Re: [SLUG] messed up yum update

[Gavin Carr]

You generally shouldn't have both epel and rpmforge enabled, as they'll 
frequently conflict. Try disabling one or other (set enabled=0 in the 
relevant repo file in /etc/yum.repos.d) and try it again.

You can also test by doing 'yum update --disablerepo=rpmforge' or 'yum
update --disablerepo=epel' to see which seems to give you the better
outcome here. But I'd still recommend disabling one of them long-term,
or you'll see this kind of thing a lot.

Cheers,
Gavin


On Sun, Dec 15, 2013 at 09:49:23AM +1100, li...@sbt.net.au wrote:
> I have a centos system, getting this[1]:

> what the best way forward ? remove clam/amavis and install again, or ?

> # amavisd -V
> amavisd-new-2.8.0 (20120630)
> # clamd --version
> ClamAV 0.98/18238/Sat Dec 14 22:44:32 2013

> [1]# yum update
> ..
> Resolving Dependencies
> --> Running transaction check
> ---> Package clamav.x86_64 0:0.98-1.el6 will be updated
> ---> Package clamav.x86_64 0:0.98-2.el6.rf will be an update
> ---> Package clamav-db.x86_64 0:0.98-1.el6 will be updated
> ---> Package clamav-db.x86_64 0:0.98-2.el6.rf will be an update
> ---> Package clamd.x86_64 0:0.98-1.el6 will be updated
> ---> Package clamd.x86_64 0:0.98-2.el6.rf will be an update
> --> Processing Dependency: /etc/clamd.d for package:
> amavisd-new-2.8.0-4.el6.noarch
> --> Finished Dependency Resolution
> Error: Package: amavisd-new-2.8.0-4.el6.noarch (@epel)
>Requires: /etc/clamd.d
>Removing: clamd-0.98-1.el6.x86_64 (@epel)
>Not found
>Updated By: clamd-0.98-2.el6.rf.x86_64 (rpmforge)
>Not found
>Available: clamd-0.98-2.el6.x86_64 (epel)
>Not found
>Available: clamd-0.96.4-1.el6.rf.x86_64 (rpmforge)
>Not found
>Available: clamd-0.96.5-1.el6.rf.x86_64 (rpmforge)
>Not found
>Available: clamd-0.97-1.el6.rf.x86_64 (rpmforge)
>Not found
>Available: clamd-0.97.1-1.el6.rf.x86_64 (rpmforge)
>Not found
>Available: clamd-0.97.2-1.el6.rf.x86_64 (rpmforge)
>Not found
>Available: clamd-0.97.3-1.el6.rf.x86_64 (rpmforge)
>Not found
>Available: clamd-0.97.4-1.el6.rf.x86_64 (rpmforge)
>Not found
>Available: clamd-0.97.5-1.el6.rf.x86_64 (rpmforge)
>Not found
>Available: clamd-0.97.5-2.el6.rf.x86_64 (rpmforge)
>Not found
>Available: clamd-0.97.6-1.el6.rf.x86_64 (rpmforge)
>Not found
>Available: clamd-0.97.7-1.el6.rf.i686 (iRedMail)
>Not found
>Available: clamd-0.98-1.el6.rf.x86_64 (rpmforge)
>Not found
>  You could try using --skip-broken to work around the problem
>  You could try running: rpm -Va --nofiles --nodigest
-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] problems with perl IO:File binmode

[Gavin Carr]

On Thu, Feb 11, 2010 at 03:24:19PM +1100, Voytek Eymont wrote:
> I am a failure in mail processing with amavisd/spamassasin
> --
> AFA05B44845  3085241 Thu Feb 11 10:07:31  s...@hotmail.com
> (host 127.0.0.1[127.0.0.1] said: 451 4.5.0 Error in processing,
> id=21117-12, mime_decode-1 FAILED: Can't locate object method "binmode"
> via package "IO::File" at /usr/lib/perl5/site_perl/5.8.5/MIME/Body.pm line
> 437. (in reply to end of DATA command))
> --
> I've reinstalled IO:File
> 
> I've googled but haven't found any resolution
> 
> any thoughts or suggestions ?
> 
> ---
> # perl -MCPAN -e shell
> 
> cpan shell -- CPAN exploration and modules installation (v1.9402)
> Enter 'h' for help.
> 
>   
>   cpan[1]>
> install
> IO::File
> CPAN: Storable loaded ok (v2.13)
> Going to read '/root/.cpan/Metadata'
>   Database was generated on Thu, 11 Feb 2010 01:32:49 GMT
> IO::File is up to date (1.14).

binmode works fine with that version of IO::File. Maybe you have another
version installed as well somewhere? What does:

  perl -MIO::File -le 'print $IO::File::VERSION'

say?

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] deleting older files from /boot

[Gavin Carr]

On Tue, Feb 09, 2010 at 05:53:48PM +1100, Jobst Schmalenbach wrote:
> On Tue, Feb 09, 2010 at 02:07:53PM +1100, Voytek Eymont (li...@sbt.net.au) 
> wrote:
> > 
> > my centos system just run out of space on /boot:
> > 
> > # df
> > Filesystem   1K-blocks  Used Available Use% Mounted on
> > /dev/mapper/VolGroup00-LogVol00
> >  101793144  28730252  67892104  30% /
> > /dev/cciss/c0d0p198747 92263  1385  99% /boot
> > none   1168044 0   1168044   0% /dev/shm
> > /dev/mapper/VolGroup00-LogVol02
> >1064312 34180976068   4% /tmp
> > 
> > 
> > if I delete older versions of init-rd*, vmlinuz*, do I need to do anything
> > else after deleteting the older files ?
>
> 
> Do
> 
>  rpm -qa | grep kernel
> 
> then do:
> 
>  yum erase kernel-XYZ
> 
> of the older ones. Do not delete the files in boot directly
> as you stuff up your rpm database and yum WILL get confused


First, you really should not be running 55.0.2, or whatever. That's a 
kernel from Centos 5.2, which is ancient. So the first thing to do is
schedule a reboot.

Then do the following:

  rpm -qa 'kernel*'

  # Decide a set of versions to delete e.g. the 67.x.y releases
  rpm -qa 'kernel*' | grep -- -67

  # Confirm that list contains only what you want to delete, and then:
  rpm -qa 'kernel*' | grep -- -67 | xargs rpm -e


You can use yum for the last step if you prefer, but it doesn't add any value
for kernel removes, and it's slower.

  
Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] where to get an Ethernet hub (NOT a switch)

[Gavin Carr]

On Sun, Jul 19, 2009 at 09:36:47AM +1000, Amos Shapira wrote:
> I'm looking for an Ethernet hub to be used for network troubleshooting
> (trying to find which of our hosts is involved in the load on our
> office uplink).

I bought a few Dlink 10MB hubs from the US quite some time back for messing
with network sniffing. Didn't reply earlier as I couldn't remember where
they were, but I've found them tonight. :-)

So you're welcome to one if you'd like Amos. Ping me off-list and we'll
tee something up.

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] useful bash tricks thread

[Gavin Carr]

On Fri, Feb 06, 2009 at 09:06:25PM +1100, Tony Sceats wrote:
> It's been a while since there's been a thread like this, so I thought it
> would be fun :)
> 
> so, have you got any?

I'm fond of the -t flag to ls, which orders by mtime desc. I have
these defined in my .bashrc:

  lead() {
ls -lt $* | head -20
  }

which is combination ls+head, giving you back the 20 most recently 
modified items in your current (or given) directory; and:

  alias l1='ls -t | head -1'

which gives you back the filename of the most recently modified 
object. I use this all the time with `` e.g.

  # Create an invoice
  invoice companyX
  # Check it (the newest file) in evince
  evince `l1`


Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: Which wireless data service should I signup to? was [SLUG] Don't buy ZTE's

[Gavin Carr]

On Thu, Dec 04, 2008 at 04:45:51PM +1100, Sridhar Dhanapalan wrote:
> 2008/12/4 david <[EMAIL PROTECTED]>:
> > I've been following this thread and checking the links... all of them refer
> > to "software" which I assume is Windows only.
> >
> > Should I assume that any of them work on either Mac or Linux, or both,
> > consecutively. Or is that a dangerous assumption?
> 
> I can't speak for them all, but the Huawei E220 and E169 (which are
> the most common 3G modems you'll see) work just fine with Linux (at
> least on the Optus network). 

And ditto the Huawei E800, which is the Express Card/PCMCIA card used by
Vodafone (and I think by 3 and Optus as well). I like the Express Card 
form factor as being less 'extrusive' than the USBs; the only real
disadvantage with them is that they're not usually usable by current
netbooks.

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: Which wireless data service should I signup to? was [SLUG] Don't buy ZTE's

[Gavin Carr]

On Thu, Dec 04, 2008 at 01:20:12PM +1100, Marghanita da Cruz wrote:
> The telstra plans are here:
> 
>
> and the rest:
> 
> 
> 

and 3:

  
http://www.three.com.au/cs/ContentServer?c=Page&pagename=Three%2FPage%2FBusinessVideoCallingTemplate&p=1156241342637&cid=1154931041257

PC Authority had a recent roundup here:

  
http://www.pcauthority.com.au/GroupTests/127633,mobile-broadband-the-big-roundup.aspx

and picked 3 as being the best value if you mainly needed urban coverage,
with Optus coming in second. AFAIK neither vodafone or 3 have capped plans 
though. Optus has pretty ordinary reviews on whirlpool though.

Also be aware that if you have an ABN available you can get 'business' 
accounts, which often have shared bandwidth limits, and cheaper overcap rates, 
etc. (3's overcap rates are 1c/MB for biz, but 10c/MB for personal accounts,
for instance).

We're about to sign up with vodafone at $dayjob, as having better coverage 
than 3, but still pretty good pricing. Vodafone may be a little slower, but 
I mainly use it via ssh though, so don't much care about speed.

Cheers,
Gavin


-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] perl equivalent for cd $(dirname $0)?

[Gavin Carr]

On Wed, Apr 23, 2008 at 05:37:42PM +1000, Sonia Hamilton wrote:
> In my bash scripts I often use this (to change to the directory where
> the script is):
> 
> cd $(dirname $0)
> 
> Is there an equivalent in perl?

  use File::Basename;
  chdir dirname $0;

:-)

Note that perl's $0 is not super-portable (see `perldoc perlvar` for details). 
If you care you want to look at the FindBin module.

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] managing .bashrc in subversion?

[Gavin Carr]

On Tue, Apr 08, 2008 at 03:04:59PM +1000, Sonia Hamilton wrote:
> A "process" rather than "technical" question about subversion.
> 
> My personal subversion repo is setup, works ok, etc. But I notice that
> subversion only lets you checkout a directory, not a file.
> 
> I want to manage (for example) my .bashrc file in subversion - how to do
> it? I could check in all of /home/sonia, but I'd have to setup heaps of
> exclusions - nasty. I could put .bashrc in /home/sonia/bin (for example)
> and link to it "cd; ln -s bin/.bashrc .bashrc) - a hack.
> 
> Any other suggestions?


I use a directory called ~/.dotfiles which contains a few *.bashrc and 
*.profile files, a vimrc, an Xdefaults, etc.

I either include them from the top-level instance e.g. in .bashrc:

  test -d $HOME/.dotfiles && . $HOME/.dotfiles/*.bashrc

or just symlink them in.

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] CPAN modules documentation

[Gavin Carr]

Hi Sonia,

On Wed, Mar 26, 2008 at 08:26:10PM +1100, Sonia Hamilton wrote:
> I'm without an internet connection at the moment (oh the pain).
> 
> I'm doing a fair bit of Perl stuff at the moment; I've downloaded the
> tgz of all the Perl documentation for easy browsing.
> 
> Anyone know if there's a similar tgz of documentation for all the CPAN
> modules for easy browsing locally? I know about wget; I'm thinking maybe
> there's an easier way...

If you've got the disk (currently 871M), the easiest way is probably 
CPAN::Mini. This lets you download all of CPAN to your disk, but only the
very latest versions of each module. It gives you not just documentation 
on every module, but the modules themselves, so you can search and install 
any module you want while offline (and fast!). I find it invaluable.

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] batch rename

[Gavin Carr]

On Sat, Nov 03, 2007 at 04:30:35PM +0900, Kevin Shackleton wrote:
> suggestions to rename multiple files in a single directory to individual
> shorter names?
> eg
>   proj1file1
>   proj1file2
>   . . .
> to
>   file1
>   file2
>   . . .

  man rename
  rename proj1 '' proj1*


Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] perl cpan reload problems

[Gavin Carr]

On Mon, Oct 08, 2007 at 01:12:20AM +1000, Voytek Eymont wrote:
> I've entered ;perl -MCPAN -e shell; to load some modules on Centos 4;
> 
> dialogue came up telling me there was a better cpan to be had, and
> suggested to install/reload cpan
> 
> which I did attempt to do, but, it didn't quite worked out to plan
> 
> now, when I restart it, it says as below, followed by a few more questions
> that mean very little to me
> 
> how to fix...?

This is just telling you that your new CPAN has a couple of extra options
it wants values for. Just answer the questions (or accept the defaults)
and you'll be back to normal.

