in
MySQL 8. The TPC-C results shown are impressive.
--
Thomas Munro
http://www.enterprisedb.com
On Tue, Jan 30, 2018 at 6:48 PM, Tatsuo Ishii wrote:
> Looks good to me. If there's no objection, especially from Thomas
> Munro, I will mark this as "ready for committer".
No objection from me.
--
Thomas Munro
http://www.enterprisedb.com
nt, as yet
unexplained, that the query couldn't be interrupted:
https://www.postgresql.org/message-id/flat/151724453314.1238.409882538067070269%40wrigleys.postgresql.org
--
Thomas Munro
http://www.enterprisedb.com
e able to shift the data to the right place successfully.
For example I think I'd want to see that a single aborted transaction
surrounded by many committed transactions shows up in the right place
after an upgrade.
--
Thomas Munro
http://www.enterprisedb.com
with-llvm LLVM_CONFIG=~/build/llvm/5/opt/install/bin/llvm-config
> is what I use, although I also add:
> LDFLAGS='-Wl,-rpath,/home/andres/build/llvm/5/opt/install/lib'
> so I don't have to install llvm anywhere the system knows about.
BTW if you're building w
ctory
$ mv ~/install/lib/llvmjit_types.bc ~/install/lib/postgresql/
postgres=# select 4 + 4;
?column?
--
8
(1 row)
--
Thomas Munro
http://www.enterprisedb.com
On Fri, Feb 2, 2018 at 5:11 PM, Thomas Munro
wrote:
> Another small thing which might be environmental... llvmjit_types.bc
> is getting installed into ${prefix}/lib here, but you're looking for
> it in ${prefix}/lib/postgresql:
Is there something broken about my installation?
s
+}
+#endif
#include
+#ifdef __cplusplus
+extern "C"
+{
+#endif
... in llvmjit.h.
--
Thomas Munro
http://www.enterprisedb.com
to populated table
ALTER TABLE itest13 ADD COLUMN c int GENERATED BY DEFAULT AS IDENTITY;
+ ERROR: column "c" contains null values
SELECT * FROM itest13;
! a | b
! ---+---
! 1 | 1
! 2 | 2
! 3 | 3
(3 rows)
-- various ALTER COLUMN tests
--
Thomas Munro
http://www.enterprisedb.com
an see from
the backtrace[1] that ExecInitExprRec is receiving a null node pointer
on this Ubuntu Trusty GCC 4.8 amd64 system.
[1] https://travis-ci.org/postgresql-cfbot/postgresql/builds/337255374
--
Thomas Munro
http://www.enterprisedb.com
g this.
+ PGTYPES_free()/ instead of free().
The "/" needs to move inside then closing tag.
--
Thomas Munro
http://www.enterprisedb.com
d. The leader process errors out and detaches
from the reference-counted DSM segment before the worker processes
manage to start up, and when they try to attach they can't because the
DSM segment is gone. This happens with anything that errors out very
quickly in the leader, like SELECT COUNT(r.n/0) FROM r using your
example table.
--
Thomas Munro
http://www.enterprisedb.com
= NULL;
/* force symbols in main binary to be loaded */
- LLVMLoadLibraryPermanently("");
+ LLVMLoadLibraryPermanently(NULL);
llvm_opt0_orc = LLVMOrcCreateInstance(llvm_opt0_targetmachine);
llvm_opt3_orc = LLVMOrcCreateInstance(llvm_opt3_targetmachine);
--
Thomas Munro
http://www.enterprisedb.com
ight where we call disable_all_timeouts().
See attached.
--
Thomas Munro
http://www.enterprisedb.com
fix-statement-timeout.patch
Description: Binary data
mpiler and library support for the ISO C++ 2011
> standard. This support must be enabled with the -std=c++11 or
> -std=gnu++11 compiler options.
Did you try passing CXXFLAGS="-std=c++11" to configure?
--
Thomas Munro
http://www.enterprisedb.com
ma, the host, and the port fields").
