Bruce Momjian [Tue, 6 Dec 2005 18:43:40 +0000 (18:43 +0000)]
Since my name has a non-ascii-letter in it, it's often spelled wrong. In
the latest release notes there is a latin1 character that shouldn't be
there so I made a patch to fix that. This patch also fixes some old
entries that uses o instead of ö (which is also wrong but not as
bad as including a latin1 character in the sgml file).
Tom Lane [Tue, 6 Dec 2005 16:50:46 +0000 (16:50 +0000)]
In a nestloop inner indexscan, it's OK to use pushed-down baserestrictinfo
clauses even if it's an outer join. This is a corner case since such
clauses could only arise from weird OUTER JOIN ON conditions, but worth
fixing. Per example from Ron at cheapcomplexdevices.com.
Tom Lane [Tue, 6 Dec 2005 02:29:27 +0000 (02:29 +0000)]
Make Win32 build use our port/snprintf.c routines, instead of depending
on libintl which may or may not provide what we need. Make a few marginal
cleanups to ensure this works. Andrew Dunstan and Tom Lane.
Tom Lane [Mon, 5 Dec 2005 02:39:43 +0000 (02:39 +0000)]
Fix a rather sizable number of problems in our homegrown snprintf, such as
incorrect implementation of argument reordering, arbitrary limit of output
size for sprintf and fprintf, willingness to access more bytes than "%.Ns"
specification allows, wrong formatting of LONGLONG_MIN, various field-padding
bugs and omissions. I believe it now accurately implements a subset of
the Single Unix Spec requirements (remaining unimplemented features are
documented, too). Bruce Momjian and Tom Lane.
Tatsuo Ishii [Sun, 4 Dec 2005 01:22:42 +0000 (01:22 +0000)]
Two fixes from Tom Lan. See the posting "[PATCHES] A couple of
proposed pgbench changes" on 2005/11/29 for more details.
The change at line 490 updates doCustom's local variable "commands"
after selecting a new file (command sequence). I think that the
existing coding will cause the thing to use the first command of the
old sequence in the remainder of the routine, which would be a bug.
I have not tried to set up a test case to prove it, though.
The other two changes cause doCustom to loop after processing a
meta-command. This might be a bit controversial, but as the code
is currently written, each meta-command "costs" one cycle of the
outer select() loop. Thus, for example, with the default TPC-B script,
once a backend returns "COMMIT" it will not receive a new command
until four cycles of issuing commands to other backends have elapsed.
(You can see this very easily by strace'ing pgbench under load.)
Bruce Momjian [Sat, 3 Dec 2005 16:45:23 +0000 (16:45 +0000)]
Allow to_char(interval) and to_char(time) to use AM/PM specifications.
Map them to a single day, so '30 hours' is 'AM'.
Have to_char(interval) and to_char(time) use "HH", "HH12" as 12-hour
intervals, rather than bypass and print the full interval hours. This
is neeeded because to_char(time) is mapped to interval in this function.
Intervals should use "HH24", and document suggestion.
Michael Meskes [Fri, 2 Dec 2005 15:04:48 +0000 (15:04 +0000)]
Added special handling of CONNECTION variable that is used by ECPG instead of given to the backend.
I failed to notice that CONNECTION had become a keyword in 8.1.
Tom Lane [Fri, 2 Dec 2005 01:30:26 +0000 (01:30 +0000)]
Rearrange code in ExecInitBitmapHeapScan so that we don't initialize the
child plan nodes until we have acquired lock on the relation to scan.
The relative order of initialization of plan nodes isn't real important in
other cases, but it's critical here because one is supposed to lock a
relation before its indexes, not vice versa. The original coding was at
least vulnerable to deadlock against DROP INDEX, and perhaps worse things.
Tom Lane [Thu, 1 Dec 2005 20:24:31 +0000 (20:24 +0000)]
Retry in FileRead and FileWrite if Windows returns ERROR_NO_SYSTEM_RESOURCES.
Also add a retry for Unixen returning EINTR, which hasn't been reported
as an issue but at least theoretically could be. Patch by Qingqing Zhou,
some minor adjustments by me.
Tom Lane [Wed, 30 Nov 2005 17:10:25 +0000 (17:10 +0000)]
Tweak choose_bitmap_and() heuristics in the light of example provided in bug
#2075: consider an index redundant if any of its index conditions were already
used, rather than if all of them were. Also, make the selectivity comparison
a bit fuzzy, so that very small differences in estimated selectivities don't
skew the results.
