FAQ
============================================================================
POSTGRESQL BUG REPORT TEMPLATE
============================================================================


Your name : David Mansfield
Your email address : david@cobite.com

Category : runtime: back-end
Severity : serious

Summary: calling notify in pl/pgsql proc causes core dump

System Configuration
--------------------
Operating System : Linux 2.2.9

PostgreSQL version : 6.5

Compiler used : gcc 2.7.2.3

Hardware:
---------
RedHat 5.2

Versions of other tools:
------------------------
RedHat 5.2 standard

--------------------------------------------------------------------------

Problem Description:
--------------------
calling notify in pl/pgsql proc causes core dump

--------------------------------------------------------------------------

Test Case:
----------
createdb bugs
psql bugs
\i /usr/src/postgresql-6.5/src/pl/plpgsql/src/mklang.sql
create function a_func()
returns
char
as
'begin
notify my_listen;
return \'a\';
end;'
language
'plpgsql'
;
select a_func();


--------------------------------------------------------------------------

Solution:
---------


--------------------------------------------------------------------------

Search Discussions

  • Henry B. Hotz at Jun 24, 1999 at 10:22 pm
    ============================================================================
    POSTGRESQL BUG REPORT TEMPLATE
    ============================================================================


    Your name : Henry B. Hotz
    Your email address : hotz@jpl.nasa.gov

    Category : runtime: back-end
    Severity : serious

    Summary: calling notify in pl/pgsql proc causes core dump

    System Configuration
    --------------------
    Operating System : Solaris 2.5

    PostgreSQL version : 6.4

    Compiler used : gcc 2.7.2.2

    Hardware:
    ---------
    SPARCstation 5

    Versions of other tools:
    ------------------------
    N/A

    --------------------------------------------------------------------------

    Problem Description:
    --------------------
    pg_dump -z creates grant commands of the form
    grant "all" on "hotz" to "hotz";
    instead of
    grant "all" on "table" to "hotz";

    Don't know if this is fixed in newer dot revisions or 6.5.

    --------------------------------------------------------------------------

    Test Case:
    ----------
    pg_dump -z <database> >db.backup

    --------------------------------------------------------------------------

    Solution:
    ---------


    --------------------------------------------------------------------------

    Signature failed Preliminary Design Review.
    Feasibility of a new signature is currently being evaluated.
    h.b.hotz@jpl.nasa.gov, or hbhotz@oxy.edu
  • Bruce Momjian at Jun 24, 1999 at 11:07 pm

    Problem Description:
    --------------------
    pg_dump -z creates grant commands of the form
    grant "all" on "hotz" to "hotz";
    instead of
    grant "all" on "table" to "hotz";

    Don't know if this is fixed in newer dot revisions or 6.5.

    --------------------------------------------------------------------------

    Test Case:
    ----------
    pg_dump -z <database> >db.backup
    I just did this:

    test=> create table test (x int);
    CREATE
    test=> grant all on test to postgres;
    CHANGE
    test=> \q
    #$ pg_dump -z -d test >/bjm/l
    pg_dump: The -z option(dump ACLs) is now the default, continuing.
    #$ less /bjm/l
    \connect - postgres
    CREATE TABLE "test" (
    "x" int4);
    REVOKE ALL on "test" from PUBLIC;
    GRANT ALL on "test" to "postgres";
    #$

    Looks OK to me.

    --
    Bruce Momjian | http://www.op.net/~candle
    maillist@candle.pha.pa.us | (610) 853-3000
    + If your life is a hard drive, | 830 Blythe Avenue
    + Christ can be your backup. | Drexel Hill, Pennsylvania 19026
  • Herouth Maoz at Jun 27, 1999 at 2:41 pm

    At 01:21 +0300 on 25/06/1999, Henry B. Hotz wrote:


    pg_dump -z creates grant commands of the form
    grant "all" on "hotz" to "hotz";
    instead of
    grant "all" on "table" to "hotz";

    Don't know if this is fixed in newer dot revisions or 6.5.
    It was fixed in 6.4.2, and I would guess 6.5 too.

