Commit112b006

committed

logical decoding: Remove TODO about unnecessary optimization.

Measurements show, and intuition agrees, that there's currently noknown cases where adding a fastpath to avoid allocating / ordering aheap for a single transaction is worthwhile.Author: Dilip KumarDiscussion:https://postgr.es/m/CAFiTN-sp701wvzvnLQJGk7JDqrFM8f--97-ihbwkU8qvn=p8nw@mail.gmail.com

1 parentf15ace7 commit112b006Copy full SHA for 112b006

File tree

-5

lines changed

-5

lines changed

Lines changed: 0 additions & 5 deletions

Original file line number	Diff line number	Diff line change
`@@ -1037,11 +1037,6 @@ ReorderBufferIterTXNInit(ReorderBuffer rb, ReorderBufferTXN txn,`
`1037`	`1037`	`nr_txns++;`
`1038`	`1038`	`}`
`1039`	`1039`
`1040`		`-/*`
`1041`		`- * TODO: Consider adding fastpath for the rather common nr_txns=1 case, no`
`1042`		`- * need to allocate/build a heap then.`
`1043`		`- */`
`1044`		`-`
`1045`	`1040`	`/* allocate iteration state */`
`1046`	`1041`	`state= (ReorderBufferIterTXNState*)`
`1047`	`1042`	`MemoryContextAllocZero(rb->context,`

Comments

(0)