|
Home > Archive > PostgreSQL Hacks > October 2005 > Sequence dependencies
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
| Author |
Sequence dependencies
|
|
| Christopher Kings-Lynne 2005-10-27, 8:14 am |
| Hi,
I notice that in the release notes there is a large query that should be
run if upgrading from prior to 8.1, to ensure that sequence dependencies
are recorded.
Should we not just make this part of contrib/adddepend?
Chris
---------------------------(end of broadcast)---------------------------
TIP 6: explain analyze is your friend
| |
| Bruce Momjian 2005-10-27, 8:14 am |
| Christopher Kings-Lynne wrote:
> Hi,
>
> I notice that in the release notes there is a large query that should be
> run if upgrading from prior to 8.1, to ensure that sequence dependencies
> are recorded.
>
> Should we not just make this part of contrib/adddepend?
Uh, I thought adddepend did more than just sequence dependencies, and I
am worried it might mess up someone's database. Also, by doing it
manually, users will see which sequences are being changed to late
binding, and perhaps catch ones they do not want changed.
--
Bruce Momjian | http://candle.pha.pa.us
pgman@candle.pha.pa.us | (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: explain analyze is your friend
| |
| Tom Lane 2005-10-27, 8:14 am |
| Christopher Kings-Lynne < chriskl@familyhealth
.com.au> writes:
> I notice that in the release notes there is a large query that should be
> run if upgrading from prior to 8.1, to ensure that sequence dependencies
> are recorded.
> Should we not just make this part of contrib/adddepend?
No ... this is not an automatic "you surely wanted this" change. It is
a fundamental change in semantics. Probably 95% of users will find it
an improvement, but that's not enough to justify having tools that
automatically screw the other 5%.
(Also, I thought we'd determined that adddepend is unmaintained and
likely broken...)
regards, tom lane
---------------------------(end of broadcast)---------------------------
TIP 1: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to majordomo@postgresql
.org so that your
message can get through to the mailing list cleanly
| |
| Christopher Kings-Lynne 2005-10-27, 8:14 am |
| >>Should we not just make this part of contrib/adddepend?
>
> Uh, I thought adddepend did more than just sequence dependencies, and I
Yes it does...
> am worried it might mess up someone's database.
Adddepend has been around for a long time - seems to work perfectly.
> Also, by doing it
> manually, users will see which sequences are being changed to late
> binding, and perhaps catch ones they do not want changed.
The addddepend script asks Y/N on each.
Also, consider people upgrading from 7.2 to 8.1. They should only have
to run the adddepend script to get ALL the dependencies, right?
How about it's in the release notes and in adddepend?
Chris
ps. I can't actually _code_ it, as I'm no Perl coder :)
---------------------------(end of broadcast)---------------------------
TIP 6: explain analyze is your friend
| |
| Tom Lane 2005-10-27, 8:14 am |
| Christopher Kings-Lynne < chriskl@familyhealth
.com.au> writes:
> Also, consider people upgrading from 7.2 to 8.1. They should only have
> to run the adddepend script to get ALL the dependencies, right?
This isn't a "dependency" though.
> How about it's in the release notes and in adddepend?
Hmm, it's currently early Wednesday morning my time, and we were
thinking of wrapping RC1 Thursday or Friday. An adddepend extension
is going to get coded and tested when exactly?
Also, I'm dubious about the assumption that adddepend still works at
all, given the author's opinion here:
http://archives.postgresql.org/pgsq...06/msg00311.php
The original plan was to pull it out of this release altogether,
along with some other contrib modules that really belong on pgfoundry.
It's currently looking like that won't get done :-( but hacking
adddepend more in the core CVS is surely pretty far down our priority
list.
regards, tom lane
---------------------------(end of broadcast)---------------------------
TIP 2: Don't 'kill -9' the postmaster
| |
| Christopher Kings-Lynne 2005-10-27, 8:14 am |
| >>How about it's in the release notes and in adddepend?
>
>
> Hmm, it's currently early Wednesday morning my time, and we were
> thinking of wrapping RC1 Thursday or Friday. An adddepend extension
> is going to get coded and tested when exactly?
Fair enough.
> Also, I'm dubious about the assumption that adddepend still works at
> all, given the author's opinion here:
> http://archives.postgresql.org/pgsq...06/msg00311.php
> The original plan was to pull it out of this release altogether,
> along with some other contrib modules that really belong on pgfoundry.
> It's currently looking like that won't get done :-( but hacking
> adddepend more in the core CVS is surely pretty far down our priority
> list.
Ok.
Chris
---------------------------(end of broadcast)---------------------------
TIP 9: In versions below 8.0, the planner will ignore your desire to
choose an index scan if your joining column's datatypes do not
match
|
|
|
|
|