Duplicate keys from different data providers are not handled properly

[kubawerlos]

No description provided.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 90.28%. Comparing base (c726f0d) to head (70dd63a).

Additional details and impacted files

@@             Coverage Diff              @@
##               10.5    #5819      +/-   ##
============================================
- Coverage     90.28%   90.28%   -0.01%     
+ Complexity     6441     6439       -2     
============================================
  Files           679      679              
  Lines         19545    19540       -5     
============================================
- Hits          17647    17642       -5     
  Misses         1898     1898              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[sebastianbergmann]

I am sorry to ask, but from looking at the patch it is not yet clear to me what issue this PR aims to solve. Can you enlighten me? Thanks!

[kubawerlos]

@sebastianbergmann currently if you run DuplicateKeyDataProvidersTest the result will be:

OK (1 test, 1 assertion)

(you can see the result of DataProvider::providedData in the 1st commit)
when:

• there are 2 tests
• what's more important: dataProvider1 provides a test case that should fail, but unfortunately it was overwritten by a case from dataProvider2, due to key duplication.

Also, the changes in DataProvider, as they are technically a bug fix, they are more simplification (it is better visible as of https://github.com/sebastianbergmann/phpunit/pull/5819/files?w=1) - it is getting rid of the additional $origData variable and calling array_merge.