OSDN Git Service
Jan Wieck [Sat, 27 Sep 2003 15:34:06 +0000 (15:34 +0000)]
Changed the logic when a CAST is dumped according to discussion
on pgsql-hackers.
A cast is included in the dump output if any of the objects does
not belong to a system namespace and all of the non-system namespace
objects belong to dumped namespaces. System namespace is defined
as nspname begins with "pg_".
Jan
Bruce Momjian [Sat, 27 Sep 2003 15:32:48 +0000 (15:32 +0000)]
Rename thread compile flag. Move thread test program to tools/thread,
and improve tests.
Bruce Momjian [Sat, 27 Sep 2003 14:06:21 +0000 (14:06 +0000)]
Small SGML cleanup from Jeroen Ruigrok/asmodai
Peter Eisentraut [Sat, 27 Sep 2003 09:29:31 +0000 (09:29 +0000)]
You can't NLS-enable a program component by just putting gettext() around
every string, especially if some of the output should be fixed-format
machine-readable. This needs to be more carefully sorted out. Also, make
the help message generated by --help-config -h be more similar in style to
the others.
Peter Eisentraut [Sat, 27 Sep 2003 09:21:26 +0000 (09:21 +0000)]
The formatting of the display of the locale names assumed that locale names
are not longer than 8 characters. But sometimes they are, and that made
the display quite ugly. So just format them vertically so that everyone
can read them.
Bruce Momjian [Sat, 27 Sep 2003 00:10:31 +0000 (00:10 +0000)]
Document that functions are checked independently of the view
permissions.
Bruce Momjian [Fri, 26 Sep 2003 17:47:44 +0000 (17:47 +0000)]
Add "tuning" section to install instructions.
Tom Lane [Fri, 26 Sep 2003 17:39:13 +0000 (17:39 +0000)]
Move -D_GNU_SOURCE hack from port header to template, so that
configure's tests see the same compilation environment as the code.
Per discussion with Stephan Szabo.
Peter Eisentraut [Fri, 26 Sep 2003 15:27:37 +0000 (15:27 +0000)]
Various message fixes, among those fixes for the previous round of fixes
Peter Eisentraut [Fri, 26 Sep 2003 15:25:39 +0000 (15:25 +0000)]
Translation update
Michael Meskes [Fri, 26 Sep 2003 15:16:29 +0000 (15:16 +0000)]
Fixed segfault after error in parsing precision argument.
Tom Lane [Fri, 26 Sep 2003 01:17:01 +0000 (01:17 +0000)]
Fix tid scan evaluation of non-constant TID values; can't try to do it
during ExecInitTidScan, because the rest of the executor isn't ready.
Tom Lane [Thu, 25 Sep 2003 23:02:12 +0000 (23:02 +0000)]
Make the world safe (more or less) for dropped columns in plpgsql rowtypes.
Tom Lane [Thu, 25 Sep 2003 19:41:49 +0000 (19:41 +0000)]
tlist_matches_tupdesc() needs to defend itself against dropped columns.
Tom Lane [Thu, 25 Sep 2003 18:58:36 +0000 (18:58 +0000)]
Get rid of ReferentialIntegritySnapshotOverride by extending Executor API
to allow es_snapshot to be set to SnapshotNow rather than a query snapshot.
This solves a bug reported by Wade Klaver, wherein triggers fired as a
result of RI cascade updates could misbehave.
Tom Lane [Thu, 25 Sep 2003 16:16:35 +0000 (16:16 +0000)]
Fix regression error messages for platforms Peter doesn't use.
Tom Lane [Thu, 25 Sep 2003 15:58:06 +0000 (15:58 +0000)]
Fix grammatical error introduced into error message.
Peter Eisentraut [Thu, 25 Sep 2003 15:48:17 +0000 (15:48 +0000)]
Forget to update the real source of the regression test results after
message editing.
Peter Eisentraut [Thu, 25 Sep 2003 06:58:07 +0000 (06:58 +0000)]
Message editing: remove gratuitous variations in message wording, standardize
terms, add some clarifications, fix some untranslatable attempts at dynamic
message building.
Bruce Momjian [Thu, 25 Sep 2003 02:27:54 +0000 (02:27 +0000)]
Update Japanese FAQ, from Jun Kuwamura
Bruce Momjian [Wed, 24 Sep 2003 19:05:32 +0000 (19:05 +0000)]
Fix creation of temporary sequences in ecpg, per report from Edmund Bacon.
