Return to BSD News archive
Path: euryale.cc.adfa.oz.au!newshost.carno.net.au!harbinger.cc.monash.edu.au!munnari.OZ.AU!spool.mu.edu!howland.erols.net!vixen.cso.uiuc.edu!sdd.hp.com!usc!news.cerf.net!news.sprintlink.net!news-ana-7.sprintlink.net!neonramp.com!cynjut.neonramp.com!cynjut.neonramp.com!not-for-mail
From: burgess@cynjut.neonramp.com (Dave Burgess)
Newsgroups: comp.unix.bsd.netbsd.announce,comp.unix.bsd.freebsd.announce,comp.answers,news.answers
Subject: [comp.unix.bsd] NetBSD, FreeBSD, and 386BSD (0.1) FAQ (Part 10 of 10)
Supersedes: <386bsd-faq-10-838497777@cynjut.neonramp.com>
Followup-To: comp.unix.bsd.netbsd.misc
Date: 28 Aug 1996 09:13:34 -0500
Organization: Dave's House in Omaha
Lines: 245
Approved: news-answers-request@MIT.Edu,cgd@sun-lamp.cs.berkeley.edu
Expires: 09/15/96 09:13:23 CDT
Message-ID: <386bsd-faq-10-841241603@cynjut.neonramp.com>
References: <386bsd-faq-1-841241603@cynjut.neonramp.com>
Reply-To: burgess@cynjut.neonramp.com (386bsd FAQ Maintainer)
NNTP-Posting-Host: cynjut.neonramp.com
Keywords: FAQ 386bsd NetBSD FreeBSD !Linux
X-Posting-Frequency: Posted on/about the 13th and the 27th of every month.
Xref: euryale.cc.adfa.oz.au comp.unix.bsd.netbsd.announce:244 comp.unix.bsd.freebsd.announce:322 comp.answers:18845 news.answers:73314
Posted-By: auto-faq 3.1.1.2
Archive-name: 386bsd-faq/part10
Section 9 ("Supported" Software List).
9.0 What GNU software has been tested and is working with Net/2 derived
BSD systems for the 386?
Just about all of it.
9.1 Has anyone ever gotten news to work?
The program 'news' running on 386bsd. Here is a quick summary of
the major places to stumble:
1) get bash, gmake, gcc 2.X, cnews, trn (or your favorite reader).
2) Make uucp work. (Read the info files that come with the
original distribution for the whole scoop on configuration
files.)
Ed Note: This step is not needed if you are implementing SLIP,
PPP, or are directly connected to a network.
3) Edit all the scripts which come with cnews and replace every
occurence of /bin/sh with /usr/local/bin/bash (or wherever you put
it).
4) Build cnews using bash, gmake and gcc 2.x
5) Install cnews in the directories you want it. Some hand-hacking
of the intall scripts is required (Too long ago to remember the
details).
6) Change the permissions on all the scripts from execute only to
read-execute for group and other. (On 386bsd, if you can't read
a script, you can't execute it).
7) Set up uucp to accept news
8) Post an article and steal it out of the uucp queue before it
gets sent. Feed it to your rnews (as user uucp) instead and make
sure that it does not bomb out with permission denied or some such.
9) Have fun!
Implementing innd is even easier. The configure script that comes
with the system has been modified to work more correctly with
Net/2 derived BSD systems. The first is that the LINTLIBSTYLE
option in config.data needs to be set to NONE, since NetBSD and
FreeBSD don't come with lint. With that changed, the system
should work right out of the box.
If you are running with memory mapped files, you will also need
to make the following patch:
--- icd.c.orig Tue Feb 7 13:36:50 1995
+++ icd.c Tue Feb 7 14:56:27 1995
@@ -366,7 +366,9 @@
ICDwriteactive()
{
#if defined(ACT_MMAP)
- /* No-op. */
+ if (msync(ICDactpointer, 0)) {
+ syslog(L_ERROR, "msync error on active file: %m");
+ }
#else
9.1.1 I want to make sure I have every set up right for my news
partition. What newfs options do I need to use to get this
information stored OK without future problems?
There has been a lot of discussion of the years about the defaul
options for newfs. If you have "modern" disks and you created
your filesystems with 1.0, or with a pre-9412 -current, then
you may want to back them up and then re-create them. u
Filesystems created with the current defaults should be much
faster.
The newfs(8) defaults are equivalent to `-a 8 -d 0 -n 1'.
To make you news server software work better, you should
increase the number of inodes available, you should include
either '-i 512' or '-i 1024' depending on the normal size of the
files in the filesystem. News partitions are often the
repository for many files which are very small, averaging less
than 512 bytes per file. By quadrupling the number of inodes
(using -i 512 instead of the default 2048) you make it more
likely that you will run out of disk SPACE before you run out of
disk INODES.
