Tom Lane [Tue, 11 Sep 2007 17:15:48 +0000 (17:15 +0000)]
Make sure that open hash table scans are cleaned up when bgwriter tries to
recover from elog(ERROR). Problem was created by introduction of hash seq
search tracking awhile back, and affects all branches that have bgwriter;
in HEAD the disease has snuck into autovacuum and walwriter too. (Not sure
that the latter two use hash_seq_search at the moment, but surely they might
someday.) Per report from Sergey Koposov.
Remove the vacuum_delay_point call in count_nondeletable_pages, because we hold
an exclusive lock on the table at this point, which we want to release as soon
as possible. This is called in the phase of lazy vacuum where we truncate the
empty pages at the end of the table.
An alternative solution would be to lower the vacuum delay settings before
starting the truncating phase, but this doesn't work very well in autovacuum
due to the autobalancing code (which can cause other processes to change our
cost delay settings). This case could be considered in the balancing code, but
it is simpler this way.
Improve page split in rtree emulation. Now if splitted result has
big misalignement, then it tries to split page basing on distribution
of boxe's centers.
Per report from Dolafi, Tom <dolafit@janelia.hhmi.org>
Tom Lane [Wed, 29 Aug 2007 16:31:51 +0000 (16:31 +0000)]
Fix aboriginal bug in _tarAddFile(): when complaining that the amount of data
read from the temp file didn't match the file length reported by ftello(),
the wrong variable's value was printed, and so the message made no sense.
Clean up a couple other coding infelicities while at it.
Tom Lane [Sat, 25 Aug 2007 19:08:31 +0000 (19:08 +0000)]
Fix brain fade in DefineIndex(): it was continuing to access the table's
relcache entry after having heap_close'd it. This could lead to misbehavior
if a relcache flush wiped out the cache entry meanwhile. In 8.2 there is a
very real risk of CREATE INDEX CONCURRENTLY using the wrong relid for locking
and waiting purposes. I think the bug is only cosmetic in 8.0 and 8.1,
because their transgression is limited to using RelationGetRelationName(rel)
in an ereport message immediately after heap_close, and there's no way (except
with special debugging options) for a cache flush to occur in that interval.
Not quite sure that it's cosmetic in 7.4, but seems best to patch anyway.
Found by trying to run the regression tests with CLOBBER_CACHE_ALWAYS enabled.
Maybe we should try to do that on a regular basis --- it's awfully slow,
but perhaps some fast buildfarm machine could do it once in awhile.
Tom Lane [Thu, 23 Aug 2007 16:16:05 +0000 (16:16 +0000)]
Fix combo_decrypt() to throw an error for zero-length input when using a
padded encryption scheme. Formerly it would try to access res[(unsigned) -1],
which resulted in core dumps on 64-bit machines, and was certainly trouble
waiting to happen on 32-bit machines (though in at least the known case
it was harmless because that byte would be overwritten after return).
Per report from Ken Colson; fix by Marko Kreen.
Tom Lane [Tue, 21 Aug 2007 02:40:18 +0000 (02:40 +0000)]
Fix potential access-off-the-end-of-memory in varbit_out(): it fetched the
byte after the last full byte of the bit array, regardless of whether that
byte was part of the valid data or not. Found by buildfarm testing.
Thanks to Stefan Kaltenbrunner for nailing down the cause.
Tom Lane [Wed, 15 Aug 2007 19:16:04 +0000 (19:16 +0000)]
Repair problems occurring when multiple RI updates have to be done to the same
row within one query: we were firing check triggers before all the updates
were done, leading to bogus failures. Fix by making the triggers queued by
an RI update go at the end of the outer query's trigger event list, thereby
effectively making the processing "breadth-first". This was indeed how it
worked pre-8.0, so the bug does not occur in the 7.x branches.
Per report from Pavel Stehule.
Neil Conway [Wed, 8 Aug 2007 18:07:02 +0000 (18:07 +0000)]
Fix a gradual memory leak in ExecReScanAgg(). Because the aggregation
hash table is allocated in a child context of the agg node's memory
context, MemoryContextReset() will reset but *not* delete the child
context. Since ExecReScanAgg() proceeds to build a new hash table
from scratch (in a new sub-context), this results in leaking the
header for the previous memory context. Therefore, use
MemoryContextResetAndDeleteChildren() instead.
Credit: My colleague Sailesh Krishnamurthy at Truviso for isolating
the cause of the leak.
