Clean up various to-do items associated with system indexes:

pg_database now has unique indexes on oid and on datname.
pg_shadow now has unique indexes on usename and on usesysid.
pg_am now has unique index on oid.
pg_opclass now has unique index on oid.
pg_amproc now has unique index on amid+amopclaid+amprocnum.
Remove pg_rewrite's unnecessary index on oid, delete unused RULEOID syscache.
Remove index on pg_listener and associated syscache for performance reasons
(caching rows that are certain to change before you need 'em again is
rather pointless).
Change pg_attrdef's nonunique index on adrelid into a unique index on
adrelid+adnum.

Fix various incorrect settings of pg_class.relisshared, make that the
primary reference point for whether a relation is shared or not.
IsSharedSystemRelationName() is now only consulted to initialize relisshared
during initial creation of tables and indexes.  In theory we might now
support shared user relations, though it's not clear how one would get
entries for them into pg_class &etc of multiple databases.

Fix recently reported bug that pg_attribute rows created for an index all have
the same OID.  (Proof that non-unique OID doesn't matter unless it's
actually used to do lookups ;-))

There's no need to treat pg_trigger, pg_attrdef, pg_relcheck as bootstrap
relations.  Convert them into plain system catalogs without hardwired
entries in pg_class and friends.

Unify global.bki and template1.bki into a single init script postgres.bki,
since the alleged distinction between them was misleading and pointless.
Not to mention that it didn't work for setting up indexes on shared
system relations.

Rationalize locking of pg_shadow, pg_group, pg_attrdef (no need to use
AccessExclusiveLock where ExclusiveLock or even RowExclusiveLock will do).
Also, hold locks until transaction commit where necessary.

Author: tglsfdc

doc/src/sgml/bki.sgml

@@ -1,4 +1,6 @@
-<!-- $Header: /cvsroot/pgsql/doc/src/sgml/bki.sgml,v 1.6 2000/12/23 16:24:29 petere Exp $ -->
+<!--
+$Header: /cvsroot/pgsql/doc/src/sgml/bki.sgml,v 1.7 2001/06/12 05:55:48 tgl Exp $
+ -->
 
 <chapter id="bki">
  <title><acronym>BKI</acronym> Backend Interface</title>
@@ -15,16 +17,14 @@
  </para>
 
  <para>
-  <application>initdb</application> uses <acronym>BKI</acronym> files
+  <application>initdb</application> uses a <acronym>BKI</acronym> file
   to do part of its job when creating a new database cluster.  The
-  input files used by <application>initbd</application> are created as
+  input file used by <application>initdb</application> is created as
   part of building and installing <productname>Postgres</productname>
   by a program named <filename>genbki.sh</filename> from some
   specially formatted C header files in the source tree.  The created
-  BKI files are called <filename>global.bki</filename> (for global
-  catalogs) and <filename>template1.bki</filename> (for the catalogs
-  initially stored in the template1 database and then duplicated in
-  every created database) and are normally installed in the
+  BKI file is called <filename>postgres.bki</filename> and is
+  normally installed in the
   <filename>share</filename> subdirectory of the installation tree.
  </para>
 
@@ -39,7 +39,7 @@
   <para>
    This section describes how the <productname>Postgres</productname>
    backend interprets <acronym>BKI</acronym> files.  This description
-   will be easier to understand if the <filename>global.bki</filename>
+   will be easier to understand if the <filename>postgres.bki</filename>
    file is at hand as an example.  You should also study the source
    code of <application>initdb</application> to get an idea of how the
    backend is invoked.
@@ -49,7 +49,7 @@
    BKI input consists of a sequence of commands.  Commands are made up
    of a number of tokens, depending on the syntax of the command.
    Tokens are usually separated by whitespace, but need not be if
-   there is no ambiguity.  There is not special command separator; the
+   there is no ambiguity.  There is no special command separator; the
    next token that syntactically cannot belong to the preceeding
    command starts a new one.  (Usually you would put a new command on
    a new line, for clarity.)  Tokens can be certain key words, special
@@ -155,13 +155,13 @@
       values and <replaceable
       class="parameter">oid_value</replaceable> for its OID.  If
       <replaceable class="parameter">oid_value</replaceable> is zero
-      (0) or the clause is ommitted, then the next available OID is
+      (0) or the clause is omitted, then the next available OID is
       used.
      </para>
 
