Pretty soon after starting to use the Mac I noticed the keyboard layout was wrong - the @ and ” were swapped, something you usually see if using a US keyboard layout with a UK keyboard. Finally I found a useful post that shows how to fix this annoying issue. At first I didn’t want to believe this was the only option, so I scoured the Apple site for an “official” fix… nothing. Thanks to that post I’m fixed… but it is poor that Apple haven’t fixed their UK keyboard layout s in over 2 years!

So, till now I have been running an IMAP server on OpenBSD using dovecot. I was using fetchmail to pull in some POP accounts and I used the built in OpenBSD mail.local delivery agent and the basic OpenBSD localhost.cf sendmail config. Everything was setup with mbox style storage - inbox was /var/mail/user and directories were in /home/user/mail/

Alas mbox is nasty. So for a long time I’ve meant to upgrade to maildirs and finally got round to it. Here’s what I did to convert:

1. Moved from mbox to Maildir for new deliveries
• Installed procmail
• Setup a global /etc/procmailrc:
  PATH=/usr/bin:/bin:/usr/local/bin:.
MAILDIR=$HOME/Maildir
DEFAULT=$MAILDIR/
  NB - I don’t think PATH is needed, but I’m not changing it now!
• Started procmail and set it to run in /etc/rc.local
2. Changed dovecot to use Maildirs
   /etc/dovecot.conf
#OLD mail_location = mbox:~/mail:INBOX=/var/mail/%u
mail_location = maildir:~/Maildir

Really that’s it, the only step remaining was to get my old mbox mail coverted to maildir format. I used a mbox to maildir converter perl script downloaded from http://batleth.sapienti-sat.org/projects/mb2md/.

(Here is also where I faffed about with CPAN to get TimeDate module installed. Turned out that lynx was causing MD5 checksum failures on everything it downloaded so I installed ncftp and configured CPAN to no longer user lynx)

The script worked a breeze, I followed the basic example.

Finally I had to tell Thunderbird to

• look at the new directories

File - Subscribe

• support subdirectories

Tools - Account Settings - Server Settings - Advanced - Server supports folder that contain sub-folders and messages

Anyway I had issues compiling the diag-client.c on Solaris 10 x86:
cc diag-client.c -o diag-client
Undefined first referenced
symbol in file
gethostbyname diag-client.o
socket diag-client.o
connect diag-client.o
ld: fatal: Symbol referencing errors. No output written to diag-client
Fixed linkage with
cc -lsocket -lnsl -lm diag-client.c -o diag-client

When I opened up my mate’s Xbox and found it was a v1.6b I knew I was in trouble because the modding is more tricky, since you need to solder a PCB rebuild to the base of the motherboard. Still, I got stuck in and had it done in no time. Imagine my horror when I closed up the Xbox and nothing worked! I thought I’d fried it! Some Christmas present…

So, all stressed I started the process of debugging. First off what was the symptom?
X3CE showed the enabled LED but the screen was blank and the LED was flashing yellow.

Next, I tried a reboot with the X3CE disabled (longer press on power button). Lo and behold the Xbox worked ok! Phew… at least the Xbox wasn’t dead. So perhaps my solder joints were no good… I inspected them all but no luck.

Next I tried live voltage readings with a multimeter… unfortunately the readings were no use when compared with those at Xbox-Scene because the v1.6b is different. Still nothing was majorly surprising… and were there was voltage it was getting to my pin header.

So I checked the joints again, retouched some of them but still no luck. Then I stumbled on a forum-post that pointed out the v1.6b video chip was not compatible with the basic 3.0.0 version of flashbios loaded into bank 1 of the X3CE. It also pointed out that the X3CE has v3.0.0 on the backup chip too (!) but should have the 3.0.3 bios in bank 5.

Using the install guide notes I selected BIOS bank 3.0.3 and was presented with the flashbios screen. From here the next step is to switch back to bios bank 1-4 (1MB) run flashbios to flash the first bank with your advanced bios (e.g. X3) and then reboot and it is done! The only thing to really cover yourself on an X3CE is to use the X3 bios utility to update the X3CE backup BIOS ROM with the 3.0.3 bios - after all there’s no point in having a backup ROM which doesn’t actually work!

Then I found a link which solved all my problems. I have the Windows Resource Kit and it does multicast generation. Magic.
C:\>mcast /intf:10.254.0.254 /send /intvl:1000 /numpkts:100000 /grps:224.1.1.1,224.2.2.2

Recently however I have rallied behind a new noble cause - venting my frustration with Virgin Media in the hope that someone might read my blog and swing their decision in favour of ANO ISP/TV/Phone company. Now, before continuing it should be said that I took the plunge - at great expense - and moved my phone line to BT. I can report that the line quality is better but this wasn’t my real reason for switching… no the real reason is so that I can have ADSL broadband and finally say goodbye to the awful Virgin ISP service.

Despite moving from Virgin for phone and ISP I do still use them for TV. Really it’s the only option available to me and recently I upgraded to their snazzy V+ service. Like Sky+ this is a DVR based service which also offers HDTV. Although I have a few issues with sound drops at the moment I’m not ready to complain yet… as a newcomer to DVRs I’m finding the whole experience has completely changed the way I watch TV.

And this is where the latest snippet in my personal online vendetta comes in.

This evening my V+ box has become semi-functional and tells me I am not subscribed to the V+ service. Customer-care said that my recent account changes had led to V+ being cancelled. OK, someone screwed-up… it happens (more to me than anyone else?).

Can it be fixed? Not now. They can’t fix it because my phone-line move order is still “open” and the order system only allows one order at a time(!!!). Fortunately after much work by a diligent customer-care person my V+ service will be back tomorrow morning… an improvement on the original estimate of 6 days.

It’s crazy that a CRM system (I presume) works in a way that prevents things being fixed. I only hope that my customer relationship with Virgin Media becomes simpler to manage when I have taken two-thirds of my business elsewhere.

http://www.youtube.com/watch?v=WOdjCb4LwQY


C:>ping www.google.com -t
Pinging www.l.google.com [64.233.183.104] with 32 bytes of data:
Request timed out.
Request timed out.
Request timed out.
Request timed out.
Request timed out.
Request timed out.
Request timed out.
Request timed out.
Request timed out.
Request timed out.
Reply from 64.233.183.104: bytes=32 time=28ms TTL=244
Reply from 64.233.183.104: bytes=32 time=20ms TTL=244
Request timed out.
Request timed out.
Request timed out.
Request timed out.
Request timed out.
Request timed out.
Request timed out.
Reply from 64.233.183.104: bytes=32 time=29ms TTL=244
Ping statistics for 64.233.183.104:
Packets: Sent = 20, Received = 3, Lost = 17 (85% loss),
Approximate round trip times in milli-seconds:
Minimum = 20ms, Maximum = 29ms, Average = 25ms

This is how bad contention is on my section of the Virgin Media cable network.