Fix up pgstats counting of live and dead tuples to recognize that committed

and aborted transactions have different effects; also teach it not to assume
that prepared transactions are always committed.

Along the way, simplify the pgstats API by tying counting directly to
Relations; I cannot detect any redeeming social value in having stats
pointers in HeapScanDesc and IndexScanDesc structures.  And fix a few
corner cases in which counts might be missed because the relation's
pgstat_info pointer hadn't been set.

src/include/utils/rel.h

@@ -7,7 +7,7 @@
  * Portions Copyright (c) 1996-2007, PostgreSQL Global Development Group
  * Portions Copyright (c) 1994, Regents of the University of California
  *
- * $PostgreSQL: pgsql/src/include/utils/rel.h,v 1.100 2007/03/29 00:15:39 tgl Exp $
+ * $PostgreSQL: pgsql/src/include/utils/rel.h,v 1.101 2007/05/27 03:50:39 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -89,15 +89,6 @@ typedef struct TriggerDesc
 } TriggerDesc;
 
 
-/*
- * Same for the statistics collector data in Relation and scan data.
- */
-typedef struct PgStat_Info
-{
-	void	   *tabentry;
-} PgStat_Info;
-
-
 /*
  * Cached lookup information for the index access method functions defined
  * by the pg_am row associated with an index relation.
@@ -200,8 +191,8 @@ typedef struct RelationData
 	List	   *rd_indpred;		/* index predicate tree, if any */
 	void	   *rd_amcache;		/* available for use by index AM */
 
-	/* statistics collection area */
-	PgStat_Info pgstat_info;
+	/* use "struct" here to avoid needing to include pgstat.h: */
+	struct PgStat_TableStatus *pgstat_info;	/* statistics collection area */
 } RelationData;
 
 typedef RelationData *Relation;