Tom Lane [Sun, 18 Mar 2007 17:57:34 +0000 (17:57 +0000)]
Fix ecpg/preproc makefile for parallel builds: parser.o must depend
on preproc.h, else make may try to build it before preproc.h is ready.
Per failures seen here and in buildfarm.
Michael Meskes [Sat, 17 Mar 2007 19:25:24 +0000 (19:25 +0000)]
- Changed some whitespacing in connect statement.
- Made some chars const as proposed by Stefan Huehner <stefan@huehner.org>.
- Synced parser and keyword lists.
- Copied two token parsing from backend parser to ecpg parser.
- Also added a test case for this.
Magnus Hagander [Sat, 17 Mar 2007 13:50:42 +0000 (13:50 +0000)]
Turn most vc build scripts into modules instead of scripts, and just have
skeleton scripts calling them. To make it easier for the buildfarm
(or other "outside callers") to use these modules directly.
Tom Lane [Sat, 17 Mar 2007 03:15:38 +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 [Sat, 17 Mar 2007 01:15:55 +0000 (01:15 +0000)]
Ooops, got only one of the two ArrayExpr variants correct in first
cut at exprTypmod support. Also, experimentation shows that we need
to label the type of Const nodes that are numeric with a specific
typmod.
Tom Lane [Sat, 17 Mar 2007 00:11:05 +0000 (00:11 +0000)]
Fix up the remaining places where the expression node structure would lose
available information about the typmod of an expression; namely, Const,
ArrayRef, ArrayExpr, and EXPR and ARRAY SubLinks. In the ArrayExpr and
SubLink cases it wasn't really the data structure's fault, but exprTypmod()
being lazy. This seems like a good idea in view of the expected increase in
typmod usage from Teodor's work to allow user-defined types to have typmods.
In particular this responds to the concerns we had about eliminating the
special-purpose hack that exprTypmod() used to have for BPCHAR Consts.
We can now tell whether or not such a Const has been cast to a specific
length, and report or display properly if so.
Tom Lane [Fri, 16 Mar 2007 16:11:49 +0000 (16:11 +0000)]
Fix race condition in parallel regression tests. The new plancache test
was expecting there to be no regular table named 'foo', but it turns out
the rules test transiently creates one, so that plancache would sometimes
fail. I couldn't reproduce that in quite a few tries here, but several
buildfarm machines have shown the failure. Fix by renaming plancache's
temp table to something nonconflicting.
Tom Lane [Thu, 15 Mar 2007 23:12:07 +0000 (23:12 +0000)]
Make use of plancache module for SPI plans. In particular, since plpgsql
uses SPI plans, this finally fixes the ancient gotcha that you can't
drop and recreate a temp table used by a plpgsql function.
Along the way, clean up SPI's API a little bit by declaring SPI plan
pointers as "SPIPlanPtr" instead of "void *". This is cosmetic but
helps to forestall simple programming mistakes. (I have changed some
but not all of the callers to match; there are still some "void *"'s
in contrib and the PL's. This is intentional so that we can see if
anyone's compiler complains about it.)
Tom Lane [Wed, 14 Mar 2007 18:48:55 +0000 (18:48 +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 17:38:06 +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.
Teodor Sigaev [Wed, 14 Mar 2007 14:21:53 +0000 (14:21 +0000)]
Add GIN support for pg_trgm. From Guillaume Smet <guillaume.smet@gmail.com>
with minor editorization by me.
Hstore improvements
* add operation hstore ? text - excat equivalent of exist()
* remove undocumented behaviour of contains operation with NULL value
* now 'key'::text=>NULL returns '"key"=>NULL' instead of NULL
* Add GIN support for contains and exist operations
* Add GiST support for exist operatiion
* improve regression tests
Tom Lane [Tue, 13 Mar 2007 22:56:48 +0000 (22:56 +0000)]
Regression makefile now needs to make separate lists of what to clean
for input/ and output/ directories, because with the addition of
largeobject_1.source, they're not the same list. Apparently the current
buildfarm process does not exercise whether 'make distclean' leaves a
clean tree behind, else the farm would have been failing for awhile.
Magnus Hagander [Tue, 13 Mar 2007 16:03:36 +0000 (16:03 +0000)]
Rewrite win32 install documentation (it's not client only anymore, and it's
now complete). Update for the MSVC6/Borland support now being only libpq.
Move most of the information about full MSVC build from README file into
documentation.
Tom Lane [Tue, 13 Mar 2007 00:33:44 +0000 (00:33 +0000)]
First phase of plan-invalidation project: create a plan cache management
module and teach PREPARE and protocol-level prepared statements to use it.
