Bruce Momjian [Mon, 4 Jul 2005 18:56:44 +0000 (18:56 +0000)]
I made the patch that improved the performance of replace_text().
The content of the patch is as follows:
(1)Create shortcut when subtext was not found.
(2)Stop using LEFT and RIGHT macro.
In LEFT and RIGHT macro, TEXTPOS is executed by the same content as
execution immediately before. The execution frequency of TEXTPOS can be
reduced by using text_substring instead of LEFT and RIGHT macro.
(3)Add appendStringInfoText, and use it instead of
appendStringInfoString.
There is an overhead of PG_TEXT_GET_STR when appendStringInfoString is
executed by text type. This can be reduced by appendStringInfoText.
(4)Reduce execution of TEXTDUP.
The effect of the patch that I measured is as follows:
- The Data for test was created by 'pgbench -i'.
- Test SQL:
select replace(aid, '9', 'A') from accounts;
- Test results: Linux(CPU: Pentium III, Compiler option: -O2)
original: 1.515s
patched: 1.250s
Tom Lane [Mon, 4 Jul 2005 04:51:52 +0000 (04:51 +0000)]
Arrange for the postmaster (and standalone backends, initdb, etc) to
chdir into PGDATA and subsequently use relative paths instead of absolute
paths to access all files under PGDATA. This seems to give a small
performance improvement, and it should make the system more robust
against naive DBAs doing things like moving a database directory that
has a live postmaster in it. Per recent discussion.
> A quick look shows that when you use --with-libraries=/foo/bar the
> generated link line for libraries says
>
> -L/foo/bar -lpq
>
> and it should probably be the other way around (as it is for the
> executables).
>
> So I suspect we need some makefile tuning.
Bruce Momjian [Mon, 4 Jul 2005 04:06:43 +0000 (04:06 +0000)]
Update wording:
< Currently, to protect against partial disk page writes, we write the
> Currently, to protect against partial disk page writes, we write 877c877
< * Turn off after-change writes if fsync is disabled
> * Turn off full page writes if fsync is disabled
Tom Lane [Sun, 3 Jul 2005 21:14:18 +0000 (21:14 +0000)]
Don't try to constant-fold functions returning RECORD. We were never
able to do this before, but I had tried to make an exception for functions
with OUT parameters. Michael Fuhr found one problem with it already, and
I found another, which was it didn't work for strict functions with a
NULL input. While both of these could be worked around, the probability
that there are more gotchas seems high; I think prudence dictates just
reverting to the former behavior for now. Accordingly, remove the kluge
added to get_expr_result_type() for Michael's case.
Support cross compilation by compiling "zic" with a native compiler. This
relies on the output of zic being platform independent, but that is
currently the case.
Bruce Momjian [Sun, 3 Jul 2005 02:32:56 +0000 (02:32 +0000)]
This patch allows contrib/pgcrypto to build with OpenSSL 0.9.8
(currently in beta) when cryptolib = openssl. According to the
following checkin message from several years ago, OpenSSL application
developers should no longer rely on <openssl/evp.h> to include
everything they need:
http://cvs.openssl.org/chngview?cn=9888
This patch adds the necessary header files. It doesn't appear to
break anything when building against OpenSSL 0.9.7.
BTW, core appears to build and work fine with OpenSSL 0.9.8. I've
built 7.3 through HEAD against 0.9.8-beta6 without noticing any
problems.
Bruce Momjian [Sat, 2 Jul 2005 23:28:22 +0000 (23:28 +0000)]
> A quick look shows that when you use --with-libraries=/foo/bar the
> generated link line for libraries says
>
> -L/foo/bar -lpq
>
> and it should probably be the other way around (as it is for the
> executables).
>
> So I suspect we need some makefile tuning.
Tom Lane [Sat, 2 Jul 2005 23:00:42 +0000 (23:00 +0000)]
Teach planner about some cases where a restriction clause can be
propagated inside an outer join. In particular, given
LEFT JOIN ON (A = B) WHERE A = constant, we cannot conclude that
B = constant at the top level (B might be null instead), but we
can nonetheless put a restriction B = constant into the quals for
B's relation, since no inner-side rows not meeting that condition
can contribute to the final result. Similarly, given
FULL JOIN USING (J) WHERE J = constant, we can't directly conclude
that either input J variable = constant, but it's OK to push such
quals into each input rel. Per recent gripe from Kim Bisgaard.
