Re: [pgsql-advocacy] [HACKERS] [GENERAL] Postgresql AMD x86-64
Bruce Momjian writes: The following patch automatically enables 64-bit mode on AMD opteron. We already had spinlock support for it, but I added some comments. This sort of thing belongs into the template file, not directly in configure.in. -- Peter Eisentraut [EMAIL PROTECTED] ---(end of broadcast)--- TIP 4: Don't 'kill -9' the postmaster
Re: [HACKERS] dblink_ora - a first shot on Oracle ...
Bruce Momjian writes: Joe, I can do the configure detection of the Oracle library needed for /contrib. I have a philosophical problem with putting Oracle detection code into the PostgreSQL build system. That way, you create a dependency that Oracle needs to be installed before you install PostgreSQL, but of course we want PostgreSQL to be first. Consider where this will be going. Soon, configure will be full with detection code for half a dozen other database systems. Perhaps a plug-in architecture for dblink is a better solution. Or make dblink a separate project. -- Peter Eisentraut [EMAIL PROTECTED] ---(end of broadcast)--- TIP 5: Have you checked our extensive FAQ? http://www.postgresql.org/docs/faqs/FAQ.html
Re: [HACKERS] Reinventing the wheel...
Sean Chittenden wrote: To prevent lib naming collisions with machines that have libevent installed, I plan on renaming all of the functions from event_* to pgevent_*. libevent also has the appropriate autoconf goo to make detection of the right library pretty seamless. It even supports the new Linux interface epoll. Please don't. For those of us using an OS that has libevent as standard, this is just pollution and perhaps you should instead look to carry the portable code in the distribution and build/link to it if the normal libevent is not available (configure will happily check for you). Peter ---(end of broadcast)--- TIP 4: Don't 'kill -9' the postmaster
[HACKERS] anoncvs problems.
When doing cvs update I get: cvs server: Updating contrib/tsearch2 cvs server: failed to create lock directory for `/projects/cvsroot/pgsql-server/contrib/tsearch2' (/projects/cvsroot/pgsql-server/contrib/tsearch2/#cvs.lock): Permission denied cvs server: failed to obtain dir lock in repository `/projects/cvsroot/pgsql-server/contrib/tsearch2' cvs [server aborted]: read lock failed - giving up Kurt ---(end of broadcast)--- TIP 7: don't forget to increase your free space map settings
Re: [HACKERS] anoncvs problems.
Mark, you don't set right umask for me, IMHO. Kurt Roeckx wrote: When doing cvs update I get: cvs server: Updating contrib/tsearch2 cvs server: failed to create lock directory for `/projects/cvsroot/pgsql-server/contrib/tsearch2' (/projects/cvsroot/pgsql-server/contrib/tsearch2/#cvs.lock): Permission denied cvs server: failed to obtain dir lock in repository `/projects/cvsroot/pgsql-server/contrib/tsearch2' cvs [server aborted]: read lock failed - giving up Kurt ---(end of broadcast)--- TIP 7: don't forget to increase your free space map settings -- Teodor Sigaev E-mail: [EMAIL PROTECTED] ---(end of broadcast)--- TIP 1: subscribe and unsubscribe commands go to [EMAIL PROTECTED]
Re: [HACKERS] anoncvs problems.
check it now, should be fine ... On Mon, 21 Jul 2003, Kurt Roeckx wrote: When doing cvs update I get: cvs server: Updating contrib/tsearch2 cvs server: failed to create lock directory for `/projects/cvsroot/pgsql-server/contrib/tsearch2' (/projects/cvsroot/pgsql-server/contrib/tsearch2/#cvs.lock): Permission denied cvs server: failed to obtain dir lock in repository `/projects/cvsroot/pgsql-server/contrib/tsearch2' cvs [server aborted]: read lock failed - giving up Kurt ---(end of broadcast)--- TIP 7: don't forget to increase your free space map settings Marc G. Fournier ICQ#7615664 IRC Nick: Scrappy Systems Administrator @ hub.org primary: [EMAIL PROTECTED] secondary: [EMAIL PROTECTED]|postgresql}.org ---(end of broadcast)--- TIP 6: Have you searched our list archives? http://archives.postgresql.org
Re: [HACKERS] dblink_ora - a first shot on Oracle ...
Of course, PostgreSQL will still install without the Oracle libraries. Of course, if you want dblink to use Oracle libraries, you have to install the Oracle libraries first, or rerun configure after you install them and reinstall dblink. I don't see the problem. --- Peter Eisentraut wrote: Bruce Momjian writes: Joe, I can do the configure detection of the Oracle library needed for /contrib. I have a philosophical problem with putting Oracle detection code into the PostgreSQL build system. That way, you create a dependency that Oracle needs to be installed before you install PostgreSQL, but of course we want PostgreSQL to be first. Consider where this will be going. Soon, configure will be full with detection code for half a dozen other database systems. Perhaps a plug-in architecture for dblink is a better solution. Or make dblink a separate project. -- Peter Eisentraut [EMAIL PROTECTED] -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 9: the planner will ignore your desire to choose an index scan if your joining column's datatypes do not match
Re: [HACKERS] dblink_ora - a first shot on Oracle ...
I don't see the problem. How about a (simple!) configure process in the dblink directory only which detects the various items. signature.asc Description: This is a digitally signed message part
Re: [HACKERS] anoncvs problems.
Teodor Sigaev [EMAIL PROTECTED] writes: Mark, you don't set right umask for me, IMHO. There seems to be more to it than that: I can do cvs update just fine from the master server. So I'd guess that the issue is not with your own permissions, but with what happens when a new directory is propagated to the anoncvs slave server. regards, tom lane ---(end of broadcast)--- TIP 6: Have you searched our list archives? http://archives.postgresql.org
Re: [HACKERS] dblink_ora - a first shot on Oracle ...
