On Tuesday, August 16, 2022 2:04 AM Melih Mutlu wrote:
> Attached patch also includes some additions to the doc along with the tests.
Hi, thank you for updating the patch. Minor review comments for the v2.
(1) whitespace issues
Please fix below whitespace errors.
$ git apply
On 13.09.22 21:02, Nathan Bossart wrote:
On Tue, Sep 13, 2022 at 08:32:43AM +0200, Peter Eisentraut wrote:
I notice that for column misspellings, the hint is phrased "Perhaps you
meant X." whereas here we have "Did you mean X?". Let's make that uniform.
Good point. I attached a new version
I have to say, to my embarrassment, after sending the previous email, I've
notice minor imperfections in a patch set caused by the last rebase.
These imperfections led to cf bot fail. I'll address this issue in the next
iteration in order not to generate excessive flow.
--
Best regards,
Maxim
Simon Riggs writes:
> A user asked me whether we prune never visible changes, such as
> BEGIN;
> INSERT...
> UPDATE.. (same row)
> COMMIT;
Didn't we just have this discussion in another thread? You cannot
do that, at least not without checking that the originating
transaction has no snapshots
Hi,
On 2022-09-16 09:14:20 +1200, Thomas Munro wrote:
> On Thu, Sep 15, 2022 at 2:26 PM Andres Freund wrote:
> > - noticed that libpgport.a had and needed a dependency on errcodes.h - that
> > seemed wrong. The dependency is due to src/port/*p{read,write}v?.c
> > including
> > postgres.h -
On Fri, 2022-09-16 at 10:26 -0400, Tom Lane wrote:
> Simon Riggs writes:
> > A user asked me whether we prune never visible changes, such as
> > BEGIN;
> > INSERT...
> > UPDATE.. (same row)
> > COMMIT;
>
> Didn't we just have this discussion in another thread?
For reference: that was
Andres Freund writes:
> On 2022-09-16 11:40:46 -0400, Tom Lane wrote:
>> I suspect this is less about clang and more about LLVM APIs,
>> but anyway it seems like we gotta fix something.
> Yea, there's definitely a bunch of llvm 15 issues that need to be fixed - this
> particular failure is
>The utility commands for cursor like DECLARE CURSOR seem to have the same
> issue
>and can cause lots of pgss entries. For example, when we use postgres_fdw
> and
>execute "SELECT * FROM WHERE id = 10" five times in the
> same
>transaction, the following commands are executed
Hi,
On 9/16/22 2:53 PM, Fujii Masao wrote:
Attached v5 to normalize 2PC commands too, so that we get things like:
+ case T_VariableSetStmt:
+ {
+ VariableSetStmt *stmt = (VariableSetStmt *) node;
+
+ /* stmt->name is NULL for RESET ALL */
+
Andres Freund writes:
> On 2022-09-16 09:14:20 +1200, Thomas Munro wrote:
>> On Thu, Sep 15, 2022 at 2:26 PM Andres Freund wrote:
>>> - noticed that libpgport.a had and needed a dependency on errcodes.h - that
>>> seemed wrong. The dependency is due to src/port/*p{read,write}v?.c including
>>>
On 08.09.22 20:18, Jacob Champion wrote:
Sounds fair. "cleartext"? "plaintext"? "plain" (like SASL's PLAIN)?
On the SASL front: In the back of my head I'd been considering adding
a "sasl:" prefix to "scram-sha-256", so that we have a namespace for
new SASL methods. That would also give us a
Hi,
On 2022-09-16 11:40:46 -0400, Tom Lane wrote:
> According to
>
> https://bugzilla.redhat.com/show_bug.cgi?id=2127503
>
> bleeding-edge clang complains thusly:
>
> llvmjit_inline.cpp: In function 'std::unique_ptr
> llvm_load_summary(llvm::StringRef)':
> llvmjit_inline.cpp:771:37: error:
Hi,
On 2022-09-16 16:07:51 -0400, Tom Lane wrote:
> Andres Freund writes:
> > On 2022-09-16 11:40:46 -0400, Tom Lane wrote:
> >> I suspect this is less about clang and more about LLVM APIs,
> >> but anyway it seems like we gotta fix something.
>
> > Yea, there's definitely a bunch of llvm 15
Ronan Dunklau writes:
> The attached does that and is much simpler. I only took into account
> entryPagesFetched, not sure if we should also charge something for data pages.
