]> granicus.if.org Git - postgresql/log
postgresql
21 years agoFix typo:
Bruce Momjian [Wed, 19 Mar 2003 03:53:21 +0000 (03:53 +0000)]
Fix typo:

< * Rod is Rod Taylor <pg@rbt.ca)>
> * Rod is Rod Taylor <pg@rbt.ca>

21 years agoReimplement create and drop scripts in C, to reduce repetitive
Peter Eisentraut [Tue, 18 Mar 2003 22:19:47 +0000 (22:19 +0000)]
Reimplement create and drop scripts in C, to reduce repetitive
connections, increase robustness, add NLS, and prepare for Windows port.
(vacuumdb and clusterdb will follow later.)

21 years agoMake the printing code somewhat more independent by not relying on
Peter Eisentraut [Tue, 18 Mar 2003 22:15:44 +0000 (22:15 +0000)]
Make the printing code somewhat more independent by not relying on
functions and global variables from the rest of psql.  Also clean up some
data type mismatches created by the last pager patch.

21 years agoImprove error message.
Peter Eisentraut [Tue, 18 Mar 2003 22:11:48 +0000 (22:11 +0000)]
Improve error message.

21 years agoFix off-by-one error in the maxlen parameter handling.
Peter Eisentraut [Tue, 18 Mar 2003 22:09:37 +0000 (22:09 +0000)]
Fix off-by-one error in the maxlen parameter handling.

21 years agoCompiling anything that uses InvalidOid under g++ yields a warning about
Bruce Momjian [Tue, 18 Mar 2003 17:21:07 +0000 (17:21 +0000)]
Compiling anything that uses InvalidOid under g++ yields a warning about
the expression using an "old-style cast."  Therefore, would it be okay
to patch postgres_ext.h as follows:

Jeroen T. Vermeulen

21 years agoRemove duplicate ANALYZE recommendation in pg_dump.
Bruce Momjian [Tue, 18 Mar 2003 17:05:01 +0000 (17:05 +0000)]
Remove duplicate ANALYZE recommendation in pg_dump.

21 years agoSome bugfixes for numerical library.
Michael Meskes [Tue, 18 Mar 2003 10:46:39 +0000 (10:46 +0000)]
Some bugfixes for numerical library.

21 years agoApplied updated translation patch
Barry Lind [Tue, 18 Mar 2003 05:26:50 +0000 (05:26 +0000)]
Applied updated translation patch

 Modified Files:
  errors_zh_TW.properties

21 years agoAdd to mmap discussion.
Bruce Momjian [Tue, 18 Mar 2003 01:36:01 +0000 (01:36 +0000)]
Add to mmap discussion.

21 years agoAdd mention of ANALYZE after object restore.
Bruce Momjian [Tue, 18 Mar 2003 00:02:11 +0000 (00:02 +0000)]
Add mention of ANALYZE after object restore.

21 years agoAdd mention of new sets returning functions capability.
Bruce Momjian [Mon, 17 Mar 2003 19:53:08 +0000 (19:53 +0000)]
Add mention of new sets returning functions capability.

21 years agoUpdate links, from Ian Barwick
Bruce Momjian [Mon, 17 Mar 2003 19:47:28 +0000 (19:47 +0000)]
Update links, from Ian Barwick

21 years agoAdd:
Bruce Momjian [Mon, 17 Mar 2003 18:50:06 +0000 (18:50 +0000)]
Add:

> * Allow UPDATE tab SET ROW (col, ...) = (...) for updating multiple columns

21 years agoReorder crypt.h include for SunOS compile problem.
Bruce Momjian [Mon, 17 Mar 2003 17:58:57 +0000 (17:58 +0000)]
Reorder crypt.h include for SunOS compile problem.

Fred Houweling

21 years agoIncrease max known database size.
Bruce Momjian [Mon, 17 Mar 2003 16:33:12 +0000 (16:33 +0000)]
Increase max known database size.

21 years agoForgot some files...
Michael Meskes [Sun, 16 Mar 2003 10:49:51 +0000 (10:49 +0000)]
Forgot some files...

21 years agoStarted working on a seperate pgtypes library. First test work. PLEASE test compilati...
Michael Meskes [Sun, 16 Mar 2003 10:42:54 +0000 (10:42 +0000)]
Started working on a seperate pgtypes library. First test work. PLEASE test compilation on iother systems.

21 years agoArrange to print the relevant key values when reporting a foreign-key
Tom Lane [Sat, 15 Mar 2003 21:19:40 +0000 (21:19 +0000)]
Arrange to print the relevant key values when reporting a foreign-key
violation.  Also, factor out some duplicate code in the RI triggers.
Patch by Dmitry Tkach, reviewed by Stephan Szabo and Tom Lane.

21 years agoA typo in src/backend/libpq/hba.c breaks local ident authentication
Bruce Momjian [Sat, 15 Mar 2003 16:18:25 +0000 (16:18 +0000)]
A typo in src/backend/libpq/hba.c breaks local ident authentication
in the SO_PEERCRED case. elif is misspelled as elsif for the test.
A patch is attached.

