]> granicus.if.org Git - postgresql/log
postgresql
21 years agoRemove mention of ALTER USER able to remove passwords.
Bruce Momjian [Fri, 6 Jun 2003 15:31:46 +0000 (15:31 +0000)]
Remove mention of ALTER USER able to remove passwords.

21 years agoImplement outer-level aggregates to conform to the SQL spec, with
Tom Lane [Fri, 6 Jun 2003 15:04:03 +0000 (15:04 +0000)]
Implement outer-level aggregates to conform to the SQL spec, with
extensions to support our historical behavior.  An aggregate belongs
to the closest query level of any of the variables in its argument,
or the current query level if there are no variables (e.g., COUNT(*)).
The implementation involves adding an agglevelsup field to Aggref,
and treating outer aggregates like outer variables at planning time.

21 years agoUpdate documentation build instructions.
Peter Eisentraut [Fri, 6 Jun 2003 14:17:08 +0000 (14:17 +0000)]
Update documentation build instructions.

21 years agoInformation schema views about functions
Peter Eisentraut [Thu, 5 Jun 2003 16:08:47 +0000 (16:08 +0000)]
Information schema views about functions

21 years agoSupport for Intel compiler on Linux
Peter Eisentraut [Thu, 5 Jun 2003 16:07:25 +0000 (16:07 +0000)]
Support for Intel compiler on Linux

21 years agoMake the world at least marginally safe for usernames with embedded spaces.
Tom Lane [Mon, 2 Jun 2003 19:00:29 +0000 (19:00 +0000)]
Make the world at least marginally safe for usernames with embedded spaces.
Per recent gripe.

21 years agoIn src/include/mb/pg_wchar.h we have:
Bruce Momjian [Mon, 2 Jun 2003 18:59:25 +0000 (18:59 +0000)]
In src/include/mb/pg_wchar.h we have:

#define PG_ENCODING_BE_LAST PG_ISO_8859_8
#define PG_ENCODING_FE_LAST PG_WIN1256

but the last client encoding in the enum list is actually PG_GB18030 and
it seems that

#define PG_ENCODING_IS_CLIEN_ONLY(_enc) \
                (((_enc) > PG_ENCODING_BE_LAST && (_enc) <= PG_ENCODING_FE_LAST)

can never be true.

I think the define should read
#define PG_ENCODING_FE_LAST PG_GB18030

On the other hand, perhaps no-one cares, because
PG_ENCODING_IS_CLIEN_ONLY is never used.

--
Oliver Elphick                                Oliver.Elphick@lfix.co.uk

21 years agoAdd:
Bruce Momjian [Mon, 2 Jun 2003 18:51:43 +0000 (18:51 +0000)]
Add:

> * Allow logging of only data definition(DDL), or DDL and modification statements

21 years agoAdd:
Bruce Momjian [Mon, 2 Jun 2003 18:42:31 +0000 (18:42 +0000)]
Add:

> * Allow a single index to index multiple tables (for inheritance and subtables)
408a410
> * Improve the planner to use CHECK constraints to prune the plan (for subtables)
418a421
> * Allow partitioning of table into multiple subtables
419a423
> T

21 years agoAdd data partitioning idea to tablespaces discussion.
Bruce Momjian [Mon, 2 Jun 2003 18:36:24 +0000 (18:36 +0000)]
Add data partitioning idea to tablespaces discussion.

21 years agoAdd Turkish FAQ, from Devrim GUNDUZ.
Bruce Momjian [Mon, 2 Jun 2003 18:16:56 +0000 (18:16 +0000)]
Add Turkish FAQ, from Devrim GUNDUZ.

21 years agoAdd:
Bruce Momjian [Mon, 2 Jun 2003 17:26:16 +0000 (17:26 +0000)]
Add:

> * Allow SET CONSTRAINTS to be qualified by schema/table

21 years agoUpdate Russian FAQ, from Viktor Vislobokov
Bruce Momjian [Mon, 2 Jun 2003 16:39:06 +0000 (16:39 +0000)]
Update Russian FAQ, from Viktor Vislobokov

21 years agoFixed segfault in forward definition parsing.
Michael Meskes [Mon, 2 Jun 2003 15:38:02 +0000 (15:38 +0000)]
Fixed segfault in forward definition parsing.

21 years agoUpdate URL to point to proper location.
Bruce Momjian [Mon, 2 Jun 2003 15:01:03 +0000 (15:01 +0000)]
Update URL to point to proper location.