Tom Lane [Sat, 4 Aug 2007 01:42:34 +0000 (01:42 +0000)]
Suppress time zone name (%Z) when logging timestamps in xlog.c startup
on Windows. This is yet another manifestation of the problem that Windows
returns time zone names that may be in a different encoding than we are using.
I've put a better solution in HEAD, but the back branches need a simple patch.
Per report from Hiroshi Saito.
Tom Lane [Tue, 31 Jul 2007 19:54:01 +0000 (19:54 +0000)]
Fix a bug in the original implementation of redundant-join-clause removal:
clauses in which one side or the other references both sides of the join
cannot be removed as redundant, because that expression won't have been
constrained below the join. Per report from Sergey Burladyan.
Tom Lane [Tue, 31 Jul 2007 15:50:01 +0000 (15:50 +0000)]
Fix security definer functions with polymorphic arguments. This case has
never worked because fmgr_security_definer() neglected to pass the fn_expr
information through. Per report from Viatcheslav Kalinin.
Tom Lane [Sat, 21 Jul 2007 22:12:17 +0000 (22:12 +0000)]
Fix elog.c to avoid infinite recursion (leading to backend crash) when
log_min_error_statement is active and there is some problem in logging the
current query string; for example, that it's too long to include in the log
message without running out of memory. This problem has existed since the
log_min_error_statement feature was introduced. No doubt the reason it
wasn't detected long ago is that 8.2 is the first release that defaults
log_min_error_statement to less than PANIC level.
Per report from Bill Moran.
Tom Lane [Fri, 20 Jul 2007 16:30:05 +0000 (16:30 +0000)]
Fix WAL replay of truncate operations to cope with the possibility that the
truncated relation was deleted later in the WAL sequence. Since replay
normally auto-creates a relation upon its first reference by a WAL log entry,
failure is seen only if the truncate entry happens to be the first reference
after the checkpoint we're restarting from; which is a pretty unusual case but
of course not impossible. Fix by making truncate entries auto-create like
the other ones do. Per report and test case from Dharmendra Goyal.
Tom Lane [Thu, 19 Jul 2007 20:34:34 +0000 (20:34 +0000)]
Make replace(), split_part(), and string_to_array() behave somewhat sanely
when handed an invalidly-encoded pattern. The previous coding could get
into an infinite loop if pg_mb2wchar_with_len() returned a zero-length
string after we'd tested for nonempty pattern; which is exactly what it
will do if the string consists only of an incomplete multibyte character.
This led to either an out-of-memory error or a backend crash depending
on platform. Per report from Wiktor Wodecki.
Andrew Dunstan [Thu, 19 Jul 2007 19:14:54 +0000 (19:14 +0000)]
Only use the pipe chunking protocol if we know the syslogger should
be catching stderr output, and we are not ourselves the
syslogger. Otherwise, go directly to stderr.
Bug noticed by Tom Lane.
Backpatch as far as 8.0.
Tom Lane [Wed, 18 Jul 2007 21:41:22 +0000 (21:41 +0000)]
Fix an old thinko in SS_make_initplan_from_plan, which is used when optimizing
a MIN or MAX aggregate call into an indexscan: the initplan is being made at
the current query nesting level and so we shouldn't increment query_level.
Though usually harmless, this mistake could lead to bogus "plan should not
reference subplan's variable" failures on complex queries. Per bug report
from David Sanchez i Gregori.
Tom Lane [Tue, 17 Jul 2007 17:45:48 +0000 (17:45 +0000)]
Fix incorrect optimization of foreign-key checks. When an UPDATE on the
referencing table does not change the tuple's FK column(s), we don't bother
to check the PK table since the constraint was presumably already valid.
However, the check is still necessary if the tuple was inserted by our own
transaction, since in that case the INSERT trigger will conclude it need not
make the check (since its version of the tuple has been deleted). We got this
right for simple cases, but not when the insert and update are in different
subtransactions of the current top-level transaction; in such cases the FK
check would never be made at all. (Hence, problem dates back to 8.0 when
subtransactions were added --- it's actually the subtransaction version of a
bug fixed in 7.3.5.) Fix, and add regression test cases. Report and fix by
Affan Salman.
