Re: 9.5 CF1

Hi.

There are 92 outstanding patches in this CommitFest, and 63 of them do
not have any reviewer. Those are very large numbers, so I hope everyone
will pitch in to keep things moving along.

There's quite a variety of patches available for review this time, and
any level of feedback about them is useful, from "no longer applies to
HEAD" or "doesn't build" to more detailed reviews.

If you don't have the time to do a full review, or are getting bogged
down, post whatever you do have (the same amount of fame and fortune
will still be yours!).

If you're wondering where to start, here are some suggestions, picked
almost at random:

Using Levenshtein distance to HINT a candidate column name
http://archives.postgresql.org/message-id/CAM3SWZS9-Xr2Ud_j9yrKDctT6xxy16h1EugtSWmLU6Or4CtGAA@mail.gmail.com

Better partial index-only scans
http://archives.postgresql.org/message-id/CABz-M-GrkvrMc9ni5S0mX53rtZg3=SZNEYrU_A8RigQ2b3MGNA@mail.gmail.com

Use unique index for longer pathkeys
http://archives.postgresql.org/message-id/20140613.164133.160845727.horiguchi.kyotaro@lab.ntt.co.jp

SQL access to database attributes
http://archives.postgresql.org/message-id/53868E57.3030908@dalibo.com

pg_resetxlog option to change system identifier
http://archives.postgresql.org/message-id/539B97FC.8040806@2ndquadrant.com

pg_xlogdump --stats
http://archives.postgresql.org/message-id/20140604104716.GA3989@toroid.org

tab completion for set search_path TO
http://archives.postgresql.org/message-id/CAMkU=1xJzK0h7=0_sOLLKGaf7zSwp_YzcKwuG41Ns+_Qcn+t=g@mail.gmail.com

idle_in_transaction_timeout
http://archives.postgresql.org/message-id/538E600E.1020708@dalibo.com

I'll post a periodic summary to the list, and will send out reminders by
private mail as usual.

Please feel free to contact me with questions.

-- Abhijit

P.S. If you tag your reviews with [REVIEW] in the Subject, it'll be
easier to keep track of them.

On Mon, Jun 16, 2014 at 2:36 AM, Abhijit Menon-Sen <ams(at)2ndquadrant(dot)com> wrote:
> P.S. If you tag your reviews with [REVIEW] in the Subject, it'll be
> easier to keep track of them.

I and, I believe, various other people hate that style, because at
least in Gmail, it breaks the threading. It is much easier to find
things if they are all posted on one thread.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

Robert Haas wrote:
> On Mon, Jun 16, 2014 at 2:36 AM, Abhijit Menon-Sen <ams(at)2ndquadrant(dot)com> wrote:
> > P.S. If you tag your reviews with [REVIEW] in the Subject, it'll be
> > easier to keep track of them.
>
> I and, I believe, various other people hate that style, because at
> least in Gmail, it breaks the threading. It is much easier to find
> things if they are all posted on one thread.

Yes, please don't do that. A simple, normal reply to the message that
submits the patch is much better from my point of view as a subsequent
reviewer and committer.

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

Alvaro Herrera <alvherre(at)2ndquadrant(dot)com> writes:
> Robert Haas wrote:
>> On Mon, Jun 16, 2014 at 2:36 AM, Abhijit Menon-Sen <ams(at)2ndquadrant(dot)com> wrote:
>>> P.S. If you tag your reviews with [REVIEW] in the Subject, it'll be
>>> easier to keep track of them.

>> I and, I believe, various other people hate that style, because at
>> least in Gmail, it breaks the threading. It is much easier to find
>> things if they are all posted on one thread.

> Yes, please don't do that. A simple, normal reply to the message that
> submits the patch is much better from my point of view as a subsequent
> reviewer and committer.

Worth noting also is that Magnus is working on a new version of the
commitfest app that will be able to automatically keep track of threads
about patches --- so long as they *are* threads according to our mailing
list archives. I'm not sure if the archives recognize replies with a
changed Subject: as being the same thread or not.

regards, tom lane

On 2014-06-17 12:47:19 -0400, Alvaro Herrera wrote:
> Robert Haas wrote:
> > On Mon, Jun 16, 2014 at 2:36 AM, Abhijit Menon-Sen <ams(at)2ndquadrant(dot)com> wrote:
> > > P.S. If you tag your reviews with [REVIEW] in the Subject, it'll be
> > > easier to keep track of them.
> >
> > I and, I believe, various other people hate that style, because at
> > least in Gmail, it breaks the threading. It is much easier to find
> > things if they are all posted on one thread.
>
> Yes, please don't do that. A simple, normal reply to the message that
> submits the patch is much better from my point of view as a subsequent
> reviewer and committer.

