diff options
Diffstat (limited to 'libmisc')
| -rw-r--r-- | libmisc/fmt.c | 34 | ||||
| -rw-r--r-- | libmisc/include/libmisc/utf8.h | 54 |
2 files changed, 67 insertions, 21 deletions
diff --git a/libmisc/fmt.c b/libmisc/fmt.c index 33788b6..6cf1d8d 100644 --- a/libmisc/fmt.c +++ b/libmisc/fmt.c @@ -6,6 +6,8 @@ #include <string.h> /* for strnlen() */ +#include <libmisc/utf8.h> + #include <libmisc/fmt.h> static const char *const hexdig = "0123456789ABCDEF"; @@ -67,19 +69,18 @@ void fmt_print_qmem(lo_interface fmt_dest w, const void *_str, size_t size) { fmt_print_byte(w, '"'); for (size_t pos = 0; pos < size;) { uint32_t ch; - uint8_t chlen; - if ((str[pos] & 0b10000000) == 0b00000000) { ch = str[pos] & 0b01111111; chlen = 1; } - else if ((str[pos] & 0b11100000) == 0b11000000) { ch = str[pos] & 0b00011111; chlen = 2; } - else if ((str[pos] & 0b11110000) == 0b11100000) { ch = str[pos] & 0b00001111; chlen = 3; } - else if ((str[pos] & 0b11111000) == 0b11110000) { ch = str[pos] & 0b00000111; chlen = 4; } - else goto invalid_utf8; - if ((ch == 0 && chlen != 1) || pos + chlen > size) goto invalid_utf8; - for (uint8_t i = 1; i < chlen; i++) { - if ((str[pos+i] & 0b11000000) != 0b10000000) goto invalid_utf8; - ch = (ch << 6) | (str[pos+i] & 0b00111111); + uint8_t chlen; + utf8_decode_codepoint(&str[pos], size-pos, &ch, &chlen); + if (!chlen) { + /* invalid UTF-8 */ + /* \xAB */ + fmt_print_byte(w, '\\'); + fmt_print_byte(w, 'x'); + fmt_print_byte(w, hexdig[(str[pos] >> 4) & 0xF]); + fmt_print_byte(w, hexdig[(str[pos] >> 0) & 0xF]); + pos++; + continue; } - if (ch > 0x10FFFF) goto invalid_utf8; - if (ch == '\0' || ch == '\b' || ch == '\f' || @@ -132,15 +133,6 @@ void fmt_print_qmem(lo_interface fmt_dest w, const void *_str, size_t size) { fmt_print_byte(w, hexdig[(ch >> 0) & 0xF]); } pos += chlen; - continue; - - invalid_utf8: - /* \xAB */ - fmt_print_byte(w, '\\'); - fmt_print_byte(w, 'x'); - fmt_print_byte(w, hexdig[(str[pos] >> 4) & 0xF]); - fmt_print_byte(w, hexdig[(str[pos] >> 0) & 0xF]); - pos++; } fmt_print_byte(w, '"'); } diff --git a/libmisc/include/libmisc/utf8.h b/libmisc/include/libmisc/utf8.h new file mode 100644 index 0000000..b5e1b0b --- /dev/null +++ b/libmisc/include/libmisc/utf8.h @@ -0,0 +1,54 @@ +/* libmisc/utf8.h - UTF-8 routines + * + * Copyright (C) 2024-2025 Luke T. Shumaker <lukeshu@lukeshu.com> + * SPDX-License-Identifier: AGPL-3.0-or-later + */ + +#ifndef _LIBMISC_UTF8_H_ +#define _LIBMISC_UTF8_H_ + +#include <stddef.h> /* for size_t */ +#include <stdint.h> /* for uint{n}_t */ + +/** + * Decode the codepoint starting at `str` and consuming at most `len` + * bytes. Invalid UTF-8 is indicated with chlen=0. For valid UTF-8, + * chlen is always in the range [1, 4]. + */ +static inline void utf8_decode_codepoint(const uint8_t *str, size_t len, uint32_t *ret_ch, uint8_t *ret_chlen) { + uint32_t ch; + uint8_t chlen; + if ((str[0] & 0b10000000) == 0b00000000) { ch = str[0] & 0b01111111; chlen = 1; } + else if ((str[0] & 0b11100000) == 0b11000000) { ch = str[0] & 0b00011111; chlen = 2; } + else if ((str[0] & 0b11110000) == 0b11100000) { ch = str[0] & 0b00001111; chlen = 3; } + else if ((str[0] & 0b11111000) == 0b11110000) { ch = str[0] & 0b00000111; chlen = 4; } + else goto invalid; + if ((ch == 0 && chlen != 1) || chlen > len) goto invalid; + for (uint8_t i = 1; i < chlen; i++) { + if ((str[i] & 0b11000000) != 0b10000000) goto invalid; + ch = (ch << 6) | (str[i] & 0b00111111); + } + if (ch > 0x10FFFF) goto invalid; + *ret_ch = ch; + *ret_chlen = chlen; + return; + invalid: + *ret_chlen = 0; +} + +static inline bool _utf8_is_valid(const uint8_t *str, size_t len, bool forbid_nul) { + for (size_t pos = 0; pos < len;) { + uint32_t ch; + uint8_t chlen; + utf8_decode_codepoint(&str[pos], len-pos, &ch, &chlen); + if (chlen == 0 || (forbid_nul && ch == 0)) + return false; + pos += chlen; + } + return true; +} + +#define utf8_is_valid(str, len) _utf8_is_valid(str, len, false) +#define utf8_is_valid_without_nul(str, len) _utf8_is_valid(str, len, true) + +#endif /* _LIBMISC_UTF8_H_ */ |