On Tue, Oct 04, 2022 at 11:17:09PM -0400, Tom Lane wrote:
> Nah, feel free.
Okay, thanks. Applied, then.
--
Michael
signature.asc
Description: PGP signature
On Tue, Oct 04, 2022 at 06:24:18PM +0530, Bharath Rupireddy wrote:
> I'm fine with doing either of these things. Let's hear from others.
>
> I've added a CF entry - https://commitfest.postgresql.org/40/3927/
About 0002, I am not sure that it is worth bothering. Sure, this
wastes a few bytes,
On Tue, Oct 4, 2022 at 7:59 PM Jeff Davis wrote:
> I am fine with that, but I'd like us all to understand what the
> downsides are.
Although I'm sure that there must be one case that loses measurably,
it's not particularly obvious where to start looking for one. I mean
it's easy to imagine
Hi Wang-san. Here are my review comments for HEAD_v12-0001 patch.
==
1. Missing documentation.
In [1] you wrote:
> I think the behaviour of multiple publications with parameter
> publish_via_partition_root could be added to the pg-doc later in a separate
> patch.
~
That doesn't seem
Michael Paquier writes:
> On Tue, Oct 04, 2022 at 09:16:44AM -0400, Tom Lane wrote:
>> I would go more for taking it out of queryjumble.h. I see no
>> reason why that constant needs to be world-visible.
> I was just looking at the patch before seeing your reply, and thought
> the exact same
On Tue, Oct 04, 2022 at 09:16:44AM -0400, Tom Lane wrote:
> I would go more for taking it out of queryjumble.h. I see no
> reason why that constant needs to be world-visible.
I was just looking at the patch before seeing your reply, and thought
the exact same thing. Perhaps you'd prefer apply
On Tue, 2022-10-04 at 11:09 -0700, Peter Geoghegan wrote:
> So a simplistic threshold
> (combined with dynamic per-page decisions about freezing) should be
> enough to avoid most of the downside of eager freezing.
...
> I want to keep
> the cost as low as possible (often "negative cost" relative
On Tue, Oct 04, 2022 at 10:35:25AM +0100, Dagfinn Ilmari Mannsåker wrote:
> LGTM, +1 to commit.
Fine by me, so done.
--
Michael
signature.asc
Description: PGP signature
At Tue, 4 Oct 2022 17:15:31 -0700, Nathan Bossart
wrote in
> On Tue, Oct 04, 2022 at 07:53:11PM -0400, Tom Lane wrote:
> > I wrote:
> >> PFA a quick-hack fix that solves this issue by making per-transaction
> >> subsidiary hash tables. That's overkill perhaps; I'm a little worried
> >> about
On Tue, Oct 04, 2022 at 07:53:11PM -0400, Tom Lane wrote:
> I wrote:
>> PFA a quick-hack fix that solves this issue by making per-transaction
>> subsidiary hash tables. That's overkill perhaps; I'm a little worried
>> about whether this slows down normal cases more than it's worth.
>> But we
I wrote:
> PFA a quick-hack fix that solves this issue by making per-transaction
> subsidiary hash tables. That's overkill perhaps; I'm a little worried
> about whether this slows down normal cases more than it's worth.
> But we ought to do something about this, because aside from the
>
Hi,
On 2022-10-03 10:01:25 -0700, Andres Freund wrote:
> On 2022-10-03 08:12:39 -0400, Robert Haas wrote:
> > On Fri, Sep 30, 2022 at 8:20 PM Andres Freund wrote:
> > I thought about trying to buy back some space elsewhere, and I think
> > that would be a reasonable approach to getting this
No, not quite.
Valid Punycode characters are `[A-Za-z0-9-]`. This proposal includes `-`,
as well as `#` and `;` for HTML entities.
I double-checked the RFC to see the valid Punycode characters and the set
above is indeed correct:
On Wed, Sep 28, 2022 at 08:16:08PM +0530, Bharath Rupireddy wrote:
> In doing so, I had to add a few Get/Set functions
> for XLogCtl variables so that xlogbackup.c can use them.
I would suggest moving this to a separate prerequisite patch that can be
reviewed independently from the patches that
[ redirecting to -hackers because patch attached ]
David Rowley writes:
> So that confirms there were 950k relations in the xl_standby_locks.
> The contents of that message seem to be produced by standby_desc().
> That should be the same WAL record that's processed by standby_redo()
> which adds
On Tue, Oct 04, 2022 at 12:54:46PM -0400, Garen Torikian wrote:
> The punycode range of characters is the exact same set as the existing
> ltree range, with the addition of a hyphen (-). Within this system, any
> human language can be encoded using just A-Za-z0-9-.