Very much agreed. Especially as many patches have several reviews over
the course of their integration. I think the separate thread suggestion
was made by some former CF manager?

I think it's sometimes, for larger/hotly debated patches, useful to
start anew at significant new version though... 300 message deep threads
get unwiedly.

Greetings,

Andres Freund

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

At 2014-06-17 12:47:19 -0400, alvherre(at)2ndquadrant(dot)com wrote:
>
> > > P.S. If you tag your reviews with [REVIEW] in the Subject, it'll
> > > be easier to keep track of them.
> >
> > I and, I believe, various other people hate that style, because at
> > least in Gmail, it breaks the threading. It is much easier to find
> > things if they are all posted on one thread.
>
> Yes, please don't do that. A simple, normal reply to the message that
> submits the patch is much better from my point of view as a subsequent
> reviewer and committer.

I find it hard to believe that gmail is incapable of threading messages
using In-Reply-To/References header fields, especially given that mail
subjects are changed all the time in the normal course of events. But
I'll take your word for it and reply to the original message (as I've
always done) without changing the Subject for any reviews I post.

Only one other person took my suggestion so far, but if anyone else is
tempted to do the same: consider the recommendation withdrawn.

-- Abhijit

P.S. I also realise now that some people may not {,be able to} change
the Subject of a message without creating an entirely new thread.

Abhijit Menon-Sen wrote:

> I find it hard to believe that gmail is incapable of threading messages
> using In-Reply-To/References header fields, especially given that mail
> subjects are changed all the time in the normal course of events. But
> I'll take your word for it and reply to the original message (as I've
> always done) without changing the Subject for any reviews I post.

They do it -- I experimented with it awhile ago. Gmail threads emails
perfectly well, but it seems they break threads on purpose when the
subject changes. Apparently, more people is used to starting a new
thread by replying to an existing email and changing to a new subject,
than people "tweak" subjects by adding tags and such.

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

On Tue, Jun 17, 2014 at 6:54 PM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:

> Alvaro Herrera <alvherre(at)2ndquadrant(dot)com> writes:
> > Robert Haas wrote:
> >> On Mon, Jun 16, 2014 at 2:36 AM, Abhijit Menon-Sen <ams(at)2ndquadrant(dot)com>
> wrote:
> >>> P.S. If you tag your reviews with [REVIEW] in the Subject, it'll be
> >>> easier to keep track of them.
>
> >> I and, I believe, various other people hate that style, because at
> >> least in Gmail, it breaks the threading. It is much easier to find
> >> things if they are all posted on one thread.
>
> > Yes, please don't do that. A simple, normal reply to the message that
> > submits the patch is much better from my point of view as a subsequent
> > reviewer and committer.
>
> Worth noting also is that Magnus is working on a new version of the
> commitfest app that will be able to automatically keep track of threads
> about patches --- so long as they *are* threads according to our mailing
> list archives. I'm not sure if the archives recognize replies with a
> changed Subject: as being the same thread or not.
>

The archives code does threading based on the headers (in-reply-to and
references, in priority order). It completely ignores the subject when it
comes to the threading.

--
Magnus Hagander
Me: http://www.hagander.net/
Work: http://www.redpill-linpro.com/

Hi.

One week into the CommitFest, we now have nine committed patches, ten
ready for committer, fourteen waiting on their author, and fifty-nine
still awaiting review.

Thanks to all the people who submitted a review (and a special mention
for MauMau for reviewing the most patches so far, of which one has been
committed and three others are ready for committer).

There are still twenty-one patches that do not have a reviewer. Here's
a selection:

KNN-GiST with recheck
http://archives.postgresql.org/message-id/CAPpHfdu_qBLNRnv-r=_tofZYYa6-r=Z5_MGF4_phAOkWcYxfRg@mail.gmail.com

XLogLockBlockRangeForCleanup
http://archives.postgresql.org/message-id/20140613084012.GA16567@toroid.org

UPDATE/DELETE .. ORDER BY .. LIMIT ..
http://archives.postgresql.org/message-id/1394662740.69204.YahooMailNeo@web163002.mail.bf1.yahoo.com

event triggers: more DROP info
http://archives.postgresql.org/message-id/20140613195049.GQ18688@eldon.alvh.no-ip.org

WAL format & API changes
http://archives.postgresql.org/message-id/539CBA75.3050802@vmware.com

Please feel free to contact me with questions.

-- Abhijit

Hi.

We're two weeks into the CommitFest now, and things are moving along
quite nicely.

Fourteen patches have been committed, and twelve more are marked ready
for committer. But most importantly, many patches have been reviewed,
and only nine patches still lack a reviewer (and most of those have
seen some discussion already).

