Thursday, 28 January 2010

How to update CyanogenMod on HTC Magic 32A [Ubuntu]

I'm running CyanogenMod on my HTC Magic, and Ubuntu 9.10 (Karmic) on my laptop computer. Most of the upgrade guides & automated tools for upgrading CyanogenMod (i.e. CMUpdater) are based on the 32B kernel. This is a guide to upgrading CyanogenMod with a HTC Magic 32A kernel. Most of the instructions are taken from the CyanogenMod wiki.
  1. Download the latest CyanogenMod (Current Stable Version) from this site.
  2. Download the latest EBI1/32A Kernel from the same page.
  3. Copy both .zip files to your SD and reboot in Recovery Mode (press Home+Power at the same time or fastboot into your recovery) 
  4. Apply Zip:
  5. Apply Kernel Port Zip: or (depends on who released the latest port: rad - Radix999,bc - bcrook ) 
  6. Reboot
  7. Wait. For a long time. =) Seriously, this will take ages to boot after an upgrade.
If you get an orange triangle & exclamation mark when you boot into recovery, just press Home+Power again to access the menu.

Ok, so what if you don't have an option to apply this zip in the recovery? If your phone only gives you the option to "apply" then you don't have the fancy recovery ROM. Download it from here (scroll down and choose the "H" version), and get the fastboot binary from here. Copy both into your android sdk "tools" directory.

Now, power down your phone, plug it into your computer via USB and press the Back and Power buttons at the same time. This should bring you to the fastboot screen. On your computer, cd into the android sdk tools directory, and type "fastboot boot ./". This will reboot the phone into the fancified recovery screen, which should let you update to any chosen zip file.

Oh, and note that this will not wipe your data!!! If you perform these steps completely and nothing goes wrong, all your apps and preferences will remain as they were. Of course, I can't guarantee anything. =) The only time you would wipe your phone is if you applied a full firmware upgrade (DRC83 in the filename, usually.) If you're unsure though, just leave a comment & I'll try to help! Or you could checkout the forums associated with some of the links above.

Wave1 Point to Point Radio (1000 series) Default Password

The default username and password details for Wave1 radios:

username: admin
password: default

Otherwise, try -

username: admin
password: admin01

This works for some of the newer LigoWave stuff.

Reseting a LigoWave Radio to Factory Defaults Using a Ping

Reseting a LigoWave Radio to Factory Defaults Using a Ping
It is possible to perform a factory reset of a LigoWave radio using a ping when you follow the instructions below.
The steps are:

  1. Obtain the Ethernet port MAC address
  1. Create a static arp entry for the Ethernet port MAC address on the same subnet you are on. For example your machine, and set radio at (radio Ethernet MAC 00-19-3B-00-06-4A)
    1. arp –s 00-19-3B-00-06-4A
    2. you can confirm this entry is in place by looking at your arp table
      arp -a (the static arp entry should exist)
  1. Restart the radio
  1. Send a 40 pings to the radio static arp entry ip address with 369 bytes and every 200ms (ping needs to be started as soon as you restart the radio). 
    1. ping –l 369 –w .2 –n 40
  1. The unit should restart using the factory defaults
  1. Remove the static arp entry
    1. arp –d


The important thing is to make sure the –w switch is used to increase the ping speed and that you are using the Ethernet MAC address, not the bridge MAC address that is seen you ping it normally.  The bridge MAC address is from the radio, and it will not respond during boot and the reset will not work.  I have put in 40 pings, as you should not send too many, -t, as if you don’t stop the pings in time you will get into a loop.  40 pings looks good from my testing.  


Tuesday, 26 January 2010

Green Cars

This is a Sponsored Post written by me on behalf of Hyundai. All opinions are 100% mine.

You've probably read about how Hyundai recently won the Global Green Challenge with their latest "ToughR" cars. Well they've now got their own Youtube channel, where you can check out videos of their prize-winning trip from Darwin to Adelaide.

Hyundai's Santa Fe R claimed the title of the biggest improvement in fuel efficiency, 5.1l/10km, against contenders including Ford, Holden and Suzuki. During the Eco challenge, contestants were allowed to implement standard fuel-saving techniques such as driving without air-conditioning, despite high mid-Australian temperatures. Most also did the trip at average speeds of 75km/h which is well under the normal speed for the Stuart Highway of 130km/h in the Northern Territory and 100km/h in South Australia.

Hyundai is setting the standard in an increasingly competitive market for fuel-efficient, low-carbon "green" cars. Their Youtube videos are certainly worth checking out. They are presented in an semi-comedy style, very easy to watch and present some interesting and useful information about the challenges this competition presented.

There were also some other prizes awarded for the Global Green Challenge, based on different metrics:
On fuel efficiency improvement: 1 HSV Maloo 48.7 per cent improvement, 7.74l/100km average, official fuel rating 15.1l/100km; 2 Ford XR6 turbo Falcon  39.8%, 7.04l/100km, 11.7l/10km; 3 Skoda Superb 33.5%, 4.59l/100km, 6.9l/100km

On actual fuel consumption:  1 Ford Fiesta ECOnetic 3.13l/100km, 2 Mini D 3.42l/100km, 3 Mini D 3.49l/100km.

Visit my sponsor: Hyundai Santa Fe - AU 2

Thursday, 21 January 2010

Microsoft Exchange 2003 Incremental Backups with NTBackup.exe

When performing an incremental backup of a Microsoft Exchange Storage Group with NTBackup.exe, you may encounter the following error:

Backup Status
Operation: Backup
Active backup destination: File
Media name: "incremental_exchange.bkf created 21/01/2010 at 3:55 PM"

Error: SERVER\Microsoft Information Store\First Storage Group is not a valid drive, or you do not have access.


