Re: UTF-8

[Matthew Szudzik]

On Thu, Aug 05, 2010 at 11:33:23AM -0700, Philip Guenther wrote:
> Umm, punycode wasn't developed because of problems with
> distinguishability.  Indeed, it does nothing to solve those, so I'm
> not sure why you would suggest that.  punycode exists to encode

Although punycode may not have been developed to solve problems with
distinguishability, it is used for that purpose.  For example, punycode
is commonly used as a defense against phishers who impersonate online
banks using URLs that are indistinguishable from the banks' actual URLs.
 http://en.wikipedia.org/wiki/IDN_homograph_attack
But with a properly-designed font, ASCII characters are all easily
distinguishable.

> > when I'm working from a terminal, I want to see the data as closely as
> > possible to the way that the computer "sees" the data.
> 
> So you want to see '41' instead of the letter 'A'?  That's "how the
> computer sees the data"...

I simply want to be able to know "at one glance" what data the computer
is using.  For that purpose, it is unnecessary to decode an "A" as 0x41.
The ASCII character is sufficient.

> experience by using LC_ALL=C.  Oops, never mind, OpenBSD hasn't
> actually implemented "plain ASCII only" for years.

The fact that OpenBSD doesn't implement "plain ASCII only" doesn't mean
that it shouldn't. ;)

And by the way, the Turing quote is from the paper in which he first
proposed the idea of a mechanical computer.  He argued that it is
sufficient for a computer to have a finite character set where each
character can be distinguished at a glance.  His argument begins at the
bottom of this page:
 http://www.turingarchive.org/viewer/?id=466&title=01u
and continues onto the top of the next.  Although it is not necessary
that we follow his proposal, it has served as a historical precedent
since the very beginning of computing.

Re: UTF-8

[Matthew Szudzik]

On Wed, Aug 04, 2010 at 01:36:17PM -0700, Matthew Dempsky wrote:
> Is there any useful documentation that explains how you're supposed to
> write C code and what's changed under the i18n New World Order?  From
> your message, it sounds like we're going to have to rewrite nearly all
> of our user-space code...

Not only does switching to unicode require a lot of work, but it
requires perpetual, unending work.  Unicode has the foolish goal of
including all known characters, so every time a country invents a new
currency symbol, for example, the unicode fonts (such as DejaVu) must be
updated to include the symbol and the C library has to be updated to
recognize that the symbol is printable, and so on.  It requires constant
maintenance.

But it's even worse, because unicode also violates the principle
(established by Alan Turing in 1936) that any two characters should be
humanly distinguishable "at a glance".  This has led to the invention of
punycode for translating unicode strings into humanly distinguishable
ASCII strings.  But then why did we switch from ASCII to unicode in the
first place?

It's my opinion that unicode shouldn't have a place in the Unix
terminal.  You might want your GUI to display unicode characters, but
when I'm working from a terminal, I want to see the data as closely as
possible to the way that the computer "sees" the data.  For example, I
don't want nvi to display unicode characters, I want to see each
individual 8-bit byte that composes the character.  I don't want unicode
on the command line.

Re: mount ffs as msdos, system hangs

[Matthew Szudzik]

On Mon, Jul 26, 2010 at 12:21:53PM -0600, Theo de Raadt wrote:
> There is no readlabelfs() for NFS filesystems.  You are being too
> specific in saying how it works.

In the case of an NFS filesystem, mount(8) just checks whether the
"special" string contains a ":" or "@" character.  So, the diff is
literally correct: "mount(8) attempts to guess a file system type using
special and readlabelfs(3)."

Re: mount ffs as msdos, system hangs

[Matthew Szudzik]

On Mon, Jul 26, 2010 at 10:07:37AM -0600, Theo de Raadt wrote:
> This diff isn't quite right.  There are no disklabels on NFS
> partitions; heck, there's no true disklabel on a MSDOS-only memory
> stick.  The language you've written is too specific.

How about the following?


--- sbin/mount/mount.8.orig Mon Jul 26 11:59:30 2010
+++ sbin/mount/mount.8  Mon Jul 26 17:02:03 2010
@@ -269,12 +269,16 @@
 The argument following the
 .Fl t
 is used to indicate the file system type.
-The type
-.Ar ffs
-is the default.
+If this option is omitted, then
+.Nm
+attempts to guess a file system type using
+.Ar special
+and
+.Xr readlabelfs 3 .
 The
 .Fl t
 option can be used
+to override this behavior and
 to indicate that the actions should only be taken on
 file systems of the specified type.
 More than one type may be specified in a comma separated list.

Re: mount ffs as msdos, system hangs

[Matthew Szudzik]

On Sun, Jul 25, 2010 at 09:48:23PM -0400, Kenneth R Westerback wrote:
> If you wanted to mount according to the partition type number, DON'T
> USE '-t '. We give you the option to OVERRIDE the partition
> type number and you made use of that override. You have taken command

I believe that this thread is the result of documentation error on the
mount(8) man page.  A patch for the man page is at the end of this
message.

The problem is that the man page says:

 The argument following the -t is used to indicate the file system type.
 The type ffs is the default.

This is incorrect.  According to lines 248 through 251 of the source
file at

 http://www.openbsd.org/cgi-bin/cvsweb/src/sbin/mount/mount.c?annotate=1.50

the actual default behavior is to check the disklabel's filesystem type:

 If -t flag has not been specified, and spec contains either
 a ':' or a '@' then assume that an NFS filesystem is being
 specified ala Sun.  If not, check the disklabel for a
 known filesystem type.

The original poster in this thread is wondering why mount(8) did not
check the disklabel's filesystem type when he specified the -t option.
The answer is that mount(8) checks the disklabel's filesystem type by
default, but that -t overrides that default behavior.


--- sbin/mount/mount.8.orig Mon Jul 26 11:59:30 2010
+++ sbin/mount/mount.8  Mon Jul 26 12:48:45 2010
@@ -269,12 +269,15 @@
 The argument following the
 .Fl t
 is used to indicate the file system type.
-The type
-.Ar ffs
-is the default.
+If this option is omitted, then
+.Nm
+attempts to guess a file system type using the 
+.Xr disklabel 5
+and other information.
 The
 .Fl t
 option can be used
+to override this behavior and
 to indicate that the actions should only be taken on
 file systems of the specified type.
 More than one type may be specified in a comma separated list.

Re: OpenBSD as a laptop OS

[Matthew Szudzik]

On Fri, Jun 18, 2010 at 11:12:44PM +, Jacob Meuser wrote:
> > And yt or youtube-dl from ports. Also, the greasmonkey scripts or
> > whatever for firefox work great for youtube, vimeo and a few others.
> 
> yeah, greasemonkey + "youtube without flash auto" is *way* better
> than swfdec or gnash for watching/downloading videos.
> 
> http://userscripts.org/scripts/show/50771

These prepackaged choices (yt, youtube-dl, and the greasemonkey script)
are all a little too bloated for my taste.  Did you notice that the
youtube-dl script

 http://bitbucket.org/rg3/youtube-dl/src/tip/youtube-dl

is over 2000 lines long!  That's longer than the Surf web browser (which
is under 1000 lines).

 http://hg.suckless.org/surf/file/e83fbd17d63a/surf.c

I would rather maintain my own 4 line shell script for downloading
youtube videos:

sed -n 's/[^v]*v.\([^&'\'']*\).*/curl '\''http:\/\/www.youtube.com\/ge'\
't_video_info?video_id=\1'\'' ; echo/p' | sh | sed -n 's/.*&video_id=\'\
'([^&'\'']*\).*&token=\([^%'\'']*\).*/curl -L -o '\''\1.mp4'\'' '\''ht'\
'tp:\/\/www.youtube.com\/get_video?video_id=\1\&t=\2=\&fmt=18'\''/p' | sh

The script takes URLs for youtube videos as standard input (one per
line), and downloads the mp4 files.  And if it doesn't do exactly what I
need, then I customize it.

Understanding youtube's API is easier than depending on somebody else's
code, in my opinion.

Re: Some secure way of updating sources?

[Matthew Szudzik]

On Tue, May 18, 2010 at 03:07:59PM -0600, Eric S. Pulley wrote:
> So you are seriously suggesting the OpenBSD folks set up a public-key
> cryptography system (SSL) to confirm their current public-key cryptography
> system (SSH)? I guess then we would need a third system to confirm the
> first two, right? This must be troll food.

Perhaps the SSH fingerprints at

 http://www.openbsd.org/anoncvs.html

should be included on the official CD's.

Re: openbsd not blob free?

[Matthew Szudzik]

On Wed, May 05, 2010 at 11:32:47AM -0400, Kent Watsen wrote:
> it's pointed out that OpenBSD source tree has a blob in it:

You've just walked into an old flame war.  See
 http://marc.info/?l=openbsd-misc&m=119983946724267

Re: /usr directory: a system or user place?

[Matthew Szudzik]

On Sat, May 01, 2010 at 10:52:54PM +0200, Harrell wrote:
> So my doubt is: Is "usr" an abbreviation of "user"? If that is so (as as

Chapter 4 of Greg Lehey's Porting Unix Software

 http://www.lemis.com/grog/Documentation/PUS/

has the following to say about the /usr directory:

 This directory used to be the other file system on a UNIX machine.  In
 the System V ABI it has lost most of its importance.  The ABI states
 uses only for /usr/bin and /usr/share, and the name /usr has lost its
 original meaning: the ABI specifies /usr only as a location for system
 files that users may wish to access.

Re: Issues with Newer Broadcom NICs on OpenBSD 4.6

[Matthew Szudzik]

On Mon, Mar 29, 2010 at 08:46:15AM -0400, David Goldsmith wrote:
> eth3: Broadcom NetXtreme II BCM5709 1000Base-SX (C0) PCI Express found
> at mem ea00, IRQ 66, node addr a4badb236d41
> 
> eth4: Broadcom NetXtreme II BCM57711 XGb (A0) PCI-E x8 2.5GHz found at
> mem ec80, IRQ 66, node addr a4badb236cdb

Incidentally, have you seen this:

 http://www.ssi.gouv.fr/site_article185.html

If not configured properly, Broadcom NetXtreme NICs are vulnerable to
remote attacks that the operating system cannot prevent, because the
bugs are in the hardware.

