Allow ApproximatelyMatcher to compare arrays

[ciaranmcnulty]

As suggested in #1367

My only concern is that the Differ doesn't know the precision the Matcher used, so in some case the verbose diff could be misleading

[TiMESPLiNTER]

Regarding your concern: that's definitely not ideal, but the current message you get if you compare an array of floats with shouldBe() is misleading anyway. It says:

expected [array:11], but got [array:11].

But if you see the use of shouldBeApproximately() in a test case you're immediately kind of aware the you're dealing with some sort of precision problem, right?

[ciaranmcnulty]

@TiMESPLiNTER If you pass the -v option you will see a diff

[TiMESPLiNTER]

@TiMESPLiNTER If you pass the -v option you will see a diff

True, but not if you compare an array with floats with shouldBe(). Then I just got the message mentioned above without a diff. Maybe a bug in the array diff?

See this example:

$ ./vendor/bin/phpspec run -vvv 


Timesplinter/Math                                                    
  17  - it calculates numbers
      expected [array:11], but got [array:11].
      
      
      
        21         $result = $this::calculate($values);
        22 
        23         $result->shouldBe([2.5, 3.5, 4.5, 5.5, 6.5, 7.5, 8.5, 9.5, 10.5, 11.5, 12.5]);
        24 
        25     }
        26 
      
       0 vendor/phpspec/phpspec/src/PhpSpec/Matcher/IdentityMatcher.php:59
         throw new PhpSpec\Exception\Example\NotEqualException("Expected [array:11...")

[ciaranmcnulty]

Odd, maybe open a new issue with a reproducer

[drupol]

@TiMESPLiNTER What's the status of this?

[TiMESPLiNTER]

Somehow lost track of it and never ended up having this problem again because I never dealt with floating arrays since then. I guess the problem still exists though.