OSDN Git Service
Hiroshi Inoue [Fri, 17 Aug 2001 23:50:00 +0000 (23:50 +0000)]
Ensure to hold an exclusive lock while reindexing a relation.
This is mainly to help developers to understand the code.
Bruce Momjian [Fri, 17 Aug 2001 15:44:17 +0000 (15:44 +0000)]
A little more code reorg for MD5/crypt.
Bruce Momjian [Fri, 17 Aug 2001 15:40:07 +0000 (15:40 +0000)]
Reorder MD5/crypt so MD5 comes first in the code.
Bruce Momjian [Fri, 17 Aug 2001 15:11:15 +0000 (15:11 +0000)]
Remove some unneeded dashes from libpq comments.
Bruce Momjian [Fri, 17 Aug 2001 15:02:18 +0000 (15:02 +0000)]
Add \n to libpq print output where needed.
Bruce Momjian [Fri, 17 Aug 2001 14:46:22 +0000 (14:46 +0000)]
This patch updates some comments in the DatabaseMetaData classes to
reflect a mail thread that discussed our conformance (or lack thereof)
to the SQL92 spec.
Barry Lind
Bruce Momjian [Fri, 17 Aug 2001 14:45:49 +0000 (14:45 +0000)]
Thanks for your feedback (and patience). Enclosed is my third
attempt at a patch to 7.1.2 to support Array.
[I think I've solved the mangled patch problem. Hotmail seems to
try to format the text file, so gzipping it should solve this
problem.]
In this patch I've incorporated Barry's feedback. Specifically:
1) OIDs are no longer hard-coded into Array.java. In order to
support this change I added a getOID(String) method to Field.java
which receives a PostgreSQL field type and returns a value from
java.sql.Types. I couldn't get away from using OIDs altogether
because the JDBC spec for Array specifies that some methods return
a ResultSet. This requires I construct Field objects,
which means I need OIDs. At least this approach doesn't hard
code these values. A Hashtable cache has been added to Field
so that an SQL lookup isn't necessary (following the model already
in Field.java).
2) Rewired the base formatting code in ResultSet.java to use 'to'
methods, which are then exposed as static methods in ResultSet.
These methods are used in Array to format the data without
duplications in the code.
3) Artifact call to first() in ResultSet.getArray() removed.
Greg Zoller
Bruce Momjian [Fri, 17 Aug 2001 13:59:29 +0000 (13:59 +0000)]
Attached is the patch requested by Tom Lane (see below). It
includes two changes in the JDBC driver:
1) When connected to a backend >= 7.2: use obj_description() and
col_description() instead of direct access to pg_description.
2) In DatabaseMetaData.getTables()/getColumns()/getProcedures():
when there is no comment on the object, return null in the
REMARKS column of the ResultSet, instead of the default string
"no remarks".
Change 2 first appeared as a side-effect of change 1, but it is
actually more compliant with the JDBC spec: "String object
containing an explanatory comment on the table/column/procedure,
which may be null". The default string "no remarks" was strictly
speaking incorrect, as it could not be distinguished from a real
user comment "no remarks". So I removed the default string
completely.
Change 2 might break existing code that doesn't follow the JDBC
spec and isn't prepared to handle a null in the REMARKS column
of getTables()/getColumns()/getProcedures.
Patch tested with jdbc2 against both a 7.1 and a CVS tip
backend. I did not have a jdbc1 environment to build and test
with, but since the touched code is identical in jdbc1 and jdbc2
I don't foresee any problems.
Regards,
Ren? Pijlman
Bruce Momjian [Fri, 17 Aug 2001 03:09:31 +0000 (03:09 +0000)]
Force crypt() salt to be null-terminated.
Bruce Momjian [Fri, 17 Aug 2001 02:59:20 +0000 (02:59 +0000)]
Add 4-byte MD5 salt.
Tom Lane [Thu, 16 Aug 2001 22:24:43 +0000 (22:24 +0000)]
geo_distance function needs to be marked strict.
From Mark Stosberg.
Bruce Momjian [Thu, 16 Aug 2001 21:53:27 +0000 (21:53 +0000)]
Update list of files to update.
Tom Lane [Thu, 16 Aug 2001 20:38:56 +0000 (20:38 +0000)]
Sequences are now based on int8, not int4, arithmetic. SERIAL pseudo-type
has an alias SERIAL4 and a sister SERIAL8. SERIAL8 is just the same
except the created column is type int8 not int4.
initdb forced. Note this also breaks any chance of pg_upgrade from 7.1,
unless we hack up pg_upgrade to drop and recreate sequences. (Which is
not out of the question, but I don't wanna do it.)