Re: questions about OpenBSD 4.7

[Matthew Szudzik]

On Tue, Mar 16, 2010 at 06:06:57PM +0200, Andreas Gerdd wrote:
> -Why is the stock Apache not getting a version update?

See the Apache section of
 http://www.openbsd.org/policy.html

Re: Hardening OpenBSD : Just delete!

[Matthew Szudzik]

There are different definitions of "security".  When OpenBSD is
advertised as being secure, it means that users do not have access to
root.

In contrast, this person appears to equate security with total
authoritarian control of what users can do.  My favorite quote from

 http://geodsoft.com/howto/harden/OpenBSD/remove_files.htm

is

 "Telnet, lynx and cvs are removed. All are dangerous... lynx will let
 anyone who can get onto a system bring in any software they want."

So, he believes that an operating system is insecure if users can access
the internet.

Incidentally, this isn't the first time that someone has confused
OpenBSD's security claims with the ability to "lock down" users.  See

 http://marc.info/?t=12641295802

Re: Is OpenBSD + PF accredited or certified in any way ?

[Matthew Szudzik]

On Mon, Feb 01, 2010 at 11:06:12PM +, Keith wrote:
> firewall that I've never used before so am desperate to find some
> details of any foreign governments that are using OpenBSD / PF as a
> firewall or any details of any certification of the PF firewall.

Did you see the "Governments" section of

 http://www.openbsd.org/users.html

?

Re: way to help: laptops and weekly

[Matthew Szudzik]

On Mon, Feb 01, 2010 at 12:24:15AM +0100, Ingo Schwarze wrote:
> It would be bad advice to disable the daily/weekly/monthly jobs
> on all non-server systems.  I mean, when the machine is switched off,
> they do no harm.  When the machine is occasionally running during the
> night, they will at least run from time to time, which usually won't
> do any harm either and might even be useful.

You must have missed my earlier message in this thread

 http://marc.info/?l=openbsd-misc&m=126436422607058

I argue that the daily/weekly scripts can be harmful on laptops.  My
laptop never runs during the middle of the night, except when I'm doing
important mission-critical work.  I have had two incidents where the
daily/weekly scripts have interfered with that work (by removing files
from /tmp and by causing a buffer underflow while CD-burning).

In my case, it is best to disable the cronjobs for the daily/weekly
scripts and to run them manually instead.

Re: way to help: laptops and weekly

[Matthew Szudzik]

On Sun, Jan 24, 2010 at 02:48:35PM -0500, Ted Unangst wrote:
> Cron runs the weekly update script every Saturday at 3:30am.  If you
> use a laptop or other desktop, your computer probably isn't on then.
> So the locate and whatis databases never get updated unless you run it
> by hand.

I run the daily and weekly scripts by hand on my laptop, and I think
that that's the way it should be.

If your laptop is on at 3:30am, then you're probably doing something
important and don't want to be bothered by the daily or weekly security
scripts.  For example, if you've placed files in /tmp then the daily
script might erase them (and you'll be left wondering "where did my
files go?").  I've also had an incident where the security script ran
while I was burning a CD and it caused a buffer underflow.

For that reason, I've disabled the cronjobs.  I run /etc/weekly every
time I install new software (to update the whatis database).  I run
/etc/daily once a week when I backup my hard drive.

Re: writing to usb very slow

[Matthew Szudzik]

On Tue, Jan 19, 2010 at 10:10:41PM +0100, T. Tofus von Blisstein wrote:
> this is an example. Attached is a 1GB (fat!) usb memory stick. It took
> 40 minutes to copy 285M.

As I mentioned previously, the performance of USB memory sticks seems to
vary greatly on OpenBSD depending on the stick's manufacturer.  A stick
that writes quickly on Windows might be terribly slow on OpenBSD.  In
that case, just buy a different memory stick.

I have found the Kingston DataTraveler to write quite quickly on
OpenBSD.
 http://www.kingston.com/flash/datatraveler.asp

Re: writing to usb very slow

[Matthew Szudzik]

On Sun, Jan 17, 2010 at 06:41:38PM +0100, Antoine Jacoutot wrote:
> In my limited experience, USB HD are usually about alright but USB 
> sticks are super slow (i.e. EHCI speed is slower than usb1.1).

My experience is that only certain USB sticks are slow on OpenBSD.
Experiment with a few different brands and models until you find one
that writes at a reasonable speed.

Re: ksh (IEEE Std 1003.1-2008 ?)

[Matthew Szudzik]

On Sat, Jan 09, 2010 at 12:43:05AM -0800, J.C. Roberts wrote:
> If I use redirection rather than piping for input to the `while` loop,
> it is not a separate process, and works like the 'for' loop.

I have seen some Unix systems where this doesn't work, either.  See

 
http://groups.google.com/group/comp.unix.programmer/browse_thread/thread/20872c6fe79dd153/f3a4f184d0c6a718

Re: testing a drive with dd -- odd results

[Matthew Szudzik]

On Thu, Dec 31, 2009 at 09:03:30AM -0500, Scott McEachern wrote:
> dd: /dev/rwd0c: Input/output error
> 268435455+0 records in
> 268435455+0 records out
> 137438952960 bytes transferred in 23740.766 secs (5789154 bytes/sec)

Note that 137438952960 + 512 = 2^37.  So, dd is failing while reading
byte number 2^37.

Re: OpenBSD sed vs GNU sed

[Matthew Szudzik]

On Fri, Dec 25, 2009 at 12:59:29PM -0500, Hugo Villeneuve wrote:
> I had been using gnu sed for a while and encountered this issue
> going with OpenBSD.
...
> It's a cute work around. My point is: if a label can't contain
> semicolons, why support "tlabel;" and not "t;"? I don't understand
> the logic of it.

Keep in mind that GNU sed is not the standard version of sed, it
contains many nonstandard features.  The standard version of sed is
specified by POSIX.  The places where OpenBSD sed deviates from the
POSIX standard are documented at

 
http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/sed/POSIX?rev=1.2;content-type=text%2Fplain

In particular, that document says that

 "Historic versions of sed permitted commands to be separated
 by semi-colons, e.g. 'sed -ne '1p;2p;3q' printed the first
 three lines of a file.  This is not specified by POSIX.
 Note, the ; command separator is not allowed for the commands
 a, c, i, w, r, :, b, t, # and at the end of a w flag in the s
 command.  This implementation follows historic practice and
 implements the ; separator."

Basically, the semicolons are a nonstandard historic feature.  If you
want your sed scripts to be compatible with all Unix systems, then stick
to the POSIX standard and use newlines or -e to separate commands.

Re: OpenBSD sed vs GNU sed

[Matthew Szudzik]

On Mon, Dec 21, 2009 at 08:37:06PM -0500, Hugo Villeneuve wrote:
> OpenBSD's sed technicaly supports command concatation with ";" on
> the command line but it breaks "branch" without label. Which correctly

Read the caveat at the end of the sed man page:

 "The use of semicolons to separate multiple commands is not permitted
 for the following commands: a, b, c, i, r, t, w, :, and #."

Either use newlines instead of semicolons, or give each line of the
script separately with the -e option.  For example, instead of

 sed 's/a/b/ ; t ; s/c/d/'

use

 sed -e 's/a/b/' -e 't' -e 's/c/d/'

Re: vi in /bin

[Matthew Szudzik]

On Fri, Dec 18, 2009 at 03:27:07PM -0800, Randal L. Schwartz wrote:
> Everything I used to know about sed, I've forgotten once learning Perl.

It seems that every decade has its own fashionable Unix scripting
language.  The original scripting language, used in the 1970's was sed.
But awk became the fashionable language in the 1980's, then perl in the
1990's, and now python in the 2000's.  Who knows what will be next?

If you have time to spare, it certainly could be useful to learn all of
these languages.  But if you're going to learn just one of them, then I
vote for sed.  You will still encounter environments where sed is the
only one of these languages available (as Ingo pointed out, the OpenBSD
installer is one such example).

It isn't really necessary to keep up with fashion, you can just stick
with the historical standards.

Re: vi in /bin

[Matthew Szudzik]

On Fri, Dec 18, 2009 at 11:30:13AM -0600, Chris Bennett wrote:
> I would like to learn to use sed, however, I did not find that the
> man page was sufficient as a tutorial. I was not able to find any

I learned sed from the book Sed & Awk by Dougherty and Robbins.

 http://amazon.com/dp/1565922255

I highly recommend it.  Of the three utilities ed, sed, and vi, sed is
probably the most challenging because it doesn't have an interactive
mode.  With ed or vi you can always do something interactively if you
don't know the relevant command, but sed forces you to understand the
commands if you want to get anything done.

Learning sed will make you a better user of ed and vi.

Re: vi in /bin

[Matthew Szudzik]

On Thu, Dec 17, 2009 at 07:47:18PM -0500, Brad Tilley wrote:
> I use ed in emergencies when /usr is inaccessible, but I'm a lot more
> comfortable with vi. Will a static vi ever live in /bin? Helping someone
> use ed remotely, who has never used ed, when I myself don't use it
> regularly is always an adventure.

If using ed is an "adventure", then you don't really know how to use vi.

ed, sed, and vi are three of the most important Unix utilities, and
there's no excuse for not learning all three.  That's because they all
use the same commands and syntax.  If you know how to use one of them,
then you know how to use the other two.

Each is useful in a slightly different situation.  vi is for interactive
editing.  ed and sed are most useful for writing editing scripts.  ed
reads a file into a buffer for editing and can write those edits back to
the file.  sed works as a filter, reading a stream from standard input
line-by-line and editing each line before writing it to standard output.
Short, one-line sed scripts are useful in command line pipes for
translating the output of one program into the input of another program.

Although most of the basic commands are the same in ed, sed, and vi,
there are subtle differences.  After you have spent some time using ed
and sed, I suggest that you ask yourself the follow questions to test
your understanding of those differences: "When is the global command
(that is, g) useful in ed and vi?  Why is the global command missing
from sed?"

Re: OT: Have you hugged your local OpenBSD dev lately?

[Matthew Szudzik]

On Mon, Dec 14, 2009 at 05:03:40PM +0100, Marc Espie wrote:
> Considering the design of X, I don't expect any valid security model to emerge
> out of it.