I have sent a number of reminders and status inquiries by private mail,
and will continue to do so for "waiting on author" and "needs review"
patches. (Living as I do at the edge of a forest, I have also curated
a fine selection of sharp sticks to poke people with if they don't
respond during the week.)

Here's a list of the patches that have no listed reviewers:

Buffer capture facility: check WAL replay consistency
http://archives.postgresql.org/message-id/CAB7nPqTFK4=fcrto=Lji4VLBX9AH+FV1Z1Ke6r98PpXuUXWeNA@mail.gmail.com

Generic atomics implementation
http://archives.postgresql.org/message-id/20140625171434.GG24114@awork2.anarazel.de

KNN-GiST with recheck
http://archives.postgresql.org/message-id/CAPpHfdu_qBLNRnv-r=_tofZYYa6-r=Z5_MGF4_phAOkWcYxfRg@mail.gmail.com

Sort support for text with strxfrm() poor man's keys
http://archives.postgresql.org/message-id/CAM3SWZQKwELa58h1R9sxwAOCJpgs=xJbiu7apDyq9pUuSfQX6g@mail.gmail.com

Use unique index for longer pathkeys
http://archives.postgresql.org/message-id/20140613.164133.160845727.horiguchi.kyotaro@lab.ntt.co.jp

CSN snapshots
http://archives.postgresql.org/message-id/539AD153.9000004@vmware.com

Postgres Hibernator contrib module
http://archives.postgresql.org/message-id/CABwTF4XGbpgwmkKt6ezSFVnpF11KmAg5m969twkGoDhnPj-QLA@mail.gmail.com

possibility to set "double" style for unicode lines
http://archives.postgresql.org/message-id/CAFj8pRCzNo6kP3vyuZAJkakpcfb_aBRhSRaLcgjUjmpXno9M3A@mail.gmail.com

Refactor SSL code to support other SSL implementations
http://archives.postgresql.org/message-id/53986CF4.1030403@vmware.com

Does anyone want to pick up one of these?

Thanks to everyone for their participation. It's especially nice to see
several reviews posted, and the authors responding quickly with updated
versions of their patch to address the review comments.

As always, please feel free to contact me with questions.

-- Abhijit

Hi.

Here's a detailed mid-week update split up by category. I've left out
patches marked returned/rejected, committed, or ready for committer.

Server features
---------------

Lag & Lead Window Functions Can Ignore Nulls
Latest patch currently pending review by Jeff and Álvaro. No
updates so far.

Foreign table inheritance
Moved from ready for committer back to waiting on author after
Noah's review comments. Should we expect an updated patch to be
posted in this CF?

http://archives.postgresql.org/message-id/20140702022327.GC1586927@tornado.leadboat.com

Using Levenshtein distance to HINT a candidate column name
No real progress in any direction. Everyone wants the feature, but
nobody is sure how many suggestions would be useful and how many
would be annoying.

Buffer capture facility: check WAL replay consistency
Some discussion, but no reviewers. Status unclear.

http://archives.postgresql.org/message-id/CAB7nPqTFK4=fcrto=Lji4VLBX9AH+FV1Z1Ke6r98PpXuUXWeNA@mail.gmail.com

Custom Plan API
Shigeru Hanada has said he plans to post a design review soon.

delta relations in AFTER triggers
No code review yet, but there's a proof-of-concept extension that
uses the feature, and Kevin is working on PL changes to make the
feature usable, and hopes to commit in this cycle.

Minmax indexes
Álvaro will respond to the design questions Heikki raised.

-- Abhijit

Miscellaneous
-------------

contrib/fastbloat - tool for quickly assessing bloat stats for a table
Pending review by Jaime.

showing index update time on EXPLAIN
Pending updated patch by Jaime.

idle_in_transaction_session_timeout
Marked as ready for committer, but as far as I can tell the patch
needs revisions per Tom's comments. It's unclear if Vik is working
on these, or Kevin, or nobody. Anyone?

logging of replication commands
Two patches posted: one generally accepted, the other contentious.
Pending status update.

Refactor SSL code to support other SSL implementations
Some discussion, but no code reviews yet. Jeff Janes tried it and it
hung; Andreas Karlsson fixed it and it worked for him. Everyone does
seem to agree that it's a good idea.

There is also one other "ready for committer" patch in this category.

Bug fixes
---------

Correctly place DLLs for ECPG apps in bin folder
Pending review by Muhammad Asif Naeem.

Per table autovacuum vacuum cost parameters behavior change
Pending review by Álvaro.

Ignore Ctrl-C/Break on Windows
Christian Ullrich has said he will post an updated patch this week.

There are also three "ready for committer" patches in this category.

