Re: Open items

Here are the open items. Feature freeze is Friday, July 1.

---------------------------------------------------------------------------

PostgreSQL 8.1 Open Items
=========================

Current version at http://candle.pha.pa.us/cgi-bin/pgopenitems.

Changes
-------
integrated auto-vacuum (Alvaro)
ICU locale patch?
Win32 signal handling patch (Magnus)
KRB4 removal patch (Magnus)
GIST concurrency fixes? (Oleg)
Roles? (Stephen Frost)
convert plperl arrays to pg arrays? (Andrew Dunstan)
plperl validator function? (Andrew Dunstan)
column-level triggers (Greg)
interval improvements (Michael Glaesemann)
move rtree_gist into core?
config file I/O?
terminate backend fix?
dbsize functions from /contrib?
fix pg_autovacuum O(n^2) behavior
remove wal siblings guc vars?
COPY performance improvements

Documentation
-------------

Fixed Since Last Beta
---------------------

--
Bruce Momjian | http://candle.pha.pa.us
pgman(at)candle(dot)pha(dot)pa(dot)us | (610) 359-1001
+ If your life is a hard drive, | 13 Roberts Road
+ Christ can be your backup. | Newtown Square, Pennsylvania 19073

On Sat, Jun 25, 2005 at 11:24:56PM -0400, Bruce Momjian wrote:

> Changes
> -------
> integrated auto-vacuum (Alvaro)

Working on it.

> Roles? (Stephen Frost)

Additionally, my shared dependencies patch :-) I'm not sure if I'll be
able to do both autovacuum and shared dependencies before freeze, but
I'll try. I'm not sure if anybody is working on roles though; if I have
to do the three of them I'm pretty sure I won't be able to meet the
deadline.

I think we should leave shared dependencies out if roles do not make it
into 8.1.

--
Alvaro Herrera (<alvherre[a]surnet.cl>)
"Por suerte hoy explotó el califont porque si no me habría muerto
de aburrido" (Papelucho)

Alvaro Herrera <alvherre(at)surnet(dot)cl> writes:
>> Roles? (Stephen Frost)

> Additionally, my shared dependencies patch :-) I'm not sure if I'll be
> able to do both autovacuum and shared dependencies before freeze, but
> I'll try. I'm not sure if anybody is working on roles though;

I'll try to finish the roles patch if Stephen sends in what he's got
within the next day or so.

regards, tom lane

On Sat, 25 Jun 2005, Bruce Momjian wrote:

> Here are the open items. Feature freeze is Friday, July 1.
>
> GIST concurrency fixes? (Oleg)
This should be
GIST concurrency support (Teodor, Oleg)

also, after moving rtree_gist into core we should add
regression tests for GIST.

Regards,
Oleg
_____________________________________________________________
Oleg Bartunov, sci.researcher, hostmaster of AstroNet,
Sternberg Astronomical Institute, Moscow University (Russia)
Internet: oleg(at)sai(dot)msu(dot)su, http://www.sai.msu.su/~megera/
phone: +007(095)939-16-83, +007(095)939-23-83

On Saturday 25 June 2005 23:24, Bruce Momjian wrote:
> Here are the open items. Feature freeze is Friday, July 1.
>
> ---------------------------------------------------------------------------
>
>
> PostgreSQL 8.1 Open Items
> =========================
>
> Current version at http://candle.pha.pa.us/cgi-bin/pgopenitems.
>
> Changes
> -------
> integrated auto-vacuum (Alvaro)
> ICU locale patch?
> Win32 signal handling patch (Magnus)
> KRB4 removal patch (Magnus)
> GIST concurrency fixes? (Oleg)
> Roles? (Stephen Frost)
> convert plperl arrays to pg arrays? (Andrew Dunstan)
> plperl validator function? (Andrew Dunstan)
> column-level triggers (Greg)
> interval improvements (Michael Glaesemann)
> move rtree_gist into core?
> config file I/O?
> terminate backend fix?
> dbsize functions from /contrib?
> fix pg_autovacuum O(n^2) behavior
> remove wal siblings guc vars?
> COPY performance improvements
>

Where does the rest of the "contrib roundup" fit into this picture? Is it
simply agreed that we will deal with them as outlined in Josh's and Tom's
email, or does that have to be *before* feature freeze if it is to be done in
8.1?

