FAQ
Hi...

Just created one of two new mailing lists, based on some of the
feedback that has been received over the past few days. These are
more directed at the hackers@postgresql.org subscribers, since its meant
to offload some of the email from that list.

This list is pgsql-ports@postgresql.org (also available as an
alias at ports@postgresql.org, so that its nice and short), with the
following 'charter' pertaining to it:

pgsql-ports@postgresql.org
- This list is for porting related and platform specific discussions.
- If PostgreSQL failed to compile on your computer then fill
out the form typically located at /usr/src/pgsql/doc/bug.template
and mail it here.
- If you get PostgreSQL running on an unsupported platform, tell
us how you did it here. And before you attempt the port, check
the archives for this area to see if someone else has already
done it.

*Please* make sure that all port/compile related are posted to here
and not to hackers@postgresql.org...let's try and cleanup the traffic a bit
and leave hackers@ for discussions?

As per usual, all the lists are/will be fully digested, so that you
don't have to see each individual packet going across.

An appropriate procmail filter for this list would be:

:0:
* ^TO_((pgsql-)?)ports@(hub|postgresql).org
$MAILDIR/folders/pg-ports

And, finally, since this is a new list, you have been auto-subscribed
to it *if* you were a member of the hackers@postgresql.org mailing list, so:

===========

How to subscribe/unsubscribe:

If you wish to subscribe or unsubscribe to any of the above
mailing lists, then send a message to pgsql-xxx-request@postgresql.org
where "xxx" is replaced by "hackers", "patches", etc. The body of the
message should contain the single line "subscribe" or "unsubscribe".

If you hate getting many mail messages per day then you should
subscribe or unsubscribe to pgsql-xxx-digest-request instead. Many
messages in pgsql-xxx will be combined into a single message in
pgsql-xxx-digest that is then mailed out. This occurs whenever the
combined message reaches a certain size or at least once per day.
DO NOT post messages to the pgsql-xxx-digest@postgresql.org address.

============

Marc G. Fournier
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org

From scrappy
Date: Thu, 17 Apr 1997 12:34:33 -0700 (PDT)
From: "James A. Hillyerd" <james@blarg.net>
Subject: [PORTS] SPARC linux port (fwd)

This is a message of mine, that was resent by David Friend, and is now
being resent by me. =) A response of a member of the team would be nice.

- -james

[ James A. Hillyerd (JH2162) james@blarg.net - Perl Programmer ]
[ http://www.blarg.net/~james/ http://www.pimedia.com/ ]
[ 1024/B11C3751 CA 1C B3 A9 07 2F 57 C9 91 F4 73 F2 19 A4 C5 88 ]
From scrappy
Date: Sat, 12 Apr 1997 19:36:55 -0400 (EDT)
From: David Friend <dfriend@atlsci.atlsci.com>
To: hackers@postgreSQL.org
Subject: [HACKERS] SPARC linux port

Was the information in the message from James Hillyerd (repeated below) ever
integrated into the tree? The updated INSTALL instructions that were
recently posted to the hackers list didn't mention that the sparc linux
platform was supported. If the changes weren't made, it would be nice if
someone implemented this before 6.1 comes out. Unfortunately, I don't have
the time. Any volunteers? (Don't forget to add sparc linux to the list of
supported compilers.)

I will volunteer to test the changes on a SPARC linux platform if someone
else can change the code.

Incidently, I am running PostgreSQL on a SPARC linux. I got it to work by
modifying the source tree so that the tas() function for the i386 was not
compiled and that src/backend/port/sparc_solaris/tas.s was compiled. I like
the solution below much better though.

Dave

From scrappy
Date: Wed, 12 Feb 1997 09:41:57 -0800 (PST)
From: "James A. Hillyerd" <james@blarg.net>
Subject: [HACKERS] Compile on SPARC Linux

I don't know if anyone else running RedHat SPARC Linux has tried, or has
had trouble compiling, but I thought I would contribute the following,
hopefully someone who Knows What They Are Doing (tm) could look this over
and see if I did it right.

diffs to include/config.h:

156,158c156,162
< # undef HAVE_CBRT
< # define NEED_I386_TAS_ASM
< # define HAS_TEST_AND_SET
- - ---
# if defined(i386)
# define NEED_I386_TAS_ASM
# define HAS_TEST_AND_SET
# endif
# if defined(sparc)
# define NEED_SPARC_TAS_ASM
# endif
It appears that SPARC Linux does have a tas() function, but it is a define
in /usr/include/asm/system.h, and defining HAS_TEST_AND_SET seems to make
it expect a real function with extern and all. =)

(/usr/include/asm)clarke% grep 'define tas' *
system.h:#define tas(ptr) (xchg((ptr),1))

And here is my Makefile.custom, setup for readline/history:

PORTNAME= linux
USE_READLINE= true
READLINE_INC+= -DHAVE_LIBREADLINE
# Remove the following line if you don't have <readline.h>
READLINE_INC+= -DHAVE_READLINE_H
# Remove the following line if you don't have <history.h>
READLINE_INC+= -DHAVE_HISTORY
READLINE_INC+= -I/usr/include
READLINE_LIB+= -L/usr/lib -lreadline
CURSES_LIB= -L/usr/lib -lncurses
READLINE_INC+= -DHAVE_LIBHISTORY
READLINE_INC+= -I/usr/include -I/usr/include/readline
READLINE_LIB+= -L/usr/lib -lhistory
CUSTOM_CC= gcc

- - -james


[ James A. Hillyerd (JH2162) james@blarg.net - Perl Programmer ]
[ http://www.blarg.net/~james/ http://www.pimedia.com/ ]
[ 1024/B11C3751 CA 1C B3 A9 07 2F 57 C9 91 F4 73 F2 19 A4 C5 88 ]

From scrappy
Date: Thu, 17 Apr 1997 16:53:28 -0300 (ADT)
From: The Hermit Hacker <scrappy@hub.org>
Subject: Re: [PORTS] SPARC linux port (fwd)
On Thu, 17 Apr 1997, James A. Hillyerd wrote:

This is a message of mine, that was resent by David Friend, and is now
being resent by me. =) A response of a member of the team would be nice.

James...

You might want to grab the current v6.1 distribution and make
changes/submit patches according to that...none of the changes you made
to config.h are pertinent anymore, as that section of config.h no longer
exists :(

-james

[ James A. Hillyerd (JH2162) james@blarg.net - Perl Programmer ]
[ http://www.blarg.net/~james/ http://www.pimedia.com/ ]
[ 1024/B11C3751 CA 1C B3 A9 07 2F 57 C9 91 F4 73 F2 19 A4 C5 88 ]

---------- Forwarded message ----------
Date: Sat, 12 Apr 1997 19:36:55 -0400 (EDT)
From: David Friend <dfriend@atlsci.atlsci.com>
To: hackers@postgreSQL.org
Subject: [HACKERS] SPARC linux port

Was the information in the message from James Hillyerd (repeated below) ever
integrated into the tree? The updated INSTALL instructions that were
recently posted to the hackers list didn't mention that the sparc linux
platform was supported. If the changes weren't made, it would be nice if
someone implemented this before 6.1 comes out. Unfortunately, I don't have
the time. Any volunteers? (Don't forget to add sparc linux to the list of
supported compilers.)

I will volunteer to test the changes on a SPARC linux platform if someone
else can change the code.

Incidently, I am running PostgreSQL on a SPARC linux. I got it to work by
modifying the source tree so that the tas() function for the i386 was not
compiled and that src/backend/port/sparc_solaris/tas.s was compiled. I like
the solution below much better though.

Dave

----------------------------------------------------------------------------

Date: Wed, 12 Feb 1997 09:41:57 -0800 (PST)
From: "James A. Hillyerd" <james@blarg.net>
Subject: [HACKERS] Compile on SPARC Linux

I don't know if anyone else running RedHat SPARC Linux has tried, or has
had trouble compiling, but I thought I would contribute the following,
hopefully someone who Knows What They Are Doing (tm) could look this over
and see if I did it right.

diffs to include/config.h:

156,158c156,162
< # undef HAVE_CBRT
< # define NEED_I386_TAS_ASM
< # define HAS_TEST_AND_SET
- ---
# if defined(i386)
# define NEED_I386_TAS_ASM
# define HAS_TEST_AND_SET
# endif
# if defined(sparc)
# define NEED_SPARC_TAS_ASM
# endif
It appears that SPARC Linux does have a tas() function, but it is a define
in /usr/include/asm/system.h, and defining HAS_TEST_AND_SET seems to make
it expect a real function with extern and all. =)

(/usr/include/asm)clarke% grep 'define tas' *
system.h:#define tas(ptr) (xchg((ptr),1))

And here is my Makefile.custom, setup for readline/history:

PORTNAME= linux
USE_READLINE= true
READLINE_INC+= -DHAVE_LIBREADLINE
# Remove the following line if you don't have <readline.h>
READLINE_INC+= -DHAVE_READLINE_H
# Remove the following line if you don't have <history.h>
READLINE_INC+= -DHAVE_HISTORY
READLINE_INC+= -I/usr/include
READLINE_LIB+= -L/usr/lib -lreadline
CURSES_LIB= -L/usr/lib -lncurses
READLINE_INC+= -DHAVE_LIBHISTORY
READLINE_INC+= -I/usr/include -I/usr/include/readline
READLINE_LIB+= -L/usr/lib -lhistory
CUSTOM_CC= gcc

- -james


[ James A. Hillyerd (JH2162) james@blarg.net - Perl Programmer ]
[ http://www.blarg.net/~james/ http://www.pimedia.com/ ]
[ 1024/B11C3751 CA 1C B3 A9 07 2F 57 C9 91 F4 73 F2 19 A4 C5 88 ]


Marc G. Fournier
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org

From scrappy
Date: Thu, 17 Apr 1997 18:21:49 -0400 (EDT)
From: System Administrator <igor@sba.miami.edu>
Subject: [PORTS] DEC Alpha compile problem

This is what I get compiling 970417:

cc -DNO_ASSERT_CHECKING -Dalpha -I../.. -I../../../include -c port.c -o
port.o
cc: Error: port.c, line 30: In this statement, "buffer" is not declared.
if (setsysinfo(SSI_NVPAIRS, buffer, 1, (caddr_t) NULL,

The buffer variable is declared as follows:
#ifdef NOFIXADE
int buffer[] = { SSIN_UACPROC, UAC_SIGBUS };
#endif /* NOFIXADE */
#ifdef NOPRINTADE
int buffer[] = { SSIN_UACPROC, UAC_NOPRINT };
#endif /* NOPRINTADE */A

However, if neither NOFIXADE nor NOPRINTADE are defined, buffer will not
be defined either. And of course neither is defined....

I guessed and declared it as the first ifdef:
(int buffer[] = {SSIN_UACPROC, UAC_SIGBUS }

However I am not entirely sure which one is correct..


=+=------------------------/\--------------------------------------=+=
Igor Natanzon |/\| E-mail: igor@homer.bus.miami.edu
System Administrator |/\| igor@cs.cs.miami.edu
--|
University of Miami |\/| Office Phone: (305) 284-1962
School of Business |\/|
=+=------------------------\/--------------------------------------=+=

From scrappy
Date: Fri, 18 Apr 1997 13:27:10 -0400
From: aixssd!darrenk@abs.net (Darren King)
Subject: [PORTS] [HACKERS] Configure and gcc.

First, backend/tcop/Makefile has a "-Wno-error" that needs the
ifeq($CC,gcc) wrapper, but...

Should the configure automatically try to use gcc if it finds it,
even if I've specified another compiler in the template for my port?

I have gcc installed for testing purposes, but prefer to use the
IBM compiler suite. I set CC= xlc in the template/aix, but when
configure finds gcc, it will create Makefile.global with it. Is
there any way to override this besides in Makefile.custom? I'd
like to keep the overrides in there to a minimum if possible.

Thanks...


Darren darrenk@insightdist.com

From scrappy
Date: Sat, 19 Apr 1997 15:35:00 +0900
From: Satoshi Ishikawa <power@sekine.densi.gifu-u.ac.jp>
Subject: [PORTS] SPARC/Linux port

From scrappy
Date: Thu, 17 Apr 1997 12:34:33 -0700 (PDT)
Message-ID: <Pine.LNX.3.95.970417123344.6678A-100000@animal.blarg.net>
Was the information in the message from James Hillyerd (repeated below) ever
integrated into the tree? The updated INSTALL instructions that were
recently posted to the hackers list didn't mention that the sparc linux
platform was supported. If the changes weren't made, it would be nice if
someone implemented this before 6.1 comes out. Unfortunately, I don't have
the time. Any volunteers? (Don't forget to add sparc linux to the list of
supported compilers.)

I will volunteer to test the changes on a SPARC linux platform if someone
else can change the code.
Please test this patch.

I successfully compiled v6.1(970418) on my SPARC/Linux box(RedHat 4.0).
but I have only one SPARC/Linux box and ruuning v6.0 on it,
so I can't do regression test etc...

- ---
diff -Nru pgsql/src/backend/storage/ipc/s_lock.c pgsql.port/src/backend/storage/ipc/s_lock.c
- --- pgsql/src/backend/storage/ipc/s_lock.c Thu Mar 13 08:01:04 1997
+++ pgsql.port/src/backend/storage/ipc/s_lock.c Sat Apr 19 14:49:30 1997
@@ -453,4 +453,37 @@

#endif

+#if defined(linux) && defined(sparc)
+
+int
+tas(slock_t *m)
+{
+ slock_t res;
+ __asm__("ldstub [%1], %0"
+ : "=&r" (res)
+ : "r" (m));
+ return (res != 0);
+}
+
+void
+S_LOCK(slock_t *lock)
+{
+ while (tas(lock))
+ ;
+}
+
+void
+S_UNLOCK(slock_t *lock)
+{
+ *lock = 0;
+}
+
+void
+S_INIT_LOCK(slock_t *lock)
+{
+ S_UNLOCK(lock);
+}
+
+#endif /* defined(linux) && defined(sparc) */
+
#endif /* HAS_TEST_AND_SET */
diff -Nru pgsql/src/include/port/linux.h pgsql.port/src/include/port/linux.h
- --- pgsql/src/include/port/linux.h Sat Apr 12 19:21:48 1997
+++ pgsql.port/src/include/port/linux.h Sat Apr 19 14:48:31 1997
@@ -12,3 +12,6 @@
# define HAS_TEST_AND_SET
typedef unsigned char slock_t;
# endif
+# if defined(sparc)
+# undef NEED_I386_TAS_ASM
+# endif

- ---

Satoshi Ishikawa
power@densi.gifu-u.ac.jp

From scrappy
Date: Sat, 19 Apr 1997 20:40:45 -0400
From: Kurt Seel <kseel@primetime.com>
Subject: [PORTS] [HACKERS] problems compiling 'load'able tcl lib

I am working on bsdi 2.1. I have installed gcc 2.7.2.2.f.2 (as reported
by gcc --version). The bsdi system port define doesn't compile, so I
used bsd44_derived instead.
The .so file needed to 'load' into tcl (7.6) aren't generated by the
gcc command so I linked by hand this way:

For libpq :
ld -r -o libpq.so fe-auth.o fe-connect.o fe-exec.o fe-misc.o fe-lobj.o
../backend/lib/dllist.o pqsignal.o -lc_s.2.1.0
For libtcl:
ld -r pgtcl.o pgtclCmds.o pgtclId.o -L/usr/local/pgsql/lib -lpq
- -L/usr/local/lib/tcl7.6 -ltcl -lm -ldl -lc_s.2.1.0 -o libpgtcl.so

But then this happens:
ls -l */*.so
- -rwxrwx--x 1 root wheel 337263 Apr 19 20:18 libpgtcl/libpgtcl.so
- -rwxrwx--x 1 root wheel 35022 Apr 19 20:19 libpq/libpq.so
tclsh7.6
% load libpq/libpq.so
couldn't find procedure Pq_Init
% load libpgtcl/libpgtcl.so
Segmentation fault (core dumped)

The first I expected, as there isn't supposed to be a 'Pq_Init' (maybe
one should be made).

Does anyone have this 'load'ing on bsdi 2.1? Help, please.

From scrappy
Date: Sun, 20 Apr 1997 18:54:34 -0300 (ADT)
From: The Hermit Hacker <scrappy@hub.org>
Subject: [PORTS] Development Freeze for v6.1

Hi...

As we previously discussed, April 23rd is this coming Wednesday, so
this is a call for any "new" feature patches that need to be applied before
that date.

As of midnight the 24th, only fixes for bugs will be applied so that
we can get v6.1 out the door by the 1st of May. This *should* generally be
restricted to compile fixes (new feature *or* port) and updating of the
regression tests.

So, if someone has something that they want to see in v6.1 and are
holding back, this is the time to send it in...

REMINDER: all patches go to patches@postgresql.org

Marc G. Fournier
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org

From scrappy
Date: Mon, 21 Apr 1997 17:01:51 +0200
From: Frits Dumortier <fdumorti@iris.ufsia.ac.be>
Subject: [PORTS] PostgreSQL fails on SGI/O2

This is a multi-part message in MIME format.

- --------------2781446B794B
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

bug report attached

- --------------2781446B794B
Content-Type: text/plain; charset=us-ascii; name="bug.report"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline; filename="bug.report"

To report a bug, please complete the following form and send it by
email to bugs@postgreSQL.org.

