This repository has been archived on 2023-07-17. You can view files and clone it, but cannot push or open issues or pull requests.
bl_mcu_sdk/components/lvgl/misc/lv_bidi.c

683 lines
22 KiB
C
Raw Normal View History

/**
* @file lv_bidi.c
*
*/
/*********************
* INCLUDES
*********************/
#include <stddef.h>
#include "lv_bidi.h"
#include "lv_txt.h"
#include "../misc/lv_mem.h"
#if LV_USE_BIDI
/*********************
* DEFINES
*********************/
#define LV_BIDI_BRACKLET_DEPTH 4
// Highest bit of the 16-bit pos_conv value specifies whether this pos is RTL or not
#define GET_POS(x) ((x) & 0x7FFF)
#define IS_RTL_POS(x) (((x) & 0x8000) != 0)
#define SET_RTL_POS(x, is_rtl) (GET_POS(x) | ((is_rtl)? 0x8000: 0))
/**********************
* TYPEDEFS
**********************/
typedef struct {
uint32_t bracklet_pos;
lv_base_dir_t dir;
} bracket_stack_t;
/**********************
* STATIC PROTOTYPES
**********************/
static uint32_t lv_bidi_get_next_paragraph(const char * txt);
static lv_base_dir_t lv_bidi_get_letter_dir(uint32_t letter);
static bool lv_bidi_letter_is_weak(uint32_t letter);
static bool lv_bidi_letter_is_rtl(uint32_t letter);
static bool lv_bidi_letter_is_neutral(uint32_t letter);
static lv_base_dir_t get_next_run(const char * txt, lv_base_dir_t base_dir, uint32_t max_len, uint32_t * len,
uint16_t * pos_conv_len);
static void rtl_reverse(char * dest, const char * src, uint32_t len, uint16_t * pos_conv_out, uint16_t pos_conv_rd_base,
uint16_t pos_conv_len);
static uint32_t char_change_to_pair(uint32_t letter);
static lv_base_dir_t bracket_process(const char * txt, uint32_t next_pos, uint32_t len, uint32_t letter,
lv_base_dir_t base_dir);
static void fill_pos_conv(uint16_t * out, uint16_t len, uint16_t index);
static uint32_t get_txt_len(const char * txt, uint32_t max_len);
/**********************
* STATIC VARIABLES
**********************/
static const uint8_t bracket_left[] = {"<({["};
static const uint8_t bracket_right[] = {">)}]"};
static bracket_stack_t br_stack[LV_BIDI_BRACKLET_DEPTH];
static uint8_t br_stack_p;
/**********************
* MACROS
**********************/
/**********************
* GLOBAL FUNCTIONS
**********************/
/**
* Convert a text to get the characters in the correct visual order according to
* Unicode Bidirectional Algorithm
* @param str_in the text to process
* @param str_out store the result here. Has the be `strlen(str_in)` length
* @param base_dir `LV_BASE_DIR_LTR` or `LV_BASE_DIR_RTL`
*/
void _lv_bidi_process(const char * str_in, char * str_out, lv_base_dir_t base_dir)
{
if(base_dir == LV_BASE_DIR_AUTO) base_dir = _lv_bidi_detect_base_dir(str_in);
uint32_t par_start = 0;
uint32_t par_len;
while(str_in[par_start] == '\n' || str_in[par_start] == '\r') {
str_out[par_start] = str_in[par_start];
par_start ++;
}
while(str_in[par_start] != '\0') {
par_len = lv_bidi_get_next_paragraph(&str_in[par_start]);
_lv_bidi_process_paragraph(&str_in[par_start], &str_out[par_start], par_len, base_dir, NULL, 0);
par_start += par_len;
while(str_in[par_start] == '\n' || str_in[par_start] == '\r') {
str_out[par_start] = str_in[par_start];
par_start ++;
}
}
str_out[par_start] = '\0';
}
/**
* Auto-detect the direction of a text based on the first strong character
* @param txt the text to process
* @return `LV_BASE_DIR_LTR` or `LV_BASE_DIR_RTL`
*/
lv_base_dir_t _lv_bidi_detect_base_dir(const char * txt)
{
uint32_t i = 0;
uint32_t letter;
while(txt[i] != '\0') {
letter = _lv_txt_encoded_next(txt, &i);
lv_base_dir_t dir;
dir = lv_bidi_get_letter_dir(letter);
if(dir == LV_BASE_DIR_RTL || dir == LV_BASE_DIR_LTR) return dir;
}
/*If there were no strong char earlier return with the default base dir*/
if(LV_BIDI_BASE_DIR_DEF == LV_BASE_DIR_AUTO) return LV_BASE_DIR_LTR;
else return LV_BIDI_BASE_DIR_DEF;
}
/**
* Get the logical position of a character in a line
* @param str_in the input string. Can be only one line.
