Tom Lane [Tue, 10 Feb 2004 03:42:45 +0000 (03:42 +0000)]
Centralize implementation of delay code by creating a pg_usleep()
subroutine in src/port/pgsleep.c. Remove platform dependencies from
miscadmin.h and put them in port.h where they belong. Extend recent
vacuum cost-based-delay patch to apply to VACUUM FULL, ANALYZE, and
non-btree index vacuuming.
By the way, where is the documentation for the cost-based-delay patch?
Tom Lane [Tue, 10 Feb 2004 01:55:27 +0000 (01:55 +0000)]
Restructure smgr API as per recent proposal. smgr no longer depends on
the relcache, and so the notion of 'blind write' is gone. This should
improve efficiency in bgwriter and background checkpoint processes.
Internal restructuring in md.c to remove the not-very-useful array of
MdfdVec objects --- might as well just use pointers.
Also remove the long-dead 'persistent main memory' storage manager (mm.c),
since it seems quite unlikely to ever get resurrected.
Neil Conway [Wed, 4 Feb 2004 01:11:47 +0000 (01:11 +0000)]
Use memmove() rather than memcpy() in set_var_from_var(). If this function
is asked to assign a variable to itself, it will result in doing a
memcpy() on an entirely-overlapping memory range, which results in
undefined behavior according to ANSI C. That said, it is unlikely to
actually do anything bad on any sane libc, but this keeps valgrind quiet.
Tom Lane [Tue, 3 Feb 2004 17:34:04 +0000 (17:34 +0000)]
Rename SortMem and VacuumMem to work_mem and maintenance_work_mem.
Make btree index creation and initial validation of foreign-key constraints
use maintenance_work_mem rather than work_mem as their memory limit.
Add some code to guc.c to allow these variables to be referenced by their
old names in SHOW and SET commands, for backwards compatibility.
Joe Conway [Tue, 3 Feb 2004 08:29:57 +0000 (08:29 +0000)]
Add function (actually an int4 and an int8 version) that generates
a series of numbers, optionally using an explicit step size other
than the default value (one). Use function in the information_schema
to replace hard-wired knowledge of INDEX_MAX_KEYS. initdb forced due
to pg_proc change. Documentation update still needed -- will be
committed separately.
Neil Conway [Mon, 2 Feb 2004 16:58:30 +0000 (16:58 +0000)]
Fix a read of uninitialized memory in next_token() of hba.c, spotted via
valgrind: a buffer passed to strncmp() had to be NUL-terminated. Original
report and patch from Dennis Bjorkland, some cleanup by Andrew Dunstan,
and finally some editorializing from Neil Conway.
Neil Conway [Mon, 2 Feb 2004 00:35:08 +0000 (00:35 +0000)]
Use Tcl_PutEnv() in place of putenv() in libpgtcl, as the latter can
apparently corrupt the Tcl runtime. Per suggestion from Michael
Brusser <michael@synchronicity.com>
Bruce Momjian [Mon, 2 Feb 2004 00:11:31 +0000 (00:11 +0000)]
Briefly,
* configure + Makefile changes
* shared memory attaching in EXEC_BACKEND case (+ minor fix for apparent
cygwin bug under cygwin/EXEC_BACKEND case only)
* PATH env var separator differences
* missing win32 rand functions added
* placeholder replacements for sync etc under port.h
To those who are really interested, and there are a few of you: the attached
patch + file will allow the source base to be compiled (and, for some
definition, "run") under MingW, with the following caveats (I wanted to
first properly fix all but the last of these, but y'all won't quit asking
for a patch :-):
* child death: SIGCHLD not yet sent, so as a minimum, you'll need to
put in some sort of delay after StartupDatabase, and handle setting
StartupPID to 0 etc (ie. the stuff the reaper() signal function is supposed
to do)
* dirmod.c: comment out the elog calls
* dfmgr.c: some hackage required to substitute_libpath_macro
* slru/xact.c: comment out the errno checking after the readdir
(fixed by next version of MingW)
Again, this is only if you *really* want to see postgres compile and start,
and is a nice leg-up for working on the other Win32 TODO list items. Just
don't expect too much else from it at this point...
Tom Lane [Sun, 1 Feb 2004 06:27:48 +0000 (06:27 +0000)]
Make length() disregard trailing spaces in char(n) values, per discussion
some time ago and recent patch from Gavin Sherry. Update documentation
to point out that trailing spaces are insignificant in char(n).