I think this is wrong, because there is already a CPU charge based on
the number of tuples visited, down at the very end
On Sat, Sep 17, 2022 at 6:45 AM Andres Freund wrote:
> On 2022-09-16 11:40:46 -0400, Tom Lane wrote:
> > According to
> >
> > https://bugzilla.redhat.com/show_bug.cgi?id=2127503
> >
> > bleeding-edge clang complains thusly:
> >
> > llvmjit_inline.cpp: In function 'std::unique_ptr
> >
On Fri, Sep 16, 2022 at 03:54:53PM +0200, Peter Eisentraut wrote:
> I don't think we need a verb there at all. I committed it without a verb.
Thanks!
--
Nathan Bossart
Amazon Web Services: https://aws.amazon.com
Hello!
Unfortunately the patch needs another rebase due to the recent split of guc.c
(0a20ff54f5e66158930d5328f89f087d4e9ab400)
I'm reviewing a patch on top of a previous commit and noticed a failed test:
# Failed test 'no parameters missing from postgresql.conf.sample'
# at
According to
https://bugzilla.redhat.com/show_bug.cgi?id=2127503
bleeding-edge clang complains thusly:
llvmjit_inline.cpp: In function 'std::unique_ptr
llvm_load_summary(llvm::StringRef)':
llvmjit_inline.cpp:771:37: error: incomplete type 'llvm::MemoryBuffer' used in
nested name specifier
On Fri, 16 Sept 2022 at 15:26, Tom Lane wrote:
>
> Simon Riggs writes:
> > A user asked me whether we prune never visible changes, such as
> > BEGIN;
> > INSERT...
> > UPDATE.. (same row)
> > COMMIT;
>
> Didn't we just have this discussion in another thread?
Not that I was aware of, but it
Simon Riggs writes:
> On Fri, 16 Sept 2022 at 15:26, Tom Lane wrote:
>> You cannot
>> do that, at least not without checking that the originating
>> transaction has no snapshots that could see the older row version.
> I'm saying this is possible only AFTER the end of the originating
> xact, so
On Fri, Sep 16, 2022 at 12:30 AM Masahiko Sawada wrote:
> After a quick benchmark, I've confirmed that the amount of WAL records
> for freezing 1 million tuples reduced to about one-fifth (1.2GB vs
> 250MB). Great.
I think that the really interesting thing about the patch is how this
changes the
On Fri, Sep 16, 2022 at 7:56 AM Peter Eisentraut
wrote:
> On 08.09.22 20:18, Jacob Champion wrote:
> After thinking about this a bit more, I think it would be best if the
> words used here match exactly with what is used in pg_hba.conf. That's
> the only thing the user cares about: reject
Thomas Munro writes:
> there's also the walker stuff[1] to address.
Yeah. I just did some experimentation with that, and it looks like
neither gcc nor clang will cut you any slack at all for declaring
an argument as "void *": given say
typedef bool (*tree_walker_callback) (Node *node, void
On 16.09.22 08:49, Marina Polyakova wrote:
But perhaps the check that --icu-locale cannot be specified unless
locale provider icu is chosen should also be moved here? So all these
checks will be in one place and it will use the provider from the
template database (which could be icu):
$
On Fri, Sep 16, 2022 at 4:19 PM Tom Lane wrote:
> I agree, this has always been a pet peeve of mine as well. I would
> have guessed there were fewer examples than you found, because I've
> generally fixed any such cases I happened to notice.
If you actually go through them all one by one you'll
Peter Geoghegan writes:
> It's possible to configure the clang-tidy tooling to tolerate various
> inconsistencies, below some kind of threshold -- it is totally
> customizable. But I think that a strict, simple rule is the way to go
> here.
Agreed; I see no need to tolerate any inconsistency.
>
On Fri, Sep 16, 2022 at 4:49 PM Tom Lane wrote:
> Agreed; I see no need to tolerate any inconsistency.
The check that I used to write the patches doesn't treat unnamed
parameters in a function declaration as an inconsistency, even when
"strict" is used. Another nearby check *could* be used to
2022-09-16 17:30 に Kyotaro Horiguchi さんは書きました:
At Fri, 16 Sep 2022 15:55:33 +0900, bt22nakamorit
wrote in
Hi,
"""\set ON_ERROR_STOP on""" stops any subsequent incoming query that
comes after an error of an SQL, but does not stop after a shell script
ran by """\! """ returning values other
@cfbot: rebased over adb466150, which did the same thing as one of the
hunks in xlogreader.c.
>From c4069bb7181b68d742d2025567f859e69d24f513 Mon Sep 17 00:00:00 2001
From: Kyotaro Horiguchi
Date: Thu, 7 Jul 2022 11:51:45 +0900
Subject: [PATCH] Make End-Of-Recovery error less scary
When recovery
Hi,
I liked this idea and after reviewing code I noticed some moments and
I'd rather ask you some questions.
