Details
-
Improvement
-
Status: Resolved
-
Major
-
Resolution: Fixed
-
None
Description
According to the current implementation, the default "next power of two" rounding policy is assumed when allocating buffers for a vector.
In particular, for fixed width vectors, this policy is applied for the validity and data buffers, and for variable width vectors, this policy is applied for the validity and offset buffers.
However, this default policy is not always used for the allocator. When an alternative policy is in use, the buffers allocated assuming the default policy will have inappropriate capacities, which may lead to waste of memory spaces.
Attachments
Issue Links
- links to