Tom Lane [Wed, 24 Sep 2003 18:54:02 +0000 (18:54 +0000)]
Repair some REINDEX problems per recent discussions. The relcache is
now able to cope with assigning new relfilenode values to nailed-in-cache
indexes, so they can be reindexed using the fully crash-safe method. This
leaves only shared system indexes as special cases. Remove the 'index
deactivation' code, since it provides no useful protection in the shared-
index case. Require reindexing of shared indexes to be done in standalone
mode, but remove other restrictions on REINDEX. -P (IgnoreSystemIndexes)
now prevents using indexes for lookups, but does not disable index updates.
It is therefore safe to allow from PGOPTIONS. Upshot: reindexing system catalogs
can be done without a standalone backend for all cases except
shared catalogs.
Marc G. Fournier [Wed, 24 Sep 2003 18:09:15 +0000 (18:09 +0000)]
another totally benign, test commit
Marc G. Fournier [Wed, 24 Sep 2003 18:08:20 +0000 (18:08 +0000)]
test 2, benign, no change
Marc G. Fournier [Wed, 24 Sep 2003 18:05:36 +0000 (18:05 +0000)]
not change, just added a space to test something
Tom Lane [Wed, 24 Sep 2003 05:38:38 +0000 (05:38 +0000)]
Add positive defense against dropped columns, per suggestion from
Christopher Kings-Lynne. Also fix a couple more schema-awareness issues.
Tom Lane [Tue, 23 Sep 2003 23:31:52 +0000 (23:31 +0000)]
Issue CREATE SCHEMA as the originally connecting user, with an
AUTHORIZATION clause to specify the desired owner. This allows a
superuser to restore schemas owned by users without CREATE-SCHEMA
permissions (ie, schemas originally created by a superuser using
AUTHORIZATION). --no-owner can be specified to suppress the
AUTHORIZATION clause if need be.
Tom Lane [Tue, 23 Sep 2003 22:48:53 +0000 (22:48 +0000)]
pg_dump/pg_restore now always use SET SESSION AUTHORIZATION, not \connect,
to control object ownership. The use-set-session-authorization and
no-reconnect switches are obsolete (still accepted on the command line,
but they don't do anything). This is a precursor to fixing handling
of CREATE SCHEMA, which will be a separate commit.
Bruce Momjian [Tue, 23 Sep 2003 21:00:18 +0000 (21:00 +0000)]
Add:
> * Allow LISTEN/NOTIFY to store info in memory rather than tables
Bruce Momjian [Tue, 23 Sep 2003 20:45:58 +0000 (20:45 +0000)]
Add:
* Allow major upgrades without dump/reload, perhaps using pg_upgrade
Tom Lane [Tue, 23 Sep 2003 19:58:50 +0000 (19:58 +0000)]
Minor copy-editing for plpgsql chapter.
Tom Lane [Tue, 23 Sep 2003 17:12:53 +0000 (17:12 +0000)]
Tweak generic_type_consistency routines to avoid loss of functionality
since 7.3: 'select array_dims(histogram_bounds) from pg_stats' used to
work and still should. Problem was that code wouldn't take input of
declared type anyarray as matching an anyarray argument. Allow this
case as long as we don't need to determine an element type (which in
practice means as long as anyelement isn't used in the function signature).
Tom Lane [Tue, 23 Sep 2003 15:11:33 +0000 (15:11 +0000)]
_SPI_cursor_operation forgot to check for failure return from
_SPI_begin_call. Per gripe from Tomasz Myrta.
Michael Meskes [Tue, 23 Sep 2003 12:56:35 +0000 (12:56 +0000)]
- Changed struct definition handling so "struct foo {}" always gets defined.
Barry Lind [Tue, 23 Sep 2003 06:13:52 +0000 (06:13 +0000)]
Fix to properly handle timezone offsets that are partial hours. If the offset
was a partial hour and less than gmt (i.e. -02:30) the code would corrupt the
minutes part.