Bruce Momjian [Thu, 16 Aug 2001 16:24:16 +0000 (16:24 +0000)]
Add new MD5 pg_hba.conf keyword. Prevent fallback to crypt.
Bruce Momjian [Thu, 16 Aug 2001 15:21:16 +0000 (15:21 +0000)]
This patch fixes the well-known but unfixed bug that fetchone() always returns
the first result in the DB-API compliant wrapper. It turned out that the bug
was way down in the C code.
Gerhard Häring
Tatsuo Ishii [Thu, 16 Aug 2001 04:30:41 +0000 (04:30 +0000)]
Fix typo. pg_dump -B --> pg_dump -b
Bruce Momjian [Thu, 16 Aug 2001 04:27:18 +0000 (04:27 +0000)]
Remove protocol version change. Try MD5 first, then crypt() on all clients.
Bruce Momjian [Wed, 15 Aug 2001 23:22:49 +0000 (23:22 +0000)]
Use malloc/palloc as appropriate.
Bruce Momjian [Wed, 15 Aug 2001 21:08:21 +0000 (21:08 +0000)]
Move md5.h contents to crypt.h.
Bruce Momjian [Wed, 15 Aug 2001 19:42:34 +0000 (19:42 +0000)]
Remove unneeded patch.
Bruce Momjian [Wed, 15 Aug 2001 19:41:59 +0000 (19:41 +0000)]
Remove UNDO dicussion. It was inconclusive, and too large.
Bruce Momjian [Wed, 15 Aug 2001 19:41:08 +0000 (19:41 +0000)]
Just a test.
Bruce Momjian [Wed, 15 Aug 2001 19:40:52 +0000 (19:40 +0000)]
update.
Bruce Momjian [Wed, 15 Aug 2001 19:24:55 +0000 (19:24 +0000)]
update
Bruce Momjian [Wed, 15 Aug 2001 19:24:40 +0000 (19:24 +0000)]
update
Bruce Momjian [Wed, 15 Aug 2001 18:42:55 +0000 (18:42 +0000)]
Add new files.
Bruce Momjian [Wed, 15 Aug 2001 18:42:16 +0000 (18:42 +0000)]
Use MD5 for wire protocol encryption for >= 7.2 client/server.
Allow pg_shadow to be MD5 encrypted.
Add ENCRYPTED/UNENCRYPTED option to CREATE/ALTER user.
Add password_encryption postgresql.conf option.
Update wire protocol version to 2.1.
Tatsuo Ishii [Wed, 15 Aug 2001 07:10:12 +0000 (07:10 +0000)]
Add convert.
Tatsuo Ishii [Wed, 15 Aug 2001 07:07:40 +0000 (07:07 +0000)]
Add conver/convert2 functions. They are similar to the SQL99's convert.
Bruce Momjian [Tue, 14 Aug 2001 23:38:39 +0000 (23:38 +0000)]
Put back changes I overwrote packaging 7.1.3.
Bruce Momjian [Tue, 14 Aug 2001 23:38:20 +0000 (23:38 +0000)]
Put back changes I overwrote in packaging 7.1.3.
Tom Lane [Tue, 14 Aug 2001 22:21:59 +0000 (22:21 +0000)]
sum() on int2 and int4 columns now uses an int8, not numeric, accumulator
for speed reasons; its result type also changes to int8. avg() on these
datatypes now accumulates the running sum in int8 for speed; but we still
deliver the final result as numeric, so that fractional accuracy is
preserved.
count() now counts and returns in int8, not int4. I am a little nervous
about this possibly breaking users' code, but there didn't seem to be
a strong sentiment for avoiding the problem. If we get complaints during
beta, we can change count back to int4 and add a "count8" aggregate.
For that matter, users can do it for themselves with a simple CREATE
AGGREGATE command; the int4inc function is still present, so no C hacking
is needed.
Also added max() and min() aggregates for OID that do proper unsigned
comparison, instead of piggybacking on int4 aggregates.
initdb forced.
Bruce Momjian [Tue, 14 Aug 2001 21:21:40 +0000 (21:21 +0000)]
Add HISTORY for 7.1.3. Packaging done.