I see there is another scheme called "cldap" (which seems to be
something like LDAP over UDP). I wonder if anyone cares about that.
--
Thomas Munro
http://www.enterprisedb.com
ight also be a good idea if you could tell 001_auth.pl where slapd
and openssl (used by 001_auth.pl to make certificates) are through
environment variables, in case anyone wants to run those tests against
an installation other than one in the hardcoded paths we told it about
for Linux, macOS (with Brew) and FreeBSD.
--
Thomas Munro
http://www.enterprisedb.com
On Sat, Feb 10, 2018 at 4:32 PM, Thomas Munro
wrote:
> I agree that it would be nice if the build farm (and my unofficial
> patch tester for that matter) could automatically test the LDAP stuff
> when running on a suitable system, but I think it would need to be
> based not
into ExecParallelHashTupleAlloc() after the
lock is acquired (the slow path where a new 32KB chunk is needed) and
run some parallel hash joins, then you'll see that pg_stat_activity
reports that as "extension" in unpatched master.
--
Thomas Munro
http://www.enterprisedb.com
0001-R
On Mon, Sep 4, 2017 at 2:18 PM, Amit Kapila wrote:
> On Sun, Sep 3, 2017 at 2:56 PM, Thomas Munro
> wrote:
>> I think it should be (size_t) 1, not UINT64CONST(1). See attached.
>
> Okay, that makes sense. Do you think we should also change type
> casting in BUCKETS_PER_PAR
On Sun, Apr 30, 2017 at 1:19 PM, Tom Lane wrote:
> Thomas Munro writes:
>> I was reading xact.c and noticed this block:
>> ...
>> Isn't this insufficient on non-TSO systems like POWER and Arm?
>
> Yeah, I think you're right. That code probably predates
If so I don't see it.
--
Thomas Munro
http://www.enterprisedb.com
0001-Remove-volatile-qualifiers-from-shm_mq.c.patch
Description: Binary data
ed in monitoring.sgml. I say that
as someone who recently screwed up and forgot to register the display
name of a new built-in tranche[1]...
[1]
https://www.postgresql.org/message-id/CAEepm%3D1WM%3D02cKT7hinnx8xRhN2QdF%2BWfE6gk5oMPorwLi%2BXiA%40mail.gmail.com
--
Thomas Munro
http://www.enterprisedb.com
except
footprint. In ancient DB2 they had a column property along those
lines called "VALUE COMPRESSION" (they also have a row-level version,
and now they have much more advanced kinds of adaptive compression
that I haven't kept up with). In some ways it'd be a bit like toast
with shared entries, but I haven't seriously looked into how such a
thing might be implemented.
--
Thomas Munro
http://www.enterprisedb.com
On Tue, Feb 13, 2018 at 9:09 AM, Robert Haas wrote:
> On Sun, Feb 11, 2018 at 6:52 PM, Thomas Munro
> wrote:
>> As far as I can see, all the volatile qualifiers in shm_mq.c have been
>> redundant since ec9037df263.
>
> That's the commit that added shm_mq.c. ITYM
ould be something like pseudo page number = hash
value % 1024. In other words, you'd use the full hash value for the
'tuple' part of the predicate lock tag, and a shorter hash value for
the 'page' part of the predicate lock tag, so you'd never have to
handle split, and p
On Tue, Oct 29, 2019 at 12:15 PM Andres Freund wrote:
> I've groused about this a few times, but to me it seems wrong that
> HashJoin and Hash are separate nodes. They're so tightly bound together
> that keeping them separate just doesn't architecturally makes sense,
> imo. So I wrote a prototype.
On Tue, Oct 29, 2019 at 1:52 AM Peter Eisentraut
wrote:
> As mentioned in [0], pg_upgrade currently does not preserve the version
> of collation objects created by initdb. Here is an attempt to fix that.
