Re: [QUESTIONS] Problems with running v6.3 on DIGITAL UNIX
Hi there,
I've just compiled postgreSQL v6.3 on one of other Alpha boxes running
DIGITAL UNIX 4.0B.Unfortunately, running initdb dumps core with the following error message:
initdb: using /cec/scratch/cecweb/postgresql/lib/local1_template1.bki.source
as
input to create the template database.
initdb: using /cec/scratch/cecweb/postgresql/lib/global1.bki.source as input
to
create the global classes.
initdb: using /cec/scratch/cecweb/postgresql/lib/pg_hba.conf.sample as the
host-
based authentication control file.We are initializing the database system with username altenhof (uid=301).
This user will own all the files and must also own the server process.initdb: creating template database in
/cec/scratch/cecweb/postgresql/data/base/t
emplate1
Running: postgres -boot -C -F -D/cec/scratch/cecweb/postgresql/data -Q
template1
ERROR: BuildFuncTupleDesc: function mkoidname(opaque, opaque) does not
exist
ERROR: BuildFuncTupleDesc: function mkoidname(opaque, opaque) does not
exist
longjmp or siglongjmp function used outside of saved context
/cec/scratch/cecweb/postgresql/bin/initdb: 10890 Abort - core dumped
Known bug. We can't get Alpha working on 6.3.
Let me mention on thing that may help the alpha developers trying to fix
this.
As part of 6.3 changes, I changed some contants in
/src/include/catalog/*.h that used 0L to just plain 0. It seemed to be
done inconsistently, and I could not figure out how the 0L could be
different than 0.
Can someone take a look at the 6.2 source and 6.3 source, and tell me if
the 0L entries in 6.2 change the alpha behavior from a plain 0. Perhaps
a way of testing would be to replace the 0L with 0 in a working 6.2, and
run initdb to see if the system still works.
--
Bruce Momjian | 830 Blythe Avenue
maillist@candle.pha.pa.us | Drexel Hill, Pennsylvania 19026
+ If your life is a hard drive, | (610) 353-9879(w)
+ Christ can be your backup. | (610) 853-3000(h)Import Notes
Reply to msg id not found: 023401bd4793ddd8cee0a170b910@vogon.kar.dec.com
Just tried the same on DEC UNIX 4.0D. I compiled postgres with the standard
C compiler instead of gcc (worked for 6.2.1) and get the same error for
initdb.
Another thing to try would be to enable assert checking:
configure --enable-cassert
--
Bruce Momjian | 830 Blythe Avenue
maillist@candle.pha.pa.us | Drexel Hill, Pennsylvania 19026
+ If your life is a hard drive, | (610) 353-9879(w)
+ Christ can be your backup. | (610) 853-3000(h)Import Notes
Reply to msg id not found: 199803042109.VAA01060@neptune.lpac.ac.uk | Resolved by subject fallback
On Wed, 4 Mar 1998, Bruce Momjian wrote:
ERROR: BuildFuncTupleDesc: function mkoidname(opaque, opaque) does not
exist
ERROR: BuildFuncTupleDesc: function mkoidname(opaque, opaque) does not
exist
longjmp or siglongjmp function used outside of saved context
/cec/scratch/cecweb/postgresql/bin/initdb: 10890 Abort - core dumpedKnown bug. We can't get Alpha working on 6.3.
Let me mention on thing that may help the alpha developers trying to fix
this.As part of 6.3 changes, I changed some contants in
/src/include/catalog/*.h that used 0L to just plain 0. It seemed to be
done inconsistently, and I could not figure out how the 0L could be
different than 0.Can someone take a look at the 6.2 source and 6.3 source, and tell me if
the 0L entries in 6.2 change the alpha behavior from a plain 0. Perhaps
a way of testing would be to replace the 0L with 0 in a working 6.2, and
run initdb to see if the system still works.
:-? I've looked at both 6.2.1 and 6.3 and have been unable to find a
single '0L' in either version. I've looked in all .h files in src/include,
not just src/include/catalog.
Pedro.
-------------------------------------------------------------------
Pedro José Lobo Perea Tel: +34 1 336 78 19
Centro de Cálculo Fax: +34 1 331 92 29
EUIT Telecomunicación - UPM e-mail: pjlobo@euitt.upm.es
:-? I've looked at both 6.2.1 and 6.3 and have been unable to find a
single '0L' in either version. I've looked in all .h files in src/include,
not just src/include/catalog.=09Pedro.
Didn't look very hard :-)
A quick:
find . -name '*.c' -exec grep 0L {} \;
finds 27 occurrences of 0L
Andrew
----------------------------------------------------------------------------
Dr. Andrew C.R. Martin University College London
EMAIL: (Work) martin@biochem.ucl.ac.uk (Home) andrew@stagleys.demon.co.uk
URL: http://www.biochem.ucl.ac.uk/~martin
Tel: (Work) +44(0)171 419 3890 (Home) +44(0)1372 275775
Import Notes
Resolved by subject fallback
On Thu, 5 Mar 1998, Andrew Martin wrote:
:-? I've looked at both 6.2.1 and 6.3 and have been unable to find a
single '0L' in either version. I've looked in all .h files in src/include,
not just src/include/catalog.=09Pedro.
Didn't look very hard :-)
A quick:
find . -name '*.c' -exec grep 0L {} \;
finds 27 occurrences of 0L
Well, not exactly. The thing is '0l', instead of '0L', both in 6.3 and
6.2.1. I'm rebuilding the beast now. I think that the problem *might* be
in some '-1L' which have been converted to '-1'. Perhaps there is a
problem with the sign expansion. I'll let you know what happens.
-------------------------------------------------------------------
Pedro José Lobo Perea Tel: +34 1 336 78 19
Centro de Cálculo Fax: +34 1 331 92 29
EUIT Telecomunicación - UPM e-mail: pjlobo@euitt.upm.es
Let me mention on thing that may help the alpha developers trying to fix
this.As part of 6.3 changes, I changed some contants in
/src/include/catalog/*.h that used 0L to just plain 0. It seemed to be
done inconsistently, and I could not figure out how the 0L could be
different than 0.Can someone take a look at the 6.2 source and 6.3 source, and tell me if
the 0L entries in 6.2 change the alpha behavior from a plain 0. Perhaps
a way of testing would be to replace the 0L with 0 in a working 6.2, and
run initdb to see if the system still works.
I've just finished the test, and it makes no difference :-( The only file
where 0L appeared is in pg_attributes.h. I have tried also to put '-1L' in
the new column, but it makes no difference. Still working...
-------------------------------------------------------------------
Pedro José Lobo Perea Tel: +34 1 336 78 19
Centro de Cálculo Fax: +34 1 331 92 29
EUIT Telecomunicación - UPM e-mail: pjlobo@euitt.upm.es
Import Notes
Resolved by subject fallback
:-? I've looked at both 6.2.1 and 6.3 and have been unable to find a
single '0L' in either version. I've looked in all .h files in src/include,
not just src/include/catalog.
Sorry, they were lowercase, and not usually 0, but -1l, or 323l.
--
Bruce Momjian | 830 Blythe Avenue
maillist@candle.pha.pa.us | Drexel Hill, Pennsylvania 19026
+ If your life is a hard drive, | (610) 353-9879(w)
+ Christ can be your backup. | (610) 853-3000(h)