* @param bidi_txt internally the text is bidi processed which buffer can be get here.
* If not required anymore has to freed with `lv_mem_free()`
* Can be `NULL` is unused
* @param len length of the line in character count
* @param base_dir base direction of the text: `LV_BASE_DIR_LTR` or `LV_BASE_DIR_RTL`
* @param visual_pos the visual character position which logical position should be get
* @param is_rtl tell the char at `visual_pos` is RTL or LTR context
* @return the logical character position
*/
uint16_t _lv_bidi_get_logical_pos(const char * str_in, char ** bidi_txt, uint32_t len, lv_base_dir_t base_dir,
uint32_t visual_pos, bool * is_rtl)
{
uint32_t pos_conv_len = get_txt_len(str_in, len);
char * buf = lv_mem_buf_get(len + 1);
if(buf == NULL) return (uint16_t) -1;
uint16_t * pos_conv_buf = lv_mem_buf_get(pos_conv_len * sizeof(uint16_t));
if(pos_conv_buf == NULL) {
lv_mem_buf_release(buf);
return (uint16_t) -1;
}
if(bidi_txt) *bidi_txt = buf;
_lv_bidi_process_paragraph(str_in, bidi_txt ? *bidi_txt : NULL, len, base_dir, pos_conv_buf, pos_conv_len);
if(is_rtl) *is_rtl = IS_RTL_POS(pos_conv_buf[visual_pos]);
if(bidi_txt == NULL) lv_mem_buf_release(buf);
uint16_t res = GET_POS(pos_conv_buf[visual_pos]);
lv_mem_buf_release(pos_conv_buf);
return res;
}
/**
* Get the visual position of a character in a line
* @param str_in the input string. Can be only one line.
* @param bidi_txt internally the text is bidi processed which buffer can be get here.
* If not required anymore has to freed with `lv_mem_free()`
* Can be `NULL` is unused
* @param len length of the line in character count
* @param base_dir base direction of the text: `LV_BASE_DIR_LTR` or `LV_BASE_DIR_RTL`
* @param logical_pos the logical character position which visual position should be get
* @param is_rtl tell the char at `logical_pos` is RTL or LTR context
* @return the visual character position
*/
uint16_t _lv_bidi_get_visual_pos(const char * str_in, char ** bidi_txt, uint16_t len, lv_base_dir_t base_dir,
uint32_t logical_pos, bool * is_rtl)
{
uint32_t pos_conv_len = get_txt_len(str_in, len);
char * buf = lv_mem_buf_get(len + 1);
if(buf == NULL) return (uint16_t) -1;
uint16_t * pos_conv_buf = lv_mem_buf_get(pos_conv_len * sizeof(uint16_t));
if(pos_conv_buf == NULL) {
lv_mem_buf_release(buf);
return (uint16_t) -1;
}
if(bidi_txt) *bidi_txt = buf;
_lv_bidi_process_paragraph(str_in, bidi_txt ? *bidi_txt : NULL, len, base_dir, pos_conv_buf, pos_conv_len);
for(uint16_t i = 0; i < pos_conv_len; i++) {
if(GET_POS(pos_conv_buf[i]) == logical_pos) {
if(is_rtl) *is_rtl = IS_RTL_POS(pos_conv_buf[i]);
lv_mem_buf_release(pos_conv_buf);
if(bidi_txt == NULL) lv_mem_buf_release(buf);
return i;
}
}
lv_mem_buf_release(pos_conv_buf);
if(bidi_txt == NULL) lv_mem_buf_release(buf);
return (uint16_t) -1;
}
/**
* Bidi process a paragraph of text
* @param str_in the string to process
* @param str_out store the result here
* @param len length of the text
* @param base_dir base dir of the text
* @param pos_conv_out an `uint16_t` array to store the related logical position of the character.