Rod Taylor wrote: -- Start of PGP signed section. I don't see the problem. How about a (simple!) configure process in the dblink directory only which detects the various items. I thought of that but configure seems so confusing that setting up another on in a contrib directory seemed pretty hard. -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 4: Don't 'kill -9' the postmaster
Re: [HACKERS] anoncvs problems.
On Mon, Jul 21, 2003 at 10:11:38AM -0300, The Hermit Hacker wrote: check it now, should be fine ... Works now, thanks. Kurt ---(end of broadcast)--- TIP 8: explain analyze is your friend
Re: [HACKERS] dblink_ora - a first shot on Oracle ...
Bruce Momjian [EMAIL PROTECTED] writes: I don't see the problem. I tend to agree with Peter: if dblink is going to start depending on stuff outside Postgres, it ought to be become a separate project, if only to simplify distribution and configuration issues. Perhaps it could be split into two parts, a PG-specific part and a cross-DBMS part? regards, tom lane PS: Has anyone looked any further at the SQL-MED standard? ISTM that's where we ought to head in the long run. ---(end of broadcast)--- TIP 4: Don't 'kill -9' the postmaster
Re: [HACKERS] dblink_ora - a first shot on Oracle ...
Tom Lane wrote: I tend to agree with Peter: if dblink is going to start depending on stuff outside Postgres, it ought to be become a separate project, if only to simplify distribution and configuration issues. Perhaps it could be split into two parts, a PG-specific part and a cross-DBMS part? regards, tom lane PS: Has anyone looked any further at the SQL-MED standard? ISTM that's where we ought to head in the long run. I think for that very reason (SQL-MED) we need to come to terms with this issue. If/when connections to external data sources is in the backend, you'll have those exact same dependencies. And in fact, we do today: consider '--with-openssl' or '--with-tcl'. I had always assumed we would need '--with-oracle', '--with-jdbc', etc (or whatever you want to call them) to support backend connections to external sources. And this discussion is the very reason I was hesitant to pursue dblink_ora or jdbclink now, because I didn't think people would be comfortable with configure options to support a contrib library. Joe ---(end of broadcast)--- TIP 7: don't forget to increase your free space map settings
Re: [HACKERS] dblink_ora - a first shot on Oracle ...
Joe Conway writes: I think for that very reason (SQL-MED) we need to come to terms with this issue. If/when connections to external data sources is in the backend, you'll have those exact same dependencies. No, SQL-MED is a framework to plug in different connectors -- exactly what some are suggesting for dblink. -- Peter Eisentraut [EMAIL PROTECTED] ---(end of broadcast)--- TIP 8: explain analyze is your friend
Re: [HACKERS] dblink_ora - a first shot on Oracle ...
Joe Conway wrote: PS: Has anyone looked any further at the SQL-MED standard? ISTM that's where we ought to head in the long run. I think for that very reason (SQL-MED) we need to come to terms with this issue. If/when connections to external data sources is in the backend, you'll have those exact same dependencies. And in fact, we do today: consider '--with-openssl' or '--with-tcl'. I had always assumed we would need '--with-oracle', '--with-jdbc', etc (or whatever you want to call them) to support backend connections to external sources. And this discussion is the very reason I was hesitant to pursue dblink_ora or jdbclink now, because I didn't think people would be comfortable with configure options to support a contrib library. I know we normally require a configure flag to look for special capabilities, like ssl, but I thought we could skip that because it was a /contrib and just look by default, but I now see that people don't want to take that step. I thought dblink was too integrated in the backend code to be a separate project. -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 5: Have you checked our extensive FAQ? http://www.postgresql.org/docs/faqs/FAQ.html
Re: [HACKERS] dblink_ora - a first shot on Oracle ...
On 7/21/2003 9:16 AM, Tom Lane wrote: Bruce Momjian [EMAIL PROTECTED] writes: I don't see the problem. I tend to agree with Peter: if dblink is going to start depending on stuff outside Postgres, it ought to be become a separate project, if only to simplify distribution and configuration issues. The ability to optionally link to another library does not necessitate a functional dependency on it. Perhaps it could be split into two parts, a PG-specific part and a cross-DBMS part? regards, tom lane PS: Has anyone looked any further at the SQL-MED standard? ISTM that's where we ought to head in the long run. ---(end of broadcast)--- TIP 4: Don't 'kill -9' the postmaster ---(end of broadcast)--- TIP 8: explain analyze is your friend
Re: [HACKERS] Bad permissions bug in 7.3 dump (and 7.4)?
Is there a TODO here? --- Peter Eisentraut wrote: On Mon, 14 Jul 2003, Tom Lane wrote: Probably the only real solution is to implement DROP-CASCADE-like checking when a privilege is revoked. Seems like rather a lot of work :-( Yes and yes. That's why the SQL standard goes on for pages and pages about REVOKE. It will be looked at eventually, just make sure someone is taking notes on the failure cases. -- Peter Eisentraut [EMAIL PROTECTED] ---(end of broadcast)--- TIP 1: subscribe and unsubscribe commands go to [EMAIL PROTECTED] -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 9: the planner will ignore your desire to choose an index scan if your joining column's datatypes do not match
Re: [HACKERS] dblink_ora - a first shot on Oracle ...
