chore: add ForceUpdate to BPTree #3993
Conversation
Signed-off-by: kostas <kostas@dragonflydb.io>
kostasrim
changed the title
kostasrim
changed the title
| score_tree->Delete(old_obj); | ||
| score_tree->Insert(new_obj); | ||
| }; | ||
| auto cb = [this](sds old_obj, sds new_obj) { score_tree->ForceUpdate(old_obj, new_obj); }; |
There was a problem hiding this comment.
when I checked if old_obj is still alive when thsi callback is called, I realised that ReallocIfNeededGeneric is really a bad name because it does not REALLOCATE. The better name would be DuplicateEntryIfFragmented
|
IMO It's not a bad idea to add a unit test so I will add one 😄 |
| @@ -585,4 +589,15 @@ template <typename T, typename Policy> void BPTree<T, Policy>::DestroyNode(BPTre | |||
| num_nodes_--; | |||
| } | |||
|
|
|||
| template <typename T, typename Policy> void BPTree<T, Policy>::ForceUpdate(KeyT old, KeyT new_obj) { | |||
There was a problem hiding this comment.
Maybe worth asserting that Policy::operator(old, new_obj) returns 0
ReallocIfNeededdeleted and reinserted the same key in theBPTreestored inSortedMap. This is not really needed since the key is actually the same and we can just update the pointer withinBPTreeinstead of doing the deletion + insertion chore.Resolves #3963