diff options
author | Shawn Landden <shawn@churchofgit.com> | 2015-10-31 20:39:15 -0700 |
---|---|---|
committer | Shawn Landden <shawn@churchofgit.com> | 2015-10-31 21:00:57 -0700 |
commit | 025b4c410577a10692c608e7dbad712655abb858 (patch) | |
tree | 3de2fe13c9a9d9fd778abdcfb1abbb97c202e834 /src/basic/utf8.h | |
parent | 534e8f89d6367cd0e9e17fff67880fb430071d01 (diff) |
utf8.[ch]: use char32_t and char16_t instead of int, int32_t, int16_t
rework C11 utf8.[ch] to use char32_t instead of uint32_t when referring
to unicode chars, to make things more expressive.
Diffstat (limited to 'src/basic/utf8.h')
-rw-r--r-- | src/basic/utf8.h | 13 |
1 files changed, 7 insertions, 6 deletions
diff --git a/src/basic/utf8.h b/src/basic/utf8.h index e745649f06..20d38f12c8 100644 --- a/src/basic/utf8.h +++ b/src/basic/utf8.h @@ -22,12 +22,13 @@ ***/ #include <stdbool.h> +#include <uchar.h> #include "macro.h" #define UTF8_REPLACEMENT_CHARACTER "\xef\xbf\xbd" -bool unichar_is_valid(uint32_t c); +bool unichar_is_valid(char32_t c); const char *utf8_is_valid(const char *s) _pure_; char *ascii_is_valid(const char *s) _pure_; @@ -38,20 +39,20 @@ bool utf8_is_printable_newline(const char* str, size_t length, bool newline) _pu char *utf8_escape_invalid(const char *s); char *utf8_escape_non_printable(const char *str); -size_t utf8_encode_unichar(char *out_utf8, uint32_t g); +size_t utf8_encode_unichar(char *out_utf8, char32_t g); char *utf16_to_utf8(const void *s, size_t length); int utf8_encoded_valid_unichar(const char *str); -int utf8_encoded_to_unichar(const char *str); +char32_t utf8_encoded_to_unichar(const char *str); -static inline bool utf16_is_surrogate(uint16_t c) { +static inline bool utf16_is_surrogate(char16_t c) { return (0xd800 <= c && c <= 0xdfff); } -static inline bool utf16_is_trailing_surrogate(uint16_t c) { +static inline bool utf16_is_trailing_surrogate(char16_t c) { return (0xdc00 <= c && c <= 0xdfff); } -static inline uint32_t utf16_surrogate_pair_to_unichar(uint16_t lead, uint16_t trail) { +static inline char32_t utf16_surrogate_pair_to_unichar(char16_t lead, char16_t trail) { return ((lead - 0xd800) << 10) + (trail - 0xdc00) + 0x10000; } |