Along the way, remove 'valid_everywhere' flag from RestrictInfo,
as on closer analysis it was not being used for anything, and was
defined backwards anyway.
Tom Lane [Fri, 1 Jul 2005 19:19:05 +0000 (19:19 +0000)]
Migrate rtree_gist functionality into the core system, and add some
basic regression tests for GiST to the standard regression tests.
I took the opportunity to add an rtree-equivalent gist opclass for
circles; the contrib version only covered boxes and polygons, but
indexing circles is very handy for distance searches.
Bruce Momjian [Fri, 1 Jul 2005 17:35:22 +0000 (17:35 +0000)]
Clarify:
< o Replace crude DELETE FROM method of pg_dumpall for cleaning of
< roles with separate DROP commands
> o Replace crude DELETE FROM method of pg_dumpall --clean for
> cleaning of roles with separate DROP commands
Bruce Momjian [Fri, 1 Jul 2005 17:32:21 +0000 (17:32 +0000)]
Update for roles:
< * Allow limits on per-db/user connections
> * Allow limits on per-db/role connections
< * Prevent default re-use of sysids for dropped users and roles
<
< Currently, if a user is removed while he still owns objects, a new
< user given might be given their user id and inherit the
< previous users objects.
< 450c444
< * Add COMMENT ON for all cluster global objects (users, roles, databases
> * Add COMMENT ON for all cluster global objects (roles, databases 609c603
< users and roles with separate DROP commands
> roles with separate DROP commands
Bruce Momjian [Fri, 1 Jul 2005 17:05:58 +0000 (17:05 +0000)]
Update for roles:
< * Prevent default re-use of sysids for dropped users and groups
> * Prevent default re-use of sysids for dropped users and roles 450c450
< * Add COMMENT ON for all cluster global objects (users, groups, databases
> * Add COMMENT ON for all cluster global objects (users, roles, databases 609c609
< users and groups with separate DROP commands
> users and roles with separate DROP commands
Fixes from Janko Richter <jankorichter@yahoo.de>
- Fix wrong index results on text, char, varchar for multibyte strings
- Fix some SIGFPE signals
- Add support for infinite timestamps
- Because of locale settings, btree_gist can not be a prefix index anymore (for text).
Each node holds now just the lower and upper boundary.
Neil Conway [Fri, 1 Jul 2005 05:13:58 +0000 (05:13 +0000)]
Adds some missing error handling to PGTYPESnumeric_div() in ecpg's
pgtypeslib: (1) we need to check the return value of sub_abs() (2) we
need to check the return value of 4 calls to digitbuf_alloc().
Per Coverity static analysis performed by EnterpriseDB.
Neil Conway [Fri, 1 Jul 2005 05:12:06 +0000 (05:12 +0000)]
Fix some minor infelicities in ecpg's pgtypeslib: (1) `pstr' must be
non-NULL in this function, so there is no need to check for it (2) we
should check the return value of pgtypes_strdup(). Patch from Eric
Astor at EnterpriseDB, with slight cleanup by myself, per a report
from the Coverity tool.
Teodor Sigaev [Thu, 30 Jun 2005 17:52:14 +0000 (17:52 +0000)]
Bug fixes for GiST crash recovery.
- add forgotten check of lsn for insert completion
- remove level of pages: hard to check in recovery
- some cleanups
Neil Conway [Thu, 30 Jun 2005 07:27:31 +0000 (07:27 +0000)]
Minor ecpg tweak: the return value of calloc() is guaranteed to be NULL
or zero-filled; therefore zero-filling it via memset() is pointless.
(I think setting `errno' is probably a waste of cycles as well, but I
haven't changed that.)
Neil Conway [Thu, 30 Jun 2005 07:08:59 +0000 (07:08 +0000)]
Warning cleanups for ecpg tests. Avoid doing pointer arithmetic on void *,
remove old-style function declarations, and mark a function "static".
There are some remaining warnings, but this fixes most of them, anyway.
Neil Conway [Thu, 30 Jun 2005 01:59:20 +0000 (01:59 +0000)]
Fix a theoretical memory leak in pg_password_sendauth(). If the first
malloc() succeeded but the second failed, the buffer allocated by the
first malloc() would be leaked. Fix this by allocating both buffers
via a single malloc(), as suggested by Tom.
