perf(ext/web): optimize atob/btoa (#13841)

Follow up to #13839, optimizing `base64_roundtrip` ~20x (~125ms => ~6.5ms)
denoland · Mar 5, 2022 · 72d593f · 72d593f
1 parent 96dc742
commit 72d593f
Show file tree

Hide file tree

Showing 3 changed files with 83 additions and 64 deletions.
diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
@@ -263,7 +263,7 @@ jobs:
             ~/.cargo/registry/index
             ~/.cargo/registry/cache
             ~/.cargo/git/db
-          key: 4-cargo-home-${{ matrix.os }}-${{ hashFiles('Cargo.lock') }}
+          key: 5-cargo-home-${{ matrix.os }}-${{ hashFiles('Cargo.lock') }}
 
       # In main branch, always creates fresh cache
       - name: Cache build output (main)
@@ -279,7 +279,7 @@ jobs:
             !./target/*/*.zip
             !./target/*/*.tar.gz
           key: |
-            4-cargo-target-${{ matrix.os }}-${{ matrix.profile }}-${{ github.sha }}
+            5-cargo-target-${{ matrix.os }}-${{ matrix.profile }}-${{ github.sha }}
 
       # Restore cache from the latest 'main' branch build.
       - name: Cache build output (PR)
@@ -295,7 +295,7 @@ jobs:
             !./target/*/*.tar.gz
           key: never_saved
           restore-keys: |
-            4-cargo-target-${{ matrix.os }}-${{ matrix.profile }}-
+            5-cargo-target-${{ matrix.os }}-${{ matrix.profile }}-
 
       # Don't save cache after building PRs or branches other than 'main'.
       - name: Skip save cache (PR)

diff --git a/ext/web/05_base64.js b/ext/web/05_base64.js
@@ -9,21 +9,10 @@
 "use strict";
 
 ((window) => {
+  const core = Deno.core;
   const webidl = window.__bootstrap.webidl;
-  const {
-    forgivingBase64Encode,
-    forgivingBase64Decode,
-  } = window.__bootstrap.infra;
   const { DOMException } = window.__bootstrap.domException;
-  const {
-    ArrayPrototypeMap,
-    StringPrototypeCharCodeAt,
-    ArrayPrototypeJoin,
-    SafeArrayIterator,
-    StringFromCharCode,
-    TypedArrayFrom,
-    Uint8Array,
-  } = window.__bootstrap.primordials;
+  const { TypeError } = window.__bootstrap.primordials;
 
   /**
    * @param {string} data
@@ -36,13 +25,17 @@
       prefix,
       context: "Argument 1",
     });
-
-    const uint8Array = forgivingBase64Decode(data);
-    const result = ArrayPrototypeMap(
-      [...new SafeArrayIterator(uint8Array)],
-      (byte) => StringFromCharCode(byte),
-    );
-    return ArrayPrototypeJoin(result, "");
+    try {
+      return core.opSync("op_base64_atob", data);
+    } catch (e) {
+      if (e instanceof TypeError) {
+        throw new DOMException(
+          "Failed to decode base64: invalid character",
+          "InvalidCharacterError",
+        );
+      }
+      throw e;
+    }
   }
 
   /**
@@ -56,20 +49,17 @@
       prefix,
       context: "Argument 1",
     });
-    const byteArray = ArrayPrototypeMap(
-      [...new SafeArrayIterator(data)],
-      (char) => {
-        const charCode = StringPrototypeCharCodeAt(char, 0);
-        if (charCode > 0xff) {
-          throw new DOMException(
-            "The string to be encoded contains characters outside of the Latin1 range.",
-            "InvalidCharacterError",
-          );
-        }
-        return charCode;
-      },
-    );
-    return forgivingBase64Encode(TypedArrayFrom(Uint8Array, byteArray));
+    try {
+      return core.opSync("op_base64_btoa", data);
+    } catch (e) {
+      if (e instanceof TypeError) {
+        throw new DOMException(
+          "The string to be encoded contains characters outside of the Latin1 range.",
+          "InvalidCharacterError",
+        );
+      }
+      throw e;
+    }
   }
 
   window.__bootstrap.base64 = {

diff --git a/ext/web/lib.rs b/ext/web/lib.rs
@@ -12,6 +12,7 @@ use deno_core::include_js_files;
 use deno_core::op_async;
 use deno_core::op_sync;
 use deno_core::url::Url;
+use deno_core::ByteString;
 use deno_core::Extension;
 use deno_core::OpState;
 use deno_core::Resource;
@@ -85,6 +86,8 @@ pub fn init<P: TimersPermission + 'static>(
     .ops(vec![
       ("op_base64_decode", op_sync(op_base64_decode)),
       ("op_base64_encode", op_sync(op_base64_encode)),
+      ("op_base64_atob", op_sync(op_base64_atob)),
+      ("op_base64_btoa", op_sync(op_base64_btoa)),
       (
         "op_encoding_normalize_label",
         op_sync(op_encoding_normalize_label),
@@ -146,21 +149,42 @@ pub fn init<P: TimersPermission + 'static>(
 }
 
 fn op_base64_decode(
-  _state: &mut OpState,
+  _: &mut OpState,
   input: String,
   _: (),
 ) -> Result<ZeroCopyBuf, AnyError> {
-  let mut input: &str = &input.replace(|c| char::is_ascii_whitespace(&c), "");
+  let mut input = input.into_bytes();
+  input.retain(|c| !c.is_ascii_whitespace());
+  Ok(b64_decode(&input)?.into())
+}
+
+fn op_base64_atob(
+  _: &mut OpState,
+  s: ByteString,
+  _: (),
+) -> Result<ByteString, AnyError> {
+  let mut s = s.0;
+  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(),
+    );
+  }
+
+  Ok(ByteString(b64_decode(&s)?))
+}
+
+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."
-  if input.len() % 4 == 0 {
-    if input.ends_with("==") {
-      input = &input[..input.len() - 2]
-    } else if input.ends_with('=') {
-      input = &input[..input.len() - 1]
-    }
-  }
+  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,
+  };
 
   // "If the length of input divides by 4 leaving a remainder of 1,
   //  throw an InvalidCharacterError exception and abort these steps."
@@ -170,38 +194,43 @@ fn op_base64_decode(
     );
   }
 
-  if input
-    .chars()
-    .any(|c| c != '+' && c != '/' && !c.is_alphanumeric())
-  {
-    return Err(
+  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",
       )
-      .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| {
-    DomExceptionInvalidCharacterError::new(&format!(
+    }
+    _ => DomExceptionInvalidCharacterError::new(&format!(
       "Failed to decode base64: {:?}",
       err
-    ))
+    )),
   })?;
-  Ok(ZeroCopyBuf::from(out))
+
+  Ok(out)
 }
 
 fn op_base64_encode(
-  _state: &mut OpState,
+  _: &mut OpState,
   s: ZeroCopyBuf,
   _: (),
 ) -> Result<String, AnyError> {
+  Ok(b64_encode(&s))
+}
+
+fn op_base64_btoa(
+  _: &mut OpState,
+  s: ByteString,
+  _: (),
+) -> Result<String, AnyError> {
+  Ok(b64_encode(&s))
+}
+
+fn b64_encode(s: impl AsRef<[u8]>) -> String {
   let cfg = base64::Config::new(base64::CharacterSet::Standard, true)
     .decode_allow_trailing_bits(true);
-  let out = base64::encode_config(&s, cfg);
-  Ok(out)
+  base64::encode_config(s.as_ref(), cfg)
 }
 
 #[derive(Deserialize)]