On Wed, May 17, 2017 at 8:38 PM, Heikki Linnakangas wrote:
>
> As a final note, I found an interesting paper called "Unnesting Arbitrary
> Queries", by Thomas Neumann and Alfons Kemper
>
Hello everybody,
While analysing the performance of TPC-H queries for the newly developed
parallel-operators, viz, parallel index, bitmap heap scan, etc. we noticed
that the time taken by gather node is significant. On investigation, as per
the current method it copies each tuple to the shared
On Thu, May 18, 2017 at 7:23 AM, Amit Langote
wrote:
> On 2017/05/18 10:49, Amit Langote wrote:
>> On 2017/05/18 2:14, Dilip Kumar wrote:
>>> On Wed, May 17, 2017 at 7:41 PM, wrote:
(gdb) bt
#0 0x0061ab1b in list_nth
I propose to allow the `-` char in allowed label characters
The reason is to allow to use more easily base64 ids, (like
https://github.com/dylang/shortid, ...), uuid's too in labels
`-` is also not used as a query operator as a plus
source:
On Fri, May 19, 2017 at 12:44:27PM +0200, Pierre-Emmanuel André wrote:
> On Fri, May 19, 2017 at 12:21:26PM +0300, Heikki Linnakangas wrote:
> > On 05/19/2017 12:03 PM, Pierre-Emmanuel André wrote:
> > > Hi,
> > >
> > > I tried to build PostgreSQL 10beta1 on OpenBSD -current and i have this
> >
On Thu, May 18, 2017 10:24 pm, Tsunakawa, Takayuki wrote:
> From: pgsql-hackers-ow...@postgresql.org
>> [mailto:pgsql-hackers-ow...@postgresql.org] On Behalf Of Michael
>> Paquier
>> On Thu, May 18, 2017 at 11:30 PM, Robert Haas
>> wrote:
>> > Because why?
>>
>> Because it
On Wed, May 17, 2017 at 6:54 PM, Ashutosh Bapat
wrote:
[...]
>
> Comments on the tests
> +#ifdef USE_ASSERT_CHECKING
> +{
> +/*
> + * Hash partition bound stores modulus and remainder at
> + * b1->datums[i][0] and
Tom, all,
* Tom Lane (t...@sss.pgh.pa.us) wrote:
> All in all, this looks pretty darn good from here, and I'm thinking
> we should push forward on it.
+1.
Thanks!
Stephen
signature.asc
Description: Digital signature
Hi,
I tried to build PostgreSQL 10beta1 on OpenBSD -current and i have this error:
gmake[3]: Entering directory
'/usr/ports/pobj/postgresql-10beta1/postgresql-10beta1/src/backend/libpq'
cc -Wall -Wmissing-prototypes -Wpointer-arith -Wdeclaration-after-statement
-Wendif-labels
On Fri, May 19, 2017 at 11:21 AM, Heikki Linnakangas
wrote:
> On 05/19/2017 12:03 PM, Pierre-Emmanuel André wrote:
>
> Mikael, could you add --with-bsd-auth to curculio's configuration, please?
> On 9.6 and above.
>
> - Heikki
>
>
Hi Heikki,
I've added -with-bsd-auth to 9.6
Hi,
There is small issue in the HINT message which we provide at the time of
dropping subscription ,where we are saying -WITH (slot_name) which need
to change with SET (slot_name).
postgres=# drop subscription sub;
ERROR: could not connect to publisher when attempting to drop the
On 05/19/2017 12:03 PM, Pierre-Emmanuel André wrote:
Hi,
I tried to build PostgreSQL 10beta1 on OpenBSD -current and i have this error:
gmake[3]: Entering directory
'/usr/ports/pobj/postgresql-10beta1/postgresql-10beta1/src/backend/libpq'
cc -Wall -Wmissing-prototypes -Wpointer-arith
On Fri, May 19, 2017 at 12:21:26PM +0300, Heikki Linnakangas wrote:
> On 05/19/2017 12:03 PM, Pierre-Emmanuel André wrote:
> > Hi,
> >
> > I tried to build PostgreSQL 10beta1 on OpenBSD -current and i have this
> > error:
> >
> > gmake[3]: Entering directory
> >
On Fri, May 19, 2017 at 3:41 PM, tushar wrote:
> postgres=# drop subscription sub;
> ERROR: could not connect to publisher when attempting to drop the
> replication slot "pub"
> DETAIL: The error was: could not connect to server: No such file or
> directory
>
On Sat, May 20, 2017 at 10:43 AM, Thomas Munro
wrote:
> On Fri, May 19, 2017 at 6:35 PM, Amit Langote
> wrote:
>> On 2017/05/19 15:16, Thomas Munro wrote:
>>> Would TransitionCaptureState be a better name for this struct?