* Can be `NULL` is unused
* @param pos_conv_len length of `pos_conv_out` in element count
*/
void _lv_bidi_process_paragraph(const char * str_in, char * str_out, uint32_t len, lv_base_dir_t base_dir,
uint16_t * pos_conv_out, uint16_t pos_conv_len)
{
uint32_t run_len = 0;
lv_base_dir_t run_dir;
uint32_t rd = 0;
uint32_t wr;
uint16_t pos_conv_run_len = 0;
uint16_t pos_conv_rd = 0;
uint16_t pos_conv_wr;
if(base_dir == LV_BASE_DIR_AUTO) base_dir = _lv_bidi_detect_base_dir(str_in);
if(base_dir == LV_BASE_DIR_RTL) {
wr = len;
pos_conv_wr = pos_conv_len;
}
else {
wr = 0;
pos_conv_wr = 0;
}
if(str_out) str_out[len] = '\0';
lv_base_dir_t dir = base_dir;
/*Empty the bracket stack*/
br_stack_p = 0;
/*Process neutral chars in the beginning*/
while(rd < len) {
uint32_t letter = _lv_txt_encoded_next(str_in, &rd);
pos_conv_rd++;
dir = lv_bidi_get_letter_dir(letter);
if(dir == LV_BASE_DIR_NEUTRAL) dir = bracket_process(str_in, rd, len, letter, base_dir);
if(dir != LV_BASE_DIR_NEUTRAL && dir != LV_BASE_DIR_WEAK) break;
}
if(rd && str_in[rd] != '\0') {
_lv_txt_encoded_prev(str_in, &rd);
pos_conv_rd--;
}
if(rd) {
if(base_dir == LV_BASE_DIR_LTR) {
if(str_out) {
lv_memcpy(&str_out[wr], str_in, rd);
wr += rd;
}
if(pos_conv_out) {
fill_pos_conv(&pos_conv_out[pos_conv_wr], pos_conv_rd, 0);
pos_conv_wr += pos_conv_rd;
}
}
else {
wr -= rd;
pos_conv_wr -= pos_conv_rd;
rtl_reverse(str_out ? &str_out[wr] : NULL, str_in, rd, pos_conv_out ? &pos_conv_out[pos_conv_wr] : NULL, 0,
pos_conv_rd);
}
}
/*Get and process the runs*/
while(rd < len && str_in[rd]) {
run_dir = get_next_run(&str_in[rd], base_dir, len - rd, &run_len, &pos_conv_run_len);
if(base_dir == LV_BASE_DIR_LTR) {
if(run_dir == LV_BASE_DIR_LTR) {
if(str_out) lv_memcpy(&str_out[wr], &str_in[rd], run_len);
if(pos_conv_out) fill_pos_conv(&pos_conv_out[pos_conv_wr], pos_conv_run_len, pos_conv_rd);
}
else rtl_reverse(str_out ? &str_out[wr] : NULL, &str_in[rd], run_len, pos_conv_out ? &pos_conv_out[pos_conv_wr] : NULL,
pos_conv_rd, pos_conv_run_len);
wr += run_len;
pos_conv_wr += pos_conv_run_len;
}
else {
wr -= run_len;
pos_conv_wr -= pos_conv_run_len;
if(run_dir == LV_BASE_DIR_LTR) {
if(str_out) lv_memcpy(&str_out[wr], &str_in[rd], run_len);
if(pos_conv_out) fill_pos_conv(&pos_conv_out[pos_conv_wr], pos_conv_run_len, pos_conv_rd);
}
else rtl_reverse(str_out ? &str_out[wr] : NULL, &str_in[rd], run_len, pos_conv_out ? &pos_conv_out[pos_conv_wr] : NULL,
pos_conv_rd, pos_conv_run_len);
}
rd += run_len;
pos_conv_rd += pos_conv_run_len;
}
}
void lv_bidi_calculate_align(lv_text_align_t * align, lv_base_dir_t * base_dir, const char * txt)
{
if(*base_dir == LV_BASE_DIR_AUTO) *base_dir = _lv_bidi_detect_base_dir(txt);
if(*align == LV_TEXT_ALIGN_AUTO) {
if(*base_dir == LV_BASE_DIR_RTL) *align = LV_TEXT_ALIGN_RIGHT;
else *align = LV_TEXT_ALIGN_LEFT;
}
}
/**********************
* STATIC FUNCTIONS
**********************/
/**
* Get the next paragraph from a text
* @param txt the text to process
* @return the length of the current paragraph in byte count
*/