21 years agoUpdate Emacs settings, from Andrew Dunstan
Bruce Momjian [Mon, 2 Jun 2003 04:35:04 +0000 (04:35 +0000)]
Update Emacs settings, from Andrew Dunstan

21 years agoAdd:
Bruce Momjian [Sun, 1 Jun 2003 04:31:33 +0000 (04:31 +0000)]
Add:

> * Add config variable to prevent auto-adding missing FROM-clause tables

21 years agomarkTargetListOrigin neglected to handle outer-scope Vars properly;
Tom Lane [Sat, 31 May 2003 19:03:34 +0000 (19:03 +0000)]
markTargetListOrigin neglected to handle outer-scope Vars properly;
per report from Joe Conway.

21 years agoAdd:
Bruce Momjian [Sat, 31 May 2003 02:54:57 +0000 (02:54 +0000)]
Add:

> * Prevent whole-row references from leaking memory, e.g. SELECT COUNT(tab.*)

21 years agoAdd:
Bruce Momjian [Sat, 31 May 2003 02:31:50 +0000 (02:31 +0000)]
Add:

> * Consider using MVCC to cache count(*) queries with no WHERE clause

21 years agoAdd item about sending success stories to http://advocacy.postgresql.org.
Bruce Momjian [Sat, 31 May 2003 02:25:11 +0000 (02:25 +0000)]
Add item about sending success stories to http://advocacy.postgresql.org.

21 years agoCause pg_dumpall to support the -a, -s, -x options of pg_dump.
Tom Lane [Fri, 30 May 2003 23:55:10 +0000 (23:55 +0000)]
Cause pg_dumpall to support the -a, -s, -x options of pg_dump.

21 years agoCause pg_dumpall to include GRANT/REVOKE for database-level permissions
Tom Lane [Fri, 30 May 2003 22:55:16 +0000 (22:55 +0000)]
Cause pg_dumpall to include GRANT/REVOKE for database-level permissions
in its output.  Make it work with server versions back to 7.0, too.

21 years agoSmall performance improvement for hash joins and hash aggregation:
Tom Lane [Fri, 30 May 2003 20:23:10 +0000 (20:23 +0000)]
Small performance improvement for hash joins and hash aggregation:
when the plan is ReScanned, we don't have to rebuild the hash table
if there is no parameter change for its child node.  This idea has
been used for a long time in Sort and Material nodes, but was not in
the hash code till now.

21 years agoCreate a real prototype function for describe.
Michael Meskes [Fri, 30 May 2003 13:22:02 +0000 (13:22 +0000)]
Create a real prototype function for describe.

21 years agoSync and some minor cleanup/fixing work plus an EXEC SQL DESCRIBE prototype.
Michael Meskes [Fri, 30 May 2003 08:39:01 +0000 (08:39 +0000)]
Sync and some minor cleanup/fixing work plus an EXEC SQL DESCRIBE prototype.

21 years agoEnsure that in all flex lexers that are part of the backend, a
Tom Lane [Thu, 29 May 2003 22:30:02 +0000 (22:30 +0000)]
Ensure that in all flex lexers that are part of the backend, a
yy_fatal_error() call results in elog(ERROR) not exit().  This was
already fixed in the main lexer and plpgsql, but extend same technique
to all the other dot-l files.  Also, on review of the possible calls
to yy_fatal_error(), it seems safe to use elog(ERROR) not elog(FATAL).

21 years agoone more change to correctly support the removeall of autocommit
Barry Lind [Thu, 29 May 2003 21:44:47 +0000 (21:44 +0000)]
one more change to correctly support the removeall of autocommit

21 years agoAdopt latest bison's spelling of 'syntax error' rather than 'parse error'
Tom Lane [Thu, 29 May 2003 20:40:36 +0000 (20:40 +0000)]
Adopt latest bison's spelling of 'syntax error' rather than 'parse error'
for grammar-detected problems.  Revert Makefile hack that kept it looking
like the pre-bison-1.875 output.

21 years agoDon't explicitly close() the session socket during backend exit; instead
Tom Lane [Thu, 29 May 2003 19:15:34 +0000 (19:15 +0000)]
Don't explicitly close() the session socket during backend exit; instead
leave it for the kernel to do after the process dies.  This allows clients
to wait for the backend to exit if they wish (after sending X message,
wait till EOF is detected on the socket).