>>
>> Yes.
Michael Paquier writes:
> Do you think that it would be better to list the letter list for each
> keyword in repl_scanner.l or have something more generic? As not that
> many commands are added in the replication protocol, I would think
> that this is more than enough
On 18 May 2017 at 20:28, David Rowley wrote:
> A vastly simplified example case is:
>
> create table fkest (a int, b int, c int unique, primary key(a,b));
> create table fkest1 (a int, b int, primary key(a,b));
>
> insert into fkest select x/10,x%10, x from
On Thu, May 18, 2017 at 12:36:46PM +0900, Masahiko Sawada wrote:
> On Thu, May 18, 2017 at 11:29 AM, Peter Eisentraut
> wrote:
> > On 5/12/17 00:30, Masahiko Sawada wrote:
> >> I got same log messages 'starting logical replication worker for
> >> subscription'
On Fri, May 19, 2017 at 5:58 PM, Robert Haas wrote:
> On Fri, May 19, 2017 at 7:55 AM, Rafia Sabih
> wrote:
>> While analysing the performance of TPC-H queries for the newly developed
>> parallel-operators, viz, parallel index, bitmap heap
On Sat, May 20, 2017 at 2:24 AM, Robert Haas wrote:
> On Tue, May 16, 2017 at 9:25 AM, Amit Kapila wrote:
>> On Tue, May 16, 2017 at 5:14 PM, Robert Haas wrote:
>>> On Tue, May 16, 2017 at 7:31 AM, Amit Kapila
Robert Haas writes:
> I guess it does seem likely that most users of the hook would need to
> do the same, but it seems confusing to pass the same function both x
> and f(x), so my vote is to not do that.
I guess what's in the back of my mind is that the password type
On Fri, May 19, 2017 at 5:31 PM, Tom Lane wrote:
>> Anyone else want to vote? So far I count 3-1 in favor of making this change.
>
> Actually, on looking at the final form of the patch, it's hard to think
> that it's not just useless API churn. The one existing hook user
On Sat, May 20, 2017 at 5:32 AM, Tom Lane wrote:
> We'd probably be better off to implement case-insensitivity the hard way.
> There is a reason why none of our other flex scanners use this switch.
Do you think that it would be better to list the letter list for each
keyword
Umm, just skimming here -- this patch shows some LOG messages using
elog() rather than ereport(), which seems bogus to me.
Also:
"logical replication table synchronization worker for subscription
\"%s\", table \"%s\" has started"
surely there is a more convenient name than "logical
On Thu, May 18, 2017 at 4:38 AM, Ashutosh Bapat
wrote:
> But in case a user has written an = operator which returns true for
> two NULL values, per description in [1], that comparison operator is
> flawed and
> using that operator is going to result in
On Thu, May 18, 2017 at 4:54 PM, Andres Freund wrote:
> There's still weird behaviour, unfortunately. If you do an ALTER
> SEQUENCE changing minval/maxval w/ restart in a transaction, and abort,
> you'll a) quite possibly not be able to use the sequence anymore,
> because it
On Fri, May 19, 2017 at 2:36 AM, Jeff Davis wrote:
> I could agree to something like that. Let's explore some of the challenges
> there and potential solutions:
>
> 1. Dump/reload of hash partitioned data.
>
> Falling back to restore-through-the-root seems like a reasonable
Tom, all,
* Tom Lane (t...@sss.pgh.pa.us) wrote:
> Robert Haas writes:
> > On Thu, May 18, 2017 at 11:00 PM, Tom Lane wrote:
> >> The reason PGSSTrackLevel is "unrecognized" is that it's not in
> >> typedefs.list, which is a deficiency in our
On Fri, May 19, 2017 at 01:32:37PM +0200, Pierre-Emmanuel André wrote:
> On Fri, May 19, 2017 at 12:44:27PM +0200, Pierre-Emmanuel André wrote:
> > On Fri, May 19, 2017 at 12:21:26PM +0300, Heikki Linnakangas wrote:
> > > On 05/19/2017 12:03 PM, Pierre-Emmanuel André wrote:
> > > > Hi,
> > > >
>
While reading some code, noticed that the headers of functions
pg_walfile_name_offset() and pg_walfile_name() incorrecty refer
pg_xlog_switch() since the inception of code in commit 704ddaaa.