Bruno Wolff III

21 years agopg_dumpall failed on groups having no members. Per report from
Tom Lane [Fri, 14 Mar 2003 22:45:49 +0000 (22:45 +0000)]
pg_dumpall failed on groups having no members.  Per report from
Nick Eskelinen.

21 years agoRemove unneeded dash blocks around function start comments.
Bruce Momjian [Fri, 14 Mar 2003 22:40:31 +0000 (22:40 +0000)]
Remove unneeded dash blocks around function start comments.

21 years agoUpdate name of GUC var:
Bruce Momjian [Fri, 14 Mar 2003 20:19:59 +0000 (20:19 +0000)]
Update name of GUC var:

< * Add GUC log_statement_duration to print statement and >= min duration
> * Add GUC log_statement_and_duration to print statement and >= min duration

21 years agoAdd:
Bruce Momjian [Fri, 14 Mar 2003 19:27:49 +0000 (19:27 +0000)]
Add:

> * Add GUC log_statement_duration to print statement and >= min duration

21 years agoFixed parsing to handle \n for updateable result sets. Bug reported by Rich Cullingford.
Barry Lind [Fri, 14 Mar 2003 05:36:58 +0000 (05:36 +0000)]
Fixed parsing to handle \n for updateable result sets.  Bug reported by Rich Cullingford.

 Modified Files:
  jdbc/org/postgresql/jdbc2/AbstractJdbc2ResultSet.java

21 years agoRepair incorrect prorettype entry for timestamptz_izone. Can't force
Tom Lane [Fri, 14 Mar 2003 04:43:52 +0000 (04:43 +0000)]
Repair incorrect prorettype entry for timestamptz_izone.  Can't force
initdb in the 7.3 branch, but we can at least make it right for people
who install 7.3.3 from scratch.

21 years agoImproved error message and added some log messages in batch update functionality
Barry Lind [Fri, 14 Mar 2003 01:21:47 +0000 (01:21 +0000)]
Improved error message and added some log messages in batch update functionality

 Modified Files:
  jdbc/org/postgresql/errors.properties
  jdbc/org/postgresql/jdbc2/AbstractJdbc2Statement.java
  jdbc/org/postgresql/util/PSQLException.java

21 years agoMake eval_const_expressions simplify FieldSelect from a whole-row Var
Tom Lane [Fri, 14 Mar 2003 00:55:17 +0000 (00:55 +0000)]
Make eval_const_expressions simplify FieldSelect from a whole-row Var
into an ordinary one-field Var.  Per example from Chris Mungall.

21 years agoPrevent infinite loop in ln_var() due to roundoff error.
Tom Lane [Fri, 14 Mar 2003 00:15:32 +0000 (00:15 +0000)]
Prevent infinite loop in ln_var() due to roundoff error.
Per report from Dave Marin.

21 years agoGROUP BY got confused if there were multiple equal() GROUP BY items.
Tom Lane [Thu, 13 Mar 2003 16:58:35 +0000 (16:58 +0000)]
GROUP BY got confused if there were multiple equal() GROUP BY items.
This bug has been latent since 7.0 or maybe even further back, but it
was only exposed when parse_clause.c stopped suppressing duplicate
items (see its rev 1.96 of 18-Aug-02).

21 years agoBig editing for consistent content and presentation.
Peter Eisentraut [Thu, 13 Mar 2003 01:30:29 +0000 (01:30 +0000)]
Big editing for consistent content and presentation.

21 years agoAdd explicit tests for division by zero to all user-accessible integer
Tom Lane [Tue, 11 Mar 2003 21:01:33 +0000 (21:01 +0000)]
Add explicit tests for division by zero to all user-accessible integer
division and modulo functions, to avoid problems on OS X (which fails to
trap 0 divide at all) and Windows (which traps it in some bizarre
nonstandard fashion).  Standardize on 'division by zero' as the one true
spelling of this error message.  Add regression tests as suggested by
Neil Conway.

21 years agoImplement SQL92-compatible FIRST, LAST, ABSOLUTE n, RELATIVE n options
Tom Lane [Tue, 11 Mar 2003 19:40:24 +0000 (19:40 +0000)]
Implement SQL92-compatible FIRST, LAST, ABSOLUTE n, RELATIVE n options
for FETCH and MOVE.

21 years agoThis patch fixes a bunch of spelling mistakes in comments throughout the
Tom Lane [Mon, 10 Mar 2003 22:28:22 +0000 (22:28 +0000)]
This patch fixes a bunch of spelling mistakes in comments throughout the
PostgreSQL source code.

Neil Conway

21 years agoUpdate German FAQ
Bruce Momjian [Mon, 10 Mar 2003 16:06:41 +0000 (16:06 +0000)]
Update German FAQ

Ian Barwick

21 years agoCleanup up psql \connect and \pset pager setting display, Dennis
Bruce Momjian [Mon, 10 Mar 2003 15:46:03 +0000 (15:46 +0000)]
Cleanup up psql \connect and \pset pager setting display, Dennis
Bj?rklund.