>
> The way I deal with this here is by having the binary-upgrade mode in
> pg_dump delete al
On Wed, Oct 30, 2019 at 9:25 AM Tom Lane wrote:
> What I'm inclined to do is go file a bug report saying that this
> behavior contradicts both POSIX and NetBSD's own man page, and
> see what they say about that.
>From a quick look at the relevant trees, isn't the problem here that
cpython thinks
On Mon, Oct 28, 2019 at 3:33 PM Craig Ringer wrote:
> I was recently surprised to notice that log_line_prefix doesn't support a
> cluster_name placeholder. I suggest adding one. If I don't hear objections
> I'll send a patch.
+1
Hello hackers,
Please feel free to edit this new page, which I'd like to use to keep
track of observations, ideas and threads relating to hash joins.
https://wiki.postgresql.org/wiki/Hash_Join
On Fri, Nov 1, 2019 at 7:42 PM Zhenghua Lyu wrote:
> It is tricky to set disable_cost a huge number. Can we come up with
> better solution?
What happens if you use DBL_MAX?
On Sat, Nov 2, 2019 at 6:15 AM Tomas Vondra
wrote:
> On Fri, Nov 01, 2019 at 12:05:12PM +0300, Павел Ерёмин wrote:
> > Hi.
> > sorry for my English.
> > I want to once again open the topic of 64 bit transaction id. I did not
> > manage to find in the archive of the option that I want to di
On Tue, Oct 29, 2019 at 5:23 PM btfujiitkp wrote:
> > Thomas Munro writes:
> >> On Sun, Sep 1, 2019 at 5:04 PM Thomas Munro
> >> wrote:
> >>> Adding to CF.
> >
> >> Rebased. An OID clashed so re-roll the dice. Also spotted a typo.
>
On Tue, Sep 10, 2019 at 11:43 AM Thomas Munro wrote:
> Rebased due to bitrot. Spotted one more place to use this, in
> src/backend/utils/adt/txid.c.
Rebased. I'm planning to commit this soon.
0001-Consolidate-code-that-makes-a-sorted-array-unique-v3.patch
Description: Binary data
On Thu, Oct 31, 2019 at 5:24 AM Melanie Plageman
wrote:
> On Wed, Oct 23, 2019 at 12:30 AM Thomas Munro wrote:
>> Of course there are some more things that could be reported in a
>> similar way eventually, such as filter counters and hash join details.
>
> This made me thin
On Mon, Nov 4, 2019 at 12:11 PM Thomas Munro wrote:
> I guess I thought of that as a debugging feature and took it out
> because it was too verbose, but maybe it just needs to be controlled
> by the VERBOSE switch. Do you think we should put that back?
By which I mean: would you like
On Wed, Oct 16, 2019 at 6:49 PM Dave Cramer wrote:
> Here's an updated patch that addresses some of Andres' concerns specifically
> does not use strtok.
Hi Dave,
I think you need to s/strncasecmp/pg_strncasecmp/ to make this build on Windows.
https://ci.appveyor.com/project/postgresql-cfbot/po
On Sat, Nov 2, 2019 at 10:38 AM Fabien COELHO wrote:
> Attached patch v4 does it.
Hi Fabien,
It looks like don't define sigint_interrupt_jmp and
sigint_interrupt_enabled on Windows, yet they are still declared and
referenced?
command.obj : error LNK2001: unresolved external symbol
sigint_interr
On Thu, Aug 1, 2019 at 6:51 PM Noah Misch wrote:
> vac_truncate_clog() instance 1 starts, considers segment ABCD eligible to
> unlink
> vac_truncate_clog() instance 2 starts, considers segment ABCD eligible to
> unlink
> vac_truncate_clog() instance 1 unlinks segment ABCD
> vac_truncate_clog() i
On Thu, Oct 17, 2019 at 3:22 PM Smith, Peter wrote:
> We have resurrected this 2 year old "SQL statements statistics counter"
> proposal from Hari.
>
> The attached patch addresses the previous review issues.
