Re: [GENERAL] Postgres users/groups

[Richard Huxton]

Adam Lang wrote:
> 
> Postgres can only do authentication with its internal scheme correct?  It is
> not possible to plug in another "module"?

it can use a file similar to /etc/passwd or kerberos detail in the html
docs (administrator section)

> Example:  I have windows desktops and run samba for my sharing and
> authentication.  Is it possible to have postgres validate based on the samba
> PAM module or something so as to avoid multiple authentication schemes?

Don't think it knows anything about PAM, but you should be able to
generate a text file of user:passwd from /etc/passwd or /etc/shadow if
you're using crypt.

- Richard Huxton

[GENERAL] CygIPC 1.08 Source Code : Where ?

[Eric Nicolas]

I don't find that nowhere on the net.
I only have a broken 1.07 binary.
THe official site for it (cygutils.netpedia.org) is password protected !!!

Help.

Eric

Re: [GENERAL] trying to back up a database

[Peter Eisentraut]

Adam Haberlach writes:

>   Sorry--I mean that when I type "pg_dumpall -h, it connects", but I guess
> that is because it is a simple script.

Nobody ever said that '-h' prints a help screen.

> > pg_dumpall must be run by postgres, it may well have problems if you
> > are running it as an unprivileged user.

Nope

> [adam@pinko docs]$ pg_dumpall valhalla -u -s > dbschema.sql
> Password: psql: Password authentication failed for user 'adam'
> Password: psql: Password authentication failed for user 'adam'
> Password: psql: Password authentication failed for user 'adam'
> Password: Password: psql: Password authentication failed for user 'adam'
> Connection to database 'valhalla' failed.
> fe_sendauth: no password supplied

Password authentication doesn't work with pg_dumpall.  Use something else.

-- 
Peter Eisentraut  [EMAIL PROTECTED]   http://yi.org/peter-e/

Re: [GENERAL] trying to back up a database

[Oliver Elphick]

Peter Eisentraut wrote:
  >> > pg_dumpall must be run by postgres, it may well have problems if you
  >> > are running it as an unprivileged user.
  >
  >Nope

Seeing that the unprivileged user may not have SELECT permission on
some of the tables, he ought to have problems!

-- 
Oliver Elphick[EMAIL PROTECTED]
Isle of Wight  http://www.lfix.co.uk/oliver
PGP: 1024R/32B8FAA1: 97 EA 1D 47 72 3F 28 47  6B 7E 39 CC 56 E4 C1 47
GPG: 1024D/3E1D0C1C: CA12 09E0 E8D5 8870 5839  932A 614D 4C34 3E1D 0C1C
 
 "O come, let us worship and bow down; let us kneel 
  before the LORD our maker."Psalms 95:6 

RE: [GENERAL] GIS-type databases using PostgreSQL

[Ken Mort]

This would be a great addition to PG. I queried about this many 
years ago (I think to Tom Lockhart) but I think there were other 
priorities at that time.

I spoke to two groups (PostgreSQL Inc and GreatBridge) at this 
year's Linux Expo in New York about this topic. I will forward 
your project's web site to them.

My interest is to create open source Java classes that could be 
used to create fat or thin mapping applications. In particular, 
the thin apps should be able to connect to geographically 
enabled servers (ideally PostgreSQL). This would enable 
organizations to do mapping on a small budget (the commercial 
vendors are charging way too much).

-- 

Regards,
Ken Mort  <[EMAIL PROTECTED]>
Brooklyn, NY, USA

[EMAIL PROTECTED] (Franck Martin) wrote in
: 

>Hi,
>
>Me and others are planning to move PG to ISO19100 compliance.
>ISO19100 is the future standard that is describing GIS
>systems. 
>
>If you visit FMaps.sourceforge.net and go in the CVS you will
>see in the directory /src/geoobj/ procedures to add
>geographic data types to PG. These procedures need to be
>rewritten as they are not ISO19100 compliant. But all the
>concepts are there and working. 
>
>The work doesn't stop here and include metadata schema,
>feature schema, but a the moment the crunch is creating  a
>geographic object type in PG and rendering it.
>
>Cheers.
>
>Franck Martin
>Network and Database Development Officer
>SOPAC South Pacific Applied Geoscience Commission
>Fiji
>E-mail: [EMAIL PROTECTED]  
>Web site: http://www.sopac.org/ 
>Support FMaps: http://fmaps.sourceforge.net/
> 
>

[GENERAL] Re: CygIPC 1.08 Source Code : Where ?

