Tom Lane [Tue, 27 Jun 2006 18:59:17 +0000 (18:59 +0000)]
Put #ifdef NOT_USED around posix_fadvise call. We may want to resurrect
this someday, but right now it seems that posix_fadvise is immature to
the point of being broken on many platforms ... and we don't have any
benchmark evidence proving it's worth spending time on.
Tom Lane [Tue, 27 Jun 2006 16:53:02 +0000 (16:53 +0000)]
Extend the MinimalTuple concept to tuplesort.c, thereby reducing the
per-tuple space overhead for sorts in memory. I chose to replace the
previous patch that tried to write out the bare minimum amount of data
when sorting on disk; instead, just dump the MinimalTuples as-is. This
wastes 3 to 10 bytes per tuple depending on architecture and null-bitmap
length, but the simplification in the writetup/readtup routines seems
worth it.
Alvaro Herrera [Tue, 27 Jun 2006 03:45:16 +0000 (03:45 +0000)]
Clamp last_anl_tuples to n_live_tuples, in case we vacuum a table without
analyzing, so that future analyze threshold calculations don't get confused.
Also, make sure we correctly track the decrease of live tuples cause by
deletes.
Per report from Dylan Hansen, patches by Tom Lane and me.
Bruce Momjian [Tue, 27 Jun 2006 03:22:45 +0000 (03:22 +0000)]
Done:
< * %Disallow changing DEFAULT expression of a SERIAL column?
<
< This should be done only if the existing SERIAL problems cannot be
< fixed.
<
> * -Disallow changing DEFAULT expression of a SERIAL column
Tom Lane [Tue, 27 Jun 2006 02:51:40 +0000 (02:51 +0000)]
Create infrastructure for 'MinimalTuple' representation of in-memory
tuples with less header overhead than a regular HeapTuple, per my
recent proposal. Teach TupleTableSlot code how to deal with these.
As proof of concept, change tuplestore.c to store MinimalTuples instead
of HeapTuples. Future patches will expand the concept to other places
where it is useful.
Tom Lane [Mon, 26 Jun 2006 17:24:41 +0000 (17:24 +0000)]
Change the row constructor syntax (ROW(...)) so that list elements foo.*
will be expanded to a list of their member fields, rather than creating
a nested rowtype field as formerly. (The old behavior is still available
by omitting '.*'.) This syntax is not allowed by the SQL spec AFAICS,
so changing its behavior doesn't violate the spec. The new behavior is
substantially more useful since it allows, for example, triggers to check
for data changes with 'if row(new.*) is distinct from row(old.*)'. Per
my recent proposal.
Tom Lane [Sun, 25 Jun 2006 18:29:49 +0000 (18:29 +0000)]
Tweak dynahash.c to avoid wasting memory space in non-shared hash tables.
palloc() will normally round allocation requests up to the next power of 2,
so make dynahash choose allocation sizes that are as close to a power of 2
as possible.
Back-patch to 8.1 --- the problem exists further back, but a much larger
patch would be needed and it doesn't seem worth taking any risks.
Bruce Momjian [Sun, 25 Jun 2006 16:27:41 +0000 (16:27 +0000)]
Add:
< * Reuse index tuples that point to rows that are not visible to anyone?
> * Reuse index tuples that point to heap tuples that are not visible to
> anyone?
Alvaro Herrera [Sun, 25 Jun 2006 04:37:55 +0000 (04:37 +0000)]
Our version of getopt_long does not set optarg upon detecting an error, as
opposed to what other versions apparently do, so it's not safe to print an
error message. Besides, getopt_long itself already did, so it's redundant
anyway.
Bruce Momjian [Sun, 25 Jun 2006 01:45:32 +0000 (01:45 +0000)]
Remove individual user copyright because the code is contributed to
PGDG:
> Yes. In fact the copyright belongs to credativ GmbH the company that
> paid Carsten for his work. As you may or may not know I'm the CEO of
> that company and can assure you that his work was contributed to the
> PostgreSQL project.