21 years agoRestructure parsetree representation of DECLARE CURSOR: now it's a
Tom Lane [Mon, 10 Mar 2003 03:53:52 +0000 (03:53 +0000)]
Restructure parsetree representation of DECLARE CURSOR: now it's a
utility statement (DeclareCursorStmt) with a SELECT query dangling from
it, rather than a SELECT query with a few unusual fields in it.  Add
code to determine whether a planned query can safely be run backwards.
If DECLARE CURSOR specifies SCROLL, ensure that the plan can be run
backwards by adding a Materialize plan node if it can't.  Without SCROLL,
you get an error if you try to fetch backwards from a cursor that can't
handle it.  (There is still some discussion about what the exact
behavior should be, but this is necessary infrastructure in any case.)
Along the way, make EXPLAIN DECLARE CURSOR work.

21 years agopg_restore failed to restore blobs if -X disable-triggers is specified.
Tom Lane [Sun, 9 Mar 2003 19:38:52 +0000 (19:38 +0000)]
pg_restore failed to restore blobs if -X disable-triggers is specified.

21 years agotuplestore_donestoring() isn't needed anymore, but provide a no-op
Tom Lane [Sun, 9 Mar 2003 03:34:10 +0000 (03:34 +0000)]
tuplestore_donestoring() isn't needed anymore, but provide a no-op
macro definition so as not to create compatibility problems.

21 years agoRevise tuplestore and nodeMaterial so that we don't have to read the
Tom Lane [Sun, 9 Mar 2003 02:19:13 +0000 (02:19 +0000)]
Revise tuplestore and nodeMaterial so that we don't have to read the
entire contents of the subplan into the tuplestore before we can return
any tuples.  Instead, the tuplestore holds what we've already read, and
we fetch additional rows from the subplan as needed.  Random access to
the previously-read rows works with the tuplestore, and doesn't affect
the state of the partially-read subplan.  This is a step towards fixing
the problems with cursors over complex queries --- we don't want to
stick in Materialize nodes if they'll prevent quick startup for a cursor.

21 years agoApplied patch from Paul Sorenson to correctly handle schema names in updateable resul...
Barry Lind [Sat, 8 Mar 2003 06:06:55 +0000 (06:06 +0000)]
Applied patch from Paul Sorenson to correctly handle schema names in updateable result sets.
Applied patch from Rich Cullingford to fix a NPE in the absolute() method of result set.
Applied patch from Tarjei Skorgenes to fix a NPE when logging is enabled.

 Modified Files:
  jdbc/org/postgresql/core/BaseResultSet.java
  jdbc/org/postgresql/jdbc1/AbstractJdbc1ResultSet.java
  jdbc/org/postgresql/jdbc2/Array.java
  jdbc/org/postgresql/util/PSQLException.java

21 years agoNot done:
Bruce Momjian [Sat, 8 Mar 2003 04:23:05 +0000 (04:23 +0000)]
Not done:

> * Precompile SQL functions to avoid overhead (Neil)

21 years agoReorder items and mark some as completed.
Bruce Momjian [Sat, 8 Mar 2003 03:03:49 +0000 (03:03 +0000)]
Reorder items and mark some as completed.

> * Allow savepoints / nested transactions [transactions] (Bruce)
215d210
<  o Add GUC parameter to control the maximum number of rewrite cycles
227,228c222
<  o Allow parameters to be specified by name and type during
<    definition
>  o Allow parameters to be specified by name and type during definition
304,305d297
< * Overhaul bufmgr/lockmgr/transaction manager
< * Allow savepoints / nested transactions [transactions] (Bruce)
386,387c378,379
< * Add checkpoint_min_warning postgresql.conf option to warn about checkpoints
<   that are too frequent
> * -Add checkpoint_min_warning postgresql.conf option to warn about checkpoints
>   that are too frequent (Bruce)
390d381
< * Allow pg_xlog to be moved without symlinks
406c397
< * Precompile SQL functions to avoid overhead (Neil)
> * -Precompile SQL functions to avoid overhead (Neil)

21 years agoCleanup and reorganization.
Barry Lind [Fri, 7 Mar 2003 18:39:46 +0000 (18:39 +0000)]
Cleanup and reorganization.
  - Added a private api layer (org.postgresql.core.Base*)
  - Cleaned up public api (org.postgresql.PG*)
  - Added consistent headers and copywrite info
  - Removed deprecated Serialize functionality
  - Cleaned up imports
  - Moved some files to more appropriate locations

