database/postgres
database/postgres
1
0
Fork 0
mirror of https://github.com/postgres/postgres.git synced 2025-05-02 11:44:50 +03:00
Code

Fix outdated comments, GIST search queue is not an RBTree anymore.

Browse Source 
The GiST search queue is implemented as a pairing heap rather than as
Red-Black Tree, since 9.5 (commit e7032610). I neglected these comments
in that commit.
This commit is contained in:
Heikki Linnakangas 2016-09-20 11:38:25 +03:00
parent 40c3fe4980
commit 45310221a9
2 changed files with 9 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
src/backend/access/gist/gistscan.c
View File

@ -125,7 +125,7 @@ gistrescan(IndexScanDesc scan, ScanKey key, int nkeys,
* which is created on the second call and reset on later calls. Thus, in * which is created on the second call and reset on later calls. Thus, in
* the common case where a scan is only rescan'd once, we just put the * the common case where a scan is only rescan'd once, we just put the
* queue in scanCxt and don't pay the overhead of making a second memory * queue in scanCxt and don't pay the overhead of making a second memory
* context. If we do rescan more than once, the first RBTree is just left * context. If we do rescan more than once, the first queue is just left
* for dead until end of scan; this small wastage seems worth the savings * for dead until end of scan; this small wastage seems worth the savings
* in the common case. * in the common case.
*/ */
@ -181,7 +181,7 @@ gistrescan(IndexScanDesc scan, ScanKey key, int nkeys,
ALLOCSET_DEFAULT_SIZES); ALLOCSET_DEFAULT_SIZES);
} }
/* create new, empty RBTree for search queue */ /* create new, empty pairing heap for search queue */
oldCxt = MemoryContextSwitchTo(so->queueCxt); oldCxt = MemoryContextSwitchTo(so->queueCxt);
so->queue = pairingheap_allocate(pairingheap_GISTSearchItem_cmp, scan); so->queue = pairingheap_allocate(pairingheap_GISTSearchItem_cmp, scan);
MemoryContextSwitchTo(oldCxt); MemoryContextSwitchTo(oldCxt);

18
src/include/access/gist_private.h
View File

@ -107,15 +107,11 @@ typedef struct GISTSTATE
* upper index pages; this rule avoids doing extra work during a search that * upper index pages; this rule avoids doing extra work during a search that
* ends early due to LIMIT. * ends early due to LIMIT.
* *
* To perform an ordered search, we use an RBTree to manage the distance-order * To perform an ordered search, we use a pairing heap to manage the
* queue. Each GISTSearchTreeItem stores all unvisited items of the same * distance-order queue. In a non-ordered search (no order-by operators),
* distance; they are GISTSearchItems chained together via their next fields. * we use it to return heap tuples before unvisited index pages, to
* * ensure depth-first order, but all entries are otherwise considered
* In a non-ordered search (no order-by operators), the RBTree degenerates * equal.
* to a single item, which we use as a queue of unvisited index pages only.
* In this case matched heap items from the current index leaf page are
* remembered in GISTScanOpaqueData.pageData[] and returned directly from
* there, instead of building a separate GISTSearchItem for each one.
*/ */
/* Individual heap tuple to be visited */ /* Individual heap tuple to be visited */
@ -298,8 +294,8 @@ typedef struct
#define GIST_ROOT_BLKNO 0 #define GIST_ROOT_BLKNO 0
/* /*
* Before PostgreSQL 9.1, we used rely on so-called "invalid tuples" on inner * Before PostgreSQL 9.1, we used to rely on so-called "invalid tuples" on
* pages to finish crash recovery of incomplete page splits. If a crash * inner pages to finish crash recovery of incomplete page splits. If a crash
* happened in the middle of a page split, so that the downlink pointers were * happened in the middle of a page split, so that the downlink pointers were
* not yet inserted, crash recovery inserted a special downlink pointer. The * not yet inserted, crash recovery inserted a special downlink pointer. The
* semantics of an invalid tuple was that it if you encounter one in a scan, * semantics of an invalid tuple was that it if you encounter one in a scan,