In PG10 implementation, actual name of the referred function is
pg_switch_wal(). So either refer the
Chapman Flack writes:
> Was my guess about the reason right? Does this PG10 announcement
> also mean it will be possible to use UNIQUE constraints with some
> pure-identifier, no-natural-ordering type that supports only hashing?
No, nobody's done anything about allowing
On 05/19/2017 06:05 PM, Stephen Frost wrote:
* Tom Lane (t...@sss.pgh.pa.us) wrote:
Robert Haas writes:
On Thu, May 18, 2017 at 11:00 PM, Tom Lane wrote:
The reason PGSSTrackLevel is "unrecognized" is that it's not in
typedefs.list, which is a
On Fri, May 19, 2017 at 11:22 AM, Tom Lane wrote:
> I certainly would rather that our version matched something that's under
> active maintenance someplace. But it seems like there are two good
> arguments for having a copy in our tree:
>
> * easy accessibility for PG
Hi,
The item on hash indexes reminded me of an old comment from years
ago that I put in the code of the first custom PG datatype I ever
built at $work:
COMMENT ON OPERATOR CLASS puid_ops USING btree IS
'As puids are only identifiers, there is no obvious reason to define
ordering operators or
On Fri, May 19, 2017 at 11:22:54AM -0400, Tom Lane wrote:
> We've had reasonably decent luck with tracking the tzcode/tzdata packages
> as local copies, so I feel like we're not taking on anything unreasonable
> if our model is that we'll occasionally (not oftener than once per year)
> update our
Heikki Linnakangas writes:
> On 05/19/2017 06:48 PM, Tom Lane wrote:
>> That's going to catch a lot of things that are just variables, though.
>> It might be all right as long as there was manual filtering after it.
> At a quick glance, there are only a couple of them. This two
Andres Freund writes:
> On 2017-05-19 12:21:52 -0400, Robert Haas wrote:
>> Yeah, but those advantages could also be gained by putting the
>> pgindent tree on git.postgresql.org in a separate repository. Having
>> it in the same repository as the actual PostgreSQL code is not
Heikki Linnakangas writes:
> You can get a pretty good typedefs list just by looking for the pattern
> "} ;".
That's going to catch a lot of things that are just variables, though.
It might be all right as long as there was manual filtering after it.
On 2017-05-19 12:21:52 -0400, Robert Haas wrote:
> On Fri, May 19, 2017 at 11:22 AM, Tom Lane wrote:
> > I certainly would rather that our version matched something that's under
> > active maintenance someplace. But it seems like there are two good
> > arguments for having a
Robert Haas writes:
> On Fri, May 19, 2017 at 11:22 AM, Tom Lane wrote:
>> I certainly would rather that our version matched something that's under
>> active maintenance someplace. But it seems like there are two good
>> arguments for having a copy in
On Fri, May 19, 2017 at 04:10:09AM +, Huong Dangminh wrote:
> Hi,
>
> > * 10 Beta Release Notes:
> > https://www.postgresql.org/docs/devel/static/release-10.html
>
> Just a minute thing, but changing of hot_standby default value is not fully
> noted in release-10.sgml.
> Please find the
On 05/19/2017 06:48 PM, Tom Lane wrote:
Heikki Linnakangas writes:
You can get a pretty good typedefs list just by looking for the pattern
"} ;".
That's going to catch a lot of things that are just variables, though.
It might be all right as long as there was manual
On Fri, May 19, 2017 at 5:32 AM, amul sul wrote:
> Updated patch attached. 0001-patch rebased against latest head.
> 0002-patch also incorporates code comments and error message changes
> as per Robert's & your suggestions. Thanks !
-if (spec->strategy !=
Stephen Frost writes:
> * Tom Lane (t...@sss.pgh.pa.us) wrote:
>> Yes, moving the goalposts on ease-of-use is an important consideration
>> here. What that says to me is that we ought to pull FreeBSD indent
>> into our tree, and provide Makefile support that makes it easy for
I propose to allow the `-` char in allowed label characters (currently
a-zA-Z0-9_ https://www.postgresql.org/docs/9.6/static/ltree.html)
The reason is to allow to use more easily base64 ids, (like
https://github.com/dylang/shortid, ...), uuid's too in labels
`-` is also not used as a query
A small patch for the release notes. The first hunk removes a duplicate
item. The second hunk adds another item to the incompatibility list.