      <para>
       NULL values can be specified using the special key word
-      <literal>_null_</literal>.  Values containing spaces should be
+      <literal>_null_</literal>.  Values containing spaces must be
       double quoted.
      </para>
     </listitem>

doc/src/sgml/catalogs.sgml

@@ -1,6 +1,6 @@
 <!--
  Documentation of the system catalogs, directed toward PostgreSQL developers
- $Header: /cvsroot/pgsql/doc/src/sgml/catalogs.sgml,v 2.17 2001/05/14 20:30:19 momjian Exp $
+ $Header: /cvsroot/pgsql/doc/src/sgml/catalogs.sgml,v 2.18 2001/06/12 05:55:48 tgl Exp $
  -->
 
 <chapter id="catalogs">
@@ -645,7 +645,8 @@
       <entry>relisshared</entry>
       <entry><type>bool</type></entry>
       <entry></entry>
-      <entry>XXX (This is not what it seems to be.)</entry>
+      <entry>True if this table is shared across all databases in the
+      cluster.</entry>
      </row>
 
      <row>

src/backend/access/common/heaptuple.c

@@ -9,7 +9,7 @@
  *
  *
  * IDENTIFICATION
- *	  $Header: /cvsroot/pgsql/src/backend/access/common/heaptuple.c,v 1.71 2001/03/22 06:16:06 momjian Exp $
+ *	  $Header: /cvsroot/pgsql/src/backend/access/common/heaptuple.c,v 1.72 2001/06/12 05:55:49 tgl Exp $
  *
  * NOTES
  *	  The old interface functions have been converted to macros
@@ -738,44 +738,46 @@ heap_freetuple(HeapTuple htup)
 }
 
 
-/* ----------------------------------------------------------------
- *						other misc functions
- * ----------------------------------------------------------------
+/* ----------------
+ *		heap_addheader
+ *
+ * This routine forms a HeapTuple by copying the given structure (tuple
+ * data) and adding a generic header.  Note that the tuple data is
+ * presumed to contain no null fields.  It is typically only useful
+ * for null-free system tables.
+ * ----------------
  */
-
 HeapTuple
-heap_addheader(uint32 natts,	/* max domain index */
-			   int structlen,	/* its length */
-			   char *structure) /* pointer to the struct */
+heap_addheader(int natts,		/* max domain index */
+			   Size structlen,	/* its length */
+			   void *structure) /* pointer to the struct */
 {
 	HeapTuple	tuple;
-	HeapTupleHeader td;			/* tuple data */
-	unsigned long len;
+	HeapTupleHeader td;
+	Size		len;
 	int			hoff;
 
 	AssertArg(natts > 0);
 
-	len = offsetof(HeapTupleHeaderData, t_bits);
+	/* header needs no null bitmap */
+	hoff = MAXALIGN(offsetof(HeapTupleHeaderData, t_bits));
+	len = hoff + structlen;
 
-	hoff = len = MAXALIGN(len); /* be conservative */
-	len += structlen;
 	tuple = (HeapTuple) palloc(HEAPTUPLESIZE + len);
-	tuple->t_datamcxt = CurrentMemoryContext;
-	td = tuple->t_data = (HeapTupleHeader) ((char *) tuple + HEAPTUPLESIZE);
 
 	tuple->t_len = len;
 	ItemPointerSetInvalid(&(tuple->t_self));
 	tuple->t_tableOid = InvalidOid;
+	tuple->t_datamcxt = CurrentMemoryContext;
+	tuple->t_data = td = (HeapTupleHeader) ((char *) tuple + HEAPTUPLESIZE);
 
 	MemSet((char *) td, 0, len);
 
 	td->t_hoff = hoff;
 	td->t_natts = natts;
-	td->t_infomask = 0;
-	td->t_infomask |= HEAP_XMAX_INVALID;
+	td->t_infomask = HEAP_XMAX_INVALID;	/* XXX sufficient? */
 
