summaryrefslogtreecommitdiff
path: root/src/backend/access/rtree/rtree.c
diff options
context:
space:
mode:
authorBruce Momjian <bruce@momjian.us>2005-11-22 18:23:31 +0000
committerBruce Momjian <bruce@momjian.us>2005-11-22 18:23:31 +0000
commitbef7764835725e5d8468da1c139e9020be689b95 (patch)
tree71075b16ab6ed5152b31757e5dd65cd2b9383ba0 /src/backend/access/rtree/rtree.c
parentc8de36352fe72ae2265eb53a6e1bf334e4f24888 (diff)
Re-run pgindent, fixing a problem where comment lines after a blank
comment line where output as too long, and update typedefs for /lib directory. Also fix case where identifiers were used as variable names in the backend, but as typedefs in ecpg (favor the backend for indenting). Backpatch to 8.1.X.
Diffstat (limited to 'src/backend/access/rtree/rtree.c')
-rw-r--r--src/backend/access/rtree/rtree.c32
1 files changed, 16 insertions, 16 deletions
diff --git a/src/backend/access/rtree/rtree.c b/src/backend/access/rtree/rtree.c
index d684101d261..652daa38203 100644
--- a/src/backend/access/rtree/rtree.c
+++ b/src/backend/access/rtree/rtree.c
@@ -8,7 +8,7 @@
*
*
* IDENTIFICATION
- * $PostgreSQL: pgsql/src/backend/access/rtree/rtree.c,v 1.92 2005/10/15 02:49:09 momjian Exp $
+ * $PostgreSQL: pgsql/src/backend/access/rtree/rtree.c,v 1.92.2.1 2005/11/22 18:23:04 momjian Exp $
*
*-------------------------------------------------------------------------
*/
@@ -527,9 +527,9 @@ rtdosplit(Relation r,
* the left page, we expect it to get smaller. This happens in the
* internal insertion routine.
*
- * 3) Insert a pointer to the right page in the parent. This may cause the
- * parent to split. If it does, we need to repeat steps one and two for
- * each split node in the tree.
+ * 3) Insert a pointer to the right page in the parent. This may cause
+ * the parent to split. If it does, we need to repeat steps one and two
+ * for each split node in the tree.
*/
/* adjust active scans */
@@ -834,8 +834,8 @@ rtpicksplit(Relation r,
/*
* Now split up the regions between the two seeds.
*
- * The cost_vector array will contain hints for determining where each tuple
- * should go. Each record in the array will contain a boolean,
+ * The cost_vector array will contain hints for determining where each
+ * tuple should go. Each record in the array will contain a boolean,
* choose_left, that indicates which node the tuple prefers to be on, and
* the absolute difference in cost between putting the tuple in its
* favored node and in the other node.
@@ -848,9 +848,9 @@ rtpicksplit(Relation r,
* First, build the cost_vector array. The new index tuple will also be
* handled in this loop, and represented in the array, with i==newitemoff.
*
- * In the case of variable size tuples it is possible that we only have the
- * two seeds and no other tuples, in which case we don't do any of this
- * cost_vector stuff.
+ * In the case of variable size tuples it is possible that we only have
+ * the two seeds and no other tuples, in which case we don't do any of
+ * this cost_vector stuff.
*/
/* to keep compiler quiet */
@@ -966,11 +966,11 @@ rtpicksplit(Relation r,
* need not worry about any other problem than failing to fit the new
* item.)
*
- * Guttman's algorithm actually has two factors to consider (in order):
- * 1. if one node has so many tuples already assigned to it that the
- * other needs all the rest in order to satisfy the condition that
- * neither node has fewer than m tuples, then that is decisive; 2.
- * otherwise, choose the page that shows the smaller enlargement of
+ * Guttman's algorithm actually has two factors to consider (in
+ * order): 1. if one node has so many tuples already assigned to it
+ * that the other needs all the rest in order to satisfy the condition
+ * that neither node has fewer than m tuples, then that is decisive;
+ * 2. otherwise, choose the page that shows the smaller enlargement of
* its union area.
*
* I have chosen m = M/2, where M is the maximum number of tuples on a
@@ -979,8 +979,8 @@ rtpicksplit(Relation r,
* tuple on a page, if it is really big. But even with variable size
* tuples we still try to get m as close as possible to M/2.)
*
- * The question of which page shows the smaller enlargement of its union
- * area has already been answered, and the answer stored in the
+ * The question of which page shows the smaller enlargement of its
+ * union area has already been answered, and the answer stored in the
* choose_left field of the SPLITCOST record.
*/
left_feasible = (left_avail_space >= item_1_sz &&