On Feb 28, 2025 at 02:27 +0800, David G. Johnston ,
wrote:
> On Thu, Feb 6, 2025 at 8:04 AM Zhang Mingli wrote:
> > On Feb 6, 2025 at 10:49 +0800, David G. Johnston
> > , wrote:
> > >
> > > Yes, and when it does it is doing so in lieu of an error, and thus it
d I think it could confuse others as well if they
encounter something like that.
> Maybe someone has a different opinion about this, and we should disallow
> setting STORAGE on an fdw table in future versions.
If that’s the case, why not today? I’d be happy to take care of it.
--
Zhang Mingli
HashData
E command, tweaked a bit the part
> at the bottom where LIKE is not part of the standard.
>
> And applied.
Thanks.
--
Zhang Mingli
HashData
On Feb 18, 2025 at 09:54 +0800, Zhang Mingli , wrote:
>
> Will address the comments later, thanks for review!
Done in patch v6.
--
Zhang Mingli
HashData
v6-0001-CREATE-FOREIGN-TABLE-LIKE.patch
Description: Binary data
On Feb 10, 2025 at 16:45 +0800, Richard Guo , wrote:
> On Sun, Feb 9, 2025 at 11:15 PM Zhang Mingli wrote:
> > By reading the codes in nodeAgg.c, I found that some comments are outdated.
> > The comments referred to the function lookup_hash_entry(), which has been
> > remo
On Feb 18, 2025 at 09:54 +0800, Zhang Mingli , wrote:
>
> Will address the comments later, thanks for review!
Done in patch v6.
--
Zhang Mingli
HashData
v6-0001-CREATE-FOREIGN-TABLE-LIKE.patch
Description: Binary data
,
Good catch.
>
> Yeah, the primary key of t3 is deferred, so only the t3_c_uidx can be used.
> The test case is inconsistent with comments.
>
> It seems that the t2 does not have a unique index on z, do you forget to
> create it in the patch?
While there isn't an unique index, an unique constraint with a NOT NULL on that
column should serve the same purpose.
+alter table t2 alter column z set not null, add constraint t2_z_uidx unique
(z);
In patch v2, I made a slight adjustment: creating a unique index on t3 will
handle multiple choice scenarios.
Thanks for the patch.
--
Zhang Mingli
HashData
v2-0001-Modify-an-incorrect-regression-test-case-in-the-grou.patch
Description: Binary data
On Feb 18, 2025 at 08:49 +0800, Michael Paquier , wrote:
> On Mon, Feb 17, 2025 at 07:14:59PM +0800, Zhang Mingli wrote:
> > On Feb 17, 2025 at 15:24 +0800, Michael Paquier ,
> > wrote:
> > > + * For foreign tables, they have no storage in Postgres.
> > > + *
rts related to
> generated, default and constraints properties. This dependency
> between both routines should be documented, I guess. Hmm.. Not sure
> where.
Comments are addressed except for this one, it might be worth considering
another patch.
--
Zhang Mingli
HashData
v5-0001-CREATE-FOREIGN-TABLE-LIKE.patch
Description: Binary data
y is
> enough.
Hmm... I copied from the cases in the same file for each option.
There's no harm in having more tests, how about we keep them?
--
Zhang Mingli
HashData
v3-0001-CREATE-FOREIGN-TABLE-LIKE.patch
Description: Binary data
they don't
actually use storage.
Virtual columns are computed when the table is read, and they should adhere to
the same rules of join semantics.
I agree with Richard, the result seems incorrect. The right outcome should be:
gpadmin=# SELECT t2.a, t2.b FROM t t1 LEFT JOIN t t2 ON FALSE;
a | b
--+--
NULL | NULL
NULL | NULL
(2 rows)
--
Zhang Mingli
HashData
Hi,
By reading the codes in nodeAgg.c, I found that some comments are outdated.
The comments referred to the function lookup_hash_entry(), which has been
removed in commit b563594.
I've adjusted the comments to refer to the correct function to make it clear.
--
Zhang Mingli
HashData
v0
on the comments.
> >
> > instead of
> >
> > + Foreign tables have no real storage in PostgreSQL.
