Partial patch status update, 3/3/13

Craig,

Here's for your summary. I stopped at --echo-hidden, working my way up
from the bottom. Overall, it seems like mostly the patches which are
still in queue are there for good reasons, although there's probably a
few things we could bounce if the authors don't respond in a couple days.

Why isn't this committed already?

- commit_delay advice doc patch
- psql \l to accept patterns

Waiting on Author for more than 2 weeks, should probably be bounced:

- Add code commentary in ProcessInterrupts

Pending Discussion:

- Peter E.'s two pkg-config patches:
-- find libxml2 using pkg-config
-- pkg-config for libpq and ecpg

- Awaiting discussion of whether we support Cygwin anymore:
-- Problematic dependency in plpython Makefile [Cygwin/Windows]

- Awaiting discussion of syntax:
-- PL/PgSQL STRICT
-- Matview WIP Patch

- Awaiting discussion of difficult issues:
-- bugfix: --echo-hidden is not supported by \sf statements

Nobody seems to want to review this:

- transforms

Reminders sent to reviewers:

- Problematic dependency in plpython Makefile [Cygwin/Windows]
- plpgsql_check_function - rebase for 9.3

Reminders set to authors:

- Problematic dependency in plpython Makefile [Cygwin/Windows]
- pg_hba error line reporting
- plpgsql_check_function - rebase for 9.3

--
Josh Berkus
PostgreSQL Experts Inc.
http://pgexperts.com

On 03/04/2013 05:35 AM, Josh Berkus wrote:
> Craig,
>
> Here's for your summary. I stopped at --echo-hidden, working my way up
> from the bottom. Overall, it seems like mostly the patches which are
> still in queue are there for good reasons, although there's probably a
> few things we could bounce if the authors don't respond in a couple days.
Thankyou.

I wonder if we need to be more willing to bounce patches that've petered
out and just re-open them if activity resumes.
> Waiting on Author for more than 2 weeks, should probably be bounced:
>
> - Add code commentary in ProcessInterrupts
Done.
> - Awaiting discussion of whether we support Cygwin anymore:
> -- Problematic dependency in plpython Makefile [Cygwin/Windows]
Per earlier discussion we do support cygwin for as long as it doesn't
create an undue maintenance burden for platforms that matter more and so
long as someone cares to maintain it. Looks like it should get committed.
> - Awaiting discussion of syntax:
> -- PL/PgSQL STRICT
> -- Matview WIP Patch
If they're still in general discussion of syntax issues, that seems like
a sign they're possibly not ready for 9.3 ... but in the end, it's up to
the people involved to determine that, certainly not I.
> - Awaiting discussion of difficult issues:
> -- bugfix: --echo-hidden is not supported by \sf statements
Small and difficult -> punt it, IMO.

--
Craig Ringer http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services

* Craig Ringer (craig(at)2ndquadrant(dot)com) wrote:
> > - Awaiting discussion of syntax:
> > -- PL/PgSQL STRICT

I agree w/ Josh and Craig that this should probably be bounced- mainly
because we need to get 9.3 properly underway and this looks to continue
to be an area of much discussion.

> > -- Matview WIP Patch

This appears to be OBE. :)

> > - Awaiting discussion of difficult issues:
> > -- bugfix: --echo-hidden is not supported by \sf statements
> Small and difficult -> punt it, IMO.

I'd really love to have this properly fixed in 9.3, but I've not been
able to find time to spend on it myself, unfortunately. I continue to
think that we should build on what we've already got in psql wrt
function handling, but others have differing opinions, apparently.

Thanks,

Stephen

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 03/04/2013 11:37 AM, Stephen Frost wrote:
>>> - Awaiting discussion of difficult issues:
>>> -- bugfix: --echo-hidden is not supported by \sf statements
>> Small and difficult -> punt it, IMO.
>
> I'd really love to have this properly fixed in 9.3, but I've not been
> able to find time to spend on it myself, unfortunately. I continue to
> think that we should build on what we've already got in psql wrt
> function handling, but others have differing opinions, apparently.
I have no strong opinion personally, I'm just interested in unblocking
the process a bit at this point.