static uint32_t lv_bidi_get_next_paragraph(const char * txt)
{
uint32_t i = 0;
_lv_txt_encoded_next(txt, &i);
while(txt[i] != '\0' && txt[i] != '\n' && txt[i] != '\r') {
_lv_txt_encoded_next(txt, &i);
}
return i;
}
/**
* Get the direction of a character
* @param letter a Unicode character
* @return `LV_BASE_DIR_RTL/LTR/WEAK/NEUTRAL`
*/
static lv_base_dir_t lv_bidi_get_letter_dir(uint32_t letter)
{
if(lv_bidi_letter_is_rtl(letter)) return LV_BASE_DIR_RTL;
if(lv_bidi_letter_is_neutral(letter)) return LV_BASE_DIR_NEUTRAL;
if(lv_bidi_letter_is_weak(letter)) return LV_BASE_DIR_WEAK;
return LV_BASE_DIR_LTR;
}
/**
* Tell whether a character is weak or not
* @param letter a Unicode character
* @return true/false
*/
static bool lv_bidi_letter_is_weak(uint32_t letter)
{
uint32_t i = 0;
static const char weaks[] = "0123456789";
do {
uint32_t x = _lv_txt_encoded_next(weaks, &i);
if(letter == x) {
return true;
}
} while(weaks[i] != '\0');
return false;
}
/**
* Tell whether a character is RTL or not
* @param letter a Unicode character
* @return true/false
*/
static bool lv_bidi_letter_is_rtl(uint32_t letter)
{
if(letter >= 0x5d0 && letter <= 0x5ea) return true;
if(letter == 0x202E) return true; /*Unicode of LV_BIDI_RLO*/
/*Check for Persian and Arabic characters [https://en.wikipedia.org/wiki/Arabic_script_in_Unicode]*/
if(letter >= 0x600 && letter <= 0x6FF) return true;
if(letter >= 0xFB50 && letter <= 0xFDFF) return true;
if(letter >= 0xFE70 && letter <= 0xFEFF) return true;
return false;
}
/**
* Tell whether a character is neutral or not
* @param letter a Unicode character
* @return true/false
*/
static bool lv_bidi_letter_is_neutral(uint32_t letter)
{
uint16_t i;
static const char neutrals[] = " \t\n\r.,:;'\"`!?%/\\-=()[]{}<>@#&$|";
for(i = 0; neutrals[i] != '\0'; i++) {
if(letter == (uint32_t)neutrals[i]) return true;
}
return false;
}
static uint32_t get_txt_len(const char * txt, uint32_t max_len)
{
uint32_t len = 0;
uint32_t i = 0;
while(i < max_len && txt[i] != '\0') {
_lv_txt_encoded_next(txt, &i);
len++;
}
return len;
}
static void fill_pos_conv(uint16_t * out, uint16_t len, uint16_t index)
{
uint16_t i;
for(i = 0; i < len; i++) {
out[i] = SET_RTL_POS(index, false);
index++;
}
}
static lv_base_dir_t get_next_run(const char * txt, lv_base_dir_t base_dir, uint32_t max_len, uint32_t * len,
uint16_t * pos_conv_len)
{
uint32_t i = 0;
uint32_t letter;
uint16_t pos_conv_i = 0;
letter = _lv_txt_encoded_next(txt, NULL);
lv_base_dir_t dir = lv_bidi_get_letter_dir(letter);
if(dir == LV_BASE_DIR_NEUTRAL) dir = bracket_process(txt, 0, max_len, letter, base_dir);
/*Find the first strong char. Skip the neutrals*/
while(dir == LV_BASE_DIR_NEUTRAL || dir == LV_BASE_DIR_WEAK) {
letter = _lv_txt_encoded_next(txt, &i);
pos_conv_i++;
dir = lv_bidi_get_letter_dir(letter);
if(dir == LV_BASE_DIR_NEUTRAL) dir = bracket_process(txt, i, max_len, letter, base_dir);
if(dir == LV_BASE_DIR_LTR || dir == LV_BASE_DIR_RTL) break;
if(i >= max_len || txt[i] == '\0' || txt[i] == '\n' || txt[i] == '\r') {
*len = i;