Hi,
No comment on the patch but I noticed that the documentation changes
don't build.
On Thu, Oct 31, 2019 at 3:03 AM Dave Cramer wrote:
> Ok, I've rebased and reverted logicalrep_read_insert
Hi Dave,
>From the code style police (actually just from cfbot, which is set up
to complain about declarations after statements, a bit of C99 we
aren't ready for):
proto.c:557:6: error: ISO
On Fri, Nov 1, 2019 at 2:21 AM Julien Rouhaud wrote:
> Are you planning to continue working on it? For the record, that's
> something needed to be able to implement a filter in REINDEX command
> [1].
Bonjour Julien,
Unfortunately I haven't had time to work on it seriously, but here's a
quick re
On Tue, Nov 5, 2019 at 8:45 AM Tomas Vondra
wrote:
> On Mon, Nov 04, 2019 at 10:44:53AM -0800, Andres Freund wrote:
> >On 2019-11-04 19:39:18 +0100, Tomas Vondra wrote:
> >> On Mon, Nov 04, 2019 at 10:04:09AM -0800, Andres Freund wrote:
> >> > And "without causing significant issues elsewhere" unf
On Tue, Nov 5, 2019 at 4:18 AM Julien Rouhaud wrote:
> On Mon, Nov 4, 2019 at 11:13 AM Julien Rouhaud wrote:
> > On Mon, Nov 4, 2019 at 4:58 AM Thomas Munro wrote:
> > > Here are some problems to think about:
> > >
> > > * We'd need to track depe
On Mon, Nov 4, 2019 at 11:13 PM Julien Rouhaud wrote:
> On Mon, Nov 4, 2019 at 4:58 AM Thomas Munro wrote:
> > * Some have expressed doubt that pg_depend is the right place for
> > this; let's see if any counter-proposals appear.
>
> When working on the REINDEX FIL
On Sat, Nov 2, 2019 at 6:57 AM Andrew Dunstan
wrote:
> On 11/1/19 11:01 AM, Robert Haas wrote:
> > On Thu, Oct 31, 2019 at 11:37 AM Andrew Dunstan
> > wrote:
> >> This patch provides a hook for a function that can supply an SSL
> >> passphrase. The hook can be filled in by a shared preloadable mo
On Thu, Sep 26, 2019 at 8:05 AM Alvaro Herrera wrote:
> This patch seems to be failing the contrib build. Please fix.
Hi Pierre,
In contrib/pg_stat_statements/pg_stat_statements.c, you need to
declare or define entry_reset_computed() before you use it. I suppose
your compiler must be warning a
On Mon, Nov 4, 2019 at 11:13 PM Julien Rouhaud wrote:
> When working on the REINDEX FILTER, I totally missed this thread and
I created a new wiki page to try to track the various moving pieces
here. Julien, Peter, Christoph, anyone interested, please feel free
to update it or add more informatio
On Sat, Oct 19, 2019 at 11:52 PM Tomas Vondra
wrote:
> I wonder if an extension could do something like that, though. It can
> install a hook after parse analysis, so I guess it could walk the CTEs
> and mark them as materialized.
I wonder if the existing pg_hint_plan extension could be extended
On Thu, Sep 12, 2019 at 10:10 AM Alvaro Herrera
wrote:
> Here's a rebased version of this patch (it had a trivial conflict).
Hi, FYI partition_prune.sql currently fails (maybe something to do
with commit d52eaa09?):
where s.a = $1 and s.b = $2 and s.c = (select 1);
explain (costs off) execute
On Wed, Nov 6, 2019 at 3:33 PM Yuya Watari wrote:
> However, this behavior depends on the platform architecture. As you
> have said, C language does not always follow IEEE-754. I think adding
> explicit checking of NaN is necessary.