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] treating \n like any other character

[Gavin Carr]

On Wed, Oct 03, 2007 at 06:52:17PM +1000, david wrote:
> On Wed, 2007-10-03 at 18:44 +1000, Amos Shapira wrote:
> > On 03/10/2007, david <[EMAIL PROTECTED]> wrote:
> > >
> > > I want to edit a multi line file as if it were all one line
> > >
> > > In other words, treat \n like any other character, and specifically
> > > doing global find and replace.
> > >
> > > I know there are various hex editors, but they are all pretty clunky as
> > > far as I can see, and none seem to be able to do that from command line.
> > >
> > > Is there a shell script way to do it?
> > 
> > 
> > Can you be more specific of what you are trying to achieve?
> > 
> 
> 
> [EMAIL PROTECTED]:~/test/testdir $ cat > test
> 1 
> 2
> 3
> [EMAIL PROTECTED]:~/test/testdir $ sed s/1\n/1/g test 
> 1
> 2
> 3
> [EMAIL PROTECTED]:~/test/testdir $ 
> 
> The output I would have liked would be:
> 
> 12
> 3


nox:~/tmp$ cat > test
1
2
3
nox:~/tmp$ 
nox:~/tmp$ perl -i -pe 's/1\n/1/' test
nox:~/tmp$ cat test
12
3
nox:~/tmp$


Cheers,
Gavin


-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Suspenseful laptops

[Gavin Carr]

On Thu, Sep 13, 2007 at 01:46:59PM +1000, James Gregory wrote:
> On Thu, 2007-09-13 at 12:30 +1000, Jeff Waugh wrote:
> > 
> > 
> > > Can anyone recommend a laptop that is small (11.1 - 13.3 inches),
> > > reasonably fast, and can do suspend (both to RAM and disk) easily? Oh, and
> > > that I can buy now; old models aren't so interesting to me. Ubuntu is my
> > > OS of preference, but if it makes a difference, I'd consider switching to
> > > something else.
> 
> > The Dell M1330 is also pretty intriguing -- backlight and SSD storage! It is
> > 13", slightly bigger than the D430.
> 
> Yeah, that does look interesting, but it seems a bit weird having a
> wedge-shaped laptop. Still a contender though.

I've got the earlier M1210 and love it. I've got everything working under linux
except the front microphone jack (which seems to need special driver support) 
and 
the webcam (which I just haven't got to yet).

We've got an M1330 coming in in about a week too, and we'll be putting linux on
it. I'll report back how we go.

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Minimal CentOS installation --- How?

[Gavin Carr]

Hi Peter,

On Tue, Sep 04, 2007 at 10:20:53AM +1000, Peter Chubb wrote:
>   I'm trying to write some howtos.  For Debian lots of things
> are easy (because that's what I use everyday).  For
> RedHat/CentOS/Fedora, I want to create an environment I can play in --
> a chroot environment using yum, and so  on.
> 
>   How can I install the bare minimum --- something like the
> result of debootstrap --- just enough to run yum?  As it'll be in a
> chroot I don't need any network utilities, etc,
> 
> Here's what I tried:
> 
>   Boot qemu on an empty disc using the first installation
>   cdrom.  Install with everything possible turned off.  The result was
>   over 1G, and stuff like X libraries, spelling checkers in x86 and
>   x86_64 versions,  etc., were still installed.  In fact there were quite
>   a few x86 versions of things installed unnecessarily.  I did a yum
>   remove *.i386 and removed 60 packages straight away.
> 
>   I could then shutdown qemu, loopback mount the filesystem and
>   copy the files into a directory ready for chroot.
> 
> But I'm sure there's other unnecessary stuff in
> there. Removing all the autosetup of networking gets rid of
> another 21 packages; and then I had to add gcc, binutils and
> rpmutils; leaving me a chroot of 1.2G.  By comparison, a
> minimal debian install is 93M -- enough to run apt-get.
> 
> Any ideas?

I did this by hand a few months ago to create a 32-bit chroot on a
64-bit machine - it's not that complicated (though rpmstrap is probably 
easier, of course), here are my notes:

  # based on owlriver caos chroot script, and planet-lab fedora build script: 
  # - ftp://ftp.owlriver.com/pub/local/ORC/ORCrebuild/ORCyum-chroot-caos
  # - http://build.planet-lab.org/build/nightly/2006.03.27/mkfedora

  export CHROOT=/export/chroot/centos4-i386
  mkdir -p $CHROOT
  mkdir -p $CHROOT/{dev,proc,tmp}
  mkdir -p $CHROOT/etc/{rpm,sysconfig}
  echo "NETWORKING=yes" > $vroot/etc/sysconfig/network
  mkdir -p $CHROOT/var/lib/rpm
  mkdir -p $CHROOT/var/{log,tmp}
  touch $CHROOT/proc/mounts
  touch $CHROOT/etc/mtab
  touch $CHROOT/etc/redhat-release
  rpm --root $CHROOT --initdb
  for i in null ptmx zero random urandom tty full; do
/dev/MAKEDEV -d $CHROOT/dev $i
  done
  # Setup yum.conf
  cat <$CHROOT/etc/yum.conf
  [main]
  cachedir=/var/cache/yum
  debuglevel=2
  logfile=/var/log/yum.log
  pkgpolicy=newest
  distroverpkg=centos-release
  tolerant=1
  exactarch=0
  obsoletes=1
  reposdir=/dev/null

  [bootstrap]
  name=CentOS-\$releasever - Base
  baseurl=http://web/mrepo/centos\$releasever-i386/RPMS.os/
  EOD
  # Prevent all locales from being installed in reference image
  mkdir -p $CHROOT/etc/rpm
  cat >$CHROOT/etc/rpm/macros 

Re: [SLUG] WAN link optimisation

[Gavin Carr]

On Thu, Jul 26, 2007 at 10:44:19AM +1000, Visser, Martin wrote:
> I think that the technology Gavin is thinking of is more about
> economising on the content being sent rather than tweaking TCP
> parameters. 

Thanks a lot for all the comments in this thread so far - been some
interesting reading. 

So far the thread's covered two areas:

1. tuning TCP parameters for WAN performance (James' comments and links)

2. content compression magic with specialist middleboxes (Martin/Glen and 
   others)

I've done a fair bit of stuff around (1) before (some based on stuff 
Glen's written). In this particular instance I'm more interested in 
latency and interactive behaviour than raw throughput though, and I 
want to minimise exposure to long-haul link congestion and dropouts.

I've found another bunch of work around this:

3. the notion of 'network striping' - opening multiple connections over 
   one or more links to improve window scaling and/or to limit exposure
   issues on a single link e.g.

   - http://www.cesnet.cz/doc/techzpravy/2006/psock/

   - http://portal.acm.org/citation.cfm?id=370413&coll=portal&dl=ACM

   - http://nms.csail.mit.edu/papers/index.php?detail=127

which look pretty interesting for my use case. Anyone have any experience 
with any of these?


What I'd _really_ like, though, (and haven't found any explicit references 
to yet) is like (3) but actually duplicating packets down multiple links, a 
sort of 'network raid 1' where (3) is network raid 0. In other words, 
something that transparently splits a stream into multiple duplicate streams 
down separate links, which are then merged/multiplexed at the other end, and
duplicates discarded. Effectively trading bandwidth for latence, given 
multiple links.

Anyone heard of anything at all like that? Or am I crazy?


Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

[SLUG] WAN link optimisation

[Gavin Carr]

Hey sluggers,

Anyone have any pointers to open source projects (or features of projects) 
around WAN link optimisation? I'm specifically looking for a way of 
duplicating traffic across multiple links to avoid resends on high latency
links, but I'm interested in the whole area.

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Has anyone any insight as to why RedHat Enterprise Level 5 is broken as far as PXE booting is concerned.

[Gavin Carr]

Hi Roger,

On Mon, May 28, 2007 at 02:32:55PM +1000, RgSalisbury wrote:
> Has anyone any insight as to why RedHat Enterprise Level 5 is broken as far 
> as PXE booting is concerned.
> 
> 
> Is there a Protocol communication problem 
> 
> It seems:
> 
> the PXE CLIENT when booting cannot access files in  the "pxelinux.cfg" dir
> 
> RHEL4 works just fine !!
> 
> But RHEL5 seems to be broken. is the TFTP SERVER broken 
> (tftp-server-0.42-3.1.i386.rpm)  
> 
> 
> Client very SLOWLY displays thus:
> 
> Trying to load: pxelinux.cfg/AC1000FD
> Trying to load: pxelinux.cfg/AC1000F
> Trying to load: pxelinux.cfg/AC1000
> Trying to load: pxelinux.cfg/AC100
> Trying to load: pxelinux.cfg/AC10
> Trying to load: pxelinux.cfg/AC1
> Trying to load: pxelinux.cfg/AC
> Trying to load: pxelinux.cfg/A
> Trying to load: pxelinux.cfg/default

Not sure if you sorted this out Roger, but I just ran into it myself. I bet 
you've 
got the rpmforge repository enabled, right?

Apparently there's a change in way the latest versions of syslinux (3.50, 3.51 
- which
are on rpmforge, not in base) query for images. I haven't yet figured out how 
to fix
it, but if you revert to syslinux 3.35 or earlier it will magically start 
working 
again. I've seen the problem with both vanilla tftp-server and dnsmasq's tftp.


Cheers,
Gavin


-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Academic research software

[Gavin Carr]

On Tue, Feb 27, 2007 at 12:34:26PM +1100, Sridhar Dhanapalan wrote:
> On Mon, 26 Feb 2007, Gavin Carr <[EMAIL PROTECTED]> wrote:
> > Hi all,
> >
> > I've just had a friend ask me whether there's anything in the free software
> > world for academic research / writing i.e. tracking bibliographic info,
> > citations, quotes etc., and then collating them into a written product.
> >
> > Any cluesticks? What do you real academics out there use (without wanting
> > to start an editor and/or word processor war!).
> 
> There are a number of tools available to aid research. OpenOffice.org has for 
> a long time had functionality to manage sources and bibliographic entries. 
> Two standalone apps which come to mind are Tomboy[1] and BasKet[2].
> 
> For Web-based research, it might make sense to manage sources within the Web 
> browser itself. There are several extensions for Firefox to do this[3], 
> including Zotero[4], Research Buddy[5], and Diigo[6].
> 
> 
> [1] http://www.gnome.org/projects/tomboy/
> [2] http://basket.kde.org/
> [3] https://addons.mozilla.org/search.php?q=research&type=E&app=firefox
> [4] http://www.zotero.org/
> [5] http://researchbuddy.mozdev.org/
> [6] https://addons.mozilla.org/firefox/2792/

Thanks a lot to all who replied - I've got a swag of things to go 
away and try out now. Should be fun!

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

[SLUG] Academic research software

[Gavin Carr]

Hi all,

I've just had a friend ask me whether there's anything in the free software
world for academic research / writing i.e. tracking bibliographic info, 
citations, quotes etc., and then collating them into a written product. He's 
used a commercial Windows product called Nota Bene before:

  http://www.notabene.com/product_tour_overview1.html

Sounds like the sort of thing that much be an itch for lots of academics, but
I've not run across anything more specialised like this in the free software 
world. 

Any cluesticks? What do you real academics out there use (without wanting to
start an editor and/or word processor war!).

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Recommendations for music download sites

[Gavin Carr]

On Fri, Dec 22, 2006 at 12:00:37AM +1100, elliott-brennan wrote:
> So, I'm interested as to whether there are sites 
> that legitimately allow me to pay for and download 
> contemporary music that I can just burn and play?

I like:

  http://www.mp3tunes.com/
  http://www.magnatune.com/

Magnatune have less mainstream stuff, but I like their values a lot: 
http://www.magnatune.com/info/whynotevil. For instance, they give 50% 
of all proceeds to the artist (compare that with your traditional 
record label!). And you can listen to any and all of their music in 
entirety before you buy.

Cheers,
Gavin


-- 
Gavin Carr
Open Fusion - Open Source Business Solutions [ Linux - Perl - Apache ]
http://www.openfusion.com.au
- Fashion is a variable, but style is a constant - Programming Perl
-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] PCIe Network Carding

[Gavin Carr]

On Tue, Nov 14, 2006 at 04:18:33PM +1100, Stephen Black wrote:
> I am looking for a PCIe Network card for the my Linux machine.
> Does anybody use such a device?
> 
> But I am finding it hard to find out whether the PCIe is also compatable

PCIe works fine. 

Intel does PCIe Gbit cards - they're their 'PT' cards e.g.

  
http://www.retailing.com.au/prod_140601_proddesc_INTEL_PRO_1000_PT_DESKTOP_ADAPTER_EXPI9300PT.html
  
http://www.retailing.com.au/prod_141060_proddesc_INTEL_PRO_1000_PT_SERVER_ADAPTER_EXPI9400PT.html

Cheers,
Gavin

-- 
Gavin Carr
Open Fusion - Open Source Business Solutions [ Linux - Perl - Apache ]
http://www.openfusion.com.au
- Fashion is a variable, but style is a constant - Programming Perl
-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Contracting stuff: wrap it into a company or PAYE through agency?

[Gavin Carr]

On Wed, Sep 27, 2006 at 05:00:26PM +1000, david wrote:
> On Wed, 2006-09-27 at 14:54 +1000, Howard Lowndes wrote:
> > Gavin Carr wrote:
> > > On Wed, Sep 27, 2006 at 12:28:16PM +1000, Jacinta Richardson wrote:
> > >> Rev Simon Rumble wrote:
> > >>> That sounds doable.  What accounting package, if you don't mind?
> > >> We use gnucash, but SQL Ledger ( http://www.sql-ledger.org/ ) has been 
> > >> getting
> > >> press.  We were toying with the idea of moving over, but we haven't yet.
> > > 
> > > I'm a SQL Ledger user too, but there's been an interesting fork in the
> > > project in the last month to six weeks, due to the perceived lack of
> > > openness and responsiveness of SQL Ledger's author. So you might want 
> > > to check out http://www.ledgersmb.org/ as well if you're considering
> > > moving.
> > 
> > This is both an interesting and concerning twist that I was not aware 
> > of.  Do you have any further info other than the link?
> 
> There was a lot of gnashing of teeth about a perceived security hole in
> SQL-Ledger. Two guys posted notice of the hole but the author did not
> appear to do anything about it, so a couple of weeks ago they decided to
> write their own fix and fork.
> 
> The author of SQL-Ledger basically writes the entire package on his own
> under GPL and sells support and documentation to make money.
> 
> He is very active developing SQL-Ledger, but I have to say that he has
> very poor human communication skills. It will be very interesting to see
> if the fork has legs. So far I'm sticking to the SQL-Ledger version, but
> there is some very interesting chatter on the ledgersmb mailing list, so
> it will be a space worth watching. 
> 
> I guess it's a classic GPL fork situation. 

