Malformed Unicode escape sequence causes NumberFormatException instead of MalformedJsonException

[Marcono1234]

Gson version

2.10.1

Description

JsonReader throws a NumberFormatException instead of a MalformedJsonException when it encounters a malformed Unicode escape sequence in the JSON data.
This actually works as designed:

gson/gson/src/main/java/com/google/gson/stream/JsonReader.java

Lines 1590 to 1591 in 1998373

	* @throws NumberFormatException if any unicode escape sequences are
	* malformed.

However, it is questionable whether that design is really a good choice because a MalformedJsonException seems to fit better here, especially since other malformed escape sequences do cause a MalformedJsonException.

Note that NumberFormatException being thrown is apparently not publicly documented, so changing this should be rather safe to do.

Expected behavior

A MalformedJsonException is thrown for malformed Unicode escape sequences.

Actual behavior

A NumberFormatException is thrown.

Reproduction steps

Reader reader = new StringReader("\"\\uXYZ\"");
JsonReader jsonReader = new JsonReader(reader);
jsonReader.nextString();

[eamonnmcmanus]

I agree. It's weird to throw NumberFormatException here yet MalformedJsonException when for example there are fewer than 4 characters after the \u. I also agree that this is unlikely to affect anyone in practice, other than tests that might have been expecting the current exception.