I'm curious about this point. C may not require IEEE 754 (for
e
On Thu, Nov 7, 2019 at 10:43 AM Tom Lane wrote:
> =?UTF-8?Q?Juan_Jos=C3=A9_Santamar=C3=ADa_Flecha?=
> writes:
> > Please find attached a patch that adds the FILE_SHARE options to
> > TestLib::slurp_file using Win32API::File.
>
> Ick. Are we going to need Windows-droppings like this all over the
Hello hackers,
Here's a draft patch that teaches PostgreSQL how to ask for collation
versions on Windows. It receives a pair of DWORDs, which, it displays
as hex. The values probably have an internal structure that is
displayed in a user-friendly way by software like Active Directory and
SQL Ser
On Thu, Nov 7, 2019 at 10:20 PM Julien Rouhaud wrote:
> Attached 4th patch handles default collation. I went with an
> ignore_systempin flag in recordMultipleDependencies.
Thanks for working on this! I haven't looked closely or tested yet,
but this seems reasonable. Obviously it assumes that t
On Thu, Nov 7, 2019 at 1:27 AM Peter Eisentraut
wrote:
> As I was working on that lately, I came to the conclusion that we should
> get *this* patch done first.
Cool. Let's aim to get this into 13!
> > * Some have expressed doubt that pg_depend is the right place for
> > this; let's see if any
On Fri, Nov 8, 2019 at 2:37 PM Michael Paquier wrote:
> On Fri, Nov 08, 2019 at 02:23:54PM +1300, Thomas Munro wrote:
> > Right, so this is basically a policy decision: do we assume that all
> > pre-13 indexes that depend on collations are potentially corrupted, or
> > ass
On Thu, Nov 7, 2019 at 11:37 PM Rafia Sabih wrote:
> I was reviewing this patch and here are a few comments,
Hi Rafia,
Thanks!
> +static void
> +ExplainNodePerProcess(ExplainState *es, bool *opened_group,
> + int worker_number, Instrumentation *instrument)
> +{
>
> A small description about th
On Thu, Aug 8, 2019 at 6:53 AM Ashwin Agrawal wrote:
>>> - I wonder if CheckForSerializableConflictOutNeeded() shouldn't have a
>>> portion of it's code as a static inline. In particular, it's a shame
>>> that we currently perform external function calls at quite the
>>> frequency when seria
Some more thoughts:
1. If you create an index on an expression that includes a COLLATE or
a partial index that has one in the WHERE clause, you get bogus
warnings:
postgres=# create table t (v text);
CREATE TABLE
postgres=# create index on t(v) where v > 'hello' collate "en_NZ";
WARNING: index
On Sat, Nov 9, 2019 at 10:20 AM Juan José Santamaría Flecha
wrote:
>> Do any Windows hackers want to help get it into shape? Some things to
>> do: test it, verify that the _WIN32_WINNT >= 0x0600 stuff makes sense
>> (why do we target such ancient Windows releases anyway?)
>
> You have to keep in
On Sat, Nov 9, 2019 at 8:41 AM Ashwin Agrawal wrote:
> On Thu, Nov 7, 2019 at 8:44 PM Thomas Munro wrote:
>> I'm planning to commit these three patches on Monday. I've attached
>> versions with whitespace-only changes from pgindent, and commit
>> messages lightl
On Tue, Nov 12, 2019 at 1:24 AM Kyotaro Horiguchi
wrote:
> Hello. While looking a patch, I found that PHJ sometimes complains for
> file leaks if accompanied by LIMIT.
Oops.
> Repro is very simple:
>
> create table t as (select a, a as b from generate_series(0, 99) a);
> analyze t;
> select
o be in
> ExecutePlan, not in PHJ.
