jduo commented on PR #38876: URL: https://github.com/apache/arrow/pull/38876#issuecomment-1828277417
> > This failure is happening for somewhat complicated reasons: https://github.com/apache/arrow/actions/runs/6983421948/job/19004513065?pr=38876#step:6:2511 > > Before refactoring, the allocator used by this test would be the Netty module's DefaultAllocationManagerFactory since TestBaseAllocator lived in the Netty module. This one correctly returns a non-zero value when given a zero-allocation. > > After refactoring, TestBaseAllocator lives in memory-core, and uses the dummy DefaultAllocationManagerFactory in memory-core's tests. This one however does return zero when given a zero-allocation. This might itself be a bug unrelated to the refactoring. This calls MemoryUtil.UNSAFE.allocateMemory(), which calls sun.misc.Unsafe.allocateMemory(), which is supposed to return a non-zero value. > > I'm going to @ignore this test for now. > > Should these allocation manager factorys return a non-zero address when the user asks for an empty allocation? I would assume so since C-allocators have this property. > > @davisusanibar , did you see anything like this? @lidavidm @danepitkin , any thoughts around this? > > Thanks for pushing this feature, James. I am observing changes. Does this problem persist? Yes this still happens. I have marked the test as ignored to get past it. It's not clear if this is a new bug or if it is exposing a bug from the memory-core test version of DefaultAllocationManagerFactory. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