The "Competitors to X" section of the X11 Wikipedia page has some
interesting comments about alternatives to X

 http://en.wikipedia.org/wiki/X_Window_System#Competitors_to_X

Unfortunately, none of them are close to becoming a reality in the near
future.

004_openssl.patch

[Matthew Szudzik]

A few days ago a patch to the highly-publicized OpenSSL vulnerability

 
http://www.pcworld.com/businesscenter/article/182304/ssl_flaw_could_have_been_used_to_hack_twitter.html

was released for OpenBSD-stable at

 ftp://ftp.openbsd.org/pub/OpenBSD/patches/4.6/common/004_openssl.patch

What exactly is the effect of this patch?  

I assume that it prevents renegotiation man-in-the-middle attacks
against webservers running OpenBSD, but does it also help OpenBSD
clients?  For example, suppose that I'm doing some online banking with
lynx, and somebody attempts a man-in-the-middle attack on my https
connection.  What will happen?

Re: Make utility documentation

[Matthew Szudzik]

On Fri, Nov 27, 2009 at 10:58:22AM +1300, Paul M wrote:
> It deals mainly with gnu make, but the principles are the same.

No.  The book Managing Projects With GNU Make
 http://www.amazon.com/dp/0596006101
by Robert Mecklenburg and published by O'Reilly in 2004 deals mainly
with GNU make, and isn't very useful for OpenBSD users.

But the book Managing Projects With Make
 http://www.amazon.com/dp/0937175900
by Andrew Oram and Steve Talbott, and last published by O'Reilly in
1993, deals with System V Release 4 make.  The authors of that book
state that they "have tried to be careful to include caveats for
features that might not have made it on to some systems--particularly
the Berkeley Standard Distribution (BSD)".  They also outline
differences with GNU make (and other versions) in the appendices.

Managing Projects With Make was one of the earliest Unix books published
by O'Reilly (maybe the first O'Reilly Unix book?), and it's probably
still the best introduction to generic make.  It was even referenced on
the man page for make in HP-UX.
 http://docs.hp.com/en/B3921-90010/make.1.html

Re: Error message from a distant site

[Matthew Szudzik]

On Sat, Nov 21, 2009 at 12:18:14PM +1100, Rod Whitworth wrote:
> I should have added that Google turned up only one hit for "ckbcintr"

Try "pckbcintr"

Minor typo OpenBSD FAQ

[Matthew Szudzik]

There is a minor typographical error in the OpenBSD FAQ at

 http://www.openbsd.org/faq/faq8.html#TimeZone

The FAQ says

 By default, OpenBSD assumes your hardware clock set to UTC (Universal

But the word "is" is missing.  It should be

 By default, OpenBSD assumes your hardware clock is set to UTC (Universal

Installing Mathematica 7 on OpenBSD

[Matthew Szudzik]

Back in March, I posted instructions for installing Mathematica 6 on
OpenBSD 4.5.
 http://marc.info/?l=openbsd-misc&m=123835793220311

Below are instructions for installing Mathematica 7 on OpenBSD 4.6.  The
biggest change is that Mathematica 7 no longer requires a
custom-compiled OpenBSD kernel.

This procedure assumes that you are running OpenBSD on the
i386 architecture, and that you have a licensed copy of Mathematica 7
For Linux.  I also assume that you have enabled compat_linux and have
installed the fedora_base package, as described at
 http://www.openbsd.org/faq/faq9.html#Interact

To install Mathematica, do the following:

 (1) Create the directory
  /tmp/mathematica
 and copy the Mathematica 7 For Linux installation files into that
 directory.

 (2) The MathInstaller file should now be located at 
  /tmp/mathematica/Unix/Installer/MathInstaller
 Type the following command to edit that file
  ed -s /tmp/mathematica/Unix/Installer/MathInstaller << 'EOF'
  1c
  #!/usr/local/emul/fedora/bin/sh
  .
  /DefaultManPageDir=/c
  DefaultManPageDir="/usr/local/man/man1"
  .
  g/`uname -s`/s/`uname -s`/Linux/
  g/chgrp -fhR/s/chgrp -fhR/chgrp -fR/
  g/chown -fhR/s/chown -fhR/chown -fR/
  g/if \[\( "${InstallerType}" != "\)App" -o\1gridMath" \]; then/c\
  if false; then\
  .
  wq
  EOF

 (3) Make sure that you have superuser privileges, and run the installer
 by typing
  /tmp/mathematica/Unix/Installer/MathInstaller
 When you are prompted for Mathematica's installation directory, enter
  /usr/local/emul/fedora/usr/local/Wolfram/Mathematica/7.0
 And when asked to create a new directory, type "y".  Accept the default
 options at all subsequent prompts.

 (4) Again, make sure that you have superuser privileges, and edit the
 Mathematica executables by typing the following commands.
  cd /usr/local/emul/fedora/usr/local/Wolfram/Mathematica/7.0/Executables
  for i in math mathematica mcc Mathematica MathKernel ; do
  ed -s $i << 'EOF'
  1c
  #!/usr/local/emul/fedora/bin/sh
  .
  g/`uname -s`/s/`uname -s`/Linux/
  g/OMP_NUM_THREADS=`cat \/proc\/cpuinfo/c\
  OMP_NUM_THREADS=`dmesg | grep '^cpu[0-9]* at ' | wc -l | tr -d ' '`\
  .
  g/ORIG_FONTCONFIG_FILE=/s/fonts.conf/conf.avail\/51-local.conf/
  wq
  EOF
  done

 (5) Mathematica's "JLink" requires Sun's Java binary, which does not
 function under Linux binary emulation on OpenBSD.  See
  http://archives.neohapsis.com/archives/openbsd/2006-11/1099.html
 To disable JLink, make sure that you have superuser privileges then
 type 
  chmod a-x 
/usr/local/emul/fedora/usr/local/Wolfram/Mathematica/7.0/SystemFiles/Links/JLink
 Note that various features of Mathematica, such as internet
 connectivity and the Mathematica Documentation Center depend on JLink.
 Fortunately, the Mathematica Documentation Center can also be accessed
 with a web browser at
  http://reference.wolfram.com

You're done!

Re: Encrypting /home on OpenBSD Laptops

[Matthew Szudzik]

On Sat, Nov 07, 2009 at 05:39:47PM -0600, Alvaro Mantilla Gimenez wrote:
> (check email, surf the net, etc...) but it is a pain in the ass the boot
> process going to the shell every time and need to type: bioctl -c C
> -l /dev/wd0d

 http://marc.info/?l=openbsd-misc&m=124187397614485

Re: Secure way to delete data in hard disc

[Matthew Szudzik]

On Tue, Oct 27, 2009 at 04:12:54PM +0100, Jordi Espasa Clofent wrote:
> to do it. Finally I've chosen a simple double-step method:
>
> First,
>
> $ dd if=/dev/urandom of=
>
> and next
>
> $ dd if=/deb/zero of=

I overwrite the disk 7 times with arandom, using the following command

 for x in `jot -s ' ' 7`; do dd if=/dev/arandom of=/dev/rDEVc bs=BLOCKSb ; done

where DEV is the abbreviated disk name (for example, sd2), and where
BLOCKS is the number of blocks to buffer.  The speed of the operation is
highly dependent on a proper choice of BLOCKS.  I have used 128 with a
160GB external hard drive, and it takes approximately 16 hours to
complete all 7 overwrites.  You'll have to experiment to see what choice
of BLOCKS is fastest on your hardware.

I use arandom instead of urandom because it's slightly faster.

Re: Forum engine

[Matthew Szudzik]

On Mon, Oct 12, 2009 at 04:42:44PM +0200, Toni Mueller wrote:
> archive, whereas, when the forum operator changes his forum software,
> or shuts down the forum, all past content is simply gone (or as good
> as).

Not true.  Whenever I read an interesting forum post, I save the html
file to my hard drive for future reference.

Re: procfs in OpenBSD

[Matthew Szudzik]

On Fri, Sep 25, 2009 at 04:34:20PM +0200, frantisek holop wrote:
> to remove unused bit rotting stuff.  what's the deal with procfs?
> 
> if it's use is far from recommended, indeed rather forbidden,
> why is it left to rot?

It's necessary for compat_linux.  See the man page.

Re: FFS/disklabel compatibility among platforms

[Matthew Szudzik]

On Mon, Sep 21, 2009 at 11:49:33PM +0200, David Vasek wrote:
>> Best way I know of to back up a lot of odd machines is to one store is  
>> running your favorite backup application so it dumps its data over ssh 
>> link to the one machine (which you have a spare of for repair purposes) 
>> which has the backup media attached to it.
>
> I agree, this is approximately what I am doing, but sometimes a directly  
> connected USB-attached disk comes handy.

You can use a combination of pax and dd to write directly to a
USB-attached disk in a platform-independent manner.  For example, to
back up the contents of the directory DIR to the drive sd2, use

 pax -w -x cpio DIR | dd of=/dev/rsd2c bs=32b

The speed of this operation will depend on the blocksize (I used 32 in
this example).  I suggest doing some experiments to see what blocksize
is the fastest for your hardware.

Re: Q: How to shop for a laptop to run OpenBSD?

[Matthew Szudzik]

On Mon, Aug 10, 2009 at 10:46:14AM +0800, Jeffrey 'jf' Lim wrote:
> better to go with AMD instead then in order to avoid this crap?

AMD has their own version of Intel AMT, included in their 780V chipset.

 http://ixbtlabs.com/articles3/mainboard/amd-780-740-chipsets-p2.html

But AMD doesn't seem to have created a marketing name for their AMT-like
technology.  Instead, they just refer to the technical standards that it
supports: IPMI, DMTF DASH, and DMTF ASF.

Re: Q: How to shop for a laptop to run OpenBSD?

[Matthew Szudzik]

On Mon, Aug 10, 2009 at 10:46:14AM +0800, Jeffrey 'jf' Lim wrote:
> is there a list, some kind of a guide, for which processors have this? Or is

Intel AMT isn't part of the processor, it's part of the chipset.  For
example, an Intel graphics controller is needed because AMT hides inside
the graphics memory.  It's not clear to me what other components it
requires, but if the computer is advertised as having "vPro", then it
has AMT.

Re: Q: How to shop for a laptop to run OpenBSD?