You are right. Here is the email I just wrote that says the same
thing, but with less efficiency:
On Tue, Nov 12, 2019 at 11:24 AM Tom Lane wrote:
> Thomas Munro writes:
> > On Tue, Nov 12, 2019 at 1:24 AM Kyotaro Horiguchi
> > wrote:
>
On Tue, Nov 12, 2019 at 4:23 PM Thomas Munro wrote:
> On Tue, Nov 12, 2019 at 4:20 PM Kyotaro Horiguchi
> wrote:
> > At Mon, 11 Nov 2019 17:24:45 -0500, Tom Lane wrote in
> > > Although the patch seems unobjectionable as far as it goes, I'd like
> > > to un
On Tue, Nov 12, 2019 at 5:03 PM Thomas Munro wrote:
> On Tue, Nov 12, 2019 at 4:23 PM Thomas Munro wrote:
> > On Tue, Nov 12, 2019 at 4:20 PM Kyotaro Horiguchi
> > wrote:
> > > The previous patch would be wrong. The root cause is a open batch so
> > > the righ
On Wed, Nov 13, 2019 at 3:27 AM Julien Rouhaud wrote:
> On Sun, Nov 10, 2019 at 10:08 AM Thomas Munro wrote:
> > That's because the 0003 patch only calls recordDependencyOnVersion()
> > for simple attribute references. When
> > recordDependencyOnSingleRelExpr() is c
Hello hackers,
>From the advanced bikeshedding department: I'd like my psql
transcripts to have the usual alignment, but be easier to copy and
paste later without having weird prompt stuff in the middle. How
about a prompt format directive %w that means "whitespace of the same
width as %/"? Then
On Wed, Nov 13, 2019 at 9:52 PM Amit Kapila wrote:
> On Wed, Nov 13, 2019 at 6:13 AM Kyotaro Horiguchi
> wrote:
> > The phatch's shape looks better. Thanks.
>
> +1. LGTM as well.
Thanks. Pushed.
Hello hackers,
I don't like this:
-> Parallel Hash (... rows=416667 ...) (... rows=33 ...)
I think the logic in get_parallel_divisor() only makes sense for
queries like this (output with nearby patch to show leader
contribution):
postgres=# set parallel_tuple_cost = 0;
SET
postgres=# set
On Sat, Sep 28, 2019 at 4:20 AM Bruce Momjian wrote:
> On Wed, Sep 4, 2019 at 06:18:31PM +1200, Thomas Munro wrote:
> > A few years back[1] I experimented with a simple readiness API that
> > would allow Append to start emitting tuples from whichever Foreign
> > Scan has
On Fri, Nov 15, 2019 at 3:58 AM Tom Lane wrote:
> Kyotaro Horiguchi writes:
> > This seems assuming %x are a kind of stable (until semicolon)
> > function. But at least %`..` can be volatile. So, I think the %w
> > thing in PROMPT2 should be able to refer the actual prompt string
> > resulted fr
On Mon, Nov 18, 2019 at 1:49 PM Alvaro Herrera wrote:
> On 2019-Nov-18, Thomas Munro wrote:
> > Nice idea. Here's one like that, that just does the counting at the
> > end and looks out for readline control codes. It's pretty naive about
> > what "width"
On Tue, Nov 19, 2019 at 6:21 AM Tom Lane wrote:
> Thomas Munro writes:
> > Yeah. Maybe pg_wcswidth() would be OK though, and it's available in
> > psql, though I guess you'd have to make a copy with the escaped bits
> > stripped out.
>
> Right, you shoul
On Tue, Nov 19, 2019 at 12:09 PM Tom Lane wrote:
> You should follow the logic in pg_wcswidth: compute PQmblen() first,
> and bail out if it's more than the remaining string length, otherwise
> it's ok to apply PQdsplen().
Got it. I was worried that it wasn't safe to call even PQmblen(),
because
On Mon, Nov 18, 2019 at 7:48 PM Jinbao Chen wrote:
> In the test case above, the small table has 3000 tuples and 100 distinct
> values on column ‘a’.
> If we use small table as inner table. The chan length of the bucket is 30.
> And we need to
> search the whole chain on probing the hash table.
