Question

我想知道下面描述的步骤是如何完成的：

如何将JS源文件（带有两个函数的test.js，funcA（）和funcB（））转换为字节码？
如何将生成的字节码加载到duktape并调用其中一个函数，比如funcA（）？

感谢。

test.js：

function funcA() {
    print("funcA()");
}

function funcB() {
    print("funcB()");
}

的main.cpp

int main() {
    duk_context* ctx = duk_create_heap_default();

    std::string byteCodeBuff; // buffer where to save the byte code;

    { // generate the bytecode from 'sourceFilePath'

        const char* sourceCodeFilePath = "test.js"; // contains the JS code
        //  Step 1 How to 'dump' the 'sourceFilePath' to bytecode buffer ???

        // get the duktape bytecode
        duk_size_t bufferSize = 0;
        const char* bytecode = (const char*)duk_get_buffer(ctx, -1, &bufferSize);

        // save the bytecode to byteCodeBuff 
        byteCodeBuff.assign(bytecode,bufferSize);

        duk_pop(ctx);  // bytecode buffer
    }

    { // load the bytecode into duktape

        const size_t length = byteCodeBuff.size();    // bytecode length
        const char* bytecode = &byteCodeBuff.front(); // pointer to bytecode 

        char* dukBuff = (char*)duk_push_fixed_buffer(ctx, length); // push a duk buffer to stack
        memcpy(dukBuff, bytecode, length); // copy the bytecode to the duk buffer

        // Step 2 ??? How start using the bytecode
        // ??? How to invoke funcA()
        // ??? How to invoke funcB()
    }

    duk_destroy_heap(ctx);

    return 0;
}

Answer 1

duk_dump_function和duk_load_function是Duktape提供的用于转换为字节码的功能：

首先，只需加载您的文件：

//  Step 1 (a):
char* sourceCode;
int sourceLen;
{ // load from file (excuse lack of error catching...)
    ifstream fscript(sourceCodeFilePath,ios::in | ios::binary);
    fscript.seekg(0,ios::end);
    sourceLen= fscript.tellg();
    buffer = new char[sourceLen];
    fscript.seekg(0);
    fscript.read(sourceCode,sourceLen);
}

然后，在编译javascript时，脚本被编译为＆＃39;功能＆＃39; （实际上是一块未执行的全局代码）到堆栈中：

//  Step 1 (b):

// compile source code
duk_compile_lstring(ctx,0,sourceCode,sourceLen); // compiles to an ECMAScript function and puts it on the stack top

// dump the code into a buffer
duk_dump_function(ctx); // replaces the stack top function with its bytecode

从这一点开始，代码的其余部分应该在堆栈顶部拾取缓冲区。

至于恢复它：

//  Step 2 (a):
// replaces the buffer on the stack top with the original function
duk_load_function(ctx);

// since this is a global script, it must be executed before calling the invidual functions
duk_call(ctx,0); // perhaps consider using duk_pcall() instead

此时，ctx堆的全局对象现在包含属性funcA和funcB。这些可以这样获得和调用：

//  Step 2 (b):
// push the global object on the stack to get its properties
duk_push_global_object(ctx);

// obtain the function with the name "funcA" and push it on the stack
duk_get_prop_string(ctx,-1,"funcA"); // key name / global function name

duk_require_function(ctx,-1); // require this is a function before calling it

// now invoke it!
// duk_push_* arguments here
duk_call(ctx,0); // 0: number of arguments

// duk_get_* (ctx,-1); to obtain return value

// pop function return value
duk_pop(ctx);

// current stack top: global object

// get next function
duk_get_prop_string(ctx,-1,"funcB");

// require it's a function
duk_require_function(ctx,-1);

// invoke it!
duk_call(ctx,0);

// pop return and global object off stack
duk_pop_2(ctx);

在Duktape中，如何将JS源文件内容转换为字节码

1 个答案: