Testing Problem
Unfortunately what is considered as whitespace character (Character.isWhitespace(...)) has changed between Java 8 (OpenJDK) and Java 11 (OpenJDK):
In particular the U+180E (Mongolian Vovel Separator) counts in Java 8 as whitespace, while it does not in Java 11.
This causes problems because when generating whitespace chars with jqwik (which contains hard-coded lists of whitespace chars based on Java 8) and actually running the functionality on Java 11.
Suggested Solution
Instead of using a prepared, hard-coded list of whitespace chars, the chars considered as whitespace by the actual runtime should be used.
Please have a look at my corresponding pull request.
Discussion
To minimize the runtime overhead, the determination of the whitespace chars could (should) be done at class-level instead of instance-level.
Testing Problem
Unfortunately what is considered as whitespace character (
Character.isWhitespace(...)) has changed between Java 8 (OpenJDK) and Java 11 (OpenJDK):In particular the U+180E (Mongolian Vovel Separator) counts in Java 8 as whitespace, while it does not in Java 11.
This causes problems because when generating whitespace chars with jqwik (which contains hard-coded lists of whitespace chars based on Java 8) and actually running the functionality on Java 11.
Suggested Solution
Instead of using a prepared, hard-coded list of whitespace chars, the chars considered as whitespace by the actual runtime should be used.
Please have a look at my corresponding pull request.
Discussion
To minimize the runtime overhead, the determination of the whitespace chars could (should) be done at class-level instead of instance-level.