21 years agoAFAICT, none of our shared libraries are anywhere near big enough to
Tom Lane [Thu, 29 May 2003 18:08:42 +0000 (18:08 +0000)]
AFAICT, none of our shared libraries are anywhere near big enough to
need -fPIC on HPPA.  Reduce to -fpic.

21 years agoChanged variable parsing so struct[n].attr works.
Michael Meskes [Thu, 29 May 2003 13:59:26 +0000 (13:59 +0000)]
Changed variable parsing so struct[n].attr works.

21 years agoParse forward definiton of structs.
Michael Meskes [Thu, 29 May 2003 12:00:22 +0000 (12:00 +0000)]
Parse forward definiton of structs.

21 years agoApplied patch from Fernando Nasser to fix up small type error
Barry Lind [Thu, 29 May 2003 04:52:44 +0000 (04:52 +0000)]
Applied patch from Fernando Nasser to fix up small type error

21 years agoApplied patch from Fernando Nasser to improve buffer sizes to avoid unnecessary resizing.
Barry Lind [Thu, 29 May 2003 04:48:33 +0000 (04:48 +0000)]
Applied patch from Fernando Nasser to improve buffer sizes to avoid unnecessary resizing.

21 years agoclean up warnings produced by Eclipse
Barry Lind [Thu, 29 May 2003 04:39:51 +0000 (04:39 +0000)]
clean up warnings produced by Eclipse

21 years agoAdding .cvsignore files for Eclipse IDE
Barry Lind [Thu, 29 May 2003 04:04:08 +0000 (04:04 +0000)]
Adding .cvsignore files for Eclipse IDE

21 years agoUp the build number to reflect the changes for V3 protocol support
Barry Lind [Thu, 29 May 2003 03:22:48 +0000 (03:22 +0000)]
Up the build number to reflect the changes for V3 protocol support

21 years agoInitial attempt to integrate in V3 protocol support. This is still a work in
Barry Lind [Thu, 29 May 2003 03:21:32 +0000 (03:21 +0000)]
Initial attempt to integrate in V3 protocol support.  This is still a work in
progress, although all RTs pass using the V3 protocol on a 7.4 database and also pass using the V2 protocol on a 7.3 database.
SSL support is known not to work.

 Modified Files:
  jdbc/org/postgresql/PGConnection.java
  jdbc/org/postgresql/errors.properties
  jdbc/org/postgresql/core/BaseConnection.java
  jdbc/org/postgresql/core/Encoding.java
  jdbc/org/postgresql/core/Field.java
  jdbc/org/postgresql/core/PGStream.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/jdbc1/AbstractJdbc1Connection.java
  jdbc/org/postgresql/test/jdbc2/BlobTest.java
  jdbc/org/postgresql/test/jdbc2/CallableStmtTest.java
  jdbc/org/postgresql/test/jdbc2/MiscTest.java
  jdbc/org/postgresql/test/jdbc3/Jdbc3TestSuite.java

21 years agoAdd a regression test to catch future silliness in the index-building
Tom Lane [Thu, 29 May 2003 01:09:13 +0000 (01:09 +0000)]
Add a regression test to catch future silliness in the index-building
area...

21 years agoFix stupid oversight :-(
Tom Lane [Thu, 29 May 2003 00:54:42 +0000 (00:54 +0000)]
Fix stupid oversight :-(

21 years agoKnock down a couple more lappend() hotspots for large WHERE clauses.
Tom Lane [Wed, 28 May 2003 23:06:16 +0000 (23:06 +0000)]
Knock down a couple more lappend() hotspots for large WHERE clauses.

21 years agoFix some planner performance problems with large WHERE clauses, by
Tom Lane [Wed, 28 May 2003 22:32:50 +0000 (22:32 +0000)]
Fix some planner performance problems with large WHERE clauses, by
introducing new 'FastList' list-construction subroutines to use in
hot spots.  This avoids the O(N^2) behavior of repeated lappend's
by keeping a tail pointer, while not changing behavior by reversing
list order as the lcons() method would do.

21 years agoFix breakage of LINUX_PROFILE code due to recent Windows changes.
Tom Lane [Wed, 28 May 2003 19:36:28 +0000 (19:36 +0000)]
Fix breakage of LINUX_PROFILE code due to recent Windows changes.