-	if (structlen > 0)
-		memcpy((char *) td + hoff, structure, (size_t) structlen);
+	memcpy((char *) td + hoff, structure, structlen);
 
 	return tuple;
 }

src/backend/catalog/Makefile

@@ -2,7 +2,7 @@
 #
 # Makefile for catalog
 #
-# $Header: /cvsroot/pgsql/src/backend/catalog/Makefile,v 1.34 2001/05/14 21:58:10 momjian Exp $
+# $Header: /cvsroot/pgsql/src/backend/catalog/Makefile,v 1.35 2001/06/12 05:55:49 tgl Exp $
 #
 #-------------------------------------------------------------------------
 
@@ -14,42 +14,36 @@ OBJS = catalog.o heap.o index.o indexing.o aclchk.o \
        pg_aggregate.o pg_largeobject.o pg_operator.o pg_proc.o \
        pg_type.o
 
-BKIFILES = global.bki template1.bki global.description template1.description
+BKIFILES = postgres.bki postgres.description
 
 all: SUBSYS.o $(BKIFILES)
 
 SUBSYS.o: $(OBJS)
 	$(LD) $(LDREL) $(LDOUT) SUBSYS.o $(OBJS)
 
+# Note: there are some undocumented dependencies on the ordering in which
+# the catalog header files are assembled into postgres.bki.  In particular,
+# indexing.h had better be last.
 
-GLOBAL_BKI_SRCS := $(addprefix $(top_srcdir)/src/include/catalog/,\
-	pg_database.h pg_shadow.h pg_group.h pg_log.h \
-    )
-
-TEMPLATE1_BKI_SRCS := $(addprefix $(top_srcdir)/src/include/catalog/,\
+POSTGRES_BKI_SRCS := $(addprefix $(top_srcdir)/src/include/catalog/,\
 	pg_proc.h pg_type.h pg_attribute.h pg_class.h \
-	pg_inherits.h pg_index.h pg_statistic.h \
+	pg_attrdef.h pg_relcheck.h pg_inherits.h pg_index.h \
 	pg_operator.h pg_opclass.h pg_am.h pg_amop.h pg_amproc.h \
-	pg_language.h pg_largeobject.h pg_aggregate.h \
-	pg_rewrite.h pg_listener.h pg_description.h indexing.h \
+	pg_language.h pg_largeobject.h pg_aggregate.h pg_statistic.h \
+	pg_rewrite.h pg_trigger.h pg_listener.h pg_description.h \
+	pg_database.h pg_shadow.h pg_group.h pg_log.h indexing.h \
     )
 
 pg_includes := $(sort -I$(top_srcdir)/src/include -I$(top_builddir)/src/include)
 
-global.bki global.description: genbki.sh $(GLOBAL_BKI_SRCS) $(top_srcdir)/src/include/catalog/indexing.h \
-    $(top_srcdir)/src/include/postgres_ext.h $(top_builddir)/src/include/config.h
-	CPP='$(CPP)' AWK='$(AWK)' $(SHELL) $< $(BKIOPTS) -o global $(pg_includes) $(GLOBAL_BKI_SRCS) --set-version=$(VERSION)
-
-template1.bki template1.description: genbki.sh $(TEMPLATE1_BKI_SRCS) \
+postgres.bki postgres.description: genbki.sh $(POSTGRES_BKI_SRCS) \
     $(top_srcdir)/src/include/postgres_ext.h $(top_builddir)/src/include/config.h
-	CPP='$(CPP)' AWK='$(AWK)' $(SHELL) $< $(BKIOPTS) -o template1 $(pg_includes) $(TEMPLATE1_BKI_SRCS) --set-version=$(VERSION)
+	CPP='$(CPP)' AWK='$(AWK)' $(SHELL) $< $(BKIOPTS) -o postgres $(pg_includes) $(POSTGRES_BKI_SRCS) --set-version=$(VERSION)
 
