Security News
JSR Working Group Kicks Off with Ambitious Roadmap and Plans for Open Governance
At its inaugural meeting, the JSR Working Group outlined plans for an open governance model and a roadmap to enhance JavaScript package management.
@yuuang/ffi-rs-darwin-x64
Advanced tools
A module written in Rust and N-APi provides interface (FFI) features for Node.js
ffi-rs is a high performance module written in Rust and N-API that provides FFI (Foreign Function Interface) features for Node.js. It allows developers to call functions written in other languages such as C++, C, and Rust directly from JavaScript without writing any C++ code.
This module aims to provide similar functionality to the node-ffi module, but with a completely rewritten underlying codebase. The node-ffi module has been unmaintained for several years and is no longer usable, which is why ffi-rs was developed.
Node.js
and c type
$ node bench/bench.js
Running "ffi" suite...
Progress: 100%
ffi-napi:
2 028 ops/s, ±4.87% | slowest, 99.24% slower
ffi-rs:
318 467 ops/s, ±0.17% | fastest
Finished 2 cases!
Fastest: ffi-rs
Slowest: ffi-napi
$ npm i ffi-rs
Currently, ffi-rs only supports there types of parameters and return values. However, support for more types will be added in the future based on actual usage scenarios.
Note: You need to make sure that the compilation environment of the dynamic library is the same as the installation and runtime environment of the ffi-rs
call.
Here is an example of how to use ffi-rs:
For below c++ code, we compile this file into a dynamic library
Note: The return value type of a function must be of type c
#include <cstdio>
#include <cstring>
#include <iostream>
#include <string>
extern "C" int sum(int a, int b) { return a + b; }
extern "C" double doubleSum(double a, double b) { return a + b; }
extern "C" const char *concatenateStrings(const char *str1, const char *str2) {
std::string result = std::string(str1) + std::string(str2);
char *cstr = new char[result.length() + 1];
strcpy(cstr, result.c_str());
return cstr;
}
extern "C" void noRet() { printf("%s", "hello world"); }
extern "C" bool return_opposite(bool input) { return !input; }
$ g++ -dynamiclib -o libsum.so cpp/sum.cpp # macos
$ g++ -shared -o libsum.so cpp/sum.cpp # linux
$ g++ -shared -o sum.dll cpp/sum.cpp # win
Then can use ffi-rs
invoke the dynamic library file contains functions.
const { equal } = require('assert')
const { load, DataType, open, close, arrayConstructor } = require('ffi-rs')
const a = 1
const b = 100
const dynamicLib = platform === 'win32' ? './sum.dll' : "./libsum.so"
// first open dynamic library with key for close
// It only needs to be opened once.
open({
library: 'libsum', // key
path: dynamicLib // path
})
const r = load({
library: "libsum", // path to the dynamic library file
funcName: 'sum', // the name of the function to call
retType: DataType.I32, // the return value type
paramsType: [DataType.I32, DataType.I32], // the parameter types
paramsValue: [a, b] // the actual parameter values
})
equal(r, a + b)
// release library memory when you're not using it.
close('libsum')
number|string|boolean|double|void
are basic types
const c = "foo"
const d = c.repeat(200)
equal(c + d, load({
library: 'libsum',
funcName: 'concatenateStrings',
retType: DataType.String,
paramsType: [DataType.String, DataType.String],
paramsValue: [c, d]
}))
equal(undefined, load({
library: 'libsum',
funcName: 'noRet',
retType: DataType.Void,
paramsType: [],
paramsValue: []
}))
equal(1.1 + 2.2, load({
library: 'libsum',
funcName: 'doubleSum',
retType: DataType.Double,
paramsType: [DataType.Double, DataType.Double],
paramsValue: [1.1, 2.2]
}))
const bool_val = true
equal(!bool_val, load({
library: 'libsum',
funcName: 'return_opposite',
retType: DataType.Boolean,
paramsType: [DataType.Boolean],
paramsValue: [bool_val],
}))
In the lateset version, ffi-rs
support modify data in place.
