Tom Lane [Fri, 15 Nov 2002 17:18:49 +0000 (17:18 +0000)]
Add relcache support for system catalogs with attnotnull columns.
The code was not making TupleConstr structs for such catalogs in
several places; with the consequence that the not-null constraint
wasn't actually enforced. With this change,
INSERT INTO pg_proc VALUES('sdf');
generates a 'Fail to add null value' error instead of a core dump.
Bruce Momjian [Fri, 15 Nov 2002 03:09:39 +0000 (03:09 +0000)]
New version attached. The following is implemented:
- CLUSTER ALL clusters all the tables that have some index with
indisclustered set and the calling user owns.
- CLUSTER tablename clusters the named table, using the index with
indisclustered set. If no index has the bit set, throws elog(ERROR).
- The multi-relation version (CLUSTER ALL) uses a multitransaction
approach, similar to what VACUUM does.
Bruce Momjian [Fri, 15 Nov 2002 03:07:52 +0000 (03:07 +0000)]
Clean up psql:
Add simple ALTER DATABASE, ALTER TRIGGER, CHECK POINT, CREATE
CONVERSION, CREATE DOMAIN, CREATE LANGUAGE, DEALLOCATE, DROP CONVERSION,
DROP DOMAIN, DROP LANGUAGE, EXECUTE, PREPARE
Complete CAST in CREATE CAST and DROP CAST but doesn't suggest what
should follow.
Add many more SET / SHOW variables to the list. Taken from SHOW ALL
output.
Complete a case sensitive search to allow \dD, \dd, \dS, \ds, \h, \H to
complete properly. But there are no matches, then try a case
insensitive search to allow case conversion. Add all missing help
options.
Tom Lane [Thu, 14 Nov 2002 19:00:36 +0000 (19:00 +0000)]
Fix planning bug introduced in recent code reorganization for hashed
aggregates: tuple_fraction has to be adjusted before passing it to
compare_fractional_path_costs().
Barry Lind [Thu, 14 Nov 2002 05:35:45 +0000 (05:35 +0000)]
Applied patch submitted by Mats Lofkvist fixing serious threading problem introduced in beta3.
Fixed bug with using setNull()(or setXXX(x, null)) and serverside prepare statements.
Improved error message when using a connection object that has already been closed.
Tom Lane [Tue, 12 Nov 2002 00:39:08 +0000 (00:39 +0000)]
Remove NO_MKTIME_BEFORE_1970. I had speculated that it was not needed
anymore given the mktime() workaround now done in DetermineLocalTimeZone.
This has now been confirmed by Robert Bruccoleri for Irix, and I'm going
to extrapolate to AIX as well.
Tom Lane [Mon, 11 Nov 2002 22:19:25 +0000 (22:19 +0000)]
Code review for ON COMMIT patch. Make the actual on-commit action happen
before commit, not after :-( --- the original coding is not only unsafe
if an error occurs while it's processing, but it generates an invalid
sequence of WAL entries. Resurrect 7.2 logic for deleting items when
no longer needed. Use an enum instead of random macros. Editorialize
on names used for routines and constants. Teach backend/nodes routines
about new field in CreateTable struct. Add a regression test.
Bruce Momjian [Sun, 10 Nov 2002 00:37:28 +0000 (00:37 +0000)]
I am attaching a patch file that will replace the previous one. It only
contains the patches to Makefile.global.in and Makefile.unixware. The
Makefile.unixware patch has been updated to include the contents of
LD_LIBRARY_PATH, if present, to the -rpath (-R) option. This change
will simplify configuring and building PostgreSQL on systems that
support LD_LIBRARY_PATH. You can set LD_LIBRARY_PATH to include all
the directorys you want to have searched for additional libraries, run
configure, then run make. The paths in LD_LIBRARY_PATH will then be
embedded in the executables via the -rpath (-R) option to the linker,
and so will not require LD_LIBRARY_PATH in order to run.
Bruce Momjian [Sun, 10 Nov 2002 00:35:58 +0000 (00:35 +0000)]
This patch makes a minor cleanup to the implementation of PERFORM in
PL/PgSQL. Previously, it had been bundled together with the assign
statement implementation, for some reason that wasn't clear to me
(they certainly don't share any code with one another). So I separated
them and made PERFORM a statement like any other. No changes in
functionality.
Along the way, I added some regression tests for PERFORM, added a
bunch more SGML tags to the PL/PgSQL docs, and removed an obsolete
comment relating to the implementation of RETURN NEXT.