Hans, I am a little confused about what you are suggesting. Are you suggesting flag to the make of the contrib module rather than configure tests? I agree this is a killer feature for many people and would like to have it in 7.4. --- Hans-Jürgen Schönig wrote: I think for that very reason (SQL-MED) we need to come to terms with this issue. If/when connections to external data sources is in the backend, you'll have those exact same dependencies. And in fact, we do today: consider '--with-openssl' or '--with-tcl'. I had always assumed we would need '--with-oracle', '--with-jdbc', etc (or whatever you want to call them) to support backend connections to external sources. And this discussion is the very reason I was hesitant to pursue dblink_ora or jdbclink now, because I didn't think people would be comfortable with configure options to support a contrib library. Joe If dblink was a core module I would say that adding the configure stuff would be very natural. Since this is contrib stuff I was not that sure about configure anymore. We will need additional flags for external data sources in the (hopefully) near future so I think we should add it. Personally I tend to think about a solution like that. dblink has a great future and many people simply love it (I cannot think of a customer who does not like it - it is a killer feature): - implement the concepts proposed by Joe on this list yesterday (I am talking about the functions dblink should provide) - add support to configure - merge dblink with dblink_ora as soon as the changes are ready - adapt jdbc_link and merge it with dblink - implement dblink_db2, dblink_csv, dblink_xml, and maybe some others - leave it in contrib because this way it will be shipped with the core distribution and people will use it more frequently I hope that I will finish the Oracle stuff (named connections, ...) within the next 3 days. Regards, Hans -- Cybertec Geschwinde u Schoenig Ludo-Hartmannplatz 1/14, A-1160 Vienna, Austria Tel: +43/2952/30706; +43/664/233 90 75 www.cybertec.at, www.postgresql.at, kernel.cybertec.at -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 6: Have you searched our list archives? http://archives.postgresql.org
Re: [HACKERS] Why are triggers semi-deferred?
Added to TODO: * Have AFTER triggers execute after the appropriate SQL statement in a function, not at the end of the function --- Philip Warner wrote: At 11:51 PM 1/06/2003 -0400, Bruce Momjian wrote: Does anyone have answers for these? I read the thread and don't 100% understand it all. My belief is that at least ROW triggers need fixing (7.3 doesn't have statement, not sure about 7.4). Currently, if you write a plpgsql procedure which calls more than one insert/update/delete statements, the AFTER triggers for all of these statements will not fire until after the procedure exits. They should fire either just after each row is updated, or just after the most immediately enclosing statement executes. I think the thread wanted the latter. So, if we have a table with two rows, and a BEFORE and AFTER trigger, and a plpgsql procedure that updates all rows twice, then we should have: procedure called procedure executes first update before trigger fires(row 1) before trigger fires(row 2) row 1 updated row 2 updated after trigger fires(row 1) after trigger fires(row 2) procedure executes second update before trigger fires(row 1) before trigger fires(row 2) row 1 updated row 2 updated after trigger fires(row 1) after trigger fires(row 2) procedure exits What we have in 7.3 is: procedure called procedure executes first update before trigger fires(row 1) before trigger fires(row 2) row 1 updated row 2 updated procedure executes second update before trigger fires(row 1) before trigger fires(row 2) row 1 updated row 2 updated procedure exits after trigger fires(row 1) after trigger fires(row 2) after trigger fires(row 1) after trigger fires(row 2) IIRC, the thread did not really discuss whether do intersperse the BEFORE executions with the updates, but doing them all before seems consistent. Apologies is this has been covered elsewhere... Philip Warner| __---_ Albatross Consulting Pty. Ltd. |/ - \ (A.B.N. 75 008 659 498) | /(@) __---_ Tel: (+61) 0500 83 82 81 | _ \ Fax: (+61) 03 5330 3172 | ___ | Http://www.rhyme.com.au |/ \| |---- PGP key available upon request, | / and from pgp5.ai.mit.edu:11371 |/ -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 8: explain analyze is your friend
Re: [HACKERS] php with postgres
Added to TODO: o Add PL/PHP (Joe, Jan) --- Joe Conway wrote: Jan Wieck wrote: I had been briefly talking with Marcus Boerger (included in CC) from the PHP team about it. He knows the PHP5 SAPI embed well. Can you include him into the team (if not already)? Sure! From what I know about this SAPI I think the PL/Tcl implementation would be a good point to start from, as it looks very similar with respect to the possibilities. I was going to start from PL/R, which is a descendent of PL/Tcl -- reason being, in PL/R I've already got SRF/table-function support and polymorphic argument/return-type support working. Also, I've done a fair amount of work to preserve arrays and composite types as they move back-and-forth. My plan is to add a few missing things to PL/R over the next couple (or so) weeks, and then start PL/PHP from that: 1) Cache lookup based on function oid and argument signature ala the patch I recently sent in (and improved by Tom) for PL/pgSQL -- this is needed to properly support polymorphic arguments. 2) Trigger support -- just haven't needed this so far, but we'll want it in PL/PHP, so I may as well add it to PL/R too. 3) Re-add nested error handling -- I removed this from PL/R early on just to simplify life. Should be easy to drop back in. I've read some examples posted regarding the PHP embed SAPI, and it looks very similar to the R interpreter also. It should be fairly easy to drop the PHP embed calls in for the libR calls. The bulk of the work will be in modifying the data conversion functions that map Postgres composite types and arrays to similar structures in PHP. Help on the PHP side of things would be most appreciated, because that's the part I'm least familiar with. Joe ---(end of broadcast)--- TIP 6: Have you searched our list archives? http://archives.postgresql.org -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 5: Have you checked our extensive FAQ? http://www.postgresql.org/docs/faqs/FAQ.html
Re: [HACKERS] php with postgres
