Tom Lane [Sun, 13 Nov 2005 19:12:05 +0000 (19:12 +0000)]
Force the second argument of SUBSTRING(foo FOR bar) to be int4, to avoid
surprising results when it's some other numeric type. This doesn't solve
the generic problem of surprising implicit casts to text, but it's a
low-impact way of making sure this particular case behaves sanely.
Per gripe from Harald Fuchs and subsequent discussion.
Tom Lane [Thu, 10 Nov 2005 00:31:49 +0000 (00:31 +0000)]
When in transaction-aborted state, reject Bind message for portals containing
anything but transaction-exiting commands (ROLLBACK etc). We already rejected
Parse and Execute in such cases, so there seems little point in allowing Bind.
This prevents at least an Assert failure, and probably worse things, since
there's a lot of infrastructure that doesn't work when not in a live
transaction. We can also simplify the Bind logic a bit by rejecting messages
with a nonzero number of parameters, instead of the former kluge to silently
substitute NULL for each parameter. Per bug #2033 from Joel Stevenson.
Tom Lane [Sat, 5 Nov 2005 03:05:05 +0000 (03:05 +0000)]
Repair an error introduced by log_line_prefix patch: it is not acceptable
to assume that the string pointer passed to set_ps_display is good forever.
There's no need to anyway since ps_status.c itself saves the string, and
we already had an API (get_ps_display) to return it.
I believe this explains Jim Nasby's report of intermittent crashes in
elog.c when %i format code is in use in log_line_prefix.
While at it, repair a previously unnoticed problem: on some platforms such as
Darwin, the string returned by get_ps_display was blank-padded to the maximum
length, meaning that lock.c's attempt to append " waiting" to it never worked.
Tom Lane [Fri, 4 Nov 2005 18:35:51 +0000 (18:35 +0000)]
Ensure that we only create one ConsoleCtrlHandler per psql process,
so as to avoid performance issues and possible ultimate crash on long
psql scripts. Per Merlin Moncure.
Tom Lane [Thu, 3 Nov 2005 00:23:43 +0000 (00:23 +0000)]
Fix longstanding race condition in transaction log management: there was a
very narrow window in which SimpleLruReadPage or SimpleLruWritePage could
think that I/O was needed when it wasn't (and indeed the buffer had already
been assigned to another page). This would result in an Assert failure if
Asserts were enabled, and probably in silent data corruption if not.
Reported independently by Jim Nasby and Robert Creager.
I intend a more extensive fix when 8.2 development starts, but this is a
reasonably low-impact patch for the existing branches.
Andrew Dunstan [Tue, 1 Nov 2005 15:32:52 +0000 (15:32 +0000)]
Provide a --no-locale option for pg_regress and a corresponding NOLOCALE=1
setting for the regression makefile, allowing Windows users to force locale
settings since Windows does not get its locale from the environment.
Bruce Momjian [Wed, 26 Oct 2005 13:43:28 +0000 (13:43 +0000)]
Properly update the 'group' flatfile when modifying the user, in case
they were added to a group. Also fix visibility of our own changes when
creating the group file. This fixes:
test=> CREATE GROUP g1;
CREATE GROUP
test=> CREATE USER u1 IN GROUP g1;
CREATE USER
test=> \! cat /u/pg/data/global/pg_group
"g1" "u1"
test=> CREATE USER u2 IN GROUP g1;
CREATE USER
test=> \! cat /u/pg/data/global/pg_group
"g1" "u1" "u2"
test=> ALTER USER u2 RENAME TO u3;
ALTER USER
test=> \! cat /u/pg/data/global/pg_group
"g1" "u1" "u3"
Tom Lane [Tue, 25 Oct 2005 20:30:35 +0000 (20:30 +0000)]
Fix longstanding bug that would sometimes let the planner generate a bad plan
for an outer join; symptom is bogus error "RIGHT JOIN is only supported with
merge-joinable join conditions". Problem was that select_mergejoin_clauses
did its tests in the wrong order. We need to force left join not right join
for a merge join when there are non-mergeable join clauses; but the test for
this only accounted for mergejoinability of the clause operator, and not
whether the left and right Vars were of the proper relations. Per report
from Jean-Pierre Pelletier.
Tom Lane [Thu, 20 Oct 2005 20:06:03 +0000 (20:06 +0000)]
Postpone pg_timezone_initialize() until after creation of postmaster.pid,
since it can take a fair amount of time and this can confuse boot scripts
that expect postmaster.pid to appear quickly. Move initialization of SSL
library and preloaded libraries to after that point, too, just for luck.
Per reports from Tony Caduto and others.