I agree with David - SQL-Ledger is a pretty impressive product for a 
one-man show, but Dieter's not been very good at managing the communication
and community sides of things, and he was supposedly alerted to this 
security issue months ago and didn't do anything until the fork happened.

The fork is likely to be a more community-oriented approach, and has 
members members of the PostgreSQL team involved, but it remains to be seen
what kind of longevity it has.

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Contracting stuff: wrap it into a company or PAYE through agency?

[Gavin Carr]

On Wed, Sep 27, 2006 at 12:28:16PM +1000, Jacinta Richardson wrote:
> Rev Simon Rumble wrote:
> > That sounds doable.  What accounting package, if you don't mind?
> 
> We use gnucash, but SQL Ledger ( http://www.sql-ledger.org/ ) has been getting
> press.  We were toying with the idea of moving over, but we haven't yet.

I'm a SQL Ledger user too, but there's been an interesting fork in the
project in the last month to six weeks, due to the perceived lack of
openness and responsiveness of SQL Ledger's author. So you might want 
to check out http://www.ledgersmb.org/ as well if you're considering
moving.

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Contracting stuff: wrap it into a company or PAYE through agency?

[Gavin Carr]

Rev Simon Rumble wrote:
>Hi folks.
>
>I'm about to start contracting and it seems it's a lot more efficient to 
>have things wrapped into a company.  I've been looking around and seeing 
>stuff about the "80/20 rule" and the like, which determines whether you 
>can apply company tax instead of personal income tax, and other stuff.
>
>So does anyone out there have any (IANAA) advice on such things?  Any 
>recommendations of accountants to get advice from?  Ideally accountants 
>who can deal with free-software ledger systems...
>
>Oh and yes, those of you wondering, I'm back in Sydney.

Do give a contrary point of view to a few others, I think doing the 
company typically _is_ worthwhile, overall. If you shop around for
accountants and are willing to do most of the book-keeping stuff 
yourself, costs can be kept pretty modest in my experience - I pay
< $1000 pa all up for accounting and company fees etc.

On the benefits side you get limited liability protection, and if you
do have multiple clients (or can stay on the right side of the ATO 
via the various 80/20 type tests) then you can retain earnings in the
company taxed at 30% max. That's not as big a deal as it was before 
the last round of tax threshold changes, but it still can be useful,
depending on how much money you make. 

As others have mentioned, super is an issue as well, but I actually
like the control DIY-super allows, for all the costs and admin
overheads.

Lots of trade-offs, of course, but I'm very happy as a P/L.

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Poor Gb network performance

[Gavin Carr]

On Tue, Aug 29, 2006 at 04:44:35PM +1000, Jeff Waugh wrote:
> 
> 
> > Speed: 1000Mb/s
> > Duplex: Full
> 
> > [   22.910730] eth0: forcedeth.c: subsystem: 01043:8141 bound to
> 
> - 8< - snip - 8< -
> 
> > eth0: Identified chip type is 'RTL8169s/8110s'.
> > eth0: RTL8169 at 0xf89ca000, 00:0f:b5:8d:59:ad, IRQ 201
> > eth0: Auto-negotiation Enabled.
> > eth0: 1000Mbps Full-duplex operation.
> 
> So you've got an Nvidia "softnic" (rather like a softmodem, so kind of icky
> if you're looking for performance) and a Realtek, which are not widely known
> for their performance. I'm not sure that explains the full extent of your
> performance problems, but it's a start.

Also, I've seen problems with the Realteks in 2.6.9 kernels, that seemed to 
go away with later 2.6 kernels (around 2.6.13, from memory). Given the debian
box is 2.6.8, that might be an issue, so you might want to try a later kernel.

The forcedeth drivers on the ubuntu side are pretty recent (and so probably 
about as good as you'll get), but you could always try the nvidia nvnet driver
for a comparison. Note that nvidia now recommend the forcedeth driver (recent
kernels only) over the nvnet one, though.

And what are you using to test this, incidentally?

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Problem in using scp

[Gavin Carr]

On Wed, Aug 09, 2006 at 04:42:13PM +1000, Michael Lake wrote:
> Hi all
> 
> I am trying to do this:
> 
> cat file | sed s/this/that/g | scp - [EMAIL PROTECTED]:/home/someplace/
> 
> As I want to replace some strings in file and copy it to another server in 
> a Makefile line. It looks like scp doesn't recognise - as "take input from 
> its stdin".
> 
> How can I do what I intend?

sed s/this/that/g file | ssh [EMAIL PROTECTED] 'cat > /home/someplace'

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] WAN simulation

[Gavin Carr]

On Thu, Jul 20, 2006 at 03:31:11PM +0930, Glen Turner wrote:
> Gavin Carr wrote:
> >Quick question - I'm wanting to do some lab-based WAN simulation i.e. 
> >have an ethernet link set up between two boxes that has wan-like 
> >latencies on it. I was thinking I could maybe just insert a linux 
> >router/forwarder into that link if I can find some means of 
> >introducing forwarding delays there - anyone know of any way to do
> >that?
> 
> Linux has the netem traffic classifier.  I use it all the
> time for exactly this scenario.
> See <http://linux-net.osdl.org/index.php/Netem>.

Thanks for all the responses. Glen's netem looks exactly what I was 
after - off to play ...

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

[SLUG] WAN simulation

[Gavin Carr]

Hi sluggers,

Quick question - I'm wanting to do some lab-based WAN simulation i.e. 
have an ethernet link set up between two boxes that has wan-like 
latencies on it. I was thinking I could maybe just insert a linux 
router/forwarder into that link if I can find some means of 
introducing forwarding delays there - anyone know of any way to do
that?

Cheers,
Gavin

-- 
Gavin Carr
Open Fusion - Open Source Business Solutions [ Linux - Perl - Apache ]
http://www.openfusion.com.au
- Fashion is a variable, but style is a constant - Programming Perl
-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] LVM Re-mounting Hell

[Gavin Carr]

On Mon, Jul 03, 2006 at 05:44:46PM +1000, Kevin Fitzgerald wrote:
> Hi All. Hoping someone can help.
> 
> First the question, then the background
> 
> Q: Can anyone help me re-mount LVM disks to a new machine
> 
> Background
> Machine 1, Fedora core 4, 3 disks in a LVM. 

What do you mean by that? Do you mean you are striping across the disks
using LVM i.e. a volume group/logical volume spanning all three disks?

If so, I suspect you are hosed. I _think_ LVM requires the entire volume
group for you to be able to bring it up on a new machine. Striping/RAID0
is dangerous in this respect - you're increasing your likelihood of 
volume failure by the number of disks in your stripe. You should only
use striped volumes for data you don't care about, or you should stripe
over an underlying mirror instead (which is RAID 10).

Cheers,
Gavin


-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [Fwd: Re: [SLUG] I so hate X windows...]

[Gavin Carr]

On Fri, May 05, 2006 at 11:32:38AM +1000, Mary Cudmore wrote:
> Re nVidia's Twinview vs Xinerama - it's my understanding that they 
> attempt to implement the same functionality, and shouldn't be used 
> together. My decision-making was based on the simplistic statement that 
> Twinview is nVidia specific and possibly faster, whereas Xinerama does 
> cooler things but might be somewhat slower. If this is true, for basic 
> cloning with an nVidia card, Twinview is the better way to go.
> 
> Anyone else have this idea?

I tend to use Twinview for dual-monitors-on-a-card, partly because (at 
least at one point) it seemed to be handle monitors with different 
resolutions better.

But FWIW, Twinview and Xinerama do seem to play fine together. One 
client has a couple of four-headed setups using 2 Nvidia cards doing
twinview over 2 monitors, and then the two twinviews stitched together
using xinerama.

Getting it working was a bit of a dog, but it's rock solid once it's
working.

Cheers,
Gavin


-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] NTP problems

[Gavin Carr]

Hi Julio,

On Wed, May 03, 2006 at 02:23:42PM +1000, Julio Cesar Ody wrote:
> But in this machine, no matter how much I want, this is what I get:
> 
> 
> # ntpq -p
> remote   refid  st t when poll reach   delay   offset  
> jitter
> ==
> hostname  220.233.180.218  3 u   23   64  3770.169  -756185 
> 80237.9

Is there anything relevant in your ntp.log file? Also check
/var/log/messages  (or whatever gentoo uses for daemon/kernel messages),
particularly for for anything about 'ntp', 'clock', 'freq', or 'tick'.
I imagine you'll at least have ntpd messages about frequency errors
exceeding the standard tolerances.

It's quite possibly a hardware issue - maybe a motherboard with a bios
issue that is losing ticks, or some other hardware (I've seen it with
some hardware raid cards) screwing up the clock ticks somehow. Do you
have any interesting hardware in that box?

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Fedora vs RH Enterprise - consultants advising to change

[Gavin Carr]

Hi Simon,

On Wed, Jan 25, 2006 at 11:53:01AM +1100, Simon wrote:
> AT the risk of starting a flamewar.I am being advised by consultants
> that I need to 'upgrade' my Fedora Core servers to RH Enterprise as it
> is 'more robust', 'better supported', 'easier to upgrade' etc etc. We
> are currently running them as our webserver (informational only - no
> transactions), mailserver and intranet webserver (this one is a bit
> slow, but just needs more RAM).
> 
> I am unaware of any major differences in the products that would require
> us to change over and start paying for what we now do for free -
> maintenance has been trivial, yum runs regularly via cron, downtime has
> been non-existent.

There are two sets of differences between FC and RHEL4 - one is that
RHEL4 is commercial software, with a price tag and commercial support,
whatever value you and your business want to place on that; if you
don't see value there, then projects like CentOS and Tao Linux 
give you robust and free (as in beer) alternatives.

The second difference is that RHEL4 is an *enterprise* OS, meaning 
longer release cycles, guaranteed product life, formal regression 
testing, certification with some hardware and software products, etc. 
FC, on the other hand, is a consumer OS, meaning it's much more 
bleeding edge, has much shorter release cycles, and (probably) a 
shorter product life.

So they _are_ different. You've got to decide whether those differences
end up making one or other more suitable for your use in your business.
Your consultants aren't necessarily being unreasonable - they're saying
that enterprise OSes (compare SLES vs. Open SuSE in the SuSE world) 
tend to be better choices for business server use than consumer ones.

That said, you've also got to weigh up the execution risk involved in 
taking nice working servers and 'upgrading' them for no particularly
immediate gain. I'd say it's not a particularly compelling case as you
present it.

Finally, one other option you might want to know about is that you can
usually do simple upgrades (i.e. not reinstalls) from FC servers to 
their closest RHEL or CentOS relative - there are various howtos 
available on the net or via the CentOS site with the details. I've 
done a few and they've all worked nicely.

Cheers,
Gavin

-- 
Open Fusion P/L - Open Source Business Solutions [ Linux - Perl - Apache ]
ph:  +612 9875 5032fax: +612 9875 4317
web: http://www.openfusion.com.au  mob: +61 403 171712
- Fashion is a variable, but style is a constant - Programming Perl
-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Linux music appliances?

[Gavin Carr]

On Sat, Dec 03, 2005 at 02:37:51PM +1100, Luke Kendall wrote:
> This is probably a dumb question ...
> I'm looking for a small, quiet, Linux-based music appliance.
> The ideal would be a small fanless PC with CD drive and hard drive,
> little 2-line LCD screen and remote, with stereo audio outputs
> to load CDs in that then get converted into MP3s on an internal hard
> drive.  So you have all your music in a handy system you can connect to
> your stereo system.

This isn't quite what you asked for, but I'm a big fan of the 
Squeezeboxes (and earlier, the SliMP3) from www.slimdevices.com. They're 
little wireless (or wired) music players that you plugin into an 
amplifier or powered speakers, and run off an open source music server 
(in Perl!) that streams your mp3/ogg/flac collection, and also handle 
Internet Radio etc. They cost US$249-299, or about $500 locally, so 
they're not cheap, but they're nice!

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] htaccess control for 200 users ?

[Gavin Carr]

On Tue, Oct 18, 2005 at 09:46:47AM +1000, Voytek wrote:
> I have Apache 1.3x running a number of vhosts, some have some htaccess
> control using maybe 10 or 15 unique usernames;
> 
> for one vhost, I'm looking at setting a 'closed shop' accessible only to
> pre-defined existing customers, like, say, oscommerce behind htaccess
> authentication;
> 
> is that a 'good idea' to look at htaccess authentication for around 200
> unique user/password ? or ?

By "htaccess authentication" you probably mean Basic Authentication,
probably with mod_auth and htpasswd files, right? The main problem with
Basic Authentication is security, since usernames and passwords are
passwd in the clear - unless you're going to do everything over SSL,
you shouldn't do it.

Scalability is a secondary issue. Since htpasswd files just use a linear
scan, even 200 users will probably start to impact performance. The docs
say [1]:

  A consequence of this is that there's a practical limit to how many 
  users you can put in one password file. This limit will vary depending 
  on the performance of your particular server machine, but you can 
  expect to see slowdowns once you get above a few hundred entries, and 
  may wish to consider a different authentication method at that time.