> >
> > Can it just be
> >
> > Foreign tables have no storage in PostgreSQL.
> >
> > the "real" is not needed.
Done.
Patch V2 addre
generation function during the rewrite stage
does not account for the Var field settings, leading to the errors we are
encountering.
--
Zhang Mingli
HashData
On Feb 7, 2025 at 22:24 +0800, Zhang Mingli , wrote:
> On Feb 6, 2025 at 18:09 +0800, Zhang Mingli , wrote:
> > On Feb 3, 2025 at 08:29 +0800, Michael Paquier , wrote:
> > > On Mon, Feb 03, 2025 at 06:22:13AM +0800, Mingli Zhang wrote:
> > > > Yes, I
On Feb 6, 2025 at 18:09 +0800, Zhang Mingli , wrote:
> On Feb 3, 2025 at 08:29 +0800, Michael Paquier , wrote:
> > On Mon, Feb 03, 2025 at 06:22:13AM +0800, Mingli Zhang wrote:
> > > Yes, I would like to provide a patch.
> > >
> > > Glad to see we have come t
we find any, the FREEZE operation won’t be applicable. However, that's a
consideration for the future.
LGTM.
--
Zhang Mingli
HashData
W will never
> produce NULL.
Hi,
OK, LGTM.
--
Zhang Mingli
HashData
rigin of the option bits.
--
Zhang Mingli
HashData
I believe that
COMPRESSION should also be copied for foreign tables, similar to STORAGE.
This might be beneficial for foreign data wrappers (FDWs) that take it into
account.
[0]
https://www.postgresql.org/message-id/6cecef0e-ee14-473c-bb0a-6aa61f539a66%40Spark
--
Zhang Mingli
HashData
v
could return NULL
actually.
gpadmin=# \pset null NULL
Null display is "NULL".
gpadmin=# select current_setting('pg_xmen', true);
current_setting
-
NULL
(1 row)
--
Zhang Mingli
HashData
tgresql.org/docs/current/sql-alterforeigntable.html
[2] https://www.postgresql.org/docs/current/sql-createforeigntable.html
--
Zhang Mingli
HashData
Hi,
Zhang Mingli
www.hashdata.xyz
On Feb 4, 2025 at 04:18 +0800, Sami Imseih , wrote:
>
> This really does not make sense as this
> optimization cannot be applied to a remote table and it
> can give a user a false impression that it was.
+1,
```
+ /*
+ * Raise an error on foreign ta
Zhang Mingli
www.hashdata.xyz
On Feb 1, 2025 at 20:20 +0800, Álvaro Herrera , wrote:
>
> Sure. Did you consider IMPORT FOREIGN SCHEMA?
Hi, Álvaro
Thank you very much for your suggestion.
I've looked into it, and it certainly can be beneficial, especially for
postgres_fdw.
However
E is not supported for creating
foreign tables")));
}
```
with some test cases and Documents changes.
Zhang Mingli
www.hashdata.xyz
Zhang Mingli
www.hashdata.xyz
On Jan 30, 2025 at 15:49 +0800, Matthias van de Meent
, wrote:
>
> Some time ago I noticed that every buffer table entry is quite large at 40
> bytes (+8): 16 bytes of HASHELEMENT header (of which the last 4 bytes are
> padding), 20 bytes of Buff
Hi
Zhang Mingli
www.hashdata.xyz
On Jan 16, 2025 at 14:04 +0800, Yugo Nagata , wrote:
>
> You will be able to log the plan during the REFRESH by using auto_explain and
> setting
> log_analyze and log_nested_statements to on.
Hi Yugo,
Thank you for your help! That’s certainly a via
execution of the REFRESH command.
Any suggestions?
Zhang Mingli
www.hashdata.xyz
ve it?