The sample code is as follows
extern int modifyData(char* buffer) {
// modify buffer data in place
}
const arr = Buffer.alloc(200) // create buffer
const res = load({
library: "libsum",
funcName: "modifyData",
retType: DataType.I32,
paramsType: [
DataType.U8Array
],
paramsValue: [arr]
})
console.log(arr) // buffer data can be updated
When use array
as retType
should use arrayConstructor
to specify array type with legal length which is important.
If the length is incorrect, program maybe exit abnormally
extern "C" int *createArrayi32(const int *arr, int size) {
int *vec = (int *)malloc((size) * sizeof(int));
for (int i = 0; i < size; i++) {
vec[i] = arr[i];
}
return vec;
}
extern "C" double *createArrayDouble(const double *arr, int size) {
double *vec = (double *)malloc((size) * sizeof(double));
for (int i = 0; i < size; i++) {
vec[i] = arr[i];
}
return vec;
}
extern "C" char **createArrayString(char **arr, int size) {
char **vec = (char **)malloc((size) * sizeof(char *));
for (int i = 0; i < size; i++) {
vec[i] = arr[i];
}
return vec;
}
let bigArr = new Array(100).fill(100)
deepStrictEqual(bigArr, load({
library: 'libsum',
funcName: 'createArrayi32',
retType: arrayConstructor({ type: DataType.I32Array, length: bigArr.length }),
paramsType: [DataType.I32Array, DataType.I32],
paramsValue: [bigArr, bigArr.length],
}))
let bigDoubleArr = new Array(5).fill(1.1)
deepStrictEqual(bigDoubleArr, load({
library: 'libsum',
funcName: 'createArrayDouble',
retType: arrayConstructor({ type: DataType.DoubleArray, length: bigDoubleArr.length }),
paramsType: [DataType.DoubleArray, DataType.I32],
paramsValue: [bigDoubleArr, bigDoubleArr.length],
}))
let stringArr = [c, c.repeat(20)]
deepStrictEqual(stringArr, load({
library: 'libsum',
funcName: 'createArrayString',
retType: arrayConstructor({ type: DataType.StringArray, length: stringArr.length }),
paramsType: [DataType.StringArray, DataType.I32],
paramsValue: [stringArr, stringArr.length],
}))
In ffi-rs
, we use DataType.External for wrap the pointer
which makes it can be passed between Node.js
and C
.
extern "C" const char *concatenateStrings(const char *str1, const char *str2) {
std::string result = std::string(str1) + std::string(str2);
char *cstr = new char[result.length() + 1];
strcpy(cstr, result.c_str());
return cstr;
}
extern "C" char *getStringFromPtr(void *ptr) { return (char *)ptr; };
// get pointer
const ptr = load({
library: "libsum",
funcName: "concatenateStrings",
retType: DataType.External,
paramsType: [DataType.String, DataType.String],
paramsValue: [c, d],
})
// send pointer
const string = load({
library: "libsum",
funcName: "getStringFromPtr",
retType: DataType.String,
paramsType: [DataType.External],
paramsValue: [ptr],
})
For create a c struct or get a c struct as a return type, you need to define the types of the parameters strictly in the order in which the fields of the c structure are defined.
