I can’t claim to be remotely unusual in this respect though – there’s a whole subculture of programmers (“hackers”) who can relate to this mindset. Sometimes when I build a new tool, I get the impulse to share it with the world in case it turns out to be of use to anyone else. In the past, these hacks have ended up on my software web page, but I think this blog might be a better medium.

So, without further ado, here’s a cute hack I just built: bping, a wrapper around ping(8) which makes it do a lot of beeping – one beep per packet received, with pitch going up an octave for every doubling of the response latency. (See below for installation instructions.) Why did I want this?

I use ethernet over power to connect the machines in my bedroom to the equipment in the lounge. In a high-rise apartment block with about 100 wireless networks fighting over the same spectrum band, this is (normally) a much more reliable option. However yesterday the connections in my home network started behaving very weirdly. I tried pinging various machines on the network from each other to narrow down the problem, but it was annoying to have to keep going between rooms to visually monitor the output from the various pings when it would have been quicker to be able to hear the quality of the connections being tested. So I wrote bping, which also turns a laptop into a sort of Geiger counter for wifi.

Currently bping uses bip (here’s a suggested approach to installation) which in turns uses sox to make the beeps. This works fine on Fedora 15 but unfortunately for some reason takes longer than one second to run on Ubuntu 11.10 and openSUSE 12.1, regardless of how short the beep is. I think it’s something to do with pulseaudio, but I haven’t bothered to figure it out yet. Answers on a e-postcard please …

By the way, many of my hacks are already publically available on my github account, but most aren’t documented yet. The current exceptions are:

• my UNIX shell environment and tools, and
• my utilities and configuration files for git.

I’ll try to document some of the others soon. Watch this space!

So plan B was to set up port redirection from the host to the guest, e.g. so that ssh’ing to localhost port 2222 would redirect to the guest’s port 22.

After a quick google, some fiddling with iptables, and a glance at the libvirt Networking wiki page, I was still having no luck. Then it hit me – my guest was using user-mode networking, and rather than getting its DHCP-allocated IP from the libvirtd-launched dnsmasq instance on the host, was receiving a hardcoded allocation of 10.0.2.15 from the host which is on 10.0.2.2. This can be extremely puzzling at first, because no network commands run on the host (such as ifconfig, iptables, brctl, route) will reveal this magic address, yet the host is still accessible from the guest via it.

After a lot more googling, I stumbled across a technique for configuring host to guest port redirection on a running VM. This sounded very promising, but virt-manager refused to accept the magic Control-Alt-2 key combination to switch to QEMU monitor mode. It turns out that this is no accident. However, since libvirt 0.8.8, the QEMU monitor can be accessed via virsh.
Note that the --hmp option is required, otherwise the monitor expects the command in JSON format, so omitting it leads to errors like error: internal error cannot parse json ... lexical error: invalid char in json text.

The final hurdle was figuring out the correct monitor command. The host_net_redir command as mentioned in the above article is no longer recognized. Luckily the QEMU monitor interface helped me out here – I spotted an encouraging sounding command hostfwd_add:

# virsh qemu-monitor-command --hmp sles11 'help hostfwd_add'
hostfwd_add [vlan_id name] [tcp|udp]:[hostaddr]:hostport-[guestaddr]:guestport -- redirect TCP or UDP connections from host to guest (requires -net user)

and google confirmed that the latter had superceded the former.

So finally we have the complete solution:

# virsh qemu-monitor-command --hmp sles11 'hostfwd_add ::2222-:22'
# ssh -p 2222 localhost
Password:
Last login: Mon Jan 23 00:37:44 2012
linux-mnsh:~ #

Hooray!

UPDATE: just found another very simple solution – add a new NIC to the VM which doesn’t use user-mode networking. Then it will get a IP (on 192.168.100.0/24 by default) which is still NAT’d but also routable via virbr0 on the host, meaning no redirection is necessary; just ssh directly to the guest’s IP from the host. A minor disadvantage of this is that the guest won’t be directly reachable from outside the host, but that’s unlikely to be an issue in most scenarios.

Reaction from the desktop development community has been swift and mostly unapologetic. Mark Shuttleworth, the founder of Canonical, responded “Once again, Linus is underestimating the importance of aesthetics in computing. You only have to look at Apple to see that people place more importance on visual beauty than the kind of efficient work processes that a flexible and reliable desktop environment enable. This is why we’re pushing our new Unity launcher as a mandatory part of Ubuntu. We’re confident that people will quickly overcome the initial shock of everything taking longer to find and access because they’ll be too busy admiring how beautiful it looks.” He then cited the latest 11.10 release as an example of this. “If you look in the release notes for [Oneiric Ocelot], you’ll see a new Alt+Tab switcher at the top of the list of highlights, and below it other radical changes such as renaming ‘Places’ to ‘Lenses’. Frankly, most people lap up this whizzbang shit, and as long as it looks cooler than their friend’s Windows 7 netbook they’ll be willing to tolerate some minor annoyances which are unavoidable when making immature software a critical component of the desktop. Sure, we could prioritise boring bug-fixing over innovation, but that just doesn’t excite the teenagers on the web forums, and we have to think about the next generation of users. Besides, if you want a dumbed down system that mostly works, there’s always Mac OS X.”

