Thread: Commitfest manager for 2022-03

Commitfest manager for 2022-03

From
Julien Rouhaud
Date:
Hi,

The final commitfest for pg15 will start in a few days, and I didn't see any
discussion on it or anyone volunteering to be a CFM.

I thought it would be a good idea to send this reminder now and avoid the same
situation as the last commitfest, to avoid unnecessary pain for the CFM(s).

Is there any volunteer?

For the record there are already 246 active patches registered.



Re: Commitfest manager for 2022-03

From
Greg Stark
Date:
I would like to volunteer.

On Fri, 25 Feb 2022 at 05:31, Julien Rouhaud <rjuju123@gmail.com> wrote:
>
> Hi,
>
> The final commitfest for pg15 will start in a few days, and I didn't see any
> discussion on it or anyone volunteering to be a CFM.
>
> I thought it would be a good idea to send this reminder now and avoid the same
> situation as the last commitfest, to avoid unnecessary pain for the CFM(s).
>
> Is there any volunteer?
>
> For the record there are already 246 active patches registered.
>
>


-- 
greg



Re: Commitfest manager for 2022-03

From
David Steele
Date:
Hi Greg,

On 2/25/22 12:39, Greg Stark wrote:
> I would like to volunteer.
> 
> On Fri, 25 Feb 2022 at 05:31, Julien Rouhaud <rjuju123@gmail.com> wrote:
>>
>> Is there any volunteer?

Since I've done the March CF for the last seven years, I thought it 
would be good if I chimed in. I've been agonizing a bit because I have 
travel planned during March and while I *could* do it I don't think I'd 
be able to do a good job.

I've been hoping somebody would volunteer, so I'm all in favor of you 
being CF.

Regards,
-David



Re: Commitfest manager for 2022-03

From
Michael Paquier
Date:
On Fri, Feb 25, 2022 at 01:58:55PM -0600, David Steele wrote:
> On 2/25/22 12:39, Greg Stark wrote:
>> I would like to volunteer.
>
> Since I've done the March CF for the last seven years, I thought it would be
> good if I chimed in. I've been agonizing a bit because I have travel planned
> during March and while I *could* do it I don't think I'd be able to do a
> good job.

That's a time-consuming task, and there have been little occasions to
enjoy travelling lately, so have a good time :)

> I've been hoping somebody would volunteer, so I'm all in favor of you being
> CF.

Greg as CFM would be fine.  It's nice to see someone volunteer.
--
Michael

Attachment

Re: Commitfest manager for 2022-03

From
Julien Rouhaud
Date:
On Sat, Feb 26, 2022 at 02:42:33PM +0900, Michael Paquier wrote:
> On Fri, Feb 25, 2022 at 01:58:55PM -0600, David Steele wrote:
> > On 2/25/22 12:39, Greg Stark wrote:
> >> I would like to volunteer.
> 
> > I've been hoping somebody would volunteer, so I'm all in favor of you being
> > CF.
> 
> Greg as CFM would be fine.  It's nice to see someone volunteer.

+1, thanks a lot Greg!

Note that the last CF of a major version is probably even more work than
"regular" CF, so if other people are interested there's probably room for
multiple CFMs.



Re: Commitfest manager for 2022-03

From
Greg Stark
Date:
On Sat, 26 Feb 2022 at 01:33, Julien Rouhaud <rjuju123@gmail.com> wrote:
>
> On Sat, Feb 26, 2022 at 02:42:33PM +0900, Michael Paquier wrote:
> > On Fri, Feb 25, 2022 at 01:58:55PM -0600, David Steele wrote:
> > > On 2/25/22 12:39, Greg Stark wrote:
> > >> I would like to volunteer.
> >
> > > I've been hoping somebody would volunteer, so I'm all in favor of you being
> > > CF.
> >
> > Greg as CFM would be fine.  It's nice to see someone volunteer.
>
> +1, thanks a lot Greg!
>
> Note that the last CF of a major version is probably even more work than
> "regular" CF, so if other people are interested there's probably room for
> multiple CFMs.

I do have the time available. What I don't necessarily have is insight
into everything that needs to be done, especially behind the scenes.
So if I could have someone to give me tips about things I'm missing or
review the things I am doing to see that I've covered everything that
would be great.


-- 
greg



Re: Commitfest manager for 2022-03

From
Justin Pryzby
Date:
Can I suggest to update the CF APP to allow:
| Target version: 16

I also suggest to update patches to indicate which are (not) being considered
for v15.