 Modified Files:
  jdbc/org/postgresql/Driver.java.in
  jdbc/org/postgresql/PGConnection.java
  jdbc/org/postgresql/PGNotification.java
  jdbc/org/postgresql/PGStatement.java
  jdbc/org/postgresql/core/Encoding.java
  jdbc/org/postgresql/core/Notification.java
  jdbc/org/postgresql/core/QueryExecutor.java
  jdbc/org/postgresql/core/StartupPacket.java
  jdbc/org/postgresql/fastpath/Fastpath.java
  jdbc/org/postgresql/fastpath/FastpathArg.java
  jdbc/org/postgresql/geometric/PGbox.java
  jdbc/org/postgresql/geometric/PGcircle.java
  jdbc/org/postgresql/geometric/PGline.java
  jdbc/org/postgresql/geometric/PGlseg.java
  jdbc/org/postgresql/geometric/PGpath.java
  jdbc/org/postgresql/geometric/PGpoint.java
  jdbc/org/postgresql/geometric/PGpolygon.java
  jdbc/org/postgresql/jdbc1/AbstractJdbc1Connection.java
  jdbc/org/postgresql/jdbc1/AbstractJdbc1DatabaseMetaData.java
  jdbc/org/postgresql/jdbc1/AbstractJdbc1ResultSet.java
  jdbc/org/postgresql/jdbc1/AbstractJdbc1ResultSetMetaData.java
  jdbc/org/postgresql/jdbc1/AbstractJdbc1Statement.java
  jdbc/org/postgresql/jdbc1/Jdbc1CallableStatement.java
  jdbc/org/postgresql/jdbc1/Jdbc1Connection.java
  jdbc/org/postgresql/jdbc1/Jdbc1DatabaseMetaData.java
  jdbc/org/postgresql/jdbc1/Jdbc1PreparedStatement.java
  jdbc/org/postgresql/jdbc1/Jdbc1ResultSet.java
  jdbc/org/postgresql/jdbc1/Jdbc1ResultSetMetaData.java
  jdbc/org/postgresql/jdbc1/Jdbc1Statement.java
  jdbc/org/postgresql/jdbc2/AbstractJdbc2Blob.java
  jdbc/org/postgresql/jdbc2/AbstractJdbc2Clob.java
  jdbc/org/postgresql/jdbc2/AbstractJdbc2Connection.java
  jdbc/org/postgresql/jdbc2/AbstractJdbc2DatabaseMetaData.java
  jdbc/org/postgresql/jdbc2/AbstractJdbc2ResultSet.java
  jdbc/org/postgresql/jdbc2/AbstractJdbc2ResultSetMetaData.java
  jdbc/org/postgresql/jdbc2/AbstractJdbc2Statement.java
  jdbc/org/postgresql/jdbc2/Array.java
  jdbc/org/postgresql/jdbc2/Jdbc2CallableStatement.java
  jdbc/org/postgresql/jdbc2/Jdbc2Connection.java
  jdbc/org/postgresql/jdbc2/Jdbc2PreparedStatement.java
  jdbc/org/postgresql/jdbc2/Jdbc2ResultSet.java
  jdbc/org/postgresql/jdbc2/Jdbc2ResultSetMetaData.java
  jdbc/org/postgresql/jdbc2/Jdbc2Statement.java
  jdbc/org/postgresql/jdbc3/AbstractJdbc3ResultSet.java
  jdbc/org/postgresql/jdbc3/Jdbc3CallableStatement.java
  jdbc/org/postgresql/jdbc3/Jdbc3Connection.java
  jdbc/org/postgresql/jdbc3/Jdbc3PreparedStatement.java
  jdbc/org/postgresql/jdbc3/Jdbc3ResultSet.java
  jdbc/org/postgresql/jdbc3/Jdbc3ResultSetMetaData.java
  jdbc/org/postgresql/jdbc3/Jdbc3Statement.java
  jdbc/org/postgresql/largeobject/BlobInputStream.java
  jdbc/org/postgresql/largeobject/BlobOutputStream.java
  jdbc/org/postgresql/largeobject/LargeObject.java
  jdbc/org/postgresql/largeobject/LargeObjectManager.java
  jdbc/org/postgresql/test/jdbc2/Jdbc2TestSuite.java
  jdbc/org/postgresql/test/jdbc2/optional/BaseDataSourceTest.java
  jdbc/org/postgresql/util/MD5Digest.java
  jdbc/org/postgresql/util/MessageTranslator.java
  jdbc/org/postgresql/util/PGbytea.java
  jdbc/org/postgresql/util/PGmoney.java
  jdbc/org/postgresql/util/PGobject.java
  jdbc/org/postgresql/util/PGtokenizer.java
  jdbc/org/postgresql/util/PSQLException.java
  jdbc/org/postgresql/util/UnixCrypt.java
 Added Files:
  jdbc/org/postgresql/core/BaseConnection.java
  jdbc/org/postgresql/core/BaseResultSet.java
  jdbc/org/postgresql/core/BaseStatement.java
  jdbc/org/postgresql/core/Field.java
  jdbc/org/postgresql/core/PGStream.java
 Removed Files:
  jdbc/org/postgresql/Field.java
  jdbc/org/postgresql/PG_Stream.java
  jdbc/org/postgresql/test/jdbc2/SerializeObject.java
  jdbc/org/postgresql/test/jdbc2/SerializeTest.java
  jdbc/org/postgresql/util/Serialize.java

21 years agoAdd to mmap emails.
Bruce Momjian [Fri, 7 Mar 2003 17:43:26 +0000 (17:43 +0000)]
Add to mmap emails.

21 years agoUpdate FAQ.
Bruce Momjian [Fri, 7 Mar 2003 05:49:10 +0000 (05:49 +0000)]
Update FAQ.

