Postgres Pro
Downloads
Home   >   mailing lists

Thread: Open 6.5 items

Open 6.5 items

From
Bruce Momjian
Date:
SELECT * FROM test WHERE test IN (SELECT * FROM test) fails with strange error
Regression test for new Numeric type
refint problems
invalidate cache on elog failure, abort is OK
large objects security problem
spinlock stuck problem
benchmark performance problem, index lookups on entries with many dup keys
add more detail inref/lock.sgml, ref/set.sgml to reflect MVCC & locking changes

Generate Admin, User, hardcopy postscript
Generate HISTORY from sgml sources.


--  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,
Pennsylvania19026

Re: [HACKERS] Open 6.5 items

From
Vadim Mikheev
Date:
Bruce Momjian wrote:
> 
> SELECT * FROM test WHERE test IN (SELECT * FROM test) fails with strange error
> Regression test for new Numeric type
> refint problems
> invalidate cache on elog failure, abort is OK

Isn't it fixed by Tom?

> large objects security problem
> spinlock stuck problem

Old one, possible when elog(FATAL) and elog(ERROR) inside bufmgr.

> benchmark performance problem, index lookups on entries with many dup keys
> add more detail inref/lock.sgml, ref/set.sgml to reflect MVCC & locking changes                       ^^^^^^^^^
Doing it right now.

> 
> Generate Admin, User, hardcopy postscript
> Generate HISTORY from sgml sources.

+ vacuum problems (two assertions) in the case of high update activity on the same table
 Seems fixed, though I wonder why it's fixed by changes I made, -:(

+ 'Deadlock detected' errors when running a row-locking query (SELECT ... ORDER BY FOR UPDATE in PL/pgSQL)
  Bug-report was posted today to ports list.  I hope to address it before release.

Vadim

Re: [HACKERS] Open 6.5 items

From
Bruce Momjian
Date:
> Bruce Momjian wrote:
> > 
> > SELECT * FROM test WHERE test IN (SELECT * FROM test) fails with strange error
> > Regression test for new Numeric type
> > refint problems
> > invalidate cache on elog failure, abort is OK
> 
> Isn't it fixed by Tom?

Removed.

> 
> > large objects security problem
> > spinlock stuck problem
> 
> Old one, possible when elog(FATAL) and elog(ERROR) inside bufmgr.

Removed. Right?

> 
> > benchmark performance problem, index lookups on entries with many dup keys
> > add more detail inref/lock.sgml, ref/set.sgml to reflect MVCC & locking changes
>                         ^^^^^^^^^
> Doing it right now.

OK.

> 
> > 
> > Generate Admin, User, hardcopy postscript
> > Generate HISTORY from sgml sources.
> 
> + vacuum problems (two assertions) in the case of high update
>   activity on the same table
> 
>   Seems fixed, though I wonder why it's fixed by changes I made, -:(
> 
> + 'Deadlock detected' errors when running a row-locking query
>   (SELECT ... ORDER BY FOR UPDATE in PL/pgSQL)
> 

I will leave these off the list.

--  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,
Pennsylvania19026

Re: [HACKERS] Open 6.5 items

From
Vadim Mikheev
Date:
Bruce Momjian wrote:
> 
> > > spinlock stuck problem
> >
> > Old one, possible when elog(FATAL) and elog(ERROR) inside bufmgr.
> 
> Removed. Right?

No time to fix it for 6.5, leave them for future.

Vadim