[Matthew Szudzik]

On Sun, Aug 09, 2009 at 11:09:49PM +0100, Edd Barrett wrote:
> mode so well. Remember if you do want to change the wifi card in a
> thinkpad, use tpwireless to unlock the bios check.

But does tpwireless work on recent ThinkPads?  According to the list of
successful BIOS modifications at

 
http://www.thinkwiki.org/wiki/Problem_with_unauthorized_MiniPCI_network_card#Successful_BIOS_Modifications

this has only worked for the X40's (or T42's or R51's) and earlier
models.  The one X60 that they tried was bricked and the modifications
failed for the W500 they tried.

Re: Q: How to shop for a laptop to run OpenBSD?

[Matthew Szudzik]

On Sun, Aug 09, 2009 at 07:39:19PM +0100, Edd Barrett wrote:
> Get a thinkpad, and replace the wireless card :)

Maybe I'm paranoid, but I've been reluctant to get a new ThinkPad
because they all have Intel AMT nowadays.

 http://marc.info/?l=openbsd-misc&m=118302016430106
 
http://software.intel.com/en-us/articles/architecture-guide-intel-active-management-technology/

And according to the following forum post, there's no way to disable it

 http://forum.thinkpads.com/viewtopic.php?f=25&t=62992

Re: wpi and firmware error

[Matthew Szudzik]

On Tue, Aug 04, 2009 at 11:13:04AM -0400, Luis Useche wrote:
> I downgraded the firmware to version 3.1 and so far the problem is gone.

Whether I'm using firmware version 3.1 with OpenBSD 4.5 (or 2.14.1.5
with OpenBSD 4.3 and 4.4), I get the firmware errors when I'm connected
to the wireless access points at my university.  But with the access
points that I've tried at home and at airports, wpi appears to work
fine.

For me, the errors appear to be specific to certain access points, but
independent of the firmware version.

Re: wpi and firmware error

[Matthew Szudzik]

On Sun, Aug 02, 2009 at 09:03:37PM -0400, Luis Useche wrote:
> From time to time my network card stop working and a error message appears:
> 
> wpi0: fatal firmware error

I've been experiencing the same problem for about a year--ever since my
university installed new access points.  The old access points worked
fine with wpi, but the new access points cause frequent firmware errors.
Some days I just give up on wireless and connect an ethernet cable...

Re: 4.6: softraid boot problems

[Matthew Szudzik]

On Thu, Jul 23, 2009 at 09:59:56PM +0200, Frank Brodbeck wrote:
> I just started to have a look at 4.6 (new installer rocks) and used the
> opportunity to setup encryption for my /home.
> 
> Everything works like a charm except that sd2 (on which my /home lives)
> doesn't get attached at boot time. This causes /etc/rc to yell at me

See
 http://marc.info/?l=openbsd-misc&m=124187397614485

Re: Activation of TrueType hinting in 4.5

[Matthew Szudzik]

On Wed, Jun 10, 2009 at 09:12:17AM +0200, Federico Giannici wrote:
> Now I tried to upgraded my desktop PC to OpenBSD 4.5, but I wasn't able  
> to activate the TrueType hinting.

Did you remove 31-nonmst.conf?  See

 http://marc.info/?l=openbsd-misc&m=124141462930332

Re: RAM not detected on HP DL580 G4

[Matthew Szudzik]

On Fri, May 29, 2009 at 02:33:48PM +0200, BARDOU Pierre wrote:
> I'm trying to set up an OpenBSD 4.5 amd64 on a HP ProLiant DL580 G4. It has 38
> GB RAM, but only ~3 GB is detected.
> Is it possible to use all the RAM ?

See

 http://marc.info/?l=openbsd-misc&m=123763424431042

Re: Disk enryption or storing data in safe

[Matthew Szudzik]

On Sat, May 09, 2009 at 04:38:41AM +0300, Cem Kayali wrote:
> Well, could you please advise me how i can automate to run bioctl  
> *before* fstab entires mounted? --- so that i can push ie; /var, /usr  
> and swap partitions into encrypted volume?

Assuming that your RAID partition is sd0d, use the following command to
edit /etc/rc

 ed -s /etc/rc
 /^# Configure raid/a
 until bioctl -c C -l /dev/sd0d softraid0; do done
 .
 wq

Then add /var, /usr, etc. to your /etc/fstab file.

Re: Trouble with "Times" fonts in OpenBSD 4.5

[Matthew Szudzik]

On Mon, May 04, 2009 at 03:47:41AM +, Matthew Szudzik wrote:
> The problem is that Bitstream Vera Serif is always being used instead of
> any variant of "Times".  For example if I set firefox or gimp to use

After a little investigation, I have discovered that this problem is
caused by the file

 /etc/fonts/conf.avail/31-nonmst.conf

that was added to xenocara in February.  See

 
http://www.openbsd.org/cgi-bin/cvsweb/xenocara/lib/fontconfig/conf.d/31-nonmst.conf?rev=1.1

Basically, it's a hack for people who prefer the Bitstream Vera fonts.
It can be disabled by removing the symbolic link at

 /etc/fonts/conf.d/31-nonmst.conf

Trouble with "Times" fonts in OpenBSD 4.5

[Matthew Szudzik]

I have just installed OpenBSD 4.5 on i386, and I am observing a problem
with fonts.  It is most noticeable with the /x11/msttcorefonts port
installed.

The problem is that Bitstream Vera Serif is always being used instead of
any variant of "Times".  For example if I set firefox or gimp to use
Times (one of the standard X11 fonts) or Times New Roman (from
msttcorefonts), then Bitstream Vera Serif is used instead.  But the
other msttcorefonts (for example, Courier New, Verdana, etc.) work fine.

I suspect that it is a problem with fontconfig.  Note that fontconfig
correctly matches the string 'Courier New' to the appropriate font

 $ fc-match :family='Courier New'
 cour.ttf: "Courier New" "Normal"

But it does not match the string 'Times New Roman'

 $ fc-match :family='Times New Roman'
 VeraSe.ttf: "Bitstream Vera Serif" "Roman"

even though the font is installed

 $ fc-list :family='Times New Roman' file
 /usr/local/lib/X11/fonts/mscorefonts/times.ttf: 
 /usr/local/lib/X11/fonts/mscorefonts/timesbd.ttf: 
 /usr/local/lib/X11/fonts/mscorefonts/timesbi.ttf: 
 /usr/local/lib/X11/fonts/mscorefonts/timesi.ttf: 

For some reason, fontconfig is giving Bitstream Vera Serif higher
priority than Times New Roman when it tries to match a string.  Here is
the prioritized list of matches:

 $ fc-match -s :family='Times New Roman'
 VeraSe.ttf: "Bitstream Vera Serif" "Roman"
 times.ttf: "Times New Roman" "Normal"
 luxirr.ttf: "Luxi Serif" "Regular"
 l049013t.pfa: "Luxi Serif" "Regular"
 n021004l.pfb: "Nimbus Roman No9 L" "Medium"
 timR12-ISO8859-1.pcf.gz: "Times" "Regular"
 timR12.pcf.gz: "Times" "Regular"
 GohaTibebZemen.ttf: "Goha-Tibeb Zemen" "Regular"
 cour.ttf: "Courier New" "Normal"
 arial.ttf: "Arial" "Normal"
 ariblk.ttf: "Arial Black" "Normal"
 SyrCOMNisibin.otf: "Estrangelo Nisibin" "Regular"
 trebuc.ttf: "Trebuchet MS" "Normal"
 SyrCOMQenNeshrin.otf: "Estrangelo Quenneshrin" "Regular"
 SyrCOMMalankara.otf: "Serto Malankara" "Regular"
 SyrCOMCtesiphon.otf: "East Syriac Ctesiphon" "Regular"
 d05l.pfb: "Dingbats" "Regular"
 cour.pfa: "Courier" "Regular"
 s05l.pfb: "Standard Symbols L" "Regular"
 12x13ja.pcf.gz: "Fixed" "ja"
 lutRS12.pcf.gz: "LucidaTypewriter" "Sans"
 6x13-ISO8859-1.pcf.gz: "Fixed" "SemiCondensed"
 6x12.pcf.gz: "Fixed" "SemiCondensed"
 6x13.pcf.gz: "Fixed" "SemiCondensed"
 7x13.pcf.gz: "Fixed" "Regular"
 8x13.pcf.gz: "Fixed" "Regular"
 term14.pcf.gz: "Terminal" "Regular"
 9x15.pcf.gz: "Fixed" "Regular"
 cu-pua12.pcf.gz: "ClearlyU PUA" "Regular"
 cu12.pcf.gz: "ClearlyU" "Regular"
 cu-alt12.pcf.gz: "ClearlyU Alternate Glyphs" "Regular"
 18x18ko.pcf.gz: "Fixed" "ko"
 10x20.pcf.gz: "Fixed" "Regular"
 
Does anybody have a workaround for this problem? 

Re: Sharing softraid volumes between machines

[Matthew Szudzik]

On Thu, Apr 30, 2009 at 08:30:25AM -0500, Marco Peereboom wrote:
> This should work but there are a few caveats.  It has to be same arch

That's the problem.  Machine A is i386 and machine B is macppc.

Sharing softraid volumes between machines

[Matthew Szudzik]

Can a softraid volume be shared between two different computers?  Here's
my scenario:

I have two OpenBSD machines, A and B.  Both are running OpenBSD 4.4.
Machine A is my primary computer, and I regularly back it up on an
external hard drive, encrypted with softraid.  My plan is to copy the
backup to machine B in the event that machine A fails.  But this doesn't
seem to work.

I can reliably read/write to the encrypted hard drive on machine A, but
when I attach the drive to machine B and type

 bioctl -c C -l /dev/sd0a softraid0

I am asked to type my passphrase twice, and then when the encrypted
volume is attached, it is completely empty.

Can anyone confirm this behavior?  Am I doing something wrong?

Re: automaticaly mount/umount encrypted $HOME or ...

[Matthew Szudzik]

On Tue, Apr 28, 2009 at 10:18:35PM +0300, Timo Myyr?? wrote:
> I encrypted my $HOME with bioctl and just put the 'bioctl -c C -l  
> /dev/sd0g softraid0' line to my /etc/rc. Simple and working solution 
> although it needs a little bit tweaking as currently I get dropped to 
> single user mode if I misstype my passphrase.  This happens quite easily 

Try

 until bioctl -c C -l /dev/sd0g softraid0; do done

Typo on OpenBSD 4.5 CD Set

[Matthew Szudzik]

There is a typographical error on the back cover of the OpenBSD 4.5 CD
set.  It says

 "This package contains the operating system and a selection of
 applications for the i386, macppc, amd64, sparc, and sparc64
 architectures."