Tom Lane [Tue, 17 Jul 2007 01:22:03 +0000 (01:22 +0000)]
Fix outfuncs.c to dump A_Const nodes representing NULLs correctly. This has
been broken since forever, but was not noticed because people seldom look
at raw parse trees. AFAIK, no impact on users except that debug_print_parse
might fail; but patch it all the way back anyway. Per report from Jeff Ross.
Joe Conway [Mon, 9 Jul 2007 01:32:44 +0000 (01:32 +0000)]
Restrict non-superusers to password authenticated connections
to prevent possible escalation of privilege. Provide new SECURITY
DEFINER functions with old behavior, but initially REVOKE ALL
from public for these functions. Per list discussion and design
proposed by Tom Lane.
Tom Lane [Sun, 8 Jul 2007 22:23:32 +0000 (22:23 +0000)]
Remove the pgstat_drop_relation() call from smgr_internal_unlink(), because
we don't know at that point which relation OID to tell pgstat to forget.
The code was passing the relfilenode, which is incorrect, and could possibly
cause some other relation's stats to be zeroed out. While we could try to
clean this up, it seems much simpler and more reliable to let the next
invocation of pgstat_vacuum_tabstat() fix things; which indeed is how it
worked before I introduced the buggy code into 8.1.3 and later :-(.
Problem noticed by Itagaki Takahiro, fix is per subsequent discussion.
Tom Lane [Mon, 2 Jul 2007 20:12:05 +0000 (20:12 +0000)]
Fix failure to restart Postgres when Linux kernel returns EIDRM for shmctl().
This is a Linux kernel bug that apparently exists in every extant kernel
version: sometimes shmctl() will fail with EIDRM when EINVAL is correct.
We were assuming that EIDRM indicates a possible conflict with pre-existing
backends, and refusing to start the postmaster when this happens. Fortunately,
there does not seem to be any case where Linux can legitimately return EIDRM
(it doesn't track shmem segments in a way that would allow that), so we can
get away with just assuming that EIDRM means EINVAL on this platform.
Per reports from Michael Fuhr and Jon Lapham --- it's a bit surprising
we have not seen more reports, actually.
Tom Lane [Fri, 29 Jun 2007 01:51:56 +0000 (01:51 +0000)]
Fix a passel of ancient bugs in to_char(), including two distinct buffer
overruns (neither of which seem likely to be exploitable as security holes,
fortunately, since the provoker can't control the data written). One of
these is due to choosing to stomp on the output of a called function, which
is bad news in any case; make it treat the called functions' results as
read-only. Avoid some unnecessary palloc/pfree traffic too; it's not
really helpful to free small temporary objects, and again this is presuming
more than it ought to about the nature of the results of called functions.
Per report from Patrick Welche and additional code-reading by Imad.
Tom Lane [Thu, 28 Jun 2007 17:50:17 +0000 (17:50 +0000)]
Fix incorrect tests for undef Perl values in some places in plperl.c.
The correct test for defined-ness is SvOK(sv), not anything involving
SvTYPE. Per bug #3415 from Matt Taylor.
Back-patch as far as 8.0; no apparent problem in 7.x.
Tom Lane [Wed, 20 Jun 2007 18:21:15 +0000 (18:21 +0000)]
transformColumnDefinition failed to complain about
create table foo (bar int default null default 3);
due to not thinking about the special-case handling of DEFAULT NULL.
Problem noticed while investigating bug #3396.
Tom Lane [Wed, 20 Jun 2007 18:16:04 +0000 (18:16 +0000)]
CREATE DOMAIN ... DEFAULT NULL failed because gram.y special-cases DEFAULT
NULL and DefineDomain didn't. Bug goes all the way back to original coding
of domains. Per bug #3396 from Sergey Burladyan.
Tom Lane [Mon, 18 Jun 2007 01:14:08 +0000 (01:14 +0000)]
Back-patch 8.2 fix that complains if trying to extend a relation encounters
a buffer containing a non-zeroed page. This seems appropriate now that the
8.2 fix has been seen to save at least one user from data loss due to a
buggy kernel (per report from Jaime Silvela 7-May-07). I'd go further
back than 8.1, except that the 8.0-to-8.1 bufmgr changes are large
enough that the patch doesn't work immediately; I'm hesitant to make a
change without more extensive analysis than I have time for now.
Andrew Dunstan [Thu, 14 Jun 2007 01:50:14 +0000 (01:50 +0000)]
Implement a chunking protocol for writes to the syslogger pipe, with messages
reassembled in the syslogger before writing to the log file. This prevents
partial messages from being written, which mucks up log rotation, and
messages from different backends being interleaved, which causes garbled
logs. Backport as far as 8.0, where the syslogger was introduced.