I can commit this myself but want to give Bruce a chance so I don't mess
up his workflow.
--
Peter Eisentraut http://www.2ndQuadrant.com/
I wrote:
> What I was just looking at is the possibility of absorbing struct
> tags ("xllist" in the above) as if they were typedef names. In
> at least 95% of our usages, if a struct has a tag then the tag is
> also the struct's typedef name. The reason this is interesting
> is that it looks
Tom Lane wrote:
> Robert Haas writes:
> > Yeah, but those advantages could also be gained by putting the
> > pgindent tree on git.postgresql.org in a separate repository. Having
> > it in the same repository as the actual PostgreSQL code is not
> > required nor, in my
On Fri, May 19, 2017 at 01:35:41PM -0400, Peter Eisentraut wrote:
> A small patch for the release notes. The first hunk removes a duplicate
> item. The second hunk adds another item to the incompatibility list.
>
> I can commit this myself but want to give Bruce a chance so I don't mess
> up
On Tue, May 16, 2017 at 9:25 AM, Amit Kapila wrote:
> On Tue, May 16, 2017 at 5:14 PM, Robert Haas wrote:
>> On Tue, May 16, 2017 at 7:31 AM, Amit Kapila wrote:
>>> I will send an updated patch once we agree on above
Robert Haas writes:
> On Thu, May 18, 2017 at 11:00 PM, Tom Lane wrote:
>> The reason PGSSTrackLevel is "unrecognized" is that it's not in
>> typedefs.list, which is a deficiency in our typedef-collection
>> technology not in indent. (I believe the
On Thu, May 18, 2017 at 2:37 AM, Heikki Linnakangas wrote:
> I still don't think this worth it to change the hook function's signature
> for this. Even though it's not a big deal for external modules to adapt,
> it's not zero effort either. And it's not that much nicer to us.
I
Robert Haas writes:
> On Thu, May 18, 2017 at 2:37 AM, Heikki Linnakangas wrote:
>> I still don't think this worth it to change the hook function's signature
>> for this. Even though it's not a big deal for external modules to adapt,
>> it's not zero
On Wed, May 17, 2017 at 5:34 PM, Noah Misch wrote:
> [Action required within three days. This is a generic notification.]
>
> The above-described topic is currently a PostgreSQL 10 open item. Robert,
> since you committed the patch believed to have created it, you own this
On 5/19/17 11:22, Tom Lane wrote:
> I certainly would rather that our version matched something that's under
> active maintenance someplace. But it seems like there are two good
> arguments for having a copy in our tree:
Is pgindent going to be indented by pgindent?
--
Peter Eisentraut
On Fri, May 19, 2017 at 6:07 AM, Ashutosh Bapat
wrote:
> We still have the same copy shared across multiple append paths and
> set_plan_refs would change change it underneath those. May not be a
> problem right now but may be a problem in the future.
I agree. I
I looked a bit more carefully at cash.c in the wake of bug #14663,
https://www.postgresql.org/message-id/20170519164653.29941.19098%40wrigleys.postgresql.org
It seems to me that there are three different bugs in the multiplication
and division operators:
1. As noted in the bug thread, applying
Peter Eisentraut writes:
> On 5/19/17 11:22, Tom Lane wrote:
>> I certainly would rather that our version matched something that's under
>> active maintenance someplace. But it seems like there are two good
>> arguments for having a copy in our tree:
> Is
On Tue, May 16, 2017 at 11:58 PM, Michael Paquier
wrote:
> Thanks for the updated patch. This looks good to me.
Committed. I also added a slight tweak to the wording of the documentation.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise
On 5/19/17 06:58, Dilip Kumar wrote:
> On Fri, May 19, 2017 at 3:41 PM, tushar wrote:
>> postgres=# drop subscription sub;
>> ERROR: could not connect to publisher when attempting to drop the
>> replication slot "pub"
>> DETAIL: The error was: could not connect to
On 5/19/17 13:31, Alvaro Herrera wrote:
> I favor having indent in a separate repository in our Git server, for
> these reasons
I am also in favor of that.
> 0. it's under our control (so we can change rules as we see fit)
> 1. we can have Piotr as a committer there
> 2. we can use the same
On 5/19/17 01:01, Masahiko Sawada wrote:
> Seems all four table sync workers are launched at the almost same
> time, but three workers of them get stuck in idle transaction state
> when creating replication slot. That is these waiting workers cannot
> proceed its work until first connected table
"Higuchi, Daisuke" writes:
> By adding flex option '-i', replication command parser could be more
> flexible.
