pgsql-server: Adjust btree index build to not use shared buffers,

Log Message:
-----------
Adjust btree index build to not use shared buffers, thereby avoiding the
locking conflict against concurrent CHECKPOINT that was discussed a few
weeks ago. Also, if not using WAL archiving (which is always true ATM
but won't be if PITR makes it into this release), there's no need to
WAL-log the index build process; it's sufficient to force-fsync the
completed index before commit. This seems to gain about a factor of 2
in my tests, which is consistent with writing half as much data. I did
not try it with WAL on a separate drive though --- probably the gain would
be a lot less in that scenario.

Modified Files:
--------------
pgsql-server/src/backend/access/nbtree:
nbtpage.c (r1.75 -> r1.76)
(http://developer.postgresql.org/cvsweb.cgi/pgsql-server/src/backend/access/nbtree/nbtpage.c.diff?r1=1.75&r2=1.76)
nbtree.c (r1.116 -> r1.117)
(http://developer.postgresql.org/cvsweb.cgi/pgsql-server/src/backend/access/nbtree/nbtree.c.diff?r1=1.116&r2=1.117)
nbtsort.c (r1.81 -> r1.82)
(http://developer.postgresql.org/cvsweb.cgi/pgsql-server/src/backend/access/nbtree/nbtsort.c.diff?r1=1.81&r2=1.82)
nbtxlog.c (r1.12 -> r1.13)
(http://developer.postgresql.org/cvsweb.cgi/pgsql-server/src/backend/access/nbtree/nbtxlog.c.diff?r1=1.12&r2=1.13)
pgsql-server/src/backend/storage/smgr:
md.c (r1.106 -> r1.107)
(http://developer.postgresql.org/cvsweb.cgi/pgsql-server/src/backend/storage/smgr/md.c.diff?r1=1.106&r2=1.107)
smgr.c (r1.72 -> r1.73)
(http://developer.postgresql.org/cvsweb.cgi/pgsql-server/src/backend/storage/smgr/smgr.c.diff?r1=1.72&r2=1.73)
pgsql-server/src/include/access:
nbtree.h (r1.77 -> r1.78)
(http://developer.postgresql.org/cvsweb.cgi/pgsql-server/src/include/access/nbtree.h.diff?r1=1.77&r2=1.78)
pgsql-server/src/include/storage:
smgr.h (r1.43 -> r1.44)
(http://developer.postgresql.org/cvsweb.cgi/pgsql-server/src/include/storage/smgr.h.diff?r1=1.43&r2=1.44)