 .PHONY: install-bki
 install-bki: $(BKIFILES) installdirs
-	$(INSTALL_DATA) global.bki            $(DESTDIR)$(datadir)/global.bki
-	$(INSTALL_DATA) global.description    $(DESTDIR)$(datadir)/global.description
-	$(INSTALL_DATA) template1.bki         $(DESTDIR)$(datadir)/template1.bki
-	$(INSTALL_DATA) template1.description $(DESTDIR)$(datadir)/template1.description
+	$(INSTALL_DATA) postgres.bki         $(DESTDIR)$(datadir)/postgres.bki
+	$(INSTALL_DATA) postgres.description $(DESTDIR)$(datadir)/postgres.description
 
 installdirs:
 	$(mkinstalldirs) $(DESTDIR)$(datadir)

src/backend/catalog/heap.c

@@ -8,7 +8,7 @@
  *
  *
  * IDENTIFICATION
- *	  $Header: /cvsroot/pgsql/src/backend/catalog/heap.c,v 1.166 2001/05/30 12:57:36 momjian Exp $
+ *	  $Header: /cvsroot/pgsql/src/backend/catalog/heap.c,v 1.167 2001/06/12 05:55:49 tgl Exp $
  *
  *
  * INTERFACE ROUTINES
@@ -218,8 +218,9 @@ heap_create(char *relname,
 	}
 
 	/*
-	 * real ugly stuff to assign the proper relid in the relation
-	 * descriptor follows.
+	 * Real ugly stuff to assign the proper relid in the relation
+	 * descriptor follows.  Note that only "bootstrapped" relations
+	 * whose OIDs are hard-coded in pg_class.h need be listed here.
 	 */
 	if (relname && IsSystemRelationName(relname))
 	{
@@ -263,12 +264,6 @@ heap_create(char *relname,
 			tblNode = InvalidOid;
 			relid = RelOid_pg_log;
 		}
-		else if (strcmp(AttrDefaultRelationName, relname) == 0)
-			relid = RelOid_pg_attrdef;
-		else if (strcmp(RelCheckRelationName, relname) == 0)
-			relid = RelOid_pg_relcheck;
-		else if (strcmp(TriggerRelationName, relname) == 0)
-			relid = RelOid_pg_trigger;
 		else
 		{
 			relid = newoid();
@@ -566,7 +561,7 @@ AddNewAttributeTuples(Oid new_rel_oid,
 
 		tup = heap_addheader(Natts_pg_attribute,
 							 ATTRIBUTE_TUPLE_SIZE,
-							 (char *) *dpp);
+							 (void *) *dpp);
 
 		heap_insert(rel, tup);
 
@@ -583,16 +578,18 @@ AddNewAttributeTuples(Oid new_rel_oid,
 	dpp = SysAtt;
 	for (i = 0; i < -1 - FirstLowInvalidHeapAttributeNumber; i++)
 	{
-		/* Fill in the correct relation OID */
-		/* HACK: we are writing on static data here */
-		(*dpp)->attrelid = new_rel_oid;
-		/* Unneeded since they should be OK in the constant data anyway */
-		/* (*dpp)->attstattarget = 0; */
-		/* (*dpp)->attcacheoff = -1; */
+		Form_pg_attribute attStruct;
 
 		tup = heap_addheader(Natts_pg_attribute,
 							 ATTRIBUTE_TUPLE_SIZE,
-							 (char *) *dpp);
+							 (void *) *dpp);
+
+		/* Fill in the correct relation OID in the copied tuple */
+		attStruct = (Form_pg_attribute) GETSTRUCT(tup);
+		attStruct->attrelid = new_rel_oid;
+		/* Unneeded since they should be OK in the constant data anyway */
+		/* attStruct->attstattarget = 0; */
+		/* attStruct->attcacheoff = -1; */
 
 		heap_insert(rel, tup);
 
@@ -603,13 +600,13 @@ AddNewAttributeTuples(Oid new_rel_oid,
 		dpp++;
 	}
 
-	heap_close(rel, RowExclusiveLock);
-
 	/*
 	 * close pg_attribute indices
 	 */
 	if (hasindex)
 		CatalogCloseIndices(Num_pg_attr_indices, idescs);
+
+	heap_close(rel, RowExclusiveLock);
 }
 
 /* --------------------------------
@@ -686,7 +683,7 @@ AddNewRelationTuple(Relation pg_class_desc,
 	 */
 	tup = heap_addheader(Natts_pg_class_fixed,
 						 CLASS_TUPLE_SIZE,
-						 (char *) new_rel_reltup);
+						 (void *) new_rel_reltup);
 
 	/* force tuple to have the desired OID */
 	tup->t_data->t_oid = new_rel_oid;