Zhang Mingli
www.hashdata.xyz
Hi,
Zhang Mingli
www.hashdata.xyz
On Oct 12, 2024 at 07:48 +0800, Joel Jacobson , wrote:
>
> Add missing checks for FORCE_NOT_NULL and FORCE_NULL when applied to
> all columns via "*". These options now correctly require CSV mode and
> are disallowed in COPY TO as a
Hi,
Zhang Mingli
www.hashdata.xyz
On Oct 9, 2024 at 20:35 +0800, Matthias van de Meent ,
wrote:
> Hi,
>
> Whilst doing some digging in parallel code, I noticed that
> max_parallel_maintenance_workers is registered as guc with a manual
> value of 1024, while max_parallel_workers_
Hi, all
I haven't paid attention to this topic in a long time, thanks all for the
advices, I will study them then update.
Thanks again.
Zhang Mingli
www.hashdata.xyz
On Sep 18, 2024 at 15:50 +0800, David Rowley , wrote:
> On Wed, 18 Sept 2024 at 19:28, Peter Eisentraut wrote:
&g
Hi,
Zhang Mingli
www.hashdata.xyz
On Jul 30, 2024 at 13:37 +0800, Kirill Reshke , wrote:
>
> Just to be clear, you are facing this on HEAD, on on REL_14_STABLE?
Postgres HEAD.
Hi,
Zhang Mingli
www.hashdata.xyz
On Jul 30, 2024 at 13:35 +0800, David Rowley , wrote:
>
> If you're only able
> to get it to misbehave by setting MAX_PARTITION_BUFFERS to less than
> 2, then my suggested fix would be to add a comment to say that values
> less than to are
NextFreeSlot() to avoid such issues.
[0] https://github.com/cloudberrydb/cloudberrydb
Zhang Mingli
www.hashdata.xyz
v0-0001-Fix-COPY-FROM-crash-due-to-buffer-flush.patch
Description: Binary data
also make it easy to cause
unnecessary troubles.
Shall we remove it before we provide an example for developers to avoid that?
Zhang Mingli
www.hashdata.xyz
in catalog when
bootstrap?
Thanks.
Zhang Mingli
www.hashdata.xyz
Hi,
Zhang Mingli
www.hashdata.xyz
On Jan 6, 2024 at 23:38 +0800, Geoff Winkless , wrote:
>
> I was hoping to see
>
> gp_n | gp_conc | n | concat
> --+-+--+
> 1 | 0 | NULL | n1x5
> 1 | 0 | NULL | n2x4
> 1 | 0 | NULL | n3x3
> 1 | 0 | NULL | n4x
Hi,
Zhang Mingli
www.hashdata.xyz
On Jan 6, 2024 at 01:38 +0800, Geoff Winkless , wrote:
>
> Am I missing some reason why the first set isn't sorted as I'd hoped?
Woo, it’s a complex order by, I try to understand your example.
And I think the order is right, what’s your expec
Hi,
Zhang Mingli
www.hashdata.xyz
On Jan 2, 2024 at 05:38 +0800, Tom Lane , wrote:
> I happened to notice that there is a not-quite-theoretical crash
> hazard in spcache_init(). If we see that SPCACHE_RESET_THRESHOLD
> is exceeded and decide to reset the cache, but then nsphash_creat
Hi,
Zhang Mingli
www.hashdata.xyz
Hi, Tomas
On Dec 31, 2023 at 07:10 +0800, Tomas Vondra ,
wrote:
> Sadly, there's no explanation why parallel scans do not allow disabling
> sync scans just like serial scans - and it's not quite obvious to me.
Feel confused too.
``
/CAApHDvrdYa%3DVhOoMe4ZZjZ-G4ALnD-xuAeUNCRTL%2BPYMVN8OnQ%40mail.gmail.com
Zhang Mingli
www.hashdata.xyz
v1-0001-Remove-useless-GROUP-BY-columns-considering-unique-i.patch
Description: Binary data
Hi, all
I see most xxxShmemInit functions have the logic to handle IsUnderPostmaster
env.
Do we need to consider it in DSMRegistryShmemInit() too? For example, add some
assertions.
Others LGTM.
Zhang Mingli
www.hashdata.xyz
On Dec 5, 2023 at 11:47 +0800, Nathan Bossart , wrote:
> Every o
Hi,
LGTM.