    Herouth

    --
    Herouth Maoz, Internet developer.
    Open University of Israel - Telem project
    http://telem.openu.ac.il/~herutma
  • Bruce Momjian at Jul 8, 1999 at 2:08 am
    Jan, this is yours.

    ============================================================================
    POSTGRESQL BUG REPORT TEMPLATE
    ============================================================================


    Your name : David Mansfield
    Your email address : david@cobite.com

    Category : runtime: back-end
    Severity : serious

    Summary: calling notify in pl/pgsql proc causes core dump

    System Configuration
    --------------------
    Operating System : Linux 2.2.9

    PostgreSQL version : 6.5

    Compiler used : gcc 2.7.2.3

    Hardware:
    ---------
    RedHat 5.2

    Versions of other tools:
    ------------------------
    RedHat 5.2 standard

    --------------------------------------------------------------------------

    Problem Description:
    --------------------
    calling notify in pl/pgsql proc causes core dump

    --------------------------------------------------------------------------

    Test Case:
    ----------
    createdb bugs
    psql bugs
    \i /usr/src/postgresql-6.5/src/pl/plpgsql/src/mklang.sql
    create function a_func()
    returns
    char
    as
    'begin
    notify my_listen;
    return \'a\';
    end;'
    language
    'plpgsql'
    ;
    select a_func();


    --------------------------------------------------------------------------

    Solution:
    ---------


    --------------------------------------------------------------------------


    --
    Bruce Momjian | http://www.op.net/~candle
    maillist@candle.pha.pa.us | (610) 853-3000
    + If your life is a hard drive, | 830 Blythe Avenue
    + Christ can be your backup. | Drexel Hill, Pennsylvania 19026
  • David Mansfield at Jul 8, 1999 at 3:49 pm
    Jan, this is yours.
    Actually, I made a 'workaround' fix to this problem (after posting this
    bug report). I'll append the patch. I don't pretend to understand the
    internals of this system, but my guess is that a simple statement like
    'notify xyz' went through the planner, got an spi_plan, but the 'plan
    list' was empty. So here's my patch, which may be fixing a symptom, not
    the cause, YMMV:

    --- pl_exec.c~ Wed May 26 03:07:39 1999
    +++ pl_exec.c Fri Jun 25 11:00:53 1999
    @@ -2482,6 +2482,10 @@

    plan = (Plan *) lfirst(spi_plan->ptlist);

    + /* it would seem as though plan can be null... --DAVID */
    + if (plan == NULL)
    + return;
    +
    /* ----------
    * 2. It must be a RESULT plan --> no scan's required
    * ----------

    ============================================================================
    POSTGRESQL BUG REPORT TEMPLATE
    ============================================================================


    Your name : David Mansfield
    Your email address : david@cobite.com

    Category : runtime: back-end
    Severity : serious

    Summary: calling notify in pl/pgsql proc causes core dump

    System Configuration
    --------------------
    Operating System : Linux 2.2.9

    PostgreSQL version : 6.5

    Compiler used : gcc 2.7.2.3

    Hardware:
    ---------
    RedHat 5.2

    Versions of other tools:
    ------------------------
    RedHat 5.2 standard

    --------------------------------------------------------------------------

    Problem Description:
    --------------------
    calling notify in pl/pgsql proc causes core dump

    --------------------------------------------------------------------------

    Test Case:
    ----------
    createdb bugs
    psql bugs
    \i /usr/src/postgresql-6.5/src/pl/plpgsql/src/mklang.sql
    create function a_func()
    returns
    char
    as
    'begin
    notify my_listen;
    return \'a\';
    end;'
    language
    'plpgsql'
    ;
    select a_func();


    --------------------------------------------------------------------------

    Solution:
    ---------


    --------------------------------------------------------------------------

    --
    /==============================\
    David Mansfield |
    david@cobite.com |
    \==============================/
  • Bruce Momjian at Sep 21, 1999 at 9:31 pm
    Can someone comment on this?