Tom Lane [Sat, 9 Jun 2007 15:52:47 +0000 (15:52 +0000)]
Allow numeric_fac() to be interrupted, since it can take quite a while for
large inputs. Also cause it to error out immediately if the result will
overflow, instead of grinding through a lot of calculation first.
Per gripe from Jim Nasby.
Magnus Hagander [Mon, 4 Jun 2007 13:39:34 +0000 (13:39 +0000)]
On win32, retry reading when WSARecv returns WSAEWOULDBLOCK. There seem
to be cases when at least Windows 2000 can do this even though select
just indicated that the socket is readable.
Tom Lane [Fri, 1 Jun 2007 23:43:23 +0000 (23:43 +0000)]
Fix aboriginal bug in BufFileDumpBuffer that would cause it to write the
wrong data when dumping a bufferload that crosses a component-file boundary.
This probably has not been seen in the wild because (a) component files are
normally 1GB apiece and (b) non-block-aligned buffer usage is relatively
rare. But it's fairly easy to reproduce a problem if one reduces RELSEG_SIZE
in a test build. Kudos to Kurt Harriman for spotting the bug.
Tom Lane [Fri, 1 Jun 2007 15:58:09 +0000 (15:58 +0000)]
Fix performance problems in multi-batch hash joins by ensuring that we select
a well-randomized batch number even when given a poorly-randomized hash value.
This is a bit inefficient but seems the only practical solution given the
constraint that we can't change the hash functions in released branches.
Per report from Joseph Shraibman.
Applied to 8.1 and 8.2 only --- HEAD is getting a cleaner fix, and 8.0 and
before use different coding that seems less vulnerable.
Tom Lane [Wed, 30 May 2007 21:01:53 +0000 (21:01 +0000)]
Fix overly-strict sanity check in BeginInternalSubTransaction that made it
fail when used in a deferred trigger. Bug goes back to 8.0; no doubt the
reason it hadn't been noticed is that we've been discouraging use of
user-defined constraint triggers. Per report from Frank van Vugt.
Neil Conway [Tue, 29 May 2007 04:59:44 +0000 (04:59 +0000)]
Fix a bug in input processing for the "interval" type. Previously,
"microsecond" and "millisecond" units were not considered valid input
by themselves, which caused inputs like "1 millisecond" to be rejected
erroneously.
Update the docs, add regression tests, and backport to 8.2 and 8.1
Tom Lane [Tue, 22 May 2007 01:40:53 +0000 (01:40 +0000)]
Fix best_inner_indexscan to return both the cheapest-total-cost and
cheapest-startup-cost innerjoin indexscans, and make joinpath.c consider
both of these (when different) as the inside of a nestloop join. The
original design was based on the assumption that indexscan paths always
have negligible startup cost, and so total cost is the only important
figure of merit; an assumption that's obviously broken by bitmap
indexscans. This oversight could lead to choosing poor plans in cases
where fast-start behavior is more important than total cost, such as
LIMIT and IN queries. 8.1-vintage brain fade exposed by an example from
Chuck D.
Tom Lane [Fri, 18 May 2007 01:20:33 +0000 (01:20 +0000)]
Remove redundant logging of send failures when SSL is in use. While pqcomm.c
had been taught not to do that ages ago, the SSL code was helpfully bleating
anyway. Resolves some recent reports such as bug #3266; however the
underlying cause of the related bug #2829 is still unclear.
Alvaro Herrera [Tue, 15 May 2007 20:20:29 +0000 (20:20 +0000)]
Avoid emitting empty role names in the GRANTED BY clause of GRANT ROLE
when the grantor has been dropped. This is a workaround for the fact
that we don't track the grantor as a shared dependency.
Tom Lane [Tue, 1 May 2007 18:54:09 +0000 (18:54 +0000)]
Fix a thinko in my patch of a couple months ago for bug #3116: it did the
wrong thing when inlining polymorphic SQL functions, because it was using the
function's declared return type where it should have used the actual result
type of the current call. In 8.1 and 8.2 this causes obvious failures even if
you don't have assertions turned on; in 8.0 and 7.4 it would only be a problem
if the inlined expression were used as an input to a function that did
run-time type determination on its inputs. Add a regression test, since this
is evidently an under-tested area.
