#include "utf8.h"
#include "util.h"
-static inline bool is_unicode_valid(uint32_t ch) {
+bool unichar_is_valid(uint32_t ch) {
if (ch >= 0x110000) /* End of unicode space */
return false;
return true;
}
-static bool is_unicode_control(uint32_t ch) {
+static bool unichar_is_control(uint32_t ch) {
/*
0 to ' '-1 is the C0 range.
val = utf8_encoded_to_unichar(p);
if (val < 0 ||
- is_unicode_control(val) ||
+ unichar_is_control(val) ||
(!newline && val == '\n'))
return false;
return (char*) str;
}
-int utf8_encode_unichar(uint16_t c, char *p) {
- uint8_t *t = (uint8_t*) p;
- int d;
+/**
+ * utf8_encode_unichar() - Encode single UCS-4 character as UTF-8
+ * @out_utf8: output buffer of at least 4 bytes or NULL
+ * @g: UCS-4 character to encode
+ *
+ * This encodes a single UCS-4 character as UTF-8 and writes it into @out_utf8.
+ * The length of the character is returned. It is not zero-terminated! If the
+ * output buffer is NULL, only the length is returned.
+ *
+ * Returns: The length in bytes that the UTF-8 representation does or would
+ * occupy.
+ */
+size_t utf8_encode_unichar(char *out_utf8, uint32_t g) {
- if (c < 0x80) {
- t[0] = (uint8_t) c;
+ if (g < (1 << 7)) {
+ if (out_utf8)
+ out_utf8[0] = g & 0x7f;
return 1;
- } else if (c < 0x800) {
- t[0] = (uint8_t) (0xc0 | (c >> 6));
- t[1] = (uint8_t) (0x80 | (c & 0x3f));
+ } else if (g < (1 << 11)) {
+ if (out_utf8) {
+ out_utf8[0] = 0xc0 | ((g >> 6) & 0x1f);
+ out_utf8[1] = 0x80 | (g & 0x3f);
+ }
return 2;
- } else {
- t[0] = (uint8_t) (0xe0 | (c >> 12));
- t[1] = (uint8_t) (0x80 | ((c >> 6) & 0x3f));
- t[2] = (uint8_t) (0x80 | (c & 0x3f));
+ } else if (g < (1 << 16)) {
+ if (out_utf8) {
+ out_utf8[0] = 0xe0 | ((g >> 12) & 0x0f);
+ out_utf8[1] = 0x80 | ((g >> 6) & 0x3f);
+ out_utf8[2] = 0x80 | (g & 0x3f);
+ }
return 3;
+ } else if (g < (1 << 21)) {
+ if (out_utf8) {
+ out_utf8[0] = 0xf0 | ((g >> 18) & 0x07);
+ out_utf8[1] = 0x80 | ((g >> 12) & 0x3f);
+ out_utf8[2] = 0x80 | ((g >> 6) & 0x3f);
+ out_utf8[3] = 0x80 | (g & 0x3f);
+ }
+ return 4;
}
+
+ return 0;
}
char *utf16_to_utf8(const void *s, size_t length) {
const uint8_t *f;
char *r, *t;
- r = new(char, (length*3+1)/2 + 1);
+ r = new(char, (length * 4 + 1) / 2 + 1);
if (!r)
return NULL;
- for (f = s, t = r; f < (const uint8_t*) s + length; f += 2)
- t += utf8_encode_unichar((f[1] << 8) | f[0], t);
+ f = s;
+ t = r;
+
+ while (f < (const uint8_t*) s + length) {
+ uint16_t w1, w2;
+
+ /* see RFC 2781 section 2.2 */
+
+ w1 = f[1] << 8 | f[0];
+ f += 2;
+
+ if (!utf16_is_surrogate(w1)) {
+ t += utf8_encode_unichar(t, w1);
+
+ continue;
+ }
+
+ if (utf16_is_trailing_surrogate(w1))
+ continue;
+ else if (f >= (const uint8_t*) s + length)
+ break;
+
+ w2 = f[1] << 8 | f[0];
+ f += 2;
+
+ if (!utf16_is_trailing_surrogate(w2)) {
+ f -= 2;
+ continue;
+ }
+
+ t += utf8_encode_unichar(t, utf16_surrogate_pair_to_unichar(w1, w2));
+ }
*t = 0;
return r;
return -EINVAL;
/* check if value has valid range */
- if (!is_unicode_valid(unichar))
+ if (!unichar_is_valid(unichar))
return -EINVAL;
return len;