21 years agoFix things so that an error occuring during standalone-backend processing
Tom Lane [Wed, 28 May 2003 18:19:09 +0000 (18:19 +0000)]
Fix things so that an error occuring during standalone-backend processing
in initdb will result in exit(1), allowing the initdb script to realize
that there's something wrong.

21 years agoIn bootstrap and standalone-backend modes, do not sort LOG elevel out
Tom Lane [Wed, 28 May 2003 17:25:02 +0000 (17:25 +0000)]
In bootstrap and standalone-backend modes, do not sort LOG elevel out
of order; the 'server log' output is actually client output in these
scenarios and we ought to treat elevels the same way as in the client
case.  This allows initdb to not send backend stderr to /dev/null anymore,
which makes it much more likely that people will notice problems during
initdb.

21 years agoReplace functional-index facility with expressional indexes. Any column
Tom Lane [Wed, 28 May 2003 16:04:02 +0000 (16:04 +0000)]
Replace functional-index facility with expressional indexes.  Any column
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.

21 years agoAdd to thread discussion.
Bruce Momjian [Wed, 28 May 2003 05:02:38 +0000 (05:02 +0000)]
Add to thread discussion.

21 years agoAdd info about selecting a random row.
Bruce Momjian [Wed, 28 May 2003 04:35:28 +0000 (04:35 +0000)]
Add info about selecting a random row.

21 years agoInternationalize interactive yes/no responses.
Peter Eisentraut [Tue, 27 May 2003 19:36:55 +0000 (19:36 +0000)]
Internationalize interactive yes/no responses.

21 years agoMake log_ GUC settings controllable only by super-users, so non-super
Bruce Momjian [Tue, 27 May 2003 17:55:50 +0000 (17:55 +0000)]
Make log_ GUC settings controllable only by super-users, so non-super
users can't turn it off.

21 years agoMake debug_ GUC varables output DEBUG1 rather than LOG, and mention in
Bruce Momjian [Tue, 27 May 2003 17:49:47 +0000 (17:49 +0000)]
Make debug_ GUC varables output DEBUG1 rather than LOG, and mention in
docs that CLIENT/LOG_MIN_MESSAGES now controls debug_* output location.
Doc changes included.

21 years agoAdd configure thread tests.
Bruce Momjian [Tue, 27 May 2003 16:36:50 +0000 (16:36 +0000)]
Add configure thread tests.

21 years agoAllow input from stdin and output to stdout.
Michael Meskes [Tue, 27 May 2003 14:36:00 +0000 (14:36 +0000)]
Allow input from stdin and output to stdout.

21 years agoFixed incorrect output for some structs.
Michael Meskes [Tue, 27 May 2003 11:31:52 +0000 (11:31 +0000)]
Fixed incorrect output for some structs.

21 years agoAdjust error-handling logic in libpq. For the first time, libpq copes
Tom Lane [Mon, 26 May 2003 20:05:20 +0000 (20:05 +0000)]
Adjust error-handling logic in libpq.  For the first time, libpq copes
sanely with running out of memory for a query result.

21 years agoAdd item for changing a column's data type.
Bruce Momjian [Mon, 26 May 2003 19:25:20 +0000 (19:25 +0000)]
Add item for changing a column's data type.

21 years agoFix misstatement in release-notes item.
Tom Lane [Mon, 26 May 2003 18:58:26 +0000 (18:58 +0000)]
Fix misstatement in release-notes item.

21 years agoUpdate:
Bruce Momjian [Mon, 26 May 2003 18:36:51 +0000 (18:36 +0000)]
Update:

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

21 years agoMake sure printtup() always sends the number of columns previously
Tom Lane [Mon, 26 May 2003 17:51:38 +0000 (17:51 +0000)]
Make sure printtup() always sends the number of columns previously
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...

21 years agoAdd documentation to 'createuser' that ALTER USER must be used to add or
Bruce Momjian [Mon, 26 May 2003 17:50:09 +0000 (17:50 +0000)]
Add documentation to 'createuser' that ALTER USER must be used to add or
modify passwords, per user confusion report.

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

21 years agoAdd:
Bruce Momjian [Mon, 26 May 2003 02:56:40 +0000 (02:56 +0000)]
Add:

> * Allow UPDATE to use SET col = DEFAULT

21 years agoRemove comment that cbrt() isn't being used.
Bruce Momjian [Mon, 26 May 2003 00:55:25 +0000 (00:55 +0000)]
Remove comment that cbrt() isn't being used.