21 years agoUpdate Russian FAQ, from Viktor Vislobokov.
Bruce Momjian [Fri, 7 Mar 2003 05:46:51 +0000 (05:46 +0000)]
Update Russian FAQ, from Viktor Vislobokov.

21 years agoTweak dependency code to suppress NOTICEs generated by new method for
Tom Lane [Thu, 6 Mar 2003 22:54:49 +0000 (22:54 +0000)]
Tweak dependency code to suppress NOTICEs generated by new method for
cleaning out temp namespaces.  We don't really want the server log to be
cluttered with 'Drop cascades to table foo' every time someone uses a
temp table...

21 years agoAdd missing --globals-only (long form of -g).
Tom Lane [Thu, 6 Mar 2003 21:45:52 +0000 (21:45 +0000)]
Add missing --globals-only (long form of -g).

21 years agoAdd:
Bruce Momjian [Thu, 6 Mar 2003 19:53:07 +0000 (19:53 +0000)]
Add:

>  o Report server version number, database encoding, client encoding

21 years agoAdd for protocol changes:
Bruce Momjian [Thu, 6 Mar 2003 19:52:20 +0000 (19:52 +0000)]
Add for protocol changes:

>  o Report server version number

21 years agoRemove:
Bruce Momjian [Thu, 6 Mar 2003 18:39:18 +0000 (18:39 +0000)]
Remove:

< * Require DROP COLUMN CASCADE for a column that is part of a multi-column index

21 years agoDone:
Bruce Momjian [Thu, 6 Mar 2003 18:05:00 +0000 (18:05 +0000)]
Done:

> * -Modify regression tests to prevent failures do to minor numeric rounding

21 years agoUse poll(2) in preference to select(2), if available. This solves
Tom Lane [Thu, 6 Mar 2003 03:16:55 +0000 (03:16 +0000)]
Use poll(2) in preference to select(2), if available.  This solves
problems in applications that may have a large number of files open,
such that libpq's socket number exceeds the range supported by fd_set.
From Chris Brown.

21 years agoAdd:
Bruce Momjian [Thu, 6 Mar 2003 02:29:44 +0000 (02:29 +0000)]
Add:

>  o Have ALTER TABLE rename SERIAL sequences

21 years agoAdd code to dump contents of free space map into $PGDATA/global/pg_fsm.cache
Tom Lane [Thu, 6 Mar 2003 00:04:27 +0000 (00:04 +0000)]
Add code to dump contents of free space map into $PGDATA/global/pg_fsm.cache
at database shutdown, and then load it again at database startup.  This
preserves our hard-won knowledge of free space across restarts (given
an orderly shutdown, that is).

21 years agoTurns out new IN implementation has got some problems in an UPDATE or
Tom Lane [Wed, 5 Mar 2003 20:01:04 +0000 (20:01 +0000)]
Turns out new IN implementation has got some problems in an UPDATE or
DELETE with inherited target table.  Fix it; add a regression test.
Also, correct ancient misspelling of 'inherited'.

21 years agoRepair bug reported by Laurent Perez: bad plan generated when UPDATE or
Tom Lane [Wed, 5 Mar 2003 18:38:14 +0000 (18:38 +0000)]
Repair bug reported by Laurent Perez: bad plan generated when UPDATE or
DELETE of an inheritance tree references another inherited relation.
This bug has been latent since 7.1; I'm still not quite sure why 7.1 and
7.2 don't manifest it (at least, they don't crash on a simple test case).

21 years agoReimplement free-space-map management as per recent discussions.
Tom Lane [Tue, 4 Mar 2003 21:51:22 +0000 (21:51 +0000)]
Reimplement free-space-map management as per recent discussions.
Adjustable threshold is gone in favor of keeping track of total requested
page storage and doling out proportional fractions to each relation
(with a minimum amount per relation, and some quantization of the results
to avoid thrashing with small changes in page counts).  Provide special-
case code for indexes so as not to waste space storing useless page
free space counts.  Restructure internal data storage to be a flat array
instead of list-of-chunks; this may cost a little more work in data
copying when reorganizing, but allows binary search to be used during
lookup_fsm_page_entry().

21 years agoPrevent clustering on incomplete indexes: partial indexes are verboten,
Tom Lane [Mon, 3 Mar 2003 04:37:37 +0000 (04:37 +0000)]
Prevent clustering on incomplete indexes: partial indexes are verboten,
as are non-amindexnulls AMs unless first column is attnotnull.

21 years agoImprove description of geometric operators.
Tom Lane [Mon, 3 Mar 2003 03:31:23 +0000 (03:31 +0000)]
Improve description of geometric operators.

21 years agoCorrect/edit comments for geometric operators. No change except to
Tom Lane [Mon, 3 Mar 2003 03:30:46 +0000 (03:30 +0000)]
Correct/edit comments for geometric operators.  No change except to
pg_description entries, so I did not force initdb.

21 years agoBe smart about outer-join qualifications that mention only one side of
Tom Lane [Sun, 2 Mar 2003 23:46:34 +0000 (23:46 +0000)]
Be smart about outer-join qualifications that mention only one side of
the join, per recent discussion on pgsql-sql.  Not clear that this will
come up often in real queries, but it's not any more expensive to do it
right, so we may as well do it right.

