7.4 status
I have emptied the patch queue, and have updated the HISTORY file for
7.4. I am looking for any improvements to that file. (I have to move
it to SGML soon.)
Perhaps it is time to start looking at a final release date for 7.4?
--
Bruce Momjian | http://candle.pha.pa.us
pgman@candle.pha.pa.us | (610) 359-1001
+ If your life is a hard drive, | 13 Roberts Road
+ Christ can be your backup. | Newtown Square, Pennsylvania 19073
Bruce Momjian <pgman@candle.pha.pa.us> writes:
Perhaps it is time to start looking at a final release date for 7.4?
At the very least we need to set a strings freeze soon, so the
translators can catch up. Peter, are you getting close to done with the
message revisions you've been making?
regards, tom lane
Should we maybe get a Beta4 out now that everything is caught up code
wise? Is anyone still sitting on something (other then the translations
stuff) that should be in v7.4?
On Mon, 29 Sep 2003, Tom Lane wrote:
Show quoted text
Bruce Momjian <pgman@candle.pha.pa.us> writes:
Perhaps it is time to start looking at a final release date for 7.4?
At the very least we need to set a strings freeze soon, so the
translators can catch up. Peter, are you getting close to done with the
message revisions you've been making?regards, tom lane
---------------------------(end of broadcast)---------------------------
TIP 8: explain analyze is your friend
On Mon, 2003-09-29 at 14:57, Bruce Momjian wrote:
I have emptied the patch queue, and have updated the HISTORY file for
7.4. I am looking for any improvements to that file. (I have to move
it to SGML soon.)
A patch that improves HISTORY is attached. I added a few bullet items to
the "highlights of this release section" that are IMHO important,
removed the mention of hash indexes (IMHO, they are still inappropriate
for production use, just less so), fixed some spelling mistakes, and
clarified a few entries.
Please apply.
-Neil
Attachments:
relnotes_cleanup-1.patchtext/x-patch; charset=ANSI_X3.4-1968; name=relnotes_cleanup-1.patchDownload
Index: HISTORY
===================================================================
RCS file: /var/lib/cvs/pgsql-server/HISTORY,v
retrieving revision 1.216
diff -c -r1.216 HISTORY
*** HISTORY 29 Sep 2003 18:06:48 -0000 1.216
--- HISTORY 29 Sep 2003 20:26:57 -0000
***************
*** 15,52 ****
Major changes in this release:
IPv6
! Full support for IPv6 connections and IPv6 data types.
SSL
! Major SSL improvements in performance and security.
Index Growth Prevention
Allow free space map to efficiently reused empty index pages,
! and other free space improvements.
New Client/Server Communication Protocol
! New protocol improves connection speed/reliability, and adds
! error codes, status information, a binary protocol, error
! reporting verbosity, and cleaner startup packets.
Performance
! IN/NOT IN subqueries now perform as efficiently as joins
! Improved GROUP BY procesing by using hack buckets
! Hash indexes now have better concurrency/reliability/performance
! New multi-key hash join capability
! ANSI joins are now fully optimized
! Faster regular expression code
! Cursors
! Allow cursors to exist outside transactions
Threads
! libpq and ecpg are now fully thread-safe with --enable-thread-safety
Contrib
! New version of full text indexing (tsearch2)
! New autovacuum tool
! Array handling has been improved and moved into the main server
_________________________________________________________________
--- 15,57 ----
Major changes in this release:
IPv6
! Full support for IPv6 connections and IPv6 address data types
SSL
! Major improvements in SSL performance and security
Index Growth Prevention
Allow free space map to efficiently reused empty index pages,
! and other free space management improvements.
!
! Standards Compliance
! Implement information schema
! Support for read-only transactions
! Make cursors comply more closely with the SQL standard
New Client/Server Communication Protocol
! New protocol improves connection speed/reliability, and adds
! error codes, status information, a binary protocol, error
! reporting verbosity, and cleaner startup packets.
