see you are
> building against a recent SDK:
>
> /Library/Developer/CommandLineTools/SDKs/MacOSX14.2.sdk/usr/include/string.h
>
> but I wonder if maybe the actual OS version is back-rev?
Skimming the releases on https://opensource.apple.com/releases/ it seems that
memset_s has been available since Mavericks (10.9) AFAICT.
--
Daniel Gustafsson
operties which are required in this codepath.
--
Daniel Gustafsson
I have pushed the fix for the calloc check for now.
--
Daniel Gustafsson
Fix check for memory allocation
Commit 61461a300c1 accidentally checked memory allocation success
using the wrong variable.
Author: Ranier Vilela
Reviewed-by: Jelte Fennema-Nio
Reviewed-by: Daniel Gustafsson
Discussion:
https://postgr.es/m/caeudqaqqfth7xcb-+k6zekjfqbhqcxcr_w4dujtxvt6h3vz
lls a few lines up.
> But, IMO, I think that has more problems.
> If any allocation fails, all allocations must be cleared.
> Or is the current behavior acceptable?
Since this is frontend library code I think we should free all the allocations
in case of OOM.
--
Daniel Gustafsson
better to add a check to pg_upgrade which detects this case and
advices the user how to deal with it.
--
Daniel Gustafsson
warrant investigation. If we
don't have a test for it already then it might be worth constructing something
to catch that.
--
Daniel Gustafsson
t privs
from SHARED_DEPENDENCY_OWNER to SHARED_DEPENDENCY_INITACL, so that these can be
mopped up with a later DROP OWNED? Trying this in a POC patch it fails with
RemoveRoleFromInitPriv not removing the rows, shortcircuiting that for a test
seems to make it work but is it the right approach?
--
Daniel Gustafsson
> On 22 May 2024, at 18:53, Tom Lane wrote:
>
> Daniel Gustafsson writes:
>> They are known to be zero, but that's not entirely equivalent though is it?
>> NIL is defined as ((List *) NULL) and NULL is typically defined as ((void *)
>> 0), so sizeof(0) would be th
ow without looking up the
definition that some of the members are list. I like that readability. That
being said, I for sure won't fight if there is concensus to remove.
--
Daniel Gustafsson
```
> since the memset has written zeroes to the object "dpns", and these members
> are known to be zero.
They are known to be zero, but that's not entirely equivalent though is it?
NIL is defined as ((List *) NULL) and NULL is typically defined as ((void *)
0), so sizeof(0) would be the si
for it to be measurable even?
--
Daniel Gustafsson
rms like Github register the patch against the tree at the
time of submitting, and won't refresh unless the user does so. Github does
detect bitrot and show a "this cannot be merged" error message, but it doesn't
alter any state on the PR etc.
--
Daniel Gustafsson
onally there is a large developer meeting shortly which many are busy
preparing for. Excercise some patience, and I'm sure there will be follow-ups
to this once development of postgres v18 picks up.
--
Daniel Gustafsson
nother scan for SSL related patches in flight at the time.
--
Daniel Gustafsson
ace in time.
Magnus' commitmessage explains it well. The way I've used it (albeit
infrequently) is to point to a specific mail in the thread where a significant
change was proposed, like the patch changhing direction or something similar.
--
Daniel Gustafsson
> On 16 May 2024, at 19:47, Tom Lane wrote:
> Yeah, it's not worth working harder than this. I do see one typo
> in your comment: s/supported then/supported when/. LGTM otherwise.
Thanks for review, I've pushed this (with the fix from above) to 14 through 12.
--
Daniel Gustafsson
Refuse upgrades from pre-9.0 clusters
Commit 695b4a113ab added a dependency on retrieving oldestxid from
pg_control, which only exists in 9.0 and onwards, but the check for
8.4 as the oldest version was retained. Since there has been few if
any complaints of 8.4 upgrades not working, fix by
Refuse upgrades from pre-9.0 clusters
Commit 695b4a113ab added a dependency on retrieving oldestxid from
pg_control, which only exists in 9.0 and onwards, but the check for
8.4 as the oldest version was retained. Since there has been few if
any complaints of 8.4 upgrades not working, fix by
Refuse upgrades from pre-9.0 clusters
Commit 695b4a113ab added a dependency on retrieving oldestxid from
pg_control, which only exists in 9.0 and onwards, but the check for
8.4 as the oldest version was retained. Since there has been few if
any complaints of 8.4 upgrades not working, fix by
e CF
app does however provide that overview. This is essentially the TODO list
aspect, but sharing one's TODO isn't all bad, especially for maintainers.
--
Daniel Gustafsson
> On 17 May 2024, at 00:03, Peter Eisentraut wrote:
> I think, if we consider the core mission of the commitfest app, we need to be
> more protective of the Needs Review state.
IMHO this is a very very good summary of what we should focus on with this
work.
--
Daniel Gustafsson
ture detail could be if the system would send an
automated email to the thread summarizing the entry when it's marked as
"committed"?
--
Daniel Gustafsson
> On 17 May 2024, at 07:57, Peter Eisentraut wrote:
>
> On 16.05.24 23:27, Daniel Gustafsson wrote:
>>> On 16 May 2024, at 11:43, Peter Eisentraut wrote:
>>> You might want to run your patch through pgperltidy. The result doesn't
>>> look bad, but a bit
on the thread to measure "activity/hotness", but starting with the
simple boolean data we already have could be a good start.
--
Daniel Gustafsson
ses the file for each call. It might be nice if
> it could accept a list. Or you can just pass the whole block as one string,
> like it was done for pg_ident.conf before.
The attached v2 pass the whole block as a here-doc which seemed like the best
option to retain readability of the config.
--
> On 16 May 2024, at 13:56, Alvaro Herrera wrote:
> I think we should also take patch 0005 in pg17, which reduces the number
> of strings to translate.
Agreed, lessening the burden on translators is always a good idea.
--
Daniel Gustafsson
errmsg("WAL generated with \"full_page_writes=off\" was replayed "
I'm not a fan of this syntax, but I at the same time can't offer a better idea
so this isn't an objection but a hope that it can be made even better during
the v18 cycle.
--
Daniel Gustafsson
IMHO, from being CMF many times,
there is a fair bit of the latter, which excacerbates the problem. This is
harder to fix with more or better software though.
> I spent a good deal of time going through the CommitFest this week
And you deserve a big Thank You for that.
--
Daniel Gustafsson
> On 16 May 2024, at 15:54, Robert Haas wrote:
>
> On Wed, May 15, 2024 at 9:33 AM Heikki Linnakangas wrote:
>> Ok, yeah, I can see that now. Here's a new version to address that. I
>> merged ENC_SSL_NEGOTIATED_SSL and ENC_SSL_DIRECT_SSL to a single method,
>> ENC_SSL. The places that need to
not attached to
any other conditional. There is no change in functionality, it's mainly for
readability (PG_TEST_EXTRA is it's own concept, not tied to library presence).
0002 ports over editing configfiles to using append_conf() instead of opening
and writing to them directly.
--
Daniel
appealing since the code will be unreachable with this check anyways.
Thoughts?
--
Daniel Gustafsson
0001-Refuse-upgrades-from-pre-9.0-clusters.patch
Description: Binary data
> On 15 May 2024, at 20:46, Tom Lane wrote:
>
> Daniel Gustafsson writes:
>> While looking at pg_dump performance today I noticed that pg_dump fails to
>> clear query results in binary_upgrade_set_pg_class_oids during binary upgrade
>> mode. 9a974cbcba00 moved t
Fix query result leak during binary upgrade
9a974cbcba00 moved the query in binary_upgrade_set_pg_class_oids to the
outer level, but left the PQclear and query buffer destruction in the
is_index conditional. 353708e1fb2d fixed the leak of the query buffer
but left the PGresult leak. This moves
Fix query result leak during binary upgrade
9a974cbcba00 moved the query in binary_upgrade_set_pg_class_oids to the
outer level, but left the PQclear and query buffer destruction in the
is_index conditional. 353708e1fb2d fixed the leak of the query buffer
but left the PGresult leak. This moves
Fix query result leak during binary upgrade
9a974cbcba00 moved the query in binary_upgrade_set_pg_class_oids to the
outer level, but left the PQclear and query buffer destruction in the
is_index conditional. 353708e1fb2d fixed the leak of the query buffer
but left the PGresult leak. This moves
it not always
be executed. 353708e1fb2d fixed the leak of the query buffer but left the
PGresult leak. The attached fixes the PGresult leak which when upgrading large
schemas can be non-trivial.
This needs to be backpatched down to v15.
--
Daniel Gustafsson
0001-Fix-query-result-leak-during
next list item under COLUMNS. Unless objected to I will remove these
parenthesis.
--
Daniel Gustafsson
is was immmediately shot down as
something the community don't want to maintain.
> And we'd need to figure out which Markdown flavor to target.
Absolutely, and as I mentioned above, we need to pick based both the final
result (text and rendered) as well as the developer experience for maintaining
this.
--
Daniel Gustafsson
> On 14 May 2024, at 08:03, Michael Paquier wrote:
>
> On Mon, May 13, 2024 at 08:06:57PM +0200, Daniel Gustafsson wrote:
>>> Any chance we'll have these fixes in v17?
>>
>> Nice timing, I was actually rebasing them today to get them committed.
>
> Looks sen
> On 14 May 2024, at 07:12, Michael Paquier wrote:
> Hence, no objections to clean up that now. Thanks for asking.
Thanks for verifying, I've pushed this now.
--
Daniel Gustafsson
Remove auth-options support from initdb
When --auth was added to initdb in commit e7029b212755 it had support
for auth options separated by space from the auth type, like:
--auth pam
--auth ident sameuser
Passing an option to the ident auth type was removed in 01c1a12a5bb4
which left
Fix memory leaks in error reporting with LOG level
When loglevel is set to LOG, allocated strings used in the error
message would leak. Fix by explicitly pfreeing them.
Author: Ranier Vilela
Reviewed-by: Daniel Gustafsson
Reviewed-by: Michael Paquier
Discussion:
https://postgr.es/m
better
to link to the archive.org entry instead.
--
Daniel Gustafsson
> On 13 May 2024, at 20:05, Ranier Vilela wrote:
> Em qua., 10 de abr. de 2024 às 15:33, Daniel Gustafsson
> escreveu:
> Thanks, I'll have a look. I've left this for post-freeze on purpose to not
> cause unnecessary rebasing. Will take a look over the next few days
> On 13 May 2024, at 10:48, Peter Eisentraut wrote:
> Patches attached.
All patches look good.
> I think the first one is a bug fix.
Agreed.
--
Daniel Gustafsson
mat option for EXPLAIN?
Since json (and yaml/xml) is intended to be machine-readable I think we use a
single unit for all values, and document this fact.
--
Daniel Gustafsson
regress_partition_merge_alice;
+ERROR: role "regress_partition_merge_alice" already exists
--
Daniel Gustafsson
ter
implementation).
Any objections to fixing this in 17 by removing it? (cc:ing Michael from the
RMT)
--
Daniel Gustafsson
v2-0001-Remove-auth-options-support-from-initdb.patch
Description: Binary data
that for SNI.
* The documentation is not polished at all and will require a more work to make
it passable I think. There are also lot's more testing that can be done, so
far it's pretty basic.
* I've so far only tested with OpenSSL and haven't yet verified how LibreSSL
handles this.
--
Daniel
to repro and fix a bug
> from the section "live issues affecting stable branches". I will
> update this BHS patch by tonight and commit it once Tomas has a chance
> to +1.
+1
--
Daniel Gustafsson
noticed for 7 major versions. Plus, based on my
> admittedly limited testing, this is unlikely to provide significant
> improvements.
Agreed, this is a nice little improvement but it's unlikely to be enough of a
speedup to warrant changing the backbranches.
--
Daniel Gustafsson
> On 7 May 2024, at 01:31, Michael Paquier wrote:
>
> On Fri, May 03, 2024 at 10:39:15AM +0200, Daniel Gustafsson wrote:
>> They are no-ops when linking against v18, but writing an extension which
>> targets all supported versions of postgres along with their respective
Fix comment regarding LibreSSL availability
SSL_AD_NO_APPLICATION_PROTOCOL is indeed available in LibreSSL, but only
in 3.4.3 and later (shipped in OpenBSD 7.0).
Discussion: https://postgr.es/m/e1s1g0z-000jec...@gemulon.postgresql.org
Branch
--
master
Details
---
> On 3 May 2024, at 21:21, Tom Lane wrote:
>
> Peter Eisentraut writes:
>> On 03.05.24 10:39, Daniel Gustafsson wrote:
>>> They are no-ops when linking against v18, but writing an extension which
>>> targets all supported versions of postgres along with their
> On 3 May 2024, at 13:48, Peter Eisentraut wrote:
> Maybe it's easier if we just replaced
>
>prototypes for functions in xxx.c
>
> with
>
>declarations for xxx.c
>
> throughout src/include/libpq/libpq.h.
+1
--
Daniel Gustafsson
> On 2 May 2024, at 12:30, Heikki Linnakangas wrote:
> On 02/05/2024 13:24, Daniel Gustafsson wrote:
>> This makes targeting 7.0 as the lowest LibreSSL version appealing in my
>> patchset for removing support for old OpenSSL and LibreSSL versions.
>
> Works for me. Alth
enSSL versions make them still required, or am I missing something?
>ERR_clear_error();
> -
> #ifdef USE_RESOWNER_FOR_HMAC
>
> Some noise diff.
Will fix with a new rebase when once the tree has settled down from the
post-freeze fixups in this area.
--
Daniel Gustafsson
> On 2 May 2024, at 11:30, Heikki Linnakangas wrote:
>
> On 02/05/2024 12:09, Daniel Gustafsson wrote:
>>> On 30 Apr 2024, at 07:26, Heikki Linnakangas
>>> wrote:
>>> Fix compilation on OpenSSL 1.0.2 and LibreSSL
>>>
>>> SSL_AD_N
1.1.0 and later, but as of this writing not in LibreSSL.
I'm a bit confused, as far as I can tell this has been in LibreSSL since the
OpenBSD 6.9 release.
https://github.com/openbsd/src/blob/master/lib/libssl/ssl_tlsext.c#L130
Or am I missing something?
--
Daniel Gustafsson
doc: Fix incorrectly spelled structname
Commit 61461a300c1 accidentally misspelled the PGcancelConn struct
using the PQ prefix (which admittedly is a very easy typo to make).
Reported off-list.
Reported-by: Alexander Lakhin
Branch
--
master
Details
---
est added here aptly cover 04e72ed617be in terms its functionality?
--
Daniel Gustafsson
it was footgun prevention.
--
Daniel Gustafsson
ethods, like clientcert to cert for
example. If we fix it we should also document that it works IMHO.
--
Daniel Gustafsson
that function existed. I agree that it's an
improvement even at the increased query complexity.
> v3 attached also has a bit more work on code comments.
LGTM.
--
Daniel Gustafsson
rts
> of "regress_dump_test_role" or to privilege strings. That
> seems unlikely enough to live with, but I wonder if anybody has
> a better idea.
I think that will be bulletproof enough to keep it working in the buildfarm and
among 99% of hackers.
--
Daniel Gustafsson
> On 27 Apr 2024, at 20:32, Daniel Gustafsson wrote:
> That's a good point, there is potential for more code removal here. The
> attached 0001 takes a stab at it while it's fresh in mind, I'll revisit before
> the July CF to see if there is more that can be done.
> On 25 Apr 2024, at 05:49, Michael Paquier wrote:
>
> On Wed, Apr 24, 2024 at 01:31:12PM +0200, Daniel Gustafsson wrote:
>> Done. Attached are the two remaining patches, rebased over HEAD, for
>> removing
>> support for OpenSSL 1.0.2 in v18. Parking them in the c
> On 21 Apr 2024, at 23:08, Tom Lane wrote:
>
> Daniel Gustafsson writes:
>>> On 6 Apr 2024, at 01:10, Tom Lane wrote:
>>> (So now I'm wondering why *only* copperhead has shown this so far.
>>> Are our other cross-version-upgrade testing animals AW
y server: %s",
app_name.data, err));
Messages like this should perhaps have translator comments to indicate what the
leading "%s" will contain?
--
Daniel Gustafsson
> On 25 Apr 2024, at 17:48, Robert Haas wrote:
>
> On Wed, Apr 24, 2024 at 3:20 PM Daniel Gustafsson wrote:
>> Patch LGTM.
>
> Thanks. Here's an updated version with fixes for the other issues you
> mentioned.
LGTM, only one small error in the commitmessage:
> On 25 Apr 2024, at 18:16, Robert Haas wrote:
>
> On Wed, Apr 24, 2024 at 3:08 PM Daniel Gustafsson wrote:
>> LGTM; only one small comment which you can ignore if you feel it's not worth
>> the extra words.
>>
>> +pg_combinebackup when the checksum statu
a. Maybe we should set an
xreflabel while at it for completeness sake?
--
Daniel Gustafsson
" -n, --dry-run don't actually do anything\n"));
printf(_(" -N, --no-sync do not wait for changes to be written
safely to disk\n"));
--
Daniel Gustafsson
> On 22 Apr 2024, at 20:52, Robert Haas wrote:
>
> On Thu, Apr 18, 2024 at 3:25 PM Daniel Gustafsson wrote:
>>> On 18 Apr 2024, at 20:11, Robert Haas wrote:
>>> 2. As (1), but make check_control_files() emit a warning message when
>>> the problem case i
> On 24 Apr 2024, at 00:20, Michael Paquier wrote:
>
> On Tue, Apr 23, 2024 at 10:08:13PM +0200, Daniel Gustafsson wrote:
>> Hearing no objections to this plan (and the posted v10), I'll go ahead with
>> 0001, 0003 and 0004 into v17 tomorrow.
>
> WFM, thanks.
> On 22 Apr 2024, at 10:45, Daniel Gustafsson wrote:
>
>> On 21 Apr 2024, at 12:57, PG Doc comments form
>> wrote:
>>
>> The following documentation comment has been logged on the website:
>>
>> Page: https://www.postgresql.org/docs/16/app-psql.html
Doc: fix prompt in psql crosstabview example
The prompt in the crosstabview example was incorrectly indicating
an open parenthesis where there is none in the example query. Fix
by changing to the normal multi-line prompt.
Reported-by: y.sabu...@gmail.com
Discussion:
FILE is close to two decades old so there might be extensions
using it (though the risk might be slim), perhaps using offering it as as well
as backwards-compatability is warranted should we introduce a new name?
--
Daniel Gustafsson
Doc: Use past tense for things which happened in the past
The paragraph on SSL compression is largely describing events which
took place many years ago, so reword with past tense.
Reviewed-by: Peter Eisentraut
Reviewed-by: Michael Paquier
Discussion:
Support disallowing SSL renegotiation when using LibreSSL
LibreSSL doesn't support the SSL_OP_NO_RENEGOTIATION macro which is
used by OpenSSL, instead it has invented a similar one for client-
side renegotiation: SSL_OP_NO_CLIENT_RENEGOTIATION. This has been
supported since LibreSSL 2.5.1 which
Support SSL_R_VERSION_TOO_LOW when using LibreSSL
The SSL_R_VERSION_TOO_LOW error reason is supported in LibreSSL since
LibreSSL 3.6.3, shipped in OpenBSD 7.2. SSL_R_VERSION_TOO_HIGH is on
the other hand not supported in any version of LibreSSL. Previously
we only checked for
> let to go through a few months of beta testing.
>
> 0004 effectively just enhances an error message for LibreSSL; there is little
> reason to backpatch this.
Hearing no objections to this plan (and the posted v10), I'll go ahead with
0001, 0003 and 0004 into v17 tomorrow.
--
Daniel Gustafsson
ocumentation is needed here.
This sounds like a question for the Homebrew maintainers, the postgres project
is not packaging postgres for Homebrew and have no insight into how they do it.
--
Daniel Gustafsson
> On 22 Apr 2024, at 18:04, Tom Lane wrote:
> Seems like a useful change
Agreed.
> ... about like this?
Patch LGTM.