Modified Files:
jdbc/org/postgresql/jdbc1/AbstractJdbc1Statement.java
Barry Lind [Tue, 23 Sep 2003 06:11:06 +0000 (06:11 +0000)]
A fix for jdbc regression test submitted yesterday by Kim Ho at Redhat
Modified Files:
ResultSetTest.java
Hiroshi Inoue [Tue, 23 Sep 2003 01:51:09 +0000 (01:51 +0000)]
Putting back the previous change must be the first thing.
ALso put back a #ifndef ENABLE_REINDEX_NAILED_RELATIONS
which was removed about a year ago.
Teodor Sigaev [Mon, 22 Sep 2003 13:32:33 +0000 (13:32 +0000)]
fix hlfinditem function. Thanks to "Stphane Bidoul" <stephane.bidoul@softwareag.com>.
The 'word' variable there is initialised from
the prs->words array, but immediately after,
that array may be reallocated, thus leaving
word pointing to unallocated memory.
Michael Meskes [Mon, 22 Sep 2003 13:19:39 +0000 (13:19 +0000)]
- Fixed order mismatch in processing "using" arguments.
- Fixed some minor things in test cases.
- Use defines for Informix error codes.
Barry Lind [Mon, 22 Sep 2003 05:38:01 +0000 (05:38 +0000)]
Additional jdbc regression tests submitted by Oliver Jowett. Some tests are
currently commented out, pending fixes for the bugs these tests uncovered.
Modified Files:
jdbc/org/postgresql/test/jdbc2/Jdbc2TestSuite.java
jdbc/org/postgresql/test/jdbc2/ServerPreparedStmtTest.java
Added Files:
jdbc/org/postgresql/test/jdbc2/CursorFetchTest.java
Barry Lind [Mon, 22 Sep 2003 04:55:00 +0000 (04:55 +0000)]
Applied jdbc patch from Kim Ho at RedHat to improve boolean support, as well
as some additional regression tests for this an other recent changes.
Modified Files:
jdbc/org/postgresql/jdbc1/AbstractJdbc1ResultSet.java
jdbc/org/postgresql/test/jdbc2/DateTest.java
jdbc/org/postgresql/test/jdbc2/ResultSetTest.java
jdbc/org/postgresql/test/jdbc2/TimeTest.java
jdbc/org/postgresql/test/jdbc2/TimestampTest.java
Tom Lane [Mon, 22 Sep 2003 00:47:23 +0000 (00:47 +0000)]
HeapTupleSatisfiesVacuum() needs to be more careful about the
difference between INSERT_IN_PROGRESS and DELETE_IN_PROGRESS for
tuples inserted and then deleted by a concurrent transaction.
Example of bug:
regression=# create table foo (f1 int);
CREATE TABLE
regression=# begin;
BEGIN
regression=# insert into foo values(1);
INSERT 195531 1
regression=# delete from foo;
DELETE 1
regression=# insert into foo values(1);
INSERT 195532 1
regression=# create unique index fooi on foo(f1);
ERROR: could not create unique index
DETAIL: Table contains duplicated values.
Peter Eisentraut [Mon, 22 Sep 2003 00:27:01 +0000 (00:27 +0000)]
Translation updates
Peter Eisentraut [Mon, 22 Sep 2003 00:23:35 +0000 (00:23 +0000)]
Message wording improvements
Peter Eisentraut [Mon, 22 Sep 2003 00:16:58 +0000 (00:16 +0000)]
Make the SQL command synopses appear less random.
Tom Lane [Sun, 21 Sep 2003 17:57:21 +0000 (17:57 +0000)]
Arrange to align shared disk buffers on at least 32-byte boundaries,
not just MAXALIGN boundaries. This makes a noticeable difference in
the speed of transfers to and from kernel space, at least on recent
Pentiums, and might help other CPUs too. We should look at making
this happen for local buffers and buffile.c too. Patch from Manfred Spraul.
Tom Lane [Sun, 21 Sep 2003 17:42:22 +0000 (17:42 +0000)]
Fix some ill-considered ifdefs, per report from Hans-JÏrgen SchÎnig.
Tom Lane [Sat, 20 Sep 2003 21:26:20 +0000 (21:26 +0000)]
Apply ranlib(1) after installing a static (.a) library on Darwin.
Per past complaints --- verified to still be a problem on OS X 10.2.6.
Tom Lane [Sat, 20 Sep 2003 21:14:57 +0000 (21:14 +0000)]
Attaching -lpq to LDFLAGS doesn't work, because the generated command
puts that before the .c file name.