IIUC ASCII characters like '!'
On Thu, 29 Sept 2022 at 04:45, Bruce Momjian wrote:
>
> We have discussed the problems caused by the use of pg_stat_reset() and
> pg_stat_reset_shared(), specifically the removal of information needed
> by autovacuum. I don't see these risks documented anywhere. Should we
> do that? Are there
Hi,
For contain_placeholders():
+ if (IsA(node, Query))
+ return query_tree_walker((Query *) node, contain_placeholders,
context, 0);
+ else if (IsA(node, PlaceHolderVar))
The `else` is not needed.
For correlated_t struct, it would be better if the fields have comments.
+
Hi hackers!
cfbot is unhappy again, with documentation package.Here's
corrected patchset.
Patchset consists of:
v21-0001-toaster-interface.patch - Pluggable TOAST API interface along with
reference TOAST mechanics - new API is introduced but
reference TOAST is still unchanged;
On Tue, Oct 4, 2022 at 2:30 PM Andres Freund wrote:
> Consider the following sequence:
>
> 1) we write WAL like this:
>
> [record A][tear boundary][record B, prev A_lsn][tear boundary][record C, prev
> B_lsn]
>
> 2) crash, the sectors with A and C made it to disk, the one with B didn't
>
> 3) We
Hi,
On 2022-10-04 17:05:40 +0900, Michael Paquier wrote:
> On Mon, Oct 03, 2022 at 04:41:11PM -0700, Andres Freund wrote:
> > There's a few further roles that seem to pose some danger goign forward:
>
> I have never seen that myself, but 0001 is a nice cleanup.
> generated.sql includes a user
Hi,
On 2022-10-04 13:36:33 -0400, Robert Haas wrote:
> On Tue, Oct 4, 2022 at 11:34 AM Andres Freund wrote:
> > > Example: Page { [ record A ] | tear boundary | [ record B ] } gets
> > > recycled and receives a new record C at the place of A with the same
> > > length.
> > >
> > > With your
On Tue, Oct 4, 2022 at 10:39 AM Jeff Davis wrote:
> On Mon, 2022-10-03 at 22:45 -0700, Peter Geoghegan wrote:
> > Once a table becomes larger than vacuum_freeze_strategy_threshold,
> > VACUUM stops marking pages all-visible in the first place,
> > consistently marking them all-frozen instead.
>
>
Thomas Munro writes:
> I tried lots of crazy stuff[1] to try to get an error out of this
> thing, but came up empty handed. Unlike realpath(), _fullpath()
> doesn't resolve symlinks (or junctions), so I guess there's less to go
> wrong. It still needs the present working directory, which is a
>
Here is an updated patch set with the following changes:
* The creation of the struct for non-shared WAL receiver state is moved to
a prerequisite 0001 patch. This should help ease review of 0002 a bit.
* I updated the nap time calculation to round up to the next millisecond,
as discussed
On Mon, 2022-10-03 at 22:45 -0700, Peter Geoghegan wrote:
> Once a table becomes larger than vacuum_freeze_strategy_threshold,
> VACUUM stops marking pages all-visible in the first place,
> consistently marking them all-frozen instead.
What are the trade-offs here? Why does it depend on table
On Tue, Oct 4, 2022 at 11:34 AM Andres Freund wrote:
> > Example: Page { [ record A ] | tear boundary | [ record B ] } gets
> > recycled and receives a new record C at the place of A with the same
> > length.
> >
> > With your proposal, record B would still be a valid record when it
> > follows
Hi,
On 2022-10-04 10:24:19 +0200, Peter Eisentraut wrote:
> On 30.09.22 06:07, Andres Freund wrote:
> > When tap tests are interrupted (e.g. with ctrl-c), we don't cancel running
> > postgres instances etc. That doesn't strike me as a good thing.
> >
> > In contrast, the postgres instances
Dear hackers,
I am submitting a patch to expand the label requirements for ltree.
The current format is restricted to alphanumeric characters, plus _.
Unfortunately, for non-English labels, this set is insufficient. Rather
than figure out how to expand this set to include characters beyond the
I wrote:
> I think what we should look at is extending the aggregate/window
> function APIs so that such functions can report where they put their
> output, and then we can nuke MemoryContextContains(), with the
> code code set up to assume that it has to copy if the called function
> didn't
Hi,
On 2022-10-04 15:05:47 +0200, Matthias van de Meent wrote:
> On Mon, 3 Oct 2022 at 23:26, Andres Freund wrote:
> > On 2022-10-03 19:40:30 +0200, Matthias van de Meent wrote:
> > > On Mon, 3 Oct 2022, 19:01 Andres Freund, wrote:
> > > > Random idea: xl_prev is large. Store a full xl_prev in
Peter Eisentraut writes:
> I was wondering why we have a definition of Abs() in c.h when there are
> more standard functions such as abs() and fabs() in widespread use. I
> think this one is left over from pre-ANSI-C days. The attached patches
> replace all uses of Abs() with more standard
bt22nakamorit writes:
> queryjumble.c and queryjumble.h both define a macro JUMBLE_SIZE = 1024.