typedef struct Person {
int age;
double *doubleArray;
Person *parent;
double doubleProps;
const char *name;
char **stringArray;
int *i32Array;
bool boolTrue;
bool boolFalse;
int64_t longVal;
char byte;
char *byteArray;
} Person;
extern "C" Person *getStruct(Person *person) {
return person;
}
extern "C" Person *createPerson() {
Person *person = (Person *)malloc(sizeof(Person));
// Allocate and initialize doubleArray
double initDoubleArray[] = {1.1, 2.2, 3.3};
person->doubleArray = (double *)malloc(sizeof(initDoubleArray));
memcpy(person->doubleArray, initDoubleArray, sizeof(initDoubleArray));
// Initialize age and doubleProps
person->age = 23;
person->doubleProps = 1.1;
person->byte = 'A';
// Allocate and initialize name
person->name = strdup("tom");
char *stringArray[] = {strdup("tom")};
person->stringArray = (char **)malloc(sizeof(stringArray));
memcpy(person->stringArray, stringArray, sizeof(stringArray));
// Allocate and initialize byteArray
char initByteArray[] = {101, 102};
person->byteArray = (char *)malloc(sizeof(initByteArray));
memcpy(person->byteArray, initByteArray, sizeof(initByteArray));
int initI32Array[] = {1, 2, 3, 4};
person->i32Array = (int *)malloc(sizeof(initI32Array));
memcpy(person->i32Array, initI32Array, sizeof(initI32Array));
person->boolTrue = true;
person->boolFalse = false;
person->longVal = 4294967296;
// Allocate and initialize parent
person->parent = (Person *)malloc(sizeof(Person));
double parentDoubleArray[] = {1.1, 2.2, 3.3};
person->parent->doubleArray = (double *)malloc(sizeof(parentDoubleArray));
memcpy(person->parent->doubleArray, parentDoubleArray,
sizeof(parentDoubleArray));
person->parent->age = 43;
person->parent->doubleProps = 3.3;
person->parent->name = strdup("tom father");
char *pstringArray[] = {strdup("tom"), strdup("father")};
person->parent->stringArray = (char **)malloc(sizeof(pstringArray));
memcpy(person->parent->stringArray, pstringArray, sizeof(pstringArray));
int parentI32Array[] = {5, 6, 7};
person->parent->i32Array = (int *)malloc(sizeof(parentI32Array));
memcpy(person->parent->i32Array, parentI32Array, sizeof(parentI32Array));
person->parent->boolTrue = true;
person->parent->boolFalse = false;
person->parent->longVal = 5294967296;
person->parent->byte = 'B';
char parentByteArray[] = {103, 104};
person->parent->byteArray = (char *)malloc(sizeof(parentByteArray));
memcpy(person->parent->byteArray, parentByteArray, sizeof(parentByteArray));
return person;
}
const parent = {
age: 43,
doubleArray: [1.1, 2.2, 3.3],
parent: {},
doubleProps: 3.3,
name: "tom father",
stringArray: ["tom", "father"],
i32Array: [5, 6, 7],
boolTrue: true,
boolFalse: false,
longVal: 5294967296,
byte: 66,
byteArray: Buffer.from([103, 104]),
};
const person = {
age: 23,
doubleArray: [1.1, 2.2, 3.3],
parent,
doubleProps: 1.1,
name: "tom",
stringArray: ["tom"],
i32Array: [1, 2, 3, 4],
boolTrue: true,
boolFalse: false,
longVal: 4294967296,
byte: 65,
byteArray: Buffer.from([101, 102]),
};
const parentType = {
age: DataType.I32,
doubleArray: arrayConstructor({
type: DataType.DoubleArray,
length: parent.doubleArray.length,
}),
parent: {},
doubleProps: DataType.Double,
name: DataType.String,
stringArray: arrayConstructor({
type: DataType.StringArray,
length: parent.stringArray.length,
}),
i32Array: arrayConstructor({
type: DataType.I32Array,
length: parent.i32Array.length,
}),
boolTrue: DataType.Boolean,
boolFalse: DataType.Boolean,
longVal: DataType.I64,
byte: DataType.U8,
byteArray: arrayConstructor({
type: DataType.U8Array,
length: parent.byteArray.length,
}),
};
const personType = {
age: DataType.I32,
doubleArray: arrayConstructor({
type: DataType.DoubleArray,
length: person.doubleArray.length,
}),
parent: parentType,
doubleProps: DataType.Double,
name: DataType.String,
stringArray: arrayConstructor({
type: DataType.StringArray,
length: person.stringArray.length,
}),
i32Array: arrayConstructor({
type: DataType.I32Array,
length: person.i32Array.length,
}),
boolTrue: DataType.Boolean,
boolFalse: DataType.Boolean,
longVal: DataType.I64,
byte: DataType.U8,