Bruce Momjian [Sun, 10 Nov 2002 00:33:43 +0000 (00:33 +0000)]
> > I'll re-check that with the ppc architecture guy here.
>
> ... he is now about to write an inlined version that can go into
> s_lock.h . I'll send the new patch later on...
OK, here it comes:
An inlined version of tas(), that works for both, powerpc and
powerpc64. The patch is against 7.3b5 and passes the test suite on
both architectures.
Bruce Momjian [Sun, 10 Nov 2002 00:14:22 +0000 (00:14 +0000)]
I just discovered, that there is missing a const when passing a buffer
to PQescapeBytea and PQunescapeBytea. I fixed it and tried to create a
usable diff (I'm not so familar to diff).
Bruce Momjian [Fri, 8 Nov 2002 20:23:57 +0000 (20:23 +0000)]
This patch removes a bunch of superfluous #include directives: if
postgres.h or c.h includes a system header (such as stdio.h or
stdlib.h), there's no need to specifically include it in any of the .c
files in the backend.
Bruce Momjian [Fri, 8 Nov 2002 20:20:22 +0000 (20:20 +0000)]
The attached patch defines functions for getting distances between
points on the surface of the earth and locating points within a
specified distance using an index based on the contrib/cube package. The
new functions are all of language type sql. A couple of bugs in the old
earthdistance function based on the point datatype are fixed. A
regression test has been added for both sets of functions. The README
file has been updated to include documentation on the new stuff. There
are comments about how this package is also useful for Astronomers.
Tom Lane [Fri, 8 Nov 2002 20:09:55 +0000 (20:09 +0000)]
Revise geometry regression testing to eliminate most cross-platform
variation. To do this, set extra_float_digits to -3 in the geometry
test, and tweak the CIRCLE_TBL dataset to avoid values that suffer
from severe cancellation error (eg, circles that just touch an axis).
We still need two geometry 'expected' files to account for the
difference between platforms that display minus zero as '-0' and those
that just say '0', but with luck that's all we'll need.
Bruce Momjian [Fri, 8 Nov 2002 19:12:21 +0000 (19:12 +0000)]
Here is a patch that does just that, while maintaining the
"traditional" behavior, so the change should be transparent. Use the
command "\pset pager always" to turn it on. Anything else does the
normal toggle between "on" and "off"
Tom Lane [Fri, 8 Nov 2002 18:32:47 +0000 (18:32 +0000)]
Replace imprecise value of PI with a better one, and tweak circle_poly
in hopes of reducing platform-to-platform variations in its results.
This will cause the geometry regression test to start failing on some
platforms. I plan to update the test later today.
Tom Lane [Fri, 8 Nov 2002 17:37:52 +0000 (17:37 +0000)]
Add extra_float_digits GUC parameter to allow adjustment of displayed
precision for float4, float8, and geometric types. Set it in pg_dump
so that float data can be dumped/reloaded exactly (at least on platforms
where the float I/O support is properly implemented). Initial patch by
Pedro Ferreira, some additional work by Tom Lane.
Bruce Momjian [Fri, 8 Nov 2002 17:27:03 +0000 (17:27 +0000)]
The "Allow easy display of usernames in a group (pg_hba.conf uses groups
now)" item on the open items, and subsequent plpgsql function I sent in,
made me realize it was too hard to get the upper and lower bound of an
array. The attached creates two functions that I think will be very
useful when combined with the ability of plpgsql to return sets.
array_lower(array, dim_num)
- and -
array_upper(array, dim_num)
They return the value (as an int) of the upper and lower bound of the
requested dim in the provided array.
Bruce Momjian [Fri, 8 Nov 2002 04:56:40 +0000 (04:56 +0000)]
I'm using libpgtcl with PostgreSQL-7.2.3. There is a discrepancy between
the documentation and behavior of "pg_result $res -oid" when $res is not
from an INSERT. The documentation says it should return an empty
string, but in fact it returns 0. I think it used to return an empty
string around PostgreSQL-7.1.3 when PQoidStatus() was used, but now it
uses PQoidValue() which returns InvalidOid, 0 in this case.