Tom Lane [Sat, 20 Sep 2003 20:12:05 +0000 (20:12 +0000)]
Fix a batch of speling misteaks identified by Peter's spell-checker tool.
Michael Meskes [Sat, 20 Sep 2003 09:10:50 +0000 (09:10 +0000)]
- Renamed my own strndup() function because of a name clash.
Michael Meskes [Sat, 20 Sep 2003 09:10:09 +0000 (09:10 +0000)]
- Applied some bug fixing patches by Dave Cramer <dave@fastcrypt.com>.
- Added protecting defines to include files.
Tom Lane [Fri, 19 Sep 2003 23:15:18 +0000 (23:15 +0000)]
pg_dump missed comments attached to primary key or unique constraints
(as distinct from their underlying indexes). Per bug report from
Hiroshi Saito.
Tom Lane [Fri, 19 Sep 2003 21:06:39 +0000 (21:06 +0000)]
Document that TRUNCATE ignores user-defined ON DELETE triggers.
Tom Lane [Fri, 19 Sep 2003 21:04:20 +0000 (21:04 +0000)]
Disallow foreign-key references from temp tables to permanent tables.
Per recent discussion, this does not work because other backends can't
reliably see tuples in a temp table and so cannot run the RI checks
correctly. Seems better to disallow this case than go back to accessing
temp tables through shared buffers. Also, disallow FK references to
ON COMMIT DELETE ROWS tables. We already caught this problem for normal
TRUNCATE, but the path used by ON COMMIT didn't check.
Tom Lane [Fri, 19 Sep 2003 20:47:59 +0000 (20:47 +0000)]
Latest version of gcc from Apple does not work well with -traditional-cpp
(it rejects some system header files...). Use -no-cpp-precomp instead.
I think it is okay to change this unconditionally, but if we hear
complaints from people still using very old compilers on Darwin,
we could put in a test to see which switch the compiler likes.
Tom Lane [Fri, 19 Sep 2003 19:57:42 +0000 (19:57 +0000)]
Seems like a bad idea that REINDEX TABLE supports (or thinks it does)
reindexing system tables without ignoring system indexes, when the
other two varieties of REINDEX disallow it. Make all three act the same,
and simplify downstream code accordingly.
Michael Meskes [Fri, 19 Sep 2003 14:13:16 +0000 (14:13 +0000)]
One small typo in preprocessor.
Michael Meskes [Fri, 19 Sep 2003 14:06:21 +0000 (14:06 +0000)]
Removed superfluous return statement in get_data.
Tom Lane [Thu, 18 Sep 2003 20:30:15 +0000 (20:30 +0000)]
Put a tad more detail in the discussion of postmaster and postgres
signal handling.
Michael Meskes [Thu, 18 Sep 2003 13:12:23 +0000 (13:12 +0000)]
- Accept output variables for FETCH in DECLARE statement.
- Synced parser.
- Allowed C variables to carry the name of prepared statements.
- Added Informix handling of datatype converion errors.
Barry Lind [Thu, 18 Sep 2003 04:14:27 +0000 (04:14 +0000)]
Applied patch from Oliver Jowett to clean up some instances where the wrong
type was being reported for PREPAREs.
Modified Files:
jdbc/org/postgresql/jdbc1/AbstractJdbc1Statement.java
jdbc/org/postgresql/test/jdbc2/ServerPreparedStmtTest.java
Barry Lind [Thu, 18 Sep 2003 04:09:02 +0000 (04:09 +0000)]
Small jdbc patch from Kim Ho at RedHat to fix some boolean problems
Modified Files:
jdbc/org/postgresql/jdbc1/AbstractJdbc1ResultSet.java
Barry Lind [Thu, 18 Sep 2003 03:58:16 +0000 (03:58 +0000)]
A minor fix to jdbc regression tests
Modified Files:
jdbc/org/postgresql/test/jdbc2/DatabaseMetaDataPropertiesTest.java
Tom Lane [Wed, 17 Sep 2003 17:19:17 +0000 (17:19 +0000)]
Disallow converting a table to a view if it has triggers, indexes, or
child tables --- all cases that will trip various sanity checks elsewhere
in the system, as well as cases that should not occur in the only intended
use of this feature, namely coping with ancient pg_dump representation
of views. Per bug report from Chris Pizzi.