21 years agoRepair memory leak introduced by recent change to make SPI return a
Tom Lane [Sun, 2 Mar 2003 20:45:47 +0000 (20:45 +0000)]
Repair memory leak introduced by recent change to make SPI return a
tupdesc even with zero tuples returned: some plpgsql routines assumed
they didn't need to do SPI_freetuptable() after retrieving no tuples.

21 years agoChange EXTRACT(EPOCH FROM timestamp) so that a timestamp without time zone
Tom Lane [Thu, 27 Feb 2003 21:36:58 +0000 (21:36 +0000)]
Change EXTRACT(EPOCH FROM timestamp) so that a timestamp without time zone
is assumed to be in local time, not GMT.  This improves consistency with
other operations, which all assume local timezone when it matters.  Per
bug #897.

21 years agoup build number to 202 for a new development build
Barry Lind [Thu, 27 Feb 2003 05:58:08 +0000 (05:58 +0000)]
up build number to 202 for a new development build

 Modified Files:
  jdbc/org/postgresql/Driver.java.in

21 years agoA little cleanup. Removing an unnecessary method.
Barry Lind [Thu, 27 Feb 2003 05:56:27 +0000 (05:56 +0000)]
A little cleanup.  Removing an unnecessary method.

Modified Files:
jdbc/org/postgresql/jdbc2/AbstractJdbc2ResultSet.java

21 years agoAdded support for SSL in the jdbc driver
Barry Lind [Thu, 27 Feb 2003 05:45:44 +0000 (05:45 +0000)]
Added support for SSL in the jdbc driver

 Modified Files:
  jdbc/build.xml jdbc/org/postgresql/Driver.java.in
  jdbc/org/postgresql/PG_Stream.java
  jdbc/org/postgresql/errors.properties
  jdbc/org/postgresql/jdbc1/AbstractJdbc1Connection.java
  jdbc/org/postgresql/util/PSQLException.java

21 years agoThe quote function should return a string suitable for dropping into a
D'Arcy J.M. Cain [Wed, 26 Feb 2003 13:16:50 +0000 (13:16 +0000)]
The quote function should return a string suitable for dropping into a
query string.  This fixes a bug where bool types sometimes returned with
a string that could not be dropped into a query.

21 years agoRemove REWRITE_INVOKE_MAX in favor of making an accurate check for
Tom Lane [Tue, 25 Feb 2003 23:47:43 +0000 (23:47 +0000)]
Remove REWRITE_INVOKE_MAX in favor of making an accurate check for
recursion in RewriteQuery(); also, detect recursion in fireRIRrules(),
so as to catch self-referential views per example from Ryan VanderBijl.
Minor code restructuring to make it easier to catch recursive case.

21 years agoAllow SET CONNECTION to be followed by connection object without leading "TO" or...
Michael Meskes [Tue, 25 Feb 2003 15:58:03 +0000 (15:58 +0000)]
Allow SET CONNECTION to be followed by connection object without leading "TO" or "=".
Allow whenever statement to list function without parameters.

21 years agoAvoid using a separate query to fetch the default expressions for
Tom Lane [Mon, 24 Feb 2003 03:54:06 +0000 (03:54 +0000)]
Avoid using a separate query to fetch the default expressions for
columns.  Improved version of patch by mallah@trade-india.com.

21 years agoDuring VACUUM FULL, truncate off any deletable pages that are at the
Tom Lane [Mon, 24 Feb 2003 00:57:17 +0000 (00:57 +0000)]
During VACUUM FULL, truncate off any deletable pages that are at the
end of a btree index.  This isn't super-effective, since we won't move
nondeletable pages, but it's better than nothing.  Also, improve stats
displayed during VACUUM VERBOSE.

21 years agoRemove no-longer-used FixBTree GUC variable.
Tom Lane [Sun, 23 Feb 2003 23:27:21 +0000 (23:27 +0000)]
Remove no-longer-used FixBTree GUC variable.

21 years agobtree page recycling can be done as soon as page's next-xact label is
Tom Lane [Sun, 23 Feb 2003 23:20:52 +0000 (23:20 +0000)]
btree page recycling can be done as soon as page's next-xact label is
older than current Xmin; we don't have to wait till it's older than
GlobalXmin.

21 years agoAdjust btbulkdelete logic so that only one WAL record is issued while
Tom Lane [Sun, 23 Feb 2003 22:43:09 +0000 (22:43 +0000)]
Adjust btbulkdelete logic so that only one WAL record is issued while
deleting multiple index entries on a single index page.  This makes for
a very substantial reduction in the amount of WAL traffic during a
large delete operation.

21 years agoImprove coding of log_heap_clean() and heap_xlog_clean().
Tom Lane [Sun, 23 Feb 2003 20:32:12 +0000 (20:32 +0000)]
Improve coding of log_heap_clean() and heap_xlog_clean().