In service of this, rearrange utility-statement processing so that parse
analysis does not assume table schemas can't change before execution for
utility statements (necessary because we don't attempt to re-acquire locks
for utility statements when reusing a stored plan). This requires some
refactoring of the ProcessUtility API, but it ends up cleaner anyway,
for instance we can get rid of the QueryContext global.
Still to do: fix up SPI and related code to use the plan cache; I'm tempted to
try to make SQL functions use it too. Also, there are at least some aspects
of system state that we want to ensure remain the same during a replan as in
the original processing; search_path certainly ought to behave that way for
instance, and perhaps there are others.
Alvaro Herrera [Sun, 11 Mar 2007 05:22:00 +0000 (05:22 +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.
Magnus Hagander [Thu, 8 Mar 2007 19:27:28 +0000 (19:27 +0000)]
Remove unsafe calling of WSAStartup and WSACleanup from DllMain. Move the
inline cleanup call around so it will be called in the right order, and
be called on errors.
Tom Lane [Thu, 8 Mar 2007 17:03:31 +0000 (17:03 +0000)]
Fix vac_update_relstats to ensure it always sends a relcache inval message,
even if none of the fields in the pg_class row change. This behavior is
necessary to ensure other backends flush rd_targblock values that might
point to truncated-away pages. We got this right pre-8.2 but it was broken
by overoptimistic change to not write out the pg_class row if unchanged.
Per report from Pavan Deolasee.
Teodor Sigaev [Wed, 7 Mar 2007 21:21:12 +0000 (21:21 +0000)]
Athough cube is a varlena type, nowhere was a detoasting of cube's value, so
fix it. Add macroses DatumGetNDBOX, PG_GETARG_NDBOX and PG_RETURN_NDBOX.
Backpatch for 8.2 too.
Previous versions use version 0 calling conventions. And fmgr code detoast
values for user-defined functions.
Tom Lane [Tue, 6 Mar 2007 22:45:16 +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.
Magnus Hagander [Mon, 5 Mar 2007 14:18:38 +0000 (14:18 +0000)]
Remove old-style win32 client-only visual c++ build infrastructure for everything except
libpq. We need to keep libpq to build static libraries and to use PQtrace
with clients using older versions of MSVC.
Neil Conway [Sat, 3 Mar 2007 19:32:55 +0000 (19:32 +0000)]
Add resetStringInfo(), which clears the content of a StringInfo, and
fixup various places in the tree that were clearing a StringInfo by hand.
Making this function a part of the API simplifies client code slightly,
and avoids needlessly peeking inside the StringInfo interface.
Tom Lane [Fri, 2 Mar 2007 23:37:23 +0000 (23:37 +0000)]
Make log_min_error_statement put LOG level at the same priority as
log_min_messages does; and arrange to suppress the duplicative output
that would otherwise result from log_statement and log_duration messages.
Bruce Momjian and Tom Lane.
Tom Lane [Fri, 2 Mar 2007 00:48:44 +0000 (00:48 +0000)]
Suppress useless searches for unused line pointers in PageAddItem. To do
this, add a 16-bit "flags" field to page headers by stealing some bits from
pd_tli. We use one flag bit as a hint to indicate whether there are any
unused line pointers; the remaining 15 are available for future use.
This is a cut-down form of an idea proposed by Hiroki Kataoka in July 2005.
At the time it was rejected because the original patch increased the size of
page headers and it wasn't clear that the benefit outweighed the distributed
cost. The flag-bit approach gets most of the benefit without requiring an
increase in the page header size.
Tom Lane [Thu, 1 Mar 2007 18:50:28 +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 [Wed, 28 Feb 2007 22:44:38 +0000 (22:44 +0000)]
Fix up several contrib modules that were using varlena datatypes in not-so-obvious
ways. I'm not totally sure that I caught everything, but at least now they pass
their regression tests with VARSIZE/SET_VARSIZE defined to reverse byte order.
Bruce Momjian [Wed, 28 Feb 2007 17:28:09 +0000 (17:28 +0000)]
Add language about rights given by posting a patch:
<li>PostgreSQL is licensed under a BSD license. By posting a patch
to the public PostgreSQL mailling lists, you are giving the PostgreSQL
Global Development Group the non-revokable right to distribute your
patch under the BSD license. If you use code that is available under
some other license that is BSD compatible (eg. public domain), please
note that in your email submission.</li>
Tom Lane [Tue, 27 Feb 2007 23:48:10 +0000 (23:48 +0000)]
Replace direct assignments to VARATT_SIZEP(x) with SET_VARSIZE(x, len).
