>>Doh. Quite right. Well, the 7.0 initdb creates pg_indexes as the last
>>thing it does. (Even though there are higher oids in pg_attribute,
>>pg_rewrite and pg_description.), so how about the attached patch?
>
>
> Seems reasonable, but the patch ought to fix the comment just above,
> preferably with the above observation that this is correct by
> examination of the now-frozen 7.0 initdb code ...
Attached.
Chris
Index: pg_dump.c
===================================================================
RCS file: /projects/cvsroot/pgsql-server/src/bin/pg_dump/pg_dump.c,v
retrieving revision 1.353
diff -c -r1.353 pg_dump.c
*** pg_dump.c 8 Oct 2003 03:52:32 -0000 1.353
--- pg_dump.c 21 Oct 2003 01:17:58 -0000
***************
*** 5864,5870 ****
/*
* findLastBuiltInOid -
* find the last built in oid
! * we do this by looking up the oid of 'template1' in pg_database,
* this is probably not foolproof but comes close
*/
--- 5864,5871 ----
/*
* findLastBuiltInOid -
* find the last built in oid
! * we do this by assuming that the last thing the now-frozen 7.0.x initdb
! * does is to create the pg_indexes view.
* this is probably not foolproof but comes close
*/
***************
*** 5876,5882 ****
int last_oid;
res = PQexec(g_conn,
! "SELECT oid from pg_database where datname = 'template1'");
if (res == NULL ||
PQresultStatus(res) != PGRES_TUPLES_OK)
{
--- 5877,5883 ----
int last_oid;
res = PQexec(g_conn,
! "SELECT oid FROM pg_class WHERE relname='pg_indexes'");
if (res == NULL ||
PQresultStatus(res) != PGRES_TUPLES_OK)
{