9.2 How did you get emacs to compile?
The problem is in the dump-emacs function. It writes the image
header and then the text section of the image overwrites the
header. This leaves you with a bad image. If you try to load it
into gdb, it will tell you that it is not an executable.
What to do?
Look back at your configuration command:
>% configure i386-intel-386bsd --with-x=no
The 386bsd qualifier means that the compiles will include the file
src/s/386bsd.h. If you go into this file and add the lines:
#define A_TEXT_OFFSET(x) (sizeof (struct exec))
#define A_TEXT_SEEK(hdr) (N_TXTOFF(hdr) + A_TEXT_OFFSET(hdr))
This tells the subroutines in src/unexec.c about the 32 byte image
header, so that they will set up the header appropriately and not step
on it while writing the emacs executable.
The second problem is that emacs tries use its own crt0.o file.
Kill that line in the Makefile and you should be able to compile
for either static or shared library operations.
9.2 Has anyone tried to get Postgres to work?
Jim Bachesta and his crew have gotten Postgres 4.2 working in
the i386 version of NetBSD 1.0. The netbsd source tree is
available from:
ftp://charon.amdahl.com:pub/agc/postgres-4.2-src-netbsd-v2.tar.gz
The regular postgres distribution is available from:
ftp://s2k-ftp.cs.berkeley.edu:pub/postgres
Get the standard distribution and then overlay the NetBSD source
distribution over it for a complete system.
There is also work in progress to get Postgres95 working.
Check the following URL for more information:
ftp://s2k-ftp.cs.berkeley.edu/pub/postgres95/postgres95-1.0.tar.gz
It works fine on NetBSD/i386 1.1. I've heard that it works
fine on the sparc port, too, so there don't seem to be any
byte-order funnies in there (although take a look in the
www/bugs/p*.html for 14 patches that should be applied to
the 1.0 sources - at least one of them deals with
order-dependencies when the backend is on a different
byte-ordered machine to the client program).
Someone mentioned that you need dynamic loading, and so you
may be out of luck if you're on one of the more esoteric
ports. I'm not sure about this, and would say that pg95
should run fine, albeit with reduced functionality, without
dynamic loading - it just means that you can't define C
functions for the backend to load at will. However, I
haven't tried this. (From memory, the previous v4r2 port
didn't have support for dynamic loading, and most of the
regression tests ran fine.)
9.3 How about the BSD Song?
In a dark dim machine room
Cool A/C in my hair
Warm smell of silicon
Rising up through the air
Up ahead in the distance
I saw a Solarian(tm) light
My kernel grew heavy, and my disk grew slim
I had to halt(8) for the night
The backup spun in the tape drive
I heard a terminal bell
And I was thinking to myself
This could be BSD or USL
Then they started a lawsuit
And they showed me the way
There were salesmen down the corridor
I thought I heard them say
Welcome to Berkeley California
Such a lovely place
Such a lovely place (backgrounded)
Such a lovely trace(1)
Plenty of jobs at Berkeley California
Any time of year
Any time of year (backgrounded)
You can find one here
You can find one here
Their code was definately twisted
But they've got the stock market trends
They've got a lot of pretty, pretty lawyers
That they call friends
How they dance in the courtroom
See BSDI sweat
Some sue to remember
Some sue to forget
So I called up Kernighan
Please bring me ctime(3)
He said
We haven't had that tm_year since 1969
And still those functions are calling from far away
Wake up Jobs in the middle of the night
Just to hear them say
Welcome to Berkeley California
Such a lovely Place
Such a lovely Place (backgrounded)
Such a lovely trace(1)
They're livin' it up suing Berkeley California
What a nice surprise
What a nice surprise (backgrounded)
Bring your alibies
Windows NT a dreaming
Pink OS on ice
And they said
We are all just prisoners here
Of a marketing device
And in the judges's chambers
They gathered for the feast
They diff(1)'d the source code listings
But they can't kill -9 the beast
Last thing I remember
I was restore(8)'ing | more(1)
I had to find the soft link back to the path I was before
sleep(3) said the pagedaemon
We are programmed to recv(2)
You can swap out any time you like
But you can never leave(1)
[ substitute whirring of disk and tape drives for guitar solo ]
Written by David Barr <barr@pop.psu.edu>
and Ken Hornstein <kenh@physci.psu.edu>
and a little help from Greg Nagy <nagy@cs.psu.edu>
and thanks to the lyrics archive at cs.uwp.edu
--
Dave Burgess (The man of a thousand E-Mail addresses)
*bsd FAQ Maintainer / SysAdmin for the NetBSD system in my spare bedroom
"Just because something is stupid doesn't mean there isn't someone that
doesn't want to do it...."