Performance
! IN/NOT IN subqueries now perform as efficiently as joins
! Improved GROUP BY processing by using hash buckets
! New multi-key hash join capability
! ANSI joins are now fully optimized
! Faster regular expression code
! Function-inlining for simple SQL functions
! Holdable Cursors
! Allow cursors to exist outside transactions
Threads
! libpq and ecpg are now fully thread-safe with --enable-thread-safety
Contrib
! New version of full text indexing (tsearch2)
! New autovacuum tool
! Array handling has been improved and moved into the main server
_________________________________________________________________
***************
*** 58,75 ****
Observe the following incompatibilities:
* The server-side autocommit setting was removed an reimplemented
in client applications and languages.
! * Error message wording has changed dramtically in this release,
and error codes have been added.
* ANSI joins may behave differently because they are now fully optimized
* A number of server variables have been renamed for clarity
* MOVE/FETCH 0 now does nothing; return value of MOVE/FETCH 1/0
varies based on the current offset in the cursor
! * COPY now can process carriage-return/line-feed and carraige-return
terminated files. Literal carriage-returns and line-feeds are no
longer accepted as data values; use \r and \n instead.
! * Auto-trim spaces when converting from CHAR() to VARCHAR()/TEXT
* FLOAT(p) now measures 'p' in bits, not digits
* Date values now must match the ordering specified by DateStyle
_________________________________________________________________
--- 63,83 ----
Observe the following incompatibilities:
* The server-side autocommit setting was removed an reimplemented
in client applications and languages.
! * Error message wording has changed dramatically in this release,
and error codes have been added.
* ANSI joins may behave differently because they are now fully optimized
* A number of server variables have been renamed for clarity
* MOVE/FETCH 0 now does nothing; return value of MOVE/FETCH 1/0
varies based on the current offset in the cursor
! * COPY now can process carriage-return/line-feed and carriage-return
terminated files. Literal carriage-returns and line-feeds are no
longer accepted as data values; use \r and \n instead.
! * Trailing spaces are now trimmed when converting from CHAR(n) to
! VARCHAR(n)/TEXT
* FLOAT(p) now measures 'p' in bits, not digits
* Date values now must match the ordering specified by DateStyle
+ * The oidrand(), oidsrand(), and userfntest() functions have been
+ removed.
_________________________________________________________________
***************
*** 177,190 ****
Object Manipulation
Make CREATE SEQUENCE grammar more SQL1999 standards compliant (Neil)
! Add FOR EACH STATEMENT statement-level triggers (Neil Conway)
Add DOMAIN CHECK constraints (Rod)
Add ALTER DOMAIN .. SET / DROP NOT NULL, SET / DROP DEFAULT, ADD / DROP
CONSTRAINT (Rod)
Fix several zero-column table bugs (Tom)
Have ALTER TABLE ... ADD PRIMARY KEY add NOT NULL constraint (Rod)
Add ALTER DOMAIN OWNER (Rod)
! Add ALTER TABLE ... WITHOUT OIDS? (Rod)
Add ALTER SEQUENCE to modify min/max/increment/cache/cycle values (Rod)
Add ALTER TABLE ... CLUSTER ON (Alvaro Herrera)
Improve DOMAIN automatic type casting (Rod, Tom)
--- 185,198 ----
Object Manipulation
Make CREATE SEQUENCE grammar more SQL1999 standards compliant (Neil)
! Add FOR EACH STATEMENT statement-level triggers (Neil)
Add DOMAIN CHECK constraints (Rod)
Add ALTER DOMAIN .. SET / DROP NOT NULL, SET / DROP DEFAULT, ADD / DROP
CONSTRAINT (Rod)
Fix several zero-column table bugs (Tom)
Have ALTER TABLE ... ADD PRIMARY KEY add NOT NULL constraint (Rod)
Add ALTER DOMAIN OWNER (Rod)