Tom Lane [Fri, 14 Oct 2005 16:41:13 +0000 (16:41 +0000)]
Pass a strdup'd ident string to openlog(), to ensure that reallocation
of GUC memory doesn't cause us to start emitting a bogus ident string.
Per report from Han Holl. Also some trivial code cleanup in write_syslog.
Tom Lane [Wed, 12 Oct 2005 17:18:15 +0000 (17:18 +0000)]
Fix longstanding bug found by Atsushi Ogawa: _bt_check_unique would mark
the wrong buffer dirty when trying to kill a dead index entry that's on
a page after the one it started on. No risk of data corruption, just
inefficiency, but still a bug.
Tom Lane [Tue, 11 Oct 2005 16:45:00 +0000 (16:45 +0000)]
Fix oversight in 8.0 modification of RestrictInfo data structures.
A RestrictInfo representing an OR clause now contains two versions of
the contained expression, one with sub-RestrictInfos and one without.
clause_selectivity() should descend to the version with sub-RestrictInfos
so that it has a chance of caching its results for the OR's sub-clauses.
Failing to do so resulted in redundant planner effort.
Tom Lane [Wed, 28 Sep 2005 21:17:24 +0000 (21:17 +0000)]
Repair planning bug introduced in 7.4: outer-join ON clauses that referenced
only the inner-side relation would be considered as potential equijoin clauses,
which is wrong because the condition doesn't necessarily hold above the point
of the outer join. Per test case from Kevin Grittner (bug#1916).
Bruce Momjian [Fri, 23 Sep 2005 21:02:37 +0000 (21:02 +0000)]
Backpatch to 8.0.X:
In several places PL/Python was calling PyObject_Str() and then
PyString_AsString() without checking if the former had returned
NULL to indicate an error. PyString_AsString() doesn't expect a
NULL argument, so passing one causes a segmentation fault. This
patch adds checks for NULL and raises errors via PLy_elog(), which
prints details of the underlying Python exception. The patch also
adds regression tests for these checks. All tests pass on my
Solaris 9 box running HEAD and Python 2.4.1.
Update Snowball. I have to update it because of
old version doesn't available on Snowball's site and new version
of stemmers can't be compiled with old interface.
Tom Lane [Mon, 12 Sep 2005 22:20:30 +0000 (22:20 +0000)]
Ensure that any memory leaked during an error inside the bgwriter is
recovered. I did not see any actual leak while testing this in CVS tip,
but 8.0 definitely has a problem with leaking the space temporarily
palloc'd by BufferSync(). In any case this seems a good idea to forestall
similar problems in future. Per report from Arjen van der Meijden.
Tom Lane [Sun, 11 Sep 2005 00:36:35 +0000 (00:36 +0000)]
Avoid changing stdin/stdout to binary mode on Windows unless that is
really the source or destination of the archive. I think this will
resolve recent complaints that password prompting is broken in pg_restore
on Windows. Note that password prompting and reading from stdin is an
unworkable combination on Windows ... but that was true anyway.
Tom Lane [Wed, 7 Sep 2005 21:39:41 +0000 (21:39 +0000)]
Update timezone data files to release 2005m of the zic database.
Among other changes, this reflects the recently passed change in USA
daylight savings rules.
Tom Lane [Sat, 27 Aug 2005 16:22:58 +0000 (16:22 +0000)]
Get the MIPS assembler syntax right. Also add a separate sync command;
the reference I consulted yesterday said SC does a SYNC, but apparently
this is not true on newer MIPS processors, so be safe.
Tom Lane [Thu, 25 Aug 2005 19:45:06 +0000 (19:45 +0000)]
Back-patch fixes for problems with VACUUM destroying t_ctid chains too soon,
and with insufficient paranoia in code that follows t_ctid links.
This patch covers the 8.0 branch.
Michael Meskes [Wed, 24 Aug 2005 10:35:12 +0000 (10:35 +0000)]
- Check for NULL before checking whether argument is an array.
- Removed stray character from string quoting.
- Fixed check to report missing varchar pointer implementation.
Tom Lane [Sat, 20 Aug 2005 19:19:31 +0000 (19:19 +0000)]
Invoke mksafefunc and mkunsafefunc with :: decoration. This seems a good
idea on consistency grounds, whether or not it really fixes bug #1831.
Michael Fuhr
Tom Lane [Tue, 16 Aug 2005 00:48:29 +0000 (00:48 +0000)]
Reject operator names >= NAMEDATALEN characters. These will not work
anyway, and in assert-enabled builds you are likely to get an assertion
failure. Backpatch as far as 7.3; 7.2 seems not to have the problem.