*pos_conv_len = pos_conv_i;
return base_dir;
}
}
lv_base_dir_t run_dir = dir;
uint32_t i_prev = i;
uint32_t i_last_strong = i;
uint16_t pos_conv_i_prev = pos_conv_i;
uint16_t pos_conv_i_last_strong = pos_conv_i;
/*Find the next char which has different direction*/
lv_base_dir_t next_dir = base_dir;
while(i_prev < max_len && txt[i] != '\0' && txt[i] != '\n' && txt[i] != '\r') {
letter = _lv_txt_encoded_next(txt, &i);
pos_conv_i++;
next_dir = lv_bidi_get_letter_dir(letter);
if(next_dir == LV_BASE_DIR_NEUTRAL) next_dir = bracket_process(txt, i, max_len, letter, base_dir);
if(next_dir == LV_BASE_DIR_WEAK) {
if(run_dir == LV_BASE_DIR_RTL) {
if(base_dir == LV_BASE_DIR_RTL) {
next_dir = LV_BASE_DIR_LTR;
}
}
}
/*New dir found?*/
if((next_dir == LV_BASE_DIR_RTL || next_dir == LV_BASE_DIR_LTR) && next_dir != run_dir) {
/*Include neutrals if `run_dir == base_dir`*/
if(run_dir == base_dir) {
*len = i_prev;
*pos_conv_len = pos_conv_i_prev;
}
/*Exclude neutrals if `run_dir != base_dir`*/
else {
*len = i_last_strong;
*pos_conv_len = pos_conv_i_last_strong;
}
return run_dir;
}
if(next_dir != LV_BASE_DIR_NEUTRAL) {
i_last_strong = i;
pos_conv_i_last_strong = pos_conv_i;
}
i_prev = i;
pos_conv_i_prev = pos_conv_i;
}
/*Handle end of of string. Apply `base_dir` on trailing neutrals*/
/*Include neutrals if `run_dir == base_dir`*/
if(run_dir == base_dir) {
*len = i_prev;
*pos_conv_len = pos_conv_i_prev;
}
/*Exclude neutrals if `run_dir != base_dir`*/
else {
*len = i_last_strong;
*pos_conv_len = pos_conv_i_last_strong;
}
return run_dir;
}
static void rtl_reverse(char * dest, const char * src, uint32_t len, uint16_t * pos_conv_out, uint16_t pos_conv_rd_base,
uint16_t pos_conv_len)
{
uint32_t i = len;
uint32_t wr = 0;
uint16_t pos_conv_i = pos_conv_len;
uint16_t pos_conv_wr = 0;
while(i) {
uint32_t letter = _lv_txt_encoded_prev(src, &i);
uint16_t pos_conv_letter = --pos_conv_i;
/*Keep weak letters (numbers) as LTR*/
if(lv_bidi_letter_is_weak(letter)) {
uint32_t last_weak = i;
uint32_t first_weak = i;
uint16_t pos_conv_last_weak = pos_conv_i;
uint16_t pos_conv_first_weak = pos_conv_i;
while(i) {
letter = _lv_txt_encoded_prev(src, &i);
pos_conv_letter = --pos_conv_i;
/*No need to call `char_change_to_pair` because there not such chars here*/
/*Finish on non-weak char*/
/*but treat number and currency related chars as weak*/
if(lv_bidi_letter_is_weak(letter) == false && letter != '.' && letter != ',' && letter != '$' && letter != '%') {
_lv_txt_encoded_next(src, &i); /*Rewind one letter*/
pos_conv_i++;
first_weak = i;
pos_conv_first_weak = pos_conv_i;
break;
}
}
if(i == 0) {
first_weak = 0;
pos_conv_first_weak = 0;
}
if(dest) lv_memcpy(&dest[wr], &src[first_weak], last_weak - first_weak + 1);
if(pos_conv_out) fill_pos_conv(&pos_conv_out[pos_conv_wr], pos_conv_last_weak - pos_conv_first_weak + 1,
pos_conv_rd_base + pos_conv_first_weak);
wr += last_weak - first_weak + 1;
pos_conv_wr += pos_conv_last_weak - pos_conv_first_weak + 1;
}
/*Simply store in reversed order*/
else {
uint32_t letter_size = _lv_txt_encoded_size((const char *)&src[i]);