============================================================================
POSTGRES95 BUG REPORT REPORT
============================================================================


Your name : Frits Dumortier
Your email address : Frits.Dumortier@ufsia.ac.be


System Configuration
- ---------------------
Architecture (example: Intel Pentium) : SGI/O2 R10000/150MHz

Operating System (example: Linux 1.3.42 ELF) : Irix 6.3

PostgreSQL version (example: Postgres95-1.01) : PostgreSQL-6.0

Compiler used (example: gcc 2.7.0) : GNU Make 3.75 (gmake)
GNU fileutils 3.12 (ginstall)
MIPSpro C compiler driver release 7.1


Please enter a FULL description of your problem:
- ------------------------------------------------

PostgreSQL-files installed in:
/usr/src/pgsql

Makefile.custom contains:
PORTNAME= irix5
SRCDIR= /usr/src/pgsql
CUSTOM_INSTALL= ginstall

gmake fails with the following error message:

...
make[1]: Entering directory `/usr/src/pgsql/src/libpq'
cc -O -Dirix5 -I../include -I../backend -c fe-auth.c -o fe-auth.o
cc -O -Dirix5 -I../include -I../backend -c fe-connect.c -o fe-connect.o
cc -O -Dirix5 -I../include -I../backend -c fe-exec.c -o fe-exec.o
cc -O -Dirix5 -I../include -I../backend -c fe-misc.c -o fe-misc.o
cc -O -Dirix5 -I../include -I../backend -c fe-lobj.c -o fe-lobj.o
cc -O -Dirix5 -I../include -I../backend -c pqsignal.c -o pqsignal.o
ar crs libpq.a `lorder fe-auth.o fe-connect.o fe-exec.o fe-misc.o fe-lobj.o ../backend/lib/dllist.o pqsignal.o | tsort`
touch libpq.a
rm -f c.h
echo "#undef PORTNAME" > c.h
echo "#define PORTNAME irix5" >> c.h
cat ../include/c.h >> c.h
make[1]: Leaving directory `/usr/src/pgsql/src/libpq'
/usr/local/bin/make -C bin all
make[1]: Entering directory `/usr/src/pgsql/src/bin'
/usr/local/bin/make -C pg_id all
make[2]: Entering directory `/usr/src/pgsql/src/bin/pg_id'
cc -I../../libpq -I../../include -O -Dirix5 -c pg_id.c -o pg_id.o
cc -o pg_id -L../../libpq pg_id.o -lpq -lm -ll
ld: FATAL 11: Object file format error (../../libpq/libpq.a): invalid header
make[2]: *** [pg_id] Error 1
make[2]: Leaving directory `/usr/src/pgsql/src/bin/pg_id'
make[1]: *** [all] Error 2
make[1]: Leaving directory `/usr/src/pgsql/src/bin'
make: *** [all] Error 2



Please describe a way to repeat the problem. Please try to provide a
concise reproducible example, if at all possible:
- ----------------------------------------------------------------------

gmake as used above to build the application.



If you know how this problem might be fixed, list the solution below:
- ---------------------------------------------------------------------


- --------------2781446B794B--

From scrappy
Date: Mon, 21 Apr 1997 13:22:48 -0300 (ADT)
From: The Hermit Hacker <scrappy@hub.org>
Subject: Re: [PORTS] PostgreSQL fails on SGI/O2

v6.1 is planned for a release of May 1st (~1week)...could you grab
the newest beta and take a look at that, to see if it works any better?

On Mon, 21 Apr 1997, Frits Dumortier wrote:

bug report attached
Marc G. Fournier
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org

From scrappy
Date: Mon, 21 Apr 1997 12:57:00 -0400 (EDT)
From: David Friend <dfriend@atlsci.atlsci.com>
Subject: Re: [PORTS] SPARC/Linux port
On Sat, 19 Apr 1997, Satoshi Ishikawa wrote:

I will volunteer to test the changes on a SPARC linux platform if someone
else can change the code.
Please test this patch.

I successfully compiled v6.1(970418) on my SPARC/Linux box(RedHat 4.0).
but I have only one SPARC/Linux box and ruuning v6.0 on it,
so I can't do regression test etc...
I will hopefully find time to do this this weekend. Do I simply grab the
latest version of 6.1 PostgreSQL, or do I grab a specific version and
apply the patch you put into the mail message I am replying to?

If I am applying the patch, could you refresh my memory on what I do
after untarring the source tree but before running make. (It's been a
while since I've run the patch command.)

David Friend ! cq995@freenet.carleton.ca
Atlantis Scientific Systems Group Inc. ! david.friend@atlsci.com
1827 Woodward Drive ! 613-727-1087 (voice)
Ottawa, Ontario, CANADA K2C 0P9 ! 800-265-3894 (voice)
ERGOvista Scientific Image Analysis ! 613-727-5853 (fax)

From scrappy
Date: Mon, 21 Apr 1997 10:57:53 -0700
From: "Padraic I. Hannon" <hannon@rt66.com>
Subject: Re: [PORTS] PostgreSQL fails on SGI/O2

I've had the same problems with both v6 and the older versions of
Postgres.. also tried it with gcc to no avail.. If you get it working
please let me know...

Thanks,
Paddy Hannon

BTW the errors I got are exactly the same (if you look at the .a file it
is 1 line long.)Frits Dumortier wrote:

From scrappy
Date: Tue, 22 Apr 1997 13:13:58 +0900
From: Satoshi Ishikawa <power@sekine.densi.gifu-u.ac.jp>
Subject: Re: [PORTS] SPARC/Linux port

From scrappy
Date: Mon, 21 Apr 1997 12:57:00 -0400 (EDT)
Message-ID: <Pine.LNX.3.95.970421125221.26276C-100000@atlsci.atlsci.com>

dfriend> > > I will volunteer to test the changes on a SPARC linux platform if someone
dfriend> > > else can change the code.
dfriend> >
dfriend> > Please test this patch.
dfriend> >
dfriend> > I successfully compiled v6.1(970418) on my SPARC/Linux box(RedHat 4.0).
dfriend> > but I have only one SPARC/Linux box and ruuning v6.0 on it,
dfriend> > so I can't do regression test etc...
dfriend>
dfriend> I will hopefully find time to do this this weekend. Do I simply grab the
dfriend> latest version of 6.1 PostgreSQL, or do I grab a specific version and
dfriend> apply the patch you put into the mail message I am replying to?
dfriend>
dfriend> If I am applying the patch, could you refresh my memory on what I do
dfriend> after untarring the source tree but before running make. (It's been a
dfriend> while since I've run the patch command.)

I grabbed and tried only 970418.tar.gz.
as all the same x86/Linux without tas(), I think the latest version of
v6.1 is no problem.

I did (or you'll do) :

% cd /usr/local/pgsql
% tar zxvf <<SRCDIR>>/970418.tar.gz
% patch -p1 < patchfile (at the end of this mail)
% cd src
% ./configure --with-template=linux-elf
just type "return" for all questions.

% make all

diff -Nru pgsql/src/backend/storage/ipc/s_lock.c pgsql.port/src/backend/storage/ipc/s_lock.c
- --- pgsql/src/backend/storage/ipc/s_lock.c Thu Mar 13 08:01:04 1997
+++ pgsql.port/src/backend/storage/ipc/s_lock.c Sat Apr 19 14:49:30 1997
@@ -453,4 +453,37 @@

#endif

+#if defined(linux) && defined(sparc)
+
+int
+tas(slock_t *m)
+{
+ slock_t res;
+ __asm__("ldstub [%1], %0"
+ : "=&r" (res)
+ : "r" (m));
+ return (res != 0);
+}
+
+void
+S_LOCK(slock_t *lock)
+{
+ while (tas(lock))
+ ;
+}
+
+void
+S_UNLOCK(slock_t *lock)
+{
+ *lock = 0;
+}
+
+void
+S_INIT_LOCK(slock_t *lock)
+{
+ S_UNLOCK(lock);
+}
+
+#endif /* defined(linux) && defined(sparc) */
+
#endif /* HAS_TEST_AND_SET */
diff -Nru pgsql/src/include/port/linux.h pgsql.port/src/include/port/linux.h
- --- pgsql/src/include/port/linux.h Sat Apr 12 19:21:48 1997
+++ pgsql.port/src/include/port/linux.h Sat Apr 19 14:48:31 1997
@@ -12,3 +12,6 @@
# define HAS_TEST_AND_SET
typedef unsigned char slock_t;
# endif
+# if defined(sparc)
+# undef NEED_I386_TAS_ASM
+# endif





From scrappy
Date: Wed, 23 Apr 1997 10:00:06 +0200
From: Andreas Jung <ajung@sz-sb.de>
Subject: [PORTS] Compiling problem PG6.1 under Solaris 2.4

Version: PG 6.1 from 20. April 97
OS: Solaris 2.4, Sun Sparcstation 20
Compiler: gcc2.7.2

I tried "configure" than "make" - compilation failes with:

make[3]: Entering directory
`/home1/postgres/pg6.1/src/backend/port/sparc_solari
s'
gcc -I../../../backend/port/sparc_solaris -Wall -Wmissing-prototypes
- -Dsparc_sol
aris -I.. -I../.. -I../../../include -c port.c -o port.o
In file included from port.c:20:
port-protos.h:50: warning: `struct rusage' declared inside parameter
list
port-protos.h:50: warning: its scope is only this definition or
declaration,
port-protos.h:50: warning: which is probably not what you want.
In file included from port.c:22:
rusagestub.h:28: conflicting types for `getrusage'
port-protos.h:50: previous declaration of `getrusage'
make[3]: *** [port.o] Error 1

- --
_\\|//_
(' O-O ')
- ------------------------------ooO-(_)-Ooo--------------------------------------

Andreas Jung, Saarbrücker Zeitung, Gutenbergstr. 11-23, D-66117
Saarbrücken
Phone: +49-(0)681-502-1528 E-Mail: ajung@sz-sb.de (PGP key
available)
SZ-Newsline: http://www.sz-sb.de/, Saarland-Online:
http://www.saar-ol.de/
- -------------------------------------------------------------------------------

From scrappy
Date: Tue, 22 Apr 1997 12:52:20 -0300 (EST)
From: "Ricardo Ferraro G. da Silva" <rferraro@ci.rnp.br>
Subject: Re: [PORTS] Compiling problem PG6.1 under Solaris 2.4

Hello Andreas,

Try to add -DNEED_RUSAGE to your compiler options. When doing this here I
could compile. I am also using Sparc_solaris (station5, version 2.4) here.

Marc told me that RUSAGE is something that is not more used in general by
all systems and will be excluded in next versions... Hope we still able to
use pg without Rusage, whatever it be... ;-)

regards, Ricardo.
On Wed, 23 Apr 1997, Andreas Jung wrote:

Version: PG 6.1 from 20. April 97
OS: Solaris 2.4, Sun Sparcstation 20
Compiler: gcc2.7.2
=20
I tried "configure" than "make" - compilation failes with:
=20
make[3]: Entering directory
`/home1/postgres/pg6.1/src/backend/port/sparc_solari
s'
gcc -I../../../backend/port/sparc_solaris -Wall -Wmissing-prototypes
-Dsparc_sol
aris -I.. -I../.. -I../../../include -c port.c -o port.o
In file included from port.c:20:
port-protos.h:50: warning: `struct rusage' declared inside parameter
list
port-protos.h:50: warning: its scope is only this definition or
declaration,
port-protos.h:50: warning: which is probably not what you want.
In file included from port.c:22:
rusagestub.h:28: conflicting types for `getrusage'
port-protos.h:50: previous declaration of `getrusage'
make[3]: *** [port.o] Error 1
=20
--
_\\|//_
(' O-O ')
------------------------------ooO-(_)-Ooo--------------------------------= - ------
=20
Andreas Jung, Saarbr=FCcker Zeitung, Gutenbergstr. 11-23, D-66117
Saarbr=FCcken
Phone: +49-(0)681-502-1528 E-Mail: ajung@sz-sb.de (PGP key
available)
SZ-Newsline: http://www.sz-sb.de/, Saarland-Online:
http://www.saar-ol.de/
-------------------------------------------------------------------------= - ------
=20
=20
=20
From scrappy
Date: Tue, 22 Apr 1997 14:02:01 -0300 (ADT)
From: The Hermit Hacker <scrappy@hub.org>
Subject: Re: [PORTS] Compiling problem PG6.1 under Solaris 2.4
On Tue, 22 Apr 1997, Ricardo Ferraro G. da Silva wrote:

Hello Andreas,

Try to add -DNEED_RUSAGE to your compiler options. When doing this here I
could compile. I am also using Sparc_solaris (station5, version 2.4) here.

Marc told me that RUSAGE is something that is not more used in general by
all systems and will be excluded in next versions... Hope we still able to
use pg without Rusage, whatever it be... ;-)
Huh?

The problem is something else, actually...someone with a better
understanding of the sparc_solaris port should send me a patch for this,
but it looks like there should be some sort of #ifdef around either or
both of the prototypes for getrusage that are found in rusagestub.h and
port-protos.h...

Someone want to check this out?
regards, Ricardo.
On Wed, 23 Apr 1997, Andreas Jung wrote:

Version: PG 6.1 from 20. April 97
OS: Solaris 2.4, Sun Sparcstation 20
Compiler: gcc2.7.2

I tried "configure" than "make" - compilation failes with:

make[3]: Entering directory
`/home1/postgres/pg6.1/src/backend/port/sparc_solari
s'
gcc -I../../../backend/port/sparc_solaris -Wall -Wmissing-prototypes
-Dsparc_sol
aris -I.. -I../.. -I../../../include -c port.c -o port.o
In file included from port.c:20:
port-protos.h:50: warning: `struct rusage' declared inside parameter
list
port-protos.h:50: warning: its scope is only this definition or
declaration,
port-protos.h:50: warning: which is probably not what you want.
In file included from port.c:22:
rusagestub.h:28: conflicting types for `getrusage'
port-protos.h:50: previous declaration of `getrusage'
make[3]: *** [port.o] Error 1

--
_\\|//_
(' O-O ')
------------------------------ooO-(_)-Ooo--------------------------------------

Andreas Jung, Saarbrücker Zeitung, Gutenbergstr. 11-23, D-66117
Saarbrücken
Phone: +49-(0)681-502-1528 E-Mail: ajung@sz-sb.de (PGP key
available)
SZ-Newsline: http://www.sz-sb.de/, Saarland-Online:
http://www.saar-ol.de/
-------------------------------------------------------------------------------

Marc G. Fournier
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org

From scrappy
Date: Tue, 22 Apr 1997 13:36:42 -0400 (EDT)
From: System Administrator <igor@sba.miami.edu>
Subject: [PORTS] Configure for DEC-Alpha

Configure script properly detects alpha-dec-osf4.0 machine, but
sets a default GENERIC template for it. I modified tempplate/.similar to
add alpha-dec-osf4.0=alpha. Then configure properly set the template to
alpha.


=+=------------------------/\--------------------------------------=+=
Igor Natanzon |/\| E-mail: igor@homer.bus.miami.edu
System Administrator |/\| igor@cs.cs.miami.edu
--|
University of Miami |\/| Office Phone: (305) 284-1962
School of Business |\/|
=+=------------------------\/--------------------------------------=+=

From scrappy
Date: Tue, 22 Apr 1997 14:49:54 -0300 (ADT)
From: The Hermit Hacker <scrappy@hub.org>
Subject: [PORTS] Binary Distributions

Hi...

Do we want to start doing binary distributions with v6.1? We'll
need to have volunteer maintainers for each on who will be responsible
for creating the appropriate distribution for each platform...

Marc G. Fournier
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org

From scrappy
Date: Tue, 22 Apr 1997 14:55:22 -0300 (ADT)
From: The Hermit Hacker <scrappy@hub.org>
Subject: Re: [PORTS] Configure for DEC-Alpha
On Tue, 22 Apr 1997, System Administrator wrote:

Configure script properly detects alpha-dec-osf4.0 machine, but
sets a default GENERIC template for it. I modified tempplate/.similar to
add alpha-dec-osf4.0=alpha. Then configure properly set the template to
alpha.
Change is applied

Marc G. Fournier
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org

From scrappy
Date: Tue, 22 Apr 1997 14:39:05 -0400 (EDT)
From: adrian@waltham.harvard.net
Subject: Re: [PORTS] Binary Distributions

Volunteer for the Sparc Solaris version.

Hi...

Do we want to start doing binary distributions with v6.1? We'll
need to have volunteer maintainers for each on who will be responsible
for creating the appropriate distribution for each platform...

Marc G. Fournier
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org

From scrappy
Date: Tue, 22 Apr 1997 15:47:25 -0300 (ADT)
From: The Hermit Hacker <scrappy@hub.org>
Subject: Re: [PORTS] Compiling problem PG6.1 under Solaris 2.4
On Tue, 22 Apr 1997 adrian@waltham.harvard.net wrote:

The -DHAVE_GETRUSAGE patch has been in configure since 970414? It was
the major butt of the patch.
What -DHAVE_GETRUSAGE patch? This shouldn't be passed as a -D
flag, it should be defined in include/config.h.in...

...and this is a ports related question, leave it on the ports list :)

Adrian
On Tue, 22 Apr 1997, Ricardo Ferraro G. da Silva wrote:

Hello Andreas,

Try to add -DNEED_RUSAGE to your compiler options. When doing this here I
could compile. I am also using Sparc_solaris (station5, version 2.4) here.

