From 83b15d0b563774f52f48bf7335152c16dd9614fd Mon Sep 17 00:00:00 2001 From: David Robillard Date: Wed, 30 Jun 2021 12:08:38 -0400 Subject: Use a lookup table to determine UTF-8 encoding length --- src/string_utils.h | 58 +++++++++++++++++++++++++++++++++++------------------- 1 file changed, 38 insertions(+), 20 deletions(-) diff --git a/src/string_utils.h b/src/string_utils.h index 07941325..5eeabc6b 100644 --- a/src/string_utils.h +++ b/src/string_utils.h @@ -112,25 +112,44 @@ serd_strncasecmp(const char* s1, const char* s2, size_t n) } static inline uint32_t -utf8_num_bytes(const uint8_t c) +utf8_num_bytes(const uint8_t leading) { - if ((c & 0x80) == 0) { // Starts with `0' - return 1; - } - - if ((c & 0xE0) == 0xC0) { // Starts with `110' - return 2; - } - - if ((c & 0xF0) == 0xE0) { // Starts with `1110' - return 3; - } - - if ((c & 0xF8) == 0xF0) { // Starts with `11110' - return 4; - } - - return 0; + static const uint8_t lengths[32] = { + 1u, // 00000xxx + 1u, // 00001xxx + 1u, // 00010xxx + 1u, // 00011xxx + 1u, // 00100xxx + 1u, // 00101xxx + 1u, // 00110xxx + 1u, // 00111xxx + 1u, // 01000xxx + 1u, // 01001xxx + 1u, // 01010xxx + 1u, // 01011xxx + 1u, // 01100xxx + 1u, // 01101xxx + 1u, // 01110xxx + 1u, // 01111xxx + 0u, // 10000xxx + 0u, // 10001xxx + 0u, // 10010xxx + 0u, // 10011xxx + 0u, // 10100xxx + 0u, // 10101xxx + 0u, // 10110xxx + 0u, // 10111xxx + 2u, // 11000xxx + 2u, // 11001xxx + 2u, // 11010xxx + 2u, // 11011xxx + 3u, // 11100xxx + 3u, // 11101xxx + 4u, // 11110xxx + 0u // 11111xxx + }; + + return lengths[leading >> 3u]; } /// Return the code point of a UTF-8 character with known length @@ -139,8 +158,7 @@ parse_counted_utf8_char(const uint8_t* utf8, size_t size) { uint32_t c = utf8[0] & ((1u << (8u - size)) - 1u); for (size_t i = 1; i < size; ++i) { - const uint8_t in = utf8[i] & 0x3Fu; - c = (c << 6) | in; + c = (c << 6) | (utf8[i] & 0x3Fu); } return c; } -- cgit v1.2.1