This is a very obscure error message, and is almost impossible to decipher at face value. I spent many days trying to fix this error, and eventually discovered that you cannot create an incremental (or differential) backup when circular logging is enabled.

Turning off circular logging will enable an incremental backup to be created.

Another symptom encountered with this problem, was whenever I started an incremental backup it would continuously prompt me "Insert the following media: incremental_exchange.bkf created 21/01/2010 at 4:46pm Media #1. Do you want to continue?" When I clicked "Yes", it prompted me again. When I eventually clicked "No", the backup failed. The report showed the error message detailed above.

More information about circular logging:
How to disable circular logging:
Other similar threads:
How to create Exchange backups with NT Backup:

Tuesday, 19 January 2010

How to open Microsoft Word files in VIrtualBox Windows from Ubuntu

There's a great article about how to setup your system to you can double click a file in Ubuntu, and have it open in a Windows machine within VirtualBox. This, along with the seamless display technology built into the latest VirtualBox release, allows some pretty amazing integration scenarios.

Saturday, 16 January 2010

How to convert a directory of videos using ffmpeg

Here is a simple command useful for using a wildcard (not a regular expression) to select filenames to convert from one video format (or audio, for that matter) to another using FFMPEG on Linux.

The command below will convert a directory of .flac files into .mp3s:

for f in *.flac; do ffmpeg -i "$f" -acodec libmp3lame -ab 320k  "${f%.flac}.mp3"; done

To convert videos, we could use this command:

for f in *.flv; do ffmpeg -i "$f" "${f%.flv}.avi"; done

This converts a directory of .flv video files into .avi files.


Monday, 4 January 2010

How to reset a Deliberant AP to Factory Defaults (CPE 2n)

From the manual available at the Deliberant Support Site:

Resetting to Factory Defaults
Deliberant products have the capability of being reset to defaults by pinging the device with a certain
packet size when the radio is booting.
During the startup of the device, when the drivers of the ethernet interfaces are loaded, the discovery
daemon is started. The daemon suspends startup process for 3 seconds and waits for ICMP "echo
request" packet of length 369 bytes. If the packet received, the discoveryd resets the device to default
Steps to reset to default settings:
Step 1. Power off the device.
Step 2. Obtain the device MAC address.
Step 3. Connect a PC to the same physical subnet as the device.
Step 4. Execute 'arp -s' command to assign the IP address (IP address should be from the same
subnet as PC) to the device MAC address:
      arp -s [ip address to assign] [device address]
Step 5. Start pinging the device:
  For linux users:
      ping [ip address] -s 369
  For Windows users:
      ping [ip address] -l 369 -t -w 0.2
Step 6. Power up device and wait about 30sec or more (depends from device hardware).

I tested these steps with Deliberant AP Solo, Duo & CPE 2-N. Sometimes it takes a long time, but eventually will work.

Sunday, 3 January 2010

Fix: Firefox 3 runs slowly keeps pausing and not responding on Ubuntu 9.10 Dual core

When I installed Ubuntu 9.10 "Karmic" on my new HP Dual-core computer, I found that Firefox was having a lot of issues. It would startup alright, and load pages pretty fast. But mainly whenever I tried to open a new tab, or had a lot of tabs open, it would randomly stop responding for a few seconds at a time. Since I have compiz enabled, Firefox would "grey out" and stop responding to mouse clicks. CPU usage never exceeded 50% of both cores though.

I trued using the CPU Frequency Scaler applet on the panel to increase the CPU speed with no effect. Also, reducing the threshold at which the CPU frequency is raised had no effect either. It turns out that Firefox just doesn't like running across multiple CPUs for some reason. The solution was to restrict Firefox to run on a single core.

This can be accomplished using the linux utility "taskset". The command I used is shown below (run as root):

taskset -c -p 1 `pidof firefox`

An explanation:
  • The -c says to use a zero-based number to identify cores (Core 1 is "0", Core 2 is "1", etc). The default is to use a bitmask (more advanced).
  • The -p says to reuse the existing pid, and not start a new task.
  • Then we specify the core (Core 2 in this case, we could separate them with commas such as "0,1" if we wanted both cores to be used). 
  • Then we have another command, between backticks (` not ') that finds us the pid (Process ID) of the running Firefox.

This command has to be run every time Firefox starts. I'm looking for a way to change this automatically when Firefox is run - perhaps by changing the menu entry? Any suggestions are welcome!

Saturday, 2 January 2010

Free Sheet Music Scanning software for Linux - Audiveris

I've been using Sibelius on Windows, along with PhotoScore Lite to scan and transcribe sheet music to date. Recently I set out to find a Linux alternative.

Audiveris is a Java-based open source (therefore free) sheet music notation scanner. It will import a PDF or image file (single page as far as I can see) and automatically transcribe the notes into an on-screen digital version. You can then export to MIDI or MusicXML files, which can be edited in another program such as Rosegarden or MuseScore.

While this program is cross-platform and runs on Windows, Linux and Mac OS, it's advanced OCR features are limited to the Windows platform due to a dependency on Microsoft's C++ libraries. This may change in the future if more developers commit time to the project, but for now I have found it to be a very capable Linux alternative to some of the commercial Optical Music Recognition software. I particularly like the fine-grained reports and control you have over the entire process. In fact, you can examine any stage of the transcription in detail thanks to the multi-step procedure. Not the prettiest display, but then who wants pretty - so long as it gets the job done?