This is incorrect, because the CDs do not contain the operating system
or applications for the sparc architecture.

This error was also on the OpenBSD 4.4 CD set.  See
 http://marc.info/?l=openbsd-misc&m=122412019502133

Re: ThinkPad T60 audible bell *very* loud

[Matthew Szudzik]

> feel.  The stupid bell should be killed dead.

Actually, I prefer the bell over the more complex "alert sounds"
produced by other operating systems--to me, anything other than a simple
bell would be needless bloat.

(I'm reminded of the fact that every time I try to shutdown a Windows
machine, I have to wait several seconds for it to finish playing the
shutdown sound.  What a waste of my time!)

The bell can also be useful in shell scripts--just echo Ctrl-G to alert
the user.

Re: Yahoo videos on OpenBSD

[Matthew Szudzik]

On Mon, Apr 20, 2009 at 07:46:42PM +0200, Tom Bodr wrote:
> someone is watching Yahoo videos on OpenBSD?

I use
 http://keepvid.com
to download the video, then play it with mplayer.

Re: [semi-OT] Can anyone recommend an OpenBSD-compatible colour laser printer?

[Matthew Szudzik]

On Sat, Apr 11, 2009 at 02:20:25PM +0200, ropers wrote:
> went on behind the scenes. I am trying to catch up on that, though I'm
> finding that comprehensive, thorough, and up-to-date documentation on
> the whole Unix printing landscape seems to be somewhat hard to find.

There is a chapter on printing in the UNIX Power Tools book:
 http://www.openbsd.org/books.html#3

Personally, I wish something simple like pdq
 http://pdq.sourceforge.net
would become the new standard for UNIX printing.  Unfortunately, the pdq
project has been dead for years.  :(

Re: Games

[Matthew Szudzik]

On Thu, Apr 09, 2009 at 02:34:20PM +0300, Aapo Lehtinen wrote:
> Very good song! Is there "original" available somewhere? I presume it is  
> somehow connected to this "Tron" but i'm unfamiliar with it. (..oh, i'll

The song "Only solutions" by Journey was featured in the movie Tron.
 http://www.youtube.com/watch?v=ZVlB6KJT2xo
"Games" is very similar in style to "Only solutions", by the lyrics and
melody are quite different.

Re: Games

[Matthew Szudzik]

On Wed, Apr 08, 2009 at 09:19:00PM +0200, Matthias Kilian wrote:
> The new release song is really catchy. Many thanks to Jonathan,

I'm in complete agreement.  It's probably the best OpenBSD song yet, and
has the potential to appeal to frustrated computer users outside the
OpenBSD community (e.g. the slashdot crowd) with lyrics like "I love to
hate my PC", "Just wanna get this job done", and "Lost my mind, it's
such a waste of time".

Installing Mathematica 6 on OpenBSD

[Matthew Szudzik]

On Wed, Mar 25, 2009 at 11:56:45AM +, Matthew Szudzik wrote:
> Incidentally, I'm running Mathematica on i386 using compat_linux.  There
> are a few non-trivial steps involved in the installation--contact me if
> you want the details.

I've received a few requests for the Mathematica installation procedure
on OpenBSD, so I thought that I'd take a moment to post it on this
mailing list.  I have tested this procedure with Mathematica 6 on
OpenBSD 4.4.  I plan to experiment with Mathematica 7 on OpenBSD 4.5 in
May, and will post my results.

This procedure assumes that you are running OpenBSD 4.2 or later on the
i386 architecture, and that you have a licensed copy of Mathematica 6
For Linux.  I also assume that you have enabled compat_linux and have
installed the fedora_base package, as described at
 http://www.openbsd.org/faq/faq9.html#Interact

To install Mathematica, do the following:

 (1) Apply the patch at
  http://marc.info/?l=openbsd-misc&m=119479722118605
 to the OpenBSD kernel.  Instructions for downloading the kernel source
 code and building the patched kernel are available at
  http://www.openbsd.org/faq/faq5.html#BldGetSrc
  http://www.openbsd.org/faq/faq5.html#BldKernel

 (2) Create the directory
  /tmp/mathematica
 and copy the Mathematica 6 For Linux installation files into that
 directory.

 (3) The MathInstaller file should now be located at 
  /tmp/mathematica/Unix/Installer/MathInstaller
 Type the following command to edit that file
  ed -s /tmp/mathematica/Unix/Installer/MathInstaller << EOF
  1c
  #!/usr/local/emul/fedora/bin/sh
  .
  /DefaultManPageDir=/c
  DefaultManPageDir="/usr/local/man/man1"
  .
  g/Linux)/s/Linux)/OpenBSD)/
  g/chgrp -fhR/s/chgrp -fhR/chgrp -fR/
  g/chown -fhR/s/chown -fhR/chown -fR/
  wq
  EOF

 (4) Make sure that you have superuser privileges, and run the installer
 by typing
  /tmp/mathematica/Unix/Installer/MathInstaller
 When you are prompted for Mathematica's installation directory, enter
  /usr/local/emul/fedora/usr/local/Wolfram/Mathematica/6.0
 And when asked to create a new directory, type "y".  Accept the default
 options at all subsequent prompts.

 (5) Again, make sure that you have superuser privileges, and edit the
 Mathematica executables by typing the following command.
  for i in math mathematica mcc Mathematica MathKernel
  do
  ed -s /usr/local/emul/fedora/usr/local/Wolfram/Mathematica/*/Executables/$i 
<< EOF
  1c
  #!/usr/local/emul/fedora/bin/sh
  .
  g/Linux)/s/Linux)/OpenBSD)/
  g/cat .proc.cpuinfo | grep processor/s/cat .proc.cpuinfo | grep 
processor/dmesg | grep '^cpu[0-9]* at '/
  wq
  EOF
  done

 (6) Mathematica's "Paclet" system requires Sun's Java binary, which
 does not function under Linux binary emulation on OpenBSD.  See
  http://archives.neohapsis.com/archives/openbsd/2006-11/1099.html
 To disable the Paclet system, remove or rename the file
  
/usr/local/emul/fedora/usr/local/Wolfram/Mathematica/*/Documentation/English/PacletInfo.bin

 (7) You are now ready to run Mathematica.  The first time you start
 Mathematica, you need to change some of the default configuration
 options.  In particular, on the menu at
  Edit > Preferences... > Internet Connectivity
 you should uncheck the box labeled "Allow Mathematica to access the
 Internet", since the Paclet system is required for that functionality.
 Furthermore, you need to configure the MathLink connection to use
 pipes rather than shared memory.  The procedure for doing this varies,
 depending on which version of Mathematica you are using.  On my
 version, I go to the
  Evaluation > Kernel Configuration Options...
 menu, edit the local kernel, click "Advanced Options" in the "Kernel
 Properties" menu, and change the word "SharedMemory" to "Pipes" in the
 "Arguments to MLOpen" box.

You're done!

Re: Ideas for Getting MATLAB/Mathematica to utilize sparc64 ram that runs openbsd

[Matthew Szudzik]

On Wed, Mar 25, 2009 at 09:10:39AM -0700, Vivek Ayer wrote:
> experiment. The Sun blade systems are all already web servers which
> have allocated 2 GB to memcached, but still have so much more to give.
> I thought maybe this was a useful way to use that extra RAM.

Another problem that you might encounter is that the maximum datasize
for a process is 1GB.  See

 http://marc.info/?l=openbsd-misc&m=120026365416264

Re: Ideas for Getting MATLAB/Mathematica to utilize sparc64 ram that runs openbsd

[Matthew Szudzik]

On Wed, Mar 25, 2009 at 03:30:56AM -0700, J.C. Roberts wrote:
> As for the actual question of getting Mathematica and/or Matlab to run
> on your sparc64 under OpenBSD, what about using SysV-R4 emulation with
> the available Solaris binaries?
> 
> At least with Mathematica, linux binaries are only available for x86
> (32 and 64 bit), but they do provide 64 bit UltraSPARC executables. 

OpenBSD's compat_svr4 and compat_linux binary emulations do not support
64-bit processors.  So, you'll have to find some other solution for
running Mathematica or Matlab binaries on your machine.

Incidentally, I'm running Mathematica on i386 using compat_linux.  There
are a few non-trivial steps involved in the installation--contact me if
you want the details.

Re: linux binary emulation (corrected kdump)

[Matthew Szudzik]

On Fri, Mar 13, 2009 at 05:40:46PM +, Matthew Szudzik wrote:
> homepage?  Other features like pf, carp, openssh, sparc64 support, etc.
> aren't mentioned at all on the front page of www.openbsd.org

Oops.  OpenSSH is mentioned on the front page, albeit in a footnote that
lists "associated projects".

Re: linux binary emulation (corrected kdump)

[Matthew Szudzik]

On Fri, Mar 13, 2009 at 12:55:18PM -0400, Ted Unangst wrote:
> On Fri, Mar 13, 2009 at 8:41 AM, Constantine Cusulos
>  wrote:
> > Maybe compat_linux needs some additional work.
> 
> No doubt about that, but interest is seriously waning.  I cared back

Although interest in compat_linux may be "seriously waning" among
developers, I think it's still important for users such as myself.
compat_linux is the most reasonable option for users who need to run
commercial software on OpenBSD.

And compat_linux is important for attracting new users to OpenBSD.  Why
else would it be mentioned in the first paragraph of the OpenBSD
homepage?  Other features like pf, carp, openssh, sparc64 support, etc.
aren't mentioned at all on the front page of www.openbsd.org

Re: halt -p does not powerdown ThinkPad X200 under 4.5beta

[Matthew Szudzik]

On Tue, Mar 10, 2009 at 09:34:38AM -0500, Marco Peereboom wrote:
> What happens here is that the bios lies and tells us to write a value to
> some register and nothing happens.  We have debugged several of these
> machines and all to the same avail; we are doing it right but nothing
> happens.  Maybe it is an idea to complain to the vendor.

Has anyone run FreeBSD, Linux, or Windows on machines that have
experienced this bug?