21 years agoCause CHAR(n) to TEXT or VARCHAR conversion to automatically strip trailing
Tom Lane [Mon, 26 May 2003 00:11:29 +0000 (00:11 +0000)]
Cause CHAR(n) to TEXT or VARCHAR conversion to automatically strip trailing
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).

21 years agoInformation schema improvements
Peter Eisentraut [Sun, 25 May 2003 09:36:09 +0000 (09:36 +0000)]
Information schema improvements

21 years agoClean up our static cbrt() implementation in float.c.
Bruce Momjian [Sun, 25 May 2003 05:30:57 +0000 (05:30 +0000)]
Clean up our static cbrt() implementation in float.c.

21 years agoMove item:
Bruce Momjian [Sun, 25 May 2003 04:44:41 +0000 (04:44 +0000)]
Move item:

< * 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

21 years agoUpdate:
Bruce Momjian [Sun, 25 May 2003 03:20:13 +0000 (03:20 +0000)]
Update:

< * 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

21 years agoAdd:
Bruce Momjian [Sat, 24 May 2003 19:03:21 +0000 (19:03 +0000)]
Add:

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

21 years agoUpdate XML items:
Bruce Momjian [Sat, 24 May 2003 18:25:35 +0000 (18:25 +0000)]
Update XML items:
> * Allow backend to output result sets in XML
> * Add XML capability to pg_dump and COPY, when backend XML capability

21 years agoAdded:
Bruce Momjian [Sat, 24 May 2003 15:33:46 +0000 (15:33 +0000)]
Added:

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

21 years agoAdd cost estimate discussion to TODO.detail.
Bruce Momjian [Sat, 24 May 2003 03:59:06 +0000 (03:59 +0000)]
Add cost estimate discussion to TODO.detail.

21 years agoAdd to TODO:
Bruce Momjian [Sat, 24 May 2003 01:28:22 +0000 (01:28 +0000)]
Add to TODO:

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

Add to trigger TODO.detail.

21 years ago With disabled triggers, allow pg_dump to use ALTER TABLE ADD FOREIGN KEY
Bruce Momjian [Sat, 24 May 2003 01:27:41 +0000 (01:27 +0000)]
 With disabled triggers, allow pg_dump to use ALTER TABLE ADD FOREIGN KEY

21 years agoImprove implementation of btrim/ltrim/rtrim: provide a special case for
Tom Lane [Fri, 23 May 2003 22:33:23 +0000 (22:33 +0000)]
Improve implementation of btrim/ltrim/rtrim: provide a special case for
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.

21 years agoOne more ecpg todo fixed and removed.
Michael Meskes [Fri, 23 May 2003 17:20:46 +0000 (17:20 +0000)]
One more ecpg todo fixed and removed.

21 years agoPlace statement_timeout in proper order in docs, from Bruno Wolff III.
Bruce Momjian [Fri, 23 May 2003 16:34:36 +0000 (16:34 +0000)]
Place statement_timeout in proper order in docs, from Bruno Wolff III.

21 years agoForgot these.
Michael Meskes [Fri, 23 May 2003 15:22:38 +0000 (15:22 +0000)]
Forgot these.

21 years agoecpg now accepts array elements as arguments.
Michael Meskes [Fri, 23 May 2003 15:19:36 +0000 (15:19 +0000)]
ecpg now accepts array elements as arguments.

21 years agoSeems pl_funcs didn't have support for plpgsql PERFORM statement.
Tom Lane [Fri, 23 May 2003 04:08:34 +0000 (04:08 +0000)]
Seems pl_funcs didn't have support for plpgsql PERFORM statement.

21 years agoUpdate for 7.3.3.
Tom Lane [Thu, 22 May 2003 20:38:50 +0000 (20:38 +0000)]
Update for 7.3.3.

21 years agoAdd define for missing EAI_ADDRFAMILY in BSD/OS 4.3.
Bruce Momjian [Thu, 22 May 2003 19:14:25 +0000 (19:14 +0000)]
Add define for missing EAI_ADDRFAMILY in BSD/OS 4.3.

21 years agoUpdate release history for impending 7.3.3 release.
Tom Lane [Thu, 22 May 2003 18:31:45 +0000 (18:31 +0000)]
Update release history for impending 7.3.3 release.

21 years agoFix compilation on Cygwin.
Peter Eisentraut [Thu, 22 May 2003 17:20:44 +0000 (17:20 +0000)]
Fix compilation on Cygwin.

