added the failing test for issue #308

[jurgenvinju]

• add failing test for Failing test for set difference #308
• probably add failing test for usethesource/capsule as well
• fix ISet.delete in the presence of hashcode collisions

[github-actions]

Test Results

     72 files   -      27       72 suites   - 27   3m 12s ⏱️ - 2m 34s
244 537 tests +      2  244 535 ✅ +      1  1 💤 ±0  1 ❌ +1 
489 272 runs   - 244 333  489 267 ✅  - 244 335  2 💤  - 1  3 ❌ +3 

For more details on these failures, see this check.

Results for commit 449f449. ± Comparison against base commit b12b40b.

[jurgenvinju]

@msteindorfer hi! we could use some help here. There seems to be an issue with PersistentHashIndexedBinaryRelation.delete of capsule.

• It's internal SetMultimap.Immutable<IValue, IValue> object seems to be confused in the presence of key hash collisions.
• Note that there are two unequal keys both with hash code 0 in the failing example.
• After one of them is removed with PersistentHashIndexedBinaryRelation.delete, the SetMultiMap is left in a uncanonical state. Namely the elements that are left in there are stored in a different order from the value when created from scratch using insert
• This makes equals fail which is how @PaulKlint discovered the issue.
• Added a test that exhibits this behavior to vallang on this PR branch in io.usethesource.vallang.issues.RegressionTests. It's the final test named setDifferenceInThePresenceOfHashCollisions