Firstly, I suggest some editing in the comment of commit. I think, it is
turned out the more laconic and the same clear. I wrote it below since I
can't think of any other way to add
Hi Thomas,
While I was working on adding the page headers to SLRU pages on your patch, I
came across this code where it seems like "MultiXactIdToMemberPage" is
mistakenly being used instead of MultiXactIdToOffsetPage in the TrimMultiXact
function.
Below is the area of concern in the patch:
On Fri, Sep 16, 2022 at 6:20 PM Tom Lane wrote:
> I think they're easily Stroustrup's worst idea ever. You're basically
> throwing away an opportunity for documentation, and that documentation
> is often sorely needed.
He could at least point to C++ pure virtual functions, where omitting
a
I wrote:
> Ugh. I wonder if we can get away with declaring the walker arguments
> as something like "bool (*walker) (Node *, void *)" without having
> to change all the actual walkers to be exactly that signature.
> Having to insert casts in the walkers would be a major pain-in-the-butt.
No joy
Peter Geoghegan writes:
> The check that I used to write the patches doesn't treat unnamed
> parameters in a function declaration as an inconsistency, even when
> "strict" is used. Another nearby check *could* be used to catch
> unnamed parameters [1] if that was deemed useful, though. How do you
Peter Geoghegan writes:
> I have to admit that these inconsistencies are a pet peeve of mine. I
> find them distracting, and have a history of fixing them on an ad-hoc
> basis. But there are real practical arguments in favor of being strict
> about it as a matter of policy -- it's not *just*
On Fri, Sep 16, 2022 at 1:09 PM houzj.f...@fujitsu.com
wrote:
>
> Attach the new version patch which addressed above comments and ran pgident.
> I also improved some codes and documents based on some comments
> given by Vignesh and Peter offlist.
>
Thanks, the patch looks mostly good to me. I
On Fri, Sep 16, 2022 at 02:54:14PM +0700, John Naylor wrote:
> v6 demonstrates why this should have been put off towards the end. (more
> below)
Since the SIMD code is fresh in my mind, I wanted to offer my review for
0001 in the "Improve dead tuple storage for lazy vacuum" thread [0].
However,
On Saturday, September 17, 2022 11:22 AM Amit Kapila
wrote:
>
> On Fri, Sep 16, 2022 at 1:09 PM houzj.f...@fujitsu.com
>
> wrote:
> >
> > Attach the new version patch which addressed above comments and ran
> pgident.
> > I also improved some codes and documents based on some comments given
>
Hi,
On 9/12/22 9:55 AM, Drouvot, Bertrand wrote:
Hi,
On 9/10/22 1:21 AM, Jacob Champion wrote:
On 8/19/22 01:12, Drouvot, Bertrand wrote:
+ wstr = palloc((strlen(tok->string + 1) + 1) * sizeof(pg_wchar));
+ wlen = pg_mb2wchar_with_len(tok->string + 1,
+
On Mon, Aug 15, 2022 at 10:39 PM John Naylor
wrote:
>
> On Mon, Aug 15, 2022 at 12:39 PM Masahiko Sawada
> wrote:
> >
> > On Fri, Jul 22, 2022 at 10:43 AM Masahiko Sawada
> > wrote:
> > >
> > > On Tue, Jul 19, 2022 at 1:30 PM John Naylor
> > > wrote:
> > > >
> > > >
> > > >
> > > > On Tue,
At Fri, 16 Sep 2022 12:10:05 +1200, Thomas Munro wrote
in
> On Wed, Sep 14, 2022 at 2:38 PM Tom Lane wrote:
> > Kyotaro Horiguchi writes:
> > > I saw the following message recently modified.
> > >> This controls the maximum distance we can read ahead in the WAL to
> > >> prefetch referenced
On 2022-09-15 09:52, Kyotaro Horiguchi wrote:
If I executed initdb as follows, I would be told to specify
--icu-locale option.
$ initdb --encoding sql-ascii --locale-provider icu hoge
...
initdb: error: ICU locale must be specified
However, when I reran the command, it complains about
Hi Hackers,
I see in the texteq() function calls to DatumGetTextPP() are followed
by conditional calls to PG_FREE_IF_COPY. e.g.
https://github.com/postgres/postgres/blob/master/src/backend/utils/adt/varlena.c#L1792
text *targ1 = DatumGetTextPP(arg1);
text *targ2 = DatumGetTextPP(arg2);
On Fri, Sep 16, 2022 at 11:46 AM Kyotaro Horiguchi
wrote:
>
> At Fri, 16 Sep 2022 12:10:05 +1200, Thomas Munro
> wrote in
> > On Wed, Sep 14, 2022 at 2:38 PM Tom Lane wrote:
> > > Kyotaro Horiguchi writes:
> > > > I saw the following message recently modified.