Something like mod_auth_dbm/mod_auth_mysql/mod_auth_ldap would be better 
for scalability, but you're still doing evil Basic Authentication, just 
faster.

You'd need something like mod_auth_tkt[2] to sidestep the problems 
with Basic Authentication if you want to avoid using SSL everywhere.


Cheers,
Gavin

[1] http://httpd.apache.org/docs/2.0/howto/auth.html
[2] http://www.openfusion.com.au/labs/mod_auth_tkt/

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Installing Perl modules on Solaris

[Gavin Carr]

On Thu, Sep 29, 2005 at 03:33:27PM +1000, saurabh shukla wrote:
> I see you're using perl 5.008004 on sun4-solaris-64int, okay.

> cc -c   -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -xarch=v8 -D_TS_ERRNO 
> -xO3 -xspace -xildoff   -DVERSION=\"1.48\" -DXS_VERSION=\"1.48\" -KPIC 
> "-I/usr/perl5/5.8.4/lib/sun4-solaris-64int/CORE"  -DDBI_NO_THREADS Perl.c
> /bin/sh: cc: not found
> make: *** [Perl.o] Error 1
>   /usr/local/bin/make  -- NOT OK
> Running make test
>   Can't test without successful make
> Running make install
>   make had returned bad status, install seems impossible
> 
> 
> I dont have cc installed on my machine however I do have gcc version 
> 3.3.2 installed. Can I change some setting to force perl to use gcc 
> instead of CC ?

Welcome to the joys of solaris! Nope, you can't use gcc to compile perl 
modules into a cc-compiled perl or the ground will open and swallow you 
up, etc. Your best bet is probably to grab the perl from sunfreeware.com, 
or to go and find yourself a cc to use. Only affects non-pure-perl 
modules, of course, but they're the best ones.

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Linux Ethernet Bridging - Is there a legitimate use?

[Gavin Carr]

On Thu, Aug 18, 2005 at 08:17:04AM +1000, Howard Lowndes wrote:
> I have been doing some reading on Linux ethernet bridging - brctl and 
> ebtables - and I can see how it could be used covertly on a network.
> 
> What I would be interested to know is any examples where it has 
> legitimate use on a fully owned and managed network that could not be 
> achieved by other means.
> 
> One that springs to mind is to extend a link beyond the 100m limit, but 
> that could be done by using an off the shelf switch.  Are there others.
> 
> I would be interested in hearing of any examples that you are able to 
> disclose...

Bridges are often useful in diagnostic or security roles where you 
want to insert a machine into a network (perhaps temporarily) without 
reconfiguring things - you can just drop it in inline and everything
Just Works.

Linux Journal also had a useful article on bridges this year:

  http://www.linuxjournal.com/article/8172

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] [OT]: Database Design Question

[Gavin Carr]

On Tue, Aug 09, 2005 at 07:35:11PM +1000, Adam W wrote:
> E.g
> The parent table is "person" and it holds common data amongst all
> people - it has primary key of personID.
> There is another table assoicated with person called "personDetails".
> This would have fields "personID", "dataType" and "dataValue". It
> might have records like the following:
> personID, dataType, dataValue
> 1,phone,123456
> 1,email,test@test.com
> 1,state,NSW
> 2,phone,987456321
> 2,state,VIC
> 3,phone,789456123
> 
> Of course this is just an example - the parent entity could be anything.
> 
> Hope someone can put a name to this sort of design. I want to research
> into this to see how people search effectively in this design and its
> performance compared to traditional methods etc etc.

I've used this type of design a number of times. I've heard it termed
'attribute tables', but I haven't been able to find much via google 
matching that term. It's a proper normalised representation though - 
both 3NF and BCNF, I believe.

As you imply, the main advantage of doing it this way is that you can
add new attributes to a object without having to change your schema - 
they just become additional rows. The disadvantage is that if you are
primarily going to want to deal with the data as a flat 'person'
record, then the joins do hit you on the performance side. Note that 
there are a number of use cases where you don't need to do that 
though - if you want to use the attributes primarily as search keys, 
for instance, then it can be a very efficient and flexible 
representation.

I confess that in one large real-world app I was involved with we
denormalised this part of the data model and did both - had the most 
common attributes as columns on the person record, _and_ as rows in 
the person attribute table. Slightly evil, but it worked out very 
well. ;-)

I'd be interested if you do turn up any discussion of this Adam.

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Looking for lazy way out

[Gavin Carr]

On Wed, Jun 22, 2005 at 04:56:13PM +1000, Gavin Carr wrote:
> On Wed, Jun 22, 2005 at 04:33:53PM +1000, Simon wrote:
> > I have uploaded a web based  CD to our Moodle setup, but all the links
> > are broken. In true sloppy MS style most of the filenames are in
> > uppercase whereas the html files refer to them in lower case. The
> > underlying webserver is Apache, (thought mod-speling might help from
> > some reading but it is already installed). I am looking for the lazy way
> > out of going through and editing all the links in the html files or
> > renaming all the disc files, is there a solution? NB there are hundreds
> > of files and corresponding links :-(
> 
> Since laziness is one of the cardinal virtues of the perl-hacker, try:
> 
>   perl -MFile::Find -e 'find sub { rename $_, lc $_ if -f $_ && $_ =~ 
> m/\.html?$/i }, "."'
> 
> from the top of your cd directory tree. This will lowercase all file names 
> ending in .htm, .html, .HTML, .HTM, etc., so be careful that's exactly 
> what you want!

BTW, if you're familiar with find(1), you can do the same thing with that:

  find . -type f -iregex '\..*html?$' -print

to check the file list, and then:

  find . -type f -iregex '\..*html?$' -print0 | perl -n -0 -e 'rename $_, lc $_'

to do the renames. Note those '0's are zeros, not letter 'O's.

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Looking for lazy way out

[Gavin Carr]

Hi Simon,

On Wed, Jun 22, 2005 at 04:33:53PM +1000, Simon wrote:
> I have uploaded a web based  CD to our Moodle setup, but all the links
> are broken. In true sloppy MS style most of the filenames are in
> uppercase whereas the html files refer to them in lower case. The
> underlying webserver is Apache, (thought mod-speling might help from
> some reading but it is already installed). I am looking for the lazy way
> out of going through and editing all the links in the html files or
> renaming all the disc files, is there a solution? NB there are hundreds
> of files and corresponding links :-(

Since laziness is one of the cardinal virtues of the perl-hacker, try:

  perl -MFile::Find -e 'find sub { rename $_, lc $_ if -f $_ && $_ =~ 
m/\.html?$/i }, "."'

from the top of your cd directory tree. This will lowercase all file names 
ending in .htm, .html, .HTML, .HTM, etc., so be careful that's exactly 
what you want!

Cheers,
Gavin

-- 
Open Fusion P/L - Open Source Business Solutions [ Linux - Perl - Apache ]
ph:  +612 9875 5032fax: +612 9875 4317
web: http://www.openfusion.com.au  mob: +61 403 171712
- Fashion is a variable, but style is a constant - Programming Perl
-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

[SLUG] Stas Bekman at Sydney Perl Mongers

[Gavin Carr]

Greetings Sluggers,

In addition to the session at Sydney University already announced here,
Stas Bekman has kindly agreed to do a mod_perl 2 workshop for the Sydney 
Perl Mongers (http://sydney.pm.org/) on this Thursday evening (2 June), 
starting 5:30 pm till at least 9:30 pm. We have 10 spots available we'd 
like to offer to SLUG, so if you'd like to attend please reply to me with 
a short paragraph about why you'd like to come. Deadline end of today 
(Tuesday 31st May).

Stas is the author of O'Reilly's Practical mod_perl, and has been one of
the core developers on both mod_perl 1 and 2. He's also apparently a
great presenter, so we're in for a treat!

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Oh goodie - again....(sober.q strikes again)

[Gavin Carr]

On Mon, May 23, 2005 at 08:45:52AM +1000, Peter Rundle wrote:
> Gavin Carr wrote:
> 
> 
> >Here's a spamassassin check that was posted on one of the rulesemporium
> >forums that is working well for me here, in case its of use to anyone:
> > 
> >
> 'scuse my ignorance but can I just add the lines to 
> /etc/mail/spamassassin/local.cf or is there a better way?
> 
> Google returns a lot of info on writing your own rules but it's not 
> clear to me on how to include those rules into
> spamassasin.

Yep, just cut and paste into local.cf (and restart spamd if you're using
that). That's the definitive place for local rules and customisations.

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Oh goodie - again....(sober.q strikes again)

[Gavin Carr]

On Sat, May 21, 2005 at 02:14:04PM +1000, James Gray wrote:
> On Sat, 21 May 2005 09:48 am, Voytek wrote:
> > 
> >
> > > http://www.techweb.com/wire/security/163106139
> > >
> > > Hold onto your mail servers folks - looks like Monday could be ground-hog
> > >
> > > I can catch some samples in the wild to modify the spamassassin filters.
> > > Will let people know if I (or other admins I know) manage to achieve
> > > this.
> >
> > isn't it simpler to DISCARD with a header check ?
> > that's what I done with the current crop
> 
> Indeed, you are correct - discarding at the MTA level is by far the better 
> option.  However, the method for doing this varies between 
> Postfix/Sendmail/Exim/Qmail/etc.  SpamAssassin is pretty universal to all of 
> these and if there's a Perl regex available, most admins will be able to mash 
> that into a header check for their particular MTA (which is what I've done 
> for my machines).

Here's a spamassassin check that was posted on one of the rulesemporium
forums that is working well for me here, in case its of use to anyone:

# German sober.q test
header SOBER_Q_SUBJECT Subject =~ /4,8 Mill\. Osteuropaeer durch Fischer-Volmer 
Erlass|Auf Streife durch den Berliner Wedding|Auslaender bevorzugt|Deutsche 
Buerger trauen sich nicht \.\.\.|Auslaenderpolitik|Blutige 
Selbstjustiz|Deutsche werden kuenftig beim Arzt ?abgezockt|Paranoider 
Deutschenmoerder kommt in Psychiatrie|Du wirst zum Sklaven gemacht!!!|Dresden 
1945|Massenhafter Steuerbetrug durch auslaendische Arbeitnehmer|Gegen das 
Vergessen|Tuerkei in die EU|Hier sind wir Lehrer die einzigen 
Auslaender|Multi-Kulturell = Multi-Kriminell|Verbrechen der deutschen 
Frau|S\.O\.S\. Kiez! Polizei schlaegt Alarm|Transparenz ist das Mindeste|Trotz 
Stellenabbau|Vorbildliche Aktion|Augen auf|Du wirst ausspioniert \.\.\.\.!|Volk 
wird nur zum zahlen gebraucht!|60 Jahre Befreiung: Wer feiert 
mit\?|Graeberschaendung auf bundesdeutsche Anordnung|Schily ueber 
Deutschland|The Whore Lived Like a German|Turkish Tabloid Enrages Germany with 
Nazi Comparisons|Dresden Bombing Is To Be Regretted Enormously|Armenian 
Genocide Plagues Ankara 90 Years On/i
describe SOBER_Q_SUBJECTContains a known Sober.Q subject
score SOBER_Q_SUBJECT 10.0

I believe it was derived from the Symantec analysis of the virus. 
Sorry - don't have the URL handy right now.

Cheers,
Gavin

--
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] web site analysis/statistics

[Gavin Carr]

On Mon, May 16, 2005 at 02:37:31PM +1000, DaZZa wrote:
> On Mon, 16 May 2005, David wrote:
> 
> > What's the latest and greatest software for web site stats?
> >
> > Just the basics... who hits and how many, where from, number pages
> > per visit, referrer
> 
> http://www.awstats.org
> 
> Gives some pretty impressive information et al from most web server logs.
> 
> It's even free {GPL}.

awstats is great, but it's had a few security holes lately. Just be 
careful that when you install it you secure it properly with _at least_
Basic Authentication passwords, and preferably something better.

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Tracking down qmail-smtpd err message output

[Gavin Carr]

On Wed, Mar 23, 2005 at 03:03:33PM +1100, Peter Rundle wrote:
> > Linux processes, including MTAs, write to 'syslogd' daemon for logging.
> [snip]
> 
> Thanks for your reply, pardon my ignorance but I'm not quite sure how to 
> make the connection between the program and syslog.

In contrast to lots of other daemons, qmail does not directly support 
syslog, and just dumps logs to stderr and expects you to do something
with them. One option is to use qmail's splogger to redirect the logs
to syslog, but I've never done it for qmail-smtpd, so can't be more
specific. Your other main option is probably to use djb's daemontools
instead of init.d, which you'll find lots of documentation on at the
standard qmail sites:

  http://www.lifewithqmail.org/
  http://www.qmailrocks.org/
  http://www.qmail.org/


Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

[SLUG] PDA/connectivity recommendations

[Gavin Carr]

I need to get hold of a pda-type form factor device for remote access to
linux servers from anywhere (AU), and was wondering if anyone has 
recommendations or success stories. I'm thinking about a Zaurus / bluetooth 
/ cellphone GPRS combination, and would love to hear of anyone doing that, 
or of any good alternatives out there. I'd prefer a linux device, of course,
but would also like to hear any "why I didn't end up going with linux" 
stories ...

