BLI_string_utf8.h

Go to the documentation of this file.

/* SPDX-FileCopyrightText: 2023 Blender Authors
 *
 * SPDX-License-Identifier: GPL-2.0-or-later */
 
#pragma once

 
#include "BLI_compiler_attrs.h"
#include "BLI_sys_types.h"
 
char *BLI_strncpy_utf8(char *__restrict dst, const char *__restrict src, size_t dst_maxncpy)
    ATTR_NONNULL(1, 2);
size_t BLI_strncpy_utf8_rlen(char *__restrict dst,
                             const char *__restrict src,
                             size_t dst_maxncpy) ATTR_WARN_UNUSED_RESULT ATTR_NONNULL(1, 2);
size_t BLI_strncpy_utf8_rlen_unterminated(char *__restrict dst,
                                          const char *__restrict src,
                                          size_t dst_maxncpy);

ptrdiff_t BLI_str_utf8_invalid_byte(const char *str, size_t str_len) ATTR_NONNULL(1);
int BLI_str_utf8_invalid_strip(char *str, size_t str_len) ATTR_NONNULL(1);
int BLI_str_utf8_invalid_substitute(char *str, size_t str_len, const char substitute)
    ATTR_NONNULL(1);

[[nodiscard]] const char *BLI_str_utf8_invalid_substitute_as_needed(
    const char *str, size_t str_len, const char substitute, char *buf, const size_t buf_maxncpy)
    ATTR_NONNULL(1, 4);

int BLI_str_utf8_size_or_error(const char *p) ATTR_WARN_UNUSED_RESULT ATTR_NONNULL(1);
int BLI_str_utf8_size_safe(const char *p) ATTR_WARN_UNUSED_RESULT ATTR_NONNULL(1);
unsigned int BLI_str_utf8_as_unicode_or_error(const char *p) ATTR_WARN_UNUSED_RESULT
    ATTR_NONNULL(1);
unsigned int BLI_str_utf8_as_unicode_safe(const char *p) ATTR_WARN_UNUSED_RESULT ATTR_NONNULL(1);
unsigned int BLI_str_utf8_as_unicode_step_safe(const char *__restrict p,
                                               size_t p_len,
                                               size_t *__restrict index) ATTR_WARN_UNUSED_RESULT
    ATTR_NONNULL(1, 3);
unsigned int BLI_str_utf8_as_unicode_step_or_error(
    const char *__restrict p, size_t p_len, size_t *__restrict index) ATTR_WARN_UNUSED_RESULT
    ATTR_NONNULL(1, 3);
 
size_t BLI_str_utf8_from_unicode_len(unsigned int c) ATTR_WARN_UNUSED_RESULT;
size_t BLI_str_utf8_from_unicode(unsigned int c, char *dst, size_t dst_maxncpy) ATTR_NONNULL(2);
size_t BLI_str_utf8_as_utf32(char32_t *__restrict dst_w,
                             const char *__restrict src_c,
                             size_t dst_w_maxncpy) ATTR_NONNULL(1, 2);
size_t BLI_str_utf32_as_utf8(char *__restrict dst,
                             const char32_t *__restrict src,
                             size_t dst_maxncpy) ATTR_NONNULL(1, 2);
size_t BLI_str_utf32_as_utf8_len_ex(const char32_t *src, size_t src_maxlen) ATTR_WARN_UNUSED_RESULT
    ATTR_NONNULL(1);
size_t BLI_str_utf32_as_utf8_len(const char32_t *src) ATTR_WARN_UNUSED_RESULT ATTR_NONNULL(1);

const char *BLI_str_find_prev_char_utf8(const char *p, const char *str_start)
    ATTR_WARN_UNUSED_RESULT ATTR_RETURNS_NONNULL ATTR_NONNULL(1, 2);
const char *BLI_str_find_next_char_utf8(const char *p, const char *str_end)
    ATTR_WARN_UNUSED_RESULT ATTR_RETURNS_NONNULL ATTR_NONNULL(1, 2);