If there's an open source operating system that doesn't have this bug,
then we can look at their code and see how they fixed it.  If all
operating systems have this bug, then there's no hope of fixing it.

Re: Can someone please suggest a replacement for xterm for me?

[Matthew Szudzik]

On Fri, Mar 06, 2009 at 11:59:49AM -0500, Marcus Watts wrote:
> Recent versions of xterm (237, 242) have some new options you might find 
> helpful.
> 
> control-center-button brings up "VT options", then
> just past "Allow 80/132 Column Switching", there are 2 new options:
> "Keep Selection", and "Select to Clipboard".  I think the 2nd of these
> might cause xterm to behave as you wish.

Excellent!  I was unaware of that option.  Adding the line

 XTerm*selectToClipboard: true

to the

 ~/.Xdefaults

file allows highlighting (and middle-clicking) to copy (and paste) with
the CLIPBOARD, rather than the PRIMARY.

Re: Can someone please suggest a replacement for xterm for me?

[Matthew Szudzik]

On Fri, Mar 06, 2009 at 02:16:05PM +0100, frantisek holop wrote:
> hmm, on Fri, Mar 06, 2009 at 11:45:49AM +0000, Matthew Szudzik said that
> > PRIMARY.  So, if you've copied something to the CLIPBOARD in firefox,
> > then you won't be able to paste it in xterm.
> 
> i disagree.
> 
> shift+insert

No, Shift-Insert does not work.  Suppose you've copied "String1" to the
CLIPBOARD in firefox.  That is, you've highlighted "String1" and pressed
Ctrl-C.  Then suppose that you highlight some other string "String2".
(For example, you may have gone to the firefox "Save Page As..." dialog
box, which automatically highlights the title of the current page.) Now,
if you go to xterm and press Shift-Insert, you do not get "String1"
which is in the CLIPBOARD, but "String2" which is in the PRIMARY.
That's the problem!

You can paste the PRIMARY to xterm, but you cannot paste the CLIPBOARD
to xterm.  (Unless you use xsel.)

Re: Can someone please suggest a replacement for xterm for me?

[Matthew Szudzik]

On Thu, Mar 05, 2009 at 11:43:58PM -0800, patrick keshishian wrote:
> cut-and-paste no. copy-and-paste yes. is that not good enough?

The main problem is that there are are 3 different copy-and-paste
operations under X Windows.  There is the PRIMARY (copies by
highlighting text and pastes by clicking the middle mouse button), the
SECONDARY, and the CLIPBOARD (copies by pressing Ctrl-C and pastes by
pressing Ctrl-V).

Most modern applications (like firefox, openoffice, etc.) can use both
the PRIMARY and CLIPBOARD.  The problem is that xterm can only use the
PRIMARY.  So, if you've copied something to the CLIPBOARD in firefox,
then you won't be able to paste it in xterm.

I solve this problem by using xsel

 http://openports.se/x11/xsel

Using my window manager, I've bound Mod4-C (Mod4 is the Windows keys on
most computers) to the command

 xsel | xsel -b

and I've bound Mod4-V to the command

 xsel -b | xsel

The effect is that Mod4-C copies the PRIMARY to the CLIPBOARD, and
Mod4-V copies the CLIPBOARD to the PRIMARY.  Therefore, I can paste from
firefox to xterm by pressing Mod4-V, then clicking the middle mouse
button.

Re: nroff -mandoc alternative

[Matthew Szudzik]

On Sat, Feb 28, 2009 at 11:01:49PM +0100, Kristaps D??onsons wrote:
> Hello, if anybody's interested in an alternative to groff for viewing
> BSD "mdoc" manual pages, I'm actively looking for patches and problem
> reports for mdocml.   From the site :

Excellent!  Was this project started in response to the discussion about
groff (in the context of pcc) on Undeadly last November?

 http://www.undeadly.org/cgi?action=article&sid=20081108135831

Re: Backup strategies

[Matthew Szudzik]

On Sun, Feb 01, 2009 at 06:34:31PM +0100, Pierre Riteau wrote:
> The 4GB file limitation is for files *INSIDE* an ISO file system, not
> for the ISO itself.

Exactly!  It's OK to have a DVD "crammed to the brim" with 4.7GB of
data, as long as the individual files on the DVD are each under the
filesize limit dictated by the operating system.  One big 4.7GB file
will probably exceed the limit, but a lot of little files whose TOTAL
SIZE is 4.7GB is OK.

Re: Backup strategies

[Matthew Szudzik]

On Sat, Jan 31, 2009 at 02:04:32PM +, Dieter wrote:
> > i kinda like cpio for fast backup of filesystems... for large media
> > files (think anime movies) -- I think its generally best to just
> > burn them on a iso..
> 
> ISO files have a 2 GB filesize limit, so large files don't fit.

I use
 pax -w -B 150k -x cpio
to avoid the 2GB filesize limit in ISO filesystems.  The command splits
the cpio archive into 1.5GB files--exactly three of these files will fit
on a standard DVD.

Re: Backup strategies

[Matthew Szudzik]

On Sun, Feb 01, 2009 at 12:12:50PM +0100, Toni Mueller wrote:
> I can fetch files that are well over 4GB and burn them on DVD. These
> files are called as "ISO" files, but I don't know exactly what's inside

See

 
http://en.wikipedia.org/wiki/ISO_9660#The_4_GiB_.28or_2_GiB_depending_on_implementation.29_file_size_limit

Some operating systems can handle files up to 4GB on an ISO 9660
filesystem, and other operating systems can handle more than 4GB.  But
if you want your ISO 9660 filesystem to be fully portable, you should
stick to the 2GB limit.

Re: Capturing kernel debugger output without serial console

[Matthew Szudzik]

On Wed, Dec 17, 2008 at 01:17:44PM +0200, Lars Nood??n wrote:
> What is a reasonable way to capture kernel debugger (ddb) output without
> a serial console?  I'm able to consistently get ifconfig to crash on the

Take a photograph of the screen.

Re: halt -p does not power off ThinkPad X61 under 4.4

[Matthew Szudzik]

> used halt -p to shutdown the machine and walked away. The next morning i 
> found that while it appears to have shutdown correctly, the machine did 
> not power off but instead showed
> 
> syncing disks...done
> uchi2: host controller halted

This bug has been supposedly fixed in OpenBSD -current. See
 http://marc.info/?l=openbsd-bugs&m=122788235218793

Re: halt -p does not power off ThinkPad X61 under 4.4

[Matthew Szudzik]

On Mon, Nov 24, 2008 at 04:17:07AM +, Matthew Szudzik wrote:
> Have you submitted a formal bug report (with a dmesg and debugger output
> from ps and trace)?  I myself have been attempting it, but every time
> that this bug occurs it seems that I'm in a hurry to get somewhere, and

I guess I spoke too soon.  The bug occurred immediately after I sent
that message, above.  I just completed a bug report.

Re: halt -p does not power off ThinkPad X61 under 4.4

[Matthew Szudzik]

On Mon, Nov 24, 2008 at 10:32:13AM +0800, David Schulz wrote:
> For the Record, I don't have a true Solution to the Issue yet, but i saw  
> a small flag in /etc/sysctl.conf that read machdep.apmhalt.  After  
> enabling that and rebooting, halt -p now powers off my Machine 9 times  
> out of ten. Not perfect, but a solution.

Have you submitted a formal bug report (with a dmesg and debugger output
from ps and trace)?  I myself have been attempting it, but every time
that this bug occurs it seems that I'm in a hurry to get somewhere, and
don't have time to carefully copy the outputs of ps and trace to paper.

Re: halt -p does not power off ThinkPad X61 under 4.4

[Matthew Szudzik]

> i installed OpenBSD 4.4 on my ThinkPad X61, played a little bit, and 
> used halt -p to shutdown the machine and walked away. The next morning i 
> found that while it appears to have shutdown correctly, the machine did 
> not power off but instead showed
> 
> syncing disks...done
> uchi2: host controller halted
> 
> and would just sit there until i manually held the power button. This 

Over the last 6 months I have experienced the same problem 2 or 3 times
with my ThinkPad T60 running OpenBSD 4.3.

Typo on OpenBSD 4.4 CD Set

[Matthew Szudzik]

I notice that there is a typographical error on the back cover of the
OpenBSD 4.4 CD set.

 http://bsdly.net/~peter/44bergen02.jpg

It says

 "This package contains the operating system and a selection of
 applications for the i386, macppc, amd64, sparc, and sparc64
 architectures."

This is incorrect, because the CDs do not contain the operating system
or applications for the sparc architecture.

Re: assembler noob question

[Matthew Szudzik]

On Thu, Jul 24, 2008 at 07:16:55PM -0600, Philip Guenther wrote:
> If you're just trying to learn the x86 instruction set, then why not
> put your code in an __asm__() block inside a C program?  That lets the
> compiler do all the heavy lifting.

Is there a man page that describes the __asm__ block?  The gcc man page
only describes how to turn it on or off, and since it's nonstandard C,
it's not described in any of the standard reference books like K&R.

Re: OpenBSD project goals

[Matthew Szudzik]

> And troff. :-)

The OpenBSD base install contains groff, not troff.  (groff is 3rd party
software maintained by Gnu.)

Re: OT: Dissertation ideas for my degree

[Matthew Szudzik]

> - A good TeX to html convertor (extensible)
> - A good TeX gui

There's a pretty good chance that TeX is going to become obsolete, and
replaced by some HTML or XML derivative.  Many technical publishers
have already made the transition.  See, for example, the following link
from Cambridge University Press

 
https://authornet.cambridge.org/information/productionguide/stm/XML_workflow.asp#xml_introduction

An interesting alternative project would be to create an HTML and MathML
GUI, with the intent of luring mathematicians and physicists away from
TeX.  And then create an HTML/MathML to TeX converter, so that they can
share their work with colleagues and journals that still insist on TeX.

Re: cwm keybindings misbehavior

[Matthew Szudzik]

On Sat, Jun 14, 2008 at 03:37:57PM +, Matthew Szudzik wrote:
> Note, I was told by one of the Fluxbox developers that I need to add the
> following line
> 
>  xmodmap -e 'add Mod4 = Super_L'
> 
> to my .xinitrc file if I want the Windows key to be well-behaved, but I
> don't understand the reason why.