Assuming the current behavior is desired, here is a patch to the
documentation doc/src/sgml/libpgtcl.sgml to match what really happens:
Bruce Momjian [Fri, 8 Nov 2002 04:31:14 +0000 (04:31 +0000)]
Add:
> * Add GUC variables extra_float_digits and extra_double_digits to
> control output digits 419a422
> * Research interaction of setitimer() and sleep() used by statement_timeout
Tom Lane [Thu, 7 Nov 2002 22:08:30 +0000 (22:08 +0000)]
Move substitute extern declarations to end of file, so that they are
not read until after we've read the port-specific header file. In
particular this should make it safer to #include system headers for
inet_aton; in general it seems that the port header file ought to be
in a position to set definitions before we do stuff based on having
a definition or not.
Bruce Momjian [Wed, 6 Nov 2002 05:55:14 +0000 (05:55 +0000)]
Remove completed items:
< A dash (-) marks changes that will appear in the upcoming 7.3 release.
> A dash (-) marks changes that will appear in the upcoming 7.4 release.
38d37
< * -Change DEBUG startup tag to LOG (Bruce)
40,51d38
< * -Add pg_backend_pid() function to backend
< * -Allow logging of query durations
<
<
< Permissions
< ===========
<
< * -Improve control over user privileges, including table creation
< * -Allow user/group names to be specified directly in pg_hba.conf (Bruce)
< * -Add ~/.pgpass to store passwords with user/host/password combinations
< * -Allow permissions for functions (Peter E)
< * -Allow object creation to be disabled for specific users
58,60d44
< * -Make it easier to create a database owned by someone who can't createdb,
< perhaps CREATE DATABASE dbname WITH OWNER = "user" (Gavin)
< * -Make equals sign optional in CREATE DATABASE WITH param = 'val'
65,66d48
< * -Prevent SIGHUP and 'pg_ctl reload' from changing command line
< specified parameters to postgresql.conf defaults (Peter E)
69,70d50
< * -Reserve last few process slots for super-user if max_connections reached
< * -Add GUC parameter to print queries that generate errors
79d58
< * -Add domain capability (Rod Taylor)
82,84d60
< * -SELECT cash_out(2) crashes because of opaque
< * -Declare typein/out functions in pg_proc with a special "C string" data type
< * -Functions returning sets do not totally work
89d64
< * -Add GUC parameter for DATESTYLE
91,93d65
< * -Allow bytea to handle LIKE with non-TEXT patterns
< * -to_char(0,'FM999.99') returns a period, to_char(1,'FM999.99') doesn't (Karel)
< * -Add floor(float8) and other missing functions
97d68
< o -Store binary-compatible type information in the system
104d74
< o -Ensure we have array-eq operators for every built-in array type
139,140d108
< * -Remove brackets as multi-statement rule grouping, must use parens (Bruce)
< * -Prevent aggregates from being used in rule WHERE clauses 154d121
< * -Allow UPDATE/DELETE on inherited table 166d132
< * -Add deleted bit to index tuples to reduce heap access 176d141
< * -Test hash index performance and discourage usage 182d146
< * -Add SIMILAR TO to allow character classes, 'pg_[a-c]%' 184d147
< * -Remove LIMIT #,# and force use LIMIT and OFFSET clauses in 7.3 (Bruce)
186,187d148
< * -Disallow TRUNCATE on tables that are involved in referential constraints
< * -Add OR REPLACE clauses to non-FUNCTION object creation 190d150
< * -Prevent create/drop scripts from allowing extra args (Bruce)
201,205d160
< o -Add ALTER TABLE DROP COLUMN feature
< o -Add ALTER TABLE DROP non-CHECK CONSTRAINT
< o -ALTER TABLE ADD PRIMARY KEY (Tom)
< o -ALTER TABLE ADD UNIQUE (Tom)
< o -ALTER TABLE ALTER COLUMN SET/DROP NOT NULL (Christopher) 210d164
< o -Have ALTER TABLE OWNER change all dependant objects like indexes
214,216d167
< o -Cluster all tables at once using pg_index.indisclustered set during
< previous CLUSTER
< o -Prevent loss of indexes, permissions, inheritance 221d171
< o -Allow specification of column names 224d173
< o -Change syntax to WITH DELIMITER, (keep old syntax around?) 228d176
< o -Generate failure on short COPY lines rather than pad NULLs
242,243d189
< o -Allow INSERT INTO my_table VALUES (a, b, c, DEFAULT, x, y, z, ...)