> Since queryjumble.c includes queryjumble.h, the JUMBLE_SIZE definition
> in queryjumble.c should be deleted.
I would go more for taking it out of queryjumble.h. I see no
reason why that constant
"Daniel Verite" writes:
> Tom Lane wrote:
>> Given the time pressure, I did not worry about installing regression
>> test coverage for this stuff, but I wonder if we shouldn't add some.
> Currently, test/regress/sql/psql.sql doesn't AFAICS write anything
> outside of stdout, but \g, \o,
On Mon, 3 Oct 2022 at 23:26, Andres Freund wrote:
>
> Hi,
>
> On 2022-10-03 19:40:30 +0200, Matthias van de Meent wrote:
> > On Mon, 3 Oct 2022, 19:01 Andres Freund, wrote:
> > > Random idea: xl_prev is large. Store a full xl_prev in the page header,
> > > but
> > > only store a 2 byte offset
Tom Lane wrote:
> Pushed after making some corrections.
Thanks!
> Given the time pressure, I did not worry about installing regression
> test coverage for this stuff, but I wonder if we shouldn't add some.
Currently, test/regress/sql/psql.sql doesn't AFAICS write anything
outside of
On Tue, Oct 4, 2022 at 2:01 PM Kyotaro Horiguchi
wrote:
>
> > 1. 0001 replaces explicit WAL file parsing code with
>
> Looks good to me.
>
> > 2. 0002 replaces MAXPGPATH with MAXFNAMELEN for WAL file names.
>
> Looks reasonable, too. I don't find other instances of the same mistake.
Thanks for
Em ter., 4 de out. de 2022 às 08:29, Ranier Vilela
escreveu:
> Em ter., 4 de out. de 2022 às 05:36, David Rowley
> escreveu:
>
>> On Tue, 4 Oct 2022 at 13:35, Ranier Vilela wrote:
>> > Revisiting my work on reducing memory consumption, I found this patch
>> left out.
>> > I'm not sure I can
Hi,
On Tue, Oct 04, 2022 at 02:15:16PM +0200, Pavel Stehule wrote:
>
> út 4. 10. 2022 v 12:48 odesílatel Никита Старовойтов
> napsal:
>
> > Hello,
> > with a view to meeting with postgres code and to get some practice with
> > it, I am making a small patch that adds the possibility of partial
Hi
út 4. 10. 2022 v 12:48 odesílatel Никита Старовойтов
napsal:
> Hello,
> with a view to meeting with postgres code and to get some practice with
> it, I am making a small patch that adds the possibility of partial tables
> dump.
> A rule of filtering is specified with standard SQL where
Em ter., 4 de out. de 2022 às 05:36, David Rowley
escreveu:
> On Tue, 4 Oct 2022 at 13:35, Ranier Vilela wrote:
> > Revisiting my work on reducing memory consumption, I found this patch
> left out.
> > I'm not sure I can help.
> > But basically I was able to write and read the block size, in
Em ter., 4 de out. de 2022 às 01:18, Michael Paquier
escreveu:
> On Mon, Oct 03, 2022 at 08:05:57AM -0300, Ranier Vilela wrote:
> > Em seg., 3 de out. de 2022 às 05:01, Masahiko Sawada <
> sawada.m...@gmail.com>
> > escreveu:
> >> On Fri, Sep 30, 2022 at 9:08 AM Ranier Vilela
> wrote:
> >>> 1.
Hello,
with a view to meeting with postgres code and to get some practice with it,
I am making a small patch that adds the possibility of partial tables dump.
A rule of filtering is specified with standard SQL where clause (without
"where" keyword)
There are three ways to send data filters over
Hi hackers!
Now cfbot is happy, but there were warnings due to recent changes in
PointerGetDatum function, so here's corrected patchset.
Sorry, forgot to attach patch files. My fault.
Patchset consists of:
v20-0001-toaster-interface.patch - Pluggable TOAST API interface along with
reference TOAST
Hi,
On Tue, Oct 04, 2022 at 05:41:12PM +0900, bt22nakamorit wrote:
>
> queryjumble.c and queryjumble.h both define a macro JUMBLE_SIZE = 1024.