Barry Lind [Wed, 17 Sep 2003 08:21:36 +0000 (08:21 +0000)]
Applied patch to jdbc from Kim Ho at RedHat, fixing improper handling of empty queries under the V3 protocol
Modified Files:
jdbc/org/postgresql/core/QueryExecutor.java
Barry Lind [Wed, 17 Sep 2003 07:00:24 +0000 (07:00 +0000)]
Applied patch for jdbc from Kim Ho at RedHat to more correctly deal with boolean values
Modified Files:
jdbc/org/postgresql/jdbc1/AbstractJdbc1ResultSet.java
Barry Lind [Wed, 17 Sep 2003 06:42:47 +0000 (06:42 +0000)]
Fixed jdbc to correctly report that constraints are supported
Modified Files:
jdbc/org/postgresql/jdbc1/AbstractJdbc1DatabaseMetaData.java
Barry Lind [Wed, 17 Sep 2003 05:14:52 +0000 (05:14 +0000)]
Applied patch from Fernando Nasser of RedHat to fix some error messages
that would not get correctly looked up in the translation files for jdbc
Modified Files:
jdbc/org/postgresql/errors.properties
jdbc/org/postgresql/jdbc2/AbstractJdbc2ResultSet.java
Barry Lind [Wed, 17 Sep 2003 05:07:38 +0000 (05:07 +0000)]
Applied patch from Kim Ho at redhat to improve boolean and bit handling
in the jdbc driver
Modified Files:
jdbc/org/postgresql/jdbc1/AbstractJdbc1Statement.java
jdbc/org/postgresql/jdbc3/AbstractJdbc3Statement.java
Tatsuo Ishii [Wed, 17 Sep 2003 04:25:29 +0000 (04:25 +0000)]
Fix typo in comment
Tom Lane [Tue, 16 Sep 2003 17:59:02 +0000 (17:59 +0000)]
Make psql correctly track the effects of SET CLIENT_ENCODING commands.
I thought I'd fixed this earlier, but I didn't get it right ...
Tatsuo Ishii [Tue, 16 Sep 2003 07:47:06 +0000 (07:47 +0000)]
fix typo in a comment
Tom Lane [Tue, 16 Sep 2003 04:32:28 +0000 (04:32 +0000)]
I don't care whether it's informix-compatible or not: #defining a word
like 'date' in a public header file is a damn fool idea.
Tom Lane [Tue, 16 Sep 2003 01:07:51 +0000 (01:07 +0000)]
Fix some problems with dropped columns in plpython trigger functions.
Tom Lane [Tue, 16 Sep 2003 00:50:09 +0000 (00:50 +0000)]
Since SPI_modifytuple's natts argument is the number of attributes to be
changed, it should allow a zero value (implying no changes to make).
Tom Lane [Mon, 15 Sep 2003 23:33:43 +0000 (23:33 +0000)]
Fix LISTEN/NOTIFY race condition reported by Gavin Sherry. While a
really general fix might be difficult, I believe the only case where
AtCommit_Notify could see an uncommitted tuple is where the other guy
has just unlistened and not yet committed. The best solution seems to
be to just skip updating that tuple, on the assumption that the other
guy does not want to hear about the notification anyway. This is not
perfect --- if the other guy rolls back his unlisten instead of committing,
then he really should have gotten this notify. But to do that, we'd have
to wait to see if he commits or not, or make UNLISTEN hold exclusive lock
on pg_listener until commit. Either of these answers is deadlock-prone,
not to mention horrible for interactive performance. Do it this way
for now. (What happened to that project to do LISTEN/NOTIFY in memory
with no table, anyway?)
Tom Lane [Mon, 15 Sep 2003 23:25:31 +0000 (23:25 +0000)]
Update regression test for message change.
Tom Lane [Mon, 15 Sep 2003 22:28:58 +0000 (22:28 +0000)]
Improve a couple of error messages per suggestions from Alvaro Herrera.
Peter Eisentraut [Mon, 15 Sep 2003 20:42:40 +0000 (20:42 +0000)]
Translation updates
Peter Eisentraut [Mon, 15 Sep 2003 20:03:37 +0000 (20:03 +0000)]
OK, some of these syntax errors should be given other codes.
PostgreSQL Daemon [Mon, 15 Sep 2003 14:32:35 +0000 (14:32 +0000)]
tag her for beta3, as announced on Friday ...