Get rid of VARATT_SIZE and VARATT_DATA, which were simply redundant with
VARSIZE and VARDATA, and as a consequence almost no code was using the
longer names. Rename the length fields of struct varlena and various
derived structures to catch anyplace that was accessing them directly;
and clean up various places so caught. In itself this patch doesn't
change any behavior at all, but it is necessary infrastructure if we hope
to play any games with the representation of varlena headers.
Greg Stark and Tom Lane
Tom Lane [Tue, 27 Feb 2007 01:11:26 +0000 (01:11 +0000)]
Get rid of the separate EState for subplans, and just let them share the
parent query's EState. Now that there's a single flat rangetable for both
the main plan and subplans, there's no need anymore for a separate EState,
and removing it allows cleaning up some crufty code in nodeSubplan.c and
nodeSubqueryscan.c. Should be a tad faster too, although any difference
will probably be hard to measure. This is the last bit of subsidiary
mop-up work from changing to a flat rangetable.
Tom Lane [Sun, 25 Feb 2007 17:44:01 +0000 (17:44 +0000)]
Put back copyObject() call I removed in a fit of brain fade. This one
is still needed despite cleanups in setrefs.c, because the point is to
let the inserted Result node compute a different tlist than its input
node does. Per example from Jeremy Drake.
Tom Lane [Fri, 23 Feb 2007 21:59:45 +0000 (21:59 +0000)]
Now that plans have flat rangetable lists, it's a lot easier to get EXPLAIN to
drill down into subplan targetlists to print the referent expression for an
OUTER or INNER var in an upper plan node. Hence, make it do that always, and
banish the old hack of showing "?columnN?" when things got too complicated.
Along the way, fix an EXPLAIN bug I introduced by suppressing subqueries from
execution-time range tables: get_name_for_var_field() assumed it could look at
rte->subquery to find out the real type of a RECORD var. That doesn't work
anymore, but instead we can look at the input plan of the SubqueryScan plan
node.
Tom Lane [Thu, 22 Feb 2007 23:44:25 +0000 (23:44 +0000)]
Change Agg and Group nodes so that Vars contained in their targetlists
and quals have varno OUTER, rather than zero, to indicate a reference to
an output of their lefttree subplan. This is consistent with the way
that every other upper-level node type does it, and allows some simplifications
in setrefs.c and EXPLAIN.
Tom Lane [Thu, 22 Feb 2007 22:49:27 +0000 (22:49 +0000)]
Fix bug I introduced in recent patch to make hash joins discard null tuples
immediately: ExecHashGetHashValue failed to restore the caller's memory
context before taking the failure exit.
Tom Lane [Thu, 22 Feb 2007 22:00:26 +0000 (22:00 +0000)]
Turn the rangetable used by the executor into a flat list, and avoid storing
useless substructure for its RangeTblEntry nodes. (I chose to keep using the
same struct node type and just zero out the link fields for unneeded info,
rather than making a separate ExecRangeTblEntry type --- it seemed too
fragile to have two different rangetable representations.)
Along the way, put subplans into a list in the toplevel PlannedStmt node,
and have SubPlan nodes refer to them by list index instead of direct pointers.
Vadim wanted to do that years ago, but I never understood what he was on about
until now. It makes things a *whole* lot more robust, because we can stop
worrying about duplicate processing of subplans during expression tree
traversals. That's been a constant source of bugs, and it's finally gone.
There are some consequent simplifications yet to be made, like not using
a separate EState for subplans in the executor, but I'll tackle that later.
Bruce Momjian [Wed, 21 Feb 2007 22:15:21 +0000 (22:15 +0000)]
Move increase FSM warning to after lazy_truncate_heap() because the
function might reduce the number of free pages in the table. Recommend
VACUUM FULL only if 20% free.
Bruce Momjian [Wed, 21 Feb 2007 20:02:17 +0000 (20:02 +0000)]
btree source code cleanups:
I refactored findsplitloc and checksplitloc so that the division of
labor is more clear IMO. I pushed all the space calculation inside the
loop to checksplitloc.
I also fixed the off by 4 in free space calculation caused by
PageGetFreeSpace subtracting sizeof(ItemIdData), even though it was
harmless, because it was distracting and I felt it might come back to
bite us in the future if we change the page layout or alignments.
There's now a new function PageGetExactFreeSpace that doesn't do the
subtraction.
findsplitloc now tries the "just the new item to right page" split as
well. If people don't like the refactoring, I can write a patch to just
add that.