Glad you got in touch with the right guys. Joe and Jan have both talked about doing PlPHP for a while. Marcus, would you check if PHP is using RESET ALL when passing persistent connection to new clients? We added that capability a few releases ago, specifically for PHP persistent connections, but I don't think that ever got into the PHP code. --- Marcus B?rger wrote: Hello ivan, Sunday, July 13, 2003, 10:12:43 PM, you wrote: i what aoubt stream ? i in plpgsql you can just write command INSERT ... or DELETE i if you want sht like this in php you need to correct zend i think . i in php all var is declared as variant type but we need look at realy type. In php we only have a few base types (int, float, bool, string) and some more complex types like array and objects. The general idea is that at least the base types can be converted without notice. This might be a problem when integrating php into postgres but i guess everything can be solved the php way. Since i also have zend commit rights i could fix things in this manner as long as the language itself doesn't change in any way. i I have other view, to first write php interpreter to postgres, and then i write a translator, which translate plphp code to C code . I cound be only i a another way, to remember about speed (compiled code is always faster i then src ) . Php source will be to testing, and to relese will be option i to translate this src to C src and then compile it. The general idea should be to leave php as is. That is, it is an interpreter. During LT we were again able to speed it up very much. So performance difference from interpreter to a real php to c compiler shouldn't be a problem for the moment. Also i thing when someone consideres using php inside his database he a) does it because he doesn't know any other language or b) he uses very advanced language features. In both cases the performance problem is no issue. Additionally there are already tokenizers out which remove the compile step. So atm we only can't get rid of the interpreter overhead. Best regards, Marcusmailto:[EMAIL PROTECTED] ---(end of broadcast)--- TIP 3: if posting/reading through Usenet, please send an appropriate subscribe-nomail command to [EMAIL PROTECTED] so that your message can get through to the mailing list cleanly -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 8: explain analyze is your friend
Re: [HACKERS] vacuumdb can't be canceled
Sorry, my testing was on CVS version of PostgreSQL. --- Kenji Sugita wrote: Vacuumdb command can't be canceled by Control-C and VACUUM is still running. When wrong database name is specified to vacuumdb, cancellation is required to stop VACUUM FULL which runs long. Option -c of psql forget to set signal handler for 7.3 or prior. Vacuumdb have no signal handler of cancellation for 7.4devel. Kenji Sugita ---(end of broadcast)--- TIP 5: Have you checked our extensive FAQ? http://www.postgresql.org/docs/faqs/FAQ.html -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 1: subscribe and unsubscribe commands go to [EMAIL PROTECTED]
Re: [HACKERS] vacuumdb can't be canceled
Control-C works here. What platform are you on? Can you reproduce it? I just tried: psql -c 'select * from pg_class, pg_proc' test and control-C terminated the query. --- Kenji Sugita wrote: Vacuumdb command can't be canceled by Control-C and VACUUM is still running. When wrong database name is specified to vacuumdb, cancellation is required to stop VACUUM FULL which runs long. Option -c of psql forget to set signal handler for 7.3 or prior. Vacuumdb have no signal handler of cancellation for 7.4devel. Kenji Sugita ---(end of broadcast)--- TIP 5: Have you checked our extensive FAQ? http://www.postgresql.org/docs/faqs/FAQ.html -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 1: subscribe and unsubscribe commands go to [EMAIL PROTECTED]
Re: [HACKERS] php with postgres
Marcus B?rger wrote: BM Marcus, would you check if PHP is using RESET ALL when passing BM persistent connection to new clients? We added that capability a few BM releases ago, specifically for PHP persistent connections, but I don't BM think that ever got into the PHP code. Unfortunately we don't do so yet. Do i need to check for errors or can i do it unconditionally on conenction start? And i'd need to know how to check if it is available (like starting with which version). It first appeared in PostgreSQL version 7.2. It doesn't generate any failures. It just resets all SET settting to their defaults, in case the previous client modified them. -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 8: explain analyze is your friend
Re: [HACKERS] did you read my mails ?
What functions are they? --- ivan wrote: someone looked at my files function ?? ---(end of broadcast)--- TIP 7: don't forget to increase your free space map settings -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 1: subscribe and unsubscribe commands go to [EMAIL PROTECTED]
[HACKERS] tsearch2 for 7.3.X
Hi there, seems we'll have 7.3.4 release. Is't worth to submit new tsearch2 module for this release ? People could play with new module without waiting 7.4 release. Regards, Oleg _ Oleg Bartunov, sci.researcher, hostmaster of AstroNet, Sternberg Astronomical Institute, Moscow University (Russia) Internet: [EMAIL PROTECTED], http://www.sai.msu.su/~megera/ phone: +007(095)939-16-83, +007(095)939-23-83 ---(end of broadcast)--- TIP 4: Don't 'kill -9' the postmaster
Re: [HACKERS] tsearch2 for 7.3.X
We don't normally issue new features in minor releases, but for a /contrib, we could consider it. --- Oleg Bartunov wrote: Hi there, seems we'll have 7.3.4 release. Is't worth to submit new tsearch2 module for this release ? People could play with new module without waiting 7.4 release. Regards, Oleg _ Oleg Bartunov, sci.researcher, hostmaster of AstroNet, Sternberg Astronomical Institute, Moscow University (Russia) Internet: [EMAIL PROTECTED], http://www.sai.msu.su/~megera/ phone: +007(095)939-16-83, +007(095)939-23-83 ---(end of broadcast)--- TIP 4: Don't 'kill -9' the postmaster -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 9: the planner will ignore your desire to choose an index scan if your joining column's datatypes do not match
Re: [HACKERS] [GENERAL] Backwards index scan