Tom Lane [Tue, 14 Aug 2001 17:12:57 +0000 (17:12 +0000)]
Fix brokenness of nested EXCEPT/INTERSECT queries. prepunion was being
a tad sloppy about generating the targetlist for some nodes, by generating
a tlist entry that claimed to be a constant when the value wasn't actually
constant. This caused setrefs.c to do the wrong thing later on.
Peter Eisentraut [Mon, 13 Aug 2001 21:34:54 +0000 (21:34 +0000)]
Make LANCOMPILER clause in CREATE LANGUAGE optional. Allow "identifier"
syntax for language names (instead of 'string').
createlang now handles the case where a second language uses the same call
handler as an already installed language (e.g., plperl/plperlu).
droplang now handles the reverse case, i.e., dropping a language where
the call handler is still used by another language. Moreover, droplang
can now be used to drop any user-defined language, not just the supplied
ones.
Tom Lane [Mon, 13 Aug 2001 19:50:11 +0000 (19:50 +0000)]
Make hashjoin give the right answer with toasted input data.
Tom Lane [Mon, 13 Aug 2001 18:45:36 +0000 (18:45 +0000)]
Add comparison operators and btree indexing support for type bytea.
From Joe Conway.
Bruce Momjian [Mon, 13 Aug 2001 15:24:46 +0000 (15:24 +0000)]
Patch NAME section of reference pages.
Tom Lane [Sun, 12 Aug 2001 21:35:19 +0000 (21:35 +0000)]
Make ALTER TABLE RENAME on a view rename the view's on-select rule too.
Needed to keep pg_dump from getting confused.
Peter Eisentraut [Sun, 12 Aug 2001 19:02:39 +0000 (19:02 +0000)]
Make pg_dump handle the new privileges.
Don't hardcode the maximum accepted server version, use PG_VERSION instead.
Install a notice processor so notices are handled like error messages.
Word smithing.
Peter Eisentraut [Sat, 11 Aug 2001 10:52:09 +0000 (10:52 +0000)]
Clean up some warnings and bugs and make things build easier.
Tom Lane [Sat, 11 Aug 2001 00:02:13 +0000 (00:02 +0000)]
CREATE VIEW with optional column name list wasn't quite right for the
case where there are resjunk columns in the query.
Tom Lane [Fri, 10 Aug 2001 23:49:20 +0000 (23:49 +0000)]
Document column-name-list option of CREATE VIEW ... which has been there
awhile, but the man page didn't know it.
Tom Lane [Fri, 10 Aug 2001 23:29:46 +0000 (23:29 +0000)]
Revert removal of relhaspkey support; fix unnecessary use of pg_index.oid.
Tom Lane [Fri, 10 Aug 2001 23:09:30 +0000 (23:09 +0000)]
Mention that COPY cannot be used on a view (per recent suggestion).
Other small improvements.
Tom Lane [Fri, 10 Aug 2001 22:50:10 +0000 (22:50 +0000)]
Since PQoidStatus is deprecated, we should probably stop using it in
our own code ...
Peter Eisentraut [Fri, 10 Aug 2001 20:58:09 +0000 (20:58 +0000)]
update
Tom Lane [Fri, 10 Aug 2001 20:52:25 +0000 (20:52 +0000)]
Cleanup some minor oversights in optional-OIDs stuff.
Tom Lane [Fri, 10 Aug 2001 18:57:42 +0000 (18:57 +0000)]
Make OIDs optional, per discussions in pghackers. WITH OIDS is still the
default, but OIDS are removed from many system catalogs that don't need them.
Some interesting side effects: TOAST pointers are 20 bytes not 32 now;
pg_description has a three-column key instead of one.
Bugs fixed in passing: BINARY cursors work again; pg_class.relhaspkey
has some usefulness; pg_dump dumps comments on indexes, rules, and
triggers in a valid order.
initdb forced.
Tom Lane [Fri, 10 Aug 2001 18:40:53 +0000 (18:40 +0000)]
Fix collateral damage from error message cleanup.
Peter Eisentraut [Fri, 10 Aug 2001 15:49:39 +0000 (15:49 +0000)]
message refinements
Bruce Momjian [Fri, 10 Aug 2001 15:43:49 +0000 (15:43 +0000)]
No longer need TODO.detail/lock.
Bruce Momjian [Fri, 10 Aug 2001 14:42:47 +0000 (14:42 +0000)]
The portion about the PQgetssl() function in the libpq documentation
appears to be duplicated (two identical <listitems> right after each
other). Here is a quick patch to remove one instance of it.