Cheers,
Gavin


-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] safe(ish) single-login from website

[Gavin Carr]

On Wed, Feb 16, 2005 at 11:14:33AM +1100, Rob Sharp wrote:
> You may run into all sorts of privacy issues if you start sending user
> passwords unencrypted over a URL... Of course, this is when the
> assymetric excryption key mentioned earlier becomes useful!
> 
> On Wed, 16 Feb 2005 11:07:11 +1100, Taryn East <[EMAIL PROTECTED]> wrote:
> > * Gavin Carr <[EMAIL PROTECTED]> spake thus:
> > > Try mod_auth_tkt: http://www.openfusion.com.au/labs/mod_auth_tkt/
> > 
> > this sounds really like a good option but...
> > 
> > > https://www.taryn.com/cgi-bin/ticket.cgi?user=foo;pass=bar

That would be where the 's' in 'https' comes in handy. :-)

-G

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] safe(ish) single-login from website

[Gavin Carr]

On Wed, Feb 16, 2005 at 11:07:11AM +1100, Taryn East wrote:
> * Gavin Carr <[EMAIL PROTECTED]> spake thus:
> > Try mod_auth_tkt: http://www.openfusion.com.au/labs/mod_auth_tkt/
> 
> this sounds really like a good option but...
> 
> > https://www.taryn.com/cgi-bin/ticket.cgi?user=foo;pass=bar
> 
> this looks like exactly the sort of thing that I can't do anymore - which
> is prompting me to make these changes...

I don't think so. I think what you are talking about is passing basic
authentication parameters in the url, which you have to do every request,
often in the clear, and is prone to leakage via referrals.

These are just CGI parameters, over SSL, done once. There's no leakage 
because all you're getting back is a text file. You could equally well
use a POST here if doing a GET makes you nervous.

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] safe(ish) single-login from website

[Gavin Carr]

On Tue, Feb 15, 2005 at 04:41:23PM +1100, Taryn East wrote:
> the issue is that our business allows some of our website to be viewable
> through the website of some of our "channel partners". These channel
> partners have a login to our website to allow them to do this.
> 
> However, the channel partners have customers that only have a login to
> the channel-partner websites... and the channel partners don't want to
> directly give them the login to our site, but do want the pages
> displayed (generally using yucky frames... but hey).
> 
> ok, now they aparrently used to do this by having a url with the
> username/password in it (ie using "basic" http authentication with the
> login details as parameters).

Try mod_auth_tkt: http://www.openfusion.com.au/labs/mod_auth_tkt/

mod_auth_tkt is a drop-in replacement for basic authentication that
uses MD5 tickets to authenticate users. Tickets are usually provided 
via cookies, but in your case it's probably easier to provide them via 
your url (like you were doing for username/password before). (This is
better because your referring site is going to be on a different domain 
than yours, which can screw cookies up.)

So in your context you could, for instance:

- modify the standard mod_auth_tkt login CGI to return the ticket it
  produces for valid users as text output, rather than setting a cookie with it

- get your channel partners to login via that CGI periodically and save 
  the ticket to a text file e.g.

wget -O ticket https://www.taryn.com/cgi-bin/ticket.cgi?user=foo;pass=bar

- get your channel partners to include that ticket on the initial referral
  to your site in the url e.g.
 
http://www.taryn.com/partners/index.html?auth_tkt=ticketgoeshere

Easy! ;-)

Cheers,
Gavin

-- 
Open Fusion P/L - Open Source Business Solutions [ Linux - Perl - Apache ]
ph:  +612 9875 5032fax: +612 9875 4317
web: http://www.openfusion.com.au  mob: +61 403 171712
- Fashion is a variable, but style is a constant - Programming Perl
-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] text to web page, adding ?

[Gavin Carr]

On Fri, Feb 11, 2005 at 12:03:12PM +1100, Voytek wrote:
> I need to place contents of word files on a web pages;
> I've sorted catdoc/charset to output correct codepage text files,
> 
> catdoc -scp1250 -d8859-2 381.rtf > 381.php
> 
> but, then I need to insert '' after every paragraph

No, you need to wrap   around every paragraph, and end up 
with nice compliant xhtml. This is the 21st century, after all. ;-)

> is there any tool that will help me here ?

Do a search for 'text html' on search.cpan.org - there's about a 100
perl modules that do different variants on this. HTML::FromText,
HTML::TextToHTML, and Text::Decorator are probably useful places to
start.

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Weird masquerading problem

[Gavin Carr]

On Tue, Jan 25, 2005 at 10:38:13AM +1100, Peter Rundle wrote:
> > The good packets do, the bad packets don't. If I add a LOG message to the
> >mangle POSTROUTING chain, though (which comes before the nat POSTROUTING
> >chain), both sets are logged there. Truly weird.
> 
> I assume that the -j LOG is at the end of the mangle chain, I.E we are sure 
> the packets are exiting that chain, very weird. 

Yeah, nothing in the mangle chain except the LOG rule. 

> It appears that iptables 
> somehow decides that the 202.125.42.141.4569:  udp packets don't need to 
> traverse the nat chain.

Indeed. :-(

> Hard to know what to try next, check the route table 202.125.x.x isn't a 
> seperate route is it?

Nope, nothing in the route table but the two local segments and the
default gateway.

I've pinged the netfilter list to see if anyone else has seen anything
like this before. Also downgraded iptables to stock woody 1.2.6, but it's
exactly the same. I'll let you know if they come up with anything.

-G

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Weird masquerading problem

[Gavin Carr]

Hi Peter,

On Tue, Jan 25, 2005 at 08:15:26AM +1100, Peter Rundle wrote:
> >For some reason these packets just ain't traversing the iptables
> >chains right. If I zero the counts in the nat table I can see
> >they never hit the POSTROUTING chain where the SNAT is happening.
> >My netcat packets, from exactly the same box, do. :-/
> 
> How many rules are in that chain *before* the masquerade statement? Is it 
> possible that the packets are matching a rule and exiting the chain?

Good suggestion, but no, there's nothing else in that chain.

> What happens if you put a -j LOG rule at the very beginning of the 
> POSTROUTING chain that matches all packets and see if the packets enter the 
> chain.

The good packets do, the bad packets don't. If I add a LOG message to the 
mangle POSTROUTING chain, though (which comes before the nat POSTROUTING
chain), both sets are logged there. Truly weird.

I'm going to ask on the netfilter list, since it's increasingly looking
like some iptables bogosity.

Thanks for your suggestions.

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Weird masquerading problem

[Gavin Carr]

On Mon, Jan 24, 2005 at 10:04:08PM +1100, David Kempe wrote:
> >I've got a very standard masquerading firewall between my internal 
> >network and the outside, done with smoothwall/iptables on Debian
> >Woody. Works beautifully, and has for a while.
> 
> is that shorewall or smoothwall?

Sorry, you're right - shorewall, 2.0.10, latest from backports.

For some reason these packets just ain't traversing the iptables
chains right. If I zero the counts in the nat table I can see
they never hit the POSTROUTING chain where the SNAT is happening.
My netcat packets, from exactly the same box, do. :-/

Anyone have any deep iptables debugging magic to share?

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

[SLUG] Weird masquerading problem

[Gavin Carr]

Hi all,

I've got a very standard masquerading firewall between my internal 
network and the outside, done with smoothwall/iptables on Debian
Woody. Works beautifully, and has for a while.

The problem is I've got one app (asterisk) on an internal server
that is connecting via UDP to a particular host, and the connection 
is failing. Same app connects fine to other hosts, and network
connectivity to the host looks fine - I can ping, I can netcat to
the udp port, etc.

Debugging this, it looks like the outward connection is 
not getting masqueraded properly - on the external interface on the
firewall a tcpdump shows:

04:35:18.203836 10.10.10.1.4569 > 202.125.42.141.4569:  udp 12 (DF) [tos 0x10] 
04:35:18.706606 10.10.10.1.4569 > 202.125.42.141.4569:  udp 12 (DF) [tos 0x10] 
04:35:18.707008 10.10.10.1.4569 > 202.125.42.141.4569:  udp 26 (DF) [tos 0x10] 
04:35:19.344564 203.213.47.14.4569 > 216.118.117.46.4569:  udp 12 (DF) [tos 
0x10] 
04:35:19.586808 216.118.117.46.4569 > 203.213.47.14.4569:  udp 12 (DF) [tos 
0x10] 
04:35:19.587876 203.213.47.14.4569 > 216.118.117.46.4569:  udp 12 (DF) [tos 
0x10] 

The first three lines are bad, where the 10.10.10.1 is my internal
address that isn't masqueraded. The next three lines are good to 
a different host (216.118.117.46), where 203.213.47.14 is my
masqueraded address. 

A 'nc -u 202.125.42.141 4569' from that same server, though, works
fine, and masquerades correctly:

04:41:59.179729 203.213.47.14.32846 > 202.125.42.141.4569:  udp 1 (DF) [tos 
0x10] 
04:41:59.332267 203.213.47.14.32846 > 202.125.42.141.4569:  udp 1 (DF) [tos 
0x10] 
04:41:59.509531 203.213.47.14.32846 > 202.125.42.141.4569:  udp 1 (DF) [tos 
0x10] 

Been batting my head against this for hours now - anyone seen
anything like this or have any pointers on what to try next?

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Most valuable free/OSS software that doesn't exist?

[Gavin Carr]

On Mon, Nov 15, 2004 at 02:20:52PM +1100, Jeff Waugh wrote:
> > OTOH, most of the server environments I encounter are much smaller than
> > that; typically 5-15 machines, and usually mostly or all Linux. In this
> > kind of environment I've found cfengine to be overkill - the learning
> > curve for the local admins is just too much to justify the benefits.
> 
> See, 5-15 machines says "one admin" to me. If that one admin has a tightly
> honed deployment / change management solution (for example, cfengine), then
> she won't be overwhelmed when projects (not maintenance or support) start to
> multiply, or when four machines die at once.
> 
> Learning cfengine is challenging, but even for small deployments, there's a
> pretty good saving to be had.

But from cfengine in particular, or from having a "tightly honed deployment / 
change management solution" in general? I'm all in favour the latter; it's 
the cfengine instance of that solution that I've found a hard sell.

cfengine is sendmail in this space - does everything, but is a dog to learn. 
I want a postfix.

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Most valuable free/OSS software that doesn't exist?

[Gavin Carr]

On Mon, Nov 15, 2004 at 01:15:39PM +1100, Jamie Wilkinson wrote:
> This one time, at band camp, James Gregory wrote:
> >I've also heard of people storing all of /etc in version control for
> >this purpose. In my opinion it would be unnecessary if you kept your
> >cfengine stuff in a source control system, but it would give you that
> >absolute confidence that you could roll back and forward, even if
> >cfengine failed.
> 
> cfengine vs all-etc-in-vc is like procedural vs functional programming;
> one of them lets you describe in detail how the operation is going to be
> done, the other one lets you describe the problem.  That's probably not a
> good analogy, but with cfengine I can describe the goal state of the system
> and have cfengine sort out what needs to be done to get it there.  

Except that that assumes describing the goal state is about the same complexity
(or easier) as making the changes directly. In my playing around with cfengine
I've found the learning curve and the extra layer of indirection mostly 
annoying, rather than helpful. Usually I know the changes I want to make in
/etc; it's faster just to go ahead and do it than figure out how to convince
cfengine to do the same thing.

> Editing
> etc and rolling out changes also means handling heterogenous systems adds
> complexity to the system you use; cfengine effectively "factors out" the
> common parts and lets you describe the small changes meaningfully.

True enough - where cfengine really seems to shine is in large heterogenous
networks (e.g. >= 25 machines). OTOH, most of the server environments I 
encounter are much smaller than that; typically 5-15 machines, and usually 
mostly or all Linux. In this kind of environment I've found cfengine to be
overkill - the learning curve for the local admins is just too much to 
justify the benefits.

I'm currently experimenting with a etc-in-vc model instead (though storing
only the deltas from base OS versions, not the full etc tree), and using 
arch branches to manage machine classes, so that changes get propogated via
inheritance down the tree i.e.

  etc--all # Common stuff
   etc--web# Webserver configs
   etc--app# Appserver configs
   etc--db # Database configs
  etc.
  
So far it's working pretty well, and much more fun that battling cfengine
indeterminacy. ;-)   

Cheers,
Gavin


-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] OT: Sydney Perl Mongers

[Gavin Carr]

Hi Michael,

Sydney.pm is, er, quiescent - there hasn't been a meeting for ages.
The man to hassle about it is [EMAIL PROTECTED], which should be 
working I think - how did you try to "contact the group"?

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Passing multiple args in Perl, with one arg being an array of chars

[Gavin Carr]

On Fri, Oct 15, 2004 at 02:10:07PM +1000, Peter Rundle wrote:
> Sluggers,
> 
> As a still wet behind the ears perler, I've googled myself into a major 
> state of confusion.
> 
> I want to pass two char strings and an array of char strings to a sub 
> routine in perl. How do I do that?
> 
> I.e
> 
> $name="Peter";
> @colours=("red",green","blue");
> mysub($name,@colours);
> 
> sub mysub()
> {
>   my($nam) = shift @-;
>   my(@col) = shift @_;
> 
>   print $nam
>   foreach $c (@col)
>   {
> print "$nam:$c";
>   }
> }
> 
> Ok so this doesn't work but what is the correct syntax?

my ($nam, @col) = @_;

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

[SLUG] Debian installed package test

[Gavin Carr]

Debian-dudes,

Is there a cleaner way of testing whether a package is installed on
debian than something like:

  dpkg -l $PKG | grep -q ^ii

Cheers,
Gavin
-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Doors and sockets

[Gavin Carr]

On Wed, Aug 11, 2004 at 03:23:49PM +1000, Rowling, Jill wrote:
> Anyone know if doorfs has been ported to Linux?

Expermental version: http://www.rampant.org/doors/

> Or, for that matter, why are doors supposed to be more efficient than
> sockets?

Doors are local only, aren't they?

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] perl POD & style sheets

[Gavin Carr]

On Thu, Jul 22, 2004 at 09:41:03PM +1000, Ken Foskey wrote:
> Is there an easy way to assign a style sheet to the html that is
> generated.  I have hunted but google turns up zero.