< o -Disallow missing columns in INSERT ... (col) VALUES, per ANSI
248,249d193
< o -Add command to display locks
< o -Add SET or BEGIN timeout parameter to cancel query 251d194
< o -Remove SET KSQO option now that OR processing is improved (Bruce)
254,256d196
< o -Add SHOW command to see locale
< o -Allow SHOW to output as a query result, like EXPLAIN
< o -Abort all SET changes made in an aborted transaction 264d203
< o -Fix PL/PgSQL to handle quoted mixed-case identifiers
287,291d225
< * -Have pg_dump use LEFT OUTER JOIN in multi-table SELECTs
< or multiple SELECTS to avoid bad system catalog entries
< * -Have pg_dump -C dump database location and encoding information
< * -Allow psql \d to show foreign keys
< * -Allow psql \d to show temporary table structure (Tom) 294d227
< * -Have pg_dump use ADD PRIMARY KEY after COPY, for performance (Neil) 302d234
< o -Updateable resultSet 307d238
< o -Implement cancel() method on Statement 309d239
< o -Add support for CallableStatements 311d240
< o -Compile under jdk 1.4 334d262
< * -Allow oid to act as a foreign key
337,338d264
< * -Allow user to control trigger firing order (Tom)
< * -Add ALTER TRIGGER ... RENAME 341d266
< * -Fix foreign key constraints to not error on intermediate db states (Stephan)
350,359d274
< * -Add pg_depend table for dependency recording; use sysrelid, oid,
< depend_sysrelid, depend_oid, name
< * -Auto-destroy sequence on DROP of table with SERIAL; perhaps a separate
< SERIAL type
< * -Prevent column dropping if column is used by foreign key
< * -Propagate column or table renaming to foreign key constraints
< * -Automatically drop constraints/functions when object is dropped
< * -Make foreign key constraints clearer in dump file
< * -Make other constraints clearer in dump file
< * -Make foreign keys easier to identify 367d281
< * -Allow autocommit so always in a transaction block 377d290
< * -Add SQL92 schemas (Tom) 400d312
< * -Cache most recent query plan(s) (Neil) [prepare] 439d350
< * -Cache system catalog information in per-database files (Tom)
450,451d360
< * -Reorder postgresql.conf WAL items in order of importance (Bruce)
< * -Remove wal_files postgresql.conf option because WAL files are now recycled
465,466d373
< * -Improve dynamic memory allocation by introducing tuple-context memory
< allocation (Tom) 468d374
< * -Nested FULL OUTER JOINs don't work (Tom)
470,472d375
< * -Add new pg_proc cachable settings to specify whether function can be
< evaluated only once or once per query
< * -Change FIXED_CHAR_SEL to 0.20 from 0.04 to give better selectivity (Bruce)
494,496d396
< * -Add Intimate Shared Memory(ISM) for Solaris
< * -Use faster flex flags for performance improvement (Peter E)
< * -Add BSD-licensed qsort() for Solaris
503,507d402
< * -Fix problems with libpq non-blocking/async code
< * -Make sure all block numbers are unsigned to increase maximum table size
< * -Merge LockMethodCtl and LockMethodTable into one shared structure (Bruce)
< * -HOLDER/HOLDERTAB rename to PROCLOCK/PROCLOCKTAG (Bruce)
< * -Remove LockMethodTable.prio field, not used (Bruce)
512,513d406
< * -Make one version of simple_prompt() in code (Bruce, Tom)
< * -Compile in syslog functionaility by default (Tatsuo) 517d409
< * -Report failure to find readline or zlib at end of configure run
519,520d410
< * -Increase identifier length (NAMEDATALEN) if small performance hit,
< * -Increase maximum number of function parameters if little wasted space
529,530d418
< * -Fix glibc's mktime() to handle pre-1970's dates
< * -Move /contrib/retep to gborg.postgresql.org
Tom Lane [Wed, 6 Nov 2002 00:00:45 +0000 (00:00 +0000)]
First phase of implementing hash-based grouping/aggregation. An AGG plan
node now does its own grouping of the input rows, and has no need for a
preceding GROUP node in the plan pipeline. This allows elimination of
the misnamed tuplePerGroup option for GROUP, and actually saves more code
in nodeGroup.c than it costs in nodeAgg.c, as well as being presumably
faster. Restructure the API of query_planner so that we do not commit to
using a sorted or unsorted plan in query_planner; instead grouping_planner
makes the decision. (Right now it isn't any smarter than query_planner
was, but that will change as soon as it has the option to select a hash-
based aggregation step.) Despite all the hackery, no initdb needed since
only in-memory node types changed.