On a somewhat relative note, I was looking at fulltextindex earlier this week,
and had thoughts to move it to pgfoundry along with some documentation from
techdocs. However after discussing with ChrisKL, we both seem to reach a
conclusion that really that is deprecated code and should be deprecated in
favor of tsearch2. That being the case does anyone think we should still move
it?

--
Robert Treat
Build A Brighter Lamp :: Linux Apache {middleware} PostgreSQL

On L, 2005-06-25 at 23:24 -0400, Bruce Momjian wrote:
> Here are the open items. Feature freeze is Friday, July 1.

I'm still hoping to get my concurrent vacuum patch in shape before
feature freeze

--
Hannu Krosing <hannu(at)skype(dot)net>

Bruce Momjian wrote:

>convert plperl arrays to pg arrays? (Andrew Dunstan)
>
>

Won't be done :-( . Too many wrinkles. Covered by recently added TODO item.

>plperl validator function? (Andrew Dunstan)
>
>

Done and merged.

Also, I am going to try to get done putting language handlers in
pg_catalog instead of public, as per recent discussion.

And I still have my fingers crossed that Abhijit Menon-Sen will submit a
patch to do row fetching via cursors for plperl in time.

cheers

andrew

Robert Treat <xzilla(at)users(dot)sourceforge(dot)net> writes:
> On Saturday 25 June 2005 23:24, Bruce Momjian wrote:
>> Here are the open items. Feature freeze is Friday, July 1.

> Where does the rest of the "contrib roundup" fit into this picture?

I think that moving rtree_gist, reindexdb, and/or userlock into core
would have to happen before feature freeze, but the other items
remaining on the to-do list are simply removal of modules, and as such
not bound by feature freeze. Besides, we've always been laxer about the
rules for contrib than for the core system.

> On a somewhat relative note, I was looking at fulltextindex earlier
> this week, and had thoughts to move it to pgfoundry along with some
> documentation from techdocs. However after discussing with ChrisKL, we
> both seem to reach a conclusion that really that is deprecated code
> and should be deprecated in favor of tsearch2. That being the case
> does anyone think we should still move it?

It's certainly deprecated in favor of tsearch2 as far as the actual
functionality goes. I think that Josh had suggested preserving it
on pgfoundry more as an example of a nontrivial trigger function.
I wouldn't cry if we just dropped it though.

regards, tom lane

Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us> writes:
> Here are the open items. Feature freeze is Friday, July 1.

This is missing the point that pg_dump output is not forward compatible
with the planned escape string changeover.

We had talked about tweaking ruleutils.c, as well as pg_dump itself,
to use E'...' syntax if the string value contains a backslash ---
but the patch so far applied does none of that.

regards, tom lane

On Sat, 2005-06-25 at 23:24 -0400, Bruce Momjian wrote:
> Here are the open items. Feature freeze is Friday, July 1.
>
> ---------------------------------------------------------------------------
>
>
> PostgreSQL 8.1 Open Items
> =========================
>
> Current version at http://candle.pha.pa.us/cgi-bin/pgopenitems.
>
> Changes
> -------
> integrated auto-vacuum (Alvaro)
> ICU locale patch?
> Win32 signal handling patch (Magnus)
> KRB4 removal patch (Magnus)
> GIST concurrency fixes? (Oleg)
> Roles? (Stephen Frost)
> convert plperl arrays to pg arrays? (Andrew Dunstan)
> plperl validator function? (Andrew Dunstan)
> column-level triggers (Greg)
> interval improvements (Michael Glaesemann)
> move rtree_gist into core?
> config file I/O?
> terminate backend fix?
> dbsize functions from /contrib?
> fix pg_autovacuum O(n^2) behavior
> remove wal siblings guc vars?
> COPY performance improvements

I've just submitted my first working version of Constraint Exclusion
(basic Partitioning) for 8.1

I also hope to complete my earlier submitted patch on WAL log file
switching also.