Tom Lane [Sat, 31 Jan 2004 00:45:21 +0000 (00:45 +0000)]
Fix text_position to not scan past end of source string in multibyte
case, per report from Korea PostgreSQL Users' Group. Also do some
cosmetic cleanup in nearby code.
Tom Lane [Wed, 28 Jan 2004 22:14:04 +0000 (22:14 +0000)]
Make USING and WITH optional to bring the syntax of \copy into exact
agreement with what the backend grammar actually accepts (which is a
bit looser than what its documentation claims). Per report from Bill
Moran, though I did not use his patch since it removed all the
undocumented flexibility that the code historically had and the backend
still has.
Tom Lane [Wed, 28 Jan 2004 21:02:40 +0000 (21:02 +0000)]
Review uses of IsUnderPostmaster, change some tests to look at
whereToSendOutput instead because they are really inquiring about
the correct client communication protocol. Update some comments.
This is pointing towards supporting regular FE/BE client protocol
in a standalone backend, per discussion a month or so back.
Michael Meskes [Wed, 28 Jan 2004 09:52:14 +0000 (09:52 +0000)]
- Issue a warning if a cursor is declared but not opened.
- Fixed prototype for ECPGprepared_statement to not moan about "const char"
- Fixed parsing of nested structures.
- Added option to parse header files.
Tom Lane [Wed, 28 Jan 2004 07:46:44 +0000 (07:46 +0000)]
Fix oversight in check_ungrouped_columns optimization that avoids
unnecessary checks for complex grouping expressions: we cannot check
whether the expressions are simple Vars until after we apply
flatten_join_alias_vars, because in the case of FULL JOIN that routine
can introduce non-Var expressions. Per example from Joel Knight.
Tom Lane [Wed, 28 Jan 2004 00:05:04 +0000 (00:05 +0000)]
simplify_function() mustn't try to evaluate functions that return
composite types, because TupleTableSlots aren't Datums and can't be
stored in Const nodes. We can remove this restriction if we ever
adopt a cleaner runtime representation for whole-tuple results, but
at the moment it's broken. Per example from Thomas Hallgren.
Tom Lane [Tue, 27 Jan 2004 00:50:33 +0000 (00:50 +0000)]
Run stats test separately, not as part of a parallel group, to avoid
possible failures due to stats collector not updating fast enough under
heavy load. Per report from Jeremy Yoder.
Bruce Momjian [Tue, 27 Jan 2004 00:45:26 +0000 (00:45 +0000)]
Here's the latest win32 signals code, this time in the form of a patch
against the latest shapshot. It also includes the replacement of kill()
with pqkill() and sigsetmask() with pqsigsetmask().
Passes all tests fine on my linux machine once applied. Still doesn't
link completely on Win32 - there are a few things still required. But
much closer than before.
At Bruce's request, I'm goint to write up a README file about the method
of signals delivery chosen and why the others were rejected (basically a
summary of the mailinglist discussions). I'll finish that up once/if the
patch is accepted.
Bruce Momjian [Mon, 26 Jan 2004 22:59:54 +0000 (22:59 +0000)]
[all] Removed call to getppid in SendPostmasterSignal, replacing with a
PostmasterPid variable, which gets set (early) in PostmasterMain
getppid would not be the postmaster?
[fork/exec] Implements processCancelRequest by keeping an array of
pid/cancel_key structs in shared mem
[fork/exec] Moves AttachSharedMemoryAndSemaphores call for backends into
SubPostmasterMain
[win32] Implements reaper/waitpid by keeping an arrays of children
pids,handles in postmaster local mem
- this item is largely untested, for reasons which should be
obvious, but appears sound
[win32/all] Added extern for pgpipe in Win32 case, and changed the second
pipe call (which seems to have been missed earlier) to pgpipe
[win32] #define'd ftruncate to chsize in the Win32 case
[win32] PG_USLEEP for Win32 has a misplaced paren. Fixed.
Tom Lane [Mon, 26 Jan 2004 22:35:32 +0000 (22:35 +0000)]
Ensure that close() and fclose() are checked for errors, at least in
cases involving writes. Per recent discussion about the possibility
of close-time failures on some filesystems. There is a TODO item for
this, too.