On Wed, Nov 20, 2019 at 12:28 AM Amit Khandekar wrote:
> On Tue, 19 Nov 2019 at 14:07, Amit Kapila wrote:
> > No, I got this before applying the patch. However, after applying the
> > patch, I got below error in the same test:
> >
> > postgres=# SELECT 1 from
> > pg_logical_slot_get_changes('reg
On Wed, Nov 20, 2019 at 1:14 AM Juan José Santamaría Flecha
wrote:
> On Tue, Nov 19, 2019 at 12:49 PM Thomas Munro wrote:
>> On Wed, Nov 20, 2019 at 12:28 AM Amit Khandekar
>> wrote:
>> > On Windows, it is documented that ReadFile() (which is called by
>> > p
On Wed, Nov 20, 2019 at 7:58 AM Thomas Munro wrote:
> On Wed, Nov 20, 2019 at 1:14 AM Juan José Santamaría Flecha
> > https://devblogs.microsoft.com/oldnewthing/20150121-00/?p=44863
>
> !?!
One thing I don't understand (besides, apparently, the documentation):
how did
by signal 6: Aborted
>
> /dev/vdb on /var/lib/pgsql type ext4 (rw,relatime,seclabel,data=ordered)
> Centos 7.7 qemu/KVM
> Linux database 3.10.0-1062.1.1.el7.x86_64 #1 SMP Fri Sep 13 22:55:44 UTC 2019
> x86_64 x86_64 x86_64 GNU/Linux
> There's no added tablespaces.
>
&g
On Wed, Nov 20, 2019 at 4:54 PM Amit Khandekar wrote:
> - BufFileLoadBuffer() seems to deliberately ignore FileRead()'s return
> value if it is -1
> if (file->nbytes < 0) file->nbytes = 0;
Ok, that's a different problem we need to fix then. But it does
explain how we didn't know. And sure e
On Wed, Nov 20, 2019 at 5:44 PM Amit Kapila wrote:
> On Wed, Nov 20, 2019 at 12:28 AM Thomas Munro wrote:
> > + if (GetLastError() == ERROR_HANDLE_EOF)
> > + return 0;
> Yes, this works for me.
Thanks, pushed.
On Wed, Nov 20, 2019 at 5:43 PM imai.yoshik...@fujitsu.com
wrote:
> Is there any agreement we can throw the wraparound problem away if we adopt
> FullTransactionId?
Here is one argument for why 64 bits ought to be enough: we use 64 bit
LSNs for the WAL, and it usually takes more than one byte of
On Fri, Nov 8, 2019 at 3:41 AM Andrew Dunstan
wrote:
> On 11/7/19 9:12 AM, Alvaro Herrera wrote:
> >>
> >> The patch says:
> >>
> >> +require Win32::API;
> >> +Win32::API->import;
> > Oh, you're right, it does. I wonder why, though:
> >
>
> On further inspection I think those line
Hi,
As noted by Amit Khandhekar yesterday[1], BufFileLoad() silently eats
pread()'s error and makes them indistinguishable from EOF.
Some observations:
1. BufFileRead() returns size_t (not ssize_t), so it's an
fread()-style interface, not a read()-style interface that could use
-1 to signal err
On Thu, Nov 21, 2019 at 10:50 AM Thomas Munro wrote:
> As noted by Amit Khandhekar yesterday[1], BufFileLoad() silently eats
Erm, Khandekar, sorry for the extra h!
On Sun, Nov 17, 2019 at 6:36 AM vignesh C wrote:
> I noticed that some of the source files does not include the copyright
> information. Most of the files have included it, but few files have
> not included it. I felt it should be included. The attached patch
> contains the fix for including the c
On Mon, Nov 25, 2019 at 8:53 AM Andreas Karlsson wrote:
> When reading the syntax documentation for SELECT I noticed a missing
> space in "[ OVERRIDING { SYSTEM | USER} VALUE ]".
Right. Pushed.
On Tue, Nov 26, 2019 at 5:21 PM Justin Pryzby wrote:
> I looked and found a new "hint".
