perf(pg-protocol): decode short ascii values without buffer.toString()

packages/pg-protocol/src/buffer-reader.ts

@@ -1,3 +1,45 @@
+const fromCharCode = String.fromCharCode
+
+// Longest value decoded without calling into buffer.toString(). Above this
+// length the per-call cost of toString() is cheaper than decoding in JS.
+const MAX_ASCII_LENGTH = 16
+
+// Most values in a result set are short: ids, flags, small numbers, dates. For
+// those the fixed cost of the C++ call behind buffer.toString() is bigger than
+// the decoding itself, so build the string from the char codes instead.
+// Returns undefined when the bytes are not ASCII, so the caller falls back to
+// buffer.toString().
+// prettier-ignore
+const decodeAscii = (b: Buffer, i: number, length: number): string | undefined => {
+  let bits = 0
+  for (let k = 0; k < length; k++) {
+    bits |= b[i + k]
+  }
+  if (bits > 127) {
+    return undefined
+  }
+  switch (length) {
+    case 0: return ''
+    case 1: return fromCharCode(b[i])
+    case 2: return fromCharCode(b[i], b[i + 1])
+    case 3: return fromCharCode(b[i], b[i + 1], b[i + 2])
+    case 4: return fromCharCode(b[i], b[i + 1], b[i + 2], b[i + 3])
+    case 5: return fromCharCode(b[i], b[i + 1], b[i + 2], b[i + 3], b[i + 4])
+    case 6: return fromCharCode(b[i], b[i + 1], b[i + 2], b[i + 3], b[i + 4], b[i + 5])
+    case 7: return fromCharCode(b[i], b[i + 1], b[i + 2], b[i + 3], b[i + 4], b[i + 5], b[i + 6])
+    case 8: return fromCharCode(b[i], b[i + 1], b[i + 2], b[i + 3], b[i + 4], b[i + 5], b[i + 6], b[i + 7])
+    case 9: return fromCharCode(b[i], b[i + 1], b[i + 2], b[i + 3], b[i + 4], b[i + 5], b[i + 6], b[i + 7], b[i + 8])
+    case 10: return fromCharCode(b[i], b[i + 1], b[i + 2], b[i + 3], b[i + 4], b[i + 5], b[i + 6], b[i + 7], b[i + 8], b[i + 9])
+    case 11: return fromCharCode(b[i], b[i + 1], b[i + 2], b[i + 3], b[i + 4], b[i + 5], b[i + 6], b[i + 7], b[i + 8], b[i + 9], b[i + 10])
+    case 12: return fromCharCode(b[i], b[i + 1], b[i + 2], b[i + 3], b[i + 4], b[i + 5], b[i + 6], b[i + 7], b[i + 8], b[i + 9], b[i + 10], b[i + 11])
+    case 13: return fromCharCode(b[i], b[i + 1], b[i + 2], b[i + 3], b[i + 4], b[i + 5], b[i + 6], b[i + 7], b[i + 8], b[i + 9], b[i + 10], b[i + 11], b[i + 12])
+    case 14: return fromCharCode(b[i], b[i + 1], b[i + 2], b[i + 3], b[i + 4], b[i + 5], b[i + 6], b[i + 7], b[i + 8], b[i + 9], b[i + 10], b[i + 11], b[i + 12], b[i + 13])
+    case 15: return fromCharCode(b[i], b[i + 1], b[i + 2], b[i + 3], b[i + 4], b[i + 5], b[i + 6], b[i + 7], b[i + 8], b[i + 9], b[i + 10], b[i + 11], b[i + 12], b[i + 13], b[i + 14])
+    case 16: return fromCharCode(b[i], b[i + 1], b[i + 2], b[i + 3], b[i + 4], b[i + 5], b[i + 6], b[i + 7], b[i + 8], b[i + 9], b[i + 10], b[i + 11], b[i + 12], b[i + 13], b[i + 14], b[i + 15])
+  }
+  return undefined
+}
+
 export class BufferReader {
   private buffer: Buffer = Buffer.allocUnsafe(0)
 
@@ -36,9 +78,15 @@ export class BufferReader {
   }
 
   public string(length: number): string {
-    const result = this.buffer.toString(this.encoding, this.offset, this.offset + length)
-    this.offset += length
-    return result
+    const start = this.offset
+    this.offset = start + length
+    if (length <= MAX_ASCII_LENGTH) {
+      const ascii = decodeAscii(this.buffer, start, length)
+      if (ascii !== undefined) {
+        return ascii
+      }
+    }
+    return this.buffer.toString(this.encoding, start, this.offset)
   }
 
   public cstring(): string {

packages/pg-protocol/src/inbound-parser.test.ts

@@ -315,6 +315,26 @@ describe('PgPacketStream', function () {
         fields: ['test'],
       })
     })
+
+    it('decodes values of any length and encoding', function () {
+      // short ascii values take a different decoding path than long or
+      // multi byte ones, so check both sides of it
+      const values = [
+        ...Array.from({ length: 20 }, (_, i) => 'x'.repeat(i)),
+        '\u00e9',
+        'citt\u00e0',
+        '\u00fcn\u00efc\u00f6d\u00e9 w\u00f6rld',
+        '\u4e2d\u6587\u5b57\u7b26\u4e32',
+        '\ud83c\udf89',
+        'a\ud83c\udf89b',
+        null,
+      ]
+      const fields: (string | null)[] = []
+      new Parser().parse(buffers.dataRow(values), (msg) => {
+        fields.push(...(msg as any).fields)
+      })
+      assert.deepStrictEqual(fields, values)
+    })
   })
 
   describe('notice message', function () {