optimization.

call for every stats packet read to adjust select() timeout.  Other
stylistic improvements.

Remove the nexus between trial_buffs and trial_conns, and don't test shared buffers lower than or equal to a value we already know is good, but use that value instead. This will make it easier to adjust the trial values and/or formulae in future if necessary.

from Andrus Moor.  The former state-machine-style coding wasn't actually
doing much except obscuring the control flow, and it didn't extend
readily to fix this case, so I just took it out.  Also, add a
YY_FLUSH_BUFFER call to ensure the lexer is reset correctly if the
previous scan failed partway through the file.

are inconsistent with the rest of the .po files, and apparently cause
problems for Sun's cc. Per report on IRC from "bitvector2".

formerly "RETRIEVE", and so forth.

a little bit, and set the minimum buffers-per-connection ratio to 10 not
5.  I folded the two test routines into one to counteract the illusion
that the tests can be twiddled independently, and added some documentation
pointing out the necessary connection between the sets of values tested.
Fixes strange choices of parameters that I noticed CVS tip making on
Darwin with Apple's undersized default SHMMAX.

they can be easily distinguished from SQL commands.

Pierre Girard

selection of a field from the result of a function returning RECORD.
I believe this case is new in 8.1; it's due to the addition of OUT parameters.
Per example from Michael Fuhr.

> * Make CLUSTER preserve recently-dead tuples per MVCC requirements

===================================================================
RCS file: /Users/neilc/postgres/cvs_root/pgsql/src/pl/plpython/plpython.c,v
retrieving revision 1.67
diff -c -r1.67 plpython.c
*** src/pl/plpython/plpython.c	26 Dec 2005 04:28:48 -0000	1.67
--- src/pl/plpython/plpython.c	29 Dec 2005 16:54:57 -0000
***************
*** 2,8 ****
   * plpython.c - python as a procedural language for PostgreSQL
   *
   * This software is copyright by Andrew Bosma
!  * but is really shameless cribbed from pltcl.c by Jan Weick, and
   * plperl.c by Mark Hollomon.
   *
   * The author hereby grants permission to use, copy, modify,
--- 2,8 ----
   * plpython.c - python as a procedural language for PostgreSQL
   *
   * This software is copyright by Andrew Bosma
!  * but is really shamelessly cribbed from pltcl.c by Jan Wieck, and
   * plperl.c by Mark Hollomon.
   *
   * The author hereby grants permission to use, copy, modify,
***************
*** 1996,2002 ****
  	int			i,
  				rv;
  	PLyPlanObject *plan;
- 	char	   *nulls;
  	MemoryContext oldcontext;

  	if (list != NULL)
--- 1996,2001 ----
***************
*** 2018,2024 ****
  	if (nargs != plan->nargs)
  	{
  		char	   *sv;
-
  		PyObject   *so = PyObject_Str(list);

  		if (!so)
--- 2017,2022 ----
***************
*** 2036,2048 ****
  	oldcontext = CurrentMemoryContext;
  	PG_TRY();
  	{
! 		nulls = palloc(nargs * sizeof(char));

  		for (i = 0; i < nargs; i++)
  		{
  			PyObject   *elem,
  					   *so;
- 			char	   *sv;

  			elem = PySequence_GetItem(list, i);
  			if (elem != Py_None)
--- 2034,2045 ----
  	oldcontext = CurrentMemoryContext;
  	PG_TRY();
  	{
! 		char	   *nulls = palloc(nargs * sizeof(char));

  		for (i = 0; i < nargs; i++)
  		{
  			PyObject   *elem,
  					   *so;

  			elem = PySequence_GetItem(list, i);
  			if (elem != Py_None)
***************
*** 2051,2070 ****
  				if (!so)
  					PLy_elog(ERROR, "function \"%s\" could not execute plan",
  							 PLy_procedure_name(PLy_curr_procedure));
! 				sv = PyString_AsString(so);

! 				/*
! 				 * FIXME -- if this elogs, we have Python reference leak
! 				 */
! 				plan->values[i] =
! 					FunctionCall3(&(plan->args[i].out.d.typfunc),
! 								  CStringGetDatum(sv),
! 							ObjectIdGetDatum(plan->args[i].out.d.typioparam),
! 								  Int32GetDatum(-1));