I've done a little Googling, and apparently this is workaround for a bug
in the X.org keycodes.  See
 http://modeemi.fi/~tuomov/ion/faq/entries/Modifier_releases.html

Re: cwm keybindings misbehavior

[Matthew Szudzik]

On Sat, Jun 14, 2008 at 02:09:38PM +, Glenn Becker wrote:
> one interesting one i have found is that M- will not work on firefox 
> if there are tabs/multiple pages open.

I always configure my window managers to use the Windows key (i.e. Mod4)
rather than Control or Alt (i.e. Meta).  This prevents conflicts with
the applications that are being managed by the window manager--since
ordinary applications, like Firefox, don't use the Windows key.

Note, I was told by one of the Fluxbox developers that I need to add the
following line

 xmodmap -e 'add Mod4 = Super_L'

to my .xinitrc file if I want the Windows key to be well-behaved, but I
don't understand the reason why.

Re: Lastet supported jdk on OpenBSD

[Matthew Szudzik]

On Fri, May 16, 2008 at 09:03:17PM -0300, John Nietzsche wrote:
> i would like to add support for java on my 4.3 openbsd desktop. Has
> anybody already done so? May you point a url where i could download
> the package(s) from?

As the previous posters have pointed out, there are no JDK binary
packages available for OpenBSD 4.3--you have to fetch and build the JDK
from source yourself.  But in OpenBSD 4.4 (which will be released in
November), that situation will change, and binary packages for Java will
be available.  See
 http://www.undeadly.org/cgi?action=article&sid=20080321023803

Re: Editing C with...

[Matthew Szudzik]

I learned emacs before vi, but I could never remember all of emacs'
complicated keystrokes.  It feels as though there are far fewer commands
to memorize in vi, and the commands that I do memorize are also useful
for writing sed scripts--so the payoff of a little memorization in vi
is greater than the payoff in emacs.

And anyway, I'm a minimalist (that's why I run OpenBSD).  nvi is
fine--vim and emacs just have too much bloat.

Re: Is NV supposed to be SLOW?

[Matthew Szudzik]

On Sun, May 04, 2008 at 04:17:58PM -0300, Nenhum_de_Nos wrote:
> radeonhd has any 3D ?

3D acceleration is not currently supported on OpenBSD, but work is being
done to ensure that it will be supported in the future.  A recent
progress report, together with a description of the status of NVIDIA,
ATI, and Intel drivers, was posted at
 http://www.undeadly.org/cgi?action=article&sid=20080416195151&pid=44

Re: Nvidia Quadro NVS 140M

[Matthew Szudzik]

> It's so fucking hard to find a correct laptop where most everything will
> work correctly these days...

Generally, the best-supported laptops on OpenBSD are the Lenovo
ThinkPads.  Detailed specifications of the individual chipsets in
current ThinkPad models can be found at
 ftp://ftp.software.ibm.com/pc/pcinstitute/psref/tabook.pdf
Before purchasing a ThinkPad, it is prudent to compare the chipsets in
those specifications to the list of OpenBSD-supported chipsets at
 http://www.openbsd.org/i386.html
(or at http://www.openbsd.org/amd64.html if you plan to run 64-bit
OpenBSD).

The two most commonly-encountered incompatibilities are with the
wireless and with the graphics chipsets.  OpenBSD does not support the
Atheros wireless chipsets in current ThinkPad models--but the Intel
wireless chipsets (currently Intel PRO/Wireless 3945ABG, and Intel
Wireless WiFi Link 4965AG) are supported.  The graphics chipsets
supported by the intel driver
 http://www.openbsd.org/cgi-bin/man.cgi?query=intel
also work well with OpenBSD.  NVIDIA and ATI graphics chipsets have
suffered from various incompatibilities with OpenBSD (but it is hoped
that support for ATI will improve now that specifications for their
graphics chipsets are available to the open source developers who write
the drivers).

If you are overwhelmed by the number of different ThinkPad models, then
you should be aware that Lenovo designs some ThinkPads for compatibility
with SUSE Linux.  A link to a list of those Linux-compatible models is
available on the Lenovo notebook main page at
 http://shop.lenovo.com/us/notebooks/
Although compatibility with SUSE Linux is no guarantee of compatibility
with OpenBSD, the hardware-compatibility issues present in both
operating systems are similar enough that this list can be a useful
place to start looking for an appropriate model.

Re: Rolling release?

[Matthew Szudzik]

On Wed, Apr 23, 2008 at 03:03:13AM +0200, Zbigniew Baniewski wrote:
> Not sure about amount of time sacrificed each time to prepare new complete
> release... but perhaps it could be spared, if the system+packages is
> "refreshed" piece-by-piece / month-by-month?

Who would buy CD's if there were a "rolling release"?  CD sales are an
important source of funding.

Every time OpenBSD makes a release, it makes news on Slashdot and other
websites.  How would OpenBSD make up for the lost publicity with a
rolling release?

Rolling releases are arguably more buggy than semi-annual releases
because less time is spent ensuring their consistency.  Some users, such
as myself, prefer the relatively bug-free semi-annual releases.

Re: 4.3 mips version?

[Matthew Szudzik]

On Thu, Apr 17, 2008 at 03:20:00PM -0600, Theo de Raadt wrote:
> BTW, you don't mean mips.  You mean sgi.  mips is a cpu architecture, not
> a machine.

This is an issue that has always confused me.  It is obvious that the
binaries for OpenBSD must be different for each cpu architecture, but
why are they sometimes different for machines that have the same cpu
architecture?

For example, there are different binaries for OpenBSD on hp300 and
mvme68k machines, even though they both have a Motorola 680x0 cpu.
There are also different binaries for zaurus and armish machines, even
though they both have an ARM cpu.  Why?

Is it because these machines have limited memory, and there is only room
for those drivers that are needed by the specific machine, or is there
something more subtle going on?

Re: install42.iso hangs....any ideas?

[Matthew Szudzik]

On Wed, Apr 16, 2008 at 05:45:56AM -0700, Unix Fan wrote:
> I thought DVD-RAM were unique in that, unlike DVD-RW, you can write to the 
> disk as if it's simply an optical hard drive...

I have only used DVD+RW and CD media in my DVDRAM drive, so I cannot
answer your question.

Re: install42.iso hangs....any ideas?

[Matthew Szudzik]

On Tue, Apr 15, 2008 at 07:33:41PM -0700, Unix Fan wrote:
> How well does OpenBSD support DVD-RAM drives? does the cd(4) driver support 
> read/write operations? - i.e: Would it be possible to use it as a normal 
> block device?

I have successfully read and written several DVDs and CDs using OpenBSD,
following the instructions at
 http://www.openbsd.org/faq/faq13.html

Re: install42.iso hangs....any ideas?

[Matthew Szudzik]

On Tue, Apr 15, 2008 at 10:03:01PM -0400, Nick Holland wrote:
> I could be wrong..it may be that only two people have ever tried to
> install OpenBSD on a machine with a DVDRAM drive into OpenBSD, you and
> me.  Or maybe you and I are the only owners of defective DVDRAM drives.

I have a DVDRAM drive with the following dmesg

 cd0 at scsibus0 targ 0 lun 0:  SCSI0 5/cdrom 
removable

and I have not had any problems installing OpenBSD 4.2 on the machine (a
ThinkPad T60).

Re: Firefox 2.0.0.12

[Matthew Szudzik]

On Mon, Apr 07, 2008 at 04:44:08PM +, Jacob Meuser wrote:
> or, quit using firefox.  it's security record is rather lousy, wouldn't
> you agree?
 
What alternatives to firefox do you suggest?

Re: Intel Core2 Dual/Quad - i386 or amd64?

[Matthew Szudzik]

> I have two computers I'm considering going amd64 on:

Another thing to keep in mind is that there are some OpenBSD packages
and ports that only run on i386.  In particular, any software that
requires Linux emulation (such as the opera-flashplugin) will not run on
amd64.

Doing a quick grep of the OpenBSD 4.2 ports tree, I've come up the
following list of packages and ports that run on i386 but not amd64.

