cJSON源码阅读（二）

JSON字符串解析函数cJSON_Parse()

cJSON * cJSON_Parse(const char *value)
{
             //第一个参数：要解析的字符串
             //第二个参数：最后一个被解析的字符（传出）
             //第三个参数：被解析的字符串是否以null结尾
    return cJSON_ParseWithOpts(value, 0, 0);
}

typedef struct
{
    const unsigned char *content;  //字符串
    size_t length;                 //字符串长度
    size_t offset;                 //偏移量（解析到哪里了）
    size_t depth; /* How deeply nested (in arrays/objects) is the input at the current offset. */
    internal_hooks hooks;          //钩子函数
} parse_buffer;

cJSON * cJSON_ParseWithOpts(const char *value, const char **return_parse_end, cJSON_bool require_null_terminated)
{
    parse_buffer buffer = { 0, 0, 0, 0, { 0, 0, 0 } };
    cJSON *item = NULL;

    /* reset error position */
    global_error.json = NULL;
    global_error.position = 0;

    if (value == NULL)
    {
        goto fail;
    }
    //初始化parse_buffer结构体
    buffer.content = (const unsigned char*)value;
    buffer.length = strlen((const char*)value) + sizeof("");
    buffer.offset = 0;
    buffer.hooks = global_hooks;

    item = cJSON_New_Item(&global_hooks);
    if (item == NULL) /* memory fail */
    {
        goto fail;
    }
    //parse_value是真正的解析函数
    if (!parse_value(item, buffer_skip_whitespace(skip_utf8_bom(&buffer))))
    {
        /* parse failure. ep is set. */
        goto fail;
    }

    /* if we require null-terminated JSON without appended garbage, skip and then check for a null terminator */
    //检查被解析的字符是否以NULL结尾
    if (require_null_terminated)
    {
        buffer_skip_whitespace(&buffer);
        if ((buffer.offset >= buffer.length) || buffer_at_offset(&buffer)[0] != '\0')
        {
            goto fail;
        }
    }
    //返回最后一个被解析的字符
    if (return_parse_end)
    {
        *return_parse_end = (const char*)buffer_at_offset(&buffer);
    }

    return item;
//异常处理
fail:
    if (item != NULL)
    {
        cJSON_Delete(item);
    }

    if (value != NULL)
    {
        //存放错误的json字符串和错误位置
        error local_error;
        local_error.json = (const unsigned char*)value;
        local_error.position = 0;

        if (buffer.offset < buffer.length)
        {
            local_error.position = buffer.offset;
        }
        else if (buffer.length > 0)
        {   
            local_error.position = buffer.length - 1;
        }
        
        if (return_parse_end != NULL)
        {   
            *return_parse_end = (const char*)local_error.json + local_error.position;
        }
        
        global_error = local_error;
    }

    return NULL;
}

真正的解析函数parse_value

/* check if the given size is left to read in a given parse buffer (starting with 1) */
#define can_read(buffer, size) ((buffer != NULL) && (((buffer)->offset + size) <= (buffer)->length))
/* check if the buffer can be accessed at the given index (starting with 0) */
#define can_access_at_index(buffer, index) ((buffer != NULL) && (((buffer)->offset + index) < (buffer)->length))
#define cannot_access_at_index(buffer, index) (!can_access_at_index(buffer, index))
/* get a pointer to the buffer at the position */
#define buffer_at_offset(buffer) ((buffer)->content + (buffer)->offset)

/* Parser core - when encountering text, process appropriately. */
static cJSON_bool parse_value(cJSON * const item, parse_buffer * const input_buffer)
{
    if ((input_buffer == NULL) || (input_buffer->content == NULL))
    {
        return false; /* no input */
    }

