snapshot won't compile on irix6.2

Has anyone confirmed these problems under Irix?

Hi,

i have just grabbed the 6.3 snapshot and tried to compile it on irix 6.2 -
without success.

The following errors occurred:

1) configure fails

% ./configure
[...]
linking ./backend/port/tas/i386_solaris.s to backend/port/tas.s
linking ./backend/port/dynloader/irix5.c to backend/port/dynloader.c
configure: error: ./backend/port/dynloader/irix5.c: File not found

This seems a little fundamental!!!

2) compiling the backend fails:
in src/backend the gcc call
gcc -mabi=64 -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 -64 -lPW -lcrypt -lm -lbsd -lreadline -lhistory
-ltermcap -lcurses

returns
[...]
ld64: ERROR 33: Unresolved text symbol "S_INIT_LOCK" -- 1st referenced by
storage/SUBSYS.o.
ld64: ERROR 33: Unresolved text symbol "S_UNLOCK" -- 1st referenced by
storage/SUBSYS.o.
ld64: ERROR 33: Unresolved text symbol "S_LOCK" -- 1st referenced by
storage/SUBSYS.o

these functions just seem to be implemented for linux/alpha ??

I'm generally a little dubious when people report a problem in actually
compiling 'cos of the Irix 6.x bug in ld. However, this guy is using
gcc, not native Irix cc/ld

Fuad Abdallah
Max-Planck-Institut fuer Zuechtungsforschung / ZWDV
Tel.: 0221/5062-739 / Priv: 0221/584563

With any luck I should have a bit of time next week to do some testing under
Irix - things have been rather hectic the last few weeks.

Andrew

----------------------------------------------------------------------------
Dr. Andrew C.R. Martin University College London
EMAIL: (Work) martin@biochem.ucl.ac.uk (Home) andrew@stagleys.demon.co.uk
URL: http://www.biochem.ucl.ac.uk/~martin
Tel: (Work) +44(0)171 419 3890 (Home) +44(0)1372 275775

On Mon, 2 Feb 1998, Andrew Martin wrote:

[...]
linking ./backend/port/tas/i386_solaris.s to backend/port/tas.s
linking ./backend/port/dynloader/irix5.c to backend/port/dynloader.c
configure: error: ./backend/port/dynloader/irix5.c: File not found

This seems a little fundamental!!!'

Fixed now...

returns
[...]
ld64: ERROR 33: Unresolved text symbol "S_INIT_LOCK" -- 1st referenced by
storage/SUBSYS.o.
ld64: ERROR 33: Unresolved text symbol "S_UNLOCK" -- 1st referenced by
storage/SUBSYS.o.
ld64: ERROR 33: Unresolved text symbol "S_LOCK" -- 1st referenced by
storage/SUBSYS.o

these functions just seem to be implemented for linux/alpha ??

I'm generally a little dubious when people report a problem in actually
compiling 'cos of the Irix 6.x bug in ld. However, this guy is using
gcc, not native Irix cc/ld

From include/storage/s_lock.h:

#if defined(irix5)

Using tools/ccsym, what shoudl this be set to. There is no longer
any concept of a "port" at the Makefile level, and hasn't been for over a
month now...From henceforth, its either determined based on what configure
knows about, or the compiler itself...in this case, tools/ccsym will have
to be used to determine what the compiler defines internally and the
settings changed...

From include/storage/s_lock.h:

#if defined(irix5)

Using tools/ccsym, what shoudl this be set to. There is no longer
any concept of a "port" at the Makefile level, and hasn't been for over a
month now...From henceforth, its either determined based on what configure
knows about, or the compiler itself...in this case, tools/ccsym will have
to be used to determine what the compiler defines internally and the
settings changed...

OK, I just ran ccsym (from PGSQL V6.1 - I know it's old...) using the native
Irix cc and with gcc. (See attached results)

I guess __sgi or _SGI_SOURCE look like the best options.

