Depending on system install scripts (was Re: COBOL)

Lists: pgsql-bugspgsql-hackers
From: Jarom Hagen <jhagen(at)telematch(dot)com>
To: pgsql-bugs(at)postgresql(dot)org
Subject: COBOL
Date: 2001-03-06 19:02:40
Message-ID: 3AA53450.3C0A43F8@telematch.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-bugs pgsql-hackers

Why do I get a message about COBOL when doing a gmake install and what
am I supposed to do about it?

This is on a Solaris 2.7 system.

Thanks,

Jarom

P.S. The make was successful.

gmake[3]: `SUBSYS.o' is up to date.
gmake[3]: Leaving directory
`/data/postgresql-7.0.3/src/backend/utils/time'
gmake[2]: Leaving directory `/data/postgresql-7.0.3/src/backend/utils'
/usr/local/bin/install -c -m 555 postgres /usr/local/pgsql/bin/postgres
You must have a COBOL system present to install this product
gmake[1]: *** [install-bin] Error 1
gmake[1]: Leaving directory `/data/postgresql-7.0.3/src/backend'
gmake: *** [install] Error 2


From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Jarom Hagen <jhagen(at)telematch(dot)com>
Cc: pgsql-bugs(at)postgresql(dot)org
Subject: Re: COBOL
Date: 2001-03-08 00:38:30
Message-ID: 22498.984011910@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-bugs pgsql-hackers

Jarom Hagen <jhagen(at)telematch(dot)com> writes:
> /usr/local/bin/install -c -m 555 postgres /usr/local/pgsql/bin/postgres
> You must have a COBOL system present to install this product

Weird. It looks like you have some exceedingly nonstandard program
in /usr/local/bin/install --- certainly not what configure thought that
that program would do, anyway. Do you know where that program came from
(perhaps a Sun COBOL package)?

A nondestructive workaround would be to hand-edit src/Makefile.global's
INSTALL variable to refer to our install-sh script (also in src/) rather
than /usr/local/bin/install. However, that install is going to bite a
lot of other open-source packages that expect to find a standard-ish
install script available, so I'd suggest deleting or at least renaming
it...

regards, tom lane


From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Jarom Hagen <jhagen(at)telematch(dot)com>
Cc: pgsql-hackers(at)postgreSQL(dot)org, pgsql-bugs(at)postgreSQL(dot)org
Subject: Depending on system install scripts (was Re: COBOL)
Date: 2001-03-08 14:59:34
Message-ID: 25534.984063574@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-bugs pgsql-hackers

Jarom Hagen <jhagen(at)telematch(dot)com> writes:
> Yes, we have an evil COBOL compiler from MicroFocus that put that install
> script there. I was really confused why postgres wanted a COBOL system. :-)

I've suggested a couple of times that since we include install-sh in our
distro anyway, it's pointless and unnecessarily risky to go looking for
a platform-supplied install program. However, I could never quite get
anyone else to see the reasoning. Now that I have this sterling example
to point to, I'm going to start rattling the cage again. Why don't we
get rid of the configure-time search for 'install', and just always use
our own script?

regards, tom lane

> On Wed, Mar 07, 2001 at 07:38:30PM -0500, Tom Lane wrote:
>> Jarom Hagen <jhagen(at)telematch(dot)com> writes:
>>> /usr/local/bin/install -c -m 555 postgres /usr/local/pgsql/bin/postgres
>>> You must have a COBOL system present to install this product
>>
>> Weird. It looks like you have some exceedingly nonstandard program
>> in /usr/local/bin/install --- certainly not what configure thought that
>> that program would do, anyway. Do you know where that program came from
>> (perhaps a Sun COBOL package)?
>>
>> A nondestructive workaround would be to hand-edit src/Makefile.global's
>> INSTALL variable to refer to our install-sh script (also in src/) rather
>> than /usr/local/bin/install. However, that install is going to bite a
>> lot of other open-source packages that expect to find a standard-ish
>> install script available, so I'd suggest deleting or at least renaming
>> it...


From: Peter Eisentraut <peter_e(at)gmx(dot)net>
To: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
Cc: Jarom Hagen <jhagen(at)telematch(dot)com>, <pgsql-hackers(at)postgreSQL(dot)org>, <pgsql-bugs(at)postgreSQL(dot)org>
Subject: Re: Depending on system install scripts (was Re: COBOL)
Date: 2001-03-08 23:13:10
Message-ID: Pine.LNX.4.30.0103090012170.1061-100000@peter.localdomain
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-bugs pgsql-hackers

Tom Lane writes:

> I've suggested a couple of times that since we include install-sh in our
> distro anyway, it's pointless and unnecessarily risky to go looking for
> a platform-supplied install program. However, I could never quite get
> anyone else to see the reasoning. Now that I have this sterling example
> to point to, I'm going to start rattling the cage again. Why don't we
> get rid of the configure-time search for 'install', and just always use
> our own script?

I've sent this to the Autoconf list for some comment, but in general I
agree with you.

--
Peter Eisentraut peter_e(at)gmx(dot)net http://yi.org/peter-e/


From: Ian Lance Taylor <ian(at)airs(dot)com>
To: Peter Eisentraut <peter_e(at)gmx(dot)net>
Cc: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, Jarom Hagen <jhagen(at)telematch(dot)com>, <pgsql-hackers(at)postgreSQL(dot)org>, <pgsql-bugs(at)postgreSQL(dot)org>
Subject: Re: [HACKERS] Re: Depending on system install scripts (was Re: COBOL)
Date: 2001-03-09 00:18:43
Message-ID: sisnkneqvw.fsf@daffy.airs.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-bugs pgsql-hackers

Peter Eisentraut <peter_e(at)gmx(dot)net> writes:

> Tom Lane writes:
>
> > I've suggested a couple of times that since we include install-sh in our
> > distro anyway, it's pointless and unnecessarily risky to go looking for
> > a platform-supplied install program. However, I could never quite get
> > anyone else to see the reasoning. Now that I have this sterling example
> > to point to, I'm going to start rattling the cage again. Why don't we
> > get rid of the configure-time search for 'install', and just always use
> > our own script?
>
> I've sent this to the Autoconf list for some comment, but in general I
> agree with you.

All the programs which use the Cygnus configure tree (e.g., gdb, GNU
binutils) always use install-sh rather than the system install
program.

The system install program can be faster. But it isn't standardized,
so if you want to be highly portable, using the shell script really is
best.

Ian

---------------------------(end of broadcast)---------------------------
TIP 83: Drop the vase and it will become a Ming of the past.
-- The Adventurer