On Sun, Apr 23, 2006 at 11:12:23PM -0400, Tom Lane wrote:
> We removed the repeated-indexscan plan type in favor of bitmap scans.
> I'm not immediately aware of any reason why we shouldn't have ... if
> there is any case where the pre-8.1 plan style wins, it'd have to be
> a pretty narrow scenario.
Michael Fuhr <[EMAIL PROTECTED]> writes:
> Is it expected that the planner in 8.1.3 (CVS) and HEAD doesn't use
> an index scan for the second query below when bitmap scans are
> disabled?
You mean this one?
> test=> EXPLAIN SELECT * FROM foo WHERE id IN (1, 2);
That's not an indexable condition,
"Tom Lane" <[EMAIL PROTECTED]> wrote
> Martijn van Oosterhout writes:
> > On Thu, Apr 20, 2006 at 12:17:07PM -0500, Jim C. Nasby wrote:
> >> Here's one to add to the list: running pgbench with a moderately heavy
> >> load on an SMP box likes to trigger a state where the database (or
> >> pgbench)
Is it expected that the planner in 8.1.3 (CVS) and HEAD doesn't use
an index scan for the second query below when bitmap scans are
disabled? Disabling sequential scans as well results in the planner
using a bitmap scan.
test=> CREATE TABLE foo (id integer PRIMARY KEY);
NOTICE: CREATE TABLE / PRI
Marc G. Fournier wrote:
>
> Try now, should be fixed ... not sure why it was hard coded with 5.8.7,
> but I've reinstalled, and seems to run fine now ...
Yeah, it worked. Thanks.
--
Alvaro Herrerahttp://www.CommandPrompt.com/
PostgreSQL Replication, Consulting,
Try now, should be fixed ... not sure why it was hard coded with 5.8.7,
but I've reinstalled, and seems to run fine now ...
On Sun, 23 Apr 2006, Alvaro Herrera wrote:
Hi,
I just tried to commit a patch to CVS and got this error:
cvs commit: cannot exec /usr/local/bin/activitymail: No such
"Magnus Hagander" <[EMAIL PROTECTED]> writes:
> I'm getting the following regression errors with a backend built using
> Visual C++:
Is HAVE_CBRT getting defined? Either their cbrt() routine or our
default one seems to be generating slightly-off answers. The default
one (at the bottom of float.c
On Apr 23, 2006, at 6:43 PM, Alvaro Herrera wrote:
Agent M wrote:
I have created a directed graph using graphviz that shows the message
flow/event stream. Perhaps this will be helpful to someone. Of course,
corrections are also welcome.
Interesting. I'm not sure how to read the graph howeve
Tom,
> the behavior of
> security-related commands should be as simple and unsurprising as
> possible. Weird special cases added in the name of improving usability
> are likely to do the opposite.
Golden rule.
Thank you :)
---(end of broadcast)---
Hi,
I just tried to commit a patch to CVS and got this error:
cvs commit: cannot exec /usr/local/bin/activitymail: No such file or directory
I went and checked /usr/local/bin/activitymail, and it's there, but the
#! header points to /usr/local/bin/perl5.8.7 while on postgresql.org I
see /usr/loc
Alvaro Herrera <[EMAIL PROTECTED]> writes:
> Gevik Babakhani wrote:
>> Of course a TABLE owner can revoke privileges from himself. But why
>> would a DATABASE owner want to lock himself out from CONNECTING to his
>> database.
> I don't know :-) If it doesn't make sense for somebody, then she won'
Gevik Babakhani <[EMAIL PROTECTED]> writes:
> On Sun, 2006-04-23 at 17:06 -0400, Alvaro Herrera wrote:
>> If I'm not mistaken, the general principle for creating objects is leave
>> their ACLs as NULLs.
> Personally I think this would create an conflict only in case of the
> CONNECT privilege. If
Agent M wrote:
> I have created a directed graph using graphviz that shows the message
> flow/event stream. Perhaps this will be helpful to someone. Of course,
> corrections are also welcome.