21 years agoAdd defense against possibility that tzname[] doesn't exist.
Tom Lane [Thu, 22 May 2003 17:13:08 +0000 (17:13 +0000)]
Add defense against possibility that tzname[] doesn't exist.

21 years agoFixed segfault when detecting undefined cursor.
Michael Meskes [Thu, 22 May 2003 17:09:00 +0000 (17:09 +0000)]
Fixed segfault when detecting undefined cursor.

21 years agoAdjust configure so that extern tzname[] will be checked for
Tom Lane [Thu, 22 May 2003 16:39:30 +0000 (16:39 +0000)]
Adjust configure so that extern tzname[] will be checked for
independently of whether the struct tm tm_zone member exists.
Also run autoheader, which seems not to have been done lately;
it added about three more things to pg_config.h.in than I was expecting...

21 years agoRepair sometimes-incorrect computation of StartUpID after a crash, per
Tom Lane [Thu, 22 May 2003 14:39:28 +0000 (14:39 +0000)]
Repair sometimes-incorrect computation of StartUpID after a crash, per
example from Rao Kumar.  This is a very corner corner-case, requiring
a minimum of three closely-spaced database crashes and an unlucky
positioning of the second recovery's checkpoint record before you'd notice
any problem.  But the consequences are dire enough that it's a must-fix.

21 years agoRemoved one TODO item for ecpg as this is done.
Michael Meskes [Thu, 22 May 2003 08:15:42 +0000 (08:15 +0000)]
Removed one TODO item for ecpg as this is done.

21 years agoecpg now recognizes named structs/unions. So you don't have to list the whole definit...
Michael Meskes [Thu, 22 May 2003 07:58:45 +0000 (07:58 +0000)]
ecpg now recognizes named structs/unions. So you don't have to list the whole definition everytime you declare a variable anymore.

21 years agoFix coredump in pg_get_triggerdef, ensure function name is schema-
Tom Lane [Tue, 20 May 2003 20:35:10 +0000 (20:35 +0000)]
Fix coredump in pg_get_triggerdef, ensure function name is schema-
qualified when necessary, simplify argument-printing code.

21 years agoReversed my incorrect fix.
Michael Meskes [Tue, 20 May 2003 11:05:27 +0000 (11:05 +0000)]
Reversed my incorrect fix.

21 years agoAdd error message style guidelines to the SGML documentation.
Tom Lane [Mon, 19 May 2003 21:38:24 +0000 (21:38 +0000)]
Add error message style guidelines to the SGML documentation.

21 years agoUse -fPIC on Sparc, per Tom Callaway.
Tom Lane [Mon, 19 May 2003 17:51:01 +0000 (17:51 +0000)]
Use -fPIC on Sparc, per Tom Callaway.

21 years agoRename Win32 include directory from win32_include to win32.
Bruce Momjian [Mon, 19 May 2003 17:31:59 +0000 (17:31 +0000)]
Rename Win32 include directory from win32_include to win32.

21 years agoFixed exec sql ifdef command.
Michael Meskes [Mon, 19 May 2003 07:19:57 +0000 (07:19 +0000)]
Fixed exec sql ifdef command.

21 years agoRepair quoting sloppiness, lack of schema awareness in reindexdb.
Tom Lane [Mon, 19 May 2003 01:00:12 +0000 (01:00 +0000)]
Repair quoting sloppiness, lack of schema awareness in reindexdb.

21 years agoAdd documentation for information schema, and some corrections on some
Peter Eisentraut [Sun, 18 May 2003 20:55:57 +0000 (20:55 +0000)]
Add documentation for information schema, and some corrections on some
views.

21 years agoAdd code to test for unknown timezone names (following some ideas from
Tom Lane [Sun, 18 May 2003 01:06:26 +0000 (01:06 +0000)]
Add code to test for unknown timezone names (following some ideas from
Ross Reedstrom, a couple months back) and to detect timezones that are
using leap-second timekeeping.  The unknown-zone-name test is pretty
heuristic and ugly, but it seems better than the old behavior of just
switching to GMT given a bad name.  Also make DecodePosixTimezone() a
tad more robust.

21 years agoFix breakage induced by recent patch (careless use of fmtId()).
Tom Lane [Sat, 17 May 2003 15:53:12 +0000 (15:53 +0000)]
Fix breakage induced by recent patch (careless use of fmtId()).