Is this a TODO? --- Tom Lane wrote: [ reply redirected to a more appropriate list ] Dmitry Tkach [EMAIL PROTECTED] writes: I am not sure if this is really a bug, but it certainly looks like one to me... It's not a bug, but I agree that _bt_first can be inefficient if there are lots of matching keys. This is because there are *lots* (a few million) of matches for x=10, and _bt_first () scans through them *all* sequentually to get to the last one. I understand that with the generic approach to operators in postgres it is, probably, not very feasible to try and teach _bt_first () to handle this situation automatically (it would need to know how to get next/previous value for every indexable type)... I think what we'd want is variant versions of _bt_search and _bt_binsrch that locate the first entry greater than the specified target key, rather than the first entry greater than or equal to it. Given such positioning, all the _bt_first cases that involve stepping over more than one entry could be improved to require no more than one step. Not sure whether it'd be better to make clone versions of these functions, or to add a parameter to tell them which behavior is wanted. regards, tom lane ---(end of broadcast)--- TIP 6: Have you searched our list archives? http://archives.postgresql.org -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 4: Don't 'kill -9' the postmaster
Re: [HACKERS] tsearch2 for 7.3.X
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On Monday 21 July 2003 22:46, Bruce Momjian wrote: We don't normally issue new features in minor releases, but for a /contrib, we could consider it. --- Oleg Bartunov wrote: Hi there, seems we'll have 7.3.4 release. Is't worth to submit new tsearch2 module for this release ? People could play with new module without waiting 7.4 release. FWIW: I would very much appreciate a tsearch2 for 7.3 for testing without having to upgrade my db's to 7.4. One question tho, is it ready for production? It's the ranking support which I'm looking forward to. - -- Andreas Joseph Krogh [EMAIL PROTECTED] gpg public_key: http://dev.officenet.no/~andreak/public_key.asc - - When there is no content, there is no crap. -BEGIN PGP SIGNATURE- Version: GnuPG v1.0.7 (GNU/Linux) iD8DBQE/HFYtUopImDh2gfQRAvj8AJ90UoHrSfumA0C4wUhkzh7bzfEN0gCfVsri NFWJfB/6ILRA6RsbMPUdcTQ= =HvZQ -END PGP SIGNATURE- ---(end of broadcast)--- TIP 2: you can get off all lists at once with the unregister command (send unregister YourEmailAddressHere to [EMAIL PROTECTED])
Re: [HACKERS] php with postgres
Marcus B?rger wrote: Hello Bruce, Monday, July 21, 2003, 9:37:08 PM, you wrote: BM Marcus B?rger wrote: BM Marcus, would you check if PHP is using RESET ALL when passing BM persistent connection to new clients? We added that capability a few BM releases ago, specifically for PHP persistent connections, but I don't BM think that ever got into the PHP code. Unfortunately we don't do so yet. Do i need to check for errors or can i do it unconditionally on conenction start? And i'd need to know how to check if it is available (like starting with which version). BM It first appeared in PostgreSQL version 7.2. It doesn't generate any BM failures. It just resets all SET settting to their defaults, in case BM the previous client modified them. Committed for PHP 5.0 if there is a version 4.4 or 4.5 i'll commit the patch for that version, too. Now we need someone with a big site to test whether it brings any new problems. If there aren't problems i could probably also commit I think Theis originally complained about it, perhaps two years ago. to PHP 4.3.3. Btw. i don't do RESET ALL when creating the link because that shouldn't be necessary. Right. -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 4: Don't 'kill -9' the postmaster
Re: [HACKERS] duplicate define in elog.h
Tom Lane wrote: Joe Conway [EMAIL PROTECTED] writes: I was looking through elog.h and noticed this at about line 36: #define ERROR 20 /* user error - abort transaction; return * to known state */ #define ERROR 20 /* user error - abort transaction; return * to known state */ Good catch. Looks like it snuck in here: http://developer.postgresql.org/cvsweb.cgi/pgsql-server/src/include/utils/elog.h.diff?r1=1.41r2=1.42 Man, I am unsafe with a command prompt. :-) -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 7: don't forget to increase your free space map settings
Re: [HACKERS] did you read my mails ?
functions to open,read,write etc files On Mon, 21 Jul 2003, Bruce Momjian wrote: What functions are they? --- ivan wrote: someone looked at my files function ?? ---(end of broadcast)--- TIP 7: don't forget to increase your free space map settings -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 1: subscribe and unsubscribe commands go to [EMAIL PROTECTED] ---(end of broadcast)--- TIP 3: if posting/reading through Usenet, please send an appropriate subscribe-nomail command to [EMAIL PROTECTED] so that your message can get through to the mailing list cleanly
Re: [HACKERS] did you read my mails ?
I haven't seen those myself. --- ivan wrote: functions to open,read,write etc files On Mon, 21 Jul 2003, Bruce Momjian wrote: What functions are they? --- ivan wrote: someone looked at my files function ?? ---(end of broadcast)--- TIP 7: don't forget to increase your free space map settings -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 1: subscribe and unsubscribe commands go to [EMAIL PROTECTED] -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 2: you can get off all lists at once with the unregister command (send unregister YourEmailAddressHere to [EMAIL PROTECTED])
Re: [HACKERS] tsearch2 for 7.3.X
On Mon, 21 Jul 2003, Andreas Joseph Krogh wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On Monday 21 July 2003 22:46, Bruce Momjian wrote: We don't normally issue new features in minor releases, but for a /contrib, we could consider it. --- Oleg Bartunov wrote: Hi there, seems we'll have 7.3.4 release. Is't worth to submit new tsearch2 module for this release ? People could play with new module without waiting 7.4 release. FWIW: I would very much appreciate a tsearch2 for 7.3 for testing without having to upgrade my db's to 7.4. One question tho, is it ready for production? It's the ranking support which I'm looking forward to. I think it's production quality. Actually, we use it in our projects with 7.3.3. You may read docs on tsearch2 home page http://www.sai.msu.su/~megera/postgres/gist/tsearch/V2/ - -- Andreas Joseph Krogh [EMAIL PROTECTED] gpg public_key: http://dev.officenet.no/~andreak/public_key.asc - - When there is no content, there is no crap. -BEGIN PGP SIGNATURE- Version: GnuPG v1.0.7 (GNU/Linux) iD8DBQE/HFYtUopImDh2gfQRAvj8AJ90UoHrSfumA0C4wUhkzh7bzfEN0gCfVsri NFWJfB/6ILRA6RsbMPUdcTQ= =HvZQ -END PGP SIGNATURE- Regards, Oleg _ Oleg Bartunov, sci.researcher, hostmaster of AstroNet, Sternberg Astronomical Institute, Moscow University (Russia) Internet: [EMAIL PROTECTED], http://www.sai.msu.su/~megera/ phone: +007(095)939-16-83, +007(095)939-23-83 ---(end of broadcast)--- TIP 7: don't forget to increase your free space map settings
Re: [HACKERS] did you read my mails ?