    Jan, this is yours.
    Actually, I made a 'workaround' fix to this problem (after posting this
    bug report). I'll append the patch. I don't pretend to understand the
    internals of this system, but my guess is that a simple statement like
    'notify xyz' went through the planner, got an spi_plan, but the 'plan
    list' was empty. So here's my patch, which may be fixing a symptom, not
    the cause, YMMV:

    --- pl_exec.c~ Wed May 26 03:07:39 1999
    +++ pl_exec.c Fri Jun 25 11:00:53 1999
    @@ -2482,6 +2482,10 @@

    plan = (Plan *) lfirst(spi_plan->ptlist);

    + /* it would seem as though plan can be null... --DAVID */
    + if (plan == NULL)
    + return;
    +
    /* ----------
    * 2. It must be a RESULT plan --> no scan's required
    * ----------

    ============================================================================
    POSTGRESQL BUG REPORT TEMPLATE
    ============================================================================


    Your name : David Mansfield
    Your email address : david@cobite.com

    Category : runtime: back-end
    Severity : serious

    Summary: calling notify in pl/pgsql proc causes core dump

    System Configuration
    --------------------
    Operating System : Linux 2.2.9

    PostgreSQL version : 6.5

    Compiler used : gcc 2.7.2.3

    Hardware:
    ---------
    RedHat 5.2

    Versions of other tools:
    ------------------------
    RedHat 5.2 standard

    --------------------------------------------------------------------------

    Problem Description:
    --------------------
    calling notify in pl/pgsql proc causes core dump

    --------------------------------------------------------------------------

    Test Case:
    ----------
    createdb bugs
    psql bugs
    \i /usr/src/postgresql-6.5/src/pl/plpgsql/src/mklang.sql
    create function a_func()
    returns
    char
    as
    'begin
    notify my_listen;
    return \'a\';
    end;'
    language
    'plpgsql'
    ;
    select a_func();


    --------------------------------------------------------------------------

    Solution:
    ---------


    --------------------------------------------------------------------------

    --
    /==============================\
    David Mansfield |
    david@cobite.com |
    \==============================/

    --
    Bruce Momjian | http://www.op.net/~candle
    maillist@candle.pha.pa.us | (610) 853-3000
    + If your life is a hard drive, | 830 Blythe Avenue
    + Christ can be your backup. | Drexel Hill, Pennsylvania 19026
  • Bruce Momjian at Sep 23, 1999 at 5:15 pm
    Is this fixed?

    Jan, this is yours.
    Actually, I made a 'workaround' fix to this problem (after posting this
    bug report). I'll append the patch. I don't pretend to understand the
    internals of this system, but my guess is that a simple statement like
    'notify xyz' went through the planner, got an spi_plan, but the 'plan
    list' was empty. So here's my patch, which may be fixing a symptom, not
    the cause, YMMV:

    --- pl_exec.c~ Wed May 26 03:07:39 1999
    +++ pl_exec.c Fri Jun 25 11:00:53 1999
    @@ -2482,6 +2482,10 @@

    plan = (Plan *) lfirst(spi_plan->ptlist);

    + /* it would seem as though plan can be null... --DAVID */
    + if (plan == NULL)
    + return;
    +
    /* ----------
    * 2. It must be a RESULT plan --> no scan's required
    * ----------

    ============================================================================
    POSTGRESQL BUG REPORT TEMPLATE
    ============================================================================