size_t BLI_wstrlen_utf8(const wchar_t *src) ATTR_NONNULL(1) ATTR_WARN_UNUSED_RESULT;
size_t BLI_strlen_utf8_ex(const char *strc, size_t *r_len_bytes)
    ATTR_NONNULL(1, 2) ATTR_WARN_UNUSED_RESULT;
size_t BLI_strlen_utf8(const char *strc) ATTR_NONNULL(1) ATTR_WARN_UNUSED_RESULT;
size_t BLI_strnlen_utf8_ex(const char *strc, size_t strc_maxlen, size_t *r_len_bytes)
    ATTR_NONNULL(1, 3);
size_t BLI_strnlen_utf8(const char *strc, size_t strc_maxlen)
    ATTR_NONNULL(1) ATTR_WARN_UNUSED_RESULT;
size_t BLI_strncpy_wchar_as_utf8(char *__restrict dst,
                                 const wchar_t *__restrict src,
                                 size_t dst_maxncpy) ATTR_NONNULL(1, 2);
size_t BLI_strncpy_wchar_from_utf8(wchar_t *__restrict dst_w,
                                   const char *__restrict src_c,
                                   size_t dst_w_maxncpy) ATTR_NONNULL(1, 2);

int BLI_wcwidth_or_error(char32_t ucs) ATTR_WARN_UNUSED_RESULT;
int BLI_wcwidth_safe(char32_t ucs) ATTR_WARN_UNUSED_RESULT;
int BLI_wcswidth_or_error(const char32_t *pwcs, size_t n) ATTR_WARN_UNUSED_RESULT ATTR_NONNULL(1);

char32_t BLI_str_utf32_char_to_upper(char32_t wc);
char32_t BLI_str_utf32_char_to_lower(char32_t wc);
 
bool BLI_str_utf32_char_is_breaking_space(char32_t codepoint);
bool BLI_str_utf32_char_is_optional_break_after(char32_t codepoint, char32_t codepoint_prev);
bool BLI_str_utf32_char_is_optional_break_before(char32_t codepoint, char32_t codepoint_prev);

int BLI_str_utf8_char_width_or_error(const char *p) ATTR_WARN_UNUSED_RESULT ATTR_NONNULL(1);
int BLI_str_utf8_char_width_safe(const char *p) ATTR_WARN_UNUSED_RESULT ATTR_NONNULL(1);
 
size_t BLI_str_partition_utf8(const char *str,
                              const unsigned int delim[],
                              const char **r_sep,
                              const char **r_suf) ATTR_NONNULL(1, 2, 3, 4);
size_t BLI_str_rpartition_utf8(const char *str,
                               const unsigned int delim[],
                               const char **r_sep,
                               const char **r_suf) ATTR_NONNULL(1, 2, 3, 4);
size_t BLI_str_partition_ex_utf8(const char *str,
                                 const char *end,
                                 const unsigned int delim[],
                                 const char **r_sep,
                                 const char **r_suf,
                                 bool from_right) ATTR_WARN_UNUSED_RESULT ATTR_NONNULL(1, 3, 4, 5);

bool BLI_str_utf8_truncate_at_size(char *str, const size_t str_size);
 
int BLI_str_utf8_offset_to_index(const char *str,
                                 size_t str_len,
                                 int offset_target) ATTR_WARN_UNUSED_RESULT ATTR_NONNULL(1);
int BLI_str_utf8_offset_from_index(const char *str,
                                   size_t str_len,
                                   int index_target) ATTR_WARN_UNUSED_RESULT ATTR_NONNULL(1);
int BLI_str_utf8_offset_to_column(const char *str,
                                  size_t str_len,
                                  int offset_target) ATTR_WARN_UNUSED_RESULT ATTR_NONNULL(1);
int BLI_str_utf8_offset_from_column(const char *str,
                                    size_t str_len,
                                    int column_target) ATTR_WARN_UNUSED_RESULT ATTR_NONNULL(1);
int BLI_str_utf8_offset_to_column_with_tabs(const char *str,
                                            size_t str_len,
                                            int offset_target,
                                            int tab_width) ATTR_WARN_UNUSED_RESULT ATTR_NONNULL(1);
int BLI_str_utf8_offset_from_column_with_tabs(const char *str,
                                              size_t str_len,
                                              int column_target,
                                              int tab_width) ATTR_WARN_UNUSED_RESULT
    ATTR_NONNULL(1);

#define BLI_UTF8_MAX 6
#define BLI_UTF8_WIDTH_MAX 2 /* columns */
#define BLI_UTF8_ERR ((unsigned int)-1)
 
/* -------------------------------------------------------------------- */
 
#define STRNCPY_UTF8(dst, src) BLI_strncpy_utf8(dst, src, ARRAY_SIZE(dst))
#define STRNCPY_UTF8_RLEN(dst, src) BLI_strncpy_utf8_rlen(dst, src, ARRAY_SIZE(dst))
 
#define STRNLEN_UTF8(str) BLI_strnlen_utf8(str, ARRAY_SIZE(str))