! Add ALTER TABLE ... WITHOUT OIDS (Rod)
Add ALTER SEQUENCE to modify min/max/increment/cache/cycle values (Rod)
Add ALTER TABLE ... CLUSTER ON (Alvaro Herrera)
Improve DOMAIN automatic type casting (Rod, Tom)
***************
*** 225,231 ****
Allow pg_dumpall to support the -a, -s, -x options of pg_dump (Tom)
Prevent pg_dump from lowercasing identifiers specified on the command line (Tom)
Allow PREPARE/bind of utility commands like FETCH and EXPLAIN (Tom)
! Allow EXPLAIN EXECUTE (Neil)
Allow pg_get_constraintdef() to support UNIQUE, PRIMARY KEY and
CHECK constraints (Christopher)
Improve VACUUM performance on indexes by reducing WAL traffic (Tom)
--- 233,239 ----
Allow pg_dumpall to support the -a, -s, -x options of pg_dump (Tom)
Prevent pg_dump from lowercasing identifiers specified on the command line (Tom)
Allow PREPARE/bind of utility commands like FETCH and EXPLAIN (Tom)
! Add EXPLAIN EXECUTE (Neil)
Allow pg_get_constraintdef() to support UNIQUE, PRIMARY KEY and
CHECK constraints (Christopher)
Improve VACUUM performance on indexes by reducing WAL traffic (Tom)
***************
*** 235,241 ****
Have SHOW TRANSACTION_ISOLATION match input to SET TRANSACTION_ISOLATION (Tom)
Have COMMENT ON DATABASE on non-local database generate a warning (Rod)
Improve reliability of LISTEN/NOTIFY (Tom)
! Allow REINDEX to reliabily reindex all indexes, except global ones (Tom)
pg_dump --use-set-session-authorization and --no-reconnect now do nothing,
all dumps use SET SESSION AUTHORIZATION
--- 243,249 ----
Have SHOW TRANSACTION_ISOLATION match input to SET TRANSACTION_ISOLATION (Tom)
Have COMMENT ON DATABASE on non-local database generate a warning (Rod)
Improve reliability of LISTEN/NOTIFY (Tom)
! Allow REINDEX to reliably reindex non-shared system catalog indexes (Tom)
pg_dump --use-set-session-authorization and --no-reconnect now do nothing,
all dumps use SET SESSION AUTHORIZATION
***************
*** 289,295 ****
'now' will no longer work as a column default, use now() (change required for
prepared statements) (Tom)
Assume NaN value to be larger than any other value in MIN()/MAX() (Tom)
! Prevent interval from supressing ':00' seconds display
New pg_get_triggerdef(prettyprint) and pg_constraint_is_visible() functions
Allow time to be specified as '040506' or '0405' (Tom)
--- 297,303 ----
'now' will no longer work as a column default, use now() (change required for
prepared statements) (Tom)
Assume NaN value to be larger than any other value in MIN()/MAX() (Tom)
! Prevent interval from suppressing ':00' seconds display
New pg_get_triggerdef(prettyprint) and pg_constraint_is_visible() functions
Allow time to be specified as '040506' or '0405' (Tom)
"Marc G. Fournier" <scrappy@postgresql.org> writes:
Should we maybe get a Beta4 out now that everything is caught up code
wise? Is anyone still sitting on something (other then the translations
stuff) that should be in v7.4?
We still have several open items in Bruce's list, but maybe we can
resolve them soon. Tentatively set a beta for Wed the 1st, maybe?
regards, tom lane
Patch applied. Thanks.
---------------------------------------------------------------------------
Neil Conway wrote:
On Mon, 2003-09-29 at 14:57, Bruce Momjian wrote:
I have emptied the patch queue, and have updated the HISTORY file for
7.4. I am looking for any improvements to that file. (I have to move
it to SGML soon.)A patch that improves HISTORY is attached. I added a few bullet items to
the "highlights of this release section" that are IMHO important,
removed the mention of hash indexes (IMHO, they are still inappropriate
for production use, just less so), fixed some spelling mistakes, and
clarified a few entries.Please apply.