Per Coverity static analysis performed by EnterpriseDB.
Tom Lane [Thu, 30 Jun 2005 00:00:52 +0000 (00:00 +0000)]
Improve the checkpoint signaling mechanism so that the bgwriter can tell
the difference between checkpoints forced due to WAL segment consumption
and checkpoints forced for other reasons (such as CREATE DATABASE). Avoid
generating 'checkpoints are occurring too frequently' messages when the
checkpoint wasn't caused by WAL segment consumption. Per gripe from
Chris K-L.
Tom Lane [Wed, 29 Jun 2005 22:51:57 +0000 (22:51 +0000)]
Clean up the rather historically encumbered interface to now() and
current time: provide a GetCurrentTimestamp() function that returns
current time in the form of a TimestampTz, instead of separate time_t
and microseconds fields. This is what all the callers really want
anyway, and it eliminates low-level dependencies on AbsoluteTime,
which is a deprecated datatype that will have to disappear eventually.
Tom Lane [Wed, 29 Jun 2005 20:34:15 +0000 (20:34 +0000)]
More cleanup on roles patch. Allow admin option to be inherited through
role memberships; make superuser/createrole distinction do something
useful; fix some locking and CommandCounterIncrement issues; prevent
creation of loops in the membership graph.
Tom Lane [Tue, 28 Jun 2005 22:16:45 +0000 (22:16 +0000)]
Fix up problems in write_auth_file and parsing of the auth file.
In particular, make hba.c cope with zero-length tokens, which it
never did properly before. Also, enforce rolcanlogin.
Tom Lane [Tue, 28 Jun 2005 19:51:26 +0000 (19:51 +0000)]
Bring syntax of role-related commands into SQL compliance. To avoid
syntactic conflicts, both privilege and role GRANT/REVOKE commands have
to use the same production for scanning the list of tokens that might
eventually turn out to be privileges or role names. So, change the
existing GRANT/REVOKE code to expect a list of strings not pre-reduced
AclMode values. Fix a couple other minor issues while at it, such as
InitializeAcl function name conflicting with a Windows system function.
Tom Lane [Tue, 28 Jun 2005 05:09:14 +0000 (05:09 +0000)]
Replace pg_shadow and pg_group by new role-capable catalogs pg_authid
and pg_auth_members. There are still many loose ends to finish in this
patch (no documentation, no regression tests, no pg_dump support for
instance). But I'm going to commit it now anyway so that Alvaro can
make some progress on shared dependencies. The catalog changes should
be pretty much done.
Teodor Sigaev [Mon, 27 Jun 2005 12:45:23 +0000 (12:45 +0000)]
Concurrency for GiST
- full concurrency for insert/update/select/vacuum:
- select and vacuum never locks more than one page simultaneously
- select (gettuple) hasn't any lock across it's calls
- insert never locks more than two page simultaneously:
- during search of leaf to insert it locks only one page
simultaneously
- while walk upward to the root it locked only parent (may be
non-direct parent) and child. One of them X-lock, another may
be S- or X-lock
- 'vacuum full' locks index
- improve gistgetmulti
- simplify XLOG records
Fix bug in index_beginscan_internal: LockRelation may clean
rd_aminfo structure, so move GET_REL_PROCEDURE after LockRelation
Tom Lane [Mon, 27 Jun 2005 02:17:47 +0000 (02:17 +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.
Neil Conway [Mon, 27 Jun 2005 02:04:26 +0000 (02:04 +0000)]
Remove support for Kerberos V4. It seems no one is using this, it has
some security issues, and upstream has declared it "dead". Patch from
Magnus Hagander, minor editorialization from Neil Conway.
Tom Lane [Mon, 27 Jun 2005 00:48:07 +0000 (00:48 +0000)]
Adjust contrib/seg &< and &> operators so that r-tree indexing logic
works properly for 1-D comparisons. Fix some other errors such as
bogus commutator specifications.
Tom Lane [Sun, 26 Jun 2005 23:32:34 +0000 (23:32 +0000)]
Tweak dynahash.c to not allocate so many entries at once when dealing
with a table that has a small predicted size. Avoids wasting several
hundred K on the timezone hash table, which is likely to have only one
or a few entries, but the entries use up 10Kb apiece ...