byteArray: arrayConstructor({
type: DataType.U8Array,
length: person.byteArray.length,
}),
};
const personObj = load({
library: "libsum",
funcName: "getStruct",
retType: personType,
paramsType: [
{
age: DataType.I32,
doubleArray: DataType.DoubleArray,
parent: {
parent: {},
age: DataType.I32,
doubleProps: DataType.Double,
name: DataType.String,
stringArray: DataType.StringArray,
doubleArray: DataType.DoubleArray,
i32Array: DataType.I32Array,
boolTrue: DataType.Boolean,
boolFalse: DataType.Boolean,
longVal: DataType.I64,
byte: DataType.U8,
byteArray: DataType.U8Array,
},
doubleProps: DataType.Double,
name: DataType.String,
stringArray: DataType.StringArray,
i32Array: DataType.I32Array,
boolTrue: DataType.Boolean,
boolFalse: DataType.Boolean,
longVal: DataType.I64,
byte: DataType.U8,
byteArray: DataType.U8Array,
},
],
paramsValue: [person],
});
deepStrictEqual(person, personObj);
const createdPerson = load({
library: "libsum",
funcName: "createPerson",
retType: personType,
paramsType: [],
paramsValue: [],
});
deepStrictEqual(createdPerson, person);
ffi-rs
supports passing js function to c, like this
typedef void (*FunctionPointer)(int a, bool b, char *c, char **d, int *e,
Person *p);
extern "C" void callFunction(FunctionPointer func) {
printf("callFunction\n");
for (int i = 0; i < 2; i++) {
int a = 100;
bool b = false;
double ddd = 100.11;
char *c = (char *)malloc(14 * sizeof(char));
strcpy(c, "Hello, World!");
char **stringArray = (char **)malloc(sizeof(char *) * 2);
stringArray[0] = strdup("Hello");
stringArray[1] = strdup("world");
int *i32Array = (int *)malloc(sizeof(int) * 3);
i32Array[0] = 101;
i32Array[1] = 202;
i32Array[2] = 303;
Person *p = createPerson();
func(a, b, c, stringArray, i32Array, p);
}
}
Corresponds to the code above,you can use ffi-rs
like
let count = 0;
const func = (a, b, c, d, e, f) => {
equal(a, 100);
equal(b, false);
equal(c, "Hello, World!");
deepStrictEqual(d, ["Hello", "world"]);
deepStrictEqual(e, [101, 202, 303]);
deepStrictEqual(f, person);
console.log("callback called");
count++;
if (count === 2) {
console.log("test succeed");
process.exit(0);
}
};
load({
library: "libsum",
funcName: "callFunction",
retType: DataType.Void,
paramsType: [
funcConstructor({
paramsType: [
DataType.I32,
DataType.Boolean,
DataType.String,
arrayConstructor({ type: DataType.StringArray, length: 2 }),
arrayConstructor({ type: DataType.I32Array, length: 3 }),
personType,
],
retType: DataType.Void,
}),
],
paramsValue: [func],
});
The function parameters supports type are all in the example above (double type is unsupported at this time), we will support more types in the future
Attention,since the vast majority of scenaros developers pass js function to c as a callback, so ffi-rs
will create threadsafe_function from jsfunction which means the jsfunction will be called asynchronous, and Node.js process will not be exited automatically
FAQs
English | 简体中文
The npm package @yuuang/ffi-rs-darwin-x64 receives a total of 9,966 weekly downloads. As such, @yuuang/ffi-rs-darwin-x64 popularity was classified as popular.
We found that @yuuang/ffi-rs-darwin-x64 demonstrated a healthy version release cadence and project activity because the last version was released less than a year ago. It has 0 open source maintainers collaborating on the project.
Did you know?
Socket for GitHub automatically highlights issues in each pull request and monitors the health of all your open source dependencies. Discover the contents of your packages and block harmful activity before you install or update your dependencies.
Security News
At its inaugural meeting, the JSR Working Group outlined plans for an open governance model and a roadmap to enhance JavaScript package management.
Security News
Research
An advanced npm supply chain attack is leveraging Ethereum smart contracts for decentralized, persistent malware control, evading traditional defenses.
Security News
Research
Attackers are impersonating Sindre Sorhus on npm with a fake 'chalk-node' package containing a malicious backdoor to compromise developers' projects.