> Since queryjumble.c includes queryjumble.h, the JUMBLE_SIZE definition in
> queryjumble.c should be deleted.
+1
Hi hackers!
Now cfbot is happy, but there were warnings due to recent changes in
PointerGetDatum function, so here's corrected patchset.
Patchset consists of:
v20-0001-toaster-interface.patch - Pluggable TOAST API interface along with
reference TOAST mechanics - new API is introduced but
vignesh C writes:
> On Tue, 4 Oct 2022 at 09:13, Michael Paquier wrote:
>>
>> On Mon, Oct 03, 2022 at 06:29:32PM +0100, Dagfinn Ilmari Mannsåker wrote:
>> > vignesh C writes:
>> >> +else if (TailMatchesCS("\\dRp*"))
>> >> +
On Tue, 2022-10-04 at 09:41 +0200, Peter Eisentraut wrote:
> In PostgreSQL 10, we added identity columns, as an alternative to serial
> columns (since 6.something). They mostly work the same. Identity
> columns are SQL-conforming, have some more features (e.g., overriding
> clause), and are a
On Tue, 4 Oct 2022 at 09:13, Michael Paquier wrote:
>
> On Mon, Oct 03, 2022 at 06:29:32PM +0100, Dagfinn Ilmari Mannsåker wrote:
> > vignesh C writes:
> >> +else if (TailMatchesCS("\\dRp*"))
> >> +COMPLETE_WITH_QUERY(Query_for_list_of_publications[0].query);
> >> +else if
On Thu, Sep 29, 2022 at 7:43 PM Bharath Rupireddy
wrote:
>
> Please see the attached v1 patch.
FWIW, I'm attaching Nathan's patch that introduced the new function
pg_walfile_offset_lsn as 0002 in the v1 patch set. Here's the CF entry
- https://commitfest.postgresql.org/40/3909/.
Please consider
Hi,
On Oct 4, 2022, 15:07 +0800, Peter Eisentraut
, wrote:
> I was wondering why we have a definition of Abs() in c.h when there are
> more standard functions such as abs() and fabs() in widespread use. I
> think this one is left over from pre-ANSI-C days. The attached patches
> replace all uses
Hi,
queryjumble.c and queryjumble.h both define a macro JUMBLE_SIZE = 1024.
Since queryjumble.c includes queryjumble.h, the JUMBLE_SIZE definition
in queryjumble.c should be deleted.
Thoughts?
Tatsudiff --git a/src/backend/utils/misc/queryjumble.c b/src/backend/utils/misc/queryjumble.c
index
On Tue, 4 Oct 2022 at 13:35, Ranier Vilela wrote:
> Revisiting my work on reducing memory consumption, I found this patch left
> out.
> I'm not sure I can help.
> But basically I was able to write and read the block size, in the chunk.
> Could it be the case of writing and reading the context
At Tue, 4 Oct 2022 13:20:54 +0530, Bharath Rupireddy
wrote in
> On Tue, Oct 4, 2022 at 12:11 PM Kyotaro Horiguchi
> wrote:
> >
> > >
> > > > static uint32 minXlogTli = 0;
> >
> > I have found other three instances of this in xlog.c and
> > pg_receivewal.c. Do they worth fixing?
> >
> >
On 30.09.22 06:07, Andres Freund wrote:
When tap tests are interrupted (e.g. with ctrl-c), we don't cancel running
postgres instances etc. That doesn't strike me as a good thing.
In contrast, the postgres instances started by pg_regress do terminate. I
assume this is because pg_regress starts
On 04.10.22 09:19, Bharath Rupireddy wrote:
We have the pg_attribute_unused() macro already. I'm not sure if
adding -Wunused-parameter for compilation plus using
pg_attribute_unused() for unused-yet-contextually-required variables
is a great idea. But it has some merits as it avoids unused
On Thu, Sep 29, 2022 at 12:50 PM shiy.f...@fujitsu.com
wrote:
>
> On Wed, Sep 28, 2022 1:49 PM Kyotaro Horiguchi
> wrote:
> >
> > At Wed, 28 Sep 2022 14:14:01 +1000, Peter Smith
> > wrote in
...
> > >
> > > 2. tab complete for GRANT
> > >
> > > test_pub=# grant
> > > ALL
I wanted to propose the attached patch to get rid of the custom pgpid_t
typedef in pg_ctl. Since we liberally use pid_t elsewhere, this seemed
plausible.