Marc told me that RUSAGE is something that is not more used in general by
all systems and will be excluded in next versions... Hope we still able to
use pg without Rusage, whatever it be... ;-)
Huh?

The problem is something else, actually...someone with a better
understanding of the sparc_solaris port should send me a patch for this,
but it looks like there should be some sort of #ifdef around either or
both of the prototypes for getrusage that are found in rusagestub.h and
port-protos.h...

Someone want to check this out?
regards, Ricardo.
On Wed, 23 Apr 1997, Andreas Jung wrote:

Version: PG 6.1 from 20. April 97
OS: Solaris 2.4, Sun Sparcstation 20
Compiler: gcc2.7.2

I tried "configure" than "make" - compilation failes with:

make[3]: Entering directory
`/home1/postgres/pg6.1/src/backend/port/sparc_solari
s'
gcc -I../../../backend/port/sparc_solaris -Wall -Wmissing-prototypes
-Dsparc_sol
aris -I.. -I../.. -I../../../include -c port.c -o port.o
In file included from port.c:20:
port-protos.h:50: warning: `struct rusage' declared inside parameter
list
port-protos.h:50: warning: its scope is only this definition or
declaration,
port-protos.h:50: warning: which is probably not what you want.
In file included from port.c:22:
rusagestub.h:28: conflicting types for `getrusage'
port-protos.h:50: previous declaration of `getrusage'
make[3]: *** [port.o] Error 1

--
_\\|//_
(' O-O ')
------------------------------ooO-(_)-Ooo--------------------------------------

Andreas Jung, Saarbr|cker Zeitung, Gutenbergstr. 11-23, D-66117
Saarbr|cken
Phone: +49-(0)681-502-1528 E-Mail: ajung@sz-sb.de (PGP key
available)
SZ-Newsline: http://www.sz-sb.de/, Saarland-Online:
http://www.saar-ol.de/
-------------------------------------------------------------------------------

Marc G. Fournier
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org

Marc G. Fournier
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org

From scrappy
Date: Tue, 22 Apr 1997 16:58:59 -0500
From: lawadmit@beta.loyno.edu
Subject: [PORTS] Compile failed

I tried to generate the system on a DEC Alpha. I got numerous compile
errors:

# gmake
gmake -C utils all
gmake[1]: Entering directory `/usr/src/pgsql/src/utils'
gmake[1]: Nothing to be done for `all'.
gmake[1]: Leaving directory `/usr/src/pgsql/src/utils'
gmake -C backend all
gmake[1]: Entering directory `/usr/src/pgsql/src/backend'
gmake -C access all
gmake[2]: Entering directory `/usr/src/pgsql/src/backend/access'
gmake -C common SUBSYS.o
gmake[3]: Entering directory `/usr/src/pgsql/src/backend/access/common'
cc -O -Dalpha -DNOFIXADE -I../.. -I../../port/alpha
- -I../../../include -c hea
ptuple.c -o heaptuple.o
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 122: Syntax
Error
extern int madvise(caddr_t addr, size_t len, int behav);
---------- ----------------^
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 122: Syntax
Error
extern int madvise(caddr_t addr, size_t len, int behav);
---------- -----------------------------^
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 123: Syntax
Error
extern caddr_t mmap(caddr_t addr, size_t len, int prot, int flags, int
filedes,
off_t off);
-------------- ^
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 123: Syntax
Error
extern caddr_t mmap(caddr_t addr, size_t len, int prot, int flags, int
filedes,
off_t off);
-------------- -------------^
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 123: Syntax
Error
extern caddr_t mmap(caddr_t addr, size_t len, int prot, int flags, int
filedes,
off_t off);
-------------- --------------------------^
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 123: Syntax
Error
extern caddr_t mmap(caddr_t addr, size_t len, int prot, int flags, int
filedes,
off_t off);
-------------- -----------------------------------------^
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 123: Syntax
Error
extern caddr_t mmap(caddr_t addr, size_t len, int prot, int flags, int
filedes,
off_t off);
-------------- ----------------------------------------------------^
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 123: Syntax
Error
extern caddr_t mmap(caddr_t addr, size_t len, int prot, int flags, int
filedes,
off_t off);
--------------
- ----------------------------------------------------------------
- -------^
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 124: Syntax
Error
extern int mprotect(caddr_t addr, size_t len, int prot);
---------- -----------------^
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 124: Syntax
Error
extern int mprotect(caddr_t addr, size_t len, int prot);
---------- ------------------------------^
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 129: Syntax
Error
extern int msync(caddr_t addr, size_t len, int flags);
---------- --------------^
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 129: Syntax
Error
extern int msync(caddr_t addr, size_t len, int flags);
---------- ---------------------------^
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 130: Syntax
Error
extern int munmap(caddr_t addr, size_t len);
---------- ---------------^
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 130: Syntax
Error
extern int munmap(caddr_t addr, size_t len);
---------- ----------------------------^
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 131: Syntax
Error
extern int mvalid(caddr_t addr, size_t len, int prot);
---------- ---------------^
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 131: Syntax
Error
extern int mvalid(caddr_t addr, size_t len, int prot);
---------- ----------------------------^
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 132: Syntax
Error
extern int shm_open(const char *path, int oflag, mode_t mode);
------------------------------------------------------^
gmake[3]: *** [heaptuple.o] Error 1
gmake[3]: Leaving directory `/usr/src/pgsql/src/backend/access/common'
gmake[2]: *** [submake] Error 2
gmake[2]: Leaving directory `/usr/src/pgsql/src/backend/access'
gmake[1]: *** [postgres_group1] Error 2
gmake[1]: Leaving directory `/usr/src/pgsql/src/backend'
gmake: *** [all] Error 2

From scrappy
Date: Tue, 22 Apr 1997 21:07:26 -0300 (ADT)
From: The Hermit Hacker <scrappy@hub.org>
Subject: Re: [PORTS] Compile failed
On Tue, 22 Apr 1997 lawadmit@beta.loyno.edu wrote:

I tried to generate the system on a DEC Alpha. I got numerous compile
errors:
Any ideas on how we fix them?
# gmake
gmake -C utils all
gmake[1]: Entering directory `/usr/src/pgsql/src/utils'
gmake[1]: Nothing to be done for `all'.
gmake[1]: Leaving directory `/usr/src/pgsql/src/utils'
gmake -C backend all
gmake[1]: Entering directory `/usr/src/pgsql/src/backend'
gmake -C access all
gmake[2]: Entering directory `/usr/src/pgsql/src/backend/access'
gmake -C common SUBSYS.o
gmake[3]: Entering directory `/usr/src/pgsql/src/backend/access/common'
cc -O -Dalpha -DNOFIXADE -I../.. -I../../port/alpha
-I../../../include -c hea
ptuple.c -o heaptuple.o
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 122: Syntax
Error
extern int madvise(caddr_t addr, size_t len, int behav);
---------- ----------------^
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 122: Syntax
Error
extern int madvise(caddr_t addr, size_t len, int behav);
---------- -----------------------------^
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 123: Syntax
Error
extern caddr_t mmap(caddr_t addr, size_t len, int prot, int flags, int
filedes,
off_t off);
-------------- ^
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 123: Syntax
Error
extern caddr_t mmap(caddr_t addr, size_t len, int prot, int flags, int
filedes,
off_t off);
-------------- -------------^
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 123: Syntax
Error
extern caddr_t mmap(caddr_t addr, size_t len, int prot, int flags, int
filedes,
off_t off);
-------------- --------------------------^
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 123: Syntax
Error
extern caddr_t mmap(caddr_t addr, size_t len, int prot, int flags, int
filedes,
off_t off);
-------------- -----------------------------------------^
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 123: Syntax
Error
extern caddr_t mmap(caddr_t addr, size_t len, int prot, int flags, int
filedes,
off_t off);
-------------- ----------------------------------------------------^
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 123: Syntax
Error
extern caddr_t mmap(caddr_t addr, size_t len, int prot, int flags, int
filedes,
off_t off);
--------------
----------------------------------------------------------------
-------^
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 124: Syntax
Error
extern int mprotect(caddr_t addr, size_t len, int prot);
---------- -----------------^
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 124: Syntax
Error
extern int mprotect(caddr_t addr, size_t len, int prot);
---------- ------------------------------^
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 129: Syntax
Error
extern int msync(caddr_t addr, size_t len, int flags);
---------- --------------^
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 129: Syntax
Error
extern int msync(caddr_t addr, size_t len, int flags);
---------- ---------------------------^
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 130: Syntax
Error
extern int munmap(caddr_t addr, size_t len);
---------- ---------------^
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 130: Syntax
Error
extern int munmap(caddr_t addr, size_t len);
---------- ----------------------------^
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 131: Syntax
Error
extern int mvalid(caddr_t addr, size_t len, int prot);
---------- ---------------^
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 131: Syntax
Error
extern int mvalid(caddr_t addr, size_t len, int prot);
---------- ----------------------------^
/usr/lib/cmplrs/cc/cfe: Error: /usr/include/sys/mman.h, line 132: Syntax
Error
extern int shm_open(const char *path, int oflag, mode_t mode);
------------------------------------------------------^
gmake[3]: *** [heaptuple.o] Error 1
gmake[3]: Leaving directory `/usr/src/pgsql/src/backend/access/common'
gmake[2]: *** [submake] Error 2
gmake[2]: Leaving directory `/usr/src/pgsql/src/backend/access'
gmake[1]: *** [postgres_group1] Error 2
gmake[1]: Leaving directory `/usr/src/pgsql/src/backend'
gmake: *** [all] Error 2
Marc G. Fournier
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org

From scrappy
Date: Wed, 23 Apr 1997 14:11:25 +1000 (EST)
From: Ross Johnson <rpj@ise.canberra.edu.au>
Subject: Re: [PORTS] Compiling problem PG6.1 under Solaris 2.4
On Tue, 22 Apr 1997, The Hermit Hacker wrote:
On Tue, 22 Apr 1997, Ricardo Ferraro G. da Silva wrote:

Hello Andreas,

Try to add -DNEED_RUSAGE to your compiler options. When doing this here I
could compile. I am also using Sparc_solaris (station5, version 2.4) here.

Marc told me that RUSAGE is something that is not more used in general by
all systems and will be excluded in next versions... Hope we still able to
use pg without Rusage, whatever it be... ;-)
Huh?

The problem is something else, actually...someone with a better
understanding of the sparc_solaris port should send me a patch for this,
but it looks like there should be some sort of #ifdef around either or
both of the prototypes for getrusage that are found in rusagestub.h and
port-protos.h...

Someone want to check this out?
The problem seems to be that at some point NEED_RUSAGE was changed (in
include/config.h) to HAVE_RUSAGE. The file backend/tcop/postgres.c still
uses NEED_RUSAGE.

I fixed this by changing all instances of

#ifdef NEED_RUSAGE

in backend/tcop/postgres.c to

#ifndef HAVE_RUSAGE


HAVE_RUSAGE is already defined for solaris in include/config.h

+----------------------+---+
Ross Johnson | | E-Mail: rpj@ise.canberra.edu.au
Info Sciences and Eng|___|
University of Canberra | FAX: +61 6 2015227
PO Box 1 |
Belconnen ACT 2616 | WWW: http://willow.canberra.edu.au/~rpj/
AUSTRALIA |
+--------------------------+

From scrappy
Date: Wed, 23 Apr 1997 01:25:59 -0300 (ADT)
From: The Hermit Hacker <scrappy@hub.org>
Subject: Re: [PORTS] Compiling problem PG6.1 under Solaris 2.4
On Wed, 23 Apr 1997, Ross Johnson wrote:
On Tue, 22 Apr 1997, The Hermit Hacker wrote:
On Tue, 22 Apr 1997, Ricardo Ferraro G. da Silva wrote:

Hello Andreas,

Try to add -DNEED_RUSAGE to your compiler options. When doing this here I
could compile. I am also using Sparc_solaris (station5, version 2.4) here.

Marc told me that RUSAGE is something that is not more used in general by
all systems and will be excluded in next versions... Hope we still able to
use pg without Rusage, whatever it be... ;-)
Huh?

The problem is something else, actually...someone with a better
understanding of the sparc_solaris port should send me a patch for this,
but it looks like there should be some sort of #ifdef around either or
both of the prototypes for getrusage that are found in rusagestub.h and
port-protos.h...

Someone want to check this out?
The problem seems to be that at some point NEED_RUSAGE was changed (in
include/config.h) to HAVE_RUSAGE. The file backend/tcop/postgres.c still
uses NEED_RUSAGE.

I fixed this by changing all instances of

#ifdef NEED_RUSAGE

in backend/tcop/postgres.c to

#ifndef HAVE_RUSAGE


HAVE_RUSAGE is already defined for solaris in include/config.h
this must have been from a very old version of v6.1 then, since
I fixed that particular problem ages ago:

- ----------------------------
revision 1.29
date: 1997/02/03 04:43:31; author: scrappy; state: Exp; lines: +7 -7
Change references to NEED_RUSAGE to HAVE_RUSAGE

And I just checked that file for any occurance of RUSAGE, and there
is not that has NEED_RUSAGE...


Marc G. Fournier
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org

From scrappy
Date: Wed, 23 Apr 1997 04:58:26 GMT
From: Thomas Lockhart <Thomas.G.Lockhart@jpl.nasa.gov>
Subject: [PORTS] Re: Bug with new date patches: wont compile

Ah, this is a porting problem; I need to bracket those references
with #ifdef HAVE_TZSET etc to get non-posix machines to work.
Will do so asap. btw scrappy, I have been submitting patches
and then waiting to do a clean install of the next distribution
before developing more code because I don't want to get out
of sync with your distributions. I assume it would not be convenient
if I went ahead and kept generating patches on the same old
distribution, since you would then have to patch -N to add
only the new incremental patch sections. Is that the case, or
do you do patch -N anyway??

- Tom

From scrappy
Date: Wed, 23 Apr 1997 07:38:34 +0200 (MET DST)
From: Andreas Jung <ajung@sz-sb.de>
Subject: Re: [PORTS] Compiling problem PG6.1 under Solaris 2.4
On Wed, 23 Apr 1997, The Hermit Hacker wrote:

On Wed, 23 Apr 1997, Ross Johnson wrote:
=20
On Tue, 22 Apr 1997, The Hermit Hacker wrote:
=20
On Tue, 22 Apr 1997, Ricardo Ferraro G. da Silva wrote:
=20
Hello Andreas,
=20
Try to add -DNEED_RUSAGE to your compiler options. When doing this=
here I
could compile. I am also using Sparc_solaris (station5, version 2.4=
) here.
=20
Marc told me that RUSAGE is something that is not more used in gen=
eral by
all systems and will be excluded in next versions... Hope we still =
able to
use pg without Rusage, whatever it be... ;-)
=20
=09Huh?
=20
=09The problem is something else, actually...someone with a better
understanding of the sparc_solaris port should send me a patch for th=
is,
but it looks like there should be some sort of #ifdef around either o=
r
both of the prototypes for getrusage that are found in rusagestub.h a=
nd
port-protos.h...
=20
=09Someone want to check this out?
=20
The problem seems to be that at some point NEED_RUSAGE was changed (in
include/config.h) to HAVE_RUSAGE. The file backend/tcop/postgres.c stil=
l
uses NEED_RUSAGE.
=20
I fixed this by changing all instances of
=20
#ifdef NEED_RUSAGE
=20
in backend/tcop/postgres.c to
=20
#ifndef HAVE_RUSAGE
=20
=20
HAVE_RUSAGE is already defined for solaris in include/config.h
=20
=09this must have been from a very old version of v6.1 then, since
I fixed that particular problem ages ago:
=20
----------------------------
revision 1.29
date: 1997/02/03 04:43:31; author: scrappy; state: Exp; lines: +7 -7
Change references to NEED_RUSAGE to HAVE_RUSAGE
=20
Nope - I toke the archiv 970420.tar.gz from the 6.1 directory :-)

Andreas

_\\|//_
(' O-O ')
- ------------------------------ooO-(_)-Ooo----------------------------------=
- ----
Andreas Jung, Saarbr=FCcker Zeitung, Gutenbergstr. 11-23, D-66117 Saarbr=
=FCcken
Phone: +49-(0)681-502-1528 E-Mail: ajung@sz-sb.de (PGP key available)
SZ-Newsline: http://www.sz-sb.de/, Saarland-Online: http://www.saar-ol.de=
/=20
- ---------------------------------------------------------------------------=
- ----

From scrappy
Date: Wed, 23 Apr 97 8:28:41 METDST"
From: "Martin S. Utesch" <utesch@aut.tu-freiberg.de>
Subject: Re: [PORTS] Binary Distributions

Hi...

Do we want to start doing binary distributions with v6.1? We'll
need to have volunteer maintainers for each on who will be responsible
for creating the appropriate distribution for each platform...

Marc G. Fournier
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org

Think I will share the binary distribution on HP-UX with the rest of the world...

I can hand it for: HP-UX A.09.01 A 9000/730

Sorry, it's a little ancient ;-)

Please tell, what binaries, documents, sources, etc., should be included in
the package or does scrappy put the rest of the available stuff to the
binaries?

What compile options shall we take?

