More Re: 7.0.2 regressions testing on Sparc running 2.5.1 - Mailing list pgsql-novice

From ghaverla@freenet.edmonton.ab.ca
Subject More Re: 7.0.2 regressions testing on Sparc running 2.5.1
Date
Msg-id Pine.A41.3.95.1000831132124.56906B-100000@freenet.edmonton.ab.ca
Whole thread Raw
In response to Re: 7.0.2 regressions testing on Sparc running 2.5.1  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-novice
On Wed, 30 Aug 2000, Tom Lane wrote:
> ghaverla@freenet.edmonton.ab.ca writes:

> > The makefile in regress/sql deletes the create_function_1.sql,
> > create_function_2.sql, copy.sql, contraints.sql and misc.sql
> > files when you run make all.  Then when you run make runtest,
> > those tests fail (cause error messages) are the executable
> > scripts are no longer there.

I still don't know why make deletes those files, maybe I'll
try to trace that later.

The errors I'm seeing upon loading the regress.so file are
a little baffling to me.  The path mentioned in the error
message to the shared object library is correct, and if I
load the file into emacs, I can find the symbol it is complaining
about (or at least a text string with that symbol, in this
case the symbol is CurrentMemoryContext).  If I run
objdump on the file, I get the following for CurrentMemoryContext:
00000000        *UND*  00000000 CurrentMemoryContext

I'm sort of guessing that somehow my mishmash of Solaris
and GNU is building a bad shared object file.  But having
never debugged anything like this before, the going is slow.

Gord

Matter Realisations     http://www.materialisations.com/
Gordon Haverland, B.Sc. M.Eng. President
101  9504 182 St. NW    Edmonton, AB, CA  T5T 3A7
780/481-8019            ghaverla @ freenet.edmonton.ab.ca



pgsql-novice by date:

Previous
From: ghaverla@freenet.edmonton.ab.ca
Date:
Subject: Re: Vacuum DB
Next
From: Tom Lane
Date:
Subject: Re: Failing to boot Postgres on Red Hat Linux