Michael Meskes [Wed, 30 Nov 2005 12:50:37 +0000 (12:50 +0000)]
- Made several variables "const char *" instead of "char *" as proposed by Qingqing Zhou <zhouqq@cs.toronto.edu>.
- Replaced all strdup() calls by ECPGstrdup().
Tom Lane [Mon, 28 Nov 2005 23:46:25 +0000 (23:46 +0000)]
Tweak hash join code to use an additional heuristic for deciding whether
it's worth probing the outer relation for emptiness before building the
hash table. To wit, if we're rescanning a join previously performed,
remember whether we found it nonempty the previous time, and don't bother
with the probe if it was nonempty. This buys back the performance lost
in examples like Mario Weilguni's.
Tom Lane [Mon, 28 Nov 2005 17:14:47 +0000 (17:14 +0000)]
Recent changes to allow hash join to exit early given empty input from
one child or the other had a problem: they did not leave the node in a
state that ExecReScanHashJoin would understand. In particular it would
tend to fail to reset the child plans when needed. Per report from
Mario Weilguni.
Tom Lane [Wed, 23 Nov 2005 20:28:05 +0000 (20:28 +0000)]
Get rid of ExecAssignResultTypeFromOuterPlan() and make all plan node types
generate their output tuple descriptors from their target lists (ie, using
ExecAssignResultTypeFromTL()). We long ago fixed things so that all node
types have minimally valid tlists, so there's no longer any good reason to
have two different ways of doing it. This change is needed to fix bug
reported by Hayden James: the fix of 2005-11-03 to emit the correct column
names after optimizing away a SubqueryScan node didn't work if the new
top-level plan node used ExecAssignResultTypeFromOuterPlan to generate its
tupdesc, since the next plan node down won't have the correct column labels.
Tom Lane [Wed, 23 Nov 2005 17:21:22 +0000 (17:21 +0000)]
Fix problems with rewriter failing to set Query.hasSubLinks when inserting
a SubLink expression into a rule query. Pre-8.1 we essentially did this
unconditionally; 8.1 tries to do it only when needed, but was missing a
couple of cases. Per report from Kyle Bateman. Add some regression test
cases covering this area.
Bruce Momjian [Tue, 22 Nov 2005 18:23:31 +0000 (18:23 +0000)]
Re-run pgindent, fixing a problem where comment lines after a blank
comment line where output as too long, and update typedefs for /lib
directory. Also fix case where identifiers were used as variable names
in the backend, but as typedefs in ecpg (favor the backend for
indenting).
Tom Lane [Sun, 20 Nov 2005 18:38:42 +0000 (18:38 +0000)]
Modify tuptoaster's API so that it does not try to modify the passed
tuple in-place, but instead passes back an all-new tuple structure if
any changes are needed. This is a much cleaner and more robust solution
for the bug discovered by Alexey Beschiokov; accordingly, revert the
quick hack I installed yesterday.
With this change, HeapTupleData.t_datamcxt is no longer needed; will
remove it in a separate commit in HEAD only.
Tom Lane [Sat, 19 Nov 2005 20:58:42 +0000 (20:58 +0000)]
Stopgap solution for problem reported by Alexey Beschiokov: after
doing heap_insert or heap_update, wipe out any extracted fields in
the TupleTableSlot containing the tuple, because they might not be valid
anymore if tuptoaster.c changed the tuple. Safe because slot must be
in the materialized state, but mighty ugly --- find a better answer!
Tom Lane [Fri, 18 Nov 2005 23:08:13 +0000 (23:08 +0000)]
Fix performance issue in exprTypmod(): for a COALESCE expression, it
recursed twice on its first argument, leading to exponential time spent
on a deep nest of COALESCEs ... such as a deeply nested FULL JOIN would
produce. Per report from Matt Carter.
Tom Lane [Thu, 17 Nov 2005 17:42:24 +0000 (17:42 +0000)]
DropRelFileNodeBuffers failed to fix the state of the lookup hash table
that was added to localbuf.c in 8.1; therefore, applying it to a temp table
left corrupt lookup state in memory. The only case where this had a
significant chance of causing problems was an ON COMMIT DELETE ROWS temp
table; the other possible paths left bogus state that was unlikely to
be used again. Per report from Csaba Nagy.
