Re: [HACKERS] XLog: how to log?

2004-05-10 Thread Simon Riggs
On Tue, 2004-05-11 at 04:05, Tom Lane wrote: > Alvaro Herrera <[EMAIL PROTECTED]> writes: > > A simple idea would be to add the committed subxact Xids to the > > xl_xact_commit and xl_xact_abort structs. However, this will be a > > variable length array, and those structs already have variable len

Re: [HACKERS] tinyint

2004-05-10 Thread Shachar Shemesh
Gaetano Mendola wrote: Hi all, someone of you sent me an example of a new data type called tinyint. Hi. I seen that is using a well directory structure just for compile a tinyint.c. Yes. The idea was to have it do more than just this one type. Maybe a bad idea, on retrospect. Is there any "empt

Re: [HACKERS] XLog: how to log?

2004-05-10 Thread Tom Lane
Alvaro Herrera <[EMAIL PROTECTED]> writes: > A simple idea would be to add the committed subxact Xids to the > xl_xact_commit and xl_xact_abort structs. However, this will be a > variable length array, and those structs already have variable length > arrays at the end (filenodes to drop). I don't

[HACKERS] tinyint

2004-05-10 Thread Gaetano Mendola
Hi all, someone of you sent me an example of a new data type called tinyint. I seen that is using a well directory structure just for compile a tinyint.c. Is there any "empty" structure that I have to use in order to delivery compile my email.c or I can just use a single line command: gcc email.c -

Re: [HACKERS] Bug in renaming view columns

2004-05-10 Thread Christopher Kings-Lynne
Take it a little further: if we extend ALTER TABLE to be able to alter view column types, would you expect CREATE OR REPLACE VIEW to stop checking that the column types didn't change? I'd argue that that's a real bad idea. If you want the view's output signature to change, you should have to use

Re: [HACKERS] email built in type

2004-05-10 Thread Gaetano Mendola
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Neil Conway wrote: | On Sat, 2004-05-08 at 13:10, Gaetano Mendola wrote: | |>However an *add on* is an *add on* with not the same eligibility |>of the main project | | | I'd suggest implementing the custom datatype first and packaging it | separately (

[HACKERS] XLog: how to log?

2004-05-10 Thread Alvaro Herrera
Hackers, I'm trying to figure out how to log the subtransaction operations. A simple idea would be to add the committed subxact Xids to the xl_xact_commit and xl_xact_abort structs. However, this will be a variable length array, and those structs already have variable length arrays at the end (f

Re: [HACKERS] dollar quoting

2004-05-10 Thread Andrew Dunstan
Andrew Dunstan wrote: Tom Lane wrote: Andrew Dunstan <[EMAIL PROTECTED]> writes: After staring at the code for a long time, I think I see how to do this. It's complicated a bit by the fact that $ is a valid identifier character. So my current thinking is to say that if we see $ not in a quo

Re: [HACKERS] PITR logging control program

2004-05-10 Thread Simon Riggs
Further design plans for PITR...as posted previously, Bruce and I had a long discussion recently to iron out the major thinking and a good deal of the detail also. In overview, major change is introducing an ARCHIVE process running under control of the Postmaster, similar to Stats collector. Due t

Re: [HACKERS] email built in type

2004-05-10 Thread Neil Conway
On Sat, 2004-05-08 at 13:10, Gaetano Mendola wrote: > However an *add on* is an *add on* with not the same eligibility > of the main project I'd suggest implementing the custom datatype first and packaging it separately (say, via pgFoundry) -- it can always be moved into the core distribution at s

Re: [HACKERS] Adding MERGE to the TODO list (resend with subject)

2004-05-10 Thread Bruce Momjian
Hannu Krosing wrote: > Bruce Momjian kirjutas E, 10.05.2004 kell 06:58: > > Added to TODO: > > > > * Add MERGE command that does UPDATE, or on failure, INSERT > > perhaps the issue raised about BEFOR/AFTER INSERT/UPDATE TRIGGERS and > RULES should get a mention in the TODO (... needs to be di

Re: [HACKERS] Adding MERGE to the TODO list (resend with subject)

2004-05-10 Thread Hannu Krosing
Bruce Momjian kirjutas E, 10.05.2004 kell 06:58: > Added to TODO: > > * Add MERGE command that does UPDATE, or on failure, INSERT perhaps the issue raised about BEFOR/AFTER INSERT/UPDATE TRIGGERS and RULES should get a mention in the TODO (... needs to be discussed, or somesuch ...) --

Re: [HACKERS] [PATCHES] patch for different join result order on regression test

2004-05-10 Thread Tom Lane
Bruce Momjian <[EMAIL PROTECTED]> writes: > Agreed it is time to get this regression test working on MinGW. I > suggest we add ORDER BY to the needed queries. Please do not do that. This is a join test, not an ORDER BY test, and I don't want it ignoring unordered join methods.