Tom Lane [Mon, 15 Sep 2003 03:32:24 +0000 (03:32 +0000)]
Okay, who left off the id= here?
Tom Lane [Mon, 15 Sep 2003 03:21:51 +0000 (03:21 +0000)]
LISTEN doesn't issue a warning for redundant listens anymore. Also,
add some 'See Also' links.
Bruce Momjian [Mon, 15 Sep 2003 02:30:29 +0000 (02:30 +0000)]
Add errno value for thread *_r function call buffer too small. Improve
thread comments.
Bruce Momjian [Mon, 15 Sep 2003 02:17:49 +0000 (02:17 +0000)]
NetBSD needs reentrant funcs.
Peter Eisentraut [Mon, 15 Sep 2003 00:26:31 +0000 (00:26 +0000)]
Remove warnings for operations that have no effect when executed repeatedly.
Peter Eisentraut [Sun, 14 Sep 2003 22:40:38 +0000 (22:40 +0000)]
Run distprep target before creating list of files that contain translatable
strings in the backend, so that .l and .y files are included. To that end,
don't make the .pot file a prerequisite on distprep.
Peter Eisentraut [Sun, 14 Sep 2003 22:37:13 +0000 (22:37 +0000)]
Allow translation of SQL help in psql.
Tom Lane [Sun, 14 Sep 2003 20:21:18 +0000 (20:21 +0000)]
Add missing support for Opteron (__x86_64__).
Tom Lane [Sun, 14 Sep 2003 18:44:23 +0000 (18:44 +0000)]
Persuade tsearch/tsearch2 to work (or at least pass their regression
tests) when using flex 2.5.31. The fix is to *not* try to use palloc
and pfree for allocations within the lexer; when you do that, the
yy_buffer_stack gets freed at inopportune times. The code is already
set up to do manual deallocation, so I see no particular advantage to
using palloc anyway.
Tom Lane [Sun, 14 Sep 2003 17:25:54 +0000 (17:25 +0000)]
Make pltcl create separate function objects when the same function is
used as trigger on different relations. I am not convinced that Tcl
actually has to have this, but it seems a good idea to make it be
parallel to the other PLs that definitely do need it.
Tom Lane [Sun, 14 Sep 2003 17:13:06 +0000 (17:13 +0000)]
Fix plpython to generate separate cached procedure data for each
relation, when the same function is used as a trigger on more than
one relation. This avoids crashes due to differing rowtypes for
different relations. Per bug report from Lance Thomas, 7-Feb-03.
Bruce Momjian [Sun, 14 Sep 2003 13:33:08 +0000 (13:33 +0000)]
Since I haven't gotten a reply, can someone REVERT the lines in
src/Makefile.shlib that adds $(libdir)/$(soname) for
port,eq,unixware to $(soname)?
Larry Rosenman
Tom Lane [Sun, 14 Sep 2003 02:30:07 +0000 (02:30 +0000)]
expected results for dblink test seem to be out of date ...
Tom Lane [Sun, 14 Sep 2003 02:18:49 +0000 (02:18 +0000)]
Make contrib/seg work with flex 2.5.31. Fix it up to have a real
btree operator class, too, since in PG 7.4 you can't GROUP without one.
Tom Lane [Sun, 14 Sep 2003 01:52:25 +0000 (01:52 +0000)]
Make contrib/cube work with flex 2.5.31. Fix it up to have a real
btree operator class, too, since in PG 7.4 you can't GROUP without one.
Tom Lane [Sun, 14 Sep 2003 00:03:32 +0000 (00:03 +0000)]
Reconsider placement of MemoryContextCheck() call --- do after commit,
not before, to avoid duplication of effort.
Tom Lane [Sat, 13 Sep 2003 22:25:38 +0000 (22:25 +0000)]
Avoid corner cases where aset.c would unnecessarily make malloc()
requests of sizes that aren't powers of 2. Per observation from
David Schultz, 28-Aug.
Tom Lane [Sat, 13 Sep 2003 21:44:50 +0000 (21:44 +0000)]
With Joe Conway's concurrence, remove srandom() call from normal_rand().
This was the last piece of code that took it upon itself to reset the
random number sequence --- now we only have srandom() in postmaster start,
backend start, and explicit setseed() operations.