Optimize ASCII-safe string rendering

sjsonnet/src-js/sjsonnet/Platform.scala

@@ -23,6 +23,18 @@ object Platform {
     }
   }
 
+  // Used only for strings already proven ASCII-safe; copies low bytes without allocation.
+  @inline def copyAsciiStringToBytes(s: String, dst: Array[Byte], dstPos: Int): Unit = {
+    var i = 0
+    var pos = dstPos
+    val len = s.length
+    while (i < len) {
+      dst(pos) = s.charAt(i).toByte
+      i += 1
+      pos += 1
+    }
+  }
+
   private def nodeToJson(node: Node): ujson.Value = node match {
     case _: Node.ScalarNode =>
       YamlDecoder.forAny.construct(node).getOrElse("") match {

sjsonnet/src-jvm/sjsonnet/Platform.scala

@@ -13,6 +13,7 @@ import org.tukaani.xz.XZOutputStream
 import org.yaml.snakeyaml.{LoaderOptions, Yaml}
 import org.yaml.snakeyaml.constructor.SafeConstructor
 
+import scala.annotation.nowarn
 import scala.collection.compat.*
 import scala.collection.mutable
 import scala.jdk.CollectionConverters.*
@@ -23,6 +24,12 @@ object Platform {
   def repeatString(s: String, count: Int): String =
     if (count <= 0) "" else s.repeat(count)
 
+  // Used only for strings already proven ASCII-safe; copies low bytes without allocation.
+  @inline
+  @nowarn("cat=deprecation")
+  def copyAsciiStringToBytes(s: String, dst: Array[Byte], dstPos: Int): Unit =
+    s.getBytes(0, s.length, dst, dstPos)
+
   def gzipBytes(b: Array[Byte]): String = {
     val outputStream: ByteArrayOutputStream = new ByteArrayOutputStream(b.length)
     val gzip: GZIPOutputStream = new GZIPOutputStream(outputStream)

sjsonnet/src-native/sjsonnet/Platform.scala

@@ -6,6 +6,7 @@ import java.util
 import java.util.Base64
 import java.util.zip.GZIPOutputStream
 import scala.scalanative.regex.Pattern
+import scala.annotation.nowarn
 import scala.collection.mutable
 import org.virtuslab.yaml.*
 
@@ -28,6 +29,12 @@ object Platform {
     }
   }
 
+  // Used only for strings already proven ASCII-safe; copies low bytes without allocation.
+  @inline
+  @nowarn("cat=deprecation")
+  def copyAsciiStringToBytes(s: String, dst: Array[Byte], dstPos: Int): Unit =
+    s.getBytes(0, s.length, dst, dstPos)
+
   def gzipBytes(b: Array[Byte]): String = {
     val outputStream: ByteArrayOutputStream = new ByteArrayOutputStream(b.length)
     val gzip: GZIPOutputStream = new GZIPOutputStream(outputStream)

sjsonnet/src/sjsonnet/BaseByteRenderer.scala

@@ -253,12 +253,8 @@ class BaseByteRenderer[T <: java.io.OutputStream](
     var pos = elemBuilder.length
     arr(pos) = '"'.toByte
     pos += 1
-    var i = 0
-    while (i < len) {
-      arr(pos) = str.charAt(i).toByte
-      pos += 1
-      i += 1
-    }
+    Platform.copyAsciiStringToBytes(str, arr, pos)
+    pos += len
     arr(pos) = '"'.toByte
     elemBuilder.length = pos + 1
   }

sjsonnet/src/sjsonnet/stdlib/EncodingModule.scala

@@ -22,6 +22,16 @@ import sjsonnet.functions.AbstractFunctionModule
 object EncodingModule extends AbstractFunctionModule {
   def name = "encoding"
 
+  @inline private def isAsciiJsonSafe(bytes: Array[Byte]): Boolean = {
+    var i = 0
+    while (i < bytes.length) {
+      val b = bytes(i) & 0xff
+      if (b < 0x20 || b >= 0x80 || b == '"' || b == '\\') return false
+      i += 1
+    }
+    true
+  }
+
   /**
    * [[https://jsonnet.org/ref/stdlib.html#std-md5 std.md5(s)]].
    *
@@ -88,9 +98,12 @@ object EncodingModule extends AbstractFunctionModule {
      * Behaves like std.base64DecodeBytes() except returns a naively encoded string instead of an
      * array of bytes.
      */
-    builtin("base64Decode", "str") { (_, _, str: String) =>
+    builtin("base64Decode", "str") { (pos, _, str: String) =>
       try {
-        new String(PlatformBase64.decode(str), UTF_8)
+        val decoded = PlatformBase64.decode(str)
+        val value = new String(decoded, UTF_8)
+        (if (isAsciiJsonSafe(decoded)) Val.Str.asciiSafe(pos, value)
+         else Val.Str(pos, value)): Val
       } catch {
         case e: IllegalArgumentException =>
           Error.fail("Invalid base64 string: " + e.getMessage)