Finish repairing 6.5's problems with r-tree indexes: create appropriate

selectivity functions and make the r-tree operators use them.  The
estimation functions themselves are just stubs, unfortunately, but
perhaps someday someone will make them compute realistic estimates.
Change pg_am so that the optimizer can reliably tell the difference
between ordered and unordered indexes --- before it would think that
an r-tree index can be scanned in '<<' order, which is not right AFAIK.
Repair broken negator links for network_sup and related ops.
Initdb forced.  This might be my last initdb force for 7.0 ... hope so
anyway ...

src/backend/utils/adt/geo_selfuncs.c

@@ -1,6 +1,6 @@
 /*-------------------------------------------------------------------------
  *
- * geo-selfuncs.c
+ * geo_selfuncs.c
  *	  Selectivity routines registered in the operator catalog in the
  *	  "oprrest" and "oprjoin" attributes.
  *
@@ -9,9 +9,10 @@
  *
  *
  * IDENTIFICATION
- *	  $Header: /cvsroot/pgsql/src/backend/utils/adt/geo_selfuncs.c,v 1.12 2000/01/26 05:57:14 momjian Exp $
+ *	  $Header: /cvsroot/pgsql/src/backend/utils/adt/geo_selfuncs.c,v 1.13 2000/02/17 03:39:42 tgl Exp $
  *
- *		XXX These are totally bogus.
+ *	XXX These are totally bogus.  Perhaps someone will make them do
+ *	something reasonable, someday.
  *
  *-------------------------------------------------------------------------
  */
@@ -19,17 +20,41 @@
 
 #include "utils/builtins.h"
 
+
+/*
+ *	Selectivity functions for rtrees.  These are bogus -- unless we know
+ *	the actual key distribution in the index, we can't make a good prediction
+ *	of the selectivity of these operators.
+ *
+ *	Note: the values used here may look unreasonably small.  Perhaps they
+ *	are.  For now, we want to make sure that the optimizer will make use
+ *	of an r-tree index if one is available, so the selectivity had better
+ *	be fairly small.
+ *
+ *	In general, rtrees need to search multiple subtrees in order to guarantee
+ *	that all occurrences of the same key have been found.  Because of this,
+ *	the estimated cost for scanning the index ought to be higher than the
+ *	output selectivity would indicate.  rtcostestimate(), over in selfuncs.c,
+ *	ought to be adjusted accordingly --- but until we can generate somewhat
+ *	realistic numbers here, it hardly matters...
+ */
+
+
+/*
+ * Selectivity for operators that depend on area, such as "overlap".
+ */
+
 float64
 areasel(Oid opid,
 		Oid relid,
 		AttrNumber attno,
-		char *value,
+		Datum value,
 		int32 flag)
 {
 	float64		result;
 
 	result = (float64) palloc(sizeof(float64data));
-	*result = 1.0 / 4.0;
+	*result = 0.05;
 	return result;
 }
 
@@ -43,81 +68,66 @@ areajoinsel(Oid opid,
 	float64		result;
 
 	result = (float64) palloc(sizeof(float64data));
-	*result = 1.0 / 4.0;
+	*result = 0.05;
 	return result;
 }
 
 /*
- *	Selectivity functions for rtrees.  These are bogus -- unless we know
- *	the actual key distribution in the index, we can't make a good prediction
- *	of the selectivity of these operators.
+ *	positionsel
  *
- *	In general, rtrees need to search multiple subtrees in order to guarantee
- *	that all occurrences of the same key have been found.  Because of this,
- *	the heuristic selectivity functions we return are higher than they would
- *	otherwise be.
+ * How likely is a box to be strictly left of (right of, above, below)
+ * a given box?
  */
 
-/*
- *	left_sel -- How likely is a box to be strictly left of (right of, above,
- *				below) a given box?
- */
-
-#ifdef NOT_USED
 float64
-leftsel(Oid opid,
-		Oid relid,
-		AttrNumber attno,
-		char *value,
-		int32 flag)
+positionsel(Oid opid,
+			Oid relid,
+			AttrNumber attno,
+			Datum value,
+			int32 flag)
 {
 	float64		result;
 
 	result = (float64) palloc(sizeof(float64data));
-	*result = 1.0 / 6.0;
+	*result = 0.1;
 	return result;
 }
 
-#endif
-
-#ifdef NOT_USED
 float64
-leftjoinsel(Oid opid,
-			Oid relid1,
-			AttrNumber attno1,
-			Oid relid2,
-			AttrNumber attno2)
+positionjoinsel(Oid opid,
+				Oid relid1,
+				AttrNumber attno1,
+				Oid relid2,
+				AttrNumber attno2)
 {
 	float64		result;
 
 	result = (float64) palloc(sizeof(float64data));
-	*result = 1.0 / 6.0;
+	*result = 0.1;
 	return result;
 }
 
-#endif
-
 /*
  *	contsel -- How likely is a box to contain (be contained by) a given box?
+ *
+ * This is a tighter constraint than "overlap", so produce a smaller
+ * estimate than areasel does.
  */
-#ifdef NOT_USED
+
 float64
 contsel(Oid opid,
 		Oid relid,
 		AttrNumber attno,
-		char *value,
+		Datum value,
 		int32 flag)
 {
 	float64		result;
 
 	result = (float64) palloc(sizeof(float64data));
-	*result = 1.0 / 10.0;
+	*result = 0.01;
 	return result;
 }
 
-#endif
-
-#ifdef NOT_USED
 float64
 contjoinsel(Oid opid,
 			Oid relid1,
@@ -128,8 +138,6 @@ contjoinsel(Oid opid,
 	float64		result;
 
 	result = (float64) palloc(sizeof(float64data));
-	*result = 1.0 / 10.0;
+	*result = 0.01;
 	return result;
 }
-
-#endif