File IO, now that's something I would like to see. I will need that REAL SOON NOW. RIck Bruce Momjian wrote: I haven't seen those myself. --- ivan wrote: functions to open,read,write etc files On Mon, 21 Jul 2003, Bruce Momjian wrote: What functions are they? --- ivan wrote: someone looked at my files function ?? ---(end of broadcast)--- TIP 7: don't forget to increase your free space map settings -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 1: subscribe and unsubscribe commands go to [EMAIL PROTECTED] -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 2: you can get off all lists at once with the unregister command (send unregister YourEmailAddressHere to [EMAIL PROTECTED]) ---(end of broadcast)--- TIP 5: Have you checked our extensive FAQ? http://www.postgresql.org/docs/faqs/FAQ.html
Re: [HACKERS] Criteria for contrib/ versus gborg?
[EMAIL PROTECTED] wrote: - allows us to say that PostgreSQL ships with field-tested replication in the source tree We have a winner! I think this one trumps all the rest. Can we say field-tested and Java in the same sentence? -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 3: if posting/reading through Usenet, please send an appropriate subscribe-nomail command to [EMAIL PROTECTED] so that your message can get through to the mailing list cleanly
Re: [HACKERS] php with postgres
Help on the PHP side of things would be most appreciated, because that's the part I'm least familiar with. SAPI/Embed in PHP is very experimental which means that it can be molded to suit PL/PHP needs. To my knowlege its only used as a plugin for irssi (irc client which you can script using PHP), and a very poorly implemented MySQL module. Anyway if you need any help with it just shout :) Edin ---(end of broadcast)--- TIP 1: subscribe and unsubscribe commands go to [EMAIL PROTECTED]
Re: [HACKERS] pgsql-server/src/backend/commands comment.c
Added to TODO: * Prevent COMMENT ON DATABASE from using a database name --- Rod Taylor wrote: -- Start of PGP signed section. For COMMENT ON DATABASE where database name is unknown or not the current database, emit a WARNING and do nothing, rather than raising ERROR. Per recent discussion in which we concluded this is the best way to deal with database dumps that are reloaded into a database of a new name. Please add a TODO note about allowing COMMENT ON DATABASE to function without providing a database name at all. -- End of PGP section, PGP failed! -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 5: Have you checked our extensive FAQ? http://www.postgresql.org/docs/faqs/FAQ.html
Re: [HACKERS] Another TODO for PL/pgSQL -- Dynamic colums
Added to TODO: o Allow PL/pgSQL to name columns by ordinal position, e.g. rec.(3) --- Josh Berkus wrote: Developers, While I realize that we already have a number of TODOs on the list for PL/pgSQL which nobody is working on, I'd like to propose one more. That way, when someday somebody takes this on, they'll have a full list of feature requests. This one gets requested about once per month on the SQL list: -- Allow naming of columns by ordinal position in PL/pgSQL, e.g. some_rec.(3) or NEW.(3) -- Josh Berkus Aglio Database Solutions San Francisco ---(end of broadcast)--- TIP 4: Don't 'kill -9' the postmaster -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 4: Don't 'kill -9' the postmaster
Re: [HACKERS] SELECT FOR UPDATE NOWAIT
Rod Taylor wrote: -- Start of PGP signed section. On Fri, 2003-07-18 at 19:46, Paulo Scardine wrote: My boss is asking for something like Oracle's SELECT FOR UPDATE NOWAIT. Is there any such feature? If no, should I look forward into implementing this? Any advice? Lookup STATEMENT_TIMEOUT and set it to a very short time. Some people have said they want to distinguish between a slow query (busy system) and waiting on a lock. I can particulary see wanting to do a NOWAIT only on exclusive locks --- not sure how many really want that, though. -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 6: Have you searched our list archives? http://archives.postgresql.org
Re: [HACKERS] php with postgres
Bruce Momjian wrote: Marcus B?rger wrote: BM Marcus, would you check if PHP is using RESET ALL when passing BM persistent connection to new clients? We added that capability a few BM releases ago, specifically for PHP persistent connections, but I don't BM think that ever got into the PHP code. Unfortunately we don't do so yet. Do i need to check for errors or can i do it unconditionally on conenction start? And i'd need to know how to check if it is available (like starting with which version). It first appeared in PostgreSQL version 7.2. It doesn't generate any failures. It just resets all SET settting to their defaults, in case the previous client modified them. It does generate the usual error if the current transaction block is in ABORT state. So the correct querystring to send would be something like ROLLBACK; RESET ALL Jan -- #==# # It's easier to get forgiveness for being wrong than for being right. # # Let's break this rule - forgive me. # #== [EMAIL PROTECTED] # ---(end of broadcast)--- TIP 5: Have you checked our extensive FAQ? http://www.postgresql.org/docs/faqs/FAQ.html
Re: [HACKERS] php with postgres
Jan Wieck wrote: Bruce Momjian wrote: Marcus B?rger wrote: BM Marcus, would you check if PHP is using RESET ALL when passing BM persistent connection to new clients? We added that capability a few BM releases ago, specifically for PHP persistent connections, but I don't BM think that ever got into the PHP code. Unfortunately we don't do so yet. Do i need to check for errors or can i do it unconditionally on conenction start? And i'd need to know how to check if it is available (like starting with which version). It first appeared in PostgreSQL version 7.2. It doesn't generate any failures. It just resets all SET settting to their defaults, in case the previous client modified them. It does generate the usual error if the current transaction block is in ABORT state. So the correct querystring to send would be something like ROLLBACK; RESET ALL Oh, I remember that now as part of the persistent connection code. As I remember, we told them to do BEGIN;COMMIT; to clear any open transaction state passed to the new client. Is that in there? If not, it has to be added too. ROLLBACK will generate an error if you are not in a transaction, so it would fill the logs with errors. -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 1: subscribe and unsubscribe commands go to [EMAIL PROTECTED]
Re: [pgsql-advocacy] [HACKERS] [GENERAL] Postgresql AMD x86-64
Peter Eisentraut wrote: Bruce Momjian writes: The following patch automatically enables 64-bit mode on AMD opteron. We already had spinlock support for it, but I added some comments. This sort of thing belongs into the template file, not directly in configure.in. I knew you were going to say that, but which template file? It is a product of the cpu/compiler, not a specific OS. -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 2: you can get off all lists at once with the unregister command (send unregister YourEmailAddressHere to [EMAIL PROTECTED])
[HACKERS] [Fwd: Re: Interesting thought....]
