ArrayBuffers (ArrayBuffer
)
Array Buffers allow highly efficient access to the same data from both JS and native.
Passing an ArrayBuffer
from native to JS and back does not involve any copies, and is therefore the fastest way of passing around data in Nitro.
- TypeScript
- C++
- Swift
- Kotlin
interface Image extends HybridObject {
getData(): ArrayBuffer
}
class HybridImage: public HybridImageSpec {
std::shared_ptr<ArrayBuffer> getData();
}
class HybridImage: HybridImageSpec {
func getData() -> ArrayBufferHolder
}
class HybridImage: HybridImageSpec() {
fun getData(): ArrayBuffer
}
It is important to understand the ownership, and threading concerns around such shared memory access.
Ownership
-
An
ArrayBuffer
that was created on the native side is owning, which means you can safely access it's data as long as theArrayBuffer
reference is alive. It can be safely held strong for longer, e.g. as a class property/member, and accessed from different Threads.func doSomething() -> ArrayBufferHolder {
let buffer = ArrayBufferHolder.allocate(1024 * 10)
let data = buffer.data // <-- ✅ safe to do because we own it!
self.buffer = buffer // <-- ✅ safe to do to use it later!
DispatchQueue.global().async {
let data = buffer.data // <-- ✅ also safe because we own it!
}
return buffer
} -
An
ArrayBuffer
that was received as a parameter from JS cannot be safely kept strong as the JS VM can delete it at any point, hence it is non-owning. It's data can only be safely accessed before the synchronous function returned, as this will stay within the JS bounds.func doSomething(buffer: ArrayBufferHolder) {
let data = buffer.data // <-- ✅ safe to do because we're still sync
DispatchQueue.global().async {
let data = buffer.data // <-- ❌ NOT safe
}
}If you need a non-owning buffer's data for longer, copy it first:
func doSomething(buffer: ArrayBufferHolder) {
let copy = ArrayBufferHolder.copy(of: buffer)
DispatchQueue.global().async {
let data = copy.data // <-- ✅ safe now because we have a owning copy
}
}
Threading
An ArrayBuffer
can be accessed from both JS and native, and even from multiple Threads at once, but they are not thread-safe.
To prevent race conditions or garbage-data from being read, make sure to not read from- and write to- the ArrayBuffer
at the same time.