    /* parse the different types of values */
    /* null */
    //解析null字符串，can_read函数判断offset+4有没有超过length，buffer_at_offset函数返回一个指针指向下一个未解析的位置。
    if (can_read(input_buffer, 4) && (strncmp((const char*)buffer_at_offset(input_buffer), "null", 4) == 0))
    {
        item->type = cJSON_NULL;
        //offset+4，offset是已经解析过的字符数
        input_buffer->offset += 4;
        return true;
    }
    /* false */
    if (can_read(input_buffer, 5) && (strncmp((const char*)buffer_at_offset(input_buffer), "false", 5) == 0))
    {
        item->type = cJSON_False;
        input_buffer->offset += 5;
        return true;
    }
    /* true */
    if (can_read(input_buffer, 4) && (strncmp((const char*)buffer_at_offset(input_buffer), "true", 4) == 0))
    {
        item->type = cJSON_True;
        item->valueint = 1;
        input_buffer->offset += 4;
        return true;
    }
    /* string */
    //解析字符串，必需以 " 字符开头
    if (can_access_at_index(input_buffer, 0) && (buffer_at_offset(input_buffer)[0] == '\"'))
    {
        return parse_string(item, input_buffer);
    }
    /* number */
    //数字，以-号或者数字开头
    if (can_access_at_index(input_buffer, 0) && ((buffer_at_offset(input_buffer)[0] == '-') || ((buffer_at_offset(input_buffer)[0] >= '0') && (buffer_at_offset(input_buffer)[0] <= '9'))))
    {
        return parse_number(item, input_buffer);
    }
    /* array */
    //数组，以 [ 开头
    if (can_access_at_index(input_buffer, 0) && (buffer_at_offset(input_buffer)[0] == '['))
    {
        return parse_array(item, input_buffer);
    }
    /* object */
    //object对象类型，以 { 开头
    if (can_access_at_index(input_buffer, 0) && (buffer_at_offset(input_buffer)[0] == '{'))
    {
        return parse_object(item, input_buffer);
    }

    return false;
}

字符串类型解析函数

/* Parse the input text into an unescaped cinput, and populate item. */
static cJSON_bool parse_string(cJSON * const item, parse_buffer * const input_buffer)
{
    const unsigned char *input_pointer = buffer_at_offset(input_buffer) + 1;
    const unsigned char *input_end = buffer_at_offset(input_buffer) + 1;
    unsigned char *output_pointer = NULL;
    unsigned char *output = NULL;

    /* not a string */
    if (buffer_at_offset(input_buffer)[0] != '\"')
    {
        goto fail;
    }
    //这一段为什么要用{}包起来，没有看懂。
    {
        /* calculate approximate size of the output (overestimate) */
        size_t allocation_length = 0;
        size_t skipped_bytes = 0;
        //如果没有超过了字符串的长度以及没有遇到结尾的 " 字符，循环继续
        while (((size_t)(input_end - input_buffer->content) < input_buffer->length) && (*input_end != '\"'))
        {
            /* is escape sequence */
            //如果是转义字符，如果字符串中有'\' 'n'这样的字符(这个是两个字符)，将它们变成'\n'（一个字符）
            if (input_end[0] == '\\')
            {
                if ((size_t)(input_end + 1 - input_buffer->content) >= input_buffer->length)
                {
                    /* prevent buffer overflow when last input character is a backslash */
                    goto fail;
                }
                skipped_bytes++;
                input_end++;
            }
            input_end++;
        }
        //如果不是以 " 字符结尾，报错
        if (((size_t)(input_end - input_buffer->content) >= input_buffer->length) || (*input_end != '\"'))
        {
            goto fail; /* string ended unexpectedly */
        }   
        
        /* This is at most how much we need for the output */
        //给output指针分配空间
        allocation_length = (size_t) (input_end - buffer_at_offset(input_buffer)) - skipped_bytes;
        output = (unsigned char*)input_buffer->hooks.allocate(allocation_length + sizeof(""));
        if (output == NULL)
        {
            goto fail; /* allocation failure */
        }   
    }   
    
    output_pointer = output;
    /* loop through the string literal */
    while (input_pointer < input_end)
    {
        if (*input_pointer != '\\')
        {
            *output_pointer++ = *input_pointer++;
        }
        /* escape sequence */
        else
        {
            unsigned char sequence_length = 2;
            if ((input_end - input_pointer) < 1)
            {
                goto fail;
            }
            // 将'\'和后面的字符变成一个字符。
            switch (input_pointer[1])
            {
                case 'b':
                    *output_pointer++ = '\b';
                    break;
                case 'f':
                    *output_pointer++ = '\f';
                    break;
                case 'n':
                    *output_pointer++ = '\n';
                    break;
                case 'r':
                    *output_pointer++ = '\r';
                    break;
                case 't':
                    *output_pointer++ = '\t';
                    break;
                case '\"':
                case '\\':
                case '/':
                    *output_pointer++ = input_pointer[1];
                    break;