Tom Lane [Mon, 15 Aug 2005 19:40:43 +0000 (19:40 +0000)]
array_in() and array_recv() need to be more paranoid about validating
their OID parameter. It was possible to crash the backend with
select array_in('{123}',0,0); because that would bypass the needed step
of initializing the workspace. These seem to be the only two places
with a problem, though (record_in and record_recv don't have the issue,
and the other array functions aren't depending on user-supplied input).
Back-patch as far as 7.4; 7.3 does not have the bug.
Tom Lane [Mon, 15 Aug 2005 19:05:30 +0000 (19:05 +0000)]
int_array_enum function should be using fcinfo->flinfo->fn_extra for
working state, not fcinfo->context. Silly oversight on my part in last
go-round of fixes.
Bruce Momjian [Fri, 12 Aug 2005 21:38:00 +0000 (21:38 +0000)]
This patch fixes the event type used to log output from the
stderr-in-service or output-from-syslogger-in-service code. Previously
everything was flagged as ERRORs there, which caused all instances to
log "LOG: logger shutting down" as error...
Please apply for 8.1. I'd also like it considered for 8.0 since logging
non-errors as errors can be cause for alarm amongst people who actually
look at their logs...
Bruce Momjian [Fri, 12 Aug 2005 21:23:18 +0000 (21:23 +0000)]
[ backpatched to 8.0.X.]
> >> 3) I restarted the postmaster both times. I got this error
> both times.
> >> :25: ERROR: could not load library "C:/Program
> >> Files/PostgreSQL/8.0/lib/testtrigfuncs.dll": dynamic load error
>
> > Yes. We really need to look at fixing that error message. I had
> > forgotten it completely :-(
>
> > Bruce, you think we can sneak that in after feature freeze? I would
> > call it a bugfix :-)
>
> Me too. That's been on the radar for awhile --- please do
> send in a patch.
Here we go, that wasn't too hard :-)
Apart from adding the error handling, it does one more thing: it changes
the errormode when loading the DLLs. Previously if a DLL was broken, or
referenced other DLLs that couldn't be found, a popup dialog box would
appear on the screen. Which had to be clicked before the backend could
continue. This patch also disables the popup error message for DLL
loads.
I think this is something we should consider doing for the entire
backend - disable those popups, and say we deal with it ourselves. What
do you other win32 hackers thinnk about this?
In the meantime, this patch fixes the error msgs. Please apply for 8.1
and please consider a backpatch to 8.0.
Tom Lane [Mon, 8 Aug 2005 23:39:14 +0000 (23:39 +0000)]
Fix crash when reading 'timezone = unknown' from postgresql.conf during
SIGHUP; it's not OK for an assign_hook to return a non-malloc'd string.
Problem was introduced during timezone library rewrite.
Tom Lane [Sun, 7 Aug 2005 18:47:38 +0000 (18:47 +0000)]
Fix count_usable_fds() to stop trying to open files once it reaches
max_files_per_process. Going further than that is just a waste of
cycles, and it seems that current Cygwin does not cope gracefully
with deliberately running the system out of FDs. Per Andrew Dunstan.
Tom Lane [Fri, 22 Jul 2005 19:12:33 +0000 (19:12 +0000)]
Fix compare_fuzzy_path_costs() to behave a bit more sanely. The original
coding would ignore startup cost differences of less than 1% of the
estimated total cost; which was OK for normal planning but highly not OK
if a very small LIMIT was applied afterwards, so that startup cost becomes
the name of the game. Instead, compare startup and total costs fuzzily
but independently. This changes the plan selected for two queries in the
regression tests; adjust expected-output files for resulting changes in
row order. Per reports from Dawid Kuroczko and Sam Mason.
Tom Lane [Mon, 18 Jul 2005 21:37:23 +0000 (21:37 +0000)]
It appears that Darwin (OS X) does not cope well with C functions that
have the same name as the containing shared library --- as best I can
tell, the compiler internally creates a function of that name, and does
not warn you about the conflict. Fix buildfarm failure in back branches
by renaming tsearch() trigger function at the C level.
Tom Lane [Mon, 18 Jul 2005 15:53:46 +0000 (15:53 +0000)]
MemSet() must not cast its pointer argument to int32* until after it has
checked that the pointer is actually word-aligned. Casting a non-aligned
pointer to int32* is technically illegal per the C spec, and some recent
versions of gcc actually generate bad code for the memset() when given
such a pointer. Per report from Andrew Morrow.
Tom Lane [Sun, 17 Jul 2005 18:29:13 +0000 (18:29 +0000)]
Make pg_regress accept a command-line option for the temporary installation's
port number, and use a default value for it that is dependent on the
configuration-time DEF_PGPORT. Should make the world safe for running
parallel 'make check' in different branches. Back-patch as far as 7.4
so that this actually is useful.