! 				Py_DECREF(so);
! 				Py_DECREF(elem);

  				nulls[i] = ' ';
  			}
  			else
--- 2048,2073 ----
  				if (!so)
  					PLy_elog(ERROR, "function \"%s\" could not execute plan",
  							 PLy_procedure_name(PLy_curr_procedure));
! 				Py_DECREF(elem);

! 				PG_TRY();
! 				{
! 					char *sv = PyString_AsString(so);

! 					plan->values[i] =
! 						FunctionCall3(&(plan->args[i].out.d.typfunc),
! 									  CStringGetDatum(sv),
! 								ObjectIdGetDatum(plan->args[i].out.d.typioparam),
! 									  Int32GetDatum(-1));
! 				}
! 				PG_CATCH();
! 				{
! 					Py_DECREF(so);
! 					PG_RE_THROW();
! 				}
! 				PG_END_TRY();

+ 				Py_DECREF(so);
  				nulls[i] = ' ';
  			}
  			else

in favor of having just one set of macros that don't do HOLD/RESUME_INTERRUPTS
(hence, these correspond to the old SpinLockAcquire_NoHoldoff case).
Given our coding rules for spinlock use, there is no reason to allow
CHECK_FOR_INTERRUPTS to be done while holding a spinlock, and also there
is no situation where ImmediateInterruptOK will be true while holding a
spinlock.  Therefore doing HOLD/RESUME_INTERRUPTS while taking/releasing a
spinlock is just a waste of cycles.  Qingqing Zhou and Tom Lane.

< * Add missing rtree optimizer selectivity
> * Improve selectivity functions for geometric operators

Move declaration of check_function_bodies to where the perl headers
haven't had a chance to mangle the definition of DLLIMPORT (thanks again, perl guys).

not be replaced with "ADD_P".

documentation.

Michael Fuhr

setup.  This protects against undesired changes in locale behavior
if someone carelessly does setlocale(LC_ALL, "") (and we know who
you are, perl guys).

wording ("table" added).

we want it to check the argument/result data types and no more.  In
particular, libperl shouldn't get initialized in this case.

get_func_arg_info() for consistency with other names there.
This code will probably be useful to other PLs when they start to
support OUT parameters, so better to have it in the main backend.
Also, fix plpgsql validator to detect bogus OUT parameters even when
check_function_bodies is off.

lacking it.  Perhaps it was suppressed because of line length
considerations, but "table" should appear.

< * Add missing optimizer selectivities for date, r-tree, etc
> * Add missing rtree optimizer selectivity

single column dump that has a \. value, so the load works properly.  I
also added documentation describing this issue.

< * %Make row-wise comparisons work per SQL spec
> * -Make row-wise comparisons work per SQL spec

(previously we only did = and <> correctly).  Also, allow row comparisons
with any operators that are in btree opclasses, not only those with these
specific names.  This gets rid of a whole lot of indefensible assumptions
about the behavior of particular operators based on their names ... though
it's still true that IN and NOT IN expand to "= ANY".  The patch adds a
RowCompareExpr expression node type, and makes some changes in the
representation of ANY/ALL/ROWCOMPARE SubLinks so that they can share code
with RowCompareExpr.

I have not yet done anything about making RowCompareExpr an indexable
operator, but will look at that soon.

initdb forced due to changes in stored rules.

Increase amount of shared buffers initdb tries to allocate to 4000,
and add logic to try max_fsm_pages up to 200000, plus accompanying minor
docs changes.

                if (c == '\\' && cstate->line_buf.len == 0)

The problem with that is the because of the input and _output_
buffering, cstate->line_buf.len could be zero even if we are not on the
first character of a line.  In fact, for a typical line, it is zero for
all characters on the line.  The proper solution is to introduce a
boolean, first_char_in_line, that we set as we enter the loop and clear
once we process a character.

I have restructured the line-reading code in copy.c by:

        o  merging the CSV/non-CSV functions into a single function
        o  used macros to centralize and clarify the buffering code
        o  updated comments
        o  renamed client_encoding_only to encoding_embeds_ascii
        o  added a high-bit test to the encoding_embeds_ascii test for
           performance
        o  in CSV mode, allow a backslash followed by a non-period to
           continue being processed as a data value

There should be no performance impact from this patch because it is
functionally equivalent.  If you apply the patch you will see copy.c is
much clearer in this area now and might suggest additional
optimizations.

I have also attached a 8.1-only patch to fix the CSV \. handling bug
with no code restructuring.