Snapshot 23Feb98 under Irix5 --- INSTALLATION BROKEN!!!

Started by Andrew Martinalmost 28 years ago4 messages
#1Andrew Martin
martin@biochemistry.ucl.ac.uk

OK, here are my experiences trying to install the latest snapshot (23Feb98)
under Irix 5.3.

In summary, installation is currently BROKEN!!!!!

I shall wait till installation works properly before I try another regression
test :-)

Andrew

Installation of Snapshot 23Feb98 under Irix 5.3
===============================================

1. .../src/template/.similar: Should say irix5 not irix
-------------------------------------------------------

*** .similar.old        Mon Feb 23 14:18:37 1998
--- .similar    Mon Feb 23 14:13:49 1998
***************
*** 13,19 ****
  i586-pc-linux-gnu=linux-elf
  i586-pc-sco3.2v=sco
  i686-pc-linux-gnu=linux-elf
! mips-sgi-irix=irix
  powerpc-ibm-aix3.2.5=aix_325
  powerpc-ibm-aix4.1.4.0=aix_41
  powerpc-ibm-aix4.1.5.0=aix_41
--- 13,19 ----
  i586-pc-linux-gnu=linux-elf
  i586-pc-sco3.2v=sco
  i686-pc-linux-gnu=linux-elf
! mips-sgi-irix=irix5
  powerpc-ibm-aix3.2.5=aix_325
  powerpc-ibm-aix4.1.4.0=aix_41
  powerpc-ibm-aix4.1.5.0=aix_41

2. configure dies (!)
---------------------
Running configure, dies with the following:

checking whether time.h and sys/time.h may both be included... yes
checking whether struct tm is in sys/time.h or time.h... time.h
checking for int timezone... yes
checking for union semun... yes
Argument expected

This occurs during the check for whether gcc requires -traditional
when I have specified --with-compiler=cc in order to use the native
IRIX cc compiler and not gcc. (If I opt to use gcc, then I don't
have this problem...)

NOTE this bug has been introduced SINCE THE LAST SNAPSHOT I tried
on the 11th Feb

3. My expected output files for Irix have not been incorporated in
.../src/test/regress/expected/ :-(
------------------------------------------------------------------

4. gmake install fails for ecpg files
-------------------------------------

The ecpg files are using
install -c -m 755
to do the installation rather than
$(INSTALL) $(INSTLOPTS)

This applies to .../src/interfaces/ecpg/include/Makefile
and .../src/interfaces/ecpg/preproc/Makefile

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

#2The Hermit Hacker
scrappy@hub.org
In reply to: Andrew Martin (#1)
Re: [HACKERS] Snapshot 23Feb98 under Irix5 --- INSTALLATION BROKEN!!!

On Mon, 23 Feb 1998, Andrew Martin wrote:

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

1. .../src/template/.similar: Should say irix5 not irix
-------------------------------------------------------

*** .similar.old        Mon Feb 23 14:18:37 1998
--- .similar    Mon Feb 23 14:13:49 1998
***************
*** 13,19 ****
i586-pc-linux-gnu=linux-elf
i586-pc-sco3.2v=sco
i686-pc-linux-gnu=linux-elf
! mips-sgi-irix=irix
powerpc-ibm-aix3.2.5=aix_325
powerpc-ibm-aix4.1.4.0=aix_41
powerpc-ibm-aix4.1.5.0=aix_41
--- 13,19 ----
i586-pc-linux-gnu=linux-elf
i586-pc-sco3.2v=sco
i686-pc-linux-gnu=linux-elf
! mips-sgi-irix=irix5
powerpc-ibm-aix3.2.5=aix_325
powerpc-ibm-aix4.1.4.0=aix_41
powerpc-ibm-aix4.1.5.0=aix_41

Fixed...

2. configure dies (!)
---------------------
Running configure, dies with the following:

checking whether time.h and sys/time.h may both be included... yes
checking whether struct tm is in sys/time.h or time.h... time.h
checking for int timezone... yes
checking for union semun... yes
Argument expected

Please investigate this one further and let me know what is
required to fix it?

3. My expected output files for Irix have not been incorporated in
.../src/test/regress/expected/ :-(
------------------------------------------------------------------

Erk...resend it to me privately?

4. gmake install fails for ecpg files
-------------------------------------

The ecpg files are using
install -c -m 755
to do the installation rather than
$(INSTALL) $(INSTLOPTS)

This applies to .../src/interfaces/ecpg/include/Makefile
and .../src/interfaces/ecpg/preproc/Makefile

Damn, thought I had got all of these :( Fixed now...

#3Noname
darrenk@insightdist.com
In reply to: The Hermit Hacker (#2)
Re: [HACKERS] Snapshot 23Feb98 under Irix5 --- INSTALLATION BROKEN!!!

2. configure dies (!)
---------------------
Running configure, dies with the following:

checking whether time.h and sys/time.h may both be included... yes
checking whether struct tm is in sys/time.h or time.h... time.h
checking for int timezone... yes
checking for union semun... yes
Argument expected

Please investigate this one further and let me know what is
required to fix it?

I complained about this one a few months back. There is a test condition
that needs to be double-quoted. In my generated configure, it's at line
#3230 in the test whether gcc needs -traditional.

I seem to recall that it was more a bug in autoconf that in the postgres
configure script.

darrenk

#4Andrew Martin
martin@biochemistry.ucl.ac.uk
In reply to: Noname (#3)
Re: [HACKERS] Snapshot 23Feb98 under Irix5 --- INSTALLATION BROKEN!!!

2. configure dies (!)
---------------------
Running configure, dies with the following:

checking whether time.h and sys/time.h may both be included... yes
checking whether struct tm is in sys/time.h or time.h... time.h
checking for int timezone... yes
checking for union semun... yes
Argument expected

Please investigate this one further and let me know what is
required to fix it?

I understand nothing about how autoconf works...
All I can say from hacking the configure script itself is that it's
during the test for whether gcc needs --traditional which is being
run even when the compiler is not gcc

3. My expected output files for Irix have not been incorporated in
.../src/test/regress/expected/ :-(
------------------------------------------------------------------

Erk...resend it to me privately?

I hope I still have it:-) If not, I'll regenerate on the next snapshot
I try - probably tomorrow

Bets wishes,

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