Best Regards, Simon Riggs

Tom Lane wrote:
> Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us> writes:
> > Here are the open items. Feature freeze is Friday, July 1.
>
> This is missing the point that pg_dump output is not forward compatible
> with the planned escape string changeover.
>
> We had talked about tweaking ruleutils.c, as well as pg_dump itself,
> to use E'...' syntax if the string value contains a backslash ---
> but the patch so far applied does none of that.

Yep, not done, but they are on the escape web page:

http://candle.pha.pa.us/cgi-bin/pgescape

I added it to the open items list.

--
Bruce Momjian | http://candle.pha.pa.us
pgman(at)candle(dot)pha(dot)pa(dot)us | (610) 359-1001
+ If your life is a hard drive, | 13 Roberts Road
+ Christ can be your backup. | Newtown Square, Pennsylvania 19073

* Alvaro Herrera (alvherre(at)surnet(dot)cl) wrote:
> Additionally, my shared dependencies patch :-) I'm not sure if I'll be
> able to do both autovacuum and shared dependencies before freeze, but
> I'll try. I'm not sure if anybody is working on roles though; if I have
> to do the three of them I'm pretty sure I won't be able to meet the
> deadline.

I've been working alot on roles over the past 5 days or so (was in Ohio,
so wasn't really checking email, but I spent 5:30am to about 10:30am
both Thursday and Friday morning, and I'm working on it right now).

> I think we should leave shared dependencies out if roles do not make it
> into 8.1.

I'd really like to see roles in 8.1 and I'm trying hard to finish it
tonight. I've finished basically all of the codeing for the parts I had
on my 'milestones' along w/ some additional things. I'm finishing up
the compile-time issues for GRANT ROLE/REVOKE ROLE support and will be
doing more testing tonight trying to find things which are still broken.

One issue that should probably be fixed before 8.1 is that in the parser
I had to write the syntax like:

GRANT ROLE <roles_list> TO <roles_list>, etc

When I tried to add it like:

GRANT <roles_list> TO <roles_list>, etc

I got 9 reduce/reduce conflicts w/ the regular GRANT. I believe this
can be fixed but I'm compilers class was a long time ago. Really a bug
fix imv, but I decided to move on from it and get the code written for
it and hope someone w/ more parser familiarity can fix it. Same
happened with REVOKE (and generated 9 more reduce/reduce conflicts).

Thanks,

Stephen

* Tom Lane (tgl(at)sss(dot)pgh(dot)pa(dot)us) wrote:
> Alvaro Herrera <alvherre(at)surnet(dot)cl> writes:
> >> Roles? (Stephen Frost)
>
> > Additionally, my shared dependencies patch :-) I'm not sure if I'll be
> > able to do both autovacuum and shared dependencies before freeze, but
> > I'll try. I'm not sure if anybody is working on roles though;
>
> I'll try to finish the roles patch if Stephen sends in what he's got
> within the next day or so.

Right, great, I'm definitely going to try and send something in tonight
with hopefully most of the codeing done. Of course, I'm sure you'll
find things to clean up and fix but I'm really trying to get it in good
shape.

Thanks,

Stephen

Hi Tom,

> I think that moving rtree_gist, reindexdb, and/or userlock into core
> would have to happen before feature freeze,
[snip]

Are you think in putting reindex at core? I was about to submit a
replacement of it but if it goes to bin/scripts (for example) I can
rearrange the patch. Could I?

Euler Taveira de Oliveira
euler[at]yahoo_com_br

__________________________________________________
Converse com seus amigos em tempo real com o Yahoo! Messenger
http://br.download.yahoo.com/messenger/

Euler Taveira de Oliveira <eulerto(at)yahoo(dot)com(dot)br> writes:
>> I think that moving rtree_gist, reindexdb, and/or userlock into core
>> would have to happen before feature freeze,
> [snip]

> Are you think in putting reindex at core? I was about to submit a
> replacement of it but if it goes to bin/scripts (for example) I can
> rearrange the patch. Could I?

The redefinition of REINDEX DATABASE is already in. A direct
replacement for reindexdb would basically need to apply this across
all DBs in the cluster, which is definitely a job for bin/scripts.
I was speculating that maybe it'd make more sense to handle it as
an option to vacuumdb or clusterdb instead of a whole new program,
but it's up to you if you're doing the work.

regards, tom lane