+ Assert(size >= sizeof(Node)); /* need the tag, at least */
+ result = (Node *) palloc0fast(size);
+ result->type = tag;
+ return result;
+}
How about moving the comments /* need the tag, at least */ after result->type =
tag; by the way?
Zha
default limit enum.
I remove it and have an install check to verify.
Are there any considerations behind this?
Shall we remove it for clear as it’s not actually the default option.
Zhang Mingli
www.hashdata.xyz
v1-0001-Remove-useless-LIMIT_OPTION_DEFAULT.patch
Description: Binary data
HI,
Zhang Mingli
www.hashdata.xyz
On Oct 31, 2023 at 03:55 +0800, Bruce Momjian , wrote:
>
> Sure, updated patch attached.
LGTM.
HI,
Zhang Mingli
www.hashdata.xyz
On Oct 30, 2023 at 10:58 +0800, Bruce Momjian , wrote:
> On Mon, Oct 30, 2023 at 05:07:48AM +0800, Mingli Zhang wrote:
> >
> > Bruce Momjian 于2023年10月30日周一03:35写道：
> >
> > On Sun, Oct 29, 2023 at 02:50:37PM +0800, Mingli Zhang wrote
wandering if it’s more reasonable to show rows_total instead of plan_rows for
Parallel Hash nodes?
For this example,
-> Parallel Hash (rows=2)
-> Parallel Seq Scan on simple s (rows=8333)
Zhang Mingli
HashData https://www.hashdata.xyz
usual, please check for problems such as wrong sorting, duplicate names
> in different variants, or names in the wrong order etc. (Our convention is
> given name followed by surname.)
>
>
Could you help me with Mingli Zhang -> Zhang Mingli
Thanks.
Zhang Mingli
HashData https://www.hashdata.xyz
HI,
> On Oct 1, 2023, at 22:54, Tom Lane wrote:
>
> For one example,
> you can't just remove the sort clause if the query uses DISTINCT ON
Hi, Tom, got it, thanks,
Zhang Mingli
HashData https://www.hashdata.xyz
mvt1_order;
REFRESH MATERIALIZED VIEW
Time: 204.398 ms
zml=# refresh materialized view mvt1_order;
REFRESH MATERIALIZED VIEW
Time: 197.510 ms
Zhang Mingli
www.hashdata.xyz
v1-01-Skip-Orderby-clause-execution-for-Materialized-Views.patch
Description: Binary data
://git.postgresql.org/gitweb/?p=pgtranslation/messages.git;a=commitdiff;h=14391f71ca61e90d52502093447fe1ee0080116f
>
> --
> Daniel Gustafsson
>
Thanks, got it~
Zhang Mingli
HashData https://www.hashdata.xyz
> On Aug 16, 2023, at 22:24, Peter Eisentraut wrote:
>
> On 16.08.23 09:34, Zhang Mingli wrote:
>> The Chinese words there are ok, but the `Unix-domian` should be
>> `Unix-domain`.
>
> fixed, thanks
>
Hi, Peter, thanks and just want to make sure tha
Hi, The Chinese words there are ok, but the `Unix-domian` should be `Unix-domain`.
v1-0001-Fix-typo-src-interfaces-libpq-po-zh_CN.po.patch
Description: Binary data
Zhang MingliHashData https://www.hashdata.xyz
On Aug 1, 2023, at 03:35, Andrew Dunstan wrote:I was hoping it be able to get to it today but that's not happening. If you want to submit a revised patch as above that will be good. I hope to get to it later this week.HI, Andrew Patch rebased and updated like above, thanks.
v4-0001-COPY-FROM-enab
Hi,
> On Jul 27, 2023, at 09:05, Michael Paquier wrote:
>
> One reason is that this increases the odds of
> conflicts when backpatching on a stable branch.
Agree. We could suggest to use OidIsValid() for new patches during review.
Zhang Mingli
https://www.hashdata.xyz
_custom.c: while (oid != 0)
```
That is another story…I would like provide a patch if it worths.
Zhang Mingli
https://www.hashdata.xyz
> Perhaps we could fix the inconsistency by changing the force_quote_all
> code to use MemSet() too. I'll defer whether to do that to Andrew's
> judgement.