Bruce Momjian [Sun, 25 Jun 2006 00:18:24 +0000 (00:18 +0000)]
Fix Win32/Cygwin problems:
After updating to the latest cvs, and also building most of the addons
(like PLs), the following patch is neededf for win32 + Visual C++.
* Switch to use the new win32 semaphore code
* Rename win32_open to pgwin32_open. win32_open collides with symbols
defined in Perl. MingW didn't detect ig, MSVC did. And it's a bit too
generic a name to export globally, imho...
* Python defines some partially broken #pragmas in the headers when
doing a debug build. Workaround.
Bruce Momjian [Sat, 24 Jun 2006 23:47:58 +0000 (23:47 +0000)]
Update entry:
< * Allow heap reuse of UPDATEd rows if old and new versions are on the
< same heap page?
> * Allow heap reuse of UPDATEd rows if no indexed columns are changed,
> and old and new versions are on the same heap page?
< This is possible for same-page updates because a single index row
< can point to both old and new values.
> While vacuum handles DELETEs fine, updating of non-indexed columns, like
> counters, are difficult for VACUUM to handle efficiently. This method
> is possible for same-page updates because a single index row can be
> used to point to both old and new values.
Bruce Momjian [Sat, 24 Jun 2006 23:45:02 +0000 (23:45 +0000)]
Add UPDATE entry for row reuse.
>
> * Allow heap reuse of UPDATEd rows if old and new versions are on the
> same heap page?
>
> This is possible for same-page updates because a single index row
> can point to both old and new values.
> http://archives.postgresql.org/pgsql-hackers/2006-06/msg01305.php
Tom Lane [Thu, 22 Jun 2006 20:42:57 +0000 (20:42 +0000)]
pg_stop_backup was calling XLogArchiveNotify() twice for the newly created
backup history file. Bug introduced by the 8.1 change to make pg_stop_backup
delete older history files. Per report from Masao Fujii.
Tom Lane [Wed, 21 Jun 2006 19:40:31 +0000 (19:40 +0000)]
Move setup_cancel_handler() up near start of psql main(), where the
setup_win32_locks() call formerly was, to ensure that cancelConnLock is
valid when it needs to be. Per Yoshiyuki Asaba.
Tom Lane [Wed, 21 Jun 2006 18:39:42 +0000 (18:39 +0000)]
Remove ancient kluge that kept nodeAgg.c from crashing on UPDATEs involving
aggregates. We just disallowed that, and AFAICS there should be no other
cases where direct (non-aggregated) references to input columns are allowed
in a query with aggregation and no GROUP BY.
Tom Lane [Wed, 21 Jun 2006 18:30:11 +0000 (18:30 +0000)]
Disallow aggregate functions in UPDATE commands (unless within a sub-SELECT).
This is disallowed by the SQL spec because it doesn't have any very sensible
interpretation. Historically Postgres has allowed it but behaved strangely.
As of PG 8.1 a server crash is possible if the MIN/MAX index optimization gets
applied; rather than try to "fix" that, it seems best to just enforce the
spec restriction. Per report from Josh Drake and Alvaro Herrera.
Joe Conway [Wed, 21 Jun 2006 16:43:11 +0000 (16:43 +0000)]
- During dblink_open, if transaction state was IDLE, force cursor count to
initially be 0. This is needed as a previous ABORT might have wiped out
an automatically opened transaction without maintaining the cursor count.
- Fix regression test expected file for the correct ERROR message, which
we now get given the above bug fix.
Tom Lane [Wed, 21 Jun 2006 16:05:11 +0000 (16:05 +0000)]
Clean up psql variable code a little: eliminate unnecessary tests in
GetVariable() and be consistent about treatment of the list header.