Servus, MAU

=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
* Martin Utesch * Institute of Automatic Control *
= = University of Mining and Technology =
* utesch@aut.tu-freiberg.de * Freiberg, Germany *
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=

From scrappy
Date: Wed, 23 Apr 1997 18:20:25 +1000
From: Philip Rhoades <philr@mail.austasia.net>
Subject: [PORTS] Anyone got a OpenLink Win95 client talking to LINUX PG?

I have the Linux (RH4.1) ver of PG (6.0?) and the OpenLink Win95 PG Agent -
but can't get the two to talk. I posted the following to the OpenLink
people with no response - nayone got any ideas?

Thanks, Phil.
I have a working database/table (phils/weather) that the postgreSQL
interactive utility works quite happily with.

I am trying to get Excell to import some external data from this database - -
I have used the ODBC32 manager to set up the connection for 'weather' but
get an error dialog saying:

"Exceeded maximum number of allowed connections"

On the Linux side I get:

[root@chutest bin]# oplrqb -d
OpenLink Request Broker
Version 2.6B (Release 1.5) as of Tue Feb 18 1997.
Compiled for linux 2.0.25 (i486-caldera-linux)
Copyright (C) OpenLink Software.

Unregistered demonstration version
This is a 5 concurrent users license
restricted to 10 concurrent connections

This is demonstration copy - Not for resale.

oplrqb: using rulebook /home/pgsql/bin/oplrqb.ini
oplrqb: bound UDP protocol to 127.0.0.1.60001
oplrqb: bound TCP protocol to 127.0.0.1.5000
oplrqb: using UDP address 192.168.0.101.0 for IPC
oplrqb: created icmp socket for connection monitoring
oplrqb: request broker started by root
oplrqb: crpc.chu.com.au called (192.168.0.110.1071)
oplrqb: request: domain=Postgres 95 /home/pgsql/data/base/phils
serveropts=
oplrqb: connectopts= user=phil opsys=win32 readonly=0
oplrqb: application=CPPDMO32 processid=4294621469
oplrqb: rejected phil@crpc.chu.com.au.CPPDMO32 (exceeded maximum number
of allowed connections)

and:

[root@chutest bin]# ./pgr95_sv -d
pgr95_sv: server started
svctcp_.c - cannot getsockname or listen: Bad file number
pgr95_sv: unable to install RPC service (Bad file number)
pgr95_sv: server shutting down

The kernel has been compiled with SYSVIPC turned on.

Philip Rhoades
Pricom Pty Ltd
http://www.pricom.com.au = http:/203.12.131.20
GPO Box 3411 Sydney NSW 2001 Australia
Ph: +61:0411:185652
Fax: +61:2:9959-3481
E-mail: philr@mail.austasia.net

From scrappy
Date: Wed, 23 Apr 1997 10:24:28 -0300 (EST)
From: "Ricardo Ferraro G. da Silva" <rferraro@ci.rnp.br>
Subject: Re: [PORTS] Compiling problem PG6.1 under Solaris 2.4

Hi Andreas,

I agree with you, the last one I toke was from the 6.1 directory
also (ftp.postgresql.org), but 970416.tar.gz.

greetings, Ricardo.


On Wed, 23 Apr 1997, Andreas Jung wrote:
[...]
=20
Nope - I toke the archiv 970420.tar.gz from the 6.1 directory :-)
=20
Andreas
=20
_\\|//_
(' O-O ')
------------------------------ooO-(_)-Ooo--------------------------------= - ------
Andreas Jung, Saarbr=FCcker Zeitung, Gutenbergstr. 11-23, D-66117 Saarb= r=FCcken
Phone: +49-(0)681-502-1528 E-Mail: ajung@sz-sb.de (PGP key available)
SZ-Newsline: http://www.sz-sb.de/, Saarland-Online: http://www.saar-ol.= de/=20
-------------------------------------------------------------------------= - ------
=20
=20
=20
From scrappy
Date: Wed, 23 Apr 1997 10:12:48 -0400 (EDT)
From: David Friend <dfriend@atlsci.atlsci.com>
Subject: Re: [PORTS] Binary Distributions
On Tue, 22 Apr 1997, The Hermit Hacker wrote:

Do we want to start doing binary distributions with v6.1? We'll
need to have volunteer maintainers for each on who will be responsible
for creating the appropriate distribution for each platform...
I will be loosing access to my postgresql server shortly so I can't
maintain a binary. However, I can probably install version 6.1 before I
leave, test it, then create a .tgz file of /usr/local/pgsql. This would
be for the sparc linux elf platform.

If I'm not too tired, I will be testing the patch Satoshi re-generated
Saturday for supporting sparc_linux. With luck, this platform will be
supported in the source tree for v6.1.

You will need some standards on how these distributions are made
available. Might I suggest the following files be made available for ftp
in ftp://ftp.postgreSQL.org/pub/bin for each binary supported platform:

1) The maintainer downloads the source, compiles it and tests it to
his satisfaction. S/he then creates a .tgz file of /usr/local/pgsql
called something like postgresql-v6_1_bin_sparc_linux_elf_tar.gz.
(I'm assuming this is all you would need, except for possibly
some separate documentation files.)

2) The maintainer creates a text file called something like
postgresql_sparc_linux_elf.README. Suggested contents of this
file include
- Notes on installing the binary.
- Notes on how the binary was generated, i.e. notes on how
the source was compiled and tested that are specific to
this platform. It will probably refer to the generic
install instructions.
- A request that people installing this binary send email
to me letting me know that it worked or didn't work,
what problems they had, etc.

The maintainer can choose to add as much or as little as they
want and feel up to doing, and can freely grab info from the
other README files.

3) A link called something like postgresql_bin_sparc_linux_elf_tar.gz
be created that points to the recommended binary for downloading.

4) If an important bug shows up (like the one for version 6.0 that
prevents dates from working on sparc systems if you don't edit
the make file to prevent an optimized compile with gcc) then
a new binary can be generated that replaces "6_1" with "6_1a",
etc. in the name. The README file would describe the differences
between each posted binary distribution. The old download
would be removed (or moved to an archive directory) in this
case.

5) If a maintainer finds that an incremental realease is stable on
their platform then they might choose to upload it, replacing
the "6_1" with something like "6_2_970525". The link will
not be modified to point to this file. The README file should
mention why you might want to use this binary in place of the
presumably more stable recommended binary.

This directory should also contain a README.txt file explaining what the
other files in this directory are, and any additional files (such as
documentation) that should be downloaded along with the files for the
desired platform.

Comments?

David Friend ! cq995@freenet.carleton.ca
Atlantis Scientific Systems Group Inc. ! david.friend@atlsci.com
1827 Woodward Drive ! 613-727-1087 (voice)
Ottawa, Ontario, CANADA K2C 0P9 ! 800-265-3894 (voice)
ERGOvista Scientific Image Analysis ! 613-727-5853 (fax)

From scrappy
Date: Wed, 23 Apr 1997 10:43:26 -0400
From: aixssd!darrenk@abs.net (Darren King)
Subject: Re: [PORTS] Binary Distributions
On Tue, 22 Apr 1997, The Hermit Hacker wrote:

Do we want to start doing binary distributions with v6.1? We'll
need to have volunteer maintainers for each on who will be responsible
for creating the appropriate distribution for each platform...
...and on Wed, 23 Apr 1997, David Friend responded with:


You will need some standards on how these distributions are made
available. Might I suggest the following files be made available for ftp
in ftp://ftp.postgreSQL.org/pub/bin for each binary supported platform:
...some very good rules to do binary distributions by deleted...
This directory should also contain a README.txt file explaining what the
other files in this directory are, and any additional files (such as
documentation) that should be downloaded along with the files for the
desired platform.

Comments?
Sound good enough to me.

I'll volunteer for the AIX platform (325 [&|] 41) if there is a need for it.


Darren darrenk@insightdist.com

From scrappy
Date: Wed, 23 Apr 1997 11:47:10 -0400 (EDT)
From: adrian@waltham.harvard.net
Subject: Re: [PORTS] Compiling problem PG6.1 under Solaris 2.4
On Tue, 22 Apr 1997, The Hermit Hacker wrote:
On Tue, 22 Apr 1997, Ricardo Ferraro G. da Silva wrote:

Hello Andreas,

Try to add -DNEED_RUSAGE to your compiler options. When doing this here I
could compile. I am also using Sparc_solaris (station5, version 2.4) here.

Marc told me that RUSAGE is something that is not more used in general by
all systems and will be excluded in next versions... Hope we still able to
use pg without Rusage, whatever it be... ;-)
Huh?

The problem is something else, actually...someone with a better
understanding of the sparc_solaris port should send me a patch for this,
but it looks like there should be some sort of #ifdef around either or
both of the prototypes for getrusage that are found in rusagestub.h and
port-protos.h...

Someone want to check this out?
The problem seems to be that at some point NEED_RUSAGE was changed (in
include/config.h) to HAVE_RUSAGE. The file backend/tcop/postgres.c still
uses NEED_RUSAGE.

I fixed this by changing all instances of

#ifdef NEED_RUSAGE

in backend/tcop/postgres.c to

#ifndef HAVE_RUSAGE


HAVE_RUSAGE is already defined for solaris in include/config.h
It was my understanding that configure looked for the existance of
getrusage() in the system libraries - its there for Solaris 2.5 and
not (apparently) for Solaris 2.4 - I dont have access to a Solaris-2.4
distribution to verify this. If configure finds it, it defines the
HAVE_GETRUSAGE to 1 in the config.h (which is generated from the
config.h.in file) - look for the lines:

/* Set to 1 if you have getrusage() */
#define HAVE_GETRUSAGE 1

Note that config.h is a GENERATED file - you shouldn't be hand editting
this file. The version of src/backend/port/sparc_solaris/port.c and
rusagestub.h that I have (which is from 970421) uses HAVE_GETRUSAGE
to define whether to compile a getrusage() stub.

Adrian

From scrappy
Date: Wed, 23 Apr 1997 14:54:59 -0300 (ADT)
From: The Hermit Hacker <scrappy@hub.org>
Subject: Re: [PORTS] Compiling problem PG6.1 under Solaris 2.4
On Wed, 23 Apr 1997 adrian@waltham.harvard.net wrote:

It was my understanding that configure looked for the existance of
getrusage() in the system libraries - its there for Solaris 2.5 and
not (apparently) for Solaris 2.4 - I dont have access to a Solaris-2.4
distribution to verify this. If configure finds it, it defines the
HAVE_GETRUSAGE to 1 in the config.h (which is generated from the
config.h.in file) - look for the lines:

/* Set to 1 if you have getrusage() */
#define HAVE_GETRUSAGE 1
You hit it right on...postgres.c was using HAVE_RUSAGE instead of
HAVE_GETRUSAGE :( Fixed

Marc G. Fournier
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org

From scrappy
Date: Wed, 23 Apr 1997 17:34:43 +0000
From: "Ronald E. Fortin" <ref@mesasys.com>
Subject: [PORTS] PostgresSQL install problem

PostgreSQL version (example: Postgres95-1.01) : PostgreSQL-6.0

Compiler used (example: gcc 2.7.0) : standard HP-UX cc


Please enter a FULL description of your problem:
- ------------------------------------------------
running regression test fails. The server log has the following
error message:

NOTICE:CommitTransaction and not in in-progress state
Bad State for Function Call("!((((Node*)CurrentMemoryContext)->type ==
T_PortalH eapMemory)):", File: "portalmem.c", Line: 882)
!((((Node*)CurrentMemoryContext)->type == T_PortalHeapMemory)) (0) [No
such file
or directory]


Please describe a way to repeat the problem. Please try to provide a
concise reproducible example, if at all possible:
- ----------------------------------------------------------------------
Well, when I built on the HP I was getting some errors about not
finding port/hpux/fixade.h when including include/c.h - so changed the
include directives to include ../backend.

Then there was the problem where srandom was defined as void
srandom(int seed) when it is really void srandom(unsigned seed).

So once, these were fixed, I just built the product and installed it.
Started the postmaster and ran make runtest.

The above error then occured.


If you know how this problem might be fixed, list the solution below:
- ---------------------------------------------------------------------

Ronald E. Fortin (ref@mesasys.com)
Mesa Systems Guild, Inc.
60 Quaker Lane
Warwick RI, 02886
401-828-8500
fax: 401-828-9550
email: info@mesasys.com
web: www.mesasys.com

From scrappy
Date: Wed, 23 Apr 1997 20:38:45 -0400 (EDT)
From: David Friend <dfriend@atlsci.atlsci.com>
Subject: Re: [PORTS] SPARC/Linux port (fwd)

I tried applying the patch repeated below to PostgreSQL v6.1 970421. I
set CASSERT to yes, otherwise I followed the instructions below. The
compile and install were successful. However, when I went to create a
database, library libpq.so failed to load.

Does anyone have any ideas on how to cure this? I have included the
output below that was generated when I ran initdb.

[postgres@appl-srv pgsql]$ initdb --pgdata=/usr/local/pgsql/data
- --pglib=/usr/local/pgsql/lib
initdb: using /usr/local/pgsql/lib/local1_template1.bki.source as input to
create the template database.
initdb: using /usr/local/pgsql/lib/global1.bki.source as input to create
the global classes.
initdb: using /usr/local/pgsql/lib/pg_hba.conf.sample as the host-based
authentication control file.

pg_id: can't load library 'libpq.so'
/usr/local/pgsql/bin/initdb: [: =: unary operator expected
pg_id: can't load library 'libpq.so'
pg_id: can't load library 'libpq.so'
/usr/local/pgsql/bin/initdb: [: too many arguments
We are initializing the database system with username postgres (uid=).
This user will own all the files and must also own the server process.

Creating Postgres database system directory /usr/local/pgsql/data/base

initdb: creating template database in /usr/local/pgsql/data/base/template1
Running: postgres -boot -C -F -D/usr/local/pgsql/data -Q template1
WARN:pg_atoi: error in "f": can't parse "f"
initdb: could not create template database
initdb: cleaning up by wiping out /usr/local/pgsql/data/base/template1
[postgres@appl-srv pgsql]$

David Friend ! cq995@freenet.carleton.ca
Atlantis Scientific Systems Group Inc. ! david.friend@atlsci.com
1827 Woodward Drive ! 613-727-1087 (voice)
Ottawa, Ontario, CANADA K2C 0P9 ! 800-265-3894 (voice)
ERGOvista Scientific Image Analysis ! 613-727-5853 (fax)
From scrappy
Date: Tue, 22 Apr 1997 13:13:58 +0900
From: Satoshi Ishikawa <power@sekine.densi.gifu-u.ac.jp>
To: dfriend@atlsci.atlsci.com
Cc: pgsql-ports@postgreSQL.org, power@sekine.densi.gifu-u.ac.jp
Subject: Re: [PORTS] SPARC/Linux port

From scrappy
Date: Mon, 21 Apr 1997 12:57:00 -0400 (EDT)
Message-ID: <Pine.LNX.3.95.970421125221.26276C-100000@atlsci.atlsci.com>

dfriend> > > I will volunteer to test the changes on a SPARC linux platform if someone
dfriend> > > else can change the code.
dfriend> >
dfriend> > Please test this patch.
dfriend> >
dfriend> > I successfully compiled v6.1(970418) on my SPARC/Linux box(RedHat 4.0).
dfriend> > but I have only one SPARC/Linux box and ruuning v6.0 on it,
dfriend> > so I can't do regression test etc...
dfriend>
dfriend> I will hopefully find time to do this this weekend. Do I simply grab the
dfriend> latest version of 6.1 PostgreSQL, or do I grab a specific version and
dfriend> apply the patch you put into the mail message I am replying to?
dfriend>
dfriend> If I am applying the patch, could you refresh my memory on what I do
dfriend> after untarring the source tree but before running make. (It's been a
dfriend> while since I've run the patch command.)

I grabbed and tried only 970418.tar.gz.
as all the same x86/Linux without tas(), I think the latest version of
v6.1 is no problem.

I did (or you'll do) :

% cd /usr/local/pgsql
% tar zxvf <<SRCDIR>>/970418.tar.gz
% patch -p1 < patchfile (at the end of this mail)
% cd src
% ./configure --with-template=linux-elf
just type "return" for all questions.

% make all

diff -Nru pgsql/src/backend/storage/ipc/s_lock.c pgsql.port/src/backend/storage/ipc/s_lock.c
- --- pgsql/src/backend/storage/ipc/s_lock.c Thu Mar 13 08:01:04 1997
+++ pgsql.port/src/backend/storage/ipc/s_lock.c Sat Apr 19 14:49:30 1997
@@ -453,4 +453,37 @@

#endif

+#if defined(linux) && defined(sparc)
+
+int
+tas(slock_t *m)
+{
+ slock_t res;
+ __asm__("ldstub [%1], %0"
+ : "=&r" (res)
+ : "r" (m));
+ return (res != 0);
+}
+
+void
+S_LOCK(slock_t *lock)
+{
+ while (tas(lock))
+ ;
+}
+
+void
+S_UNLOCK(slock_t *lock)
+{
+ *lock = 0;
+}
+
+void
+S_INIT_LOCK(slock_t *lock)
+{
+ S_UNLOCK(lock);
+}
+
+#endif /* defined(linux) && defined(sparc) */
+
#endif /* HAS_TEST_AND_SET */
diff -Nru pgsql/src/include/port/linux.h pgsql.port/src/include/port/linux.h
- --- pgsql/src/include/port/linux.h Sat Apr 12 19:21:48 1997
+++ pgsql.port/src/include/port/linux.h Sat Apr 19 14:48:31 1997
@@ -12,3 +12,6 @@
# define HAS_TEST_AND_SET
typedef unsigned char slock_t;
# endif
+# if defined(sparc)
+# undef NEED_I386_TAS_ASM
+# endif





From scrappy
Date: Wed, 23 Apr 1997 20:06:05 -0500
From: Brian Weber <brian@cbiowa.com>
Subject: Re: [PORTS] SPARC/Linux port (fwd)

David Friend wrote:
I tried applying the patch repeated below to PostgreSQL v6.1 970421. I
set CASSERT to yes, otherwise I followed the instructions below. The
compile and install were successful. However, when I went to create a
database, library libpq.so failed to load.