[Fred Yankowski]

http://www.neuro.gatech.edu/users/cwilson/cygutils/V1.1/cygipc/index.html

-- 
Fred Yankowski   [EMAIL PROTECTED]  tel: +1.630.879.1312
Principal Consultant www.OntoSys.com   fax: +1.630.879.1370
OntoSys, Inc 38W242 Deerpath Rd, Batavia, IL 60510, USA

Re: [GENERAL] PostgreSQL profiling?

[Tom Lane]

Alex Pilosov <[EMAIL PROTECTED]> writes:
> Have someone made effort to do profiling of pgsql during execution of
> certain things (inserts, selects, sorting, indices)? 

Yes ...

> I have a feeling (based on stopping postgres from gdb periodically), that
> a lot of time is used in strcoll() (if table and index has string
> columns).
> Column in question is declared char(3).
> So, why's postgres collating anything at all?

Because textual comparisons are defined in terms of strcoll() if you've
enabled locale support.  There is no way around this; either don't use
locales or write a faster version of strcoll().

regards, tom lane

[GENERAL] php & postgresql data type

[Denni]

hi

I'm newbie here, I've a little knowledge about postgresql.
I'm planning to make a project for my school's assignment
about php using PostgreSQL as the database server.
I'm planning to use arrays data type and inheritance in
PostgreSQL. My question is can php3/php support this feature?

Thanks in advanced.

--
Best Regard,

Denni Pidono
[EMAIL PROTECTED]

Re: [GENERAL] Query plan: varchar vs char indexes

[Tom Lane]

"Richard Huxton" <[EMAIL PROTECTED]> writes:
> With indexed varchar fields the explain changes - performing a seq-scan on
> users rather than using the index.
> Is this because the estimator guesses costs differently for char vs varchar
> or is it because I'm not explicitly casting the id fields?

The estimator has no special ideas about either char or varchar.
However there are some discrepancies in the sets of available functions
for the two datatypes, so what appears to be the same expression may
translate into different function invocations --- especially if you are
doing random combinations of datatypes and expecting the system to pick
an operator for you.  I suspect it is picking a combination that doesn't
work out to be semantically equivalent to the '=' operator in the
index's operator class, so it doesn't think it can use the index.

> And yes - I know I probably shouldn't be joining an int to a varchar.

Not without being pretty darn careful.  You didn't actually say what the
datatype of tag_list.id is, however.

regards, tom lane

Re: [GENERAL] 7.0 configuration

[Tom Lane]

"Belcher, Jim" <[EMAIL PROTECTED]> writes:
> When the system boots, and Postgres is attempting to 
> start, I get the following error: 564, unary operator expected.

I don't believe any part of Postgres itself issues such an error code.
Perhaps the message is coming out of your shell, which would suggest
some portability problem in the pg_ctl shell script.  Try running
pg_ctl by hand and/or inserting debugging echo commands to see if you
can narrow down where the problem is. 

What platform are you on, anyway?

regards, tom lane

Re: [GENERAL] DEBUG strings

[Tom Lane]

Ferruccio Zamuner <[EMAIL PROTECTED]> writes:
> Then I've restored some db dumps from (A) to (B) without any
> problem,

Just exactly how did you do the restore?  It looks to me like your
pg_log (transaction commit info) might be out of sync with 
transaction numbers in the database ...

regards, tom lane

Re: [GENERAL] Postgres SQL 7.1 a thank you and a possible bug

[Tom Lane]

"Nic Ferrier" <[EMAIL PROTECTED]> writes:
> The synopsis of the problem is that date columns don't seem to be
> working as dates when they're selected through a union view.

No, the problem is that restriction clauses applied to views that
contain unions are busted in beta4 :-(.  Thanks for the bug report!
It's fixed in current CVS, or use tomorrow morning's snapshot.

regards, tom lane

[GENERAL] automatic type cast to numeric ?

[Peter Vazsonyi]

Gretings!

I try to change from using of float8 to numeric... For quantitative
calculations and storage the float8 wasn't the best choice;)
For some reasons i need automatic or explicit type cast from text to
numeric and vice versa.
But:
test=# select '123'::text::numeric;
ERROR:  Cannot cast type 'text' to 'numeric'
;(
test=# CREATE FUNCTION numeric(text) as 'SELECT float8($1)::numeric;'
LANGUAGE 'SQL';
ERROR:  parser: parse error at or near "numeric"
;(((
The another side (numeric to text) works...

I had some more clumsy shots with the type '_numeric', but i don't now
enough about the working of parametric types...

So i need Help;)

--
 thanx
 nek.

i use 7.0.2/7.0.3 i386.rpm-s