/usr/ports/archivers/rar/Makefile:ONLY_FOR_ARCHS=   i386 sparc
/usr/ports/audio/gogo/Makefile:ONLY_FOR_ARCHS= i386
/usr/ports/audio/rio500/Makefile:ONLY_FOR_ARCHS=i386
/usr/ports/audio/wmtune/Makefile:ONLY_FOR_ARCHS=i386
/usr/ports/chinese/rxvt-big5/Makefile:ONLY_FOR_ARCHS=   i386
/usr/ports/chinese/xcin25/Makefile:ONLY_FOR_ARCHS=  i386
/usr/ports/devel/ald/Makefile:ONLY_FOR_ARCHS=   i386
/usr/ports/devel/jad/Makefile:ONLY_FOR_ARCHS=   i386
/usr/ports/devel/jdk/1.3/Makefile:ONLY_FOR_ARCHS=   arm i386 powerpc sparc
/usr/ports/devel/jdk/1.3-linux/Makefile:ONLY_FOR_ARCHS= i386
/usr/ports/devel/jdk/1.4/Makefile:ONLY_FOR_ARCHS=   i386
/usr/ports/devel/nasm/Makefile:ONLY_FOR_ARCHS=  i386
/usr/ports/devel/prc-tools/Makefile:ONLY_FOR_ARCHS= i386
/usr/ports/devel/py-psyco/Makefile:ONLY_FOR_ARCHS=  i386 
/usr/ports/emulators/freebsd_lib/Makefile:ONLY_FOR_ARCHS=   i386
/usr/ports/emulators/ines/Makefile:ONLY_FOR_ARCHS=  i386
/usr/ports/emulators/mastergear/Makefile:ONLY_FOR_ARCHS=i386
/usr/ports/emulators/redhat/Makefile.inc:ONLY_FOR_ARCHS=i386
/usr/ports/emulators/snes9x/Makefile:ONLY_FOR_ARCHS=i386
/usr/ports/emulators/vgb/Makefile:ONLY_FOR_ARCHS=   i386
/usr/ports/emulators/vmware/3/Makefile:ONLY_FOR_ARCHS= i386
/usr/ports/emulators/wine/Makefile:ONLY_FOR_ARCHS= i386
/usr/ports/emulators/zsnes/Makefile:ONLY_FOR_ARCHS= i386
/usr/ports/emulators/fedora/Makefile.inc:ONLY_FOR_ARCHS=i386
/usr/ports/games/adom/Makefile:ONLY_FOR_ARCHS=  i386
/usr/ports/games/quake2/Makefile:ONLY_FOR_ARCHS=i386
/usr/ports/graphics/cqcam/Makefile:ONLY_FOR_ARCHS=  i386
/usr/ports/graphics/win32-codecs/Makefile:ONLY_FOR_ARCHS=   i386
/usr/ports/lang/clisp/Makefile:ONLY_FOR_ARCHS=  i386
/usr/ports/lang/ezm3/Makefile:ONLY_FOR_ARCHS=   i386
/usr/ports/lang/gcc/3.3/Makefile:ONLY_FOR_ARCHS=alpha i386 m68k sparc 
sparc64 powerpc vax
/usr/ports/lang/gcc/3.4/Makefile:ONLY_FOR_ARCHS=alpha i386 m68k sparc 
sparc64 powerpc vax
/usr/ports/lang/gcc/4.0/Makefile:ONLY_FOR_ARCHS=alpha i386 m68k sparc 
sparc64 powerpc vax
/usr/ports/lang/gcc/4.1/Makefile:ONLY_FOR_ARCHS=alpha i386 m68k sparc 
sparc64 powerpc vax
/usr/ports/lang/gprolog/Makefile:ONLY_FOR_ARCHS=i386
/usr/ports/lang/pm3/Makefile:ONLY_FOR_ARCHS=i386
/usr/ports/lang/smlnj/Makefile:ONLY_FOR_ARCHS=  i386
/usr/ports/lang/tendra/Makefile:ONLY_FOR_ARCHS= i386
/usr/ports/lang/jamvm/Makefile:ONLY_FOR_ARCHS=  i386 powerpc arm
/usr/ports/lang/scheme48/Makefile:ONLY_FOR_ARCHS=   i386
/usr/ports/math/maple/Makefile:ONLY_FOR_ARCHS=  i386
/usr/ports/math/maple-share/Makefile:ONLY_FOR_ARCHS=i386
/usr/ports/math/wingz/Makefile:ONLY_FOR_ARCHS=  i386
/usr/ports/misc/dnetc/Makefile:ONLY_FOR_ARCHS=  i386 m68k sparc
/usr/ports/misc/tpwireless/Makefile:ONLY_FOR_ARCHS= i386
/usr/ports/net/cvsup/Makefile:ONLY_FOR_ARCHS=   i386
/usr/ports/net/dss/Makefile:ONLY_FOR_ARCHS= i386
/usr/ports/net/libst/Makefile:ONLY_FOR_ARCHS=   alpha i386 sparc
/usr/ports/net/openafs/Makefile:ONLY_FOR_ARCHS= i386
/usr/ports/net/pdnsd/Makefile:ONLY_FOR_ARCHS=   i386
/usr/ports/net/ser/Makefile:ONLY_FOR_ARCHS= i386 sparc
/usr/ports/palm/isilo/Makefile:ONLY_FOR_ARCHS=  i386
/usr/ports/palm/pose/Makefile:ONLY_FOR_ARCHS=   i386
/usr/ports/print/acroread/Makefile:ONLY_FOR_ARCHS=  i386
/usr/ports/security/l0phtcrack/Makefile:ONLY_FOR_ARCHS= i386
/usr/ports/sysutils/libretto-config/Makefile:ONLY_FOR_ARCHS=i386
/usr/ports/sysutils/perfd/Makefile:ONLY_FOR_ARCHS=  i386 macppc sparc zaurus
/usr/ports/sysutils/tphdisk/Makefile:ONLY_FOR_ARCHS=i386
/usr/ports/sysutils/wdsetup/Makefile:ONLY_FOR_ARCHS= i386
/usr/ports/sysutils/wmapm/Makefile:ONLY_FOR_ARCHS=  arm i386 macppc
/usr/ports/sysutils/xbatt/Makefile:ONLY_FOR_ARCHS=  arm i386 macppc 
sparc
/usr/ports/sysutils/grub/Makefile:ONLY_FOR_ARCHS=   i386
/usr/ports/www/mod_frontpage/Makefile:ONLY_FOR_ARCHS=   i386
/usr/ports/www/opera/Makefile:ONLY_FOR_ARCHS=   i386
/usr/ports/www/opera-flashplugin/Makefile:ONLY_FOR_ARCHS=   i386
/usr/ports/x11/asapm/Makefile:ONLY_FOR_ARCHS=   arm i386 macppc
/usr/ports/x11/gatos-bin/Makefile:ONLY_FOR_ARCHS=   i386
/usr/ports/x11/xfce4/xfce4-battery/Makefile:ONLY_FOR_ARCHS= arm i386 
powerpc sparc
/usr/ports/x11/gnustep/gnustep.port.mk:ONLY_FOR_ARCHS=  i386

Re: Intel Core2 Dual/Quad - i386 or amd64?

[Matthew Szudzik]

> I have two computers I'm considering going amd64 on:
>
> 1. My Thinkpad X61s (1GB RAM)

It appears that the wireless driver your Thinkpad requires

> wpi0 at pci2 dev 0 function 0 "Intel PRO/Wireless 3945ABG" rev 0x02: apic 2 

is not supported in OpenBSD/amd64.  See

 http://www.openbsd.org/amd64.html

Re: minimac on openbsd

[Matthew Szudzik]

> doesn't support firewire.  I don't know if the built in wireless works
> as I've never tried it.  It is recognized by the kernel:
> 
>   ath0 at pci2 dev 0 function 0 "Atheros AR5424" rev 0x01: apic 2 int 17 (irq 
> 11)

Atheros AR5424 wireless devices are not supported in OpenBSD.  See
 http://marc.info/?l=openbsd-misc&m=118896011519883

Re: using openbsd to make presentations

[Matthew Szudzik]

> Question: Do you have any recommendation / suggestion to prepare talks
> to be shown in a projector including mathematical equations, plots
> and, eventually, movies (I can live without this last point)?

HTML is probably the most portable solution for your problem, and movies
would work fine too (using VLC's Mozilla plug-in).  Graphics display
quickly and Firefox has MathML for displaying equations, but special
fonts are required, and I'm unsure if anyone has ever tried to install
them on OpenBSD (I certainly haven't).  An example of MathML used in
HTML is at

 http://pear.math.pitt.edu/mathzilla/Examples/markupOftheWeek.mhtml

Personally, I use Mathematica on my OpenBSD laptop--it has a nice
presentation mode and renders equations beautifully.  Of course, it's
proprietary software that costs money, so it's not for everyone.

Re: ath0 - not reachable - system hangs

[Matthew Szudzik]

> ath0 at pci0 dev 12 function 0 "Atheros AR2413" rev 0x01: irq 9
> ath0: AR2413 7.8 phy 4.5 rf 5.6, FCC2A*, address 00:1d:0f:af:98:88

According to the CVS log at
 http://www.openbsd.org/cgi-bin/cvsweb/src/sys/dev/ic/ath.c#rev1.56
"support is still incomplete" for the AR2413 chipset.

Re: [OFFTOPIC] Naming convention for programs

[Matthew Szudzik]

> I was thinking of verb-subject, or verb_subject, or viceversa.

If you are writing a family of related programs, then the word that they
have in common should be first.  For example, in OpenBSD there are

 mount_cd9660
 mount_ffs
 mount_msdos

and so on.  The fact that these programs perform related operations is
immediately apparent, and it is easy to find all of the related programs
simply by typing

 mount_

followed by TAB on the command-line.

Note that the word that is common could be a verb or a subject.
Although it is a verb in the case of "mount", it is a subject in the
case of "pkg":

 pkg_add
 pkg_create
 pkg_delete
 pkg_merge

> no separator between subject and verb.

Historically, "no separator" seems to win.  The best example is the
World Wide Web.  When it started in the 1990's, there were a plethora of
hyphen-separated URLs, but over time they have disappeared and been
replaced with no separator.

Mathematicians (who have dealt with this issue for centuries) also
prefer no separator:  "sinh" for "sine hyperbolic", "ln" for "logarithm
natural", "arcsin" for "arc sine", and "arcsinh" for "arc sine
hyperbolic".

The oldest and most-used Unix commands also follow this convention:
"lpq", "lpr", "lprm", "whereis", "mkdir", and so on.

Re: The Dilbert Problem...

[Matthew Szudzik]

> I've been noticing a similar problem with Firefox on OpenBSD...

I've also experienced this problem, but was never able to reproduce it.
It would happen maybe once every month or two during normal web browsing
(which in my case means 5 or more tabs simultaneously open in Firefox).

Re: compat_linux(8) has 2GB filesize limit in 4.2-stable

[Matthew Szudzik]

> The COMPAT_LINUX code is in dire need of an upgrade to match more recent
> linux kernels...

I certainly agree.  For example, I need to compile a custom OpenBSD
kernel with this patch

 http://marc.info/?l=openbsd-misc&m=119479722118605

just to get certain programs running under compat_linux.

Error in softraid documentation?

[Matthew Szudzik]

I recently noticed that the examples in the softraid man page

 http://www.openbsd.org/cgi-bin/man.cgi?query=softraid

contain many lines such as

 echo "d a\na\n\n\n\nRAID\nw\nq\n" | disklabel -E wd1

Of course, not every version of echo interprets "\n" as a newline.  In
fact, /bin/echo treats "\n" as a literal backslash followed by a literal
n.  The version of echo that is built into csh also interprets it as a
literal backslash followed by a literal n.  But the softraid man page
certainly intends it to be interpreted as a newline.

So, is the man page in error?  Or are the examples in man pages only
intended for use in the default shell?  (Note that ksh is the default
shell, and ksh has a built-in echo command that interprets "\n" as a
newline.)

Re: OSS v4.0 released under BSD license

[Matthew Szudzik]

> There is also the question of ALSA compatibility layer which is in my
> understanding slowly incorporated into OSS. Is it really important to have
> ALSA compatibility layer? Can somebody give me an example of the software
> which  requires ALSA (please exclude Skype although there is OSS version)

Yes, an ALSA compatibility would allow us to use Linux binaries (through
compat_linux) which require ALSA.  The most infamous example is the
Adobe Flash 9 Plugin.