- --
Craig Ringer http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.13 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/

iQEcBAEBAgAGBQJRNBoyAAoJELBXNkqjr+S2LDMH/0/gIYKd/g8PSbntmO9B71rx
JW/DpyRHp8vB+Hyq0msy5m2KAnHI6rkLDjTQT4laGtLPxo+SwE0tBOeSL41KjNMz
UKlL43/4CsTYJDyQZ1LkewlV2cS6JMikvIxGApVRNMLN1j4tOYr3ytGv6eNPu84o
yl7UqsNArokcvnhFD7Tjz7OyrUNOUhf8chCCIPt3Opbwdp1PxNLBWkDodDkRTiXs
OUGMjN77h/ZPpQImXea9+Hn1bH/UHd3Ya6KEDGw7SHj4IR8gPJJ3OOEwq71GYmxr
37ZVrX3GtmbkeCODVXOPVALXUd6QhivZegamtFBUE8yby71H1PWVtChADay+Cvo=
=hZr0
-----END PGP SIGNATURE-----

I've gone through all the Server Features entries and partway through the performance category. So far:

"Writable Foreign Tables" is waiting for Tom to have a chance to check it out and do some detailed work on it. Anyone else who wants to step in and help might be able to improve its chances, as right now it looks like it'll probably slip.

Stalled waiting for final committer check and apply:

- Identity projection (partitioning)
- Writable foreign tables
(but needs significant work not just apply, test, commit)

Nobody seems to want to review this:

- The SEPostgreSQL patches

I still need to test this on Windows:

- parallel pg_dump

Discussion ongoing, uncertain if it'll make the 9.3 cut:

- Timeout framework extension and lock_timeout
- the rest of the event triggers patch
- checksums

Rejected per thread discussion:

- Add code commentary in ProcessInterrupts

RWF per thread discussion:

- Store additional information in GIN index

Pending update after response to reminders/queries:

- Extension templates
- Adjacent in SP-GiST for range-types
- Performance Improvement by reducing WAL for Update Operation

--
Craig Ringer http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services

2013/3/4 Craig Ringer <craig(at)2ndquadrant(dot)com>:
> - The SEPostgreSQL patches
>
> I still need to test this on Windows:
>
SEPostgreSQL is designed to perform on Linux with SELinux
(security-enhanced Linux) feature enabled, because its design
target is centralization of access control rules between DBMS
and operating system.
Unless Windows support SELinux compatible feature in (far)
future, it does not make sense to run SE-PostgreSQL on windows.

Thanks,
--
KaiGai Kohei <kaigai(at)kaigai(dot)gr(dot)jp>

On 03/04/2013 01:02 PM, Kohei KaiGai wrote:
> 2013/3/4 Craig Ringer <craig(at)2ndquadrant(dot)com>:
>> - The SEPostgreSQL patches
>>
>> I still need to test this on Windows:
>>
> SEPostgreSQL is designed to perform on Linux with SELinux
> (security-enhanced Linux) feature enabled, because its design
> target is centralization of access control rules between DBMS
> and operating system.
> Unless Windows support SELinux compatible feature in (far)
> future, it does not make sense to run SE-PostgreSQL on windows.
Heh, yeah. Thankfully you just misread this; the colon indicates that
this statement applies to the following line (parallel pg_dump).

--
Craig Ringer http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services

For the patch: Patch to compute Max LSN of Data Pages

Reviewer has already given his opinion in mail at below link, that this is good from his side and he has no more further concerns. He has one minor suggestion which I also fixed in later provided patch.
http://www.postgresql.org/message-id/CAEJvTzW555_nHi1g_R3TrvgiG7Cy66UqrqhyaFwJNUspCRX-ag@mail.gmail.com

Can you recommend what should be the way for this patch.

With Regards,
Amit Kapila.