Havoc Pennington, a GNOME developer well-known for initiating the war on Linux desktop flexibility by drastically reducing the number of preferences and replacing GNOME’s default window manager, the high-performance scriptable Sawfish, with Metacity, commented: “It’s about achieving the right work/play balance. If your desktop allowed you to get stuff done too quickly, it would just increase your stress levels. Some ‘power’ users think they want to be able to stream-line their workflows, but we know better, so we are doing them a favour by making this customizability harder. After all, everyone needs basically the same things. Rather than trying to be different, these people should instead learn to enjoy the cute visuals and focus more on having fun. Life’s more than just work, work, work, you know.”

The KDE camp has been slightly less vociferous, perhaps because it’s old hat for them – back in 2008 they pioneered the concept of intrusive redesigns and ended up the wrong end of one of Linus’ rants as a result. “With KDE 4.0, we did our best to prevent people achieving real work, and I think we largely succeeded”, one of the KDE team leaders recalls. “I mean, there was a significant period of time where neither the KDE3 version of knetworkmanager nor its KDE4 rewrite worked properly, so for many wireless networks, the only way you could connect was to disable NetworkManager and write a shell-script to interface directly with wpa-supplicant and ifup. And that’s just one small example.”

Despite Torvalds’ comments, the move towards form over function has been witnessed elsewhere outside the desktop software space. For example, Apple have introduced the MacWheel, a move so bold that it makes innovations such as Unity and the GNOME Shell look positively conservative. However there is no clear industry-wide consensus; in fact companies such as Ebay and Sony are beginning to experiment with rejecting both form and function, turning conventional wisdom on its head.

Anyway, in the mean time a quick google brought up the following solution:

• http://www.dgf64art.com/2011/07/05/installing-amazon-mp3-downloader-on-ubuntu-11-04/ (unfortunately this link no longer works)

Unfortunately it doesn’t work – the step which installs the manually downloaded .deb files fails due to broken dependencies. However further googling found a post from 2008 which revealed a technique based on the very useful getlibs utility.

So here’s my solution:

1. Download the 32-bit Amazon downloader app for Ubuntu 9.10.
2. Run sudo dpkg -i --force-all AmazonMP3DownloaderInstall.deb
3. Run sudo apt-get install getlibs if you don’t already have getlibs installed.
4. Run sudo getlibs /usr/bin/amazonmp3 and answer yes to the confirmation.

At this point if you try to run /usr/bin/amazonmp3 you’ll probably hit Ubuntu bug 781870. The workaround is as follows:

export GDK_PIXBUF_MODULE_FILE=/usr/lib32/gdk-pixbuf-2.0/2.10.0/loaders.cache
/usr/bin/amazonmp3

You’ll still get an error that it’s trying to load the 64-bit version of libgvfsdbus.so thanks to Ubuntu bug 369498. I had hopes that export GIO_EXTRA_MODULES=/usr/lib32/gio/modules would fix this, but it seems that this variable only gets honoured too late. However, apparently this issue doesn’t stop the program working so can be ignored.

Another option is to use Banshee’s built-in Amazon downloader, but even without all the politics surrounding Ubuntu’s version of Banshee this didn’t suit my tastes.

UPDATE: Wow. Just found out Amazon doesn’t support re-downloading stuff you’ve already bought. This is truly pathetic, especially considering their Android app kind of implements a locker service. From now on I’ll be using 7digital whenever I can – unfortunately their selection isn’t as big though. The quest for the perfect music services continues … :-/

Finally started using my CompuTrainer a bit more. Just rode 20km in the lounge with the computer pacing me at 200 Watts. After so much riding with guys way stronger than me, the temptation to draft the computer is difficult to resist. However in the last 500 metres it spontaneously decided to make a sprint for the finish line – the bloody cheek! Obviously I couldn’t allow a bunch of transistors to beat me so I sprinted after it and hit the finish line 0.05 seconds ahead, narrowly avoiding embarassment. Next step is to start doing brick sessions …

By the way, in case anyone googles for computrainer linux, I am actually running it under VMware Workstation on LinuxMCE 0710 which is based on Ubuntu Gutsy. It works great with the unfortunate exception of the Interactive Real Course Videos (IRCV), which seems to be due to missing support in VMware Workstation for VMR9. I’ve posted about this in both VMware Communities and the RacerMate Forum but noone has been able to suggest a solution so far.

UPDATE 15th February 2010 – great news! Check out this fantastic new piece of software called Golden Cheetah!

There were other things, like korganizer getting slower and more buggy, phonon not working right, plasma’s bizarre customisation UI and tendency to leak memory like a sieve, a device notifier pop-up which often makes new devices invisible and unclickable … but enough is enough, life is too short. GNOME does what I need (which isn’t much, thanks to the flexibility of openbox).