perf(ext/web): use base64-simd for atob/btoa (#14992)

Cargo.lock

@@ -242,6 +242,15 @@ version = "0.13.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "904dfeac50f3cdaba28fc6f57fdcddb75f49ed61346676a78c4ffe55877802fd"
 
+[[package]]
+name = "base64-simd"
+version = "0.6.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "278c7ba87265587c4823cf1b2fdf57834151540b2e509574adb03627f8c7f22d"
+dependencies = [
+ "simd-abstraction",
+]
+
 [[package]]
 name = "base64ct"
 version = "1.5.0"
@@ -1202,7 +1211,7 @@ name = "deno_web"
 version = "0.89.0"
 dependencies = [
  "async-trait",
- "base64 0.13.0",
+ "base64-simd",
  "deno_bench_util",
  "deno_core",
  "deno_url",
@@ -3893,6 +3902,12 @@ dependencies = [
  "rand_core",
 ]
 
+[[package]]
+name = "simd-abstraction"
+version = "0.6.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c2880f3f7b392823ee65bbcc681961cd8e698c6a30e91ab9b4eef1f9c6c226d8"
+
 [[package]]
 name = "siphasher"
 version = "0.3.10"

Cargo.toml

@@ -110,6 +110,8 @@ opt-level = 3
 opt-level = 3
 [profile.bench.package.zstd-sys]
 opt-level = 3
+[profile.bench.package.base64-simd]
+opt-level = 3
 
 # NB: the `bench` and `release` profiles must remain EXACTLY the same.
 [profile.release.package.rand]
@@ -174,3 +176,5 @@ opt-level = 3
 opt-level = 3
 [profile.release.package.zstd-sys]
 opt-level = 3
+[profile.release.package.base64-simd]
+opt-level = 3

ext/web/Cargo.toml

@@ -15,7 +15,7 @@ path = "lib.rs"
 
 [dependencies]
 async-trait = "0.1.51"
-base64 = "0.13.0"
+base64-simd = "0.6.2"
 deno_core = { version = "0.140.0", path = "../../core" }
 encoding_rs = "0.8.31"
 flate2 = "1"

ext/web/lib.rs

@@ -124,74 +124,41 @@ pub fn init<P: TimersPermission + 'static>(
 
 #[op]
 fn op_base64_decode(input: String) -> Result<ZeroCopyBuf, AnyError> {
-  let mut input = input.into_bytes();
-  input.retain(|c| !c.is_ascii_whitespace());
-  Ok(b64_decode(&input)?.into())
+  Ok(forgiving_base64_decode(input.into_bytes())?.into())
 }
 
 #[op]
-fn op_base64_atob(mut s: ByteString) -> Result<ByteString, AnyError> {
-  s.retain(|c| !c.is_ascii_whitespace());
-
-  // If padding is expected, fail if not 4-byte aligned
-  if s.len() % 4 != 0 && (s.ends_with(b"==") || s.ends_with(b"=")) {
-    return Err(
-      DomExceptionInvalidCharacterError::new("Failed to decode base64.").into(),
-    );
+fn op_base64_atob(s: ByteString) -> Result<ByteString, AnyError> {
+  Ok(forgiving_base64_decode(s.into())?.into())
 }
 
-  Ok(b64_decode(&s)?.into())
-}
+/// See <https://infra.spec.whatwg.org/#forgiving-base64>
+fn forgiving_base64_decode(mut input: Vec<u8>) -> Result<Vec<u8>, AnyError> {
+  let error: _ =
+    || DomExceptionInvalidCharacterError::new("Failed to decode base64");
 
-fn b64_decode(input: &[u8]) -> Result<Vec<u8>, AnyError> {
-  // "If the length of input divides by 4 leaving no remainder, then:
-  //  if input ends with one or two U+003D EQUALS SIGN (=) characters,
-  //  remove them from input."
-  let input = match input.len() % 4 == 0 {
-    true if input.ends_with(b"==") => &input[..input.len() - 2],
-    true if input.ends_with(b"=") => &input[..input.len() - 1],
-    _ => input,
-  };
+  let decoded = base64_simd::Base64::forgiving_decode_inplace(&mut input)
+    .map_err(|_| error())?;
 
-  // "If the length of input divides by 4 leaving a remainder of 1,
-  //  throw an InvalidCharacterError exception and abort these steps."
-  if input.len() % 4 == 1 {
-    return Err(
-      DomExceptionInvalidCharacterError::new("Failed to decode base64.").into(),
-    );
-  }
-
-  let cfg = base64::Config::new(base64::CharacterSet::Standard, true)
-    .decode_allow_trailing_bits(true);
-  let out = base64::decode_config(input, cfg).map_err(|err| match err {
-    base64::DecodeError::InvalidByte(_, _) => {
-      DomExceptionInvalidCharacterError::new(
-        "Failed to decode base64: invalid character",
-      )
-    }
-    _ => DomExceptionInvalidCharacterError::new(&format!(
-      "Failed to decode base64: {:?}",
-      err
-    )),
-  })?;
-
-  Ok(out)
+  let decoded_len = decoded.len();
+  input.truncate(decoded_len);
+  Ok(input)
 }
 
 #[op]
 fn op_base64_encode(s: ZeroCopyBuf) -> String {
-  b64_encode(&s)
+  forgiving_base64_encode(s.as_ref())
 }
 
 #[op]
 fn op_base64_btoa(s: ByteString) -> String {
-  b64_encode(s)
+  forgiving_base64_encode(s.as_ref())
 }
 
-fn b64_encode(s: impl AsRef<[u8]>) -> String {
-  let cfg = base64::Config::new(base64::CharacterSet::Standard, true)
-    .decode_allow_trailing_bits(true);
-  base64::encode_config(s.as_ref(), cfg)
+/// See <https://infra.spec.whatwg.org/#forgiving-base64>
+fn forgiving_base64_encode(s: &[u8]) -> String {
+  let base64 = base64_simd::Base64::STANDARD;
+  base64.encode_to_boxed_str(s).into_string()
 }
 
 #[derive(Deserialize)]