u16str, u16len, NULL, NULL, err);
+ g_free (u32str);
+ g_free (u16str);
+ return (gunichar*)u8str;
+}
+
+gchar*
+g_utf8_strup (const gchar *str, gssize len)
+{
+ return (gchar*)utf8_case_conv (str, len, TRUE);
+}
+
+gchar*
+g_utf8_strdown (const gchar *str, gssize len)
+{
+ return (gchar*)utf8_case_conv (str, len, FALSE);
+}
+
gunichar2*
g_utf8_to_utf16 (const gchar *str, glong len, glong *items_read, glong *items_written, GError **error)
{
if (--mb_remain == 0) {
/* multi byte character is fully consumed now. */
if (codepoint < 0x10000) {
- ret [out_pos++] = codepoint % 0x10000;
+ ret [out_pos++] = (gunichar2)(codepoint % 0x10000);
} else if (codepoint < 0x110000) {
/* surrogate pair */
codepoint -= 0x10000;
- ret [out_pos++] = (codepoint >> 10) + 0xD800;
- ret [out_pos++] = (codepoint & 0x3FF) + 0xDC00;
+ ret [out_pos++] = (gunichar2)((codepoint >> 10) + 0xD800);
+ ret [out_pos++] = (gunichar2)((codepoint & 0x3FF) + 0xDC00);
} else {
/* invalid utf-8 sequence (excess) */
codepoint = 0;
gchar *ret;
glong in_pos, out_pos;
gunichar2 ch;
- guint32 codepoint;
+ guint32 codepoint = 0;
gboolean surrogate;
in_pos = 0;
while (len < 0 ? str [in_pos] : in_pos < len) {
ch = str [in_pos];
if (surrogate) {
- surrogate = 0;
- if (ch >= 0xDC00 && ch <= 0xDFFF)
+ if (ch >= 0xDC00 && ch <= 0xDFFF) {
codepoint = 0x10000 + (ch - 0xDC00) + ((surrogate - 0xD800) << 10);
- else
+ surrogate = 0;
+ } else {
+ surrogate = 0;
/* invalid surrogate pair */
continue;
+ }
} else {
/* fast path optimization */
if (ch < 0x80) {
for (; len < 0 ? str [in_pos] : in_pos < len; in_pos++) {
if (str [in_pos] < 0x80)
- ret [out_pos++] = str [in_pos];
+ ret [out_pos++] = (gchar)(str [in_pos]);
else
break;
}
}
in_pos++;
+ if (surrogate != 0)
+ continue;
if (codepoint < 0x80)
ret [out_pos++] = (gchar) codepoint;
else if (codepoint < 0x0800) {
*items_read = in_pos;
return ret;
}
+
+static glong
+g_ucs4_to_utf16_len (const gunichar *str, glong len, glong *items_read, GError **error)
+{
+ glong retlen = 0;
+ glong errindex = 0;
+ const gunichar *lstr = str;
+
+ if (!str)
+ return 0;
+
+ while (*lstr != '\0' && len--) {
+ gunichar ch;
+ ch = *lstr++;
+ if (ch <= 0x0000FFFF) {
+ if (ch >= 0xD800 && ch <= 0xDFFF) {
+ errindex = (glong)(lstr - str)-1;
+ if (error)
+ g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
+ "Invalid sequence in conversion input");
+ if (items_read)
+ *items_read = errindex;
+ return 0;
+ } else {
+ retlen++;
+ }
+ } else if (ch > 0x10FFFF) {
+ errindex = (glong)(lstr - str)-1;
+ if (error)
+ g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
+ "Character out of range for UTF-16");
+ if (items_read)
+ *items_read = errindex;
+ return 0;
+
+ } else {
+ retlen+=2;
+ }
+ }
+
+ if (items_read)
+ *items_read = (glong)(lstr - str);
+ return retlen;
+}
+
+gunichar2*
+g_ucs4_to_utf16 (const gunichar *str, glong len, glong *items_read, glong *items_written, GError **error)
+{
+ glong allocsz;
+ gunichar2 *retstr = 0;
+ gunichar2 *retch = 0;
+ glong nwritten = 0;
+ GError *lerror =0 ;
+
+ allocsz = g_ucs4_to_utf16_len (str, len, items_read, &lerror);
+
+ if (!lerror) {
+ retch = retstr = g_malloc ((allocsz+1) * sizeof (gunichar2));
+ retstr[allocsz] = '\0';
+
+ while (*str != '\0' && len--) {
+ gunichar ch;
+ ch = *str++;
+ if (ch <= 0x0000FFFF && (ch < 0xD800 || ch > 0xDFFF)) {
+ *retch++ = (gunichar2)ch;
+ nwritten ++;
+ } else {
+ ch -= 0x0010000UL;
+ *retch++ = (gunichar2)((ch >> 10) + 0xD800);
+ *retch++ = (gunichar2)((ch & 0x3FFUL) + 0xDC00);
+ nwritten +=2;
+ }
+ }
+ }
+
+ if (items_written)
+ *items_written = nwritten;
+ if (error)
+ *error = lerror;
+
+ return retstr;
+}
+
+static glong
+g_utf16_to_ucs4_len (const gunichar2 *str, glong len, glong *items_read, GError **error)
+{
+ glong retlen = 0;
+ glong errindex = 0;
+ const gunichar2 *lstr = str;
+ gunichar2 ch,ch2;
+
+ if (!str)
+ return 0;
+
+ while (*lstr != '\0' && len--) {
+ ch = *lstr++;
+ if (ch >= 0xD800 && ch <= 0xDBFF) {
+ if (!len--) {
+ lstr--;
+ break;
+ }
+ ch2 = *lstr;
+ if (ch2 >= 0xDC00 && ch2 <= 0xDFFF) {
+ lstr++;
+ } else {
+ errindex = (glong)(lstr - str);
+ if (error)
+ g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
+ "Invalid sequence in conversion input");
+ if (items_read)
+ *items_read = errindex;
+ return 0;
+ }
+ } else {
+ if (ch >= 0xDC00 && ch <= 0xDFFF) {
+ errindex = (glong)(lstr - str)-1;
+ if (error)
+ g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
+ "Invalid sequence in conversion input");
+ if (items_read)
+ *items_read = errindex;
+ return 0;
+ }
+ }
+ retlen++;
+ }
+
+ if (items_read)
+ *items_read = (glong)(lstr - str);
+
+ return retlen;
+}
+
+gunichar*
+g_utf16_to_ucs4 (const gunichar2 *str, glong len, glong *items_read, glong *items_written, GError **error)
+{
+ glong allocsz;
+ gunichar *retstr = 0;
+ gunichar *retch = 0;
+ glong nwritten = 0;
+ GError *lerror =0 ;
+ gunichar ch,ch2;
+
+ allocsz = g_utf16_to_ucs4_len (str, len, items_read, &lerror);
+
+ if (!lerror) {
+ retch = retstr = g_malloc ((allocsz+1) * sizeof (gunichar));
+ retstr[allocsz] = '\0';
+ nwritten = allocsz;
+
+ while (*str != '\0' && allocsz--) {
+ ch = *str++;
+ if (ch >= 0xD800 && ch <= 0xDBFF) {
+ ch2 = *str++;
+ ch = ((ch - (gunichar)0xD800) << 10)
+ + (ch2 - (gunichar)0xDC00) + (gunichar)0x0010000UL;
+ }
+ *retch++ = ch;
+ }
+ }
+
+ if (items_written)
+ *items_written = nwritten;
+ if (error)
+ *error = lerror;
+
+ return retstr;
+}
projects / mono.git / blobdiff