diff options
Diffstat (limited to 'src/nvim/mbyte.c')
| -rw-r--r-- | src/nvim/mbyte.c | 69 |
1 files changed, 50 insertions, 19 deletions
diff --git a/src/nvim/mbyte.c b/src/nvim/mbyte.c index 62cc3b56ed..cc488d486f 100644 --- a/src/nvim/mbyte.c +++ b/src/nvim/mbyte.c @@ -346,7 +346,6 @@ static int enc_canon_search(const char_u *name) } - /* * Find canonical encoding "name" in the list and return its properties. * Returns 0 if not found. @@ -675,16 +674,16 @@ static int utf_safe_read_char_adv(const char_u **s, size_t *n) } if (k <= *n) { - /* We have a multibyte sequence and it isn't truncated by buffer - * limits so utf_ptr2char() is safe to use. Or the first byte is - * illegal (k=0), and it's also safe to use utf_ptr2char(). */ + // We have a multibyte sequence and it isn't truncated by buffer + // limits so utf_ptr2char() is safe to use. Or the first byte is + // illegal (k=0), and it's also safe to use utf_ptr2char(). c = utf_ptr2char(*s); - /* On failure, utf_ptr2char() returns the first byte, so here we - * check equality with the first byte. The only non-ASCII character - * which equals the first byte of its own UTF-8 representation is - * U+00C3 (UTF-8: 0xC3 0x83), so need to check that special case too. - * It's safe even if n=1, else we would have k=2 > n. */ + // On failure, utf_ptr2char() returns the first byte, so here we + // check equality with the first byte. The only non-ASCII character + // which equals the first byte of its own UTF-8 representation is + // U+00C3 (UTF-8: 0xC3 0x83), so need to check that special case too. + // It's safe even if n=1, else we would have k=2 > n. if (c != (int)(**s) || (c == 0xC3 && (*s)[1] == 0x83)) { // byte sequence was successfully decoded *s += k; @@ -1582,12 +1581,12 @@ void show_utf8(void) int clen; int i; - /* Get the byte length of the char under the cursor, including composing - * characters. */ + // Get the byte length of the char under the cursor, including composing + // characters. line = get_cursor_pos_ptr(); len = utfc_ptr2len(line); if (len == 0) { - MSG("NUL"); + msg("NUL"); return; } @@ -1610,7 +1609,7 @@ void show_utf8(void) } } - msg(IObuff); + msg((char *)IObuff); } /// Return offset from "p" to the first byte of the character it points into. @@ -1625,8 +1624,8 @@ int utf_head_off(const char_u *base, const char_u *p) return 0; } - /* Skip backwards over trailing bytes: 10xx.xxxx - * Skip backwards again if on a composing char. */ + // Skip backwards over trailing bytes: 10xx.xxxx + // Skip backwards again if on a composing char. const char_u *q; for (q = p;; --q) { // Move s to the last byte of this char. @@ -1883,6 +1882,40 @@ int mb_tail_off(char_u *base, char_u *p) return i; } + +/// Return the offset from "p" to the first byte of the character it points +/// into. Can start anywhere in a stream of bytes. +/// +/// @param[in] base Pointer to start of string +/// @param[in] p Pointer to byte for which to return the offset to the previous codepoint +// +/// @return 0 if invalid sequence, else offset to previous codepoint +int mb_head_off(char_u *base, char_u *p) +{ + int i; + int j; + + if (*p == NUL) { + return 0; + } + + // Find the first character that is not 10xx.xxxx + for (i = 0; p - i > base; i--) { + if ((p[i] & 0xc0) != 0x80) { + break; + } + } + + // Find the last character that is 10xx.xxxx + for (j = 0; (p[j + 1] & 0xc0) == 0x80; j++) {} + + // Check for illegal sequence. + if (utf8len_tab[p[i]] == 1) { + return 0; + } + return i; +} + /* * Find the next illegal byte sequence. */ @@ -1915,8 +1948,8 @@ void utf_find_illegal(void) } while (*p != NUL) { - /* Illegal means that there are not enough trail bytes (checked by - * utf_ptr2len()) or too many of them (overlong sequence). */ + // Illegal means that there are not enough trail bytes (checked by + // utf_ptr2len()) or too many of them (overlong sequence). len = utf_ptr2len(p); if (*p >= 0x80 && (len == 1 || utf_char2len(utf_ptr2char(p)) != len)) { @@ -2398,8 +2431,6 @@ static char_u *iconv_string(const vimconv_T *const vcp, char_u *str, size_t slen #endif // HAVE_ICONV - - /* * Setup "vcp" for conversion from "from" to "to". * The names must have been made canonical with enc_canonize(). |