Magnus Hagander
Bruce Momjian [Fri, 10 Aug 2001 14:42:07 +0000 (14:42 +0000)]
Attached is a patch to remove some redundant code in the JDBC driver.
* Merges identical code from org.postgresql.jdbc[1|2].Statement into
org.postgresql.Statement.
* Moves escapeSQL() method from Connection to Statement (the only place
it's used)
* Minor cleanup of the new isolation level stuff.
* Minor cleanup of version string handling.
Anders Bengtsson
Bruce Momjian [Fri, 10 Aug 2001 14:34:28 +0000 (14:34 +0000)]
1. null-safe interface to GiST
(as proposed in http://fts.postgresql.org/db/mw/msg.html?mid=
1028327)
2. support for 'pass-by-value' arguments - to test this
we used special opclass for int4 with values in range [0-2^15]
More testing will be done after resolving problem with
index_formtuple and implementation of B-tree using GiST
3. small patch to contrib modules (seg,cube,rtree_gist,intarray) -
mark functions as 'isstrict' where needed.
Oleg Bartunov
Bruce Momjian [Fri, 10 Aug 2001 14:30:15 +0000 (14:30 +0000)]
Patch to LOCK multiple tables in one LOCK command.
Neil Padgett
Bruce Momjian [Fri, 10 Aug 2001 14:29:30 +0000 (14:29 +0000)]
Update TODO list.
Bruce Momjian [Fri, 10 Aug 2001 14:23:33 +0000 (14:23 +0000)]
Update TODO list.
Hiroshi Inoue [Fri, 10 Aug 2001 01:24:15 +0000 (01:24 +0000)]
Add new files for the preparation.
Peter Eisentraut [Thu, 9 Aug 2001 21:03:11 +0000 (21:03 +0000)]
This should be the right expected file.
Peter Eisentraut [Thu, 9 Aug 2001 18:28:18 +0000 (18:28 +0000)]
Use format_type sibling in backend error messages, so the user sees
consistent type naming.
Peter Eisentraut [Thu, 9 Aug 2001 18:13:23 +0000 (18:13 +0000)]
No longer a need for -Wno-error
Bruce Momjian [Thu, 9 Aug 2001 16:38:59 +0000 (16:38 +0000)]
Update branding for 7.1.3, no HISTORY yet.
Peter Eisentraut [Thu, 9 Aug 2001 16:20:43 +0000 (16:20 +0000)]
corrections from the DocNotes
Tom Lane [Thu, 9 Aug 2001 13:52:06 +0000 (13:52 +0000)]
Fix Cygwin build, per Jason Tishler.
Tom Lane [Thu, 9 Aug 2001 03:32:16 +0000 (03:32 +0000)]
Add TOAST table to the set of relkinds known to \d.
Peter Eisentraut [Wed, 8 Aug 2001 13:30:20 +0000 (13:30 +0000)]
0 -> O
Peter Eisentraut [Tue, 7 Aug 2001 22:41:49 +0000 (22:41 +0000)]
some clarifications inspired by the DocNotes
Bruce Momjian [Tue, 7 Aug 2001 18:18:33 +0000 (18:18 +0000)]
Update README for fuzzystrmatch description.
Bruce Momjian [Tue, 7 Aug 2001 18:16:01 +0000 (18:16 +0000)]
Sorry - I should have gotten to this sooner. Here's a patch which you should
be able to apply against what you just committed. It rolls soundex into
fuzzystrmatch.
Remove soundex/metaphone and merge into fuzzystrmatch.
Joe Conway
Bruce Momjian [Tue, 7 Aug 2001 17:45:29 +0000 (17:45 +0000)]
I think you replaced too many things with put(...
Here is a context diff from latest cvs
And I see why you couldn't apply the last diff, the setCatalog diff has
been backed out, that was causing the compile problem in the first
place.
This following one needs to be applied to allow the current cvs to
compile
Dave Cramer
Bruce Momjian [Tue, 7 Aug 2001 16:56:17 +0000 (16:56 +0000)]
A small patch to keep postgres working on the latest BeOS.
Cyril VELTER
Bruce Momjian [Tue, 7 Aug 2001 16:47:43 +0000 (16:47 +0000)]
Per this discussion, here's a patch to implement both levenshtein() and
metaphone() in a contrib. There seem to be a fair number of different
approaches to both of these algorithms. I used the simplest case for
levenshtein which has a cost of 1 for any character insertion, deletion, or
substitution. For metaphone, I adapted the same code from CPAN that the PHP
folks did.