/*Swap arithmetical symbols*/
if(letter_size == 1) {
uint32_t new_letter = letter = char_change_to_pair(letter);
if(dest) dest[wr] = (uint8_t)new_letter;
if(pos_conv_out) pos_conv_out[pos_conv_wr] = SET_RTL_POS(pos_conv_rd_base + pos_conv_letter, true);
wr++;
pos_conv_wr++;
}
/*Just store the letter*/
else {
if(dest) lv_memcpy(&dest[wr], &src[i], letter_size);
if(pos_conv_out) pos_conv_out[pos_conv_wr] = SET_RTL_POS(pos_conv_rd_base + pos_conv_i, true);
wr += letter_size;
pos_conv_wr++;
}
}
}
}
static uint32_t char_change_to_pair(uint32_t letter)
{
uint8_t i;
for(i = 0; bracket_left[i] != '\0'; i++) {
if(letter == bracket_left[i]) return bracket_right[i];
}
for(i = 0; bracket_right[i] != '\0'; i++) {
if(letter == bracket_right[i]) return bracket_left[i];
}
return letter;
}
static lv_base_dir_t bracket_process(const char * txt, uint32_t next_pos, uint32_t len, uint32_t letter,
lv_base_dir_t base_dir)
{
lv_base_dir_t bracket_dir = LV_BASE_DIR_NEUTRAL;
uint8_t i;
/*Is the letter an opening bracket?*/
for(i = 0; bracket_left[i] != '\0'; i++) {
if(bracket_left[i] == letter) {
/*If so find its matching closing bracket.
*If a char with base dir. direction is found then the brackets will have `base_dir` direction*/
uint32_t txt_i = next_pos;
while(txt_i < len) {
uint32_t letter_next = _lv_txt_encoded_next(txt, &txt_i);
if(letter_next == bracket_right[i]) {
/*Closing bracket found*/
break;
}
else {
/*Save the dir*/
lv_base_dir_t letter_dir = lv_bidi_get_letter_dir(letter_next);
if(letter_dir == base_dir) {
bracket_dir = base_dir;
}
}
}
/*There were no matching closing bracket*/
if(txt_i > len) return LV_BASE_DIR_NEUTRAL;
/*There where a strong char with base dir in the bracket so the dir is found.*/
if(bracket_dir != LV_BASE_DIR_NEUTRAL && bracket_dir != LV_BASE_DIR_WEAK) break;
/*If there were no matching strong chars in the brackets then check the previous chars*/
txt_i = next_pos;
if(txt_i) _lv_txt_encoded_prev(txt, &txt_i);
while(txt_i > 0) {
uint32_t letter_next = _lv_txt_encoded_prev(txt, &txt_i);
lv_base_dir_t letter_dir = lv_bidi_get_letter_dir(letter_next);
if(letter_dir == LV_BASE_DIR_LTR || letter_dir == LV_BASE_DIR_RTL) {
bracket_dir = letter_dir;
break;
}
}
/*There where a previous strong char which can be used*/
if(bracket_dir != LV_BASE_DIR_NEUTRAL) break;
/*There were no strong chars before the bracket, so use the base dir.*/
if(txt_i == 0) bracket_dir = base_dir;
break;
}
}
/*The letter was an opening bracket*/
if(bracket_left[i] != '\0') {
if(bracket_dir == LV_BASE_DIR_NEUTRAL || br_stack_p == LV_BIDI_BRACKLET_DEPTH) return LV_BASE_DIR_NEUTRAL;
br_stack[br_stack_p].bracklet_pos = i;
br_stack[br_stack_p].dir = bracket_dir;
br_stack_p++;
return bracket_dir;
}
else if(br_stack_p > 0) {
/*Is the letter a closing bracket of the last opening?*/
if(letter == bracket_right[br_stack[br_stack_p - 1].bracklet_pos]) {
bracket_dir = br_stack[br_stack_p - 1].dir;
br_stack_p--;
return bracket_dir;
}
}
return LV_BASE_DIR_NEUTRAL;
}
#endif /*LV_USE_BIDI*/