> > > >> This controls the
CK Tan writes:
> I see in the texteq() function calls to DatumGetTextPP() are followed
> by conditional calls to PG_FREE_IF_COPY. e.g.
That's because texteq() is potentially usable as a btree index
function, and btree isn't too forgiving about leaks.
> However, in textlike(), PG_FREE_IF_COPY
On Wed, Sep 14, 2022 at 02:24:12PM +0530, Bharath Rupireddy wrote:
> I'm attaching the v4 patch that's rebased on to the latest HEAD.
> Please consider this for review.
I have been looking at this patch.
- StringInfo labelfile;
- StringInfo tblspc_map_file;
backup_manifest_info
At Fri, 16 Sep 2022 09:15:58 +0530, Bharath Rupireddy
wrote in
> On Thu, Sep 15, 2022 at 1:52 PM Kyotaro Horiguchi
> wrote:
> >
> > At Thu, 15 Sep 2022 10:28:12 +0530, Bharath Rupireddy
> > wrote in
> > > I'm attaching the v6 patch that's rebased on to the latest HEAD.
> > > Please consider
On Tue, Sep 6, 2022 at 12:39 AM Zhang Mingli wrote:
> Macro exec_subplan_get_plan is not used anymore.
> Attach a patch to remove it.
>
How about add it to the CF to not lose track of it?
Thanks
Richard
Hi,
"""\set ON_ERROR_STOP on""" stops any subsequent incoming query that
comes after an error of an SQL, but does not stop after a shell script
ran by """\! """ returning values other than 0, -1, or
127, which suggests a failure in the result of the shell script.
For example, suppose that
On Sep 16, 2022, 14:47 +0800, Richard Guo , wrote:
>
> On Tue, Sep 6, 2022 at 12:39 AM Zhang Mingli wrote:
> > Macro exec_subplan_get_plan is not used anymore.
> > Attach a patch to remove it.
>
> How about add it to the CF to not lose track of it?
Will add it, thanks~
Regards,
Zhang Mingli
On 2022-09-16 07:55, Kyotaro Horiguchi wrote:
At Thu, 15 Sep 2022 18:41:31 +0300, Marina Polyakova
wrote in
P.S. While working on the patch, I discovered that UTF8 encoding is
always used for the ICU provider in initdb unless it is explicitly
specified by the user:
if (!encoding &&
On Friday, September 16, 2022 1:42 PM Amit Kapila
wrote:
>
> On Thu, Sep 15, 2022 at 6:27 PM houzj.f...@fujitsu.com
> wrote:
> >
> > Attach the new version patch which added suggested restriction for
> > column list and merged Vignesh's patch.
> >
>
> Few comments:
>
> 1.
>
On Fri, Sep 16, 2022 at 1:01 PM Masahiko Sawada wrote:
>
> On Mon, Aug 15, 2022 at 10:39 PM John Naylor
> wrote:
> >
> > bool, buth = and <=. Should be pretty close. Also, i believe if you
> > left this for last as a possible refactoring, it might save some work.
v6 demonstrates why this should
On Wed, Jul 27, 2022 at 5:10 PM Xing Guo wrote:
> The bounded heap sorting status flag is set twice in sort_bounded_heap()
> and tuplesort_performsort(). This patch helps remove one of them.
>
Revisiting this patch I think maybe it's better to remove the setting of
Tuplesort status from
Got it. It is a leak-by-design for efficiency.
Thanks,
-cktan
On Fri, Sep 16, 2022 at 12:03 AM Tom Lane wrote:
>
> CK Tan writes:
> > I see in the texteq() function calls to DatumGetTextPP() are followed
> > by conditional calls to PG_FREE_IF_COPY. e.g.
>
> That's because texteq() is
On Tue, Sep 13, 2022 at 6:02 AM Peter Geoghegan wrote:
>
> My ongoing project to make VACUUM more predictable over time by
> proactive freezing [1] will increase the overall number of tuples
> frozen by VACUUM significantly (at least in larger tables). It's
> important that we avoid any new
On Fri, Sep 16, 2022 at 12:06 PM Kyotaro Horiguchi
wrote:
>
> In other words, it seems to me that the macro name doesn't manifest
> the condition correctly.
>
> I don't think we don't particularly want to do that unconditionally.