Does anyone have any ideas on how to cure this? I have included the
output below that was generated when I ran initdb.
You have to go to /etc/ld.so.conf and add the directory that the
library is installed to. Then you run ldconfig. That will configure
the loadable library into the system and it will work like a charm after
that
- --
__________________
Brian Weber
brian@cbiowa.com
Partners In Computing
http://www.cbiowa.com/

From scrappy
Date: Wed, 23 Apr 1997 22:56:15 -0300 (ADT)
From: The Hermit Hacker <scrappy@hub.org>
Subject: Re: [PORTS] PostgresSQL install problem
On Wed, 23 Apr 1997, Ronald E. Fortin wrote:

PostgreSQL version (example: Postgres95-1.01) : PostgreSQL-6.0
As v6.1 will be released in less then a week, could you please
grab the newest beta and confirm whether this problem still exists? There
won't be anything 'new' put into the current source tree, only bug/port
related fixes, so its "stable" from that standpoint...

Compiler used (example: gcc 2.7.0) : standard HP-UX cc


Please enter a FULL description of your problem:
------------------------------------------------
running regression test fails. The server log has the following
error message:

NOTICE:CommitTransaction and not in in-progress state
Bad State for Function Call("!((((Node*)CurrentMemoryContext)->type ==
T_PortalH eapMemory)):", File: "portalmem.c", Line: 882)
!((((Node*)CurrentMemoryContext)->type == T_PortalHeapMemory)) (0) [No
such file
or directory]


Please describe a way to repeat the problem. Please try to provide a
concise reproducible example, if at all possible:
----------------------------------------------------------------------
Well, when I built on the HP I was getting some errors about not
finding port/hpux/fixade.h when including include/c.h - so changed the
include directives to include ../backend.

Then there was the problem where srandom was defined as void
srandom(int seed) when it is really void srandom(unsigned seed).

So once, these were fixed, I just built the product and installed it.
Started the postmaster and ran make runtest.

The above error then occured.


If you know how this problem might be fixed, list the solution below:
---------------------------------------------------------------------

Ronald E. Fortin (ref@mesasys.com)
Mesa Systems Guild, Inc.
60 Quaker Lane
Warwick RI, 02886
401-828-8500
fax: 401-828-9550
email: info@mesasys.com
web: www.mesasys.com
Marc G. Fournier
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org

From scrappy
Date: Wed, 23 Apr 1997 23:10:16 -0300 (ADT)
From: The Hermit Hacker <scrappy@hub.org>
Subject: Re: [PORTS] SPARC/Linux port
On Sat, 19 Apr 1997, Satoshi Ishikawa wrote:

From: "James A. Hillyerd" <james@blarg.net>
Subject: [PORTS] SPARC linux port (fwd)
Date: Thu, 17 Apr 1997 12:34:33 -0700 (PDT)
Message-ID: <Pine.LNX.3.95.970417123344.6678A-100000@animal.blarg.net>
Was the information in the message from James Hillyerd (repeated below) ever
integrated into the tree? The updated INSTALL instructions that were
recently posted to the hackers list didn't mention that the sparc linux
platform was supported. If the changes weren't made, it would be nice if
someone implemented this before 6.1 comes out. Unfortunately, I don't have
the time. Any volunteers? (Don't forget to add sparc linux to the list of
supported compilers.)

I will volunteer to test the changes on a SPARC linux platform if someone
else can change the code.
Please test this patch.

I successfully compiled v6.1(970418) on my SPARC/Linux box(RedHat 4.0).
but I have only one SPARC/Linux box and ruuning v6.0 on it,
so I can't do regression test etc...
Applied, will be in tonights run...

From scrappy
Date: Thu, 24 Apr 1997 15:47:45 +0900
From: Satoshi Ishikawa <power@sekine.densi.gifu-u.ac.jp>
Subject: Re: [PORTS] SPARC/Linux port (fwd)

From scrappy
Date: Wed, 23 Apr 1997 20:06:05 -0500
Message-ID: <335EB1FD.3A77022B@cbiowa.com>

brian> David Friend wrote:
brian> >
brian> > I tried applying the patch repeated below to PostgreSQL v6.1 970421. I
brian> > set CASSERT to yes, otherwise I followed the instructions below. The
brian> > compile and install were successful. However, when I went to create a
brian> > database, library libpq.so failed to load.
brian> >
brian> > Does anyone have any ideas on how to cure this? I have included the
brian> > output below that was generated when I ran initdb.
brian> You have to go to /etc/ld.so.conf and add the directory that the
brian> library is installed to. Then you run ldconfig. That will configure
brian> the loadable library into the system and it will work like a charm after
brian> that

or please try :

export LD_LIBRARY_PATH=/usr/local/pgsql/lib (bash)
or
setenv LD_LIBRARY_PATH /usr/local/pgsql/lib (csh)

- ---
Satoshi Ishikawa
power@densi.gifu-u.ac.jp





From scrappy
Date: Thu, 24 Apr 1997 14:15:46 +0200
From: Frits Dumortier <fdumorti@iris.ufsia.ac.be>
Subject: [PORTS] /usr/local/bin/make -C page SUBSYS.o make[3]: Entering directory `/usr/src/pgsql/src/backend/storage/page' make[3]: `SUBSYS.o' is up to date. make[3]: Leaving directory `/usr/src/pgsql/src/backend/storage/page' /usr/local/bin/make -C smgr SUBSYS.o make[3]: Entering directory `/usr/src/pgsql/src/backend/storage/smgr' make[3]: `SUBSYS.o' is up to date. make[3]: Leaving directory `/usr/src/pgsql/src/backend/storage/smgr' make[2]: Leaving directory `/usr/src/pgsql/src/backend/storage' /usr/local/bin/make -C tcop all PORTNAME=irix5 make[2]: Entering directory `/usr/src/pgsql/src/backend/tcop' cc -O -Dirix5 -I.. -I../port/irix5 -I../../include -Wno-error -c variable.c -o variable.o cc: Error: malformed or unknown option: -Wno-error make[2]: *** [variable.o] Error 2 make[2]: Leaving directory `/usr/src/pgsql/src/bPostgresSQL 6.1beta fails on SGI/O2

============================================================================
POSTGRES95 BUG REPORT REPORT
============================================================================


Your name : Frits Dumortier
Your email address : Frits.Dumortier@ufsia.ac.be
fdumorti@Iris.ufsia.ac.be


System Configuration
- ---------------------
Architecture (example: Intel Pentium) : SGI/O2
R10000/150MHz

Operating System (example: Linux 1.3.42 ELF) : Irix 6.3

PostgreSQL version (example: Postgres95-1.01) :
PostgreSQL-6.1beta (970421)

Compiler used (example: gcc 2.7.0) : GNU Make 3.75
(gmake)
GNU fileutils
3.12 (ginstall)
MIPSpro C
compiler driver release 7.1 (cc)


Please enter a FULL description of your problem:
- ------------------------------------------------

PostgreSQL-files installed in:
/usr/src/pgsql

Makefile.custom contains:
PORTNAME= irix5
SRCDIR= /usr/src/pgsql
CUSTOM_INSTALL= ginstall

gmake fails with the following error message:

...
/usr/local/bin/make -C page SUBSYS.o
make[3]: Entering directory `/usr/src/pgsql/src/backend/storage/page'
make[3]: `SUBSYS.o' is up to date.
make[3]: Leaving directory `/usr/src/pgsql/src/backend/storage/page'
/usr/local/bin/make -C smgr SUBSYS.o
make[3]: Entering directory `/usr/src/pgsql/src/backend/storage/smgr'
make[3]: `SUBSYS.o' is up to date.
make[3]: Leaving directory `/usr/src/pgsql/src/backend/storage/smgr'
make[2]: Leaving directory `/usr/src/pgsql/src/backend/storage'
/usr/local/bin/make -C tcop all PORTNAME=irix5
make[2]: Entering directory `/usr/src/pgsql/src/backend/tcop'
cc -O -Dirix5 -I.. -I../port/irix5 -I../../include -Wno-error -c
variable.c -o variable.o
cc: Error: malformed or unknown option: -Wno-error
make[2]: *** [variable.o] Error 2
make[2]: Leaving directory `/usr/src/pgsql/src/backend/tcop'
make[1]: *** [tcop.dir] Error 2
make[1]: Leaving directory `/usr/src/pgsql/src/backend'
make: *** [all] Error 2


Please describe a way to repeat the problem. Please try to provide a
concise reproducible example, if at all possible:
- ----------------------------------------------------------------------

gmake as used above to build the application.



If you know how this problem might be fixed, list the solution below:
- ---------------------------------------------------------------------

From scrappy
Date: Thu, 24 Apr 1997 14:19:23 +0200
From: Frits Dumortier <fdumorti@iris.ufsia.ac.be>
Subject: [PORTS] PostgresSQL 6.1beta fails on SGI/O2

============================================================================
POSTGRES95 BUG REPORT REPORT
============================================================================


Your name : Frits Dumortier
Your email address : Frits.Dumortier@ufsia.ac.be
fdumorti@Iris.ufsia.ac.be


System Configuration
- ---------------------
Architecture (example: Intel Pentium) : SGI/O2
R10000/150MHz

Operating System (example: Linux 1.3.42 ELF) : Irix 6.3

PostgreSQL version (example: Postgres95-1.01) :
PostgreSQL-6.1beta (970421)

Compiler used (example: gcc 2.7.0) : GNU Make 3.75
(gmake)
GNU fileutils
3.12 (ginstall)
MIPSpro C
compiler driver release 7.1 (cc)


Please enter a FULL description of your problem:
- ------------------------------------------------

PostgreSQL-files installed in:
/usr/src/pgsql

Makefile.custom contains:
PORTNAME= irix5
SRCDIR= /usr/src/pgsql
CUSTOM_INSTALL= ginstall

gmake fails with the following error message:

...
/usr/local/bin/make -C page SUBSYS.o
make[3]: Entering directory `/usr/src/pgsql/src/backend/storage/page'
make[3]: `SUBSYS.o' is up to date.
make[3]: Leaving directory `/usr/src/pgsql/src/backend/storage/page'
/usr/local/bin/make -C smgr SUBSYS.o
make[3]: Entering directory `/usr/src/pgsql/src/backend/storage/smgr'
make[3]: `SUBSYS.o' is up to date.
make[3]: Leaving directory `/usr/src/pgsql/src/backend/storage/smgr'
make[2]: Leaving directory `/usr/src/pgsql/src/backend/storage'
/usr/local/bin/make -C tcop all PORTNAME=irix5
make[2]: Entering directory `/usr/src/pgsql/src/backend/tcop'
cc -O -Dirix5 -I.. -I../port/irix5 -I../../include -Wno-error -c
variable.c -o variable.o
cc: Error: malformed or unknown option: -Wno-error
make[2]: *** [variable.o] Error 2
make[2]: Leaving directory `/usr/src/pgsql/src/backend/tcop'
make[1]: *** [tcop.dir] Error 2
make[1]: Leaving directory `/usr/src/pgsql/src/backend'
make: *** [all] Error 2


Please describe a way to repeat the problem. Please try to provide a
concise reproducible example, if at all possible:
- ----------------------------------------------------------------------

gmake as used above to build the application.



If you know how this problem might be fixed, list the solution below:
- ---------------------------------------------------------------------

From scrappy
Date: Thu, 24 Apr 1997 08:38:13 -0400 (EDT)
From: David Friend <dfriend@atlsci.atlsci.com>
Subject: [PORTS] SPARC/Linux Port

I am busy until Friday night. At that time I will download the latest
code again, recompile on my SPARC Linux box, and run the regression tests.
Thanks for the info on loading the .so files. I've heard this information
before but it's not in any of my Linux books (the books are too old).

David Friend ! cq995@freenet.carleton.ca
Atlantis Scientific Systems Group Inc. ! david.friend@atlsci.com
1827 Woodward Drive ! 613-727-1087 (voice)
Ottawa, Ontario, CANADA K2C 0P9 ! 800-265-3894 (voice)
ERGOvista Scientific Image Analysis ! 613-727-5853 (fax)

From scrappy
Date: Thu, 24 Apr 1997 08:19:17 +0000
From: "Ronald E. Fortin" <ref@mesasys.com>
Subject: Re: [PORTS] PostgresSQL install problem

Thank you for the quick response.

However, I downloaded the 970421 release and tried to install it and
am now
getting a problem with BYTE_ORDER. HPUX-10 does not have an endian.h

any ideas how to get around this?

cc -W l,-E -Ae -DNOFIXADE -Dhpux -I.. -I../port/hpux -I../../include
- -c pqcomp rim.c -o pqcomprim.o cpp: "pqcomprim.c", line 17: error
4062: BYTE_ORDER must be defined as LITTLE_EN DIAN, BIG_ENDIAN or
PDP_ENDIAN
cpp: "pqcomprim.c", line 35: warning 2013: Unknown preprocessing
directive.
cpp: "pqcomprim.c", line 37: warning 2013: Unknown
preprocessing directive.
make[2]: *** [pqcomprim.o] Error 1
make[2]: Leaving directory
`/sos/web/tools/database/postgres-v6.0/src/backend/li bpq' make[1]:
*** [libpq.dir] Error 2 make[1]: Leaving directory
`/sos/web/tools/database/postgres-v6.0/src/backend' make: *** [all]
Error 2
Ronald E. Fortin (ref@mesasys.com)
Mesa Systems Guild, Inc.
60 Quaker Lane
Warwick RI, 02886
401-828-8500
fax: 401-828-9550
email: info@mesasys.com
web: www.mesasys.com

From scrappy
Date: Thu, 24 Apr 1997 10:26:21 -0300 (ADT)
From: The Hermit Hacker <scrappy@hub.org>
Subject: Re: [PORTS] /usr/local/bin/make -C page SUBSYS.o make[3]:
On Thu, 24 Apr 1997, Frits Dumortier wrote:

/usr/local/bin/make -C page SUBSYS.o
make[3]: Entering directory `/usr/src/pgsql/src/backend/storage/page'
make[3]: `SUBSYS.o' is up to date.
make[3]: Leaving directory `/usr/src/pgsql/src/backend/storage/page'
/usr/local/bin/make -C smgr SUBSYS.o
make[3]: Entering directory `/usr/src/pgsql/src/backend/storage/smgr'
make[3]: `SUBSYS.o' is up to date.
make[3]: Leaving directory `/usr/src/pgsql/src/backend/storage/smgr'
make[2]: Leaving directory `/usr/src/pgsql/src/backend/storage'
/usr/local/bin/make -C tcop all PORTNAME=irix5
make[2]: Entering directory `/usr/src/pgsql/src/backend/tcop'
cc -O -Dirix5 -I.. -I../port/irix5 -I../../include -Wno-error -c
variable.c -o variable.o
cc: Error: malformed or unknown option: -Wno-error
make[2]: *** [variable.o] Error 2
make[2]: Leaving directory `/usr/src/pgsql/src/backend/tcop'
make[1]: *** [tcop.dir] Error 2
make[1]: Leaving directory `/usr/src/pgsql/src/backend'
make: *** [all] Error 2
Fixed in the source tree, apply:

Index: backend/tcop/Makefile
===================================================================
RCS file: /usr/local/cvsroot/postgres95/src/backend/tcop/Makefile,v
retrieving revision 1.11
retrieving revision 1.12
diff -r1.11 -r1.12
7c7
< # $Header: /usr/local/cvsroot/postgres95/src/backend/tcop/Makefile,v 1.11 1997/04/04 10:40:34 scrappy Exp $
- ---
# $Header: /usr/local/cvsroot/postgres95/src/backend/tcop/Makefile,v 1.12 1997/04/24 13:25:07 scrappy Exp $
18c18,22
< CFLAGS+= $(INCLUDE_OPT) -Wno-error
- ---
CFLAGS+= $(INCLUDE_OPT)

ifeq ($(CC), gcc)
CFLAGS+= -Wno-error
endif
Marc G. Fournier
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org

From scrappy
Date: Thu, 24 Apr 1997 10:30:23 -0300 (ADT)
From: The Hermit Hacker <scrappy@hub.org>
Subject: [PORTS] Re: postgresql 6.0
On Thu, 24 Apr 1997, Michael Reifenberger wrote:
On Wed, 23 Apr 1997, The Hermit Hacker wrote:

Date: Wed, 23 Apr 1997 22:59:30 -0300 (ADT)
From: The Hermit Hacker <scrappy@hub.org>
To: Michael Reifenberger <root@totum.plaut.de>
Subject: Re: postgresql 6.0
On Wed, 23 Apr 1997, Michael Reifenberger wrote:

Hi,
starting with 6.0 I see:

make all install
...
...
Adding template1 database to pg_database...
Running: postgres -boot -C -F -D/usr/local/pgsql/data -Q template1 < /tmp/create.20587

vacuuming template1
Segmentation fault - core dumped
...