A few specific ones from myself:

|https://commitfest.postgresql.org/37/2573/
|pg_dump - read data for some options from external file     Ready for Committer     15
Marked RFC by Daniel Gustafsson since 2021-10-01.
Is it more likely than not to be included in v15, or should the "v15" be
removed ?

|https://commitfest.postgresql.org/37/3499/
|libpq compression     Waiting on Author     15
I re-raised the same concerns made ~12 months ago but haven't heard back.
WOA since 35 days.  Unlikely to be considered/included in v15.

|https://commitfest.postgresql.org/37/3571/
|Add LZ4 compression in pg_dump     Needs review
This patch was just submitted on 2022-02-25.
I did a lot of the same things this patch does for a previous patch submission
(32/2888) for pg_dump/ZSTD, so I could review this, if there were interest to
include it in v15.

|https://commitfest.postgresql.org/37/2349/
|Global temporary table     Needs review     15
The handling/hijacking of pg_class and pg_statistic is not likely to be
acceptable.  I doubt this will be progressed for v15.

|https://commitfest.postgresql.org/37/3448/
|reduce impact of lengthy startup and checkpoint tasks     Needs review     15
I think this is in the design/concept phase, and not likely for v15, except
maybe in reduced scope.

|https://commitfest.postgresql.org/37/2906/
|Row filtering for logical replication
If I'm not wrong, this is merged and should be closed?

For my own patches, it'd be helpful if someone would suggest if any are (not)
likely to be progressed for v15.

These are marked RFC, but have no "committer" set.  Are they all targetting
v15?

    https://commitfest.postgresql.org/37/2863/
    Function to log backtrace of postgres processes     Ready for Committer
    Should target v15 ?

    https://commitfest.postgresql.org/37/2897/
    Faster pglz compression     Ready for Committer     15
    Tomas, are you still planning to merge this one ?

    https://commitfest.postgresql.org/37/2992/
    Allow batched insert during cross-partition updates     Ready for Committer

    https://commitfest.postgresql.org/37/3073/
    Add callback table access method to reset filenode when dropping relation     Ready for Committer

    https://commitfest.postgresql.org/37/3384/
    use has_privs_for_role for predefined roles     Ready for Committer     15

    Do all of the RFC patches target v15 (except bugfixes) ?
    https://commitfest.postgresql.org/37/?status=3

These are not marked as targetting any version .. are any of them 
being considered for v15 ?

    https://commitfest.postgresql.org/37/2903/
    Parallel Hash Full Join     Needs review     

    https://commitfest.postgresql.org/37/3508/
    Avoid smgrimmedsync() during index build and add unbuffered IO API     Needs review

    https://commitfest.postgresql.org/37/3182/
    automatically generating node support functions     Needs review     

    https://commitfest.postgresql.org/37/3183/
    Detectable crashes and unlogged table resets     Needs review         Jeff Davis (jdavis)

These are set as "targetting v15", but have no committer set.  If someone
thinks it's unrealistic they'll be included in v15, I suggest to update not to
say so.

    https://commitfest.postgresql.org/37/3272/
    Add system view tracking shared buffer actions     Needs review     15

    https://commitfest.postgresql.org/37/3224/
    Fix ExecRTCheckPerms() inefficiency with many prunable partitions     Needs review     15

    https://commitfest.postgresql.org/37/3270/
    Cache tuple routing info during bulk loads into partitioned tables     Needs review     15

    https://commitfest.postgresql.org/37/3461/
    In-place persistence change of a relation (fast ALTER TABLE ... SET LOGGED with wal_level=minimal)     Needs review
   15
 

    https://commitfest.postgresql.org/37/3539/
    Allow parallel plan for referential integrity checks     Needs review     15

    https://commitfest.postgresql.org/37/3397/
    Prefetching in recovery, take II     Needs review     15     Thomas Munro (macdice)

    https://commitfest.postgresql.org/37/2482/
    jsonpath syntax extensions     Needs review     15

    https://commitfest.postgresql.org/37/2901/
    SQL/JSON: functions     Needs review     15

    https://commitfest.postgresql.org/37/2902/
    SQL/JSON: JSON_TABLE     Needs review     15

    https://commitfest.postgresql.org/37/3099/
    Asymmetric partition-wise JOIN     Needs review     15

    https://commitfest.postgresql.org/37/3293/
    Tags in errordata     Needs review     15

    https://commitfest.postgresql.org/37/3358/
    Update relfrozenxmin when truncating temp tables     Needs review     15     Greg Stark (stark)

    https://commitfest.postgresql.org/37/3490/
    Pluggable toaster     Needs review     15

    https://commitfest.postgresql.org/37/3500/
    Collecting statistics about contents of JSONB columns     Needs review     15     Tomas Vondra (fuzzycz),

    https://commitfest.postgresql.org/37/2138/
    Incremental Materialized View Maintenance     Needs review     15

    https://commitfest.postgresql.org/37/3565/
    Add relation and block-level filtering to pg_waldump     Needs review     15     David Christensen (dwc-pgguru)

    https://commitfest.postgresql.org/37/3436/
    remove exclusive backup mode     Needs review     15     David Steele (dsteele)

    https://commitfest.postgresql.org/37/3457/
    pg_stat_toast     Needs review     15



Re: Commitfest manager for 2022-03

From
Julien Rouhaud
Date:
On Sat, Feb 26, 2022 at 06:37:21PM -0600, Justin Pryzby wrote:
> Can I suggest to update the CF APP to allow:
> | Target version: 16
> 
> I also suggest to update patches to indicate which are (not) being considered
> for v15.