Motivated by noticing strspn() taking an unreasonable percentage of
runtime --- the call removed from GetVariable() was the only one that
could be in a high-usage path ...
Tom Lane [Tue, 20 Jun 2006 22:52:00 +0000 (22:52 +0000)]
Remove redundant gettimeofday() calls to the extent practical without
changing semantics too much. statement_timestamp is now set immediately
upon receipt of a client command message, and the various places that used
to do their own gettimeofday() calls to mark command startup are referenced
to that instead. I have also made stats_command_string use that same
value for pg_stat_activity.query_start for both the command itself and
its eventual replacement by <IDLE> or <idle in transaction>. There was
some debate about that, but no argument that seemed convincing enough to
justify an extra gettimeofday() call.
Tom Lane [Tue, 20 Jun 2006 19:56:52 +0000 (19:56 +0000)]
Split definitions for md5.c out of crypt.h and into their own header
libpq/md5.h, so that there's a clear separation between backend-only
definitions and shared frontend/backend definitions. (Turns out this
is reversing a bad decision from some years ago...) Fix up references
to crypt.h as needed. I looked into moving the code into src/port, but
the headers in src/include/libpq are sufficiently intertwined that it
seems more work than it's worth to do that.
Tom Lane [Mon, 19 Jun 2006 01:51:22 +0000 (01:51 +0000)]
Take the statistics collector out of the loop for monitoring backends'
current commands; instead, store current-status information in shared
memory. This substantially reduces the overhead of stats_command_string
and also ensures that pg_stat_activity is fully up to date at all times.
Per my recent proposal.
Tom Lane [Sun, 18 Jun 2006 18:30:21 +0000 (18:30 +0000)]
Don't try to call posix_fadvise() unless <fcntl.h> supplies a declaration
for it. Hopefully will fix core dump evidenced by some buildfarm members
since fadvise patch went in. The actual definition of the function is not
ABI-compatible with compiler's default assumption in the absence of any
declaration, so it's clearly unsafe to try to call it without seeing a
declaration.
Tom Lane [Sun, 18 Jun 2006 17:30:48 +0000 (17:30 +0000)]
Fix saveHistory() to not emit bogus complaint during psql exit when using
Darwin's libedit; per my proposal last August. Also, suppress cast-away-
const warnings.
Tom Lane [Sun, 18 Jun 2006 16:21:23 +0000 (16:21 +0000)]
Increase timeout in statement_timeout test from 1 second to 2 seconds.
We have once or twice seen failures suggesting that control didn't get
to the exception block before the timeout elapsed, which is unlikely
but not impossible in a parallel regression test (with a dozen other
backends competing for cycles). This change doesn't completely prevent
the problem of course, but it should reduce the probability enough that
we don't see it anymore. Per buildfarm results.
Peter Eisentraut [Sun, 18 Jun 2006 15:38:37 +0000 (15:38 +0000)]
Merge postmaster and postgres command into just postgres. postmaster
symlink is kept for now for compatibility. To call single-user mode, use
postgres --single.
Bruce Momjian [Fri, 16 Jun 2006 22:08:46 +0000 (22:08 +0000)]
Add:
> o Allow PL/python to composite types and result sets
> once buggy assert-enabled versions of python can be detected
>
> http://archives.postgresql.org/pgsql-patches/2006-04/msg00087.php
Tom Lane [Fri, 16 Jun 2006 18:42:24 +0000 (18:42 +0000)]
Fix problems with cached tuple descriptors disappearing while still in use
by creating a reference-count mechanism, similar to what we did a long time
ago for catcache entries. The back branches have an ugly solution involving
lots of extra copies, but this way is more efficient. Reference counting is
only applied to tupdescs that are actually in caches --- there seems no need
to use it for tupdescs that are generated in the executor, since they'll go
away during plan shutdown by virtue of being in the per-query memory context.
