psql \d does not work

has anyone noticed that the backend still dies when you do a "\d" from
psql ? This is with 6.4.1 on bsdi 3.1

Just checked it on bsdi 4.0, and it works fine. We have also had no
other problem reports like yours.

test=> create table test (x int);
CREATE
test=> \d

	Database    = test
	 +------------------+----------------------------------+----------+
	 |  Owner           |             Relation             |   Type   |
	 +------------------+----------------------------------+----------+
	 | postgres         | test                             | table    |
	 +------------------+----------------------------------+----------+

-- 
  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, Pennsylvania 19026

On Sat, 26 Dec 1998, Angelos Karageorgiou wrote:

has anyone noticed that the backend still dies when you do a "\d" from
psql ? This is with 6.4.1 on bsdi 3.1

Hate to respond to myself, but Bruce noted that other people's \d on the
back end did not crash.

I had compiled postgres with shlicc2 which is the shared lib version of gcc2
on bsdi. That version was crashing the back end. I recompiled with the
default gcc2 and it works magnificently

Any clues ?

--
Incredible Networks LTD Angelos Karageorgiou
20 Karea st, +30.1.92.12.312 (voice)
116 36 Athens, Greece. +30.1.92.12.314 (fax)
http://www.incredible.com angelos@incredible.com (e-mail)

On Sat, 26 Dec 1998, Angelos Karageorgiou wrote:

has anyone noticed that the backend still dies when you do a "\d" from
psql ? This is with 6.4.1 on bsdi 3.1

Hate to respond to myself, but Bruce noted that other people's \d on the
back end did not crash.

I had compiled postgres with shlicc2 which is the shared lib version of gcc2
on bsdi. That version was crashing the back end. I recompiled with the
default gcc2 and it works magnificently

Any clues ?

Strange. I thought I used shlicc on 6.3 and bsdi 3.1, but I am not
sure.

Can you send a backtrace of the crash through gdb? It will be tricky
because you have to get the query from psql.c that is \d, start up gdb
on a backend, and run the query. The backtrace may show that if you
sent it to me.

Do all queries fail, or just \d? Strange. I don't know what it is
about shlicc that would cause a failure.

-- 
  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, Pennsylvania 19026

On Sun, 27 Dec 1998, Bruce Momjian wrote:

Strange. I thought I used shlicc on 6.3 and bsdi 3.1, but I am not
sure.

Well you are right, after a lot of investigating I found out that it was
neither shlicc2's fault nor postgres's what had happened is that I was using
6.4.beta2 and have moved things about a bit. As a result the pg_shadow
table existed but the pg_user did NOT.

So when psql was doing a select from pg_user the back end would crash, not
crash really but disconnect.

I redid the initdb and reloaded all my postgress databases and everything is
fine.

BTW the reason I was using a beta is that 6.3.2 and locales did not work
together all to well. Now I am happy

--
Incredible Networks LTD Angelos Karageorgiou
20 Karea st, +30.1.92.12.312 (voice)
116 36 Athens, Greece. +30.1.92.12.314 (fax)
http://www.incredible.com angelos@incredible.com (e-mail)