Do you have an idea?

Unfortunately I cant debug postgres.
gdb says:
(gdb) core-file postgres.core
"/usr/local/pgsql/data/base/template1/postgres.core" is not a core dump: File truncated
This is using the 'ports' version of PostgreSQL?
Yes.
But I have seen this failure before too using the official version.

Maybe it is related:
I'm using current on a 586 w 64MB mainmem, msun's libm.
I'm using 3.0-CURRENT on a P133 with 32Meg of RAM *shrug*

Let me try rebuilding with ports and see what I come out with,
I've been inside of v6.1 too long now :(

Marc G. Fournier
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org

From scrappy
Date: Thu, 24 Apr 1997 16:01:22 +0200
From: Frits Dumortier <fdumorti@iris.ufsia.ac.be>
Subject: [PORTS] PostgresSQL 6.1beta fails on SGI/O2 - sequel

============================================================================
POSTGRES95 BUG REPORT REPORT
============================================================================


Your name : Frits Dumortier
Your email address : Frits.Dumortier@ufsia.ac.be
fdumorti@Iris.ufsia.ac.be


System Configuration
- ---------------------
Architecture (example: Intel Pentium) : SGI/O2
R10000/150MHz

Operating System (example: Linux 1.3.42 ELF) : Irix 6.3

PostgreSQL version (example: Postgres95-1.01) :
PostgreSQL-6.1beta (970421)

Compiler used (example: gcc 2.7.0) : GNU Make 3.75
(gmake)
GNU fileutils
3.12 (ginstall)
MIPSpro C
compiler driver release 7.1 (cc)


Please enter a FULL description of your problem:
- ------------------------------------------------

PostgreSQL-files installed in:
/usr/src/pgsql

Makefile.custom contains:
PORTNAME= irix5
SRCDIR= /usr/src/pgsql
CUSTOM_INSTALL= ginstall

gmake fails with the following error message:

...cc -o postgres access/SUBSYS.o bootstrap/SUBSYS.o catalog/SUBSYS.o
commands/SUBSYS.o executor/SUBSYS.o lib/SUBSYS.o libpq/SUBSYS.o
main/SUBSYS.o nodes/SUBSYS.o optimizer/SUBSYS.o parser/SUBSYS.o
port/SUBSYS.o postmaster/SUBSYS.o regex/SUBSYS.o rewrite/SUBSYS.o
storage/SUBSYS.o tcop/SUBSYS.o utils/SUBSYS.o ../utils/version.o -lgen
- -lcrypt -lbsd -lnsl -ldl -lm -ltermcap -lcurses
ld: WARNING 84: /usr/lib/libgen.so is not used for resolving any symbol.
ld: WARNING 84: /usr/lib/libdl.so is not used for resolving any symbol.
ld: WARNING 84: /usr/lib/libtermcap.so is not used for resolving any
symbol.
ld: WARNING 134: weak definition of setkey in /usr/lib/libcrypt.so
preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of encrypt in /usr/lib/libcrypt.so
preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of crypt in /usr/lib/libcrypt.so
preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 85: definition of _crypt in /usr/lib/libcrypt.so preempts
that definition in /usr/lib/libc.so.
ld: WARNING 85: definition of _ffs in /usr/lib/libnsl.so preempts that
definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of ffs in /usr/lib/libnsl.so preempts
that weak definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of clnt_create in /usr/lib/libnsl.so
preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of clnt_create_vers in
/usr/lib/libnsl.so preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 85: definition of _clnt_create_vers in /usr/lib/libnsl.so
preempts that definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of clnt_sperror in /usr/lib/libnsl.so
preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of clnt_perror in /usr/lib/libnsl.so
preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of clnt_sperrno in /usr/lib/libnsl.so
preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 85: definition of _clnt_perrno in /usr/lib/libnsl.so
preempts that definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of clnt_perrno in /usr/lib/libnsl.so
preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of clnt_spcreateerror in
/usr/lib/libnsl.so preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of clnt_pcreateerror in
/usr/lib/libnsl.so preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 85: definition of _clntraw_create in /usr/lib/libnsl.so
preempts that definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of clntraw_create in /usr/lib/libnsl.so
preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of callrpc in /usr/lib/libnsl.so
preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 85: definition of _callrpc in /usr/lib/libnsl.so preempts
that definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of clnttcp_create in /usr/lib/libnsl.so
preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of clntudp_bufcreate in
/usr/lib/libnsl.so preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of clntudp_create in /usr/lib/libnsl.so
preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 85: definition of _netname2user in /usr/lib/libnsl.so
preempts that definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of netname2user in /usr/lib/libnsl.so
preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of netname2host in /usr/lib/libnsl.so
preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 85: definition of _netname2host in /usr/lib/libnsl.so
preempts that definition in /usr/lib/libc.so.
ld: WARNING 85: definition of _getnetname in /usr/lib/libnsl.so preempts
that definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of getnetname in /usr/lib/libnsl.so
preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of user2netname in /usr/lib/libnsl.so
preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of host2netname in /usr/lib/libnsl.so
preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of pmap_set in /usr/lib/libnsl.so
preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of pmap_unset in /usr/lib/libnsl.so
preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 85: definition of _pmap_getmaps in /usr/lib/libnsl.so
preempts that definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of pmap_getmaps in /usr/lib/libnsl.so
preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of pmap_getport in /usr/lib/libnsl.so
preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of xdr_pmap in /usr/lib/libnsl.so
preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of xdr_pmaplist in /usr/lib/libnsl.so
preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of pmap_rmtcall in /usr/lib/libnsl.so
preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 85: definition of _pmap_rmtcall in /usr/lib/libnsl.so
preempts that definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of xdr_rmtcall_args in
/usr/lib/libnsl.so preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of xdr_rmtcallres in /usr/lib/libnsl.so
preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of clnt_broadcast in /usr/lib/libnsl.so
preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 85: definition of _clnt_broadcast in /usr/lib/libnsl.so
preempts that definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of xdr_callmsg in /usr/lib/libnsl.so
preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of xdr_opaque_auth in
/usr/lib/libnsl.so preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 85: definition of _xdr_des_block in /usr/lib/libnsl.so
preempts that definition in /usr/lib/libc.so.
ld: Giving up after printing 50 warnings. Use -wall to print all
warnings.
cp catalog/global1.bki.source .
cp catalog/local1_template1.bki.source .
make[1]: Leaving directory `/usr/src/pgsql/src/backend'
/usr/local/bin/make -C libpq all
make[1]: Entering directory `/usr/src/pgsql/src/libpq'
cc -Dirix5 -I../include -I../backend -c fe-auth.c -o fe-auth.o
cc -Dirix5 -I../include -I../backend -c fe-connect.c -o fe-connect.o
cc -Dirix5 -I../include -I../backend -c fe-exec.c -o fe-exec.o
cc -Dirix5 -I../include -I../backend -c fe-misc.c -o fe-misc.o
cc -Dirix5 -I../include -I../backend -c fe-lobj.c -o fe-lobj.o
ln -s ../backend/lib/dllist.c .
cc -Dirix5 -I../include -I../backend -c dllist.c -o dllist.o
cc -Dirix5 -I../include -I../backend -c pqsignal.c -o pqsignal.o
ln -s ../backend/libpq/pqcomprim.c .
cc -Dirix5 -I../include -I../backend -c pqcomprim.c -o pqcomprim.o
ar crs libpq.a `lorder fe-auth.o fe-connect.o fe-exec.o fe-misc.o
fe-lobj.o dllist.o pqsignal.o pqcomprim.o | tsort`
touch libpq.a
rm -f c.h
echo "#undef PORTNAME" > c.h
echo "#define PORTNAME irix5" >> c.h
cat ../include/c.h >> c.h
make[1]: Leaving directory `/usr/src/pgsql/src/libpq'
/usr/local/bin/make -C bin all
make[1]: Entering directory `/usr/src/pgsql/src/bin'
/usr/local/bin/make -C pg_id all
make[2]: Entering directory `/usr/src/pgsql/src/bin/pg_id'
cc -Dirix5 -I/usr/src/pgsql/libpq -I/usr/src/pgsql/include -c
pg_id.c -o pg_id.o
/usr/local/bin/make -C /usr/src/pgsql/libpq libpq.a
make: Entering an unknown directory
make: *** /usr/src/pgsql/libpq: No such file or directory. Stop.
make: Leaving an unknown directory
make[2]: *** [/usr/src/pgsql/libpq/libpq.a] Error 2
make[2]: Leaving directory `/usr/src/pgsql/src/bin/pg_id'
make[1]: *** [all] Error 2
make[1]: Leaving directory `/usr/src/pgsql/src/bin'
make: *** [all] Error 2



Please describe a way to repeat the problem. Please try to provide a
concise reproducible example, if at all possible:
- ----------------------------------------------------------------------

gmake as used above to build the application.



If you know how this problem might be fixed, list the solution below:
- ---------------------------------------------------------------------

From scrappy
Date: Thu, 24 Apr 1997 11:10:18 -0300 (ADT)
From: The Hermit Hacker <scrappy@hub.org>
Subject: [PORTS] PostgreSQL v6.0 as /usr/ports under FreeBSD

Hi...

After getting a report that the port of PostgreSQL v6.0 that is
part of the ports section doesn't work, from Michael, I just re-tried to
install it, and it seems to work correctly on install:

Initializing PostgreSQL Databases - this may take a few minutes...
initdb: using /usr/local/pgsql/lib/local1_template1.bki.source as input to create the template database.
initdb: using /usr/local/pgsql/lib/global1.bki.source as input to create the global classes.
initdb: using /usr/local/pgsql/lib/pg_hba.conf.sample as the host-based authentication control file.

We are initializing the database system with username pgsql (uid=1005).
This user will own all the files and must also own the server process.

Creating Postgres database system directory /usr/local/pgsql/data

Creating Postgres database system directory /usr/local/pgsql/data/base

initdb: creating template database in /usr/local/pgsql/data/base/template1
Running: postgres -boot -C -F -D/usr/local/pgsql/data -Q template1

Creating global classes in /base
Running: postgres -boot -C -F -D/usr/local/pgsql/data -Q template1

Adding template1 database to pg_database...
Running: postgres -boot -C -F -D/usr/local/pgsql/data -Q template1 < /tmp/create.2069

vacuuming template1
Installing /usr/local/etc/rc.d/postgresql.sh startup file.

Now that PostgreSQL is installed, you should read the documentation and
implementation guides. These can be found at:

http://www.PostgreSQL.org/docs

===> Compressing manual pages for postgresql-6.0
===> Registering installation for postgresql-6.0


Michael...what version of FreeBSD are you using? Is anyone else out
there running FreeBSD that can try an install from ports?


Marc G. Fournier
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org

From scrappy
Date: Thu, 24 Apr 1997 11:16:15 -0300 (ADT)
From: The Hermit Hacker <scrappy@hub.org>
Subject: [PORTS] Missing .tar.gz files

Hi...

After having just been informed that there are no files for
the 22nd and 23rd, I've just manually created one for the 24th...there
is, unfortunately, no chngd file for the intervening period, so you'll
have to grab the whole thing..but in beta test mode, you should be grabbing
it all anyway, just in case :)


Marc G. Fournier
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org

From scrappy
Date: Thu, 24 Apr 1997 11:31:04 -0300 (ADT)
From: The Hermit Hacker <scrappy@hub.org>
Subject: Re: [PORTS] PostgresSQL 6.1beta fails on SGI/O2 - sequel
On Thu, 24 Apr 1997, Frits Dumortier wrote:

make[1]: Leaving directory `/usr/src/pgsql/src/libpq'
/usr/local/bin/make -C bin all
make[1]: Entering directory `/usr/src/pgsql/src/bin'
/usr/local/bin/make -C pg_id all
make[2]: Entering directory `/usr/src/pgsql/src/bin/pg_id'
cc -Dirix5 -I/usr/src/pgsql/libpq -I/usr/src/pgsql/include -c
pg_id.c -o pg_id.o
/usr/local/bin/make -C /usr/src/pgsql/libpq libpq.a
make: Entering an unknown directory
make: *** /usr/src/pgsql/libpq: No such file or directory. Stop.
This should fix that, which will be in 970425, but you might want
to grab 970424 before you go any further:

Index: bin/Makefile.global
===================================================================
RCS file: /usr/local/cvsroot/postgres95/src/bin/Makefile.global,v
retrieving revision 1.14
retrieving revision 1.15
diff -r1.14 -r1.15
10c10
< # $Header: /usr/local/cvsroot/postgres95/src/bin/Makefile.global,v 1.14 1997/04/04 10:40:56 scrappy Exp $
- ---
# $Header: /usr/local/cvsroot/postgres95/src/bin/Makefile.global,v 1.15 1997/04/24 14:30:09 scrappy Exp $ 13a14
SRCDIR=..

Marc G. Fournier
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org

From scrappy
Date: Thu, 24 Apr 1997 18:03:05 +0200 (MET DST)
From: "Pedro J. Lobo" <pjlobo@euitt.upm.es>
Subject: [PORTS] How to add compiler flags?

Hi, all.

I'm trying to figure out how to enable some compiler flags if and only if
the platform is xxx (alpha/DU 3.2x in this case) and the compiler is yyy
(DEC cc).

Patching templates/alpha would do the trick for me, but it would fail for
people using GNU's gcc because it doesn't recognize one of the flags I
want to add (-Olimit 1000).

Any ideas? Marc?

=09Pedro.

- -------------------------------------------------------------------
Pedro Jos=E9 Lobo Perea Tel: +34 1 336 78 19
Centro de C=E1lculo Fax: +34 1 331 92 29
EUIT Telecomunicaci=F3n - UPM e-mail: pjlobo@euitt.upm.es

From scrappy
Date: Thu, 24 Apr 1997 15:09:21 -0300 (ADT)
From: The Hermit Hacker <scrappy@hub.org>
Subject: Re: [PORTS] How to add compiler flags?
On Thu, 24 Apr 1997, Pedro J. Lobo wrote:

Hi, all.

I'm trying to figure out how to enable some compiler flags if and only if
the platform is xxx (alpha/DU 3.2x in this case) and the compiler is yyy
(DEC cc).

Patching templates/alpha would do the trick for me, but it would fail for
people using GNU's gcc because it doesn't recognize one of the flags I
want to add (-Olimit 1000).

Any ideas? Marc?
Great a 'alpha-cc' template file :)

Marc G. Fournier
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org

From scrappy
Date: Thu, 24 Apr 1997 17:31:44 -0300 (ADT)
From: The Hermit Hacker <scrappy@hub.org>
Subject: [PORTS] Mail Folders Cleaned Out...

Hi...

Well, I just went through all my different mail folders just in case
I've missed any patches, and believe I've got everyone's submissions included
(after a brief review, of course *grin*)...

Other then patches forthcoming from Thomas for *BSD systems, things
looking okay?


Marc G. Fournier
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org

From scrappy
Date: Thu, 24 Apr 1997 17:44:07 -0400 (EDT)
From: David Friend <dfriend@atlsci.atlsci.com>
Subject: Re: [PORTS] Mail Folders Cleaned Out...
On Thu, 24 Apr 1997, The Hermit Hacker wrote:

Other then patches forthcoming from Thomas for *BSD systems, things
looking okay?
If I find the energy (my time is already reserved for it) I will be
spending Friday evening and Saturday morning testing out the the source
code on a SPARC linux box. (I'm not anticipating any problems.)

I then intend to update the documentation in /usr/src/pgsql (save that the
HISTORY file won't be properly updated because I don't intend to spend the
time poring over several months of posts <grin>). I intend to pay special
attention to file INSTALL.

Offhand, does anyone on a unix system have an example of exactly how you
modify the boot file so that the postmaster is started? I haven't done it
yet myself so I don't have a tested example to put in the INSTALL file.
(When responding, please mention which platform your example is for.)

David Friend ! cq995@freenet.carleton.ca
Atlantis Scientific Systems Group Inc. ! david.friend@atlsci.com
1827 Woodward Drive ! 613-727-1087 (voice)
Ottawa, Ontario, CANADA K2C 0P9 ! 800-265-3894 (voice)
ERGOvista Scientific Image Analysis ! 613-727-5853 (fax)

From scrappy
Date: Thu, 24 Apr 1997 18:57:30 -0300 (ADT)
From: The Hermit Hacker <scrappy@hub.org>
Subject: Re: [PORTS] Mail Folders Cleaned Out...
On Thu, 24 Apr 1997, David Friend wrote:

Offhand, does anyone on a unix system have an example of exactly how you
modify the boot file so that the postmaster is started? I haven't done it
yet myself so I don't have a tested example to put in the INSTALL file.
(When responding, please mention which platform your example is for.)
The one that is in the FAQ should work for just about any platform,
actually...what we use as default in FreeBSD in the Ports collection is:

#!/bin/sh
[ -x /usr/local/pgsql/bin/postmaster ] && su -l pgsql -c '/usr/local/pgsql/bin/postmaster -D/usr/local/pgsql/data -o -F > /usr/local/pgsql/errlog &' && echo -n ' pgsql'


Marc G. Fournier
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org

From scrappy
Date: Thu, 24 Apr 1997 18:18:44 -0400 (EDT)
From: darcy@druid.net (D'Arcy J.M. Cain)
Subject: Re: [PORTS] Mail Folders Cleaned Out...