> This option is already used for syncrep_scanner.c, so it is not strange to
> add for repl_scanner.c too.
Really? That wasn't an especially bright
On Thu, May 18, 2017 at 8:11 PM, Michael Paquier
wrote:
> On Thu, May 18, 2017 at 11:30 PM, Robert Haas wrote:
>> On Thu, May 18, 2017 at 7:06 AM, Michael Paquier
>> wrote:
>>> FWIW, I am of the opinion to not have an
On Fri, May 19, 2017 at 12:51 AM, Michael Paquier
wrote:
> On Fri, May 19, 2017 at 1:16 PM, Higuchi, Daisuke
> wrote:
>> The reason of this problem is that sending 'show transaction_read_only' is
>> failed.
>> 'show' must be in
On 2017/05/19 15:16, Thomas Munro wrote:
> On Fri, May 19, 2017 at 5:51 PM, Amit Langote
> wrote:
>> I saw in the latest patch that now ExecSetupTriggerTransitionState() looks
>> at mtstate->mt_partition_dispatch_info when setting up the transition
>> conversion
On Thursday, May 18, 2017, Robert Haas wrote:
> My experience with this area has led
> me to give up on the idea of complete uniformity as impractical, and
> instead look at it from the perspective of "what do we absolutely have
> to ban in order for this to be sane?".
I
From: Michael Paquier [mailto:michael.paqu...@gmail.com]
> if (!PQsendQuery(conn,
>- "show transaction_read_only"))
>+ "SHOW transaction_read_only"))
>Or perhaps the replication command parser could be made
On Fri, May 19, 2017 at 11:05 AM, Michael Paquier
wrote:
> On Thu, May 18, 2017 at 1:43 PM, Thomas Munro
> wrote:
>> On Thu, May 11, 2017 at 1:48 PM, Michael Paquier
>> wrote:
>>> I had my eyes on the WAL
On Fri, May 19, 2017 at 5:51 PM, Amit Langote
wrote:
> I saw in the latest patch that now ExecSetupTriggerTransitionState() looks
> at mtstate->mt_partition_dispatch_info when setting up the transition
> conversion map. In the case where it's non-NULL, you may have
On 19/05/17 21:47, Peter Eisentraut wrote:
> On 5/19/17 01:01, Masahiko Sawada wrote:
>> Seems all four table sync workers are launched at the almost same
>> time, but three workers of them get stuck in idle transaction state
>> when creating replication slot. That is these waiting workers cannot
In match_eclasses_to_foreign_key_col(), there is this:
if (em->em_is_child)
continue; /* ignore children here */
ISTM, it might as well be:
Assert(!em->em_is_child);/* no children yet */
That's because, I think it's still too early in
On Fri, May 19, 2017 at 6:35 PM, Amit Langote
wrote:
> On 2017/05/19 15:16, Thomas Munro wrote:
>> Would TransitionCaptureState be a better name for this struct?
>
> Yes. Although, losing the Trigger prefix might make it sound a bit
> ambiguous though. Right above
A couple of thoughts I've been having that relate to this:
The traditional meaning of "installcheck" in GNU packages is to test
against the installed code, whereas "check" tests before installation.
Our concept of testing against a running server obviously does not apply
to many kinds of
On 4/24/17 22:26, Florin Asavoaie wrote:
> If there's nobody against this, I can try to do the patch myself,
> doesn't look too difficult (I expect it to simply work by
> calling SSL_set_tlsext_host_name(SSL_context, PQhost(conn))) somewhere
> in initialize_SSL in fe-secure-openssl.c.
I had to
On Thu, May 18, 2017 at 11:00 PM, Tom Lane wrote:
> * Improvements in formatting around sizeof and related constructs,
> for example:
>
> * Likewise, operators after casts work better than before:
>
> * Sane formatting of function typedefs, for example:
>
> * Non-typedef
On Fri, May 19, 2017 at 7:55 AM, Rafia Sabih
wrote:
> While analysing the performance of TPC-H queries for the newly developed
> parallel-operators, viz, parallel index, bitmap heap scan, etc. we noticed
> that the time taken by gather node is significant. On
On 18/05/17 16:16, Robert Haas wrote:
> On Wed, May 17, 2017 at 6:58 AM, Masahiko Sawada
> wrote:
>> I think the above changes can solve this issue but It seems to me that
>> holding AccessExclusiveLock on pg_subscription by DROP SUBSCRIPTION
>> until commit could lead
78 matches
Mail list logo