Neil Conway and Tom Lane
Bruce Momjian [Fri, 16 Jun 2006 17:44:43 +0000 (17:44 +0000)]
Add:
> o Fix problems with RETURN NEXT on tables with
> dropped/added columns after function creation
>
> http://archives.postgresql.org/pgsql-patches/2006-02/msg00165.php
Bruce Momjian [Thu, 15 Jun 2006 16:54:13 +0000 (16:54 +0000)]
Clarify dynamic pl/pgsql item and add URLs. Restructure server-side
section into PL/pgSQL and non-PL/pgSQL sections:
< o Fix PL/pgSQL RENAME to work on variables other than OLD/NEW
< o Allow function parameters to be passed by name,
< get_employee_salary(emp_id => 12345, tax_year => 2001)
< o Add Oracle-style packages
< o Add table function support to pltcl, plpython
< o Add capability to create and call PROCEDURES
< o Allow PL/pgSQL to handle %TYPE arrays, e.g. tab.col%TYPE[]
< o Allow function argument names to be statements from PL/PgSQL
< o Add MOVE to PL/pgSQL
< o Add support for polymorphic arguments and return types to
< languages other than PL/PgSQL
< o Add support for OUT and INOUT parameters to languages other
< than PL/PgSQL
< o Add single-step debugging of PL/PgSQL functions
< o Allow PL/PgSQL to support WITH HOLD cursors
< o Allow PL/PgSQL RETURN to return row or record functions
<
< http://archives.postgresql.org/pgsql-patches/2005-11/msg00045.php
> o PL/pgSQL
> o Fix RENAME to work on variables other than OLD/NEW
> o Allow function parameters to be passed by name,
> get_employee_salary(emp_id => 12345, tax_year => 2001)
> o Add Oracle-style packages
> o Allow handling of %TYPE arrays, e.g. tab.col%TYPE[]
> o Allow listing of record column names, and access to
> record columns via variables, e.g. columns := r.(*),
> tval2 := r.(colname)
>
> http://archives.postgresql.org/pgsql-patches/2005-07/msg00458.php
> http://archives.postgresql.org/pgsql-patches/2006-05/msg00302.php
> http://archives.postgresql.org/pgsql-patches/2006-06/msg00031.php
>
> o Add MOVE
> o Add single-step debugging of functions
> o Add support for WITH HOLD cursors
> o Allow PL/RETURN to return row or record functions
>
> http://archives.postgresql.org/pgsql-patches/2005-11/msg00045.php
>
>
> o Other
> o Add table function support to pltcl, plpython
> o Add support for polymorphic arguments and return types to
> languages other than PL/PgSQL
> o Add capability to create and call PROCEDURES
> o Add support for OUT and INOUT parameters to languages other
> than PL/PgSQL
Bruce Momjian [Thu, 15 Jun 2006 02:23:13 +0000 (02:23 +0000)]
Done, removed:
<
< * Consider padding the spinlock structure to have spinlock registers
< on different CPU cache line
<
< http://archives.postgresql.org/pgsql-hackers/2005-10/msg00498.php
Tom Lane [Thu, 15 Jun 2006 02:08:09 +0000 (02:08 +0000)]
Remove the limit on the number of entries allowed in catcaches, and
remove the infrastructure needed to enforce the limit, ie, the global
LRU list of cache entries. On small-to-middling databases this wins
because maintaining the LRU list is a waste of time. On large databases
this wins because it's better to keep more cache entries (we assume
such users can afford to use some more per-backend memory than was
contemplated in the Berkeley-era catcache design). This provides a
noticeable improvement in the speed of psql \d on a 10000-table
database, though it doesn't make it instantaneous.
While at it, use per-catcache settings for the number of hash buckets
per catcache, rather than the former one-size-fits-all value. It's a
bit silly to be using the same number of hash buckets for, eg, pg_am
and pg_attribute. The specific values I used might need some tuning,
but they seem to be in the right ballpark based on CATCACHE_STATS
results from the standard regression tests.