However, this patch fails the CompilerWarnings job on Cirrus, because
apparently under mingw, pid_t is "volatile long long int", so all
On Mon, Oct 03, 2022 at 04:41:11PM -0700, Andres Freund wrote:
> There's a few further roles that seem to pose some danger goign forward:
I have never seen that myself, but 0001 is a nice cleanup.
generated.sql includes a user named "regress_user11". Perhaps that's
worth renaming while on it?
>
On Tue, Oct 4, 2022 at 12:11 PM Kyotaro Horiguchi
wrote:
>
> >
> > > static uint32 minXlogTli = 0;
>
> I have found other three instances of this in xlog.c and
> pg_receivewal.c. Do they worth fixing?
>
> (pg_upgarade.c has "uint32 tli/logid/segno but I'm not sure they need
> to be "fixed". At
In PostgreSQL 10, we added identity columns, as an alternative to serial
columns (since 6.something). They mostly work the same. Identity
columns are SQL-conforming, have some more features (e.g., overriding
clause), and are a bit more robust in schema management. Some of that
was described
On Fri, Sep 30, 2022 at 7:48 PM Japin Li wrote:
>
> On Fri, 30 Sep 2022 at 22:09, Bharath Rupireddy
> wrote:
> > On Fri, Sep 30, 2022 at 7:30 PM Japin Li wrote:
> >>
> >> When I try to use -Wunused-parameter, I find there are many warnings :-( .
> >
> > Great!
> >
> > I think we can't just
I was wondering why we have a definition of Abs() in c.h when there are
more standard functions such as abs() and fabs() in widespread use. I
think this one is left over from pre-ANSI-C days. The attached patches
replace all uses of Abs() with more standard functions.
The first patch
On Tue, Oct 4, 2022 at 5:48 PM Michael Paquier wrote:
>
> On Tue, Oct 04, 2022 at 04:20:36PM +1100, Peter Smith wrote:
> > The v2 patches are updated as follows:
> >
> > 0001 - Now this patch only fixes a comment that had a wrong enum name.
>
> This was wrong, so fixed.
Thanks for pushing!
>
>
On Tue, Oct 4, 2022 at 12:00 PM Peter Eisentraut
wrote:
>
> On 16.05.22 10:27, Peter Eisentraut wrote:
> > Inspired by [0], I looked to convert more macros to inline functions.
>
> Here is another one from the same batch of work that I somehow didn't
> send in last time.
>
I think assertion can
On Tue, Oct 04, 2022 at 04:20:36PM +1100, Peter Smith wrote:
> The v2 patches are updated as follows:
>
> 0001 - Now this patch only fixes a comment that had a wrong enum name.
This was wrong, so fixed.
> 0002 - Removes unnecessary whitespace (same as v1-0002)
This one does not seem worth
At Tue, 04 Oct 2022 15:23:48 +0900 (JST), Kyotaro Horiguchi
wrote in
> This is not directly related to this patch, pg_resetwal.c has the
> following line..
>
> > static uint32 minXlogTli = 0;
I have found other three instances of this in xlog.c and
pg_receivewal.c. Do they worth fixing?
On 16.05.22 10:27, Peter Eisentraut wrote:
Inspired by [0], I looked to convert more macros to inline functions.
Here is another one from the same batch of work that I somehow didn't
send in last time.
(IMO it's questionable whether this one should be an inline function or
macro at all,
On Tue, Oct 04, 2022 at 03:17:06PM +0900, Kyotaro Horiguchi wrote:
> Nice finding. I found a few '%08X%08X's but they don't seem to fit
> similar fix.
Nice cleanup.
> Couldn't we use XLByteToSeg() here?
>
> Other than that, it looks good to me.
Yep. It looks that you're right here.
--
Michael
At Tue, 04 Oct 2022 15:17:06 +0900 (JST), Kyotaro Horiguchi
wrote in
> Other than that, it looks good to me.
Sorry I have another comment.
> - unsigned int tli,
> - log,
> - seg;
> +
On Tue, Oct 4, 2022 at 11:47 AM Kyotaro Horiguchi
wrote:
>
> > - segs_per_xlogid = (UINT64CONST(0x0001) /
> > ControlFile.xlog_seg_size);
> > newXlogSegNo = ControlFile.checkPointCopy.redo /
> > ControlFile.xlog_seg_size;
>
> Couldn't we use XLByteToSeg() here?
Yes, we
At Tue, 4 Oct 2022 11:06:15 +0530, Bharath Rupireddy
wrote in
> It looks like there's an opportunity to replace explicit WAL file
> parsing code with XLogFromFileName() in pg_resetwal.c. This was not
> done then (in PG 10) because the XLogFromFileName() wasn't accepting
> file size as an input
72 matches
Mail list logo