Tom Lane [Sun, 26 Jun 2005 19:16:07 +0000 (19:16 +0000)]
Code review for escape-strings patch. Sync psql and plpgsql lexers
with main, avoid using a SQL-defined SQLSTATE for what is most definitely
not a SQL-compatible error condition, fix documentation omissions,
adhere to message style guidelines, don't use two GUC_REPORT variables
when one is sufficient. Nothing done about pg_dump issues.
Tom Lane [Sat, 25 Jun 2005 22:47:29 +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 16:53:49 +0000 (16:53 +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.
Bruce Momjian [Sat, 25 Jun 2005 01:32:02 +0000 (01:32 +0000)]
Add item:
> * Add NUMERIC division operator that doesn't round?
>
> Currently NUMERIC _rounds_ the result to the specified precision.
> This means division can return a result that multiplied by the
> divisor is greater than the dividend, e.g. this returns a value > 10:
>
> SELECT (10::numeric(2,0) / 6::numeric(2,0))::numeric(2,0) * 6;
>
> The positive modulus result returned by NUMERICs might be considered
> inaccurate, in one sense.
>
Tom Lane [Fri, 24 Jun 2005 20:53:34 +0000 (20:53 +0000)]
Extend r-tree operator classes to handle Y-direction tests equivalent
to the existing X-direction tests. An rtree class now includes 4 actual
2-D tests, 4 1-D X-direction tests, and 4 1-D Y-direction tests.
This involved adding four new Y-direction test operators for each of
box and polygon; I followed the PostGIS project's lead as to the names
of these operators.
NON BACKWARDS COMPATIBLE CHANGE: the poly_overleft (&<) and poly_overright
(&>) operators now have semantics comparable to box_overleft and box_overright.
This is necessary to make r-tree indexes work correctly on polygons.
Also, I changed circle_left and circle_right to agree with box_left and
box_right --- formerly they allowed the boundaries to touch. This isn't
actually essential given the lack of any r-tree opclass for circles, but
it seems best to sync all the definitions while we are at it.
Tom Lane [Fri, 24 Jun 2005 14:28:06 +0000 (14:28 +0000)]
Remove overspecification of precision of CURRENT_TIMESTAMP, LOCALTIMESTAMP,
CURRENT_TIME, and LOCALTIME: now they just produce "timestamptz" not
"timestamptz(6)", etc. This makes the behavior more consistent with our
choice to not assign a specific default precision to column datatypes.
It should also save a few cycles at runtime due to not having to invoke
the round-to-given-precision functions.
I also took the opportunity to translate CURRENT_TIMESTAMP into "now()"
instead of an invocation of the timestamptz input converter --- this should
save a few cycles too.
Tom Lane [Fri, 24 Jun 2005 00:18:52 +0000 (00:18 +0000)]
Fix rtree and contrib/rtree_gist search behavior for the 1-D box and
polygon operators (<<, &<, >>, &>). Per ideas originally put forward
by andrew@supernews and later rediscovered by moi. This patch just
fixes the existing opclasses, and does not add any new behavior as I
proposed earlier; that can be sorted out later. In principle this
could be back-patched, since it changes only search behavior and not
system catalog entries nor rtree index contents. I'm not currently
planning to do that, though, since I think it could use more testing.
Bruce Momjian [Thu, 23 Jun 2005 21:28:10 +0000 (21:28 +0000)]
Update text of terminate:
< * Allow administrators to safely terminate individual sessions
<
< Right now, SIGTERM will terminate a session, but it is treated as
< though the postmaster has paniced and shared memory might not be
< cleaned up properly. A new signal is needed for safe termination
< because backends must first do a query cancel, then exit once they
< have run the query cancel cleanup routine.
<
> * Allow administrators to safely terminate individual sessions either
> via an SQL function or SIGTERM
Tom Lane [Thu, 23 Jun 2005 02:33:28 +0000 (02:33 +0000)]
Move findoidjoins out of contrib and into src/tools, which is a more
logical place for it since it is of no use to users. Per recent
discussions on cleaning up contrib.
Tom Lane [Thu, 23 Jun 2005 00:06:37 +0000 (00:06 +0000)]
Cleanup the contrib/lo module: there is no need anymore to implement
a physically separate type. Defining 'lo' as a domain over OID works
just fine and is more efficient. Improve documentation and fix up the
test script. (Would like to turn test script into a proper regression
test, but right now its output is not constant because of numeric OIDs;
plus it makes Unix-specific assumptions about files it can import.)