A couple of questions:
1. Does it make sense to fold the soundex contrib together with this one?
2. I was debating trying to add multibyte support to levenshtein (it would
make no sense at all for metaphone), but a quick search through the contrib
directory found no hits on the word MULTIBYTE. Should worry about adding
multibyte support to levenshtein()?
Joe Conway
Tom Lane [Tue, 7 Aug 2001 15:55:16 +0000 (15:55 +0000)]
Fix thinko (revealed by gcc warning).
Peter Eisentraut [Tue, 7 Aug 2001 11:41:17 +0000 (11:41 +0000)]
Russian translation by Serguei Mokhov
Peter Eisentraut [Tue, 7 Aug 2001 11:28:18 +0000 (11:28 +0000)]
Czech translation for psql from Karel Zak
Peter Eisentraut [Tue, 7 Aug 2001 10:44:16 +0000 (10:44 +0000)]
Add a check for end of client connection before expecting a password
response, to avoid noise in the server log.
Tom Lane [Mon, 6 Aug 2001 22:55:00 +0000 (22:55 +0000)]
Doesn't seem to be much point in keeping this README up to date anymore,
since it's completely redundant with regress.sgml. I think we agreed to
remove it awhile back, actually, but no one got around to doing it.
Tom Lane [Mon, 6 Aug 2001 22:53:26 +0000 (22:53 +0000)]
Update now-obsolete example of platform-specific regression comparison
files.
Peter Eisentraut [Mon, 6 Aug 2001 21:55:16 +0000 (21:55 +0000)]
Use a fixed error message for ERANGE to avoid duplicate test result files.
Add some resultmap entries for SCO OpenServer.
Tom Lane [Mon, 6 Aug 2001 18:17:42 +0000 (18:17 +0000)]
Seems like a bad idea to free() a string we are about to use in an
error message.
Tom Lane [Mon, 6 Aug 2001 18:09:45 +0000 (18:09 +0000)]
Modify partial-index-predicate applicability tester to test whether
clauses are equal(), before trying to match them up using btree opclass
inference rules. This allows it to recognize many simple cases involving
non-btree operations, for example 'x IS NULL'. Clean up code a little.
Tom Lane [Mon, 6 Aug 2001 18:05:07 +0000 (18:05 +0000)]
Evaluate LIMIT/OFFSET expressions with ExecEvalExprSwitchContext, not
ExecEvalExpr, to avoid possible memory leak.
Peter Eisentraut [Mon, 6 Aug 2001 15:46:44 +0000 (15:46 +0000)]
Do not pre-expand localedir as substituted in the makefile, so that 'make
install prefix=elsewhere' works.
Peter Eisentraut [Mon, 6 Aug 2001 13:58:26 +0000 (13:58 +0000)]
update from upstream
Peter Eisentraut [Mon, 6 Aug 2001 13:45:15 +0000 (13:45 +0000)]
Check that the data directory does not have group or world access; remove
a similar check on postgresql.conf.
Bruce Momjian [Mon, 6 Aug 2001 05:42:48 +0000 (05:42 +0000)]
Add QueryIsRule gram.y reset.
Tom Lane [Mon, 6 Aug 2001 01:25:32 +0000 (01:25 +0000)]
Surely the pg_stat interrogation functions must not be marked proiscachable.
Tom Lane [Sun, 5 Aug 2001 22:13:46 +0000 (22:13 +0000)]
Show index predicate when doing \d on a partial index.
Tom Lane [Sun, 5 Aug 2001 02:06:50 +0000 (02:06 +0000)]
Endeavor to make pgstats buffer process (a) safe and (b) useful.
Make sure it exits immediately when collector process dies --- in old code,
buffer process would hang around and compete with the new buffer process
for packets. Make sure it doesn't block on writing the pipe when the
collector falls more than a pipeload behind. Avoid leaking pgstats FDs
into every backend.
Tom Lane [Sun, 5 Aug 2001 01:22:16 +0000 (01:22 +0000)]
Remove no-longer-needed fcntl call (I'm not sure it *ever* did anything
useful, in fact).
Bruce Momjian [Sat, 4 Aug 2001 22:01:39 +0000 (22:01 +0000)]
Back out LOCK A,B,C patch at Tom's suggestion.
Bruce Momjian [Sat, 4 Aug 2001 19:46:56 +0000 (19:46 +0000)]
Compile fix for jdbc1.
Bruce Momjian [Sat, 4 Aug 2001 19:42:34 +0000 (19:42 +0000)]
This patch is because Hurd does not support NOFILE. It is against current
cvs.