Generated how? pod2html should accept a --css arg, IIRC.

Cheers,
Gavin

-- 
Open Fusion Pty. Ltd.[ Linux - Web - Databases - Security ]
Open Source Business Solutions http://www.openfusion.com.au
p: 02 9875 5032f: 02 9875 4317   m: 0403 171712
- Fashion is a variable, but style is a constant - Programming Perl
-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] OT: Perl: CGI.pm

[Gavin Carr]

On Mon, Jul 05, 2004 at 12:53:38PM +1000, Michael Kraus wrote:
> Thanks for correcting my terms... Yes, URI escaped... But is there a way
> of using the two functions together?
> 
> (Or would I need to use an intermediate file?)

Sorry, I thought you were asking how to unescape the values.

There's no unescaped version of save_parameters, so you'd have to use that
intermediate file, or just do it yourself - isn't it just this:

  foreach my $p ($q->param) {
printf $filehandle "%s=%s\n", $p, $q->param($p);
  }

since CGI handles all the unescaping for your within param.

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] OT: Perl: CGI.pm

[Gavin Carr]

Hi Michael,

On Mon, Jul 05, 2004 at 12:21:09PM +1000, Michael Kraus wrote:
> Using the Perl CGI module is there a function similar to
> save_parameters() (save if using OO) that decodes the HTML'ised text?
> (Ie. Will print "[EMAIL PROTECTED]" rather than
> "Email=mkraus%40wildtechnology.net")
> 
> I'm wanting to save the values of forms without having to have separate
> scripts for different forms, in a human-readable fashion.

Not in CGI (I think it only handles html escaping, not unescaping). You probably
want URI::Escape's uri_unescape here, since form parameters are URI escaped, not
HTML escaped (which is & -> & style - for that you'd use HTML::Entities).

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Help Configuring CPAN and wget Issue on Solaris !

[Gavin Carr]

On Wed, Jun 30, 2004 at 12:39:05PM +1000, Louis wrote:
> [Louis] I have tried what you said and it still won't work. I paste below
> what I did based on what you said. Please help.

Does wget work from the commandline now with your http_proxy and ftp_proxy
environment variables set? Can you wget 

  http://mirror.aarnet.edu.au/pub/perl/CPAN/authors/01mailrc.txt.gz

When you go back into the cpan shell, is your urllist still okay? What does
'o conf' show?

If not, try making those changes you made, then doing 'o conf commit', 
and then exiting and going back in. Check your settings again, and then 
try the install.

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Help Configuring CPAN and wget Issue on Solaris !

[Gavin Carr]

On Wed, Jun 30, 2004 at 10:40:13AM +1000, Louis wrote:
> Note: To connect to http, ftp I use a proxy server. I specified it here
> with format :port number . Anyway please have a look, and let
> me know why I cannot connect at all.
> 
> I've also tried wget for ftp, http to get files from the same server, and
> this as well won't work. Output is below after CPAN.

Two problems: 

1. You want to use 'http://152.76.224.188:3128' for your proxy settings, not 
the bare hostname and port (this is probably also your wget issue)

2. CPAN is trying to talk proper ftp to your proxy, when you probably need 
it to use ftp-over-http instead (since you're using a http proxy, not a 
proper ftp one). One way to hack around it is to try and get your cpan init
to complete, even with bogus settings, and then setup your urllist manually,
using something like:

  o conf urllist [ http://mirror.aarnet.edu.au/pub/perl/CPAN/

(from in 'perl -MCPAN -e shell'). Once you get http urls in your urllist 
your proxy should work as you'd expect.

Cheers,
Gavin

--
Open Fusion Pty. Ltd.[ Linux - Web - Databases - Security ]
Open Source Business Solutions http://www.openfusion.com.au
p: 02 9875 5032f: 02 9875 4317   m: 0403 171712
- Fashion is a variable, but style is a constant - Programming Perl

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] arch "rm foo; cvs up foo" equiv

[Gavin Carr]

On Thu, Jun 24, 2004 at 10:27:59AM +1000, Benno wrote:
> On Thu Mar 18, 2004 at 13:32:50 +1100, Benno wrote:
> >Sorry this is probably a dumb question, but what is the 
> >"arch way" of reverting a file back to its latest copy.
> >
> >In cvs I do:
> >
> >rm foo; cvs up foo
> 
> Following up my own question...
> 
> The answer is that tla can't do this. 

Not quite true - this works fine:

  cp `tla file-find foo` .

(assuming foo is a unique name, of course)

The aba version seems better on the fingers though.

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] HA PostgreSQL

[Gavin Carr]

On Thu, Jun 17, 2004 at 08:24:44PM +1000, Stuart Guthrie wrote:
> No but be sure to post to the list if you get some goodies. I'm also
> interested..
> 
> I found this postgres db reference, no idea of it's effectiveness:
> 
> http://www.commandprompt.com/entry.lxp?lxpe=304

Yep, Command Prompt's Replicator seems to be the most mature of the 
asynchronous replicators out there. The others we're considering 
are eRserver, which is now open sourced with PostgreSQL as of 7.3.x,
and the new kid on the block, Slony, which looks very nice but is
only in beta right now, release date 1 July:

  http://gborg.postgresql.org/project/slony1/projdisplay.php

> What's the language? 

Java, Weblogic.

> There are splitters in JDBC terms that enable
> database updates to be replicated automagically. Budget solutions and
> not-so budget solutions are available.
> 
> http://www.budget-ha.com/ha7.jsp
> http://c-jdbc.objectweb.org/
> http://cocoonhive.org/articles/jboss/20031125/JBoss32-hajms.html

These might be interesting alternatives to doing replication at the pg 
level then - thanks for the pointers. 

But who's actually doing/done this stuff?

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

[SLUG] HA PostgreSQL

[Gavin Carr]

Hi all,

Do any sluggers know of anyone using PostgreSQL in any high availability 
modes? I'm interested in both lan-based cluster configurations or using 
more wan-based disaster-recovery-type replication? 

I've got a client who would love to use PostgreSQL in place of Oracle on 
a project with some nasty availability requirements, but we haven't been 
able to find too many reference sites for this kind of thing.

Any pointers much appreciated. Aussie ideally, but I'd take anything at
this point.

Cheers,
Gavin


-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Re: perl gurus

[Gavin Carr]

On Mon, May 31, 2004 at 12:42:50PM +1000, Ashley Maher wrote:
> In a 1D array I'd order using sort. This is 2D. My brain wave was as 
> this was a 2D array write a routine to work through the 2D and insert 
> the new data in the correct position and return.

Still use sort e.g.


  #!/usr/bin/perl -wT

  use strict;
  use Data::Dumper qw(Dumper);

  my @original = (
[ "3", "test1" ],
[ "2", "test2" ],
[ "7", "test3" ],
[ "9", "test4" ],
  );
  my @new = qw(6 testing);

  # Add to original
  push @original, [EMAIL PROTECTED];

  # Sort
  my @sorted = sort { $a->[0] <=> $b->[0] } @original;

  print Dumper [EMAIL PROTECTED];


Cheers,
Gavin

-- 
Open Fusion P/L - Open Source Business Solutions [ Linux - Perl - Apache ]
ph:  02 9875 5032fax: 02 9875 4317
web: http://www.openfusion.com.aumob: 0403 171712
- Fashion is a variable, but style is a constant - Programming Perl
-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] perl gurus

[Gavin Carr]

Hi Ashley,

On Mon, May 31, 2004 at 10:04:33AM +1000, Ashley Maher wrote:
> sub insert_order_array {
> my(
>   @new_value,
>   $new_value,
>   @current_array,
>   $current_array,
>   @ordered_array,
>   $i,
>   $j,
>   );
> 
> print "\nIn subroutine \n\n";
> 
>   @new_value = ();
>   @current_array = ();
>   @ordered_array = ();
> 
>   $new_value = $_[0];
> print "\nthe new array values\n";
> print $new_value->[0];
> print "\t";
> print $new_value->[1];
> print "\n";
> 
> #
> print "\ncurrent array\n";
>   $current_array = $_[1];
> 
> # print the whole thing one at a time
> for $i ( 0 .. $#current_array ) {
>for $j ( 0 .. $#{$current_array[$i]} ) {
>print "element $i $j is $current_array[$i][$j]\n";
>}
> }

Your problem is here, inside the subroutine - you're not passing an
array (@current_array), you're passing a reference to it, so there's
another layer of indirection required. So your print loop becomes:

  # print the whole thing one at a time
  for $i ( 0 .. $#$current_array ) {
 for $j ( 0 .. $#{$current_array->[$i]} ) {
 print "element $i $j is $current_array->[$i][$j]\n";
 }
  }

and you'd return the array like:

  return @$current_array;


Does that help?

Cheers,
Gavin

-- 
Open Fusion P/L - Open Source Business Solutions [ Linux - Perl - Apache ]
ph:  02 9875 5032fax: 02 9875 4317
web: http://www.openfusion.com.aumob: 0403 171712
- Fashion is a variable, but style is a constant - Programming Perl
-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] CSS and web content management

[Gavin Carr]

On Wed, May 26, 2004 at 03:00:35PM +1000, Peter Rundle wrote:
> following on from the thread on building and maintaining a school web site, 
> I'd like to ask the collective wisdom for input on web site content 
> development and management.
> 
> Last time I was involved in this stuff web sites used frames, these seemed 
> to have fallen from favour and most sites appear to be using cascading 
> style sheets to achieve a header, navigation and main page look and feel. 
> These sites appear to be made up of distinct html pages which have common 
> areas. Thus a bookmark of a particular page works properly unlike with 
> frames.
> 
> However, what sort of techniques are being used to put these pages 
> together? I played around and managed to successfully create such a site by 
> "hand" but obviously unless the process of merging the main content of the 
> page with the header and the navigate  areas was automated this would 
> become painful and lead to typo's etc.
> 
> So are most sites generating such pages dynamically at http get time, or 
> are there software packages available to put such content together (Linux 
> based of course)?

You might want to check out ewok (http://www.openfusion.com.au/labs/ewok),
which is a small-medium CMS built on top of Embperl. One of my goals with
ewok was that it would be easy to publish these kind of static html sites - 
you can run ewok on a real box somewhere (e.g. your linux box at home), and 
then publish the whole site as static html that you upload to your cheap
and nasty hosting provider, no bells or whistles required. 

It requires apache 1, mod_perl, Embperl, and a couple of other perl 
modules - no database required. Like most others, it's got a bit of a 
learning curve to get it humming just the way you want.

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Postgresql database replication

[Gavin Carr]

On Thu, Apr 08, 2004 at 11:48:19AM +1000, Jamie Wilkinson wrote:
> This one time, at band camp, Howard Lowndes wrote:
> >Is anyone doing database replication from Postgresql?
> 
> There is a bit of software called Mammoth which I believe has been
> donated to the postgresql project, but that's all I know about it.

Almost. Mammoth is a commercial release of postgresql, and they have a replicator
option, which looks quite nice, but is non-free (although source available):

  http://www.commandprompt.com/entry.lxp?lxpe=304

We spent some time looking at postgresql replication again recently, and it
depends a lot on what you're wanting it for - there's at least this many
kinds of driver:

* HA/Failover (roll over to replica, usually LAN-based)
  (Master/slave, sync or async)
* Load balancing (use a number of smaller boxen)
  (Multi-master, sync or async + conflict resolution)
* Data warehousing (read replicas for insane read queries)
  (Async master/slave)
* Disaster recovery (replication to remote WAN-based servers)
  (Async master/slave or multi-master)
* Mobile servers (laptops!)
  (Async multi-master + conflict resolution)

Most of the current postgresql replication options (including the
commandprompt one) seem to handle async master-slave stuff reasonably
well, but multi-master and/or synchronous versions are non-existent
or experimental.

If that's what you need, then you probably want to go the shared
storage route instead of the application-level replication route, 
whether hardware or software based.

Cheers,
Gavin

--
Open Fusion P/L - Open Source Business Solutions [ Linux - Perl - Apache ]
ph:  02 9875 5032fax: 02 9875 4317
web: http://www.openfusion.com.aumob: 0403 171712
- Fashion is a variable, but style is a constant - Programming Perl

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Linux/Mozilla / internet banking

[Gavin Carr]

On Wed, Apr 07, 2004 at 08:19:45PM +1000, Martin Ellison wrote:
> St George Bank appears to no longer accept Mozilla on Linux.
> 
> 1. any experience? (confirmation, work-arounds)
> 
> 2. alternatives?

CBA? NAB? ANZ? Westpac? :-)

-G

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Squirrelmail Q

[Gavin Carr]

On Tue, Apr 06, 2004 at 09:05:24AM +1000, Grant Parnell wrote:
> On Fri, 2 Apr 2004 [EMAIL PROTECTED] wrote:
> > - is there a plug in to use Squirrelmail to retrive email from a remote
> > pop server ?

SM has the mail_fetch plugin which does what you want.

> If squirrelmail doesen't have one itself, fetchmail will do the job.

Depends. We use mail_fetch on a shared imap server to allow users (who 
don't have shell access) to pop mail into their imap mailbox from 
arbitrary pop accounts. That's non-trivial with fetchmail, for instance.