-- Abhijit

On Tue, Jul 1, 2014 at 10:06 PM, Abhijit Menon-Sen <ams(at)2ndquadrant(dot)com> wrote:
> Using Levenshtein distance to HINT a candidate column name
> No real progress in any direction. Everyone wants the feature, but
> nobody is sure how many suggestions would be useful and how many
> would be annoying.

Expect a revised patch soon.

--
Peter Geoghegan

Functions
---------

min/max support for inet datatypes
Pending review by Muhammad Asif Naeem.

Selectivity estimation for inet operators
Dilip Kumar plans to post a review this week.

There are two "ready for committer" patches in this category.

Clients
-------

Gaussian distribution pgbench
Fabien Coelho plans to post a review of the latest patch.

possibility to set "double" style for unicode lines
No reviews, no reviewers. \N{DOUBLE LINED UNHAPPY FACE}[1]

Allow parallel cores to be used by vacuumdb
Status not entirely clear. Do we want this? Does the patch have
too much duplicated code? Can it be fixed in this cycle?

add line number as prompt option to psql
Sawada Masahiko plans to post an updated patch soon.

There are three "ready for committer" patches in this category.

-- Abhijit

1. Not a real Unicode character. Sorry.

System administration
---------------------

pg_hibernator
Nice feature, some discussion, no code reviews. Status not entirely
clear, but updated patch available.

Monitoring & control
--------------------

Reducing impact of hints/cleanup for SELECTs
Pending performance review by Emanuel Calvo. (He said he's posted a
review, but I couldn't find it. He's travelling right now, will send
me a pointer later.)

Send transaction commit/rollback stats to the stats collector
unconditionally
Currently marked waiting on author, but I think Kevin is in the
process of getting it committed.

issue log message to suggest VACUUM FULL if a table is nearly empty
Euler has said he plans to post a review this week.

pg_shmem_allocations view
Euler has said he plans to post a review this week.

pg_xlogdump --stats
Pending review by Dilip Kumar (who plans to post a review this
week), but Osamu Furuya and Marti Raudsepp have said it looks
OK generally.

There is one "ready for committer" patch in this category.

-- Abhijit

Security
--------

Row-security based on Updatable security barrier views
Lots of discussion that I haven't dared to look at properly yet. I
gather there's still plenty of design-level work needed, and this
is not in any imminent danger of being committed.

Replication & Recovery
----------------------

pg_receivexlog add synchronous mode
Patch works, is generally acceptable. Fujii-san proposed a
refactoring patch to be applied before this one, and plans
to commit it soon.

Compression of Full Page Writes
Currently under development, timeline unclear. Probably needs to be
marked returned with feedback and moved to August CF.

WAL format & API changes
I'm not sure what's happening here. Will look more closely, but
updates are welcome from the people who've been participating in
the discussion/review.

There are three "ready for committer" patches in this category.

-- Abhijit

SQL commands
------------

Event triggers: object creation support
Enormous patch, no reviews, no reviewers, but it's known to work
already. Does anyone want to have a look at this? (I thought it
was being reviewed, and didn't realise otherwise until a couple
of days ago. Sorry about that.)

change alter user to be a true alias for alter role
Original patch worked (Vik) but didn't go far enough towards
preventing future repetitions of the mistake (Tom) but the suggested
change didn't turn out to be easy (Vik). I don't think anyone is
working on this at the moment.

IMPORT FOREIGN SCHEMA statement
Lots of discussion and various revisions; currently marked waiting
on author. Updates welcome.

"RETURNING PRIMARY KEY" syntax extension
Patch works and has been revised as suggested. Some questions about
how useful it is. Updates welcome.

Allow an unlogged table to be changed to logged GSoC 2014
Stephen has looked at the patch a little, but isn't sure how much
time he'll have to complete the review. More eyes are welcome. I
recommend this patch to anyone who's looking to get started with
reviewing something substantial: it looks very nicely-done and
straightforward, but still non-trivial.

-- Abhijit

Performance
-----------

scalable LWLocks
Generic atomics implementation
Both under active discussion.

KNN-GiST with recheck
Partial sort
Some earlier discussion, but no conclusions, and as far as I know
nobody is working on these patches at the moment.

lowering array_agg memory requirements
I'm waiting to hear from Tomas if he expects to post an updated
version of the patch soon. Otherwise I'll move it to 2014-08.

Don't require a NBuffer sized PrivateRefCount array of local buffer pins
Pending performance tests before commit. Is anyone working on this?

Allow more join types to be removed
Updated patch pending review from Simon. Updates welcome.

Sort support for text with strxfrm() poor man's keys
Lots of earlier discussion, but no conclusions as far as I can tell.
Does anyone want to take a look at this?

Scaling shared buffer eviction
Pending review by Andres. Any updates?

