diff options
Diffstat (limited to 'src/backend/access/nbtree/nbtsearch.c')
| -rw-r--r-- | src/backend/access/nbtree/nbtsearch.c | 43 |
1 files changed, 20 insertions, 23 deletions
diff --git a/src/backend/access/nbtree/nbtsearch.c b/src/backend/access/nbtree/nbtsearch.c index 45f4f7604d0..a78acd0f04b 100644 --- a/src/backend/access/nbtree/nbtsearch.c +++ b/src/backend/access/nbtree/nbtsearch.c @@ -7,7 +7,7 @@ * * * IDENTIFICATION - * $Header: /cvsroot/pgsql/src/backend/access/nbtree/nbtsearch.c,v 1.20 1997/05/30 18:35:37 vadim Exp $ + * $Header: /cvsroot/pgsql/src/backend/access/nbtree/nbtsearch.c,v 1.21 1997/06/10 07:28:50 vadim Exp $ * *------------------------------------------------------------------------- */ @@ -159,6 +159,7 @@ _bt_moveright(Relation rel, BTPageOpaque opaque; ItemId hikey; BlockNumber rblkno; + int natts = rel->rd_rel->relnatts; page = BufferGetPage(buf); opaque = (BTPageOpaque) PageGetSpecialPointer(page); @@ -195,26 +196,9 @@ _bt_moveright(Relation rel, * on this page to do not lose "good" tuples if number * of attrs > keysize. Example: (2,0) - last items on * this page, (2,1) - first item on next page (hikey), - * our scankey is x = 2. Scankey >= (2,1) because of + * our scankey is x = 2. Scankey == (2,1) because of * we compare first attrs only, but we shouldn't to move * right of here. - vadim 04/15/97 - * - * XXX - * This code changed again! Actually, we break our - * duplicates handling in single case: if we insert - * new minimum key into leftmost page with duplicates - * and splitting doesn't occure then _bt_insertonpg doesn't - * worry about duplicates-rule. Fix _bt_insertonpg ? - * But I don't see why don't compare scankey with _last_ - * item on the page instead of first one, in any cases. - * So - we do it in that way now. - vadim 05/26/97 - * - * Also, if we are on an "pseudo-empty" leaf page (i.e. there is - * only hikey here) and scankey == hikey then we don't move - * right! It's fix for bug described in _bt_insertonpg(). It's - * right - at least till index cleanups are perfomed by vacuum - * in exclusive mode: so, though this page may be just splitted, - * it may not be "emptied" before we got here. - vadim 05/27/97 */ if ( _bt_skeycmp (rel, keysz, scankey, page, hikey, @@ -227,14 +211,27 @@ _bt_moveright(Relation rel, } if ( offmax > P_HIKEY ) { - if ( _bt_skeycmp (rel, keysz, scankey, page, + if ( natts == keysz ) /* sanity checks */ + { + if ( _bt_skeycmp (rel, keysz, scankey, page, + PageGetItemId (page, P_FIRSTKEY), + BTEqualStrategyNumber) ) + elog (FATAL, "btree: BTP_CHAIN flag was expected"); + if ( _bt_skeycmp (rel, keysz, scankey, page, + PageGetItemId (page, offmax), + BTEqualStrategyNumber) ) + elog (FATAL, "btree: unexpected equal last item"); + if ( _bt_skeycmp (rel, keysz, scankey, page, + PageGetItemId (page, offmax), + BTLessStrategyNumber) ) + elog (FATAL, "btree: unexpected greater last item"); + /* move right */ + } + else if ( _bt_skeycmp (rel, keysz, scankey, page, PageGetItemId (page, offmax), BTLessEqualStrategyNumber) ) break; } - else if ( offmax == P_HIKEY && - ( opaque->btpo_flags & BTP_LEAF ) ) - break; } /* step right one page */ |