Tom Lane [Mon, 19 Aug 2002 00:11:53 +0000 (00:11 +0000)]
Add missing copyfuncs/equalfuncs entries, including T_Null which has
been missing forever; surprising it wasn't noticed before. The other
additions are, um, sloppiness in certain recent feature additions.
Tom Lane [Sun, 18 Aug 2002 18:46:15 +0000 (18:46 +0000)]
Remove optimization whereby parser would make only one sort-list entry
when two equal() targetlist items were to be added to an ORDER BY or
DISTINCT list. Although indeed this would make sorting fractionally
faster by sometimes saving a comparison, it confuses the heck out of
later stages of processing, because it makes it look like the user
wrote DISTINCT ON rather than DISTINCT. Bug reported by joe@piscitella.com.
Bruce Momjian [Sun, 18 Aug 2002 01:39:43 +0000 (01:39 +0000)]
Back out tcl patch, per Tom Lane:
Everytime if I do PQconsumeInput (when the backend channel gets
readable) I check for the return value. (0 == error) and generate a
notification manually, e.g. fixed string connection_closed) and pass it to the
Bruce Momjian [Sat, 17 Aug 2002 13:06:50 +0000 (13:06 +0000)]
In that case, attached is a patch which locates the beginning of the
offending token more efficiently (per your suggestion of using
scanbuf). The new patch does the same as before:
template1=# select * frum pg_class;
ERROR: parser: parse error at or near "frum" at character 10
It also implement's Tom's suggestion:
template1=# select * from pg_class where\g
ERROR: parse: parse error at end of input
Bruce Momjian [Sat, 17 Aug 2002 13:04:19 +0000 (13:04 +0000)]
[ Newest version of patch applied.]
This patch is an updated version of the lock listing patch. I've made
the following changes:
- write documentation
- wrap the SRF in a view called 'pg_locks': all user-level
access should be done through this view
- re-diff against latest CVS
One thing I chose not to do is adapt the SRF to use the anonymous
composite type code from Joe Conway. I'll probably do that eventually,
but I'm not really convinced it's a significantly cleaner way to
bootstrap SRF builtins than the method this patch uses (of course, it
has other uses...)
Bruce Momjian [Sat, 17 Aug 2002 12:19:31 +0000 (12:19 +0000)]
What I have done for libpgtcl:
Everytime if I do PQconsumeInput (when the backend channel gets
readable) I check for the return value. (0 == error) and generate a
notification manually, e.g. fixed string connection_closed) and pass it to the
TCL event queue. The only other thing I had to do is to comment out removing
all pending events in PgStopNotifyEventSource whenever the connection was
unexpectedly closed (so the manually generated event will not be deleted).
A broken backend connection triggers a notify event to the client (fixed
notification string "connection_closed") so proper action can be taken to switch
to another database server etc. Remember that this is event driven. If you have
applications, that have idle database connections most of the time, you'll get
immediate feedback of a dying server. Upon connection to the server issue a
pg_notify for notify event "connection_closed" and whenever the backend crashes
(which it does do in very very rare cases) you get an event driven recovery. (of
course the Tcl-Event loop has to be processed). Issuing a notification
"connection_closed" on a still working database could be used for switching to
another db-server (which I've actually impelemented right now).
Tom Lane [Fri, 16 Aug 2002 23:01:21 +0000 (23:01 +0000)]
Modify pg_dump to dump foreign-key constraints as constraints, not as
sets of triggers. Also modify psql \d command to show foreign key
constraints as such and hide the triggers. pg_get_constraintdef()
function added to backend to support these. From Rod Taylor, code
review and some editorialization by Tom Lane.
Bruce Momjian [Fri, 16 Aug 2002 04:47:43 +0000 (04:47 +0000)]
This patch improves the "Database Users and Permissions" section of
the Administrator's Guide. This section needs a lot more work, but
this is a start anyway...
Bruce Momjian [Fri, 16 Aug 2002 04:45:44 +0000 (04:45 +0000)]
Added:
<
> * Prevent mismatch of frontend/backend encodings from converting bytea
> data from being interpreted as encoded strings 512a514,515
> * Fix glibc's mktime() to handle pre-1970's dates
>
Bruce Momjian [Thu, 15 Aug 2002 17:46:46 +0000 (17:46 +0000)]
Done:
> * -Improve control over user privileges, including table creation
> * -Add PGPASSWORDFILE environment variable or ~/.pgpass to store
> o -Compile under jdk 1.4
Bruce Momjian [Thu, 15 Aug 2002 16:36:08 +0000 (16:36 +0000)]
Tom Lane wrote:
> There's no longer a separate call to heap_storage_create in that routine
> --- the right place to make the test is now in the storage_create
> boolean parameter being passed to heap_create. A simple change, but
> it passeth patch's understanding ...