Spread shared memory across NUMA memory nodes
Marked waiting on author, but status unclear. Any updates?

Use unique index for longer pathkeys
No reviews, no reviewers. I took a quick look, and it's not clear if
this is useful in itself, or if it just enables more interesting
optimisations later. Does anyone want to look at this?

XLogLockBlockRangeForCleanup
Amit Khandekar plans to post a review this week.

SKIP LOCKED
Two patches available: original by Simon, updated/merged version by
Thomas Munro. Simon said he'll look into both with a view towards
committing the functionality. Timeline not yet clear.

Allow NOT IN to use anti joins
Revised patch marked pending review. Updates welcome.

CSN snapshots
Work in progress, plenty of discussion, no reviews. Probably nothing
to do here. (Heikki, are you waiting for feedback or anything?)

There's not even one "ready for committer" patch in this category.

-- Abhijit

On Wed, Jul 2, 2014 at 2:44 PM, Abhijit Menon-Sen <ams(at)2ndquadrant(dot)com>
wrote:

> WAL format & API changes
> I'm not sure what's happening here. Will look more closely, but
> updates are welcome from the people who've been participating in
> the discussion/review.
>
Patch has been reviewed once. Heikki has submitted an updated that still
has some issues, nothing unbearable though IMO. I imagine that there are
still some discussions needed for the potention performance drop that this
patch could create because of the individual memcpy calls that are
performed for each WAL record registration.
Regards,
--
Michael

(2014/07/02 14:06), Abhijit Menon-Sen wrote:

> Foreign table inheritance
> Moved from ready for committer back to waiting on author after
> Noah's review comments. Should we expect an updated patch to be
> posted in this CF?
>
> http://archives.postgresql.org/message-id/20140702022327.GC1586927@tornado.leadboat.com

Noah, thank you for the review!

In addition to the issues pointed out by Noah, the following should be
resolved, I think. And ISTM we should work on the latter first, for
which I plan to add a separate patch to the next CF. So, I think it
would be better to move this to the next CF.

http://www.postgresql.org/message-id/3492.1404136084@sss.pgh.pa.us

Thanks,

Best regards,
Etsuro Fujita

Thanks for the update. I have marked the patch "returned with feedback"
and moved it to the 2014-08 CF.

-- Abhijit

(2014/07/02 18:19), Abhijit Menon-Sen wrote:
> Thanks for the update. I have marked the patch "returned with feedback"
> and moved it to the 2014-08 CF.

OK I've changed the status from "returned with feedback" to "Waiting on
Author".

Thanks,

Best regards,
Etsuro Fujita

On Wed, Jul 2, 2014 at 10:08 AM, Abhijit Menon-Sen <ams(at)2ndquadrant(dot)com>
wrote:

> KNN-GiST with recheck
> Partial sort
> Some earlier discussion, but no conclusions, and as far as I know
> nobody is working on these patches at the moment.
>

I'm now working on next revision of KNN-GiST with recheck. Also, I think
partial sort need a look of somebody more aware of planner than me and
Marti.

------
With best regards,
Alexander Korotkov.

Abhijit Menon-Sen <ams(at)2ndQuadrant(dot)com> wrote:

> Spread shared memory across NUMA memory nodes
>     Marked waiting on author, but status unclear. Any updates?

The review was a little sparse:

http://www.postgresql.org/message-id/CADyhKSXs+oUetngSbeiM0tVSRy=QeCaSNBQBDbM=SFQTDg+Zog@mail.gmail.com

In particular, there was no benchmarking of any sort in the review.
I did some benchmarking on a 16 core Power PC machine with 4 NUMA
memory nodes, and found that the benefit was overall about 2% in
the unbalanced memory node tests I was able to devise.  Unless
memory usage was unbalanced I saw no overall difference.  Timings
with the patch varied less from run to run than without the patch.
Where I saw worse performance in the field (prompting me to work on
this patch) was on an Intel 40 core 4 memory node system.  Perhaps
that is necessary to see the really bad behavior.

I am not sure that the modest worst-case benefits I have seen
justify applying the patch, especially since to see that I also
needed to set up a custom cpuset to run PostgreSQL under so that
the OS also used interleaving for its cache.  I did have one
outlier for the unpatched code which was about 20% worse than
average, and no such outliers for patched code.  I have seen
reports of much worse performance hits from an unbalanced load than
I was able to reliably produce, so I was hoping that someone with
more creativity in creating worst case tests would see what they
could do.  IMV, the best argument for the patch is as insurance
against such pessimal events.

The suggestion that there be a GUC to suppress the interleaving of
the main shared memory segment among NUMA memory nodes is
interesting.  I'm dubious, but I guess it would allow an opt-out
escape hatch if someone found unexpected performance problems with
it in the field.  It's certainly trivial to add; the main argument
against it is that it is another knob that might confuse users.