    Your name : David Mansfield
    Your email address : david@cobite.com

    Category : runtime: back-end
    Severity : serious

    Summary: calling notify in pl/pgsql proc causes core dump

    System Configuration
    --------------------
    Operating System : Linux 2.2.9

    PostgreSQL version : 6.5

    Compiler used : gcc 2.7.2.3

    Hardware:
    ---------
    RedHat 5.2

    Versions of other tools:
    ------------------------
    RedHat 5.2 standard

    --------------------------------------------------------------------------

    Problem Description:
    --------------------
    calling notify in pl/pgsql proc causes core dump

    --------------------------------------------------------------------------

    Test Case:
    ----------
    createdb bugs
    psql bugs
    \i /usr/src/postgresql-6.5/src/pl/plpgsql/src/mklang.sql
    create function a_func()
    returns
    char
    as
    'begin
    notify my_listen;
    return \'a\';
    end;'
    language
    'plpgsql'
    ;
    select a_func();


    --------------------------------------------------------------------------

    Solution:
    ---------


    --------------------------------------------------------------------------

    --
    /==============================\
    David Mansfield |
    david@cobite.com |
    \==============================/

    --
    Bruce Momjian | http://www.op.net/~candle
    maillist@candle.pha.pa.us | (610) 853-3000
    + If your life is a hard drive, | 830 Blythe Avenue
    + Christ can be your backup. | Drexel Hill, Pennsylvania 19026
  • David Mansfield at Sep 23, 1999 at 8:55 pm
    Is this fixed?
    Yes. This patch (with a different comment) was applied to one of the
    recent versions. Call it closed!

    David

    Jan, this is yours.
    Actually, I made a 'workaround' fix to this problem (after posting this
    bug report). I'll append the patch. I don't pretend to understand the
    internals of this system, but my guess is that a simple statement like
    'notify xyz' went through the planner, got an spi_plan, but the 'plan
    list' was empty. So here's my patch, which may be fixing a symptom, not
    the cause, YMMV:

    --- pl_exec.c~ Wed May 26 03:07:39 1999
    +++ pl_exec.c Fri Jun 25 11:00:53 1999
    @@ -2482,6 +2482,10 @@

    plan = (Plan *) lfirst(spi_plan->ptlist);

    + /* it would seem as though plan can be null... --DAVID */
    + if (plan == NULL)
    + return;
    +
    /* ----------
    * 2. It must be a RESULT plan --> no scan's required
    * ----------

    ============================================================================
    POSTGRESQL BUG REPORT TEMPLATE
    ============================================================================


    Your name : David Mansfield
    Your email address : david@cobite.com

    Category : runtime: back-end
    Severity : serious

    Summary: calling notify in pl/pgsql proc causes core dump

    System Configuration
    --------------------
    Operating System : Linux 2.2.9

    PostgreSQL version : 6.5

    Compiler used : gcc 2.7.2.3

    Hardware:
    ---------
    RedHat 5.2

    Versions of other tools:
    ------------------------
    RedHat 5.2 standard

    --------------------------------------------------------------------------

    Problem Description:
    --------------------
    calling notify in pl/pgsql proc causes core dump

    --------------------------------------------------------------------------

    Test Case:
    ----------
    createdb bugs
    psql bugs
    \i /usr/src/postgresql-6.5/src/pl/plpgsql/src/mklang.sql
    create function a_func()
    returns
    char
    as
    'begin
    notify my_listen;
    return \'a\';
    end;'
    language
    'plpgsql'
    ;
    select a_func();


    --------------------------------------------------------------------------

    Solution:
    ---------


    --------------------------------------------------------------------------

    --
    /==============================\
    David Mansfield |
    david@cobite.com |
    \==============================/
    --
    /==============================\
    David Mansfield |
    david@cobite.com |
    \==============================/

Related Discussions

Discussion Navigation
viewthread | post
Discussion Overview
grouppgsql-ports @
categoriespostgresql
postedJun 24, '99 at 8:50p
activeSep 23, '99 at 8:55p
posts9
users4
websitepostgresql.org
irc#postgresql

People

Translate

site design / logo © 2022 Grokbase