Assertj-core: Multidimensional arrays assertion

[radistao]

Summary

Pursuing SO question AssertJ multidimensional arrays assertion (deep equals)
i'd like to try to make an assertion for multidimensional arrays.

I think i'd implement it similar to native Java Arrays#deepEquals(), just "extending" it with nicer assertion exceptions, including indexes and actual/expected values.

Example

int[][] expected = {
    new int[]{1, 0, 1},
    new int[]{0, 0, 0},
    new int[]{1, 1, 1}
};

int[][] actual = gameOfLife.play(board);

assertThat(actual).isDeepEqualTo(expected);

[joel-costigliola]

That would be a nice addition, bear in mind that it would need to be added for all primitive types and Object which is some substantial work.

[mwajcht]

Two years and nothing happened to this issue, so I took it and results can be seen in this quite a big pull request.

[joel-costigliola]

@mwajcht let's start with one of the array type (int for example) and then when this is done, we can replicate it to other types

[mwajcht]

Ok, I created the second, much smaller pull request. It contains only changes for int two-dimensional arrays. It should be easier to make the code review.

[joel-costigliola]

thanks, that will make it much easier!

[mwajcht]

@joel-costigliola As assertions for two-dimensional int arrays are already merged, I can apply your comments to other types. Do you prefer separate pull requests for each type or all-in-one?

[joel-costigliola]

@mwajcht I would prefer separate PRs, I might have missed things to improve (hopefully mot many) and seeing more than 100 files ina PR just sucks the energy out of me ;-)

[mwajcht]

I added a few pull requests with assertions for other types.

[joel-costigliola]

@mwajcht waht would you think of renaming hasSize to hasDimensions, the latter feels more natural and shows we are checking for multiple dimensions whereas the singular size?

BTW, I have renamed the master branch to main, your PR have been updated accordingly in this repo but not yours, don't be surprised.

[mwajcht]

@joel-costigliola Your suggestion regarding change hasSize->hasDimensions sounds reasonable. I created a new pull request which changes the existing methods (for int[][] and char[][]) to avoid changing them in each of the remaining pull requests. Once it's merged, I will update the rest of pull requests.

[joel-costigliola]

awesome thank you so much @mwajcht

[joel-costigliola]

@mwajcht for consistency, we now need to change is hasSameSizeAs to hasSameDimensionsAs, I can take care of it if you are busy.

[mwajcht]

@joel-costigliola Ok, please change hasSameSizeAs to hasSameDimensionsAs. Once it's done, I will rebase and change the rest of waiting pull requests.

[joel-costigliola]

@mwajcht I renamed hasSameSizeAs to hasSameDimensionsAs and made quite a lot of changes too:

• Fix isEmpty andisNotEmpty that should check there elements absence/presence, ex{{}, {}} is empty as it does no have any elements.
• Use a better error message for hasDimensions assertion.
• Improve SubarraysShouldHaveSameSize error message, remove duplication in it (at the cost of being less strongly typed)
• Remove internal comparison strategy that was applied to rows but the underlying comparator was comparing 2D arrays
• Refactor tests to check business logic in Arrays2D and verify Char2DArrays and Int2DArrays delegate to it
• Rename MultidimensionalArrayShouldBeEqual to Array2dElementShouldBeDeepEqual and simplify the implementation

[mwajcht]

I updated the pull requests for long[][], short[][] and byte[][] to contain all the changes introduced above. In the upcoming days I will also update other pull requests.

[mwajcht]

Now all pull requests should be up-to-date.

[joel-costigliola]

Finally done, thanks for such huge and good work @mwajcht!