-Neil
[ Attachment, skipping... ]
--
Bruce Momjian | http://candle.pha.pa.us
pgman@candle.pha.pa.us | (610) 359-1001
+ If your life is a hard drive, | 13 Roberts Road
+ Christ can be your backup. | Newtown Square, Pennsylvania 19073
Tom Lane writes:
At the very least we need to set a strings freeze soon, so the
translators can catch up. Peter, are you getting close to done with the
message revisions you've been making?
Yes, I think we're ready for a string freeze. Alvaro, do you have
anything you still want to submit in that area?
--
Peter Eisentraut peter_e@gmx.net
On Mon, Sep 29, 2003 at 11:50:23PM +0200, Peter Eisentraut wrote:
Tom Lane writes:
At the very least we need to set a strings freeze soon, so the
translators can catch up. Peter, are you getting close to done with the
message revisions you've been making?Yes, I think we're ready for a string freeze. Alvaro, do you have
anything you still want to submit in that area?
The only things left that I can see are
#: commands/tablecmds.c:4093
msgid "tables \"%s\" already has a TOAST table"
"tables" -> "table"
#: commands/user.c:651 commands/user.c:1357
msgid "sysid %d is already assigned"
"sysid" -> "user ID"? Not sure, maybe it can be a group ID.
--
Alvaro Herrera (<alvherre[a]dcc.uchile.cl>)
"I think my standards have lowered enough that now I think 'good design'
is when the page doesn't irritate the living f*ck out of me." (JWZ)
Neil Conway <neilc@samurai.com> writes:
SSL
! Major improvements in SSL performance and security
Did we actually add any "security" to the SSL code? Performance and
reliability, maybe, but I didn't think we'd done anything to the
security algorithms per se. Did I miss something?
Index Growth Prevention
Allow free space map to efficiently reused empty index pages,
! and other free space management improvements.
"reused" -> "reuse"
! ANSI joins are now fully optimized
"better" optimized maybe; I wouldn't say we're done with that issue yet.
(In particular, I want to look into reordering outer joins sometime.)
! Faster regular expression code
We could tout more functionality too, since the new regex package
has a lot of advanced stuff that wasn't there before.
* The server-side autocommit setting was removed an reimplemented
in client applications and languages.
"an" -> "and"
! * Error message wording has changed dramatically in this release,
and error codes have been added.
"Substantially" might be a better term anyway?
* ANSI joins may behave differently because they are now fully optimized
See above
* Date values now must match the ordering specified by DateStyle
... except that YYYY-MM-DD (with 3 or more digits for year) is always
accepted.
regards, tom lane
Tom Lane wrote:
Neil Conway <neilc@samurai.com> writes:
SSL
! Major improvements in SSL performance and securityDid we actually add any "security" to the SSL code? Performance and
reliability, maybe, but I didn't think we'd done anything to the
security algorithms per se. Did I miss something?
I changed the wording to reliability. We did so some security by
changing the rate of key negotiation, but it isn't substantial.
Index Growth Prevention
Allow free space map to efficiently reused empty index pages,
! and other free space management improvements."reused" -> "reuse"
Done.
! ANSI joins are now fully optimized
"better" optimized maybe; I wouldn't say we're done with that issue yet.
(In particular, I want to look into reordering outer joins sometime.)
Yep.
! Faster regular expression code
We could tout more functionality too, since the new regex package
has a lot of advanced stuff that wasn't there before.
Added "more powerful"
* The server-side autocommit setting was removed an reimplemented
in client applications and languages."an" -> "and"
Done.
! * Error message wording has changed dramatically in this release,
and error codes have been added."Substantially" might be a better term anyway?
Yep.
* ANSI joins may behave differently because they are now fully optimized
See above
Changed to "better optimized"
* Date values now must match the ordering specified by DateStyle
... except that YYYY-MM-DD (with 3 or more digits for year) is always
accepted.