Cheers,
Gavin

-- 
Open Fusion P/L - Open Source Business Solutions [ Linux - Perl - Apache ]
ph:  02 9875 5032fax: 02 9875 4317
web: http://www.openfusion.com.aumob: 0403 171712
- Fashion is a variable, but style is a constant - Programming Perl
-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] clamav on RH7.3

[Gavin Carr]

On Wed, Mar 24, 2004 at 10:33:36PM +1100, Del wrote:
> Does anyone have compiled RPMs for clamav on Red Hat 7.3?
> 
> The current src/SRPM won't compile because it requires too
> new a version of automake, libwrap, etc.

http://dag.wieers.com/packages/clamav/

Cheers,
Gavin

-- 
Open Fusion P/L - Open Source Business Solutions [ Linux - Perl - Apache ]
ph:  02 9875 5032fax: 02 9875 4317
web: http://www.openfusion.com.aumob: 0403 171712
- Fashion is a variable, but style is a constant - Programming Perl
-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] updates to RH7.3 ?

[Gavin Carr]

On Wed, Mar 17, 2004 at 10:14:02PM +1100, John Clarke wrote:
> On Wed, Mar 17, 2004 at 10:16:32 +1100, [EMAIL PROTECTED] wrote:
> 
> > now that RedHat stopped support for RH7.3, are there any sources to keep
> > RH7.3 current ?
> 
> fedoralegacy.org.

www.progeny.com, US$5/month

Cheers,
Gavin
-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] temporary email addresses

[Gavin Carr]

Hi Gottfried,

On Fri, Mar 12, 2004 at 08:32:01PM +1100, Gottfried Szing wrote:
> is there a way to generate email addresses, which have an "expiry date". 
> what i want to achieve is to use these email addresses for posting in 
> news groups and for site registration.

I know of a couple of qmail solutions that should be generalisable to
other environments + toolchains:

  http://jclement.ca/software/datedmail.py/
  http://www.palomine.net/qdated/

I've also seen per-sender-encoded addresses which are similar, but
can't lay my hands on a link at the moment.

Cheers,
Gavin

-- 
Open Fusion P/L - Open Source Business Solutions [ Linux - Perl - Apache ]
ph:  02 9875 5032fax: 02 9875 4317
web: http://www.openfusion.com.aumob: 0403 171712
- Fashion is a variable, but style is a constant - Programming Perl
-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Mutt, exim and maildir

[Gavin Carr]

On Mon, Mar 08, 2004 at 02:38:06PM +1100, Alexander Samad wrote:
> On Mon, Mar 08, 2004 at 02:32:32PM +1100, Jamie Wilkinson wrote:
> > This one time, at band camp, Alexander Samad wrote:
> > >thanks nicier, but as a side question any ideas on how I lost lines,
> > >there were there before the change to Maildir
> > 
> > You got the lines count because in mbox format it's easy to work out
> > when parsing the mailbox file.  In maildir format, it requires more
> > processing and since you don't need to know how long the messages are to
> > parse them all, it's not done.
> > 
> > Line counting in mbox parsing is a trivial extension, in maildir it's
> > unnecessary overhead.  Reading the Lines or Content-Length headers
> > (which I think is what Jeff is referring to) and relying on that is bad
> > practice for mail readers[1], but displaying the values it thinks are
> > correct is merely cosmetic.
> > 
> > [1]
> > http://wp.netscape.com/eng/mozilla/2.0/relnotes/demo/content-length.html
> 
> Okay agree, but still the underly question I have is my only change has
> been in my procmail rc file and the change has gone from 
> 
> ~alex/imap/mboxfile
> 
> to
> 
> ~alex/Maildir/.mboxfile/
> 
> The jist off it is, my the change in the mbox format the lines header
> was there and in the Maildir format it is not.
> 
> If you are saying that procmail says hey its maildir and I don't need to
> insert that attribute, thats cool.  But this hasn't been made clear.

My understanding is that when using mbox format, having a Lines: header 
makes parsing the mbox faster, so MUAs like mutt add it if it's missing.
And since you have to scan the whole mailbox anyway, it doesn't add any
processing overhead.

For Maildir mailboxes, having a Lines header doesn't help with parsing,
and you get better performance by initially just reading headers rather
than bodies as well, so MUAs like mutt *don't* add said Lines header.
OTOH, it does seem to generate more than it's fair share of questions/
complaints.

So I think it's a mutt thing, not a procmail thing, and supposedly a 
feature rather a bug. :-) 

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] archiving email from mbox files

[Gavin Carr]

On Fri, Mar 05, 2004 at 03:33:55PM +1100, Peter Hardy wrote:
> On Fri, 2004-03-05 at 15:08, Alexander Samad wrote:
> > No the last step is I want to be able to run a program/script against a
> > mbox file and seperate into two files (or append to other files) emails
> > that are older than a certain date and younger than a certain date.
> > 
> > so before I venture off in to formail,perl and write my own does any
> > body know of a package module that does this ?
> 
> Yep! Called, funnily enough, archivemail.
> 
> Nice little python jobbie that'll rotate anything older than a certain
> date out of a given mail box and in to a compressed mbox.

Or you could consider converting from mbox format to Maildir format
(which mutt and procmail both handle fine), which stores each message
as a separate file instead of your monster mbox file. Rolling your own 
stuff for this sort of archiving then becomes almost trivial.

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Don't like your HELO/EHLO. Hostname must contain a dot.

[Gavin Carr]

On Tue, Mar 02, 2004 at 07:11:23PM +1100, Nick Croft wrote:
> My humble home desktop machine goes by the name 'artarmon' (I like the place).
> 
> I just had a posting rejected from a list at sourceforge, with the
> explanation that they want to see a dot in my hostname.
> 
> I've since renamed the machine artarmon.net, and the message now seems to
> have reached the list. 
> 
> Just the same looking through the mail headers of pfaedit-users, I notice
> that most emails originate on machines named 'localhost' or some other
> single-word hostname, without dots.
> 
> Any ideas?

About why they're rejecting you, or about how others are getting through?

For the first, RFC2821 (SMTP) says the argument to HELO and/or EHLO must 
be a fully qualified domain name, or an IP address literal (like 
[66.35.250.206]).  So you're not being RFC compliant by using bare 
'artarmon', and they're being strict about it, done usually as a spam/virus
filtering measure.

For the second, the MTA can do a regex/string check or can do an actual
dns lookup - if they're doing a dns lookup things like localhost will pass,
of course, unless specifically excluded. Other generic barenames sometimes
get passed too that way - 'mail', 'mailhost', etc.

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Maildir layout with IMAP?

[Gavin Carr]

Hi Sonia,

On Wed, Feb 11, 2004 at 09:09:59PM +1100, Sonia Hamilton wrote:
> I'm using Maildir style folders with an IMAP server (bincimap), and I
> wondering how I should setup the Maildirs so that on the mail client the
> inbox emails appear at the top of the tree, rather than as a subfolder
> of '/'. IMAP mail retrieval is working, but the users on my network
> don't like the current setup. Server folder layout is:
> 
> ~user/Maildir/
> INBOX/
> cur/
> new/
> tmp/
> folder1/
> cur/
> ..
> 
> On the client (running the 'other' operation system), this appears as:
> 
> /? error
> INBOX/
> folder1/
> folder2/
> ..
> 
> What I'd like is:
> 
> / (inbox emails)
> folder1/
> folder2/
> ..

I haven't played with bincimap yet, but on Courier (from memory) you do 
it like this:

  ~user/Maildir/
 cur/
 new/
 tmp/
 .folder1
cur/
new/
tmp/
 .folder2
cur/
new/
tmp/

By the looks you don't need leading dots on your subfolders, so try
nesting them *inside* your INBOX folder. You might also experiment 
with whether the two-stage Maildir/INBOX separation is required - on
Courier the top Maildir is just treated as INBOX directly on the 
client side.

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] convert src to binary rpm

[Gavin Carr]

On Tue, Dec 09, 2003 at 03:25:04PM +1100, Jamie Wilkinson wrote:
> This one time, at band camp, Simon Bryan wrote:
> >HOw do I convert a src rpmm to the correct rpm for my system? (i386 i686? RH9)
> >
> >I tried rpm --rebuild but it says it is an invalid option.
> 
> Try
> 
>   rpmbuild -ba thingy.src.rpm
> 
> rpm lost the ability to build SRPMs when RH9 was released.

And 'rpmbuild --target i686 -ba thingy.src.rpm' for a particular target 
system ...

Cheers,
Gavin

-- 
Open Fusion P/L - Open Source Business Solutions [ Linux - Perl - Apache ]
ph:  02 9875 5032fax: 02 9875 4317
web: http://www.openfusion.com.aumob: 0403 171712
- Fashion is a variable, but style is a constant - Programming Perl
-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] djbdns opinions?

[Gavin Carr]

On Sun, Dec 07, 2003 at 09:00:49PM +1100, Andrew Cowie wrote:
> On Sun, 2003-12-07 at 20:33, Sonia Hamilton wrote:
> > I've been looking at DJ Bernstein's djbdns server at
> > http://cr.yp.to/djbdns.html.
> > 
> > I'm thinking of using it instead of bind9; just wondering if anyone has
> > any experiences with using it, and opinions (positive or negative).
> 
> You said you've used qmail. That probably means you've used daemontools
> as well. If you're comfortable with (and using) svscan/svc/multilog/etc
> and the whole /service symlink thing, then plugging djbdns in is a snap.

I'd agree - if you're comfortable with qmail then djbdns is super easy to 
get up and running and great to work with.

> That said, I'm really impressed with djbdns (and dnscachx which goes
> with it). It's a bit of a learning curve as is usual with DJB's stuff
> (what is it with that guy compulsively making hard to use software?) but
> the fact that it automatically generates the PTR records, for example,
> is terrific.

I reckon almost the opposite is true - BIND is so *ugly* to work with 
that it screws up your intuitions about dns and how it should work (e.g.
djbdns decoupling zone serving - tinydns - from caching - dnscache - 
seems really weird at first, but after a while it's obviously The Right
Thing to do). 

> I seem to recall that it either a) has issues with, or b) doesn't do,
> zone transfers. I'm probably wrong, but you might want to check.

I think it does support bind-style zone transfers, but you have to run
a separate daemon - axfrdns - and is discouraged unless you really have
to interface with bind servers (for hosting, not caching, of course).
If you're running djbdns on both primaries and secondaries djb 
recommends you use something tried-and-true like scp or rsync to 
transfer your zone files.

Hope that helps.

Cheers,
Gavin

-- 
Open Fusion P/L - Open Source Business Solutions [ Linux - Perl - Apache ]
ph:  02 9875 5032fax: 02 9875 4317
web: http://www.openfusion.com.aumob: 0403 171712
- Fashion is a variable, but style is a constant - Programming Perl
-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] Milter Advice

[Gavin Carr]

Hi Scott,

On Tue, Nov 25, 2003 at 03:06:49PM +1100, [EMAIL PROTECTED] wrote:
> What I would like to do is this kind of network setup:
> 
>   internet
> |
>  Gateway
>  /\
>  MX/HTTP  LAN
> 
> Now From the firewall, I would not allow the MX/HTTP server access the LAN 
> at all, via iptables, and they will be on different subnets.
> 
> So what I want to do is setup the MX (Currently going to be exim, but open 
> to suggestions) to accept incoming mail for our domains, then have 
> fetchmail on the LAN pickup the mail from MX every few minutes, after its 
> been scanned for viruses/spam etc.

The problem with using fetchmail to pull mail onto the LAN is that you're
introducing support for another protocol (POP, ETRN, ODMR or something) onto 
your MX/HTTP box - it's probably just simpler to use SMTP and open up just
the SMTP port inward from MX -> Domino. I guess that's somethign more like:

 internet
|
  proxy (MX/HTTP)
|
   LAN

> How would I accomplish this?
> What would be the best tools in debian woody?
> Are there better/easier ways to do this?

I personally would never use sendmail in this kind of situation because of
the recurring security SNAFUs with it - any of the other standard MTAs
around would probably be better. I'm currently using qmail + qpsmtpd
(http://develooper.com/code/qpsmtpd/) for this at the moment - qpsmtpd
is a very nice pluggable smtp daemon written in perl that lets you do 
very sexy things on the spam/virus/validation fronts.

Cheers,
Gavin

-- 
Open Fusion P/L - Open Source Business Solutions [ Linux - Perl - Apache ]
ph:  02 9875 5032fax: 02 9875 4317
web: http://www.openfusion.com.aumob: 0403 171712
- Fashion is a variable, but style is a constant - Programming Perl
-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] Logging to a program with Syslog

[Gavin Carr]

On Fri, Nov 21, 2003 at 05:06:49AM +, Mike MacCana wrote:
> On Fri, 2003-11-21 at 04:59, Gavin Carr wrote:
> > Yep, misreading the man page. :-) A pipe can be used to send the output
> > to a named pipe/fifo, not direct to an executable. You then can have 
> > a program reading from the fifo and doing whatever you want, of course.
> 
> Hehe. I did actually read that man page on my system, but I had heard
> (or maybe seen) that this could be done directly.
> 
> When I went online, I found:
> http://www.gsp.com/cgi-bin/man.cgi?section=5&topic=syslog.conf
> 
> Which mentions "
> A vertical bar ("|"), followed by a command to pipe the selected
> messages to. The command is passed to sh(1) for evaluation, so usual
> shell metacharacters or input/output redirection can occur."
> 
> Is there another version of syslog also called just plain `syslog'?

I believe at least some of the BSD syslogds support this - your page above
is from FreeBSD, by the looks. The standard Linux version doesn't appear 
to. Depending on what you're trying to do, you might also try syslog-ng,
which definitely does support it. 

Cheers,
Gavin
-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] Logging to a program with Syslog

[Gavin Carr]

Hi Mike,

On Fri, Nov 21, 2003 at 03:27:16AM +, Mike MacCana wrote:
> According to the syslog man page, I should be able to use a pipe
> character to send a log to a program. Adding a line like:
> 
> user.notice   |exec wall
> WORKDAMNYOU
> 
> to /etc/syslog.conf, restarting syslog, then running "logger whatever"
> (which sends messages from the user facility with a priority of notice)
> just doesn't seem to work. 
> 
> Logging to a file from this facility / priority does though.
> 
> Anyone know what I'm doing wrong?

