On Tue, 15 Jul 2025 16:43:49 GMT, Alan Bateman <al...@openjdk.org> wrote:
>> Yes. The OS-dependent `mkdirs` sub-tests reflect the actual behavior. > > Okay, but if I read this correctly then these methods will test mkdirs with a > non-empty path as child will be "/child" or "\child". In order to test > mkdirs with an empty path then I think we need `assertFalse(f.mkdirs());` > [I]f I read this correctly then these methods will test mkdirs with a > non-empty path as child will be "/child" or "\child". Right. Yes, I agree this needs modification. Also, I don't trust the stability of the `mkdirs` sub-test on Windows as conceivably it could fail with no permission to create a folder `C:\folder`. ------------- PR Review Comment: https://git.openjdk.org/jdk/pull/26224#discussion_r2208057888
