I was a bit disappointed with html2ps's output quality, so I went ahead and hacked a small xulapp that converts an .html file to a postscript without showing any graphics and in an unattended way. It's far from ideal but it (almost) works ;)

Here it is: mozilla2ps

Mark Lanegan - Playing Ground

Kernels pre 2.6.14 will need this patch:

--- linux-2.6.13.4/drivers/scsi/scsi_devinfo.c.largelun 2005-11-17 17:52:49.424914776 +0100
+++ linux-2.6.13.4/drivers/scsi/scsi_devinfo.c  2005-11-17 17:53:12.938400491 +0100
@@ -135,7 +135,7 @@
{"COMPAQ", "MSA1000 VOLUME", NULL, BLIST_SPARSELUN | BLIST_NOSTARTONADD},
{"COMPAQ", "HSV110", NULL, BLIST_REPORTLUN2 | BLIST_NOSTARTONADD},
{"DDN", "SAN DataDirector", "*", BLIST_SPARSELUN},
-   {"DEC", "HSG80", NULL, BLIST_SPARSELUN | BLIST_NOSTARTONADD},
+   {"DEC", "HSG80", NULL, BLIST_SPARSELUN | BLIST_LARGELUN | BLIST_NOSTARTONADD},
{"DELL", "PV660F", NULL, BLIST_SPARSELUN},
{"DELL", "PV660F   PSEUDO", NULL, BLIST_SPARSELUN},
{"DELL", "PSEUDO DEVICE .", NULL, BLIST_SPARSELUN}, /* Dell PV 530F */

Fairly busy with things, lots of work on pretty much every front. Took some time off this saturday to go get some cd's and to hit the movies.

Bought two cds: Ballad of the Broken Seas - Isobel Campbell, Mark Lanegan Negramaro - Tutto scorre

Last night we watched : Walk the Line

And I finally completed Chan-wook Park's trilogy:

Couple of quick notes.. Get Dan Kegel's crosstool. Unpack it, edit demo-ppc970.sh (G5 - Powerpc toolchain) and make sure the only uncommented eval line reads like:

eval  `cat powerpc-970.dat gcc-4.0.1-glibc-2.3.5-hdrs-2.6.11.2.dat`
sh all.sh --notest

Once the new toolchain is in place, unpack your linux-2.6.16-rc1 file

export PATH=/home/devel/ppc/gcc-4.0.1-glibc-2.3.5/powerpc64-unknown-linux-gnu/bin:$PATH
wget http://michele.pupazzo.org/files/xserve-2.6.16-rc1.config -O .config




make ARCH=powerpc AS=powerpc64-unknown-linux-gnu-as
LD=powerpc64-unknown-linux-gnu-ld
CC=powerpc64-unknown-linux-gnu-gcc
STRIP=powerpc64-unknown-linux-gnu-strip
OBJCOPY=powerpc64-unknown-linux-gnu-objcopy
OBJDUMP=powerpc64-unknown-linux-gnu-objdump
RANLIB=powerpc64-unknown-linux-gnu-ranlib

Proceed with the usual kernel compiling lingo...

I needed to download LARTC's mailing list archives to check some things while on my way home by train. Here's the script I used to turn them into a single mbox file:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19

#!/bin/sh
BASE=http://mailman.ds9a.nl/pipermail/lartc/

for i in `seq 2002 2006`; do
for j in `seq 1 4`; do
FINAL="$i""q""$j.txt.gz"
wget "$BASE""$FINAL"
done
done

echo "Done downloading"

for i in `ls *.gz`; do
gunzip $i
NEW=${i%.gz}
formail -ds > "lartc-2002-2006.mbox"
done

echo "Done formatting archives"

The Zephyrs

Two years ago in Hamburg, I discovered what became my favourite drink (albeit quite hard to find). Sol got me a couple of bottles when she returned from Berlin and I opened them up for my birthday. I somehow need to get a store or a place where I can buy this drink. It's great stuff, plus it has this irresistible cover on the bottle.

Death Cab Cuties

While compiling the latest Bacula on SCO Unix I spotted this little gem which makes the compile fail:

/usr/include/sys/file.h
...
#define FDIRECT         0x20            /* perform direct I/O/*/
...

So I got a little annoyed by the constant downloading of my blog (almost twice per second) by broken versions of the liferea RSS reader. Went ahead and added this to my configuration:

RewriteEngine on
RewriteCond %{HTTP_USER_AGENT} ^Liferea/0.*     [OR]
RewriteCond %{HTTP_USER_AGENT} ^Liferea/1.0-RC[1-3].*
RewriteRule .*$         not-existant [G,L]

This way any version less than 1.0RC4 will stop wasting bandwith.

Cake - Satan is my Motor

I stumbled upon this issue while configuring bacula on a 2.6.x based system on an IBM 4560SLX tape library.

http://www.ussg.iu.edu/hypermail/linux/kernel/0302.2/0066.html

The open() behaviour of st was changed at 2.5.3 to conform with SUS
(blocking) and what the other Unices do (timeout). If the device is opened
without O_NONBLOCK, the driver waits for some time (default 2 minutes) for
the device to become ready. If it does not become ready, an error is returned.

Bacula 1.36.x does not open the device with O_NONBLOCK and thus isn't able to cope with this change. The solution is either to upgrade bacula to a 1.37.x version or to use a 2.4.x kernel.

I'll probably clean this up and ask to add it to the bacula FAQ list.