Tom Lane [Thu, 26 Apr 2007 23:25:09 +0000 (23:25 +0000)]
Fix dynahash.c to suppress hash bucket splits while a hash_seq_search() scan
is in progress on the same hashtable. This seems the least invasive way to
fix the recently-recognized problem that a split could cause the scan to
visit entries twice or (with much lower probability) miss them entirely.
The only field-reported problem caused by this is the "failed to re-find
shared lock object" PANIC in COMMIT PREPARED reported by Michel Dorochevsky,
which was caused by multiply visited entries. However, it seems certain
that mdsync() is vulnerable to missing required fsync's due to missed
entries, and I am fearful that RelationCacheInitializePhase2() might be at
risk as well. Because of that and the generalized hazard presented by this
bug, back-patch all the supported branches.
Along the way, fix pg_prepared_statement() and pg_cursor() to not assume
that the hashtables they are examining will stay static between calls.
This is risky regardless of the newly noted dynahash problem, because
hash_seq_search() has never promised to cope with deletion of table entries
other than the just-returned one. There may be no bug here because the only
supported way to call these functions is via ExecMakeTableFunctionResult()
which will cycle them to completion before doing anything very interesting,
but it seems best to get rid of the assumption. This affects 8.2 and HEAD
only, since those functions weren't there earlier.
Tom Lane [Fri, 20 Apr 2007 02:38:05 +0000 (02:38 +0000)]
Support explicit placement of the temporary-table schema within search_path.
This is needed to allow a security-definer function to set a truly secure
value of search_path. Without it, a malicious user can use temporary objects
to execute code with the privileges of the security-definer function. Even
pushing the temp schema to the back of the search path is not quite good
enough, because a function or operator at the back of the path might still
capture control from one nearer the front due to having a more exact datatype
match. Hence, disable searching the temp schema altogether for functions and
operators.
Tom Lane [Thu, 19 Apr 2007 20:24:18 +0000 (20:24 +0000)]
Repair PANIC condition in hash indexes when a previous index extension attempt
failed (due to lock conflicts or out-of-space). We might have already
extended the index's filesystem EOF before failing, causing the EOF to be
beyond what the metapage says is the last used page. Hence the invariant
maintained by the code needs to be "EOF is at or beyond last used page",
not "EOF is exactly the last used page". Problem was created by my patch
of 2006-11-19 that attempted to repair bug #2737. Since that was
back-patched to 7.4, this needs to be as well. Per report and test case
from Vlastimil Krejcir.
Tom Lane [Tue, 17 Apr 2007 20:03:16 +0000 (20:03 +0000)]
Rewrite choose_bitmap_and() to make it more robust in the presence of
competing alternatives for indexes to use in a bitmap scan. The former
coding took estimated selectivity as an overriding factor, causing it to
sometimes choose indexes that were much slower to scan than ones with a
slightly worse selectivity. It was also too narrow-minded about which
combinations of indexes to consider ANDing. The rewrite makes it pay more
attention to index scan cost than selectivity; this seems sane since it's
impossible to have very bad selectivity with low cost, whereas the reverse
isn't true. Also, we now consider each index alone, as well as adding
each index to an AND-group led by each prior index, for a total of about
O(N^2) rather than O(N) combinations considered. This makes the results
much less dependent on the exact order in which the indexes are
considered. It's still a lot cheaper than an O(2^N) exhaustive search.
A prefilter step eliminates all but the cheapest of those indexes using
the same set of WHERE conditions, to keep the effective value of N down in
scenarios where the DBA has created lots of partially-redundant indexes.
Tom Lane [Thu, 12 Apr 2007 17:11:07 +0000 (17:11 +0000)]
Rearrange mdsync() looping logic to avoid the problem that a sufficiently
fast flow of new fsync requests can prevent mdsync() from ever completing.
This was an unforeseen consequence of a patch added in Mar 2006 to prevent
the fsync request queue from overflowing. Problem identified by Heikki
Linnakangas and independently by ITAGAKI Takahiro; fix based on ideas from
Takahiro-san, Heikki, and Tom.
Back-patch as far as 8.1 because a previous back-patch introduced the problem
into 8.1 ...
Tom Lane [Thu, 12 Apr 2007 15:04:47 +0000 (15:04 +0000)]
Cancel pending fsync requests during WAL replay of DROP DATABASE, per bug
report from David Darville. Back-patch as far as 8.1, which may or may not
have the problem but it seems a safe change anyway.