Andrew

Native C compiler
=================

Irix5 Irix6
----- -----
LANGUAGE_C -Ac=xansi
MIPSEB LANGUAGE_C
SYSTYPE_SVR4 MIPSEB
_ABIO32=1 _ABIN32=2
_CFE _COMPILER_VERSION=710
_LANGUAGE_C _LANGUAGE_C
_LONGLONG _LONGLONG
_MIPSEB _MIPSEB
_MIPS_FPSET=16 _MIPS_FPSET=32
_MIPS_ISA=1 _MIPS_ISA=4
_MIPS_SIM=_ABIO32 _MIPS_SIM=_ABIN32
_MIPS_SZINT=32 _MIPS_SZINT=32
_MIPS_SZLONG=32 _MIPS_SZLONG=32
_MIPS_SZPTR=32 _MIPS_SZPTR=32
_MODERN_C _MODERN_C
_PIC _PIC
_SGI_SOURCE _SGI_SOURCE
_SVR4_SOURCE _SVR4_SOURCE
_SYSTYPE_SVR4 _SYSTYPE_SVR4
__DSO__ __DSO__
__EXTENSIONS__ __EXTENSIONS__
__INLINE_INTRINSICS __INLINE_INTRINSICS
__STDC__=1 __MATH_HAS_NO_SIDE_EFFECTS
__host_mips __host_mips
__mips=1 __mips=4
__sgi __sgi
__unix __unix
host_mips host_mips
machine=mips machine=mips
mips mips
sgi sgi
system=unix system=unix
unix unix

GCC
===

Irix5 Irix6
----- -----
LANGUAGE_C LANGUAGE_C
MIPSEB MIPSEB
SYSTYPE_SVR4 SYSTYPE_SVR4
_LANGUAGE_C _LANGUAGE_C
_LONGLONG _LONGLONG
_MIPSEB _MIPSEB
_MIPSEB _MIPSEB
_MIPS_FPSET=16 _MIPS_FPSET=16
_MIPS_ISA=_MIPS_ISA_MIPS1 _MIPS_ISA=_MIPS_ISA_MIPS1
_MIPS_SIM=_MIPS_SIM_ABI32 _MIPS_SIM=_MIPS_SIM_ABI32
_MIPS_SIM=_MIPS_SIM_ABI32 _MIPS_SIM=_MIPS_SIM_ABI32
_MIPS_SZINT=32 _MIPS_SZINT=32
_MIPS_SZLONG=32 _MIPS_SZLONG=32
_MIPS_SZPTR=32 _MIPS_SZPTR=32
_MIPS_SZPTR=32 _MIPS_SZPTR=32
_MODERN_C _MODERN_C
_MODERN_C _MODERN_C
_SGI_SOURCE _SGI_SOURCE
_SVR4_SOURCE _SVR4_SOURCE
_SVR4_SOURCE _SVR4_SOURCE
__CHAR_UNSIGNED__ __CHAR_UNSIGNED__
__DSO__ __DSO__
__DSO__ __DSO__
__EXTENSIONS__ __EXTENSIONS__
__GNUC_MINOR__=7 __GNUC_MINOR__=7
__GNUC__=2 __GNUC__=2
__MIPSEB __MIPSEB
__MIPSEB__ __MIPSEB__
__PTRDIFF_TYPE__=int __PTRDIFF_TYPE__=int
__SIZE_TYPE__=unsigned __SIZE_TYPE__=unsigned
__SYSTYPE_SVR4 __SYSTYPE_SVR4
__SYSTYPE_SVR4__ __SYSTYPE_SVR4__
__host_mips __host_mips
__host_mips__ __host_mips__
__mips __mips
__mips__ __mips__
__sgi __sgi
__sgi__ __sgi__
__unix __unix
__unix__ __unix__
cpu=mips cpu=mips
host_mips host_mips
machine=sgi machine=sgi
mips mips
sgi sgi
system=svr4 system=svr4
system=unix system=unix
unix unix
----------------------------------------------------------------------------
Dr. Andrew C.R. Martin University College London
EMAIL: (Work) martin@biochem.ucl.ac.uk (Home) andrew@stagleys.demon.co.uk
URL: http://www.biochem.ucl.ac.uk/~martin
Tel: (Work) +44(0)171 419 3890 (Home) +44(0)1372 275775

On Tue, 3 Feb 1998, Andrew Martin wrote:

From include/storage/s_lock.h:

#if defined(irix5)

Using tools/ccsym, what shoudl this be set to. There is no longer
any concept of a "port" at the Makefile level, and hasn't been for over a
month now...From henceforth, its either determined based on what configure
knows about, or the compiler itself...in this case, tools/ccsym will have
to be used to determine what the compiler defines internally and the
settings changed...

OK, I just ran ccsym (from PGSQL V6.1 - I know it's old...) using the native
Irix cc and with gcc. (See attached results)

I guess __sgi or _SGI_SOURCE look like the best options.

Done

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