C++ 将字符数组解析为脚本文件(语法)

Question

我用 C++ 制作了一个简单的脚本阅读类，它允许我阅读和解析脚本。基本上有一个 FILE 类，然后我继续使用“fopen”打开它。

在函数中，我继续调用“fgetc”和“ftell”以根据需要解析脚本文件，请注意这不是解释器。每个脚本文件都应该遵循某种语法，但这就是我在这里寻求解决方案的原因。

这是一个脚本的样子:

# Script File Comment
USERNAME    = "Joe"
PASSWORD    = "pw0001"
ACCESSLEVEL = 3

DATABASE    = ("localhost",3306,"db","user","password")

基本上我有几个功能:

// This function searches for "variables"
nextToken();

// After I have the variable, e.g: USERNAME, PASSWORD, ACCESSLEVEL or DATABASE
// I proceed to call this function
// This function reads the char array for (,-{}()[]=) these are symbols
readSymbol();

// In a condition I check what "token/variable" I got and proceed to read 
// it accordingly
// e.g; for USERNAME I do:
readString(); // reads text inside "
// e.g; for ACCESSLEVEL I do:
readNumber(); // reads digits until the next char ain't a digit
// e.g; for DATABASE I do:
readSymbol(); // (
readString(); // 127.0.0.1
readSymbol(); // ,
readNumber(); // 3306
readSymbol(); // ,
readString(); // db
readSymbol(); // ,
readString(); // user
readSymbol(); // ,
readString(); // password
readSymbol(); // )

我希望能够读取这样的变量声明:

DATABASELIST = {"data1","data2","data3"}
or
DATABASELIST = {"data1"}

我可以轻松地执行 readSymbol 和 readString 来读取变量内的 3 个不同的字符串定义，但是这个列表应该具有自定义用户数据，例如 5 个不同的字符串，或 8 个不同的字符串 - 取决于。

而且我真的不知道如何使用我编写的解析器来做到这一点。请注意，我是基于我从脚本编写者那里获取的这种格式的一些伪代码，我有从 IDA 中提取的伪代码，如果你想看到它以便更好地理解这里的帖子

这是我的“readSymbol”函数的示例。

READSYMBOL

int TReadScriptFile::readSymbol()
{
    int currentData = 0;
    int stringStart = -1;

    // Check if we can't read anymore
    if (end)
        return 0;

    while (true)
    {
        // Basically get chars in the script
        currentData = fgetc(File);

        // Check for end of file
        if (currentData == -1)
        {
            end = true;
            break;
        }

        if (stringStart == -1)
        {
            if (isdigit(currentData) || isalpha(currentData))
            {   
                printf("TReadScriptFile::readSymbol: Symbol expected\n");
                close();
                return 0;
            }
            else if 
                (
                currentData == '=' || currentData == ',' || 
                currentData == '(' || currentData == ')' || 
                currentData == '{' || currentData == '}' || 
                currentData == '>' || currentData == '<' ||
                currentData == ':' || currentData == '-'
                )
            {
#ifdef __DEBUG__
                printf("Symbol: %c\n", currentData);
#endif
                stringStart = ftell(File);
                break;
            }
        }
    }

    return 1;
}

NEXTTOKEN

int TReadScriptFile::nextToken()
{
    int currentData = 0;
    int stringStart = -1;
    int stringEnd = -1;
    RecursionDepth = -1;
    memset(String, 0, 4000);

    // Check if we can't read anymore
    if (end)
        return 0;

    while (true)
    {
        // ** Syntax ** 

        if (isdigit(getNext()) || getNext() == -1)
        {
            printf("No more tokens left.\n");
            end = true;
            close();
            return 0;
        }
        // End
        // Basically get chars in the script
        currentData = fgetc(File);

        // Check for end of file
        if (currentData == -1)
        {
            end = true;
            break;
        }

        // Syntax Checking Part, this really isn't needed but w/e
        if (stringStart == -1)
        {
            if (currentData == '=' || isdigit(currentData))
            {
                printf("TReadScriptFile::nextToken: Syntax Error: string expected\n");
                close();
                return 0;
            }
        }

        // End Syntax Checking

        // It's a comment line, we should skip
        if (currentData == '#')
        {
            seekNewLn();
            continue;
        }

        // There are no variables, yet
        if (stringStart == -1)
        {
            // We found a letter, we are near a token!
            if (isalpha(currentData))
            {
                stringStart = ftell(File);

                // We might as well add the letter to the string
                RecursionDepth++;
                String[RecursionDepth] = currentData;
                continue;
            }
        }
        else if (stringStart != -1)
        {
            // Let's wait until we get an identifier or space

            // We found a digit, error
            if (isdigit(currentData))
            {
                printf("TReadScriptFile::nextToken: string expected\n");
                close();
                return 0;
            }
            // We found a space, maybe we should stop looking for tokens?
            else if (isspace(currentData))
            {
#ifdef __DEBUG__
                printf("Token: %s\n", String);
#endif
                break;
            }

            RecursionDepth++;
            String[RecursionDepth] = currentData;
        }
    }

    return 1;
}

我在这里找到了一个很好的方法示例:

http://llvm.org/docs/tutorial/LangImpl1.html

Answer 1

一种处理 DATABASE_LIST 的机制是这样的:

找到变量 DATABASE_LIST 后，使用 readSymbol() 读取一个符号，检查它是否是一个 { 然后在循环中执行 readString() 将其添加到 std::vector(或其他一些合适的容器)，然后检查 、 或 (使用 readSymbol())。如果它是一个 ,(逗号)，那么你返回并读取另一个字符串添加到 vector 中等等，直到你最终到达 } 。完成后，您将拥有一个表示 DATABASE_LIST

的字符串 vector (动态数组)