Repeatead chars not working as expected about jqwik

[jlink]

New features available in 1.8.0-SNAPSHOT

[jlink]

@esfomeado Thanks for raising that issue.
I see why you‘d expect that repeatChars(0) to not create any duplicate chars at all. Currently it just sets the probability of explicitly adding duplicate chars to 0. The effect you see is due to edge case resolution.

I’ll have to think about if changing the behaviour or documentation of repeatChars is the right thing to do. Maybe an additional uniqueChars() would be even clearer.

To solve your problem short-term, you can use withoutEdgeCases() on the strings arbitrary:

Arbitraries.strings().alpha()
		   .ofMinLength(5).ofMaxLength(25)
		   .repeatChars(0.0)
		   .withoutEdgeCases();

This will prevent „AAAAA“ to be generated on purpose. Duplicate chars will still occur by chance though.

If you definitely want no duplicates, you’d currently have to take a detour through something like:

Arbitraries.chars().alpha().list()
		   .ofMinSize(5).ofMaxSize(25)
		   .uniqueElements()
		   .map(chars -> chars.stream().map(String::valueOf).collect(Collectors.joining()));

[jlink]

I decided to introduce an experimental new API method: StringArbitrary.uniqueChars()

The open question is if StringArbitrary.repeatChars(0.0) should log a warning or just call uniqueChars implicitly, or both?

@esfomeado What do you think?

[esfomeado]

Thanks for the quick response.

@jlink Personally I would expect to have the same behaviour was uniqueChars when set to 0 so I would say to just call uniqueChars implicitly.

[jlink]

The StringArbitrary.uniqueChars() functionality and the fix for repeatChars(0.0) is available in 1.8.0-SNAPSHOT.

[jlink]

In addition, there's now an annotation @UniqueChars that can be applied to any String parameter in a property.