tustvold commented on PR #4522: URL: https://github.com/apache/arrow-datafusion/pull/4522#issuecomment-1356224225
> I had a suggestion about the API design (MemoryPool::register()) rather than creating TrackedAllocations directly. > I think having the MemoryPool create TrackedAllocations makes more logical sense as they are so tightly bound and you need a MemoryPool for a tracked allocation anyways I originally had this prior to https://github.com/apache/arrow-datafusion/pull/4522/commits/718a94b72a0b65486d26ceb482e9b3393ee3dd5e, however, it results in passing round this `MemoryManager` struct that has no state of its own which seemed a bit pointless. In order for MemoryPool to be able to construct a `TrackedAllocation` the construction methods on `TrackedAllocation` need to be public regardless, so that they can be used downstream, it seemed simpler just to not bother? -- 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]