Tom Lane [Sun, 17 Jul 2005 04:05:49 +0000 (04:05 +0000)]
Back-patch recent changes to alter the order of -L flags inserted from
LDFLAGS versus those built into the Makefiles. This looks like it will
fix several buildfarm failures in the back branches.
Tom Lane [Fri, 15 Jul 2005 18:40:20 +0000 (18:40 +0000)]
Check for out-of-range varoattno in deparse_context_for_subplan.
I have seen this case in CVS tip due to new "physical tlist" optimization
for subqueries. I believe it probably can't happen in existing releases,
but the check is not going to hurt anything, so backpatch to 8.0 just
in case.
Tom Lane [Fri, 15 Jul 2005 17:09:50 +0000 (17:09 +0000)]
Fix overenthusiastic optimization of 'x IN (SELECT DISTINCT ...)' and related
cases: we can't just consider whether the subquery's output is unique on its
own terms, we have to check whether the set of output columns we are going to
use will be unique. Per complaint from Luca Pireddu and test case from
Michael Fuhr.
Tom Lane [Mon, 27 Jun 2005 02:18:14 +0000 (02:18 +0000)]
Modify pg_dump to assume that a check constraint is inherited if its
name matches the name of any parent-table constraint, without looking
at the constraint text. This is a not-very-bulletproof workaround for
the problem exhibited by Berend Tober last month. We really ought to
record constraint inheritance status in pg_constraint, but it's looking
like that may not get done for 8.1 --- and even if it does, we will
need this kluge for dumping from older servers.
Tom Lane [Sat, 25 Jun 2005 22:47:49 +0000 (22:47 +0000)]
Force a checkpoint before committing a CREATE DATABASE command. This
should fix the recent reports of "index is not a btree" failures,
as well as preventing a more obscure race condition involving changes
to a template database just after copying it with CREATE DATABASE.
Tom Lane [Sat, 25 Jun 2005 16:54:00 +0000 (16:54 +0000)]
Fix ancient memory leak in index_create(): RelationInitIndexAccessInfo
was being called twice in normal operation, leading to a leak of one set
of relcache subsidiary info. Per report from Jeff Gold.
Neil Conway [Wed, 22 Jun 2005 02:02:09 +0000 (02:02 +0000)]
Correct some code in pg_restore when reading the header of a tar archive:
(1) The code doesn't initialize `sum', so the initial "does the checksum
match?" test is wrong.
(2) The loop that is intended to check for a "null block" just checks
the first byte of the tar block 512 times, rather than each of the
512 bytes one time (!), which I'm guessing was the intent.
It was only through sheer luck that this worked in the first place.
Per Coverity static analysis performed by EnterpriseDB.
Neil Conway [Tue, 21 Jun 2005 01:23:25 +0000 (01:23 +0000)]
Fix a potential backend crash during authentication when parsing a
malformed ident map file. This was introduced by the linked list
rewrite in 8.0 -- mea maxima culpa.
Per Coverity static analysis performed by EnterpriseDB.
Tom Lane [Mon, 20 Jun 2005 22:51:49 +0000 (22:51 +0000)]
exec_eval_datum leaks memory when dealing with ROW or REC values.
It never leaked memory before PG 8.0, so none of the callers are
expecting this. Cleanest fix seems to be to make it allocate the needed
memory in estate->eval_econtext, where it will be cleaned up by
the next exec_eval_cleanup. Per report from Bill Rugolsky.
Tom Lane [Mon, 20 Jun 2005 20:44:50 +0000 (20:44 +0000)]
plpgsql's exec_assign_value() freed the old value of a variable before
copying/converting the new value, which meant that it failed badly on
"var := var" if var is of pass-by-reference type. Fix this and a similar
hazard in exec_move_row(); not sure that the latter can manifest before
8.0, but patch it all the way back anyway. Per report from Dave Chapeskie.
Tom Lane [Sat, 18 Jun 2005 20:51:44 +0000 (20:51 +0000)]
When using C-string lookup keys in a dynahash.c hash table, use strncpy()
not memcpy() to copy the offered key into the hash table during HASH_ENTER.
This avoids possible core dump if the passed key is located very near the
end of memory. Per report from Stefan Kaltenbrunner.
Tom Lane [Tue, 14 Jun 2005 14:21:23 +0000 (14:21 +0000)]
The random selection in function linear() could deliver a value equal to max
if geqo_rand() returns exactly 1.0, resulting in failure due to indexing
off the end of the pool array. Also, since this is using inexact float math,
it seems wise to guard against roundoff error producing values slightly
outside the expected range. Per report from bug@zedware.org.