Hi guys, Wasn't sure if this is a valid idea, but Bruce's response made me think it was worth forwarding here to see if anyone had/has considered it before. As everyone is aware, I'm not up for coding it, but am mentioning it Just In Case it's deemed worthy of adding to the TODO list in case someone wants to pick up. :-) Regards and best wishes, Justin Clift Original Message Subject: Re: Interesting thought Date: Mon, 21 Jul 2003 14:13:49 -0400 (EDT) From: Bruce Momjian [EMAIL PROTECTED] To: Justin Clift [EMAIL PROTECTED] This would fix all our inheritance problems, and allow sequences to span multiple tables with an index constraint --- Justin Clift wrote: Hi Bruce, Here's an interesting thought Multi-table indexes... like multi-column, but instead of specifying columns in one table, they specify columns in several tables. That leads into thoughts about multi-table primary and secondary keys and multi-table foreign keys too. Wonder if it'd be useful in the real world? I reckon so. :-) Regards and best wishes, Justin Clift -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 3: if posting/reading through Usenet, please send an appropriate subscribe-nomail command to [EMAIL PROTECTED] so that your message can get through to the mailing list cleanly
Re: [HACKERS] [GENERAL] LinkServer
Here is someone looking for dblink with Oracle capability. It is not in CVS yet, but we are trying to get it for 7.4. You might be able to run /contrib/dblink from CVS in a few weeks without waiting for a 7.4 release. --- Richard Huxton wrote: On Tuesday 08 Jul 2003 7:58 am, Shridhar Daithankar wrote: On 8 Jul 2003 at 12:03, Madhavi Daroor wrote: Hi All, How do I link a postgres server and an Oracle server and get the data from an Oracle sever from a Postgres server? I'm using Postgres 7.2.3 on Red Hat linux 7.2 Does postgres have any function to do so? And what other configuration changes are need to achieve this? Well no. There is nothing out of box that would help here. But I did see mention of a dblink for oracle a week or two ago. The dblink package is in contrib and I'd suggest a search of the archives for dblink and oracle. -- Richard Huxton ---(end of broadcast)--- TIP 8: explain analyze is your friend -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 9: the planner will ignore your desire to choose an index scan if your joining column's datatypes do not match
Re: [HACKERS] php with postgres
Edin Kadribasic wrote: Help on the PHP side of things would be most appreciated, because that's the part I'm least familiar with. SAPI/Embed in PHP is very experimental which means that it can be molded to suit PL/PHP needs. To my knowlege its only used as a plugin for irssi (irc client which you can script using PHP), and a very poorly implemented MySQL module. Anyway if you need any help with it just shout :) Great! I'm likely to take you up on that before it's over ;-) It will probably be a few more weeks before I can get started on this, but it's definitely on my short list. Joe ---(end of broadcast)--- TIP 2: you can get off all lists at once with the unregister command (send unregister YourEmailAddressHere to [EMAIL PROTECTED])
Re: [HACKERS] tsearch2 for 7.3.X
Bruce Momjian [EMAIL PROTECTED] writes: We don't normally issue new features in minor releases, but for a /contrib, we could consider it. I can't see sticking code that hasn't been through any public beta testing into 7.3.4. Not even as contrib material --- how embarrassed would you be if the contrib tree then fails to build on some platform? 7.3 is long past the point where we should be adding new features to it. If there are people out there who want to try tsearch2 with 7.3, let them grab a separate tarball for it. regards, tom lane ---(end of broadcast)--- TIP 1: subscribe and unsubscribe commands go to [EMAIL PROTECTED]
Re: [HACKERS] PostgreSQL 7.3.3 and Intel C compiler
Bruce Momjian wrote: Hans-J?rgen Sch?nig wrote: This week I have done some performance tuning at a customer's office. We have beaten (demoralized) MS SQL and DB2 in serializable mode and DB2 in any transaction isolation level :). By the way: In case of very simple statements SERIALIZABLE is about 3 times faster than READ COMMITTED. I expected it to be faster but I did not expect this difference. Why was SERIALIZABLE faster? I know SERIALIZABLE doesn't have the rollback penalty in read-only queries, but I don't understand why it would be faster. To be honest I don't have the slightest idea. Maybe it has to do with snapshotting but I don't know precisely. In case of SERIALIZABLE all snapshots inside a transaction are the same - maybe this makes the big difference. I have no other explanation for that. There is one nifty detail which seems VERY strange to me: If serializable mode is set in postgresql.conf the system was 3 times faster (~ 7.5 sec. vs. 2.5sec). If serializable mode was set for every transaction (using set at the beginning of the transaction) serializable mode was as fast as read committed. We have done 90% cursor work and very simple queries (mostly queries such as DECLARE CURSOR x FOR SELECT * FROM ... WHERE a = b). I have no idea why PostgreSQL behaves like that but it seems to be a really good tweak because in this mode we beat any other database including SQL server on Windows 2003 (2.9sec) and IBM DB2 on Linux (12.6 seconds). I am sorry but I cannot provide you the tools we have used because we have a non disclosure agreement with the customer. I will try to verify this with my machines and a simple self-made benchmark. Regards, Hans -- Cybertec Geschwinde u Schoenig Ludo-Hartmannplatz 1/14, A-1160 Vienna, Austria Tel: +43/2952/30706; +43/664/233 90 75 www.cybertec.at, www.postgresql.at, kernel.cybertec.at ---(end of broadcast)--- TIP 7: don't forget to increase your free space map settings
Re: [HACKERS] dblink_ora - a first shot on Oracle ...