Interesting. I'm not sure how to read the graph however -- for example
what does it mean to have ReadyF
Alvaro Herrera <[EMAIL PROTECTED]> writes:
> Gevik Babakhani wrote:
>> At this moment the owner of the database CAN REVOKE himself form the
>> ACL_OBJECT_DATABASE. If the implementation above is acceptable then I
>> can work on this one :)
> Hmm, what do you want to do about it? ISTM the owner sh
> I don't know :-) If it doesn't make sense for somebody, then she won't
> do it.
>
> It's not like we are going out of our way to allow somebody to revoke
> the privileges from oneself. We are just keeping the thing as simple as
> possible. As I said, maybe a reasonable option would be to rais
Gevik Babakhani wrote:
> I was wondering whether there is a backend function to help finding the
> dependencies on a relation.column inside pg_depend. Something like
> dependency.c:0255:findAutoDeletableObjects also taking the "refobjsubid"
> with the search. This I ask in order to avoid reinventi
Gevik Babakhani wrote:
> > > Personally I think it would be better for the database owner not have
> > > the option to REVOKE himself from the CONNECTION privilege of his own
> > > database.
> >
> > Why? A table owner can revoke privileges from himself.
>
> Of course a TABLE owner can revoke p
Cool... will get them added.
On 4/23/06, Alvaro Herrera <[EMAIL PROTECTED]> wrote:
> I hope I'm not too late.
>
> Jonah H. Harris wrote:
> > On 4/19/06, John DeSoi <[EMAIL PROTECTED]> wrote:
> > > Alvaro indicated he would be willing to provide direction on this
> > > with testing support from me.
Hi,
> if (acl in pg_database == NULL)
> acl = acldefault
> else
> acl = acl in pg_database
> if (has_permission(acl, user, ACL_CONNECT))
> can connect
> else
> can't connect
>
>
> To revoke a privilege you do this:
>
> if (acl in pg_datbase == NULL)
> acl = acldefa
Gevik Babakhani wrote:
> > If I'm not mistaken, the general principle for creating objects is leave
> > their ACLs as NULLs. Later, when a privilege is going to be checked, a
> > NULL is treated as if it contained whatever default privilege the object
> > class has. So you should leave this code
I have created a directed graph using graphviz that shows the message
flow/event stream. Perhaps this will be helpful to someone. Of course,
corrections are also welcome.
http://www.themactionfaction.com/pg/PGXProtocol.dot
http://www.themactionfaction.org/pg/PGXProtocol.svg
http://www.themactio
On Sun, Apr 23, 2006 at 11:40:08PM +0200, Gevik Babakhani wrote:
> > Also I'm not sure if we have discussed what's the default (initial)
> > privilege state. Do we want PUBLIC to have CONNECT privilege?
>
> If I where a DBA I would rather explicitly give people connect
> permission to my database
I hope I'm not too late.
Jonah H. Harris wrote:
> On 4/19/06, John DeSoi <[EMAIL PROTECTED]> wrote:
> > Alvaro indicated he would be willing to provide direction on this
> > with testing support from me. He also said there are several other
> > possible PL/PHP issues that would warrant a SoC proje
Hi,
On Sun, 2006-04-23 at 17:06 -0400, Alvaro Herrera wrote:
> Gevik Babakhani wrote:
>
> Hi,
>
> > 2) The file parsenodes.h is updated to support
> > #define ACL_DATABASE_CONNECT
>
> I wouldn't call it ACL_DATABASE_CONNECT, just ACL_CONNECT. Currently we
> don't have any objects other than da
Gevik Babakhani wrote:
Hi,
> 2) The file parsenodes.h is updated to support
> #define ACL_DATABASE_CONNECT
I wouldn't call it ACL_DATABASE_CONNECT, just ACL_CONNECT. Currently we
don't have any objects other than databases that need connecting to, but
you never know.
