*
* (c) 1997--2004 Martin Mares <mj@ucw.cz>
* (c) 2004 Robert Spalek <robert@ucw.cz>
+ * (c) 2007 Pavel Charvat <pchar@ucw.cz>
*
* This software may be freely distributed and used according to the terms
* of the GNU Lesser General Public License.
#ifndef _UCW_UNICODE_H
#define _UCW_UNICODE_H
+#include "lib/unaligned.h"
+
/* Macros for handling UTF-8 */
#define UNI_REPLACEMENT 0xfffc
+/* Encode a character from the basic multilingual plane [0, 0xFFFF]
+ * (subset of Unicode 4.0); up to 3 bytes needed (RFC2279) */
static inline byte *
utf8_put(byte *p, uns u)
{
return p;
}
+/* Encode a value from the range [0, 0x7FFFFFFF];
+ * (superset of Unicode 4.0) up to 6 bytes needed (RFC2279) */
static inline byte *
utf8_32_put(byte *p, uns u)
{
#define UTF8_GET_NEXT if (unlikely((*p & 0xc0) != 0x80)) goto bad; u = (u << 6) | (*p++ & 0x3f)
-static inline const byte *
-utf8_get(const byte *p, uns *uu)
+/* Decode a character from the basic multilingual plane [0, 0xFFFF]
+ * or return 'repl' if the encoding has been corrupted */
+static inline byte *
+utf8_get_repl(const byte *p, uns *uu, uns repl)
{
uns u = *p++;
if (u < 0x80)
{
/* Incorrect byte sequence */
bad:
- u = UNI_REPLACEMENT;
+ u = repl;
}
else if (u < 0xe0)
{
else
goto bad;
*uu = u;
- return p;
+ return (byte *)p;
}
+/* Decode a value from the range [0, 0x7FFFFFFF]
+ * or return 'repl' if the encoding has been corrupted */
static inline byte *
-utf8_32_get(byte *p, uns *uu)
+utf8_32_get_repl(const byte *p, uns *uu, uns repl)
{
uns u = *p++;
if (u < 0x80)
{
/* Incorrect byte sequence */
bad:
- u = UNI_REPLACEMENT;
+ u = repl;
}
else if (u < 0xe0)
{
else
goto bad;
*uu = u;
- return p;
+ return (byte *)p;
+}
+
+/* Decode a character from the basic multilingual plane [0, 0xFFFF]
+ * or return UNI_REPLACEMENT if the encoding has been corrupted */
+static inline byte *
+utf8_get(const byte *p, uns *uu)
+{
+ return utf8_get_repl(p, uu, UNI_REPLACEMENT);
+}
+
+/* Decode a value from the range [0, 0x7FFFFFFF]
+ * or return UNI_REPLACEMENT if the encoding has been corrupted */
+static inline byte *
+utf8_32_get(const byte *p, uns *uu)
+{
+ return utf8_32_get_repl(p, uu, UNI_REPLACEMENT);
}
#define PUT_UTF8(p,u) p = utf8_put(p, u)
return 6;
}
+/* Encode a character from the range [0, 0xD7FF] or [0xE000,0x11FFFF];
+ * up to 4 bytes needed */
+static inline void *
+utf16_le_put(void *p, uns u)
+{
+ if (u < 0xd800 || (u < 0x10000 && u >= 0xe000))
+ {
+ put_u16_le(p, u);
+ return p + 2;
+ }
+ else if ((u -= 0x10000) < 0x100000)
+ {
+ put_u16_le(p, 0xd800 | (u >> 10));
+ put_u16_le(p + 2, 0xdc00 | (u & 0x3ff));
+ return p + 4;
+ }
+ else
+ ASSERT(0);
+}
+
+static inline void *
+utf16_be_put(void *p, uns u)
+{
+ if (u < 0xd800 || (u < 0x10000 && u >= 0xe000))
+ {
+ put_u16_be(p, u);
+ return p + 2;
+ }
+ else if ((u -= 0x10000) < 0x100000)
+ {
+ put_u16_be(p, 0xd800 | (u >> 10));
+ put_u16_be(p + 2, 0xdc00 | (u & 0x3ff));
+ return p + 4;
+ }
+ else
+ ASSERT(0);
+}
+
+/* Decode a character from the range [0, 0xD7FF] or [0xE000,11FFFF]
+ * or return `repl' if the encoding has been corrupted */
+static inline void *
+utf16_le_get_repl(const void *p, uns *uu, uns repl)
+{
+ uns u = get_u16_le(p), x, y;
+ x = u - 0xd800;
+ if (x < 0x800)
+ if (x < 0x400 && (y = get_u16_le(p + 2) - 0xdc00) < 0x400)
+ {
+ u = 0x10000 + (x << 10) + y;
+ p += 2;
+ }
+ else
+ u = repl;
+ *uu = u;
+ return (void *)(p + 2);
+}
+
+static inline void *
+utf16_be_get_repl(const void *p, uns *uu, uns repl)
+{
+ uns u = get_u16_be(p), x, y;
+ x = u - 0xd800;
+ if (x < 0x800)
+ if (x < 0x400 && (y = get_u16_be(p + 2) - 0xdc00) < 0x400)
+ {
+ u = 0x10000 + (x << 10) + y;
+ p += 2;
+ }
+ else
+ u = repl;
+ *uu = u;
+ return (void *)(p + 2);
+}
+
+/* Decode a character from the range [0, 0xD7FF] or [0xE000,11FFFF]
+ * or return UNI_REPLACEMENT if the encoding has been corrupted */
+static inline void *
+utf16_le_get(const void *p, uns *uu)
+{
+ return utf16_le_get_repl(p, uu, UNI_REPLACEMENT);
+}
+
+static inline void *
+utf16_be_get(const void *p, uns *uu)
+{
+ return utf16_be_get_repl(p, uu, UNI_REPLACEMENT);
+}
+
+static inline uns
+unicode_sanitize_char(uns u)
+{
+ if (u >= 0x10000 || // We don't accept anything outside the basic plane
+ u >= 0xd800 && u < 0xf900 || // neither we do surrogates
+ u >= 0x80 && u < 0xa0 || // nor latin-1 control chars
+ u < 0x20 && u != '\t')
+ return UNI_REPLACEMENT;
+ return u;
+}
+
/* unicode-utf8.c */
-uns utf8_strlen(byte *str);
-uns utf8_strnlen(byte *str, uns n);
-uns utf8_check(byte *str);
+uns utf8_strlen(const byte *str);
+uns utf8_strnlen(const byte *str, uns n);
#endif