I think for that very reason (SQL-MED) we need to come to terms with this issue. If/when connections to external data sources is in the backend, you'll have those exact same dependencies. And in fact, we do today: consider '--with-openssl' or '--with-tcl'. I had always assumed we would need '--with-oracle', '--with-jdbc', etc (or whatever you want to call them) to support backend connections to external sources. And this discussion is the very reason I was hesitant to pursue dblink_ora or jdbclink now, because I didn't think people would be comfortable with configure options to support a contrib library. Joe If dblink was a core module I would say that adding the configure stuff would be very natural. Since this is contrib stuff I was not that sure about configure anymore. We will need additional flags for external data sources in the (hopefully) near future so I think we should add it. Personally I tend to think about a solution like that. dblink has a great future and many people simply love it (I cannot think of a customer who does not like it - it is a killer feature): - implement the concepts proposed by Joe on this list yesterday (I am talking about the functions dblink should provide) - add support to configure - merge dblink with dblink_ora as soon as the changes are ready - adapt jdbc_link and merge it with dblink - implement dblink_db2, dblink_csv, dblink_xml, and maybe some others - leave it in contrib because this way it will be shipped with the core distribution and people will use it more frequently I hope that I will finish the Oracle stuff (named connections, ...) within the next 3 days. Regards, Hans -- Cybertec Geschwinde u Schoenig Ludo-Hartmannplatz 1/14, A-1160 Vienna, Austria Tel: +43/2952/30706; +43/664/233 90 75 www.cybertec.at, www.postgresql.at, kernel.cybertec.at ---(end of broadcast)--- TIP 5: Have you checked our extensive FAQ? http://www.postgresql.org/docs/faqs/FAQ.html
Re: [HACKERS] PostgreSQL 7.3.3 and Intel C compiler
On Mon, 21 Jul 2003, [ISO-8859-1] Hans-Jürgen Schönig wrote: Why was SERIALIZABLE faster? I know SERIALIZABLE doesn't have the rollback penalty in read-only queries, but I don't understand why it would be faster. To be honest I don't have the slightest idea. Maybe it has to do with snapshotting but I don't know precisely. In case of SERIALIZABLE all snapshots inside a transaction are the same - maybe this makes the big difference. I have no other explanation for that. There is one nifty detail which seems VERY strange to me: If serializable mode is set in postgresql.conf the system was 3 times faster (~ 7.5 sec. vs. 2.5sec). If serializable mode was set for every transaction (using set at the beginning of the transaction) serializable mode was as fast as read committed. Hmm. We have done 90% cursor work and very simple queries (mostly queries such as DECLARE CURSOR x FOR SELECT * FROM ... WHERE a = b). I have no idea why PostgreSQL behaves like that but it seems to be a really good tweak because in this mode we beat any other database including SQL server on Windows 2003 (2.9sec) and IBM DB2 on Linux (12.6 seconds). Are you testing the same type of cursors? Cursors do not behave according to READ COMMITTED principles under Postgres. What are the results for READ ONLY INSENSITIVE cursors with DB2 and MS SQL? Thanks, Gavin ---(end of broadcast)--- TIP 2: you can get off all lists at once with the unregister command (send unregister YourEmailAddressHere to [EMAIL PROTECTED])
Re: [HACKERS] dblink_ora - a first shot on Oracle ...
Bruce Momjian wrote: Hans, I am a little confused about what you are suggesting. Are you suggesting flag to the make of the contrib module rather than configure tests? I agree this is a killer feature for many people and would like to have it in 7.4. Under these circumstances I was thinking of integrating it into the main configuration mechanism - not just for contrib. We will need it for cross db suff later on anyway. Sorry for the confusion. Regards, Hans -- Cybertec Geschwinde u Schoenig Ludo-Hartmannplatz 1/14, A-1160 Vienna, Austria Tel: +43/2952/30706; +43/664/233 90 75 www.cybertec.at, www.postgresql.at, kernel.cybertec.at ---(end of broadcast)--- TIP 1: subscribe and unsubscribe commands go to [EMAIL PROTECTED]
Re: [HACKERS] php with postgres
Hello Bruce, Monday, July 21, 2003, 9:15:05 PM, you wrote: BM Glad you got in touch with the right guys. Joe and Jan have both talked BM about doing PlPHP for a while. :-) BM Marcus, would you check if PHP is using RESET ALL when passing BM persistent connection to new clients? We added that capability a few BM releases ago, specifically for PHP persistent connections, but I don't BM think that ever got into the PHP code. Unfortunately we don't do so yet. Do i need to check for errors or can i do it unconditionally on conenction start? And i'd need to know how to check if it is available (like starting with which version). Best regards, Marcusmailto:[EMAIL PROTECTED] ---(end of broadcast)--- TIP 9: the planner will ignore your desire to choose an index scan if your joining column's datatypes do not match
Re: [HACKERS] php with postgres
Hello Bruce, Monday, July 21, 2003, 9:37:08 PM, you wrote: BM Marcus B?rger wrote: BM Marcus, would you check if PHP is using RESET ALL when passing BM persistent connection to new clients? We added that capability a few BM releases ago, specifically for PHP persistent connections, but I don't BM think that ever got into the PHP code. Unfortunately we don't do so yet. Do i need to check for errors or can i do it unconditionally on conenction start? And i'd need to know how to check if it is available (like starting with which version). BM It first appeared in PostgreSQL version 7.2. It doesn't generate any BM failures. It just resets all SET settting to their defaults, in case BM the previous client modified them. Committed for PHP 5.0 if there is a version 4.4 or 4.5 i'll commit the patch for that version, too. Now we need someone with a big site to test whether it brings any new problems. If there aren't problems i could probably also commit to PHP 4.3.3. Btw. i don't do RESET ALL when creating the link because that shouldn't be necessary. -- Best regards, Marcusmailto:[EMAIL PROTECTED] ---(end of broadcast)--- TIP 6: Have you searched our list archives? http://archives.postgresql.org