Changed to:
* Ambiguous date values now must match the ordering specified by DateStyle
Committed to CVS.
--
Bruce Momjian | http://candle.pha.pa.us
pgman@candle.pha.pa.us | (610) 359-1001
+ If your life is a hard drive, | 13 Roberts Road
+ Christ can be your backup. | Newtown Square, Pennsylvania 19073
Bruce Momjian kirjutas T, 30.09.2003 kell 02:16:
Tom Lane wrote:
! Faster regular expression code
We could tout more functionality too, since the new regex package
has a lot of advanced stuff that wasn't there before.Added "more powerful"
This wording covers nicely possible incompatibilities too ;)
--------------
Hannu
On Mon, Sep 29, 2003 at 11:50:23PM +0200, Peter Eisentraut wrote:
Tom Lane writes:
At the very least we need to set a strings freeze soon, so the
translators can catch up. Peter, are you getting close to done with the
message revisions you've been making?Yes, I think we're ready for a string freeze. Alvaro, do you have
anything you still want to submit in that area?
Does this count a string change? :)
Patrick
Attachments:
pattext/plain; charset=us-asciiDownload
? psql
Index: help.c
===================================================================
RCS file: /projects/cvsroot/pgsql-server/src/bin/psql/help.c,v
retrieving revision 1.80
diff -u -r1.80 help.c
--- help.c 14 Sep 2003 22:37:13 -0000 1.80
+++ help.c 1 Oct 2003 19:58:15 -0000
@@ -234,8 +234,8 @@
ON(pset.popt.topt.format == PRINT_HTML));
fprintf(output, _(" \\pset NAME [VALUE]\n"
" set table output option\n"
- " (NAME := {format|border|expanded|fieldsep|null|recordsep|\n"
- " tuples_only|title|tableattr|pager})\n"));
+ " (NAME := {format|border|expanded|fieldsep|footer|null|\n"
+ " recordsep|tuples_only|title|tableattr|pager})\n"));
fprintf(output, _(" \\t show only rows (currently %s)\n"),
ON(pset.popt.topt.tuples_only));
fprintf(output, _(" \\T [STRING] set HTML <table> tag attributes, or unset if none\n"));
Alvaro Herrera writes:
The only things left that I can see are
#: commands/tablecmds.c:4093
msgid "tables \"%s\" already has a TOAST table"
"tables" -> "table"
Fixed.
#: commands/user.c:651 commands/user.c:1357
msgid "sysid %d is already assigned"
"sysid" -> "user ID"? Not sure, maybe it can be a group ID.
Changed to "user ID" and "group ID". Sounds better to me too.
--
Peter Eisentraut peter_e@gmx.net
Patrick Welche writes:
Does this count a string change? :)
Yes, but it's also a documentation bug fix.
Anyway, you got it in before the freeze. :)
--
Peter Eisentraut peter_e@gmx.net
While reviewing someone else's translation of pg_dump I noted that the
phrase "ACL list" is used in a couple of places. However ACL stands for
"Access Control List", so the term "ACL list" seems redundant.
Maybe it should be replaced with plain "ACL"?
--
Alvaro Herrera (<alvherre[a]dcc.uchile.cl>)
"I think my standards have lowered enough that now I think 'good design'
is when the page doesn't irritate the living f*ck out of me." (JWZ)
Alvaro Herrera writes:
While reviewing someone else's translation of pg_dump I noted that the
phrase "ACL list" is used in a couple of places. However ACL stands for
"Access Control List", so the term "ACL list" seems redundant.
These kinds of redundancies are pretty common for the sake of clarity and
the flow of the language. Even major so-called "Usage Panels" accept
them.
The thing being parsed here is an array of datums of type "aclitem", so
the term "ACL list" is rather unclear anyway. But I hesitate to change it
now because we wanted to call a string freeze.
--
Peter Eisentraut peter_e@gmx.net