Tom Lane [Wed, 16 Nov 2005 17:08:12 +0000 (17:08 +0000)]
make_restrictinfo() failed to attach the specified required_relids to
its result when the clause was an OR clause. Brain fade exposed by
example from Sebastian BÎck.
Bruce Momjian [Tue, 15 Nov 2005 14:45:14 +0000 (14:45 +0000)]
Prevent certain symbols that are used for both typedefs and variable
names from being added to pgindent's typedef list. The existance of
them caused weird formatting in the date/type files, and in keywords.c.
Bruce Momjian [Tue, 15 Nov 2005 01:02:13 +0000 (01:02 +0000)]
Fix recent problems with BSD indent, including indenting past 80
columns, shifting comment to the right when more than 150 'else if'
clauses were used, and update typedefs for 8.1.X.
Tom Lane [Mon, 14 Nov 2005 23:54:36 +0000 (23:54 +0000)]
Restore the former RestrictInfo field valid_everywhere (but invert the flag
sense and rename to "outerjoin_delayed" to more clearly reflect what it
means). I had decided that it was redundant in 8.1, but the folly of this
is exposed by a bug report from Sebastian Böck. The place where it's
needed is to prevent orindxpath.c from cherry-picking arms of an outer-join
OR clause to form a relation restriction that isn't actually legal to push
down to the relation scan level. There may be some legal cases that this
forbids optimizing, but we'd need much closer analysis to determine it.
Tom Lane [Mon, 14 Nov 2005 17:43:13 +0000 (17:43 +0000)]
Prevent ExecInsert() and ExecUpdate() from scribbling on the result tuple
slot of the topmost plan node when a trigger returns a modified tuple.
These appear to be the only places where a plan node's caller did not
treat the result slot as read-only, which is an assumption that nodeUnique
makes as of 8.1. Fixes trigger-vs-DISTINCT bug reported by Frank van Vugt.
Tom Lane [Sun, 13 Nov 2005 19:11:45 +0000 (19:11 +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 00:31:40 +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 03:04:53 +0000 (03:04 +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 18:35:40 +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.
Bruce Momjian [Fri, 4 Nov 2005 18:16:50 +0000 (18:16 +0000)]
We haven't seen any agreement emerge as to what is causing AIX 5.3 ML3
to fail to successfully build the release candidates.
However, a patch has emerged (thanks, Seneca!) that does allow it to
work, and which I'd expect to be portable (better still!).
We are still actively pursuing why it breaks, but supposing that still
remains outstanding, at least the following would allow AIX users to
better survive a build...
Tom Lane [Fri, 4 Nov 2005 17:25:15 +0000 (17:25 +0000)]
Disregard superuserness when checking to see if a role GRANT would
create circularity of role memberships. This is a minimum-impact fix
for the problem reported by Florian Pflug. I thought about removing
the superuser_arg test from is_member_of_role() altogether, as it seems
redundant for many of the callers --- but not all, and it's way too late
in the 8.1 cycle to be making large changes. Perhaps reconsider this
later.
Tom Lane [Fri, 4 Nov 2005 02:56:31 +0000 (02:56 +0000)]
Avoid referring to a specific version of the SQL standard except where
necessary, and be careful to refer to the right version where it is
useful to do so. This partially reverts an ill-considered search and
replace from a few months ago.
Tom Lane [Thu, 3 Nov 2005 17:45:29 +0000 (17:45 +0000)]
Thinking further, it seems we had better also copy down resorigtbl/resorigcol
to ensure that SubqueryScan elimination doesn't change the behavior of
reporting of original column sources.
Tom Lane [Thu, 3 Nov 2005 17:34:03 +0000 (17:34 +0000)]
Fix the recently-added code that eliminates unnecessary SubqueryScan nodes
from a finished plan tree. We have to copy the output column names
(resname fields) from the SubqueryScan down to its child plan node;
else, if this is the topmost level of the plan, the wrong column names
will be delivered to the client. Per bug #2017 reported by Jolly Chen.
Alvaro Herrera [Thu, 3 Nov 2005 17:11:40 +0000 (17:11 +0000)]
Rename the members of CommandDest enum so they don't collide with other uses of
those names. (Debug and None were pretty bad names anyway.) I hope I catched
all uses of the names in comments too.
Tom Lane [Thu, 3 Nov 2005 00:23:36 +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 15:09:11 +0000 (15:09 +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.