Problem with building PostgreSQL 7.4 from source

From: Mark Iszler <iszler(at)midrivers(dot)com>
To: pgsql-bugs(at)postgresql(dot)org
Subject: Problem with building PostgreSQL 7.4 from source
Date: 2003-11-27 00:11:45
Message-ID: 3FC54141.1080202@midrivers.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-bugs

I am having some difficulty with building PostgreSQL 7.4 from source code.
I am not using any special options -- just running ./configure and gmake
(using GNU Make version 3.79.1). I am using an x86 based system running
Slackware Linux 8.1 (kernel version 2.4.18).

Running the regression tests with 'gmake check' produces the following
messages:
============== creating temporary installation ==============
============== initializing database system ==============
============== starting postmaster ==============

About 45 seconds later the following messages are displayed:
running on port 65432 with pid 3192
============== creating database "regression" ==============

/pgsrc/postgresql-7.4/src/test/regress/./tmp_check/install//usr/local/pgsql\
/bin/createdb: relocation error:
/pgsrc/postgresql-7.4/src/test/regress/./\
tmp_check/install//usr/local/pgsql/bin/createdb: undefined symbol:
get_progname

There are two issues here:
1. The long pause after the "starting postmaster" message.
2. The error messages from createdb.

It appears that createdb was copied from postgresql-7.4/src/bin/scripts
for the
temporary installation. I noticed that this is now a compiled C program
instead of a shell script like in version 7.3.4. I believe that it was not
properly linked during the build. In fact, none of the programs in this
directory appear to be properly linked. They all produce the same kind
of error
message.

To see if I was on the right track, I modified the Makefile in the scripts
directory, explicitly adding "$(libpq_srcdir)/path.o" as a parameter to the
"$(CC)" compilation command. After deleting all of the executable
programs from
the scripts directory and running "gmake" to rebuild them, I tried the
regression
tests (gmake ckeck) again. The "relocation error" and "undefined
symbol" errors
went away. However, now a different error message was displayed and the
regression tests were aborted. (I am away from that computer as I write
this and
I don't remember the error message. I will send more info on that later.)

I hope this all helps. Let me know if you need any more info.

Mark Iszler

Browse pgsql-bugs by date

  From Date Subject
Next Message Ngo Cong Do 2003-11-27 08:40:22 Some troubles from Postgrres_beta4
Previous Message Steve Thames 2003-11-26 23:36:53 Problem using LIMIT 1 when only 1 record exists.