Sure, let’s wait for Andrew and I will put everything in one pot then.
Zhang Mingli
https://www.hashdata.xyz
tter, thanks again.
Zhang Mingli
www.hashdata.xyz
HI,
On Jun 29, 2023 at 13:24 +0800, Nathan Bossart ,
wrote:
>
> Connecting to different databases with the same
> host/port/user information seems okay.
Have a look, yeah,
cluster_all_databases/vacuum_all_databases/reindex_all_databases will get there.
LGTM.
Regards,
Zhang Mingli
Hi,
On Jul 9, 2023 at 11:51 +0800, Zhang Mingli , wrote:
HI,
Regards,
Zhang Mingli
On Jul 7, 2023, 18:00 +0800, Damir Belyalov , wrote:
The patch does not work for the current version of postgres, it needs to be
updated.
I tested your patch. Everything looks simple and works well
Hi,
Regards,
Zhang Mingli
On Jul 17, 2023 at 21:09 +0800, Zhang Mingli , wrote:
> sequence_options
And inside pg_sequence_parameters:
pgstuple = SearchSysCache1(SEQRELID, relid);
Hi
Regards,
Zhang Mingli
On Jul 17, 2023 at 19:10 +0800, Michael Paquier , wrote:
> Hi all,
>
> While scanning the code, I have noticed that a couple of code paths
> that do syscache lookups are passing down directly Oids rather than
> Datums. I think that we'd better be cons
Hi,
Regards,
Zhang Mingli
On Jul 16, 2023 at 09:57 +0800, Andres Freund , wrote:
> Hi,
>
> Several loops which are important for query performance, like heapgetpage()'s
> loop over all tuples, have to call functions like
> HeapCheckForSerializableConflictOut() and Predica
HI,
Regards,
Zhang Mingli
On Jul 7, 2023, 18:00 +0800, Damir Belyalov , wrote:
>
> The patch does not work for the current version of postgres, it needs to be
> updated.
> I tested your patch. Everything looks simple and works well.
>
> There is a suggestion to simplify th
uces one output column for each of the listed column pairs (in
> the listed order), followed by any remaining columns from T1, followed by any
> remaining columns from T2.
Thanks for your help.
Regards,
Zhang Mingli
Hi
Regards,
Zhang Mingli
On Jun 28, 2023, 17:17 +0800, Julien Rouhaud , wrote:
> This is working as intended. When using a USING clause you "merge" both
> columns so the final target list only contain one version of the merged
> columns, which doesn't happen if you
ut: t1.c1, t1.c2
-> Hash (cost=0.02..0.02 rows=1 width=8)
Output: cte1.c1, cte1.c2
-> CTE Scan on cte1 (cost=0.00..0.02 rows=1 width=8)
Output: cte1.c1, cte1.c2
(14 rows)
Regards,
Zhang Mingli
oin t1 using(c1);
c1 | c2 | c1 | c2
+++
(0 rows)
Table t1 and t2 both has 2 columns: c1, c2, when CTE join select *, the result
target list seems to lost one’s column c1.
But it looks good when select cte1.* and t1.* explicitly .
Is it a bug?
Regards,
Zhang Mingli
L);
+ else if (childrel->pathlist != NIL &&
childrel->cheapest_total_path->param_info == NULL)
accumulate_append_subpath(childrel->cheapest_total_path,
&subpaths, NULL);
Could we also consider tuple_fraction in partial_pathlist for parallel append?
Regards,
Zhang Mingli
at t2 has empty rows but not null. Is it an expected behavior?
And what’s the semantic of SELECT INTO without any columns?
I also see lots of that SELECT INTO in out test cases like:
-- SELECT INTO doesn't support USING
SELECT INTO tableam_tblselectinto_heap2 USING heap2 FROM tableam_tbl_heap2;
Regards,
Zhang Mingli
Hi, all
Found several typos like plgsql, I think it should be plpgsql.
Regards,
Zhang Mingli
0001-Fix-typo-plgsql-to-plpgsql.patch
Description: Binary data
HI,
On Jan 16, 2023, 00:10 +0800, Nathan Bossart , wrote:
> On Mon, Jan 16, 2023 at 12:04:56AM +0800, Zhang Mingli wrote:
> > So, they are all dead codes, provide a patch to remove them.