21 years agoFirst cut at recycling space in btree indexes. Still some rough edges
Tom Lane [Sun, 23 Feb 2003 06:17:13 +0000 (06:17 +0000)]
First cut at recycling space in btree indexes.  Still some rough edges
to fix, but it seems to basically work...

21 years agoIf a shutdown request comes in while we're still starting up, don't
Tom Lane [Sun, 23 Feb 2003 04:48:19 +0000 (04:48 +0000)]
If a shutdown request comes in while we're still starting up, don't
service it until after we execute SetThisStartUpID().  Else shutdown
process will write the wrong SUI into the shutdown checkpoint, which
seems likely to be trouble --- although I've not quite figured out
how significant it really is.

21 years agoSimplify timezone-handling code per proposal to pghackers: get rid of
Tom Lane [Sat, 22 Feb 2003 05:57:45 +0000 (05:57 +0000)]
Simplify timezone-handling code per proposal to pghackers: get rid of
setting timezone-related variables during transaction start.  They were
not used anyway in platforms that HAVE_TM_ZONE or HAVE_INT_TIMEZONE,
which it appears is *all* the platforms we are currently supporting.
For platforms that have neither, we now only support UTC or numeric-
offset-from-UTC timezones.

21 years agoMore infrastructure for btree compaction project. Tree-traversal code
Tom Lane [Sat, 22 Feb 2003 00:45:05 +0000 (00:45 +0000)]
More infrastructure for btree compaction project.  Tree-traversal code
now knows what to do upon hitting a dead page (in theory anyway, it's
untested...).  Add a post-VACUUM-cleanup entry point for index AMs, to
provide a place for dead-page scavenging to happen.
Also, fix oversight that broke btpo_prev links in temporary indexes.
initdb forced due to additions in pg_am.

21 years agoRevert patch that broke \d commands, until it can be fixed.
Tom Lane [Fri, 21 Feb 2003 21:34:27 +0000 (21:34 +0000)]
Revert patch that broke \d commands, until it can be fixed.

21 years agoForgot to commit Makefile change.
Michael Meskes [Fri, 21 Feb 2003 15:19:18 +0000 (15:19 +0000)]
Forgot to commit Makefile change.

21 years agoMake btree index structure adjustments and WAL logging changes needed to
Tom Lane [Fri, 21 Feb 2003 00:06:22 +0000 (00:06 +0000)]
Make btree index structure adjustments and WAL logging changes needed to
support btree compaction, as per proposal of a few days ago.  btree index
pages no longer store parent links, instead they have a level indicator
(counting up from zero for leaf pages).  The FixBTree recovery logic is
removed, and replaced by code that detects missing parent-level insertions
during WAL replay.  Also, generate appropriate WAL entries when updating
btree metapage and when building a btree index from scratch.  I believe
btree indexes are now completely WAL-legal for the first time.
initdb forced due to index and WAL changes.

21 years agoFix timestamptz_in so that parsing of 'now'::timestamptz gives right
Tom Lane [Thu, 20 Feb 2003 05:24:55 +0000 (05:24 +0000)]
Fix timestamptz_in so that parsing of 'now'::timestamptz gives right
answer when SET TIMEZONE has been done since the start of the current
transaction.  Per bug report from Robert Haas.
I plan some futher cleanup in HEAD, but this is a low-risk patch for
the immediate issue in 7.3.

21 years agoBack out LOCKTAG changes by Rod Taylor, pending code review. Sorry.
Bruce Momjian [Wed, 19 Feb 2003 23:41:15 +0000 (23:41 +0000)]
Back out LOCKTAG changes by Rod Taylor, pending code review. Sorry.

21 years agoFix for GUC client_encoding variable not being handled
Tatsuo Ishii [Wed, 19 Feb 2003 14:31:26 +0000 (14:31 +0000)]
Fix for GUC client_encoding variable not being handled
correctly. See following thread for more details.

Subject: [HACKERS] client_encoding directive is ignored in postgresql.conf
From: Tatsuo Ishii <t-ishii@sra.co.jp>
Date: Wed, 29 Jan 2003 22:24:04 +0900 (JST)

21 years agoAdded DATABASE command as alias to CONNECT TO.
Michael Meskes [Wed, 19 Feb 2003 12:36:39 +0000 (12:36 +0000)]
Added DATABASE command as alias to CONNECT TO.

21 years agoThis patch makes the following changes to the documentation:
Bruce Momjian [Wed, 19 Feb 2003 04:06:28 +0000 (04:06 +0000)]
This patch makes the following changes to the documentation:

- more work from the SGML police

- some grammar improvements: rewriting a paragraph or two, replacing
contractions where (IMHO) appropriate

- fix missing utility commands in lock mode docs

- improve CLUSTER, REINDEX, SET SESSION AUTHORIZATION ref pages

Neil Conway

21 years agoAutoconf run for readline/libedit patch.
Bruce Momjian [Wed, 19 Feb 2003 04:05:07 +0000 (04:05 +0000)]
Autoconf run for readline/libedit patch.