                /* UTF-16 literal */
                case 'u':
                    sequence_length = utf16_literal_to_utf8(input_pointer, input_end, &output_pointer);
                    if (sequence_length == 0)
                    {
                        /* failed to convert UTF16-literal to UTF-8 */
                        goto fail;
                    }
                    break;
                default:
                    goto fail;
            }
            input_pointer += sequence_length;
        }
    }

    /* zero terminate the output */
    *output_pointer = '\0';

    item->type = cJSON_String;
    item->valuestring = (char*)output;

    input_buffer->offset = (size_t) (input_end - input_buffer->content);
    //为什么还要++？
    input_buffer->offset++;

    return true;

fail:
    if (output != NULL)
    {
        input_buffer->hooks.deallocate(output);
    }

    if (input_pointer != NULL)
    {
        input_buffer->offset = (size_t)(input_pointer - input_buffer->content);
    }

    return false;
}

解析数字

/* Parse the input text to generate a number, and populate the result into item. */
static cJSON_bool parse_number(cJSON * const item, parse_buffer * const input_buffer)
{
    double number = 0;
    unsigned char *after_end = NULL;
    unsigned char number_c_string[64];
    unsigned char decimal_point = get_decimal_point();
    size_t i = 0;

    if ((input_buffer == NULL) || (input_buffer->content == NULL))
    {
        return false;
    }

    /* copy the number into a temporary buffer and replace '.' with the decimal point
     * of the current locale (for strtod)
     * This also takes care of '\0' not necessarily being available for marking the end of the input */
    for (i = 0; (i < (sizeof(number_c_string) - 1)) && can_access_at_index(input_buffer, i); i++)
    {
        switch (buffer_at_offset(input_buffer)[i])
        {
            case '0':
            case '1':
            case '2':
            case '3':
            case '4':
            case '5':
            case '6':
            case '7':
            case '8':
            case '9':
            case '+':
            case '-':
            case 'e':
            case 'E':
                number_c_string[i] = buffer_at_offset(input_buffer)[i];
                break;

            case '.':
                number_c_string[i] = decimal_point;
                break;

            default:
                goto loop_end;
        }
    }
loop_end:
    number_c_string[i] = '\0';

    number = strtod((const char*)number_c_string, (char**)&after_end);
    if (number_c_string == after_end)
    {
        return false; /* parse_error */
    }

    item->valuedouble = number;

    /* use saturation in case of overflow */
    if (number >= INT_MAX)
    {
        item->valueint = INT_MAX;
    }
    else if (number <= (double)INT_MIN)
    {
        item->valueint = INT_MIN;
    }
    else
    {
        item->valueint = (int)number;
    }

    item->type = cJSON_Number;

    input_buffer->offset += (size_t)(after_end - number_c_string);
    return true;
}

数组解析

#define CJSON_NESTING_LIMIT 1000
/* Build an array from input text. */
static cJSON_bool parse_array(cJSON * const item, parse_buffer * const input_buffer)
{
    cJSON *head = NULL; /* head of the linked list */
    cJSON *current_item = NULL;
    
    //防止嵌套超过1000层，防止堆栈溢出
    if (input_buffer->depth >= CJSON_NESTING_LIMIT)
    {
        return false; /* to deeply nested */
    }
    input_buffer->depth++;

    if (buffer_at_offset(input_buffer)[0] != '[')
    {
        /* not an array */
        goto fail;
    }

    input_buffer->offset++;
    //跳过空白字符
    buffer_skip_whitespace(input_buffer);
    if (can_access_at_index(input_buffer, 0) && (buffer_at_offset(input_buffer)[0] == ']'))
    {
        /* empty array */
        goto success;
    }

    /* check if we skipped to the end of the buffer */
    if (cannot_access_at_index(input_buffer, 0))
    {
        input_buffer->offset--;
        goto fail;
    }

    /* step back to character in front of the first element */
    //为什么要回退一个字符？
    input_buffer->offset--;
    /* loop through the comma separated array elements */
    do
    {
        /* allocate next item */
        cJSON *new_item = cJSON_New_Item(&(input_buffer->hooks));
        if (new_item == NULL)
        {
            goto fail; /* allocation failure */
        }

