This first part of the background writer does no syncing at all.
It's only purpose is to keep the LRU heads clean so that regular
backends seldom to never have to call write().

Jan

> * Add GUC setting to make created tables default to WITHOUT OIDS

< 	o --Add ALTER TABLE tab SET WITHOUT OIDS (Rod)
> 	o -Add ALTER TABLE tab SET WITHOUT OIDS (Rod)

which had been unintentionally broken by recent changes to tighten up the
DateStyle rules for all-numeric date input.  Add documentation and
regression tests for this, too.

    debug_shared_buffers = <seconds>

as per previous discussion.


Jan

to try additional addresses returned from getaddrinfo() if the first one
fails at the bind() or connect() steps.  Per yesterday's discussion.

a comma between multiple column_constraint's for a single column.
Per report from Tomislaw Kitynski.

can terminate the bootstrap run.

double or out-of-sequence output with child process.

data directory.  Also fix handling of error conditions associated with
data directory checking step (can't use a boolean to distinguish four
possible result states...)

usage of PGSemaMagic, the minimum required value is actually 537, but
say 1000 just to use a round number.  255 is *not* good enough.

and the cache replacement strategy as well as a description of the
ARC algorithm and the special tailoring of that done for PostgreSQL.

Jan

of option switches for backend, fix handling of COPY from data files so
that we won't have the newline-after-\. issue back again, add back some
comments and printouts lost from the shell script, etc.  Still needs work
for error handling; in particular the shell version worked much more
nicely for the case of a postgres executable that fails on invocation.

add a header comment/copyright notice.

Andrew Dunstan