Yep, misreading the man page. :-) A pipe can be used to send the output
to a named pipe/fifo, not direct to an executable. You then can have 
a program reading from the fifo and doing whatever you want, of course.

Cheers,
Gavin

-- 
Open Fusion P/L - Open Source Business Solutions [ Linux - Perl - Apache ]
ph:  02 9875 5032fax: 02 9875 4317
web: http://www.openfusion.com.aumob: 0403 171712
- Fashion is a variable, but style is a constant - Programming Perl
-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] YahooIM Protocol

[Gavin Carr]

On Fri, Sep 26, 2003 at 02:58:55PM +1000, James Gray wrote:
> Quick question.  Has Yahoo changed their IM protocol?

Yep. There was a slashdot thread discussing it a week ago:
http://slashdot.org/article.pl?sid=03/09/17/2358240&mode=thread&tid=185&tid=187

> I've worked around it by downloading the Yahoo Linux 
> (http://messenger.yahoo.com/messenger/download/unix.html) client and it's 
> working (complete with banner ads and other nastyness!).  However, I 
> preferred my single jabber client (Psi) being able to access all my IM 
> accounts.  Yes I know about "EveryBuddy" and other all-in-one clients, but 
> the other reason I'm running a jabber server is simply because it's open 
> (as in the Jabber protocol is an open source/standard).  

Well, both Everybuddy (ayttm) and Gaim stopped working too. Apparently Gaim 0.69 
works - I'm just compiling it up now. 

Cheers,
Gavin

-- 
Open Fusion P/L - Open Source Business Solutions [ Linux - Perl - Apache ]
ph:  02 9875 5032fax: 02 9875 4317
web: http://www.openfusion.com.aumob: 0403 171712
- Fashion is a variable, but style is a constant - Programming Perl
-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] (OT) IP number geographic locations

[Gavin Carr]

One fast way to do this is via Perl - try installing any of these from CPAN:

  Geo::IP
  Geo::IP::PurePerl
  Geo::IPfree

Cheers,
Gavin

On Wed, Oct 01, 2003 at 10:11:46AM +1000, Tim White wrote:
> > ARIN keeps a list of first-level allocations.  A whois query 
> > to whois.arin.net for the IP address of interest should 
> > provide you with a record, from which you can retrieve the 
> > country code and see if it evaluates to AU.  Try it - run 
> > 'whois 203.1.2.3 | grep "^country:"' in a terminal. 
> > Lovely.   Repeat with any IP address you like, maybe even 
> > write a bit of awk
> > or sed or something to strip out everything except the country code.
> 
> Once you find out who ARIN has allocated the IP range to, you can usually go 
> and check their whois server.
> 
> An example would be to do 'whois 203.1.2.3 @whois.apnic.net'. This will give
> you the name and address that APNIC have allocated the IP address to. Often
> these will be ISPs, some of whom have their own whois servers and so on.

-- 
Open Fusion P/L - Open Source Business Solutions [ Linux - Perl - Apache ]
ph:  02 9875 5032fax: 02 9875 4317
web: http://www.openfusion.com.aumob: 0403 171712
- Fashion is a variable, but style is a constant - Programming Perl
-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] Redhat 9 install

[Gavin Carr]

On Mon, Sep 08, 2003 at 09:50:07AM +1000, Edwin Humphries wrote:
> After a cumulative 15 man-hours wasted time, we've finally found out there are known 
> problems with installation on systems with useful amounts of RAM. We had a 
> successful install on a 64Mb router/firewall, but when we try to install on a 512Mb 
> system, the install kept coming up with CD read errors - on three separate sets of 
> CDs and three separate systems. I understand that anaconda has memory management 
> issues.

This sounds extremely dubous to me - what's your source for it? I've installed several 
RH9 systems lately with up to 3GB of RAM with no issues at all. If you have more than 
4GB you have to use the bigmem kernel, but anything less than that should work OOTB 
just 
fine. 

The only RAM problem I had turned out to be a flaky stick, and we got the install
working by explicitly limiting the kernel memory during the install - you might try
something like 'linux install mem=64M' at the install prompt to see if that helps.

Cheers,
Gavin

-- 
Open Fusion P/L - Open Source Business Solutions [ Linux - Perl - Apache ]
ph:  02 9875 5032fax: 02 9875 4317
web: http://www.openfusion.com.aumob: 0403 171712
- Fashion is a variable, but style is a constant - Programming Perl
-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] Red Hat Linux 9 ISOs will be available by March 31st 2003 !!!

[Gavin Carr]

On Tue, Mar 25, 2003 at 06:53:01PM +1100, LS wrote:
> Just found out about this from RHN.

Only available to paying RHN subscribers on 31 March though. Mere mortals 
have to wait for another week.

Cheers,
Gavin

-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] secure imap -> outlook

[Gavin Carr]

On Tue, Mar 18, 2003 at 07:10:45AM +1100, Tony Green wrote:
> > I'm running a UOW imap setup (on a mandrake box) for a client and 
> > recently got it working with SSL, port 993 and all that. Everything was 
> > working fine with a variety of client/OS combinations until the new 
> > windows XP machine turned up today and its Outlook started complaining 
> > about invalid server certificates. I haven't looked into the problem in 
> > detail yet but just wanted a heads up to see if anyone has had this 
> > problem.
> 
> I have a similar problem with courier-(imap|pop)-ssl.  I don't think
> that outlook likes self signed certs.
> 
> If you point IE to https://yourmail.server:993, you can save the cert
> and in some versions that works.
> 
> Anyone else have anything?

We decided to get a commercial certificate - < A$100 from 
www.instantssl.com - and used that with courier-imap|pop with great 
success. I haven't used the UW one though.

Cheers,
Gavin
-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] /boot partitions

[Gavin Carr]

Hey sluggers,

Is there any sort of consensus on whether /boot partitions are necessary
these days (at least on modern hardware)? I thought they were mostly a 
workaround to the old 1024 cylinder problem with flaky BIOSes, but distros 
seem to still like using them when autopartitioning. The Large Disk HOWTO [1] 
seems to imply that problems with large disks are largely a thing of the
past now, so is there any other good reason for having a /boot?

Cheers,
Gavin

[1] http://www.tldp.org/HOWTO/Large-Disk-HOWTO-2.html
-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] [OT] Creating userid and passwords for website

[Gavin Carr]

On Mon, Feb 10, 2003 at 04:17:13PM +1100, Richard Hayes wrote:
> Using a .htaccess file it is very easy to password protected areas of a 
> website.
> 
> Is there any painless way to drop in authentication for user with PHP or 
> any other system?
> 
> 1. What I would like is password based on email address.
> 
> 2. Change passwords
> 
> 3. Issue new one is case of lost password.

Are you after a full solution or a library around which to roll your own?

Libraries:
 - PHPaccess: http://sourceforge.net/projects/phpaccess/
 - http://aspn.activestate.com/ASPN/Cookbook/PHP/Recipe/108479
 - (Perl) Apache::Htpasswd, CPAN

There's a huge array of full solutions, but matching them with your exact 
requirements is always the hard part. I've found it's often faster to roll
your own with a good library than hack someone else's 80% solution. But 
some places to start:

 - http://php.resourceindex.com/Complete_Scripts/User_Management/Password_Protection/
 - http://www.scriptdex.com/dex/php_user_management.shtml
 
HTH.

Cheers,
Gavin

-- 
Open Fusion P/L - Open Source Business Solutions [ Linux - Perl - Apache ]
ph:  02 9875 5032fax: 02 9875 4317
web: http://www.openfusion.com.aumob: 0403 171712
- Fashion is a variable, but style is a constant - Programming Perl
-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] Using ANSI escape sequences with RedHat8

[Gavin Carr]

Hi Greg,

On Fri, Jan 24, 2003 at 09:53:55AM +1100, [EMAIL PROTECTED] wrote:
> Now instead the CSI code is dropped and the remaining 
> string is dumped to the display.
> 
> The environment variable TERM is set to "linux". The 
> program "cat" can be used to dump the above sequence to 
> the terminal with the intended result.

Not sure if this is your problem, but one RH8 gotcha is their use of
UTF8-compatible locales, which seems to screw lots of other things up.
Check your LANG and LC* settings, and try resetting LANG to en_AU and
see if that makes a difference.

Cheers,
Gavin

-- 
Open Fusion P/L - Open Source Business Solutions [ Linux - Perl - Apache ]
http://www.openfusion.com.au
- Fashion is a variable, but style is a constant - Programming Perl
-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [Re: [SLUG] Changing Perl Cpan Install Repositories]

[Gavin Carr]

On Tue, Dec 31, 2002 at 12:57:51PM +1100, Louis Selvon wrote:
> *** You are right an "echo $PATH" returned:
> 
> << /usr/local/bin:/bin:/usr/bin:/usr/X11R6/bin:/home/admin/bin <<
> 
> >If this looks right, either change your path around or just do 
> '/usr/bin/perl -MCPAN -e shell' instead.
> 
> *** I actually was doing /usr/bin/perl to force it to install in the 5.6.0
> directory.

Strange. What does '/usr/bin/perl -v' say?
> 
> Another question where is these PATHs stored for Red Hat Linux 7.1. I cannot
> edit the file as I am not sure what it is ??

It's set from your profile in your home directory, probably .bash_profile on RH.

Cheers,
Gavin

-- 
Open Fusion P/L - Open Source Business Solutions [ Linux - Perl - Apache ]
http://www.openfusion.com.au
- Fashion is a variable, but style is a constant - Programming Perl
-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] Where to get sharemarket data?

[Gavin Carr]

On Tue, Dec 31, 2002 at 08:33:47AM +1100, Holroyd Engineering Services wrote:
> Looking to set up my RH7.3 to retreive end of day or if possible live
> sharemarket data. does anyone know of a good source (free or otherwise).

There's a really nice perl module - Finance::Quote - that picks up the data 
(20 minutes delayed) from Yahoo or the ASX. It depends a bit what you want
to do with the data.

I was playing around with this a year or so ago and started building a 
website to do some fun slicing and dicing on the All Ords data, but got waylaid
fairly early on. The site is still running though, at http://www.sharebot.net,
if you're interested.

Cheers,
Gavin

-- 
Open Fusion P/L - Open Source Business Solutions [ Linux - Perl - Apache ]
http://www.openfusion.com.au
- Fashion is a variable, but style is a constant - Programming Perl
-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] Changing Perl Cpan Install Repositories

[Gavin Carr]

On Mon, Dec 30, 2002 at 11:34:44PM +1100, Louis Selvon wrote:
> I am downloading some Perl modules via "perl -MCPAN -e shell".
> 
> My problem is when installing using the standard
> 
> perl Makefile.PL 
> make 
> make test 
> make install 
> 
> All the modules are getting installed to "/usr/local/bin/perl/perl5/5.8.0"
> 
> In "Makefile" after "perl Makefile.PL" I can see all the paths are pointing to
> "/usr/local/bin/perl/perl5/5.8.0".
> 
> I want all install(s) to go to "/usr/lib/perl5/5.6.0" base directory. What do
> I have to change to make all CPAN install go to this directory instead ?

Do a 'which perl'. You probably have /usr/local/bin before /usr/bin in your
PATH, so doing the 'perl Makefile.PL' picks up the /usr/local/bin one instead
of the /usr/bin one.

If this looks right, either change your path around or just do 
'/usr/bin/perl -MCPAN -e shell' instead.

Cheers,
Gavin
-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] fustrations with mysql (TIMESTAMP)

[Gavin Carr]

On Tue, Dec 17, 2002 at 03:51:03PM +, James Gregory wrote:
> On Tue, 2002-12-17 at 04:41, Tony Green wrote:
> > Your using 'INSERT INTO blah VALUES( 'blah','blah','');' and it doesn't
> > work.  Try 'INSERT INTO blah VALUES( 'blah','blah',NULL);' and see if
> > you get any better results.
> 
> I've also had issues with it, but if you want it to try it then
> specifying a value for it won't let it set it itself.
> 
> If it is going to work, the way to do it will be to not specify that
> field at all in your list of fields to insert into. It's supposed to
> operate like a default value essentially. If you specify something else
> it will use that. both NULL and the empty string are "something".

Nope, that's wrong. Not inserting the field OR inserting NULL are
equivalent for timestamps - both set it to the current time. But '' and 
'\n' won't work - you have to use NULL proper.

I've never had any problems like this with timestamps. What version are
you using Michael? You're absolutely sure you're inserting null and not
a quoted 'null'?

Cheers,
Gavin

-- 
Open Fusion P/L - Open Source Business Solutions [ Linux - Perl - Apache ]
http://www.openfusion.com.au
- Fashion is a variable, but style is a constant - Programming Perl
-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] Question about text computer game from late seventies/eighties

[Gavin Carr]

rogue?

:-)

-G

-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] Setting Up a Website

[Gavin Carr]

On Thu, Dec 12, 2002 at 03:43:15PM +1100, Michael Fox wrote:
> > PS. forget the .au domain names, they're a rip off.
> > I used to be keen on them but not when a .com is AU$30
> > per year and you have a choice of hundreds of registrars.
> 
> I've seen .com.au's for $77 for 2 years. I believe..

Namescout (www.namescout.com) has .com.au's for $59.99 for 2 years.
Their support sucks, but if you can get away without it ...

-G

-- 
Open Fusion P/L - Open Source Business Solutions [ Linux - Perl - Apache ]
http://www.openfusion.com.au
- Fashion is a variable, but style is a constant - Programming Perl
-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug