* Andres Freund (and...@anarazel.de) wrote:
> I'd rather see those split into separate commits. Doing polishing and
> active bugs in one commit imo isn't a good idea if the polishing goes
> beyond a line or two.
>
> Otherwise this looks ok to me.
Done that way.
Thanks!
S
On Fri, Jul 3, 2015 at 3:22 AM, Robbie Harwood wrote:
> Hello -hackers,
>
> As previously discussed on this list, I have coded up GSSAPI encryption
> support. If it is easier for anyone, this code is also available for
> viewing on my github:
>
> https://github.com/postgres/postgres/compare/mast
On Tue, Aug 18, 2015 at 3:06 PM, Merlin Moncure wrote:
> On Tue, Aug 18, 2015 at 8:17 AM, Robert Haas wrote:
>> On Sat, Aug 15, 2015 at 6:47 PM, Noah Misch wrote:
>>> CREATE TABLE t (c) AS SELECT 1;
>>> BEGIN;
>>> UPDATE t SET c = 2 WHERE c = 1;
>>> BEGIN_AUTONOMOUS;
>>> UPDATE t SET c = 3 WHERE
On Tue, Aug 18, 2015 at 11:27 PM, Amit Kapila wrote:
>> Here is one other issue I found. Existing code assumes a TOC segment has
>> only one contents per node type, so it uses pre-defined key (like
>> PARALLEL_KEY_SCAN) per node type, however, it is problematic if we put
>> multiple PlannedStmt or
I split this into a new thread to talk about archiving best practices.
On 8/18/15 2:25 PM, Stephen Frost wrote:
> * Stephen Frost (sfr...@snowman.net) wrote:
>> * Robert Haas (robertmh...@gmail.com) wrote:
>>> On Mon, Aug 17, 2015 at 2:50 PM, Peter Eisentraut
wrote:
The commit message for de
On Tue, Aug 18, 2015 at 8:36 PM, Tom Lane wrote:
> I wrote:
>> Just thinking about this ... I wonder why we need to call
>> TransactionIdIsInProgress() at all rather than believing the answer from
>> the snapshot? Under what circumstances could TransactionIdIsInProgress()
>> return true where Xid
On Wed, Aug 19, 2015 at 2:53 AM, Amit Kapila wrote:
> I think one case where the patch can impact is for aborted transactions.
> In TransactionIdIsInProgress(), we check for aborted transactions before
> consulting pg_subtrans whereas with patch it will consult pg_subtrans
> without aborted transa
Michael Paquier wrote:
> if (needlabel)
> appendStringInfo(buf, "::%s",
> -
> format_type_with_typemod(node->consttype,
> -
> node->consttypmod));
> +
> format_type_be_qualified(node->consttype));
> Pondering more about this one, I think that we are going to need a new
> ro
On Fri, Aug 21, 2015 at 12:55:39PM -0300, Alvaro Herrera wrote:
> Michael Paquier wrote:
>
> > if (needlabel)
> > appendStringInfo(buf, "::%s",
> > -
> > format_type_with_typemod(node->consttype,
> > -
> > node->consttypmod));
> > +
> > format_type_be_qualified(node->constt
Robert Haas writes:
> On Tue, Aug 18, 2015 at 8:36 PM, Tom Lane wrote:
>> I experimented with the attached patch, which replaces
>> HeapTupleSatisfiesMVCC's calls of TransactionIdIsInProgress with
>> XidInMVCCSnapshot, and then as a cross-check has all the "return false"
>> exits from XidInMVCCSn
Piotr Stefaniak writes:
> during my testing I've found cases of left-shifting negative integers
> during run-time and I was recently encouraged to post a report of them,
> so here it is (done against 960ea971e66bcd621ba88841b4cb85c7f0e7c383).
What's your concern exactly? The behavior is well-d
Alvaro Herrera wrote:
> I don't see any reason not to remove this. It's been unused since
> a191a169d6d0b9558da4519e66510c4540204a51, dated Jan 10 2007.
Pushed.
--
Álvaro Herrerahttp://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services
On 2015-08-21 13:03:42 -0400, Tom Lane wrote:
> The behavior is well-defined, at least as long as we don't shift far
> enough to have integer overflow
Unfortunately not:
5.8.2: The value of E1 << E2 is E1 left-shifted E2 bit positions;
vacated bits are zero-filled. If E1 has an unsigned type, the
Andres Freund writes:
> On 2015-08-21 13:03:42 -0400, Tom Lane wrote:
>> The behavior is well-defined, at least as long as we don't shift far
>> enough to have integer overflow
> Unfortunately not:
> 5.8.2: The value of E1 << E2 is E1 left-shifted E2 bit positions;
> vacated bits are zero-filled.
On Friday, August 21, 2015, Piotr Stefaniak
wrote:
> If I'm not mistaken, the roles introduced in this test are never dropped,
> which will cause the test to fail on consequent runs.
>
Ah, I was thinking there was a reason to not leave them around but couldn't
think of it and liked the idea of t
Michael Paquier wrote:
> TransactionIdGetCommitTsData@commit_ts.c does the following:
> if (ts)
> *ts = entry.time;
> [...]
> return *ts != 0;
> This is a bad idea, because if TransactionIdGetCommitTsData is called
> with ts == NULL this would simply crash. It seems to me that
Stephen Frost writes:
> On Friday, August 21, 2015, Piotr Stefaniak
> wrote:
>> If I'm not mistaken, the roles introduced in this test are never dropped,
>> which will cause the test to fail on consequent runs.
> Ah, I was thinking there was a reason to not leave them around but couldn't
> think
On Friday, August 21, 2015, Tom Lane wrote:
> Stephen Frost > writes:
> > On Friday, August 21, 2015, Piotr Stefaniak >
> > wrote:
> >> If I'm not mistaken, the roles introduced in this test are never
> dropped,
> >> which will cause the test to fail on consequent runs.
>
> > Ah, I was thinking
On Thu, Aug 20, 2015 at 11:56 PM, Simon Riggs wrote:
> This will give more runs, so merging those needs some thought. It will also
> give a more predictable number of runs, so we'll be able to predict any
> merging issues ahead of time. We can more easily find out the min/max tuple
> in each run,
On 2015-08-21 13:27:22 -0400, Tom Lane wrote:
> Andres Freund writes:
> > On 2015-08-21 13:03:42 -0400, Tom Lane wrote:
> >> The behavior is well-defined, at least as long as we don't shift far
> >> enough to have integer overflow
>
> > Unfortunately not:
> > 5.8.2: The value of E1 << E2 is E1 le
On Wed, Aug 19, 2015 at 11:39 AM, Andres Freund wrote:
> There's a bunch of issues with those two blocks afaics:
>
> 1) The first block (in one backend) might read INVALID_PGPROCNO before
>ever locking the semaphore if a second backend quickly enough writes
>INVALID_PGPROCNO. That way the
Stephen Frost writes:
> On Friday, August 21, 2015, Tom Lane wrote:
>> It is not really acceptable to leave roles hanging around after "make
>> installcheck"; that would be a security hazard for the installation.
>> Please drop them.
> The only ones which were left were intentionally all NOLOGIN
On Tue, Aug 18, 2015 at 11:23 PM, Tom Lane wrote:
> More generally, what would you hope to accomplish with such a construct
> that wouldn't be better done by writing the cursor's underlying query
> directly in the WITH clause?
Maybe I'm stupid today, but it seems like the obvious use case would
b
On 2015-08-21 14:08:36 -0400, Robert Haas wrote:
> On Wed, Aug 19, 2015 at 11:39 AM, Andres Freund wrote:
> > There's a bunch of issues with those two blocks afaics:
> >
> > 1) The first block (in one backend) might read INVALID_PGPROCNO before
> >ever locking the semaphore if a second backend
Tomas Vondra writes:
> On 08/20/2015 11:04 PM, Stefan Kaltenbrunner wrote:
>> On 08/20/2015 06:09 PM, Tom Lane wrote:
>>> (The reason I say "lobotomize" is that there's no particularly
>>> good reason to assume that the first N lines will tell you what you
>>> need to know. And the filter rule wou
Robert Haas writes:
> On Tue, Aug 18, 2015 at 11:23 PM, Tom Lane wrote:
>> More generally, what would you hope to accomplish with such a construct
>> that wouldn't be better done by writing the cursor's underlying query
>> directly in the WITH clause?
> Maybe I'm stupid today, but it seems like
On Fri, Aug 21, 2015 at 2:31 PM, Andres Freund wrote:
> No, if it's paired like that, I don't think it's allowed to fail.
>
> But, as the code stands, there's absolutely no guarantee you're not
> seeing something like:
> P1: a = 0;
> P1: b = 0;
> P1: PGSemaphoreLock(&P1);
> P2: a = 1;
> P2: PGSema
On Fri, Aug 21, 2015 at 2:39 PM, Tom Lane wrote:
> Robert Haas writes:
>> On Tue, Aug 18, 2015 at 11:23 PM, Tom Lane wrote:
>>> More generally, what would you hope to accomplish with such a construct
>>> that wouldn't be better done by writing the cursor's underlying query
>>> directly in the WI
Robert Haas writes:
> The existing syntax for FETCH already includes a way to specify the
> number of rows you want to fetch, as in this example from the
> documentation:
> FETCH FORWARD 5 FROM liahona;
> Why wouldn't that work here too?
Mm, okay, but you still have the other objections to addres
On Fri, Aug 21, 2015 at 2:53 PM, Tom Lane wrote:
> Robert Haas writes:
>> The existing syntax for FETCH already includes a way to specify the
>> number of rows you want to fetch, as in this example from the
>> documentation:
>> FETCH FORWARD 5 FROM liahona;
>> Why wouldn't that work here too?
>
>
Michael Paquier writes:
> On Fri, Jul 3, 2015 at 3:22 AM, Robbie Harwood wrote:
>
>> There are 8 commits in this series; I have tried to err on the side of
>> creating too much separation rather than too little. A patch for each
>> is attached. This is v1 of the series.
>
> I just had a quick
diff --git a/src/backend/commands/trigger.c b/src/backend/commands/trigger.c
index d169027..43421d6 100644
--- a/src/backend/commands/trigger.c
+++ b/src/backend/commands/trigger.c
@@ -4126,7 +4126,7 @@ AfterTriggerEndXact(bool isCommit)
/*
-* Forget the query stack and constrant-related
Hello KaiGai-san,
On 08/21/2015 02:28 AM, Kouhei Kaigai wrote:
...
But what is the impact on queries that actually need more than 1GB
of buckets? I assume we'd only limit the initial allocation and
still allow the resize based on the actual data (i.e. the 9.5
improvement), so the queries would
* Tom Lane (t...@sss.pgh.pa.us) wrote:
> Stephen Frost writes:
> > On Friday, August 21, 2015, Tom Lane wrote:
> >> It is not really acceptable to leave roles hanging around after "make
> >> installcheck"; that would be a security hazard for the installation.
> >> Please drop them.
>
> > The onl
On 8/17/15 4:46 PM, Tom Lane wrote:
Jim Nasby writes:
On 8/17/15 9:48 AM, Tom Lane wrote:
I'm inclined to think that if we wanted to make this better, the way to
improve it would be to detect the error*at compile time*, and get rid of
this hack in plpgsql_exec_function altogether.
So split
On Wed, Aug 5, 2015 at 3:17 AM, Robert Abraham <
robert.abraha...@googlemail.com> wrote:
> Hello,
>
> we are using gin indexes on big tables. these tables happen to have
> several billion rows.
> the index creation fails in ginCombineData in src/backend/access/ginbulk.c
> because repalloc is limit
I wrote:
> I thought a bit more about this. Generally, what you want to know about
> a given situation is which contexts have a whole lot of allocations
> and/or a whole lot of child contexts. What you suggest above won't work
> very well if the problem is buried more than about two levels down i
The following review has been posted through the commitfest application:
make installcheck-world: tested, passed
Implements feature: tested, failed
Spec compliant: not tested
Documentation:not tested
The feature doesn't seem to work:
pg_dump -t t -t 'ii*' --strict-name
On Fri, May 1, 2015 at 2:39 PM, Tom Lane wrote:
> * Multiply-referenced WITH item (unless the outer query applies
> identical constraints to each reference, which seems silly and not
> worth the cycles to check for).
>
Not sure if I understand this correctly. Look at this query, CTE q is
reference
[- the vax lists since they cause majordomo confirmation emails for
anyone responding]
On Thu, Aug 20, 2015 at 3:29 PM, Tom Lane wrote:
>
>> There are some planner tests that fail with floating point exceptions
>> -- that's probably a bug on our part. And I've seen at least one
>> server crash (m
Jim Nasby writes:
> On 8/17/15 9:48 AM, Tom Lane wrote:
>>> I'm inclined to think that if we wanted to make this better, the way to
>>> improve it would be to detect the error*at compile time*, and get rid of
>>> this hack in plpgsql_exec_function altogether.
> Here's a patch that does that. This
On 8/21/15 7:21 PM, Tom Lane wrote:
Applied with some fixes. The major oversight was that EXIT does*not*
have the same rules as CONTINUE, as is clearly documented (though in your
defense, there was no regression test verifying the behavior ... there is
now).
Yay more tests.
>I refactored the
Neil Conway writes:
> On Mon, Aug 17, 2015 at 7:56 AM, Tom Lane wrote:
>> Should we back-patch this change? Seems like it's arguably a
>> performance bug.
> Sounds good to me.
Committed and back-patched.
regards, tom lane
--
Sent via pgsql-hackers mailing list (pgsq
On Sat, Aug 22, 2015 at 12:55 AM, Alvaro Herrera
wrote:
> Michael Paquier wrote:
>
> > if (needlabel)
> > appendStringInfo(buf, "::%s",
> > -
> > format_type_with_typemod(node->consttype,
> > -
> > node->consttypmod));
> > +
> > format_type_be_qualified(node->consttype));
> Hello KaiGai-san,
>
> On 08/21/2015 02:28 AM, Kouhei Kaigai wrote:
> ...
> >>
> >> But what is the impact on queries that actually need more than 1GB
> >> of buckets? I assume we'd only limit the initial allocation and
> >> still allow the resize based on the actual data (i.e. the 9.5
> >> impro
On 8/18/15 12:31 PM, Josh Berkus wrote:
Also this would be useful for range
partitions:
CREATE PARTITION ON parent_table USING ( start_value );
... where start_value is the start range of the new partition. Again,
easier for users to get correct.
Instead of that, I think it would be more foo
On 08/21/2015 08:37 PM, Tom Lane wrote:
Tomas Vondra writes:
I also don't think logging just subset of the stats is a lost case.
Sure, we can't know which of the lines are important, but for example
logging just the top-level contexts with a summary of the child contexts
would be OK.
I thoug
On Fri, Aug 21, 2015 at 8:22 PM, Robert Haas wrote:
>
> On Wed, Aug 19, 2015 at 2:53 AM, Amit Kapila
wrote:
> > I think one case where the patch can impact is for aborted transactions.
> > In TransactionIdIsInProgress(), we check for aborted transactions before
> > consulting pg_subtrans whereas
On Sat, Aug 22, 2015 at 4:06 AM, Robbie Harwood wrote:
>
> Michael Paquier writes:
> > Going through the docs, the overall approach taken by the patch looks neat,
> > and the default values as designed for both the client and the server are
> > good things to do. Now actually looking at the code I
On Fri, Aug 21, 2015 at 7:40 PM, Robert Haas wrote:
>
> On Tue, Aug 18, 2015 at 11:27 PM, Amit Kapila
wrote:
> >> Here is one other issue I found. Existing code assumes a TOC segment
has
> >> only one contents per node type, so it uses pre-defined key (like
> >> PARALLEL_KEY_SCAN) per node type,
On Tue, Aug 18, 2015 at 02:03:19PM +0100, Greg Stark wrote:
> On Tue, Aug 18, 2015 at 6:57 AM, Noah Misch wrote:
> > My own position is based on having maintained a pg_regress suite an order of
> > magnitude larger than that. I don't know why that outcome was so different.
> And does your pg_reg
On Fri, Aug 21, 2015 at 10:06:44AM -0400, Robert Haas wrote:
> On Tue, Aug 18, 2015 at 3:06 PM, Merlin Moncure wrote:
> > On Tue, Aug 18, 2015 at 8:17 AM, Robert Haas wrote:
> >> On Sat, Aug 15, 2015 at 6:47 PM, Noah Misch wrote:
> >>> [1] That's not to say it must use the shmem lock structures
Hello Tomas,
from time to time I need to correlate PostgreSQL logs to other logs,
containing numeric timestamps - a prime example of that is pgbench. With
%t and %m that's not quite trivial, because of timezones etc.
I propose adding two new log_line_prefix escape sequences - %T and %M,
doing
53 matches
Mail list logo