> I wanted just to get rid of the macro from the usage site. Even
I have a question about displaying NestLoopParam. In the plan below,
# explain (costs off)
select * from a, lateral (select sum(i) as i from b where exists (select
sum(c.i) from c where c.j = a.j and c.i = b.i) ) ss where a.i = ss.i;
QUERY PLAN
On 16.09.22 09:31, Marina Polyakova wrote:
IMO it is hardly understantable from the program output either - it
looks like I manually chose the encoding UTF8. Maybe first inform about
selected encoding?..
Yes, I included something like that in the patch just committed.
diff --git
On 15.09.22 17:41, Marina Polyakova wrote:
I agree with you. Here's another version of the patch. The
locale/encoding checks and reports in initdb have been reordered,
because now the encoding is set first and only then the ICU locale is
checked.
I committed something based on the first
At Fri, 16 Sep 2022 12:29:52 +0530, Amit Kapila wrote
in
> So, the change related to wal_decode_buffer_size needs to be
> backpatched to 15 whereas other message changes will be HEAD only, am
> I correct?
Has 15 closed the entry? IMHO I supposed that all changes are applied
back(?) to 15.
On Fri, Sep 16, 2022 at 2:07 PM Kyotaro Horiguchi
wrote:
>
> At Fri, 16 Sep 2022 12:29:52 +0530, Amit Kapila
> wrote in
> > So, the change related to wal_decode_buffer_size needs to be
> > backpatched to 15 whereas other message changes will be HEAD only, am
> > I correct?
>
> Has 15 closed the
At Fri, 16 Sep 2022 09:49:28 +0300, Marina Polyakova
wrote in
> On 2022-09-15 09:52, Kyotaro Horiguchi wrote:
> > However, when I reran the command, it complains about incompatible
> > encoding this time. I think it's more user-friendly to check for the
> > encoding compatibility before the
At Fri, 16 Sep 2022 09:56:31 +0200, Peter Eisentraut
wrote in
> On 15.09.22 17:41, Marina Polyakova wrote:
> > I agree with you. Here's another version of the patch. The
> > locale/encoding checks and reports in initdb have been reordered,
> > because now the encoding is set first and only then
At Fri, 16 Sep 2022 15:55:33 +0900, bt22nakamorit
wrote in
> Hi,
>
> """\set ON_ERROR_STOP on""" stops any subsequent incoming query that
> comes after an error of an SQL, but does not stop after a shell script
> ran by """\! """ returning values other than 0, -1, or
> 127, which suggests a
A user asked me whether we prune never visible changes, such as
BEGIN;
INSERT...
UPDATE.. (same row)
COMMIT;
Once committed, the original insert is no longer visible to anyone, so
"ought to be able to be pruned", sayeth the user. And they also say
that changing the app is much harder, as ever.
On 2022/09/09 19:11, Drouvot, Bertrand wrote:
IME if your application relies on 2PC it's very likely that you will hit the
exact same problems described in your original email.
The utility commands for cursor like DECLARE CURSOR seem to have the same issue
and can cause lots of pgss
Hi,
The release team is planning to release PostgreSQL 15 Release Candidate
1 (RC1) on 2022-09-29. Please ensure all open items[1] are resolved no
later than 2022-09-24 0:00 AoE.
Following recent release patterns, we planning for 2022-10-06 to be the
GA date. This may change based on what
Someone on general list recently complained that the error message
from trying to use options on a partitioned table was misleading,
which it definitely is:
CREATE TABLE parted_col_comment (a int, b text) PARTITION BY LIST (a)
WITH (fillfactor=100);
ERROR: unrecognized parameter "fillfactor"
On Thu, 15 Sept 2022 at 12:36, Japin Li wrote:
>
>
> On Thu, 15 Sep 2022 at 18:04, Simon Riggs
> wrote:
> > On Wed, 14 Sept 2022 at 15:21, Alvaro Herrera
> > wrote:
> >>
> >> On 2022-Aug-30, Simon Riggs wrote:
> >>
> >> > 001_new_isolation_tests_for_subxids.v3.patch
> >> > Adds new test cases
On Fri, 16 Sep 2022 at 20:13, Simon Riggs wrote:
> Someone on general list recently complained that the error message
> from trying to use options on a partitioned table was misleading,
> which it definitely is:
>
> CREATE TABLE parted_col_comment (a int, b text) PARTITION BY LIST (a)
> WITH
I wrote:
> On Fri, 16 Sep 2022 at 20:13, Simon Riggs
> wrote:
>> Patch 002 replaces this with a more meaningful error message, which
>> matches our fine manual.
>> https://www.postgresql.org/docs/current/sql-createtable.html
>>
>> ERROR: cannot specify storage options for a partitioned table
72 matches
Mail list logo