>
> I am proposing a new use of dsa_create, dsa_attach, and dsa_get_ha
HI,
On Jan 15, 2023, 23:43 +0800, Zhang Mingli , wrote:
> Hi, hackers
>
> Found some functions in dsa.c are not used anymore.
>
> dsa_create
> dsa_attach
> dsa_get_handle
> dsa_trim
> dsa_dump
>
> We once used dsa_create to create DSA and it ’s all replace
.
dsa_trim and dsa_dump are introduced by DSA original commit 13df76a537 , but
not used since then.
So, they are all dead codes, provide a patch to remove them.
Regards,
Zhang Mingli
v0-0001-Remove-unused-functions-in-dsa.c.patch
Description: Binary data
Hi,
Regards,
Zhang Mingli
On Jan 12, 2023, 16:54 +0800, Richard Guo , wrote:
>
> On Thu, Jan 12, 2023 at 2:50 PM Zhang Mingli wrote:
> > On Jan 12, 2023, 14:34 +0800, Zhang Mingli , wrote:
> > > Some conditions in shm_toc_insert and shm_toc_allocate are bogus, like:
> >
= cstate->attnumlist;
else
attnums = CopyGetAttnums(tupDesc, cstate->rel, cstate->opts.force_notnull);
// process force_notnull columns
}
```
Any other suggestions?
Regards,
Zhang Mingli
Hi,
On Jan 12, 2023, 14:34 +0800, Zhang Mingli , wrote:
> Hi, hackers
>
> Some conditions in shm_toc_insert and shm_toc_allocate are bogus, like:
> if (toc_bytes + nbytes > total_bytes || toc_bytes + nbytes <
> toc_bytes)
> Remove the condition `toc_bytes +
se(&toc->toc_mutex);
shm_toc_freespace is introduced with shm_toc by original commit 6ddd5137b2, but
is not used since then, so remove it.
Regards,
Zhang Mingli
v0-0001-Fix-condition-in-shm_toc-and-remove-unused-functi.patch
Description: Binary data
Hi, hackers
Param RangeTblEntry *rte in function set_plain_rel_pathlist is not used at all.
I look at the commit e2fa76d80b(10 years ago), it’s useless since then.
Add a path to remove it.
Regards,
Zhang Mingli
v1-0001-Remove-unused-param-rte-in-set_plain_rel_pathlist.patch
Description
find a place to handle that option.
Do we miss something? ex: forget handle it in table_endscan.
Else, it’s not used at all.
The first commit introduced this option is c3b23ae457.
Other commits modify this option just changed the enum order.
Regards,
Zhang Mingli
h("Shared Buffer Lookup Table”
```
But it will help for other Databases built on Postgres or later Postgres in
case of forgetting to update SHMEM_INDEX_KEYSIZE
when new shmem added with a name longer than current SHMEM_INDEX_KEYSIZE.
And we don’t have such assertion now.
So, +1 for the patch.
Regards,
Zhang Mingli
HI,
Regards,
Zhang Mingli
On Nov 7, 2022, 14:26 +0800, Tom Lane , wrote:
> Andrey Lepikhov writes:
> > I'm still in review of your patch now. At most it seems ok, but are you
> > really need both eq_sources and eq_derives lists now?
>
> Didn't we just have this co
that it's not reached at all in your example, while the
> regression tests give
>
> 49 LOG: freeing list of length 0
> 2 LOG: freeing list of length 1
>
Thanks for the investigation.
Yeah, this patch is negligible. I’ll withdraw it in CF.
Regards,
Zhang Mingli
e->oidCount?
>
> If the lock is unnecessary, I think adding some comments is better.
>
> --
> Regrads,
> Japin Li.
> ChengDu WenWu Information Technology Co.,Ltd.
>
>
As its name BootStrapXLOG, it’s used in BootStrap mode to initialize the
template database.
The process doesn’t speak SQL and the database is not ready.
There won’t be concurrent access to variables.
