Skip to main content
Deno 2 is finally here 🎉️
Learn more
Module

std/_wasm_crypto/src/lib.rs

Deno standard library
Go to Latest
The Standard Library has been moved to JSR. See the blog post for details.
File
// Copyright 2018-2021 the Deno authors. All rights reserved. MIT license.use derive_more::{From, Into};use js_sys::Uint8Array;use wasm_bindgen::prelude::*;
mod digest;
/// Returns the digest of the given `data` using the given hash `algorithm`.////// `length` will usually be left `undefined` to use the default length for/// the algorithm. For algorithms with variable-length output, it can be used/// to specify a non-negative integer number of bytes.////// An error will be thrown if `algorithm` is not a supported hash algorithm or/// `length` is not a supported length for the algorithm.#[wasm_bindgen]pub fn digest( algorithm: String, data: js_sys::Uint8Array, length: Option<usize>,) -> Result<Box<[u8]>, JsValue> { let mut context = DigestContext::new(algorithm)?; context.update(data)?; context.digest_and_drop(length)}
/// A context for incrementally computing a digest using a given hash algorithm.#[wasm_bindgen]#[derive(Clone, Into, From)]pub struct DigestContext(digest::Context);
#[wasm_bindgen]impl DigestContext { /// Creates a new context incrementally computing a digest using the given /// hash algorithm. /// /// An error will be thrown if `algorithm` is not a supported hash algorithm. #[wasm_bindgen(constructor)] pub fn new(algorithm: String) -> Result<DigestContext, JsValue> { Ok( digest::Context::new(&algorithm) .map_err(|message| JsValue::from(js_sys::TypeError::new(message)))? .into(), ) }
/// Update the digest's internal state with the additional input `data`. /// /// If the `data` array view is large, it will be split into subarrays (via /// JavaScript bindings) which will be processed sequentially in order to /// limit the amount of memory that needs to be allocated in the WASM heap. #[wasm_bindgen] pub fn update(&mut self, data: js_sys::Uint8Array) -> Result<(), JsValue> { // Every method call on `data` has to go through the JavaScript bindings, so // try to minimize them. Splitting on the JavaScript side could be more // efficient, but this is called from multiple places in JavaScript so it // seems simpler to keep it here for now.
let length = data.byte_length(); let base_offset = data.byte_offset(); let buffer = data.buffer();
let chunk_size: u32 = 65_536;
if length <= chunk_size { let chunk = data.to_vec(); self.0.update(&chunk); } else { let mut offset = 0; while offset < length { let chunk = Uint8Array::new_with_byte_offset_and_length( &buffer, base_offset + offset, chunk_size.min(length - offset), ) .to_vec(); self.0.update(&chunk); offset += chunk_size; } }
Ok(()) }
/// Returns the digest of the input data so far. This may be called repeatedly /// without side effects. /// /// `length` will usually be left `undefined` to use the default length for /// the algorithm. For algorithms with variable-length output, it can be used /// to specify a non-negative integer number of bytes. /// /// An error will be thrown if `algorithm` is not a supported hash algorithm or /// `length` is not a supported length for the algorithm. #[wasm_bindgen] pub fn digest(&self, length: Option<usize>) -> Result<Box<[u8]>, JsValue> { self .0 .digest(length) .map_err(|message| JsValue::from(js_sys::TypeError::new(message))) }
/// Returns the digest of the input data so far, and resets this context to /// its initial state, as though it has not yet been provided with any input /// data. (It will still use the same algorithm.) /// /// `length` will usually be left `undefined` to use the default length for /// the algorithm. For algorithms with variable-length output, it can be used /// to specify a non-negative integer number of bytes. /// /// An error will be thrown if `algorithm` is not a supported hash algorithm or /// `length` is not a supported length for the algorithm. #[wasm_bindgen(js_name=digestAndReset)] pub fn digest_and_reset( &mut self, length: Option<usize>, ) -> Result<Box<[u8]>, JsValue> { self .0 .digest_and_reset(length) .map_err(|message| JsValue::from(js_sys::TypeError::new(message))) }
/// Returns the digest of the input data so far, and then drops the context /// from memory on the WASM side. This context must no longer be used, and any /// further method calls will result in null pointer errors being thrown. /// https://github.com/rustwasm/wasm-bindgen/blob/bf39cfd8/crates/backend/src/codegen.rs#L186 /// /// `length` will usually be left `undefined` to use the default length for /// the algorithm. For algorithms with variable-length output, it can be used /// to specify a non-negative integer number of bytes. /// /// An error will be thrown if `algorithm` is not a supported hash algorithm or /// `length` is not a supported length for the algorithm. #[wasm_bindgen(js_name=digestAndDrop)] pub fn digest_and_drop( mut self, length: Option<usize>, ) -> Result<Box<[u8]>, JsValue> { self .0 .digest_and_reset(length) .map_err(|message| JsValue::from(js_sys::TypeError::new(message))) }
/// Resets this context to its initial state, as though it has not yet been /// provided with any input data. (It will still use the same algorithm.) #[wasm_bindgen] pub fn reset(&mut self) -> Result<(), JsValue> { self.0.reset();
Ok(()) }
/// Returns a new `DigestContext` that is a copy of this one, i.e., using the /// same algorithm and with a copy of the same internal state. /// /// This may be a more efficient option for computing multiple digests that /// start with a common prefix. #[wasm_bindgen] #[allow(clippy::should_implement_trait)] pub fn clone(&self) -> DigestContext { Clone::clone(self) }}