Creating codex for Map[Integer, String]

[iblislin]

Hi,

I got errors on using Java Integer as Map key

scala> JsonCodecMaker.make[Map[java.lang.Integer, String]]
-- Error: ----------------------------------------------------------------------
  1 |JsonCodecMaker.make[Map[java.lang.Integer, String]]
    |^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    |Exception occurred while executing macro expansion.
    |java.lang.Exception: Expr cast exception: in.readKeyAsInt()
    |of type: scala.Int
    |did not conform to type: java.lang.Integer
    |
    |
    |---------------------------------------------------------------------------
    |Inline stack trace
    |- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
    |This location contains code that was inlined from JsonCodecMaker.scala:511
     ---------------------------------------------------------------------------

The java.lang.Long got simliar issue

scala> JsonCodecMaker.make[Map[java.lang.Long, String]]
-- Error: ----------------------------------------------------------------------
  1 |JsonCodecMaker.make[Map[java.lang.Long, String]]
    |^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
...

Since I'm working with Kafka Java library, I got some Java Integer from it.

[plokhotnyuk]

@iblislin Hi, Iblis! Thanks for reporting!

I've reproduced it locally for Scala 3 (Scala 2 is not affected) and going to fix and cut a release this evening.

As a workaround you can try the following custom codec for keys:

implicit val keyCodec: JsonKeyCodec[_root_.java.lang.Integer] = new JsonKeyCodec[_root_.java.lang.Integer] {
  override def decodeKey(in: JsonReader): _root_.java.lang.Integer = _root_.java.lang.Integer.valueOf(in.readKeyAsInt())

  override def encodeKey(x: _root_.java.lang.Integer, out: JsonWriter): _root_.scala.Unit = out.writeKey(x.intValue())
}

[plokhotnyuk]

@iblislin Please peek and try the latest release with a fix: https://github.com/plokhotnyuk/jsoniter-scala/releases/tag/v2.23.3

[iblislin]

I tested it, works fine for me. Thanks! 🙏