diff options
author | David Robillard <d@drobilla.net> | 2023-02-05 12:42:52 -0500 |
---|---|---|
committer | David Robillard <d@drobilla.net> | 2023-12-02 18:49:08 -0500 |
commit | ea4b6e7d109ae3abc7f8ecdf99e3eb33e3484b77 (patch) | |
tree | 9b050faf6286c055d2fc78729eb4b56a12e3746c /src/string_utils.h | |
parent | d35082a57adac79703f2c9bb72da468172a209c5 (diff) | |
download | serd-ea4b6e7d109ae3abc7f8ecdf99e3eb33e3484b77.tar.gz serd-ea4b6e7d109ae3abc7f8ecdf99e3eb33e3484b77.tar.bz2 serd-ea4b6e7d109ae3abc7f8ecdf99e3eb33e3484b77.zip |
Strengthen handling of corrupt UTF-8 input
Diffstat (limited to 'src/string_utils.h')
-rw-r--r-- | src/string_utils.h | 20 |
1 files changed, 20 insertions, 0 deletions
diff --git a/src/string_utils.h b/src/string_utils.h index 8f7ea083..9de03fa0 100644 --- a/src/string_utils.h +++ b/src/string_utils.h @@ -48,6 +48,26 @@ is_xdigit(const int c) return is_hexdig(c) || in_range(c, 'a', 'f'); } +/** UTF-8: Leading bytes start with 0, or two to four 1s followed by a 0 */ +static inline bool +is_utf8_leading(const uint8_t c) +{ + static const uint8_t m1 = 0x80U; // 10000000 + static const uint8_t m2 = 0xC0U; // 11000000 + static const uint8_t m3 = 0xE0U; // 11100000 + static const uint8_t m4 = 0xF0U; // 11110000 + static const uint8_t m5 = 0xF8U; // 11111000 + + return (c & m1) == 0U || (c & m3) == m2 || (c & m4) == m3 || (c & m5) == m4; +} + +/** UTF-8: Continuation bytes start with 10 */ +static inline bool +is_utf8_continuation(const uint8_t c) +{ + return (c & 0xC0U) == 0x80U; +} + static inline bool is_space(const char c) { |