Anyway, Waiting on Author is probably not the best state for this
patch; there's nothing that I feel there is a consensus to change.
I'm OK with either Needs Review or Returned with Feedback. If
someone can produce a benchmark showing more benefit than I was
able to show, it can be revived in a later CF.  Perhaps as more
machines with high core counts and multiple NUMA memory nodes
become more common the cases where we run into trouble will become
more clear.

--
Kevin Grittner
EDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

* Abhijit Menon-Sen (ams(at)2ndQuadrant(dot)com) wrote:
> Row-security based on Updatable security barrier views
> Lots of discussion that I haven't dared to look at properly yet. I
> gather there's still plenty of design-level work needed, and this
> is not in any imminent danger of being committed.

I feel like we're actually getting pretty close on the design side..
That said, the design we're arriving at will require a bit more work to
implement that seems unlikely to be completed in the next two weeks,
though I could be wrong. I'd be fine leaving this open for now at least
('waiting on author' perhaps) and then bumping it to August only if
necessary. I'm still very interested in getting this committed early in
this cycle to allow for a good bit of testing, of course.

Thanks,

Stephen

Abhijit Menon-Sen wrote:

> Minmax indexes
> Álvaro will respond to the design questions Heikki raised.

I'm currently reworking the patch so that things that need to be
abstract per discussion are abstract, without enlarging the scope
excessively. I'm working on an opclass implementation that can be used
for "sortable" datatypes, so that it covers the datatypes it covers
today (basically anything with a btree opclass). I don't intend to
implement opclasses for any other type, but with the design I expect to
have it should be possible to cover stuff like geometric types usefully.
(I guess it will be possible to index arrays in a general way also, but
I don't think it will be of any usefulness).

Once I post that version, I will focus on reviewing other patches, and
then maybe someone interested can try to implement other opclasses on
top of it, to ensure that the API I propose is sensible or at least
useful.

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

At 2014-07-02 11:19:10 -0400, sfrost(at)snowman(dot)net wrote:
>
> I'd be fine leaving this open for now at least ('waiting on author'
> perhaps) and then bumping it to August only if necessary.

Since there's active discussion/development happening, I wasn't planning
to change the status anyway. It looks fine as-is.

-- Abhijit

Hi.

We're into the last full week of this CommitFest. Here's an update on
the current state of the queue.

Needs review: 31
Waiting on author: 18
Ready for committer: 13
Committed: 19

(I'll send separate topic-wise updates too.)

Reviewers, please try to post whatever you have in the next few days to
give the authors a chance to respond. Even if you haven't completed the
review, please consider posting what you have now and updating it later.
If for any reason you don't expect to be able to complete a review you
are working on, please let me know as soon as possible.

Patch authors, please let me know if you intend to resubmit revisions of
patches that are currently marked "waiting on author" during this CF. If
it's likely to take more than a few days, it should be moved to the
August CF.

Committers… well, have fun looking at the queue, I guess!

Thanks again to everyone for their help in keeping things moving during
this review cycle. Starting at 97 patches, this wasn't the largest CF
we've ever had, but it's right up there near the top.

-- Abhijit

At 2014-07-02 10:36:23 +0530, ams(at)2ndQuadrant(dot)com wrote:
>
> Server features
> ---------------
>
> Lag & Lead Window Functions Can Ignore Nulls
> Latest patch currently pending review by Jeff and Álvaro. No
> updates so far.

Jeff has posted a review, and is working on the patch further.

> Using Levenshtein distance to HINT a candidate column name
> No real progress in any direction. Everyone wants the feature, but
> nobody is sure how many suggestions would be useful and how many
> would be annoying.

Peter posted a new patch. Nobody has reviewed it yet.

> Buffer capture facility: check WAL replay consistency
> Some discussion, but no reviewers. Status unclear.
>
> http://archives.postgresql.org/message-id/CAB7nPqTFK4=fcrto=Lji4VLBX9AH+FV1Z1Ke6r98PpXuUXWeNA@mail.gmail.com

No updates.

> Custom Plan API
> Shigeru Hanada has said he plans to post a design review soon.

Any updates? Should this be moved to the next CF?

> delta relations in AFTER triggers
> No code review yet, but there's a proof-of-concept extension that
> uses the feature, and Kevin is working on PL changes to make the
> feature usable, and hopes to commit in this cycle.

Ditto.

> Minmax indexes
> Álvaro will respond to the design questions Heikki raised.

Álvaro is still working on this and hopes to post an update soon.

