Return-Path: postman 
Delivery-Date: Mon, 11 Oct 93 04:52:31 PDT
Return-Path: postman
Received: by postgres.Berkeley.EDU (5.61/1.29)
	id AA00946; Mon, 11 Oct 93 04:47:41 -0700
Resent-From: postman (POSTGRES mailing list)
Resent-Message-Id: <9310111147.AA00946@postgres.Berkeley.EDU>
Sender: owner-postman@postgres.Berkeley.EDU
X-Return-Path: tom@izf.tno.nl
Received: from izfcs.izf.tno.nl by postgres.Berkeley.EDU (5.61/1.29)
	id AA00938; Mon, 11 Oct 93 04:47:32 -0700
Received: from rainbow.izf.tno.nl by izfcs.izf.tno.nl (4.1/SMI-4.1)
	id AA11461; Mon, 11 Oct 93 12:50:44 +0100
Received: by rainbow.izf.tno.nl (5.0/SMI-SVR4)
	id AA02326; Mon, 11 Oct 93 12:48:01 +0100
Date: Mon, 11 Oct 93 12:48:01 +0100
From: tom@izf.tno.nl (Tom Vijlbrief)
Message-Id: <9310111148.AA02326@rainbow.izf.tno.nl>
To: shlam@ie.cuhk.hk, postgres@postgres.berkeley.edu
Subject: Re: system v rel 4 port now available!
X-Sun-Charset: US-ASCII
Content-Length: 1452
Resent-To: postgres-dist
Resent-Date: Mon, 11 Oct 93 04:47:40 PDT


>> robert withrow (witr@rwwa.COM) has ported 4.1 to intel svr4 and made 
>> the results publically available.  (this port includes a considerable 
>> number of bug fixes for 4.1 as well.)
>
>> you can obtain it from:
>> toe.cs.berkeley.edu:/pub/postgres/unofficial-ports/svr4/postgres4.1-svr4.shar.z
>
>Do you know anyone has successfully ported postgres 4.1 to SunOS 5.2
>(Solaris 2.2)? We are currently using postgres 4.1 on SunOS 4.1.3.
>However, the new Sun workstations are migrating to SunOS 5.2 (Solaris
>2.2). We need to port the postgres 4.1 to SunOS 5.2 (Solaris 2.2).
>
>Alan S. H. Lam
>Department of Information Engineering, CUHK, Hong Kong
>E-mail: shlam@ie.cuhk.hk 
>Tel: (852) 609 8364 Fax: (852) 603 5032
>

I am currently testing such a port. I'll be finished in one or two days.
The port is fairly straight forward. I expect it will be almost identical
to the intel svr4 port.

The only real problem is that gcc (I don't use Sun's cc) expects
doubles to be aligned on 8 byte boundaries and this assumption is
violated by Postgres.

There exists a gcc flag '-munaligned-doubles' which fixes part of this
problem. A full fix requires a minor gcc patch which I have made.

I am currently testing this patch and the postgres dynamic loader modifications.

I don't have the time to merge my fixes with the original postgres source tree
or to give any other support. If there is a volunteer for this job then
let him contact me......

Tom
