Add support for checking subsets that contain duplicate values to ArraySubset

[pfrenssen]

In #3161 support was added for checking indexed arrays in ArraySubset. The edge case of checking subsets that contain multiple identical values was left out of the scope of that PR. This means that in the current implementation the following example will generate a false positive:

• Array: [0, 1, 2, 3]
• Subset: [0, 1, 1, 2]
• Expected result: FALSE since the value 1 is only present once in the array, but twice in the subset.
• Actual result: TRUE

This PR adds support for this edge case, so that it will become possible to check subsets with duplicate values.

[codecov-io]

Codecov Report

Merging #3186 into master will increase coverage by <.01%.
The diff coverage is 100%.

@@             Coverage Diff              @@
##             master    #3186      +/-   ##
============================================
+ Coverage     82.37%   82.37%   +<.01%     
- Complexity     3508     3512       +4     
============================================
  Files           140      140              
  Lines          9243     9245       +2     
============================================
+ Hits           7614     7616       +2     
  Misses         1629     1629

Impacted Files	Coverage Δ	Complexity Δ
src/Framework/Constraint/ArraySubset.php	98.63% <100%> (+0.03%)	37 <0> (+4)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 995a578...c0c94e9. Read the comment docs.

[KristopherWindsor]

We have a few tests that started failing when phpunit 7.3 was released.
Not good to have a minor phpunit version break tests. 😭 Anyway, we think this PR will make our tests pass again.

07:16:26 There were 4 failures:
07:16:26 
07:16:26 1) CompanyTest\Integration\Lib\Model\Test\ClassNameTest::testHandleFollowUpCacheKeyCallSettingOffAwaySettingOnNoSentReply
07:16:26 Failed asserting that an array has the subset Array &0 (
07:16:26     0 => '30'
07:16:26     1 => '14097220017'
07:16:26     2 => 'A'
07:16:26     3 => 'F'
07:16:26     4 => 'F'
07:16:26     5 => 'N'
07:16:26     6 => 'P'
07:16:26     7 => 'SMS'
07:16:26     8 => 'ProductName'
07:16:26 ).
07:16:26 --- Expected
07:16:26 +++ Actual
07:16:26 @@ @@
07:16:26  Array
07:16:26  (
07:16:26 -    [0] => 30
07:16:26 -    [1] => 14097220017
07:16:26 +    [1] => 30
07:16:26      [2] => A
07:16:26 -    [3] => F
07:16:26 -    [4] => F
07:16:26 -    [5] => N
07:16:26 -    [6] => P
07:16:26 -    [7] => SMS
07:16:26 -    [8] => ProductName
07:16:26 +    [3] => 14097220007
07:16:26 +    [4] => P
07:16:26 +    [5] => F
07:16:26 +    [6] => ProductName
07:16:26 +    [7] => F
07:16:26 +    [8] => N
07:16:26 +    [9] => SMS
07:16:26  )

[pfrenssen]

@KristopherWindsor did you try it with this PR? I see you have duplicate values in your subset. That's what this PR is fixing. If it is not, could you provide a test case so I can have a look?

[pfrenssen]

In #3240 a regression was reported against assertArraySubset(), but this problem no longer occurs in this PR. I have included the test case for that issue to prove that this is fixed here.

[RikSomers]

@pfrenssen We use the Laravel framework here at work, and serveral tests stopped working when upgrading to PHPUnit 7.3.0 due to the ArraySubset change.

I can confirm this PR fixes any failing tests that were introduced with the 7.3.0 changes in our companies codebase.

[sebastianbergmann]

I have reverted #3161 because of #3240.

@pfrenssen Please send a new pull request against master that contains both #3161 and #3186. I am sorry for having to ask this of you -- and that you have to wait until PHPUnit 7.4 to get these improvements into a stable release of PHPUnit.

[kmbt]

Having some background in mathematics and also in python programming language I find this way of thinking really confusing.

In mathematical set theory, an object just may be or may not be a member of a set. In set theory there is no concept of an object being a member "more than once" of a set.

This reasoning is applied in python, where the following expression evaluates as True:

set([[0, 1, 1, 2]) == set([0, 1, 2])

Please kindly take my remarks into your consideration.

[Stadly]

@pfrenssen Did you have a chance to create a new PR containing both #3161 and #3186? Would you like me to look into it?

[pfrenssen]

@Stadly I have this on my to-do list which is admittedly very long :)

I don't think I will get round to this very soon, so feel free to open the PR if you want. I would be glad if you could keep my git history so I can get credit for the work, but if this is not practical for you then don't worry.

Would be nice to see this merged indeed!