Re: [HACKERS] Bug in renaming view columns

2004-05-10 Thread Tom Lane
Christopher Kings-Lynne <[EMAIL PROTECTED]> writes: > Are we supposed to be able to rename view columns or not? You can't if > you replace the view, but you can if you rename the column. I think that's a definition disagreement, not a bug. Replacing a view is not supposed to change the external

Re: [HACKERS] [PATCHES] patch for different join result order on regression

2004-05-10 Thread Bruce Momjian
Tom Lane wrote: > Bruce Momjian <[EMAIL PROTECTED]> writes: > > Agreed it is time to get this regression test working on MinGW. I > > suggest we add ORDER BY to the needed queries. > > Please do not do that. This is a join test, not an ORDER BY test, > and I don't want it ignoring unordered join

Re: [HACKERS] [PATCHES] patch for different join result order on regression test

2004-05-10 Thread Bruce Momjian
Agreed it is time to get this regression test working on MinGW. I suggest we add ORDER BY to the needed queries. Tom as hoping to come up with a better solution, but at this point I think we should just add the ORDER BY rather than adding a different output file. Comments?

Re: [HACKERS] signal 11 on AIX: 7.4.2

2004-05-10 Thread Andrew Sullivan
On Wed, Apr 28, 2004 at 03:56:55PM -0400, Andrew Sullivan wrote: > On Mon, Apr 26, 2004 at 03:19:21PM -0400, Bruce Momjian wrote: > > > > Has this been resolved? > it elsewhere. I've been trying some alternative approaches to > causing it today, and so far no luck. On the weekend, we ran a set

[HACKERS] Linux 2.6.6 also

2004-05-10 Thread Gregory Stark
This patch also looks relevant to Postgres for two reasons. This part seems like it might expose some bugs that otherwise might have remained hidden: This affects I/O scheduling potentially quite significantly. It is no longer the case that the kernel will submit pages for

[HACKERS] Linux 2.6.6 changes

2004-05-10 Thread Gregory Stark
I'm not sure how important these changes are, but perhaps we'll find fsync faster than O_SYNC on linux 2.6.6 where O_SYNC was faster before? <[EMAIL PROTECTED]> [PATCH] ext3 fsync() and fdatasync() speedup ext3's fsync/fdatasync implementation is currently syncing the in

[HACKERS] Bug in renaming view columns

2004-05-10 Thread Christopher Kings-Lynne
Hi, Are we supposed to be able to rename view columns or not? You can't if you replace the view, but you can if you rename the column. test=# create view test as select 1 as a; CREATE VIEW test=# \d test View "public.test" Column | Type | Modifiers +-+--- a

Re: [HACKERS] CLUSTER locking

2004-05-10 Thread Tom Lane
Christopher Kings-Lynne <[EMAIL PROTECTED]> writes: >> No. Committing the relfilenode swap would cut the knees off of any >> active scan on the old file. > Could it upgrade its lock to exclusive just before doing the swap? That is a recipe for deadlock failures. regards,

Re: [HACKERS] CLUSTER locking

2004-05-10 Thread Christopher Kings-Lynne
No. Committing the relfilenode swap would cut the knees off of any active scan on the old file. Could it upgrade its lock to exclusive just before doing the swap? Chris ---(end of broadcast)--- TIP 3: if posting/reading through Usenet, please send

Re: [HACKERS] Comments on all system objects

2004-05-10 Thread Euler Taveira de Oliveira
Hi all, > > The ability to comment on all types of catalog objects or actual > > comments on all predefined catalog objects? Both are more or less > > reasonable. But I think we should have some sort of > > internationalization mechanism for the actual comments. > > One of my first commits f

Re: [HACKERS] CLUSTER locking

2004-05-10 Thread Tom Lane
Christopher Kings-Lynne <[EMAIL PROTECTED]> writes: > Is there a way the CLUSTER command can be changed to not take an > exclusive lock on the table, No. Committing the relfilenode swap would cut the knees off of any active scan on the old file. regards, tom lane --

[HACKERS] CLUSTER locking

2004-05-10 Thread Christopher Kings-Lynne
Hi, Is there a way the CLUSTER command can be changed to not take an exclusive lock on the table, and instead allow reads on the old table and index, just preventing writes? Chris ---(end of broadcast)--- TIP 2: you can get off all lists at once

Re: [HACKERS] Comments on all system objects

2004-05-10 Thread Fabien COELHO
> Bulk? Redundancy with the SGML documentation? > > There are already more than enough places to have to edit documentation > when we change/add a column in a system catalog. I don't see the value > of adding another, especially not one that's constrained to one line. > And what will comments on