Thus spake David Friend
Offhand, does anyone on a unix system have an example of exactly how you
modify the boot file so that the postmaster is started? I haven't done it
yet myself so I don't have a tested example to put in the INSTALL file.
(When responding, please mention which platform your example is for.)
I have the following in rc.local on my NetBSD box as well as in an rc2.d
startup script on a Sparc Solaris 2.5.1 box:

su postgres -c "/usr/local/pgsql/bin/postmaster -S -D /usr/local/pgsql/data"

- --
D'Arcy J.M. Cain | Democracy is three wolves
darcy@{druid.net|vex.net} | and a sheep voting on
+1 416 424 2871 (DoD#0082) (eNTP) | what's for dinner.
-- http://www.druid.net/darcy --

From scrappy
Date: Fri, 25 Apr 1997 07:27:33 +0800 (GMT+0800)
From: "Michael J. Maravillo" <mmj@philonline.com.ph>
Subject: Re: [PORTS] Mail Folders Cleaned Out...
On Thu, 24 Apr 1997, D'Arcy J.M. Cain wrote:

Thus spake David Friend
Offhand, does anyone on a unix system have an example of exactly how you
modify the boot file so that the postmaster is started? I haven't done it
yet myself so I don't have a tested example to put in the INSTALL file.
(When responding, please mention which platform your example is for.)
I have the following in rc.local on my NetBSD box as well as in an rc2.d
startup script on a Sparc Solaris 2.5.1 box:

su postgres -c "/usr/local/pgsql/bin/postmaster -S -D /usr/local/pgsql/data"
I have this entry in my (Linux RH 4.0) /etc/inittab just to make sure
it gets "revived" in case it dies.

pg:2345:respawn:/bin/su - postgres -c "/usr/local/pgsql/bin/postmaster -D/usr/local/pgsql/data >> /usr/local/pgsql/server.log 2>&1" > /dev/null

However, I'm really not sure if this has side effects.


Mike
__
Michael J. Maravillo Philippines Online
System Administrator InfoDyne, Incorporated
http://www.philonline.com.ph/~mmj/ 4A N&M Bldg, 1184 Chino Roces, Makati

From scrappy
Date: Fri, 25 Apr 1997 06:02:21 -0400 (EDT)
From: Christopher Hicks <chicks@chicks.net>
Subject: Re: [PORTS] Mail Folders Cleaned Out...
On Thu, 24 Apr 1997, The Hermit Hacker wrote:
#!/bin/sh
[ -x /usr/local/pgsql/bin/postmaster ] && su -l pgsql -c
'/usr/local/pgsql/bin/postmaster -D/usr/local/pgsql/data -o -F >
/usr/local/pgsql/errlog &' && echo -n ' pgsql'
That's cool, but it'd be more comprehensible, easuer to configure, and
more portable to do:

PGUSER=postgres # Change to your taste
PGDIR=/home/postgres # Customize as appropriate

# You may need to customize these:
POSTMASTER=$PGDIR/bin/postmaster
PGDATA=$PGDIR/data
PGLOG=$PGDIR/postmaster.log

if [ -x $POSTMASTER && -d $PGDATA && -e $PGLOG ]
then
su -c '$POSTMASTER -D$PGDATA -o -F > $PGLOG &' $PGUSER && echo " pg"
else
echo "Postgres didn't start: no postmaster, no data dir, or no log"
fi

That was off the top of my head, but you get the idea.

</chris>

Any chance of those paper cups and string being upgraded to tin cans
and wire? Or as a coworker has said . . . I've seen better throughput
from a pair of gorillas and flash cards. -Jon Lewis <jlewis@fdt.net>

From scrappy
Date: Fri, 25 Apr 1997 11:07:23 -0700
From: "Arthur M. Kang" <arthur@netrominc.com>
Subject: [PORTS] Solaris x86 Compile Fail

I'm trying to compile the latest beta of Postgresql 6.1
(970425.tar.gz). I tried both the i386_solaris-cc and generic templates
and both of them failed with the failure appearing like:

bootstrap.c:169: initializer element for `Procid[15].outproc' is not
constant
bootstrap.c:170: `F_ARRAY_IN' undeclared here (not in a function)
bootstrap.c:170: initializer element for `Procid[16].inproc' is not
constant
bootstrap.c:170: `F_ARRAY_OUT' undeclared here (not in a function)
bootstrap.c:170: initializer element for `Procid[16].outproc' is not
constant
bootstrap.c:171: `F_ARRAY_IN' undeclared here (not in a function)
bootstrap.c:171: initializer element for `Procid[17].inproc' is not
constant
bootstrap.c:171: `F_ARRAY_OUT' undeclared here (not in a function)
bootstrap.c:171: initializer element for `Procid[17].outproc' is not
constant
gmake[2]: *** [bootstrap.o] Error 1
gmake[2]: Leaving directory
`/spider/downloads/postgres/src/backend/bootstrap'
gmake[1]: *** [bootstrap.dir] Error 2
gmake[1]: Leaving directory `/spider/downloads/postgres/src/backend'
gmake: *** [all] Error 2
$

Is there a i386_solaris-gcc template to be??

Any help in solving this and getting Postgresql to compile would be
great!

Arthur
- --
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Arthur M. Kang, Webmaster - Netrom, Inc.
mailto:arthur@netrominc.com http://www.netrominc.com
1770 Kettner Boulevard San Diego, California 92101
Phone : (619) 234-9283 - FAX : (619) 234-2821
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=

From scrappy
Date: Sat, 26 Apr 1997 13:38:23 +0900
From: "=?EUC-KR?B?wOW0673E?=" <dsjang@vision.soongsil.ac.kr>
Subject: [PORTS] Porting problem

============================================================================

POSTGRES95 BUG REPORT TEMPLATE
============================================================================



Your name : Daesik Jang
Your email address : dsjang@vision.soongsil.ac.kr


System Configuration
- ---------------------
Architecture (example: Intel Pentium) : Sun Sparc Classic

Operating System (example: Linux 1.3.42 ELF) : Sunos 5.4

PostgreSQL version (example: Postgres95-1.01) : PostgreSQL-6.0

Compiler used (example: gcc 2.7.0) : gcc


Please enter a FULL description of your problem:
- ------------------------------------------------
I tried to port PostgreSQL-6.0 to our machine but I couldn't.
I set the PORTNAME to sparc_solaris.
The error occured in the file 'progres.c'.
The error is 'invalid use of undefined type "struct rusage"' at the line
1570.
There are several more errors but they are all derived the the error.




Please describe a way to repeat the problem. Please try to provide a
concise reproducible example, if at all possible:
- ----------------------------------------------------------------------
I followed the INSTALL file.
At first, I modified the Makefile.global. I changed the PORTNAME to
sparc_solaris.
And then, I ran 'gmake'.
In the process of 'gmake' there was a compile error.





If you know how this problem might be fixed, list the solution below:
- ---------------------------------------------------------------------
I think there is no definition about structure 'rusage'.

===========================================
Daesik Jang
Vision Laboratory
School of Computing, Soongsil University
Phone : +82-2-825-1087
email : dsjang@vision.soongsil.ac.kr
===========================================

From scrappy
Date: Sat, 26 Apr 1997 02:02:09 -0300 (ADT)
From: The Hermit Hacker <scrappy@hub.org>
Subject: Re: [PORTS] Porting problem
On Sat, 26 Apr 1997, [EUC-KR] Àå´ë½Ä wrote:

============================================================================

POSTGRES95 BUG REPORT TEMPLATE
============================================================================



Your name : Daesik Jang
Your email address : dsjang@vision.soongsil.ac.kr


System Configuration
---------------------
Architecture (example: Intel Pentium) : Sun Sparc Classic

Operating System (example: Linux 1.3.42 ELF) : Sunos 5.4

PostgreSQL version (example: Postgres95-1.01) : PostgreSQL-6.0

Compiler used (example: gcc 2.7.0) : gcc


Please enter a FULL description of your problem:
------------------------------------------------
I tried to port PostgreSQL-6.0 to our machine but I couldn't.
I set the PORTNAME to sparc_solaris.
The error occured in the file 'progres.c'.
The error is 'invalid use of undefined type "struct rusage"' at the line
1570.
There are several more errors but they are all derived the the error.
Please grab a copy of v6.1 BETA and try that out, as these problems,
and many others, are fixed in this version. Release date for a full release
is set for May 1st, if you want to wait for a Released version...

From scrappy
Date: Sat, 26 Apr 1997 23:02:07 +1000 (EST)
From: Ben Elliston <bje@air.net.au>
Subject: [PORTS] Win32 port

Has anyone managed to port PortgreSQL to Win32? I see lots of conditional
code for Win32, but it's not listed as a supported platform in the
documentation.

If I have to attempt the port myself, is the stuff in the src/test/
directory sufficient to prove that it works? Are there more extensive
test suites out there that could be used?

Thanks,
Ben.

- ---
Ben Elliston
<bje@air.net.au>

"For my birthday I got a humidifier and a de-humidifier. I put them in the same
room and let them fight it out." -- Steven Wright

From scrappy
Date: Sat, 26 Apr 1997 10:19:41 -0300 (ADT)
From: The Hermit Hacker <scrappy@hub.org>
Subject: Re: [PORTS] Win32 port
On Sat, 26 Apr 1997, Ben Elliston wrote:

Has anyone managed to port PortgreSQL to Win32? I see lots of conditional
code for Win32, but it's not listed as a supported platform in the
documentation.

If I have to attempt the port myself, is the stuff in the src/test/
directory sufficient to prove that it works? Are there more extensive
test suites out there that could be used?
src/test is all we have right now...but if someone wants to come up
with addittional tests to be included, I'm willing to look at it :)


Marc G. Fournier
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org

From scrappy
Date: Sat, 26 Apr 1997 23:23:14 +1000 (EST)
From: Ben Elliston <bje@air.net.au>
Subject: Re: [PORTS] Win32 port
On Sat, 26 Apr 1997, The Hermit Hacker wrote:

If I have to attempt the port myself, is the stuff in the src/test/
directory sufficient to prove that it works? Are there more extensive
test suites out there that could be used?
src/test is all we have right now...but if someone wants to come up
with addittional tests to be included, I'm willing to look at it :)
Sure.

So, do I take this to mean there is no working Win32 port right now? Is
there any technical reason why it doesn't work?

Ben

From scrappy
Date: Sat, 26 Apr 1997 10:31:40 -0300 (ADT)
From: The Hermit Hacker <scrappy@hub.org>
Subject: Re: [PORTS] Win32 port
On Sat, 26 Apr 1997, Ben Elliston wrote:
On Sat, 26 Apr 1997, The Hermit Hacker wrote:

If I have to attempt the port myself, is the stuff in the src/test/
directory sufficient to prove that it works? Are there more extensive
test suites out there that could be used?
src/test is all we have right now...but if someone wants to come up
with addittional tests to be included, I'm willing to look at it :)
Sure.

So, do I take this to mean there is no working Win32 port right now? Is
there any technical reason why it doesn't work?
Nope, in fact, I thought we took *out* all the Win32 related
stuff...the only reason it doesn't work...nobody was interested in making
it work. Other then that, I know of no technical reason.

I'd recommend not putting too much work into it until after v6.1
is released, or, at least, don't submit any patches on it, since we're
trying to get v6.1 'releasably' by the 1st and are trying to make sure we
don't throw any new instabilities in :)

But after that, all and any patches are more then welcome :)

Marc G. Fournier
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org

From scrappy
Date: Sat, 26 Apr 1997 13:52:44 -0400 (EDT)
From: David Friend <dfriend@atlsci.atlsci.com>
Subject: Re: [PORTS] Win32 port
On Sat, 26 Apr 1997, Ben Elliston wrote:

Has anyone managed to port PortgreSQL to Win32? I see lots of conditional
code for Win32, but it's not listed as a supported platform in the
documentation.

If I have to attempt the port myself, is the stuff in the src/test/
directory sufficient to prove that it works? Are there more extensive
test suites out there that could be used?
Go to the hackers archives at
ftp://ftp.postgresql.org/pub/majordomo/pgsql-hackers-digest. Look for
messages I mailed on March 27 and April 12th. Look for a message from
Martin.H.Ludwig@rz.ruhr_uni_bachan.de on April 11th. (I forget who I was
replying to in March.)

Martin ported the libpq or libpq++ library (I forget which). He wants to
port the whole thing. The fellow from March (I assume he is a different
person) also wants to do a port. If the three of you got together, you
could start a sub-team to port the whole thing. I would suggest you start
with the client side, since the project would be much smaller and would be
useful even if you never finish the server side port.

Good luck!

David Friend ! cq995@freenet.carleton.ca
Atlantis Scientific Systems Group Inc. ! david.friend@atlsci.com
1827 Woodward Drive ! 613-727-1087 (voice)
Ottawa, Ontario, CANADA K2C 0P9 ! 800-265-3894 (voice)
ERGOvista Scientific Image Analysis ! 613-727-5853 (fax)

From scrappy
Date: Sun, 27 Apr 1997 08:11:21 +1000 (EST)
From: Ben Elliston <bje@air.net.au>
Subject: Re: [PORTS] Win32 port
Go to the hackers archives at
ftp://ftp.postgresql.org/pub/majordomo/pgsql-hackers-digest. Look for
messages I mailed on March 27 and April 12th. Look for a message from
Martin.H.Ludwig@rz.ruhr_uni_bachan.de on April 11th. (I forget who I was
replying to in March.)
Thanks for the info..
Martin ported the libpq or libpq++ library (I forget which). He wants to
port the whole thing. The fellow from March (I assume he is a different
person) also wants to do a port. If the three of you got together, you
could start a sub-team to port the whole thing. I would suggest you start
with the client side, since the project would be much smaller and would be
useful even if you never finish the server side port.
Agreed, but in my case, if I don't get the server side ported to Win32,
PostgreSQL will be a whole lot less appealing for someone trying to run my
software on WinNT. They'd need a UNIX box as well! ;-)

Cheers,
Ben.

- ---
Ben Elliston
<bje@air.net.au>

"For my birthday I got a humidifier and a de-humidifier. I put them in the same
room and let them fight it out." -- Steven Wright

From scrappy
Date: Mon, 28 Apr 97 20:53:59 KST
From: radioOn@pl.ewha.ac.kr
Subject: [none]



From scrappy
Date: Mon, 28 Apr 97 20:56:02 KST
From: radioOn@pl.ewha.ac.kr
Subject: [none]

To report a bug, please complete the following form and send it by
email to bugs@postgreSQL.org.

============================================================================
POSTGRES95 BUG REPORT TEMPLATE
============================================================================


Your name :Hyun-Sook,Choi
Your email address :radioOn@pl.ewha.ac.kr


System Configuration
- ---------------------
Architecture (example: Intel Pentium) :SunSparc

Operating System (example: Linux 1.3.42 ELF) :SunOS 4.1.3

PostgreSQL version (example: Postgres95-1.01) : PostgreSQL-6.0

Compiler used (example: gcc 2.7.0) :gcc 2.7.2


Please enter a FULL description of your problem:
- ------------------------------------------------
I have failed install PostgreSQL.
After editing Makefile.global,and executing 'make' in /usr/src/pgsql/src, I got the messages following...

(omitted)
make -C nbtree SUBSYS.o
make[3]: Entering directory `/usr/share/src/pgsql/src/backend/access/nbtree'
gcc -Wall -Wmissing-prototypes -O2 -Dsunos4 -I../.. -I../../port/sunos4 -I../.
./../include -c nbtsort.c -o nbtsort.o
In file included from nbtsort.c:61:
../../../include/regex/utils.h:57: warning: `memmove' redefined
../../../include/c.h:741: warning: this is the location of the previous definiti
on
nbtsort.c: In function `_bt_taperewind':
nbtsort.c:272: `SEEK_SET' undeclared (first use this function)
nbtsort.c:272: (Each undeclared identifier is reported only once
nbtsort.c:272: for each function it appears in.)
nbtsort.c: In function `_bt_spoolinit':
nbtsort.c:457: warning: implicit declaration of function `memset'
make[3]: *** [nbtsort.o] Error 1
make[3]: Leaving directory `/usr/share/src/pgsql/src/backend/access/nbtree'
make[2]: *** [submake] Error 2
make[2]: Leaving directory `/usr/share/src/pgsql/src/backend/access'
make[1]: *** [postgres_group1] Error 2
make[1]: Leaving directory `/usr/share/src/pgsql/src/backend'
make: *** [all] Error 2

I think header files are linked in the directory of /usr/include.
Please tell me what's wrong.
Thank you.




Please describe a way to repeat the problem. Please try to provide a
concise reproducible example, if at all possible:
- ----------------------------------------------------------------------





If you know how this problem might be fixed, list the solution below:
- ---------------------------------------------------------------------

From scrappy
Date: Mon, 28 Apr 97 20:46:30 KST
From: radioOn@pl.ewha.ac.kr
Subject: [none]



From scrappy
Date: Mon, 28 Apr 97 20:50:04 KST
From: radioOn@pl.ewha.ac.kr
Subject: [none]



From scrappy
Date: Mon, 28 Apr 1997 11:53:26 -0400 (EDT)
From: David Friend <dfriend@atlsci.atlsci.com>
Subject: [PORTS] Support for Sparc/Linux platform

It didn't work! :(

I just installed 970428.tar.gz on my Sparc Linux platform. Most of the
regression tests failed. I couldn't use the system either. (Although I
was able to load in my database okay.)