Thanks.
Attached is a patch against cvs tip as of 8:30 PM PST or so. Turned out
that even after fixing the failed hunks, there was a new spot in
bufmgr.c which needed to be fixed (related to temp relations;
RelationUpdateNumberOfBlocks). But thankfully the regression test code
caught it :-)
Bruce Momjian [Thu, 15 Aug 2002 03:32:36 +0000 (03:32 +0000)]
Well, that certainly appeared to be very straight forward. pg.py and
syscat.py scripts were both modified. pg.py uses it to cache a list of
pks (which is seemingly does for every db connection) and various
attributes. syscat uses it to walk the list of system tables and
queries the various attributes from these tables.
In both cases, it seemingly makes sense to apply what you've requested.
Bruce Momjian [Thu, 15 Aug 2002 03:31:45 +0000 (03:31 +0000)]
http://archives.postgresql.org/pgsql-bugs/2002-06/msg00086.php and never
saw a fix offered up. Since I'm gearing up to use Postgres and Python
soon, I figured I'd have a hand at trying to get this sucker addressed.
Apologies if this has already been plugged. I looked in the archives
and never saw a response.
At any rate, I must admit I don't think I fully understand the
implications of some of the changes I made even though they appear to be
straight forward. We all know the devil is in the details. Anyone more
knowledgeable is requested to review my changes. :(
I also updated the advanced.py script in a somewhat nonsensical fashion
to make use of an int8 field in an effort to test this change. It seems
to run okay, however, this is by no means an all exhaustive test. So,
it's possible that a bumpy road may lay ahead for some. On the other
hand...overflows (hopefully) previously lurked (long -> int conversion).
Bruce Momjian [Thu, 15 Aug 2002 03:00:59 +0000 (03:00 +0000)]
PostgreSQL 7.2.1 Documentation
Chapter 1. libpq - C Library
1.3. Command Execution Functions
1.3.3. Escaping binary strings for inclusion in SQL queries
I found the line
"The result string length does not include the terminating zero byte of the result."
is not right.
The result string length does indeed include the terminating zero byte.
Bruce Momjian [Thu, 15 Aug 2002 02:59:18 +0000 (02:59 +0000)]
> Neil Conway <nconway@klamath.dyndns.org> writes:
> > This patch improves the documentation of the UPDATE and ALTER TABLE
> > commands to elaborate on the effect of specifying an "ONLY" clause.
>
> Unfortunately this is still only half the truth ... see the
> SQL_INHERITANCE configuration variable.
Okay, I've attached an updated patch with more information on
SQL_INHERITANCE and inheritance behavior in prior releases.
Bruce Momjian [Thu, 15 Aug 2002 02:58:29 +0000 (02:58 +0000)]
The attached patch changes most of the usages of sprintf() to
snprintf() in contrib/. I didn't touch the places where pointer
arithmatic was being used, or other areas where the fix wasn't
trivial. I would think that few, if any, of the usages of sprintf()
were actually exploitable, but it's probably better to be paranoid...
Bruce Momjian [Thu, 15 Aug 2002 02:56:19 +0000 (02:56 +0000)]
I'm giving a try at some TODO items. Currently it's the turn of the
PGPASSWORDFILE environment variable. I have modified libpq to make use
of this variable. I present the first cut here.
Bruce Momjian [Thu, 15 Aug 2002 02:53:35 +0000 (02:53 +0000)]
Appears there is a problem finding the opclass when indexing a domain.
CREATE DOMAIN newint as int4;
CREATE TABLE tab (col newint unique);
ERROR: data type newint has no default operator class for access method
"btree"
You must specify an operator class for the index or define a
default operator class for the data type
Specifically, GetDefaultOpClass() finds 0 exact matches and 3 binary
compatible matches. Fetching getBaseType() of the attribute fixes the
problem for domains (see attachment).