21 years agoHere's the patch I promised over on HACKERS - it alters the
Bruce Momjian [Wed, 19 Feb 2003 04:04:04 +0000 (04:04 +0000)]
Here's the patch I promised over on HACKERS - it alters the
implementation
of '\e' history tracking for systems that have a readline compatability
library without replace_history_entry.  I fall back to pushing the query
onto the history stack after the \e, rather than replacing it.

The patch adds one more place to look for readline headers, and a test
for replace_history_entry. I've only included the patch for configure.in

Ross J. Reedstrom

21 years ago- Modifies LOCKTAG to include a 'classId'. Relation receive a classId of
Bruce Momjian [Wed, 19 Feb 2003 04:02:54 +0000 (04:02 +0000)]
- Modifies LOCKTAG to include a 'classId'. Relation receive a classId of
RelOid_pg_class, and transaction locks XactLockTableId. RelId is renamed
to objId.

- LockObject() and UnlockObject() functions created, and their use
sprinkled throughout the code to do descent locking for domains and
types. They accept lock modes AccessShare and AccessExclusive, as we
only really need a 'read' and 'write' lock at the moment.  Most locking
cases are held until the end of the transaction.

This fixes the cases Tom mentioned earlier in regards to locking with
Domains.  If the patch is good, I'll work on cleaning up issues with
other database objects that have this problem (most of them).

Rod Taylor

21 years agoAllow PQcmdTuples to return row counts for MOVE and FETCH.
Bruce Momjian [Wed, 19 Feb 2003 03:59:02 +0000 (03:59 +0000)]
Allow PQcmdTuples to return row counts for MOVE and FETCH.

Neil Conway

21 years agoHere's some changes I made last night to psql's common.c (as found in
Bruce Momjian [Wed, 19 Feb 2003 03:54:39 +0000 (03:54 +0000)]
Here's some changes I made last night to psql's common.c (as found in
7.3.2).  It removes some code duplication and #ifdeffing, and some
unstructured ugliness such as tacky breaks and an unneeded continue.
Breaks up a large function into smaller functions and reduces required
nesting levels, and kills a variable or two.

Jeroen T. Vermeulen

21 years agoAt present it is impossible to kill createuser except with kill -9
Bruce Momjian [Wed, 19 Feb 2003 03:52:57 +0000 (03:52 +0000)]
At present it is impossible to kill createuser except with kill -9

Oliver Elphick

21 years agoPlease apply patches for contrib/ltree.
Bruce Momjian [Wed, 19 Feb 2003 03:50:09 +0000 (03:50 +0000)]
Please apply patches for contrib/ltree.

ltree_73.patch.gz - for 7.3 :
        Fix ~ operation bug: eg '1.1.1' ~ '*.1'

ltree_74.patch.gz - for current CVS
    Fix ~ operation bug: eg '1.1.1' ~ '*.1'
    Add ? operation
    Optimize index storage

Last change needs drop/create all ltree indexes, so only for 7.4

Teodor Sigaev

21 years agoThe following patches eliminate the overflows in the j2date() and date2j()
Bruce Momjian [Wed, 19 Feb 2003 03:48:11 +0000 (03:48 +0000)]
The following patches eliminate the overflows in the j2date() and date2j()
functions which limited the maximum date for a timestamp to AD 1465001.
The new limit is AD 5874897.
The files affected are:

doc/src/sgml/datatype.sgml:
    Documentation change due to patch. Included is a notice about
    the reduced range when using an eight-byte integer for timestamps.

src/backend/utils/adt/datetime.c:
    Replacement functions for j2date() and date2j() functions.

src/include/utils/datetime.h:
    Corrected a bug with the limit on the earliest possible date,
    Nov 23,-4713 has a Julian day count of -1. The earliest possible
    date should be Nov 24, -4713 with a day count of 0.

src/test/regress/expected/horology-no-DST-before-1970.out:
src/test/regress/expected/horology-solaris-1947.out:
src/test/regress/expected/horology.out:
    Copies of expected output for regression testing.
    Note: Only horology.out has been physically tested. I do not have access
    to a Solaris box and I don't know how to provoke the "pre-1970" test.

src/test/regress/sql/horology.sql:
    Added some test cases to check extended range.

John Cochran

21 years agoWe just released new version of contrib/btree_gist
Bruce Momjian [Wed, 19 Feb 2003 03:46:00 +0000 (03:46 +0000)]
We just released new version of contrib/btree_gist
(7.3 and current CVS) with support of int8, float4, float8
in addition to int4. Thanks Janko Richter for contribution.

Oleg Bartunov

21 years agoThis patch includes more SGML markup fixes as well as a few minor
Bruce Momjian [Wed, 19 Feb 2003 03:13:25 +0000 (03:13 +0000)]
This patch includes more SGML markup fixes as well as a few minor
additions to the docs.

Neil Conway

21 years agoThis patch adds a note to the documentation describing why the
Bruce Momjian [Wed, 19 Feb 2003 03:12:22 +0000 (03:12 +0000)]
This patch adds a note to the documentation describing why the
performance of min() and max() is slow when applied to the entire table,
and suggesting the simple workaround most experienced Pg users
eventually learn about (SELECT xyz ... ORDER BY xyz LIMIT 1).

Neil Conway