Tom Lane [Mon, 2 Apr 2007 18:49:41 +0000 (18:49 +0000)]
Fix check_sql_fn_retval to allow the case where a SQL function declared to
return void ends with a SELECT, if that SELECT has a single result that is
also of type void. Without this, it's hard to write a void function that
calls another void function. Per gripe from Peter.
Tom Lane [Sat, 17 Mar 2007 03:15:55 +0000 (03:15 +0000)]
SPI_cursor_open failed to enforce that only read-only queries could be
executed in read_only mode. This could lead to various relatively-subtle
failures, such as an allegedly stable function returning non-stable results.
Bug goes all the way back to the introduction of read-only mode in 8.0.
Per report from Gaetano Mendola.
Tom Lane [Wed, 14 Mar 2007 18:49:12 +0000 (18:49 +0000)]
Fix a longstanding bug in VACUUM FULL's handling of update chains. The code
did not expect that a DEAD tuple could follow a RECENTLY_DEAD tuple in an
update chain, but because the OldestXmin rule for determining deadness is a
simplification of reality, it is possible for this situation to occur
(implying that the RECENTLY_DEAD tuple is in fact dead to all observers,
but this patch does not attempt to exploit that). The code would follow a
chain forward all the way, but then stop before a DEAD tuple when backing
up, meaning that not all of the chain got moved. This could lead to copying
the chain multiple times (resulting in duplicate copies of the live tuple at
its end), or leaving dangling index entries behind (which, aside from
generating warnings from later vacuums, creates a risk of wrong query
results or bogus duplicate-key errors once the heap slot the index entry
points to is repopulated).
The fix is to recheck HeapTupleSatisfiesVacuum while following a chain
forward, and to stop if a DEAD tuple is reached. Each contiguous group
of RECENTLY_DEAD tuples will therefore be copied as a separate chain.
The patch also adds a couple of extra sanity checks to verify correct
behavior.
Tom Lane [Wed, 14 Mar 2007 17:38:22 +0000 (17:38 +0000)]
Arrange to install a "posixrules" entry in our timezone database, so that
POSIX-style timezone specs that don't exactly match any database entry will
be treated as having correct USA DST rules. Also, document that this can
be changed if you want to use some other DST rules with a POSIX zone spec.
We could consider changing localtime.c's TZDEFRULESTRING, but since that
facility can only deal with one DST transition rule, it seems fairly useless
now; might as well just plan to override it using a "posixrules" entry.
Backpatch as far as 8.0. There isn't much we can do in 7.x ... either your
libc gets it right, or it doesn't.
Alvaro Herrera [Sun, 11 Mar 2007 06:43:23 +0000 (06:43 +0000)]
Fix a race condition that caused pg_database_size() and pg_tablespace_size()
to fail if an object was removed between calls to ReadDir() and stat().
Per discussion in pgsql-hackers.
Tom Lane [Tue, 6 Mar 2007 22:45:29 +0000 (22:45 +0000)]
Fix oversight in original coding of inline_function(): since
check_sql_fn_retval allows binary-compatibility cases, the expression
extracted from an inline-able SQL function might have a type that is only
binary-compatible with the declared function result type. To avoid possibly
changing the semantics of the expression, we should insert a RelabelType node
in such cases. This has only been shown to have bad consequences in recent
8.1 and up releases, but I suspect there may be failure cases in the older
branches too, so patch it all the way back. Per bug #3116 from Greg Mullane.
Along the way, fix an omission in eval_const_expressions_mutator: it failed
to copy the relabelformat field when processing a RelabelType. No known
observable failures from this, but it definitely isn't intended behavior.
Tom Lane [Thu, 1 Mar 2007 18:50:42 +0000 (18:50 +0000)]
Fix markQueryForLocking() to work correctly in the presence of nested views.
It has been wrong for this case since it was first written for 7.1 :-(
Per report from Pavel HanĂ¡k.
Tom Lane [Sun, 18 Feb 2007 19:49:35 +0000 (19:49 +0000)]
Fix portal management code to support non-default command completion tags for
portals using PORTAL_UTIL_SELECT strategy. This is currently significant only
for FETCH queries, which are supposed to include a count in the tag. Seems
it's been broken since 7.4, but nobody noticed before Knut Lehre.
