of an index can now be a computed expression instead of a simple variable.
Restrictions on expressions are the same as for predicates (only immutable
functions, no sub-selects).  This fixes problems recently introduced with
inlining SQL functions, because the inlining transformation is applied to
both expression trees so the planner can still match them up.  Along the
way, improve efficiency of handling index predicates (both predicates and
index expressions are now cached by the relcache) and fix 7.3 oversight
that didn't record dependencies of predicate expressions.

users can't turn it off.

docs that CLIENT/LOG_MIN_MESSAGES now controls debug_* output location.
Doc changes included.

sanely with running out of memory for a query result.

< * Add now("transaction|statement|clock") functionality
> * Add transaction_timestamp(), statement_timestamp(), clock_timestamp() functionality

advertised in RowDescription message.  Depending on the physical tuple's
column count is not really correct, since according to heap_getattr()
conventions the tuple may be short some columns, which will automatically
get read as nulls.  Problem has been latent since forever, but was only
exposed by recent change to skip a projection step in SELECT * FROM...

modify passwords, per user confusion report.

Also clean up wording that command line utils need database access.

> * Allow UPDATE to use SET col = DEFAULT

blanks, in hopes of reducing the surprise factor for newbies.  Remove
redundant operators for VARCHAR (it depends wholly on TEXT operations now).
Clean up resolution of ambiguous operators/functions to avoid surprising
choices for domains: domains are treated as equivalent to their base types
and binary-coercibility is no longer considered a preference item when
choosing among multiple operators/functions.  IsBinaryCoercible now correctly
reflects the notion that you need *only* relabel the type to get from type
A to type B: that is, a domain is binary-coercible to its base type, but
not vice versa.  Various marginal cleanup, including merging the essentially
duplicate resolution code in parse_func.c and parse_oper.c.  Improve opr_sanity
regression test to understand about binary compatibility (using pg_cast),
and fix a couple of small errors in the catalogs revealed thereby.
Restructure "special operator" handling to fetch operators via index opclasses
rather than hardwiring assumptions about names (cleans up the pattern_ops
stuff a little).

< * Update clients to use data types, typmod, schema.table.column names of
<   result sets using new query protocol
453a452,453
> 	o Update clients to use data types, typmod, schema.table.column names of
> 	  result sets using new query protocol

< * Allow clients to get data types, typmod, schema.table.column names from
<   result sets, either via the backend protocol or a new QUERYINFO command

to:

> * Update clients to use data types, typmod, schema.table.column names of
>   result sets using new query protocol

> * Allow clients to get data types, typmod, schema.table.column names from
>   result sets, either via the backend protocol or a new QUERYINFO command

> * Allow backend to output result sets in XML
> * Add XML capability to pg_dump and COPY, when backend XML capability

> * Find a way to reduce rotational delay when repeatedly writing
>   last WAL page

  * With disabled triggers, allow pg_dump to use ALTER TABLE ADD FOREIGN KEY

Add to trigger TODO.detail.

single-byte encodings, and a direct C implementation of the single-argument
forms (where spaces are always what gets trimmed).  This is in preparation
for using rtrim1() as the bpchar-to-text cast operator, but is a useful
performance improvement even if we decide not to do that.