-- Abhijit

At 2014-07-02 11:38:31 +0530, ams(at)2ndQuadrant(dot)com wrote:
>
> Performance
> -----------
>
> scalable LWLocks
> Generic atomics implementation
> Both under active discussion.

Work continues on these.

> KNN-GiST with recheck
> Partial sort
> Some earlier discussion, but no conclusions, and as far as I know
> nobody is working on these patches at the moment.

Heikki said he won't be able to look at these, so unless someone else
plans to review them, they should be moved to the next CF. Alexander is
working on an updated version of the KNN-GiST patch anyway.

> Don't require a NBuffer sized PrivateRefCount array of local buffer pins
> Pending performance tests before commit. Is anyone working on this?

No updates.

> Allow more join types to be removed
> Updated patch pending review from Simon. Updates welcome.

Tom is looking at this now.

> Sort support for text with strxfrm() poor man's keys
> Lots of earlier discussion, but no conclusions as far as I can tell.
> Does anyone want to take a look at this?
>
> Scaling shared buffer eviction
> Pending review by Andres. Any updates?
>
> Use unique index for longer pathkeys
> No reviews, no reviewers. I took a quick look, and it's not clear if
> this is useful in itself, or if it just enables more interesting
> optimisations later. Does anyone want to look at this?

No updates for any of these.

> XLogLockBlockRangeForCleanup
> Amit Khandekar plans to post a review this week.

Pending performance results from me, but also a final conclusion from
Amit's review.

> SKIP LOCKED
> Two patches available: original by Simon, updated/merged version by
> Thomas Munro. Simon said he'll look into both with a view towards
> committing the functionality. Timeline not yet clear.
>
> Allow NOT IN to use anti joins
> Revised patch marked pending review. Updates welcome.
>
> CSN snapshots
> Work in progress, plenty of discussion, no reviews. Probably nothing
> to do here. (Heikki, are you waiting for feedback or anything?)

No updates for any of these.

-- Abhijit

>> Custom Plan API
>> Shigeru Hanada has said he plans to post a design review soon.
>
> Any updates? Should this be moved to the next CF?
>
Now I'm working to revise the patch according to his suggestion;
will be completed within a couple of days.
A few issues needs design-level suggestion from committers.

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

At 2014-07-02 11:21:53 +0530, ams(at)2ndQuadrant(dot)com wrote:
>
> SQL commands
> ------------
>
> Event triggers: object creation support
> Enormous patch, no reviews, no reviewers, but it's known to work
> already. Does anyone want to have a look at this? (I thought it
> was being reviewed, and didn't realise otherwise until a couple
> of days ago. Sorry about that.)

No updates. I'll look at the patch myself, but I don't think I'll be
able to post any useful review comments until the beginning of next
week.

> change alter user to be a true alias for alter role
> Original patch worked (Vik) but didn't go far enough towards
> preventing future repetitions of the mistake (Tom) but the suggested
> change didn't turn out to be easy (Vik). I don't think anyone is
> working on this at the moment.

No updates.

> "RETURNING PRIMARY KEY" syntax extension
> Patch works and has been revised as suggested. Some questions about
> how useful it is. Updates welcome.

Ian will respond to comments soon.

> Allow an unlogged table to be changed to logged GSoC 2014
> Stephen has looked at the patch a little, but isn't sure how much
> time he'll have to complete the review. More eyes are welcome. I
> recommend this patch to anyone who's looking to get started with
> reviewing something substantial: it looks very nicely-done and
> straightforward, but still non-trivial.

Christoph Berg is looking at this, but doesn't expect to be able to post
a review for a few days yet. (The recommendation still stands.)

-- Abhijit

At 2014-07-02 11:14:52 +0530, ams(at)2ndQuadrant(dot)com wrote:
>
> Security
> --------
>
> Row-security based on Updatable security barrier views
> Lots of discussion that I haven't dared to look at properly yet. I
> gather there's still plenty of design-level work needed, and this
> is not in any imminent danger of being committed.

Work continues.

> Replication & Recovery
> ----------------------
>
> pg_receivexlog add synchronous mode
> Patch works, is generally acceptable. Fujii-san proposed a
> refactoring patch to be applied before this one, and plans
> to commit it soon.

Refactoring patch committed, other patch still pending.

> Compression of Full Page Writes
> Currently under development, timeline unclear. Probably needs to be
> marked returned with feedback and moved to August CF.

Revised patch posted, but lots of work still needed.

> WAL format & API changes
> I'm not sure what's happening here. Will look more closely, but
> updates are welcome from the people who've been participating in
> the discussion/review.

Waiting on Heikki.

-- Abhijit

At 2014-07-02 11:09:19 +0530, ams(at)2ndQuadrant(dot)com wrote:
>
> System administration
> ---------------------
>
> pg_hibernator
> Nice feature, some discussion, no code reviews. Status not entirely
> clear, but updated patch available.