(It should be very easy to
I'm getting the following regression errors with a backend built using
Visual C++:
*** ./expected/float8-exp-three-digits-win32.outWed Jun 8
23:15:29 2005
--- ./results/float8.outSun Apr 23 20:44:44 2006
***
*** 324,331
|0 |
On Sunday 23 April 2006 11:42, Alvaro Herrera wrote:
> Robert Treat wrote:
> > On Saturday 22 April 2006 11:24, Alvaro Herrera wrote:
> > > Dhanaraj M wrote:
> > > > I saw the following in the TODO list. I am currently trying to work
> > > > on them. I could not understand clearly what needs to be
Uff, I finally got it. It's very simple. I was trying to use dlltool
all the time, but the solution is much simplier - just use directly
the .lib file with gcc and it will handle it. So easy that it didn't
cross my mind:)
gcc "." atomrtget.o atomrtree.o atomrtscan.o atomrtcostestimate.o
rtree
Yes, you got me right. I'm just trying the 2) and it doesn't work. So
I'll try the "depends" tool.
Thx,
David
23. dubna 2006, 20:05:00, napsal jste:
>> 5. Copy the dll to the bin directory (when I let it in the
>> contrib directory, where also the libatomrtree.dl
> 5. Copy the dll to the bin directory (when I let it in the
> contrib directory, where also the libatomrtree.dll file with
> the AM functions is, so it writes out, that the modul can't
> be found, when loading it).
>
> 6. When there is "void *header = f_header()" anywhere in the
> source, so
David Hoksza <[EMAIL PROTECTED]> writes:
> I do following steps:
> 1. Compiling the dll in Microsoft Visual Studio.
> 2. Creating the def file (in mingw32):
I don't think you can use VS for building PG shared libraries --- it's
got incompatible conventions about symbol spelling, or some such.
The
Hi, I'm trying to implement my own access method, which uses c++
framework for saving tuples. The framework consists of set of
functions, which are implemented in win32 dll.
All the functions which are desired by PG (ambuild, aminsert, ...) are
implemented and functional. But when I want to use any
* Robert Treat ([EMAIL PROTECTED]) wrote:
> On Saturday 22 April 2006 13:34, Tom Lane wrote:
> > Also, the TODO item could be worded
> >
> > * Make psql's \d commands more consistent
> >
> > because that's really what Neil is on about ...
> >
>
> like making \df only show user functions and \d
Robert Treat wrote:
> On Saturday 22 April 2006 11:24, Alvaro Herrera wrote:
> > Dhanaraj M wrote:
> > > I saw the following in the TODO list. I am currently trying to work on
> > > them. I could not understand clearly what needs to be done. Can anybody
> > > give me the details for the following s
On Saturday 22 April 2006 11:24, Alvaro Herrera wrote:
> Dhanaraj M wrote:
> > I saw the following in the TODO list. I am currently trying to work on
> > them. I could not understand clearly what needs to be done. Can anybody
> > give me the details for the following so that I can work on?
> >
> >
On Saturday 22 April 2006 13:34, Tom Lane wrote:
> Also, the TODO item could be worded
>
> * Make psql's \d commands more consistent
>
> because that's really what Neil is on about ...
>
like making \df only show user functions and \dfS show system functions, like
all the other objects? :-
On Sat, Apr 22, 2006 at 14:20:32 -0700,
daveg <[EMAIL PROTECTED]> wrote:
> On Sat, Apr 22, 2006 at 01:49:25PM -0700, David Fetter wrote:
> > On Sat, Apr 22, 2006 at 01:14:42PM -0700, David Gould wrote:
> >
> > > To avoid running out of swap and triggering the oom killer we have
> > > had to redu
Hi,
I was wondering whether there is a backend function to help finding the
dependencies on a relation.column inside pg_depend. Something like
dependency.c:0255:findAutoDeletableObjects also taking the "refobjsubid"
with the search. This I ask in order to avoid reinventing the wheel :)
Perhaps th
38 matches
Mail list logo
