While reviewing the parallel CREATE INDEX patch, I noticed that commit 8b304b8b omitted to remove a comment that it made obsolete.
Attached patch removes the comment. -- Peter Geoghegan
From a52ca388cb21f797bcfe3a1b15cc9168e5d26f41 Mon Sep 17 00:00:00 2001 From: Peter Geoghegan <p...@bowt.ie> Date: Fri, 8 Dec 2017 15:32:55 -0800 Subject: [PATCH] Remove obsolete comment about single run case. It is no longer possible to reach the "merge runs" stage of an external tuplesort and find that there is only a single run to merge (i.e., that no merge is actually required). This was missed by commit 8b304b8b, which removed replacement selection sort. --- src/backend/utils/sort/tuplesort.c | 5 ----- 1 file changed, 5 deletions(-) diff --git a/src/backend/utils/sort/tuplesort.c b/src/backend/utils/sort/tuplesort.c index 3c23ac7..35eebad 100644 --- a/src/backend/utils/sort/tuplesort.c +++ b/src/backend/utils/sort/tuplesort.c @@ -2459,11 +2459,6 @@ mergeruns(Tuplesortstate *state) * Use all the remaining memory we have available for read buffers among * the input tapes. * - * We do this only after checking for the case that we produced only one - * initial run, because there is no need to use a large read buffer when - * we're reading from a single tape. With one tape, the I/O pattern will - * be the same regardless of the buffer size. - * * We don't try to "rebalance" the memory among tapes, when we start a new * merge phase, even if some tapes are inactive in the new phase. That * would be hard, because logtape.c doesn't know where one run ends and -- 2.7.4