I don't know if the failure is due to problems with the patch supplied by
Satoshi Ishikawa <power@sekine.densi.gifu-u.ac.jp> or if some other
problem developed between version 6.0 and version 6.1.

Unless other Linux people or SPARC people are having trouble with this
version, I would say that we will have to leave support for the
Sparc/Linux platform out of version 6.1 and get it fixed later. <sigh>

David Friend ! cq995@freenet.carleton.ca
Atlantis Scientific Systems Group Inc. ! david.friend@atlsci.com
1827 Woodward Drive ! 613-727-1087 (voice)
Ottawa, Ontario, CANADA K2C 0P9 ! 800-265-3894 (voice)
ERGOvista Scientific Image Analysis ! 613-727-5853 (fax)

From scrappy
Date: Mon, 28 Apr 1997 17:49:01 -0400
From: Lee Roth <roth@tpusa.com>
Subject: [PORTS] Fail: 6.1beta on DEC Alpha does not like 'Point'

============================================================================
POSTGRESQL BUG REPORT TEMPLATE
============================================================================


Your name : Lee Roth
Your email address : roth@tpusa.com


System Configuration
- ---------------------
Architecture (example: Intel Pentium) : DEC Alpha

Operating System (example: Linux 1.3.42 ELF) : Digital Unix V3.2g

PostgreSQL version (example: PostgreSQL-6.1) : PostgreSQL-6.1 28-Apr-97

Compiler used (example: gcc 2.7.0) : Digital 'C' complier, V?.?


Please enter a FULL description of your problem:
- ------------------------------------------------

Build fails when trying to include 'geo_decls.h'.


Please describe a way to repeat the problem. Please try to provide a
concise reproducible example, if at all possible:
- ----------------------------------------------------------------------

Excerpt from build:

cc -I../../../include -DNOFIXADE -Dalpha -I../.. -I../../port/alpha
- -I../../../include
-c scankey.c
cc -I../../../include -DNOFIXADE -Dalpha -I../.. -I../../port/alpha
- -I../../../include
-c tupdesc.c
/usr/lib/cmplrs/cc/cfe: Error: ../../../include/utils/geo_decls.h, line 87:
member types cannot be function types or incomplete type.
Point p[0];
----- ^
/usr/lib/cmplrs/cc/cfe: Warning: ../../../include/utils/geo_decls.h, line 87:
Must be an integer type greater than zero.
Point p[0];
----- --^
/usr/lib/cmplrs/cc/cfe: Error: ../../../include/utils/geo_decls.h, line 122:
member types cannot be function types or incomplete type.
Point p[0];
------^
/usr/lib/cmplrs/cc/cfe: Warning: ../../../include/utils/geo_decls.h, line 122:
Must be an integer type greater than zero.
Point p[0];
--------^
gmake[3]: *** [tupdesc.o] Error 1
gmake[3]: Leaving directory `/usr/local/postgres6_1/src/backend/access/common'
gmake[2]: *** [submake] Error 2
gmake[2]: Leaving directory `/usr/local/postgres6_1/src/backend/access'
gmake[1]: *** [access.dir] Error 2
gmake[1]: Leaving directory `/usr/local/postgres6_1/src/backend'
gmake: *** [all] Error 2



If you know how this problem might be fixed, list the solution below:
- ---------------------------------------------------------------------

Unknown.



Lee Roth
Trade Point USA
roth@tpusa.com

From scrappy
Date: Mon, 28 Apr 1997 19:34:21 -0400 (EDT)
From: David Friend <dfriend@atlsci.atlsci.com>
Subject: Re: [PORTS] Support for Sparc/Linux platform
On Mon, 28 Apr 1997, The Hermit Hacker wrote:
On Mon, 28 Apr 1997, Thomas Lockhart wrote:

I think that there are other problems with v6.1alpha/beta
which are _not_ Sparc-specific. My guess is that folks
will be willing to postpone the release until the
problems (or suspicion of such) are resolved. Don't give
up hope!
My question is is anyone actually looking into this, or are the
bugs just being reported? If nobody is looking into it, then there is
no point in holding off on a release, since will be on hold from here
till eternity :(
So far, the sparc linux platform has not been supported. Several of us
discovered that if you take the v6.0 release and make it use either the
SPARC SunOS version of tas() or the one that comes with linux then it
would work.

Since there are only two or three of us using a Sparc Linux platform it
doesn't make sense to wait for this to be supported. I know I don't have
time. I can live without the new features for a while (although multiple
field unique indexes would be REAL nice).

David Friend ! cq995@freenet.carleton.ca
Atlantis Scientific Systems Group Inc. ! david.friend@atlsci.com
1827 Woodward Drive ! 613-727-1087 (voice)
Ottawa, Ontario, CANADA K2C 0P9 ! 800-265-3894 (voice)
ERGOvista Scientific Image Analysis ! 613-727-5853 (fax)

From scrappy
Date: Tue, 29 Apr 1997 00:44:17 GMT
From: Thomas Lockhart <Thomas.G.Lockhart@jpl.nasa.gov>
Subject: Re: [PORTS] Fail: 6.1beta on DEC Alpha does not like 'Point'

The alpha box apparently does not like zero-indexed
declarations. It could possibly be fixed by removing the
zeros or by replacing them with a one (1). Can you
try both solutions and let me know what works? Don't know
what would be the impact on other compilers, but will
test with gcc tonight. Will see if I can fix it for v6.1.
Thanks for the info.

- Tom (tgl@mythos.jpl.nasa.gov)

From scrappy
Date: Tue, 29 Apr 1997 11:07:38 +0200 (MET DST)
From: "Pedro J. Lobo" <pjlobo@euitt.upm.es>
Subject: Re: [PORTS] Fail: 6.1beta on DEC Alpha does not like 'Point'
On Tue, 29 Apr 1997, Thomas Lockhart wrote:

The alpha box apparently does not like zero-indexed
declarations. It could possibly be fixed by removing the
zeros or by replacing them with a one (1). Can you
try both solutions and let me know what works? Don't know
what would be the impact on other compilers, but will
test with gcc tonight. Will see if I can fix it for v6.1.
Thanks for the info.
I sent a patch to patches@postgresql.org some days ago (I think it was
last Friday, don't remember exactly). The patch substitutes "Point p[0]"
with "Point *p", and works nicely on any compiler (I don't know if an
array with 0 length is standard C, but using a pointer as an array is).

=09Pedro.

- -------------------------------------------------------------------
Pedro Jos=E9 Lobo Perea Tel: +34 1 336 78 19
Centro de C=E1lculo Fax: +34 1 331 92 29
EUIT Telecomunicaci=F3n - UPM e-mail: pjlobo@euitt.upm.es

From scrappy
Date: Tue, 29 Apr 1997 06:59:27 -0400 (EDT)
From: Unprivileged user <nobody>
Subject: [PORTS] Port Bug Report: TEST

============================================================================
POSTGRESQL BUG REPORT TEMPLATE
============================================================================


Your name :
Your email address :

Category : unknown
Severity : non-critical

Summary: TEST

System Configuration
- --------------------
Operating System :

PostgreSQL version : 6.0

Compiler used :

Hardware:
- ---------


Versions of other tools:
- ------------------------


- --------------------------------------------------------------------------

Problem Description:
- --------------------


- --------------------------------------------------------------------------

Test Case:
- ----------


- --------------------------------------------------------------------------

Solution:
- ---------


- --------------------------------------------------------------------------

From scrappy
Date: Tue, 29 Apr 1997 07:02:46 -0400 (EDT)
From: Unprivileged user <nobody>
Subject: [PORTS] Port Bug Report: TEST

============================================================================
POSTGRESQL BUG REPORT TEMPLATE
============================================================================


Your name :
Your email address :

Category : unknown
Severity : non-critical

Summary: TEST

System Configuration
- --------------------
Operating System :

PostgreSQL version : 6.0

Compiler used :

Hardware:
- ---------


Versions of other tools:
- ------------------------


- --------------------------------------------------------------------------

Problem Description:
- --------------------


- --------------------------------------------------------------------------

Test Case:
- ----------


- --------------------------------------------------------------------------

Solution:
- ---------


- --------------------------------------------------------------------------

From scrappy
Date: Tue, 29 Apr 1997 07:03:08 -0400 (EDT)
From: Unprivileged user <nobody>
Subject: [PORTS] Port Bug Report: TEST

============================================================================
POSTGRESQL BUG REPORT TEMPLATE
============================================================================


Your name :
Your email address :

Category : unknown
Severity : non-critical

Summary: TEST

System Configuration
- --------------------
Operating System :

PostgreSQL version : 6.0

Compiler used :

Hardware:
- ---------


Versions of other tools:
- ------------------------


- --------------------------------------------------------------------------

Problem Description:
- --------------------


- --------------------------------------------------------------------------

Test Case:
- ----------


- --------------------------------------------------------------------------

Solution:
- ---------


- --------------------------------------------------------------------------

From scrappy
Date: Tue, 29 Apr 1997 08:55:39 -0400
From: Lee Roth <roth@tpusa.com>
Subject: Re: [PORTS] Fail: 6.1beta on DEC Alpha does not like 'Point'
At 11:07 AM 4/29/97 +0200, Pedro J. Lobo wrote:
On Tue, 29 Apr 1997, Thomas Lockhart wrote:

The alpha box apparently does not like zero-indexed
declarations. It could possibly be fixed by removing the
zeros or by replacing them with a one (1). Can you
try both solutions and let me know what works? Don't know
what would be the impact on other compilers, but will
test with gcc tonight. Will see if I can fix it for v6.1.
Thanks for the info.
I sent a patch to patches@postgresql.org some days ago (I think it was
last Friday, don't remember exactly). The patch substitutes "Point p[0]"
with "Point *p", and works nicely on any compiler (I don't know if an
array with 0 length is standard C, but using a pointer as an array is).

Pedro.
I'm going to give Pedro's suggestion a try and see what happens.

Thanks,



Lee Roth
Trade Point USA
roth@tpusa.com

From scrappy
Date: Wed, 30 Apr 1997 05:35:57 -0400 (EDT)
From: Unprivileged user <nobody>
Subject: [PORTS] Port Bug Report: warnings and compile crash at specified location

============================================================================
POSTGRESQL BUG REPORT TEMPLATE
============================================================================


Your name : Frits Dumortier
Your email address : Frits.Dumortier@UFSIA.ac.be

Category : install: compile
Severity : critical

Summary: warnings and compile crash at specified location

System Configuration
- --------------------
Operating System : IRIX 6.3

PostgreSQL version : 6.0

Compiler used : MIPSpro C 7.1

Hardware:
- ---------
Silicon Graphics O2
IRIX Iris 6.3 12161207 IP32

Versions of other tools:
- ------------------------
gmake: GNU Make version 3.75
ginstall: GNU fileutils 3.12

- --------------------------------------------------------------------------

Problem Description:
- --------------------
1) warnings: 3 types

cfe: Warning 654: ../../include/utils/geo_decls.h, line 87: Must be an integer type greater than zero.
Point p[0];
----- --^
cfe: Warning 654: ../../include/utils/geo_decls.h, line 122: Must be an integer type greater than zero.
Point p[0];

--------^
ld: WARNING 15: multiply defined:(PoolSize) in geqo_eval.o and geqo_main.o (2nd definition ignored).
ld: WARNING 15: multiply defined:(Generations) in geqo_eval.o and geqo_main.o (2nd definition ignored).
ld: WARNING 15: multiply defined:(RandomSeed) in geqo_eval.o and geqo_main.o (2nd definition ignored).
ld: WARNING 15: multiply defined:(SelectionBias) in geqo_eval.o and geqo_main.o (2nd definition ignored).

ld: WARNING 84: /usr/lib/libgen.so is not used for resolving any symbol.
ld: WARNING 84: /usr/lib/libdl.so is not used for resolving any symbol.
ld: WARNING 85: definition of box in utils/SUBSYS.o preempts that definition in /usr/lib/libtermcap.so.
ld: WARNING 84: /usr/lib/libtermcap.so is not used for resolving any symbol.
ld: WARNING 134: weak definition of setkey in /usr/lib/libcrypt.so preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of encrypt in /usr/lib/libcrypt.so preempts that weak definition in /usr/lib/libc.so.
ld: WARNING 134: weak definition of crypt in /usr/lib/libcrypt.so preempts that weak definition in /usr/lib/libc.so.

2) compiler crash:

cc -I/usr/src/pgsql/include -Dirix5 -c pg_id.c -o pg_id.o
cc -o pg_id -L../../libpq pg_id.o -lpq -lgen -lcrypt -lbsd -lnsl -ldl -lm -ltermcap -lcurses
ld: FATAL 11: Object file format error (../../libpq/libpq.a): invalid header
make[2]: *** [pg_id] Error 1
make[2]: Leaving directory `/usr/src/pgsql/src/bin/pg_id'
make[1]: *** [all] Error 2
make[1]: Leaving directory `/usr/src/pgsql/src/bin'
make: *** [all] Error 2


- --------------------------------------------------------------------------

Test Case:
- ----------
gmake always reproduces the same sequence

- --------------------------------------------------------------------------

Solution:
- ---------


- --------------------------------------------------------------------------

From scrappy
Date: Wed, 30 Apr 1997 10:37:57 -0400 (EDT)
From: Unprivileged user <nobody>
Subject: [PORTS] Port Bug Report: what compression tool did you use?

============================================================================
POSTGRESQL BUG REPORT TEMPLATE
============================================================================


Your name : john.makar
Your email address : john.makar@state.co.us

Category : install: other
Severity : serious

Summary: what compression tool did you use?

System Configuration
- --------------------
Operating System : OSF/1 & SunOS4.1.3

PostgreSQL version : 6.0

Compiler used :

Hardware:
- ---------
dec alpha and sparc2

Versions of other tools:
- ------------------------


- --------------------------------------------------------------------------

Problem Description:
- --------------------
I am unfamiliar with the compression tool that creates .gz compression files. My main box is an alpha running osf/1 and any number of GNU things that people have 'contributed'. How do I unpack this thing.

- --------------------------------------------------------------------------

Test Case:
- ----------


- --------------------------------------------------------------------------

Solution:
- ---------


- --------------------------------------------------------------------------

From scrappy
Date: Wed, 30 Apr 1997 11:39:07 -0400 (EDT)
From: Unprivileged user <nobody>
Subject: [PORTS] Port Bug Report: problems with grammatical accent

============================================================================
POSTGRESQL BUG REPORT TEMPLATE
============================================================================


Your name : Manuel Leboucher
Your email address : webmaster@unicaen.fr

Category : runtime: back-end: SQL
Severity : serious

Summary: problems with grammatical accent

System Configuration
- --------------------
Operating System : OSF1 no V2.0 240 alpha

PostgreSQL version : 95

Compiler used :

Hardware:
- ---------
OSF1 no V2.0 240 alpha

Versions of other tools:
- ------------------------


- --------------------------------------------------------------------------

Problem Description:
- --------------------
when making a query like
select * from Table where Label ~*'e';

(where the e is an acute accent)

I get
FATAL: unexpected results from the backend, it probably dumped core.FATAL: unexpected results from the backend, it probably dumped core.

- --------------------------------------------------------------------------

Test Case:
- ----------
just make query using grammatical accent as above

- --------------------------------------------------------------------------

Solution:
- ---------


- --------------------------------------------------------------------------

From scrappy
Date: Wed, 30 Apr 1997 16:12:18 -0400 (EDT)
From: Unprivileged user <nobody>
Subject: [PORTS] Port Bug Report: Can't make further connections to PostgreSQL

============================================================================
POSTGRESQL BUG REPORT TEMPLATE
============================================================================


Your name : Craig Anderson
Your email address : csa@transmeta.com

Category : runtime: back-end
Severity : non-critical

Summary: Can't make further connections to PostgreSQL

System Configuration
- --------------------
Operating System : Linux 2.0.29 ELF (Red Hat?)

PostgreSQL version : 6.0

Compiler used : gcc 2.6.2

Hardware:
- ---------
Linux vlfarm101.transmeta.com 2.1.29 #3-smp Thu Apr 10 15:24:52 PDT 1997 i686
128M RAM, Pentium Pro 200MHz

Versions of other tools:
- ------------------------
GNU Make version 3.74, by Richard Stallman and Roland McGrath.
flex version 2.5.4


- --------------------------------------------------------------------------

Problem Description:
- --------------------
Here is the error message:
NOTICE:SIAssignBackendId: discarding tag 2147483645
FATAL 1:Backend cache invalidation initialization failed

- --------------------------------------------------------------------------

Test Case:
- ----------
Sorry.....

- --------------------------------------------------------------------------

Solution:
- ---------


- --------------------------------------------------------------------------

Search Discussions

Related Discussions

Discussion Navigation
viewthread | post
Discussion Overview
grouppgsql-ports @
categoriespostgresql
postedApr 17, '97 at 4:32p
activeApr 17, '97 at 4:32p
posts1
users1
websitepostgresql.org
irc#postgresql

1 user in discussion

The Hermit Hacker: 1 post

People

Translate

site design / logo © 2022 Grokbase