The Debian bug report says, "The upstream source makes use of NOFILE
unconditionalized. As the Hurd doesn't have an arbitrary limit on the
number of open files, this is not defined. But _SC_OPEN_MAX works fine
and returns 1024 (applications can increase this as they want), so I
suggest the below diff. Please forward this upstream, too."
Oliver Elphick
Bruce Momjian [Sat, 4 Aug 2001 19:41:00 +0000 (19:41 +0000)]
This patch adds the following to the FTI module:
* The ability to index more than one column in a table with a single
trigger.
* All uses of sprintf changed to snprintf to prevent users from crashing
Postgres.
* Error messages made more consistent
* Some changes made to bring it into line with coding requirements for
triggers specified in the docs. (ie. check you're a trigger before casting
your context)
* The perl script that generate indices has been updated to support indexing
multiple columns in a table.
* Fairly well tested in our development environment indexing a food
database's brand and description fields. The size of the fti index is
around 300,000 rows.
* All docs and examples upgraded. This includes specifying more efficient
index usage that was specified before, better examples that don't produce
duplicates, etc.
Christopher Kings-Lynne & Brett
Bruce Momjian [Sat, 4 Aug 2001 19:39:00 +0000 (19:39 +0000)]
Add LOCK A,B,C functionality as LOCK A;LOCK B;LOCK C; as agreed.
Neil Padgett
Bruce Momjian [Sat, 4 Aug 2001 19:36:45 +0000 (19:36 +0000)]
Looks okay in a quick glance, except error message spelling is poor:
! #define ARRISNULL(x) ( (x) ? ( ( ARR_NDIM(x) == NDIM ) ? ( ( ARRNELEMS( x ) )
? 0 : 1 ) : ( ( ARR_NDIM(x) ) ? (elog(ERROR,"Array is not one-dimentional: %d di
mentions", ARR_NDIM(x)),1) : 1 ) ) : 1 )
Should be "one-dimensional" and "dimensions". Bruce, would you fix that
when you apply it?
Tom
Bruce Momjian [Sat, 4 Aug 2001 19:35:32 +0000 (19:35 +0000)]
1. Fixed error with empty array ( '{}' ),
test data changed to include such data
2. Test a dimension of an array ( we support only one-dimension)
Oleg Bartunov
Bruce Momjian [Sat, 4 Aug 2001 19:33:49 +0000 (19:33 +0000)]
> 1) When a row is retrieved, and then a SQL_FETCH_FIRST is issued, the
check
> in convert.c
> does not consider the fact that the value in the field has been altered to
> be a '1' if the
> backend handed it a 't'. The net result being that the first row on any
> subsequent queries
> has all it's boolean set to 0.
Aidan Mountford
Bruce Momjian [Sat, 4 Aug 2001 19:32:04 +0000 (19:32 +0000)]
Attached is a patch that does the following:
1) improves performance of commit/rollback by reducing number of round
trips to the server
2) uses 7.1 functionality for setting the transaction isolation level
3) backs out a patch from 11 days ago because that code failed to
compile under jdk1.1
Details:
1) The old code was doing the following for each commit:
commit
begin
set transaction isolation level xxx
thus a call to commit was performing three round trips to the database.
The new code does this in one round trip as:
commit; begin; set transaction isolation level xxx
In a simple test program that performs 1000 transactions (where each
transaction does one simple select inside that transaction) has the
following before and after timings:
Client and Server on same machine
old new
--- ---
1.877sec 1.405sec 25.1% improvement
Client and Server on different machines
old new
--- ---
4.184sec 2.927sec 34.3% improvement
(all timings are an average of four different runs)
2) The driver was using 'set transaction isolation level xxx' at the
begining of each transaction, instead of using the new 7.1 syntax of
'set session characteristics as transaction isolation level xxx' which
only needs to be done once instead of for each transaction. This is
done conditionally (i.e. if server is 7.0 or older do the old behaviour,
else do the new behaviour) to not break backward compatibility. This
also required the movement of some code to check/test database version
numbers from the DatabaseMetaData object to the Connection object.
3) Finally while testing, I discovered that the code that was checked in
11 days ago actually didn't compile. The code in the patch for
Connection.setCatalog() used Properties.setProperty() which only exists
in JDK1.2 or higher. Thus compiling the JDBC1 driver failed as this
method doesn't exist. Thus I backed out that patch.
Barry Lind