Neil Conway [Mon, 26 Jan 2004 22:13:21 +0000 (22:13 +0000)]
The attached patch clarifies (or, rather, makes explicit) to readers how
to handle memory management for char pointers returned by libpq functions.
Original patch by Gavin Sherry, some tweaking and consistency improvements
by Neil Conway.
Tom Lane [Mon, 26 Jan 2004 19:15:59 +0000 (19:15 +0000)]
Repair incorrect order of operations in GetNewTransactionId(). We must
complete ExtendCLOG() before advancing nextXid, so that if that routine
fails, the next incoming transaction will try it again. Per trouble
report from Christopher Kings-Lynne.
Tom Lane [Sat, 24 Jan 2004 23:45:13 +0000 (23:45 +0000)]
Support named parameters in information_schema.parameters,
per Dennis Bjorklund. I did not force initdb for this, but
you'd need to do one to get the improved view.
Neil Conway [Sat, 24 Jan 2004 20:43:26 +0000 (20:43 +0000)]
Revert the change to print.c, as this breaks src/bin/scripts. The proper
way to fix this is probably implementing safe memory handling functions
once in a static lib and then using that in the various client apps,
but for the moment I've just reverted the change to un-break the tree.
Neil Conway [Sat, 24 Jan 2004 19:38:49 +0000 (19:38 +0000)]
This patch makes some of the memory manipulation performed by psql a
little more sane. Some parts of the code was using a static function
xmalloc() that did safe memory allocation (where "safe" means "bail
out on OOM"), but most of it was just invoking calloc() or malloc()
directly. Now almost everything invokes xmalloc() or xcalloc().
Tom Lane [Sat, 24 Jan 2004 00:37:28 +0000 (00:37 +0000)]
Repair planner failure for cases involving Cartesian products inside
IN (sub-SELECT) constructs. We must force a clauseless join of the
sub-select member relations, but it wasn't happening because the code
thought it would be able to use the join clause arising from the IN.
Tom Lane [Fri, 23 Jan 2004 23:54:21 +0000 (23:54 +0000)]
Revise GEQO planner to make use of some heuristic knowledge about SQL, namely
that it's good to join where there are join clauses rather than where there
are not. Also enable it to generate bushy plans at need, so that it doesn't
fail in the presence of multiple IN clauses containing sub-joins. These
changes appear to improve the behavior enough that we can substantially reduce
the default pool size and generations count, thereby decreasing the runtime,
and yet get as good or better plans as we were getting in 7.4. Consequently,
adjust the default GEQO parameters. I also modified the way geqo_effort is
used so that it affects both population size and number of generations;
it's now useful as a single control to adjust the GEQO runtime-vs-plan-quality
tradeoff. Bump geqo_threshold to 12, since even with these changes GEQO
seems to be slower than the regular planner at 11 relations.
Neil Conway [Fri, 23 Jan 2004 02:13:12 +0000 (02:13 +0000)]
Fix a minor bug introduced by the recent CREATE TABLE AS / WITH OIDS
patch: a 3-value enum was mistakenly assigned directly to a 'bool'
in transformCreateStmt(). Along the way, change makeObjectName()
to be static, as it isn't used outside analyze.c
Tom Lane [Thu, 22 Jan 2004 02:23:21 +0000 (02:23 +0000)]
Fix oversight in optimization that avoids an unnecessary projection step
when scanning a table that we need all the columns from. In case of
SELECT INTO, we have to check that the hasoids flag matches the desired
output type, too. Per report from Mike Mascari.
Tom Lane [Wed, 21 Jan 2004 23:33:34 +0000 (23:33 +0000)]
Repair error apparently introduced in the initial coding of GUC: the
default value for geqo_effort is supposed to be 40, not 1. The actual
'genetic' component of the GEQO algorithm has been practically disabled
since 7.1 because of this mistake. Improve documentation while at it.
Tom Lane [Wed, 21 Jan 2004 19:22:19 +0000 (19:22 +0000)]
If we don't have shared libraries, we don't have conversions. Make
conversion_create.sql be empty (except for a helpful comment) in this
case. Allows initdb to succeed with --disable-shared.
Tom Lane [Wed, 21 Jan 2004 19:04:11 +0000 (19:04 +0000)]
Fix bit-rot in support for building with --disable-shared. This patch
gets us past 'make install', but initdb still fails for lack of conversion
libraries ...