>
> On Tue, Nov 19, 2019 at 05:57:59AM -0600, Justin Pryzby wrote:
> > < 2019-11-15 22:16:07.098 EST >PANIC: could not fsync file
> > "base/16491/1731839470.2": No such file or directory
> > < 2019-11-15 22:
On Fri, Nov 8, 2019 at 12:44 PM Thomas Munro wrote:
> The reason for returning an empty string for "C" and "POSIX" is the
> following comment for get_collation_actual_version():
>
> * A particular provider must always either return a non-NULL string or return
On Fri, Nov 8, 2019 at 5:40 PM Laurenz Albe wrote:
> On Fri, 2019-11-08 at 15:04 +1300, Thomas Munro wrote:
> > 3. We don't know if pre-13 indexes are corrupted or not, and we'll
> > record that with a special value just as in proposal #1, except that
> > we could
On Thu, Nov 28, 2019 at 3:45 PM Michael Paquier wrote:
> On Tue, Sep 17, 2019 at 10:03:20AM +1200, Thomas Munro wrote:
> > Oops, right. So it should just be added to the if condition. Will do.
>
> It's been a couple of months and the discussion has stale. It seems
>
On Wed, Nov 27, 2019 at 7:53 PM Justin Pryzby wrote:
> 2019-11-26 23:41:50.009-05 | could not fsync file
> "pg_tblspc/16401/PG_12_201909212/16460/973123799.10": No such file or
> directory
I managed to reproduce this (see below). I think I know what the
problem is: mdsyncfiletag() uses _mdfd_
On Fri, Nov 29, 2019 at 9:08 AM Alexander Lakhin wrote:
> So for now it seems dangerous to use predefined collations as their old
> versions are not preserved by pg_upgrade and the user doesn't know which
> indexes affected by the actual ICU collation changes.
Yeah, we noticed this while working
On Fri, Nov 29, 2019 at 3:13 AM Thomas Munro wrote:
> On Wed, Nov 27, 2019 at 7:53 PM Justin Pryzby wrote:
> > 2019-11-26 23:41:50.009-05 | could not fsync file
> > "pg_tblspc/16401/PG_12_201909212/16460/973123799.10": No such file or
> > directory
>
> I
On Fri, Nov 29, 2019 at 11:14 AM Justin Pryzby wrote:
> On Fri, Nov 29, 2019 at 10:50:36AM +1300, Thomas Munro wrote:
> > On Fri, Nov 29, 2019 at 3:13 AM Thomas Munro wrote:
> > > On Wed, Nov 27, 2019 at 7:53 PM Justin Pryzby
> > > wrote:
> > > > 2019-
On Fri, Nov 29, 2019 at 12:34 PM Thomas Munro wrote:
> ... or stop using
> _mdfd_getseg() for this so that you can remove segments independently
> without worrying about sync requests for other segments (it was
> actually like that in an earlier version of the patch for commit
>
On Thu, Nov 21, 2019 at 11:31 AM Tom Lane wrote:
> Thomas Munro writes:
> > I think the choices are: (1) switch to ssize_t and return -1, (2) add
> > at least one of BufFileEof(), BufFileError(), (3) have BufFileRead()
> > raise errors with elog(). I lean towards (2)
On Sun, Dec 1, 2019 at 12:22 PM Mark Dilger wrote:
> These two patches (v3) no longer apply cleanly. Could you please
> rebase?
Hi Mark,
Thanks. Here's v4.
0001-Add-SQL-type-xid8-to-expose-FullTransactionId-to--v4.patch
Description: Binary data
0002-Introduce-xid8-variants-of-the-txid_XXX-f
On Mon, Dec 2, 2019 at 5:58 PM Michael Paquier wrote:
> I was just looking at some callers of pg_fsync(), to notice that some
> code paths don't use data_sync_elevel(). For some code paths, that's
> actually better to never PANIC (say backup_label file, logical
> decoding snapshot, lock file wher
1001 - 1100 of 4324 matches
Mail list logo