Bruce Momjian [Thu, 15 Aug 2002 02:51:27 +0000 (02:51 +0000)]
As discussed on several occasions previously, the new anonymous
composite type capability makes it possible to create a system view
based on a table function in a way that is hopefully palatable to
everyone. The attached patch takes advantage of this, moving
show_all_settings() from contrib/tablefunc into the backend (renamed
all_settings(). It is defined as a builtin returning type RECORD. During
initdb a system view is created to expose the same information presently
available through SHOW ALL. For example:
test=# select * from pg_settings where name like '%debug%';
name | setting
-----------------------+---------
debug_assertions | on
debug_pretty_print | off
debug_print_parse | off
debug_print_plan | off
debug_print_query | off
debug_print_rewritten | off
wal_debug | 0
(7 rows)
Additionally during initdb two rules are created which make it possible
to change settings by updating the system view -- a "virtual table" as
Tom put it. Here's an example:
Barry Lind [Wed, 14 Aug 2002 20:35:40 +0000 (20:35 +0000)]
Added support for JDBC3. The driver will now build under JDBC3 (i.e. Java 1.4).
This concludes my changes that restructured the code to support JDBC3.
The jdbc unit tests were also resturctured to allow different tests between
jdbc2 and jdbc3, although currently make check (aka ant test) for JDBC3 just
runs the JDBC2 tests. Of special note the largeobject/PGblob and PGclob
classes have been moved under the jdbc2/jdbc3 specific directories as they
now differ by jdbc version. Also note that this checkin removes the
PostgresqlDataSource and files in the xa directory. A recent checkin has
added new datasource support that replaces the functionality provided by these
classes.
(note the added quotes to provide a little protection against spaces
and such). Then it's perfectly obvious what the calculation is doing.
I don't care about wasting 20-some bytes, but confusing readers of the
code is worth avoiding.
Bruce Momjian [Wed, 14 Aug 2002 03:40:41 +0000 (03:40 +0000)]
Updated:
< * Allow temporary views
< * Require view using temporary tables to be temporary views
> * -Have views on temporary tables exist in the temporary namespace
Bruce Momjian [Wed, 14 Aug 2002 02:49:42 +0000 (02:49 +0000)]
Added:
<
> o Improve PL/PgSQL exception handling
> o Allow PL/PgSQL parameters to be specified by name and type during
> definition
> o Allow PL/PgSQL function parameters to be passed by name,
> get_employee_salary(emp_id => 12345, tax_year => 2001)
> o Add PL/PgSQL packages
> o Allow array declarations and other data types in PL/PgSQl DECLARE
> o Add PL/PgSQL PROCEDURES that can return multiple values
Tom Lane [Tue, 13 Aug 2002 20:14:24 +0000 (20:14 +0000)]
Fix tuple-chain-moving tests to handle marked-for-update tuples correctly
(they are not part of a chain). When failing to find a parent tuple in
an update chain, emit a warning and abandon repair_frag, but do not give
an error as before. This should eliminate the infamous 'No one parent tuple
was found' failure, which we now realize is not a can't-happen condition
but a perfectly valid database state. Per recent pghackers discussion.
Tom Lane [Tue, 13 Aug 2002 20:11:03 +0000 (20:11 +0000)]
Make sure that t_ctid is reset to equal t_self in heap_delete and
heap_mark4update; this avoids situations where a deleted tuple might
look like it is chained to something else. Also, cause all the WAL
redo routines to set t_ctid to equal t_self, rather than leaving it
undefined as before. Make heap_xlog_clean set the page's LSN and SUI
correctly. All per past discussions in pghackers, ranging back to
last December.
Tom Lane [Mon, 12 Aug 2002 14:25:07 +0000 (14:25 +0000)]
Restructure rowtype-parameter handling to eliminate need for possibly-
overflowable buffer for 'name%rowtype'; not to mention avoid problems
with mixed-case type names and other special cases.
Tom Lane [Mon, 12 Aug 2002 00:36:12 +0000 (00:36 +0000)]
Change memory-space accounting mechanism in tuplesort.c and tuplestore.c
to make a reasonable attempt at accounting for palloc overhead, not just
the requested size of each memory chunk. Since in many scenarios this
will make for a significant reduction in the amount of space acquired,
partially compensate by doubling the default value of SORT_MEM to 1Mb.
Per discussion in pgsql-general around 9-Jun-2002..