Tom Lane [Fri, 16 Feb 2007 00:14:16 +0000 (00:14 +0000)]
Restructure code that is responsible for ensuring that clauseless joins are
considered when it is necessary to do so because of a join-order restriction
(that is, an outer-join or IN-subselect construct). The former coding was a
bit ad-hoc and inconsistent, and it missed some cases, as exposed by Mario
Weilguni's recent bug report. His specific problem was that an IN could be
turned into a "clauseless" join due to constant-propagation removing the IN's
joinclause, and if the IN's subselect involved more than one relation and
there was more than one such IN linking to the same upper relation, then the
only valid join orders involve "bushy" plans but we would fail to consider the
specific paths needed to get there. (See the example case added to the join
regression test.) On examining the code I wonder if there weren't some other
problem cases too; in particular it seems that GEQO was defending against a
different set of corner cases than the main planner was. There was also an
efficiency problem, in that when we did realize we needed a clauseless join
because of an IN, we'd consider clauseless joins against every other relation
whether this was sensible or not. It seems a better design is to use the
outer-join and in-clause lists as a backup heuristic, just as the rule of
joining only where there are joinclauses is a heuristic: we'll join two
relations if they have a usable joinclause *or* this might be necessary to
satisfy an outer-join or IN-clause join order restriction. I refactored the
code to have just one place considering this instead of three, and made sure
that it covered all the cases that any of them had been considering.
Backpatch as far as 8.1 (which has only the IN-clause form of the disease).
By rights 8.0 and 7.4 should have the bug too, but they accidentally fail
to fail, because the joininfo structure used in those releases preserves some
memory of there having once been a joinclause between the inner and outer
sides of an IN, and so it leads the code in the right direction anyway.
I'll be conservative and not touch them.
Tom Lane [Tue, 13 Feb 2007 19:39:55 +0000 (19:39 +0000)]
Disallow committing a prepared transaction unless we are in the same database
it was executed in. Someday it might be nice to allow cross-DB commits, but
work would be needed in NOTIFY and perhaps other places. Per Heikki.
Tom Lane [Thu, 8 Feb 2007 18:37:55 +0000 (18:37 +0000)]
Fix an ancient logic error in plpgsql's exec_stmt_block: it thought it could
get away with not (re)initializing a local variable if the variable is marked
"isconst" and not "isnull". Unfortunately it makes this decision after having
already freed the old value, meaning that something like
for i in 1..10 loop
declare c constant text := 'hi there';
leads to subsequent accesses to freed memory, and hence probably crashes.
(In particular, this is why Asif Ali Rehman's bug leads to crash and not
just an unexpectedly-NULL value for SQLERRM: SQLERRM is marked CONSTANT
and so triggers this error.)
The whole thing seems wrong on its face anyway: CONSTANT means that you can't
change the variable inside the block, not that the initializer expression is
guaranteed not to change value across successive block entries. Hence,
remove the "optimization" instead of trying to fix it.
Tom Lane [Thu, 8 Feb 2007 18:37:52 +0000 (18:37 +0000)]
Rearrange use of plpgsql_add_initdatums() so that only the parsing of a
DECLARE section needs to know about it. Formerly, everyplace besides DECLARE
that created variables needed to do "plpgsql_add_initdatums(NULL)" to prevent
those variables from being sucked up as part of a subsequent DECLARE block.
This is obviously error-prone, and in fact the SQLSTATE/SQLERRM patch had
failed to do it for those two variables, leading to the bug recently exhibited
by Asif Ali Rehman: a DECLARE within an exception handler tried to reinitialize
SQLERRM.
Although the SQLSTATE/SQLERRM patch isn't in any pre-8.1 branches, and so
I can't point to a demonstrable failure there, it seems wise to back-patch
this into the older branches anyway, just to keep the logic similar to HEAD.
Tom Lane [Tue, 6 Feb 2007 22:49:36 +0000 (22:49 +0000)]
Fix an error in the original coding of holdable cursors: PersistHoldablePortal
thought that it didn't have to reposition the underlying tuplestore if the
portal is atEnd. But this is not so, because tuplestores have separate read
and write cursors ... and the read cursor hasn't moved from the start.
This mistake explains bug #2970 from William Zhang.
Note: the coding here is pretty inefficient, but given that no one has noticed
this bug until now, I'd say hardly anyone uses the case where the cursor has
been advanced before being persisted. So maybe it's not worth worrying about.