Now being reviewed by MauMau.

> Monitoring & control
> --------------------
>
> Reducing impact of hints/cleanup for SELECTs
> Pending performance review by Emanuel Calvo. (He said he's posted a
> review, but I couldn't find it. He's travelling right now, will send
> me a pointer later.)

No updates.

> issue log message to suggest VACUUM FULL if a table is nearly empty
> Euler has said he plans to post a review this week.
>
> pg_shmem_allocations view
> Euler has said he plans to post a review this week.

No updates from Euler.

> pg_xlogdump --stats
> Pending review by Dilip Kumar (who plans to post a review this
> week), but Osamu Furuya and Marti Raudsepp have said it looks
> OK generally.

Latest patch posted, some discussion needed on the calling conventions
for the new rm_identify method.

-- Abhijit

At 2014-07-02 10:54:05 +0530, ams(at)2ndQuadrant(dot)com wrote:
>
> Functions
> ---------
>
> min/max support for inet datatypes
> Pending review by Muhammad Asif Naeem.

No updates.

> Selectivity estimation for inet operators
> Dilip Kumar plans to post a review this week.

Reviewed, and updated patch from Emre available.

> Clients
> -------
>
> Gaussian distribution pgbench
> Fabien Coelho plans to post a review of the latest patch.

Some questions raised about whether we want this feature in this form at
all. Status unclear.

> possibility to set "double" style for unicode lines
> No reviews, no reviewers. \N{DOUBLE LINED UNHAPPY FACE}[1]

No updates.

> Allow parallel cores to be used by vacuumdb
> Status not entirely clear. Do we want this? Does the patch have
> too much duplicated code? Can it be fixed in this cycle?

I'm not sure what the status of this patch is.

> add line number as prompt option to psql
> Sawada Masahiko plans to post an updated patch soon.

New patch posted, pending review. Should be ready for committer soon.

-- Abhijit

At 2014-07-02 10:44:23 +0530, ams(at)2ndQuadrant(dot)com wrote:
>
> Miscellaneous
> -------------
>
> contrib/fastbloat - tool for quickly assessing bloat stats for a table
> Pending review by Jaime.
>
> showing index update time on EXPLAIN
> Pending updated patch by Jaime.
>
> idle_in_transaction_session_timeout
> Marked as ready for committer, but as far as I can tell the patch
> needs revisions per Tom's comments. It's unclear if Vik is working
> on these, or Kevin, or nobody. Anyone?
>
> logging of replication commands
> Two patches posted: one generally accepted, the other contentious.
> Pending status update.
>
> Refactor SSL code to support other SSL implementations
> Some discussion, but no code reviews yet. Jeff Janes tried it and it
> hung; Andreas Karlsson fixed it and it worked for him. Everyone does
> seem to agree that it's a good idea.
>
> There is also one other "ready for committer" patch in this category.
>
> Bug fixes
> ---------
>
> Correctly place DLLs for ECPG apps in bin folder
> Pending review by Muhammad Asif Naeem.
>
> Per table autovacuum vacuum cost parameters behavior change
> Pending review by Álvaro.
>
> Ignore Ctrl-C/Break on Windows
> Christian Ullrich has said he will post an updated patch this week.
>
> There are also three "ready for committer" patches in this category.

No updates at all.

-- Abhijit

Hi.

We're now at the "nearly done" stage of the first 9.5 CF.

Twenty-four patches have been committed so far, and five more are ready
for committer (and of those, four are small). Thirty patches are still
marked "needs review", and twenty-one are waiting on author.

I'll go through the patches and follow up or move them to the 2014-08
CF as appropriate, so that this one can be closed out as scheduled.

Many thanks to everyone for their participation in this CF.

-- Abhijit

On Mon, Jul 14, 2014 at 3:21 AM, Abhijit Menon-Sen <ams(at)2ndquadrant(dot)com>
wrote:
>
> Hi.
>
> We're now at the "nearly done" stage of the first 9.5 CF.
>
> Twenty-four patches have been committed so far, and five more are ready
> for committer (and of those, four are small). Thirty patches are still
> marked "needs review", and twenty-one are waiting on author.
>

Actually thirty one maked "needs review" because I already send a new
reviewed patch but I forgot to mark "needs review"... sorry!

Greetings,

--
Fabrízio de Royes Mello
Consultoria/Coaching PostgreSQL
>> Timbira: http://www.timbira.com.br
>> Blog sobre TI: http://fabriziomello.blogspot.com
>> Perfil Linkedin: http://br.linkedin.com/in/fabriziomello
>> Twitter: http://twitter.com/fabriziomello