Regards,
Zhang Mingli
>
Hi,
On Oct 27, 2022, 21:29 +0800, Tom Lane , wrote:
> Zhang Mingli writes:
> > How about combine ec->ec_sources and ec->derives as one list for less codes?
>
> Keeping them separate is required for the broken-EC code paths.
> Even if it weren't, I wouldn't mer
_union(ec->ec_sources, ec->ec_derives))
{
rinfo = (RestrictInfo *) lfirst(lc);
if (rinfo->left_em == leftem &&
rinfo->right_em == rightem &&
rinfo->parent_ec == parent_ec)
return rinfo;
if (rinfo->left_em == rightem &&
rinfo->right_em == leftem &&
rinfo->parent_ec == parent_ec)
return rinfo;
}
```
I have a try, it will change some in join.out and avoid changes in tidscan.out.
Regards,
Zhang Mingli
>
Hi, Andres
On Oct 24, 2022, 01:16 +0800, Andres Freund , wrote:
> Hi,
>
> On 2022-10-22 11:32:47 +0800, Zhang Mingli wrote:
> > Hi, hackers
> >
> > I don't quite understand FullTransactionIdAdvance(), correct me if I’m
> > wrong.
> >
> >
Another confusion is the comments: /* see FullTransactionIdAdvance() */, is its
own itself.
Could anyone explain this? Thanks in advance.
Regards,
Zhang Mingli
make it more "tidy".
> >
> > Reviews and opinions are very welcome!
> >
> > --
> > Best regards,
> > Maxim Orlov.
Found some outdate code comments around several variables, such as
xidWrapLimit/xidWarnLimit/xidStopLimt.
These variables are not used any more.
I
ommit 56788d2156, each parallel worker will try to get ranges of
blocks “chunks".
Access to the table remains sequential inside each worker’s process, but not
across all workers or the parallel query.
Shall we update the documents?
Regards,
Zhang Mingli
e the first XLogRegisterBuffer() call.
>
> Would any of you like to write a patch?
> --
> Michael
Patch added.
Regards,
Zhang Mingli
v2-0001-Fix-GIN-fast-path-XLogBeginInsert-and-Read-LockBu.patch
Description: Binary data
also found some places where could use fabsf instead of fabs if
possible, add a patch to replace them.
Regards,
Zhang Mingli
v2-0005-replace-fabs-with-fabsf-if-possible.patch
Description: Binary data
CRIT_SECTION();
```
Shall we adjust that too?
Regards,
Zhang Mingli
> PFA a patch that rectifies this issue, by moving the XLogBeginInsert()
> down to where 1.) we have all relevant buffers pinned and locked, and
> 2.) we're in a critical section, making that part of the code
> consistent with the general scheme for XLog insertion.
+1, Make sense.
Regards,
Zhang Mingli
t; Thanks,
> Melih
Looks good to me.
Regards,
Zhang Mingli
ntry()? From my quick skim of the code, it
> seems like it should be safe, but I thought I'd ask the question.
Same question, have a look, it doesn’t seem to matter.
Regards,
Zhang Mingli
Hi,
On Sep 28, 2022, 22:41 +0800, Melih Mutlu , wrote:
>
>
> Zhang Mingli , 28 Eyl 2022 Çar, 17:31 tarihinde şunu
> yazdı:
> > Why compute usagecount_avg twice?
>
> I should have removed the first one, but I think I missed it.
> Nice catch.
>
> Attached an upd
Regards,
Zhang Mingli
On Sep 28, 2022, 21:50 +0800, Melih Mutlu , wrote:
> Hi all,
>
> The patch needed a rebase due to recent changes on pg_buffercache.
> You can find the updated version attached.
>
> Best,
> Melih
>
>
```
+
+ if (buffers_used != 0)
+ usag
Regards,
Zhang Mingli
On Sep 27, 2022, 00:47 +0800, Ashutosh Sharma , wrote:
>
> And further it looks like you are doing an
> experiment on undamaged relation which is not recommended as
> documented.
Yeah.
> If the relation would have been damaged, you probably may
> not be
1 - 100 of 140 matches
Mail list logo