        /* attach next item to list */
        if (head == NULL)
        {
            /* start the linked list */
            current_item = head = new_item;
        }
        else
        {
            /* add to the end and advance */
            current_item->next = new_item;
            new_item->prev = current_item;
            current_item = new_item;
        }

        /* parse next value */
        input_buffer->offset++;
        buffer_skip_whitespace(input_buffer);
        //递归调用
        if (!parse_value(current_item, input_buffer))
        {
            goto fail; /* failed to parse value */
        }
        buffer_skip_whitespace(input_buffer);
    }
    while (can_access_at_index(input_buffer, 0) && (buffer_at_offset(input_buffer)[0] == ','));

    if (cannot_access_at_index(input_buffer, 0) || buffer_at_offset(input_buffer)[0] != ']')
    {
        goto fail; /* expected end of array */
    }

success:
    input_buffer->depth--;

    item->type = cJSON_Array;
    item->child = head;

    input_buffer->offset++;

    return true;

fail:
    if (head != NULL)
    {
        cJSON_Delete(head);
    }

    return false;
}

object类型解析

/* Build an object from the text. */
static cJSON_bool parse_object(cJSON * const item, parse_buffer * const input_buffer)
{
    cJSON *head = NULL; /* linked list head */
    cJSON *current_item = NULL;
    //和数组一样要判断深度
    if (input_buffer->depth >= CJSON_NESTING_LIMIT)
    {
        return false; /* to deeply nested */
    }
    input_buffer->depth++;
    //如果不是以 '{' 开头，不是object类型
    if (cannot_access_at_index(input_buffer, 0) || (buffer_at_offset(input_buffer)[0] != '{'))
    {
        goto fail; /* not an object */
    }

    input_buffer->offset++;
    buffer_skip_whitespace(input_buffer);
    if (can_access_at_index(input_buffer, 0) && (buffer_at_offset(input_buffer)[0] == '}'))
    {
        goto success; /* empty object */
    }

    /* check if we skipped to the end of the buffer */
    if (cannot_access_at_index(input_buffer, 0))
    {
        input_buffer->offset--;
        goto fail;
    }

    /* step back to character in front of the first element */
    input_buffer->offset--;
    /* loop through the comma separated array elements */
    do
    {
        /* allocate next item */
        cJSON *new_item = cJSON_New_Item(&(input_buffer->hooks));
        if (new_item == NULL)
        {
            goto fail; /* allocation failure */
        }

        /* attach next item to list */
        if (head == NULL)
        {
            /* start the linked list */
            current_item = head = new_item;
        }
        else
        {
            /* add to the end and advance */
            current_item->next = new_item;
            new_item->prev = current_item;
            current_item = new_item;
        }

        /* parse the name of the child */
        //解析key的名称
        input_buffer->offset++;
        buffer_skip_whitespace(input_buffer);
        if (!parse_string(current_item, input_buffer))
        {
            goto fail; /* failed to parse name */
        }
        buffer_skip_whitespace(input_buffer);

        /* swap valuestring and string, because we parsed the name */
        current_item->string = current_item->valuestring;
        current_item->valuestring = NULL;

        if (cannot_access_at_index(input_buffer, 0) || (buffer_at_offset(input_buffer)[0] != ':'))
        {
            goto fail; /* invalid object */
        }

        /* parse the value */
        //解析value
        input_buffer->offset++;
        buffer_skip_whitespace(input_buffer);
        if (!parse_value(current_item, input_buffer))
        {
            goto fail; /* failed to parse value */
        }
        buffer_skip_whitespace(input_buffer);
    }
    while (can_access_at_index(input_buffer, 0) && (buffer_at_offset(input_buffer)[0] == ','));

    if (cannot_access_at_index(input_buffer, 0) || (buffer_at_offset(input_buffer)[0] != '}'))
    {
        goto fail; /* expected end of object */
    }

success:
    input_buffer->depth--;

    item->type = cJSON_Object;
    item->child = head;

    input_buffer->offset++;
    return true;

fail:
    if (head != NULL)
    {
        cJSON_Delete(head);
    }

    return false;
}

关于将json对象转换成字符串的部分放到下一篇。