I don't really understand what that field is supposed to mean.  But now that
we're in the final pg15 commit fest, wouldn't it be simpler to actually move
the patches for which there's a agreement that they can't make it to pg15?
Tagging them and letting them rot for a month isn't helpful for the authors or
the CFMs, especially when there are already 250 patches that needs to be
handled.

There's already an on-going a discussion for the PGTT patchset, maybe it should
also happen for some of the thread you mentioned.



Re: Commitfest manager for 2022-03

From
Michael Paquier
Date:
On Sun, Feb 27, 2022 at 04:51:16PM +0800, Julien Rouhaud wrote:
> I don't really understand what that field is supposed to mean.  But now that
> we're in the final pg15 commit fest, wouldn't it be simpler to actually move
> the patches for which there's a agreement that they can't make it to pg15?
> Tagging them and letting them rot for a month isn't helpful for the authors or
> the CFMs, especially when there are already 250 patches that needs to be
> handled.

Yes, I don't see any point in having a new tag just to mark patches
that will have to be moved to the next CF anyway.  These should just
be moved to the July one rather than stay in the March one.
--
Michael

Attachment

Re: Commitfest manager for 2022-03

From
"Euler Taveira"
Date:
On Sat, Feb 26, 2022, at 9:37 PM, Justin Pryzby wrote:
|Row filtering for logical replication
If I'm not wrong, this is merged and should be closed?
I think Amit forgot to mark it as committed. Done.


--
Euler Taveira

Re: Commitfest manager for 2022-03

From
Amit Kapila
Date:
On Sun, Feb 27, 2022 at 11:37 PM Euler Taveira <euler@eulerto.com> wrote:
>
> On Sat, Feb 26, 2022, at 9:37 PM, Justin Pryzby wrote:
>
> |https://commitfest.postgresql.org/37/2906/
> |Row filtering for logical replication
> If I'm not wrong, this is merged and should be closed?
>
> I think Amit forgot to mark it as committed. Done.
>

I was waiting for some build farm cycles to complete but it is fine to
mark it now.

-- 
With Regards,
Amit Kapila.



Re: Commitfest manager for 2022-03

From
Justin Pryzby
Date:
On Sun, Feb 27, 2022 at 04:51:16PM +0800, Julien Rouhaud wrote:
> On Sat, Feb 26, 2022 at 06:37:21PM -0600, Justin Pryzby wrote:
> > Can I suggest to update the CF APP to allow:
> > | Target version: 16
> > 
> > I also suggest to update patches to indicate which are (not) being considered
> > for v15.
> 
> I don't really understand what that field is supposed to mean.

It can mean whatever we decide it's convenient for it to mean.  I suppose its
primary purpose may have been to indicate which backbranch a bugfix applies to
(?)

I find it useful to indicate whether or not a patch is intended to be
considered for the next release.  If a patch is targetting v15, it's more
interesting to review.

@Thomas Munro: I think it'd be useful if cfbot would also show 1) the "target
version" (if any); and, 2) whether a patch has a committer set.

-- 
Justin



Re: Commitfest manager for 2022-03

From
Andres Freund
Date:
Hi,

On 2022-02-26 16:12:27 -0500, Greg Stark wrote:
> I do have the time available. What I don't necessarily have is insight
> into everything that needs to be done, especially behind the scenes.

One thing that desperately needs doing, particularly during the last
commitfest, is looking through CF entries and pruning stuff that shouldn't be
there anymore or that are in the wrong state.

I just went through the list of patches that are failing on
http://cfbot.cputube.org/index.html

There were several CF entries that haven't made progress in months marked as
"needs review", despite the last things on the thread being asks of the author(s).

One can't just automatically mark all failing runs as "waiting on author"
because sometimes there are issues with somebody else posting an incremental
diff confusing cfbot or spurious test failures...

If a patch fails to apply and it looks to be a "real patch" clearly some
action has to be taken by the author, so marking entries as "waiting on
author" is good.

Greetings,

Andres Freund



Re: Commitfest manager for 2022-03

From
Greg Stark
Date:
On Mon, 21 Mar 2022 at 21:48, Andres Freund <andres@anarazel.de> wrote:
>
> Hi,
>
> On 2022-02-26 16:12:27 -0500, Greg Stark wrote:
> > I do have the time available. What I don't necessarily have is insight
> > into everything that needs to be done, especially behind the scenes.
>
> One thing that desperately needs doing, particularly during the last
> commitfest, is looking through CF entries and pruning stuff that shouldn't be
> there anymore or that are in the wrong state.

Thanks

> One can't just automatically mark all failing runs as "waiting on author"
> because sometimes there are issues with somebody else posting an incremental
> diff confusing cfbot or spurious test failures...

What I'm seeing is patches that are failing with either the
027_stream_regress.pl failure that I see is being actively worked on
in another thread or simply a timeout which I'm not sure but may be
the same issue?

But I'll do a pass and then do another pass later in the week when
those problems may have been ironed out.


--
greg