--
Daniel Gustafsson
> On 21 Apr 2024, at 23:08, Tom Lane wrote:
> ... were you going to look at it? I can take a whack if it's too far down
> your priority list.
Yeah, I’m working on a patchset right now.
./daniel
testdb(> FROM my_table t1 CROSS JOIN my_table t2 ORDER BY 1 DESC
> testdb(> \crosstabview "A" "B" "AxB" ord
> ```
That's an impressively eagle-eyed observation, it is indeed showing an open
parenthesis where none is. Will fix.
--
Daniel Gustafsson
> On 20 Apr 2024, at 01:23, Michael Paquier wrote:
>
> On Fri, Apr 19, 2024 at 10:12:14AM +0200, Daniel Gustafsson wrote:
>> Off the cuff that seems to make sense, it does make it easier to grep for
>> uses
>> of the control file.
>
> +1 for switching to the ma
w for new features opens again I am sure this patch will get reviews (I
have it on my personal TODO and hope to get to it).
--
Daniel Gustafsson
> On 19 Apr 2024, at 13:49, Daniel Gustafsson wrote:
>> On 19 Apr 2024, at 12:31, Aleksander Alekseev
>> wrote:
>> Here is the patch.
>
> Nice catch, and thanks for the patch. I'll apply it with a backpatch to when
> they were removed.
Done, thanks for the re
Doc: Remove mention of @ and ~ GiST operators
These operators were removed by 2f70fdb0644c in the v14 cycle but they were
accidentally left in the table of build-in operator classes. Backpatch down
to v14 where the operators where removed.
Author: Aleksander Alekseev
Reported-by: Colin Caine
Doc: Remove mention of @ and ~ GiST operators
These operators were removed by 2f70fdb0644c in the v14 cycle but they were
accidentally left in the table of build-in operator classes. Backpatch down
to v14 where the operators where removed.
Author: Aleksander Alekseev
Reported-by: Colin Caine
Doc: Remove mention of @ and ~ GiST operators
These operators were removed by 2f70fdb0644c in the v14 cycle but they were
accidentally left in the table of build-in operator classes. Backpatch down
to v14 where the operators where removed.
Author: Aleksander Alekseev
Reported-by: Colin Caine
Doc: Remove mention of @ and ~ GiST operators
These operators were removed by 2f70fdb0644c in the v14 cycle but they were
accidentally left in the table of build-in operator classes. Backpatch down
to v14 where the operators where removed.
Author: Aleksander Alekseev
Reported-by: Colin Caine
These
>> operators were removed by 2f70fdb0644c back in 2020.
>>
>> I will submit a patch for the documentation shortly. Thanks for reporting.
>
> Here is the patch.
Nice catch, and thanks for the patch. I'll apply it with a backpatch to when
they were removed.
--
Daniel Gustafsson
> On 16 Apr 2024, at 15:37, Nazir Bilal Yavuz wrote:
> I realized two small typos: 'sgmr' -> 'smgr'. You may want to include
> them in 0001.
Thanks, I incorporated these into 0001 before pushing. All the commits in this
patchset are now applied.
--
Daniel Gustafsson
> On 19 Apr 2024, at 05:50, Anton A. Melnikov wrote:
> May be better use this macro everywhere in C code?
Off the cuff that seems to make sense, it does make it easier to grep for uses
of the control file.
--
Daniel Gustafsson
Remove unused function prototype
Commit aafc05de1bf5 removed StartSlotSyncWorker() but mistakenly left
the prototype in slotsync.h. Fix by removing.
Reported-by: Alexander Lakhin
Discussion: https://postgr.es/m/3f577953-a29e-4722-98ad-2da9eff2c...@yesql.se
Branch
--
master
Details
Fix incorrect parameter name in prototype
The function declaration for select_next_encryption_method use the
variable name have_valid_connection, so fix the prototype in the
header to match that.
Reported-by: Alexander Lakhin
Discussion:
ut a backpatch? Or perhaps
> you disagree?
If we want to 0001 can be baclpatched to v15, 0004 to v13 and 0003 all the way.
I don't have strong opinions either way.
--
Daniel Gustafsson
1 - 100 of 3272 matches
Mail list logo