Implicit conversion from integral numbers to byte array

[scordio]

I'd like to propose the addition of implicit conversion from integral numbers (byte, short, int, long) to byte array. The concrete use case is code dealing with bitwise operations, mostly related to cryptography.

My wish is to write something like this:

  @ParameterizedTest
  @ValueSource(ints = {Integer.MIN_VALUE, 0, 0xFF, 0xFFFF, 0xFFFFFF, Integer.MAX_VALUE})
  void test(byte[] byteArray) {
    assertThat(byteArray).hasSize(4);
  }

where I can leverage binary and hexadecimal literals which are automatically converted to a byte array using a ByteBuffer with Big Endian order.

I put together a small poc with a custom argument converter. If the proposal is accepted, I'm happy to raise a PR about it.

Deliverables

• A new implicit conversion from byte / short / int / long to byte[]

[sbrannen]

For that use case, you wouldn't use an ArgumentConverter since that can only convert a single argument.

Rather, you would use an ArgumentsAggregator to aggregate all arguments into a single array.

[sbrannen]

In addition, I don't see how we could introduce implicit support for that without breaking existing use cases, but perhaps I'm overlooking something.

[scordio]

@sbrannen I meant something different. Probably a more detailed example would've explained it better.

I don't want to aggregate all the arguments in a single array, I want each argument to be converted into a dedicated byte array, with array size depending on the primitive type.

Following my example:

@ValueSource(ints = {Integer.MIN_VALUE, 0, 0xFF, 0xFFFF, 0xFFFFFF, Integer.MAX_VALUE})

each integer would be converted to a separate byte array of size 4, being the size in bytes of the integer type (i.e., Integer.BYTES). In case of byte values, each array would have size 1; shorts would have arrays of size 2, and longs would have arrays of size 8.

The test cases for the custom converter might explain it better than my words 🙂

[scordio]

Written in a style similar to the official docs, this is what I'd like to have:

Source type	Target type	Example
byte / Byte	byte[]	0xF → ByteBuffer.allocate(Byte.BYTES).put((byte) 0xF).array()
short / Short	byte[]	0xFFF → ByteBuffer.allocate(Short.BYTES).putShort((short) 0xFFF).array()
int / Integer	byte[]	0xFFFFFFF → ByteBuffer.allocate(Integer.BYTES).putInt(0xFFFFFFF).array()
long / Long	byte[]	0xFFFFFFFFL → ByteBuffer.allocate(Long.BYTES).putLong(0xFFFFFFFFL).array()

[scordio]

I raised #2711 to help evaluate this feature.

[marcphilipp]

I'm not yet convinced how generally useful such a built-in conversion would be. How about making it opt-in via an annotation like @JavaTimeConversionPattern?

[scordio]

I think that bitwise operations are not the most common use case for Java development, so I'm probably targeting a niche area. My thoughts were that such an implicit conversion would not introduce side effects for other use cases and test authors could always define their own conversion logic, if there is any.

Having an annotation would also be nice. What I'd like to avoid is defining a custom converter and sharing it with multiple projects, while this could be a generally usable feature. The annotation could also have an optional parameter to change the endianness, which I would keep as big endian by default.

Do you have any idea about naming? Something like @IntegralToBytesConversion?

[scordio]

As the stale bot just commented on #2711, I wanted to check if any decisions have been made about how this topic might move forward. Would @IntegralToBytesConversion be the right way to go?

[stale]

This issue has been automatically marked as stale because it has not had recent activity. Given the limited bandwidth of the team, it will be automatically closed if no further activity occurs. Thank you for your contribution.

[scordio]

As the stale bot just commented on #2711, I wanted to check if any decisions have been made about how this topic might move forward. Would @IntegralToBytesConversion be the right way to go?

Bump

[marcphilipp]

@scordio Sorry for the delay! I think @IntegralToBytesConversion or @IntegralToByteArrayConversion sounds fine. Have you considered submitting this to JUnit Pioneer instead?

[scordio]

I was actually thinking about it today 😄 maybe combining this topic with the proposal at #3141 (comment).

As far as I remember, there are no argument converters in JUnit Pioneer. If the project is open to having them, that would work for me. I'll raise an issue and let you know.

[stale]

If you would like us to be able to process this issue, please provide the requested information. If the information is not provided within the next 3 weeks, we will be unable to proceed and this issue will be closed.

[marcphilipp]

@scordio Can we close this issue in favor of the Pioneer one?

[scordio]

Superseded by junit-pioneer/junit-pioneer#734.