In both cases, I was contacted out of the blue by a Google recruiter. The first time I had been considering looking for a new role and pursued it immediately; the second time I hadn't been and put off the recruitment process for several months, during which the same recruiter contacted me again twice to follow up. If nothing else, that's a nice ego boost, but a more cynical mind might be considering the shotgun approach to a narrow recruiting filter and commissions :D

First, a quick data point, I was applying for an SRE(SA) position on both occasions - Site Reliability Engineer (System Administration), because in most of my roles to date, I've been doing both sysadmin and development work and I've never seemed to drift towards one pigeonhole or another. SRE(SA) seemed optimal - interesting sysadmin work on large-scale systems and quite a bit of tool-writing to boot. This was decided on between myself and the recruiter, based on the self-assessment form you are given to fill out. I would love to know how they get around illusory superiority and the Dunning-Kruger effect with those forms, especially given the wierd bias they'd have in the dataset from having so many of the best in their fields working there.

Both times, the process proceeded in the same way:

Dagda, the robot I spent several years of my PhD building - this post is on the cameras and computer system on the robot.

A new toy arrived today – the VisTablet, which is just a rebadged Aiptek 600U (model number WCK-C121). They sell as Aiptek or Waltop or Medion or any one of a few other names for about €65 on ebay.ie at the moment, so I thought what the heck, I do a fair bit of stuff in Gimp so maybe it’s worth a look. A few clicks and a week or so later and it arrives. Cue tearing open the box and unwrapping!

Plugging it into the laptop (Thinkpad R61 running Debian Squeeze amd64) got an instant result in that it found the device and allowed cursor movement… but no actual buttons or pressure sensor stuff, so it couldn’t work as a tablet. So off to google we go and after trying some frustrating messing about with udev rules (to create /dev/apitektablet - which isn’t necessary it turns out) and hal fdi files (to set up X.org options, which are needed) and I’d advanced in a posterior direction.

By now we were through the phase where plugging in the device crashed X and had reached the phase where plugging the device out crashed X. At this point, I found that the problem lay in the wacom drivers. See, just because it’s aiptek hardware doesn’t mean the aiptek drivers actually work – in this case, the wacom driver works much better.

But, just to be annoying, the new version of the wacom driver doesn’t work because it picks up that it’s not a wacom device and commits seppku to avoid being of any use. Note to the coder who thought that was a good idea – go choke on a hairball please.

Solution? Go to linuxwacom and download the latest driver, then untar the tarball, go into linuxwacom-0.8.4-2/src/xdrv and edit wcmUSB.c. … Read the rest

So for the last few days I’ve been banging my head up against the same problem in the lab. We currently use Chillispot to run our access points for the Metakall user trials, which is grand, but it’s also fairly out of date – it hasn’t been actively developed on since 1997 – and it has some serious security issues. So with the new academic year looming, we’re moving over to its successor, Coova-Chilli.

Coova has a few neat features we like, and it’s also being actively maintained, which is important. So I took our testbed server, installed FreeRADIUS and MySQL and Apache and all the custom stuff we’ve written and generally got things set up, and then went to install the latest version of coova-chilli (1.0.14). There was an hour of swearing at iptables as usual (I really am coming to hate that program), and then I thought that’d be that.

But of course not. The Metakall client wouldn’t log in. At first I thought the iptables setup was wrong, but no, all’s well there. FreeRADIUS was fine, so it had to be something in Coova-Chilli. Ah-ha, says I, it’s that new-fangled JSON captive portal setup, that’s the problem, I’ll dig into that. A day and a half later, and I’ve gone through the JSON portal stuff and can’t find anything wrong.

Exasperated, I switch Coova back to the old Chillispot configuration files (which are compatible), bypassing all the new JSON stuff and still there’s a problem.

It’s at this point that I try to log in manually, which no-one’s done on the system since we got the client working. And I notice that usernames are passed perfectly to FreeRADIUS. And so are passwords… up until you feed in more than 15 characters to the password field. Ah-ha. Check … Read the rest

One of the problems with running a dual boot WinXP/Linux system, apart from the fact that you rarely boot into Windows except to play games (and so you tend not to play games much – though FreeSpace2 SCP is helping there!), is that if you have a shared media folder that sees frequent creation and deletion of large files (say, for example, if you were bittorrenting cookery shows or news shows on a daily basis), the shared media folder tends to lead to a high fragmentation rate on the Windows partition (it has to be on the Windows side because while Linux can read and write NTFS with ease thanks to ntfs-3g, Windows has… issues with ext3). As in, 62% fragmented.

And then you start to realise why your dual-core 64-bit 3GHz machine with the 4Gb of RAM is stuttering while you’re trying to learn how to make pork wellington (like beef wellington but with pork tenderloin).

So you boot into Windows, flag the partition as dirty because the defragger won’t work until chkdsk runs, reboot to run chkdsk (which takes three hours to complete), then log in and fire off the defragger. A dozen times. And then, with fragmentation at 59%, decide to try a better debugger. Download the free trial of O&O and fire that off, and wait….

and wait….

and wait….

…*sigh*

That’s a whole day and all night so far, and still the damn disk isn’t happy.

I’m seriously thinking of just saying “Feck it” and erasing the Windows partition, expanding the Linux partition to take the whole disk and just using Virtualbox (which I do 95% of the time that I need anything on windows – which is literally to maintain one single diagram that hasn’t been translated from Visio to Dia/Inkspace/Xfig yet, and to … Read the rest

The more you look at legislation in this country, the nastier an opinion you develop about it. You’d be able to forgive minor errors, small awkwardnesses, even larger problems so long as the common good was served, but the more I look at statute law in Ireland and more critically, at how it is drafted, the less charitable I feel about the drafters. Much of the stuff I see, I see through the Firearms Acts – that stuff I talk about elsewhere.

This time, though, it’s more apropos to here. The new Broadcasting Bill 2009, currently on it’s last stages in the Oireachtas and about to become the Broadcasting Act 2009, has a lovely little sting in it.

In section 140 (the definitions), it defines “television set” to mean:

any electronic apparatus capable of receiving and exhibiting television broadcasting services broadcast for general reception (whether or not its use for that purpose is dependent on the use of anything else in conjunction with it) and any software or assembly comprising such apparatus and other apparatus;

Nice little bit there. What it basicly means is that if you have no TV and you watch Youtube over your broadband connection (or download video footage and watch that), then you need to have a TV licence.

Yeah, that’s a bit of a surprise, isn’t it? We’ve the worst broadband rollout in the EU, in a nation where we prided ourselves on being the gateway to the EU for IT companies, where we have fancy plans for rolling out high-speed broadband to every sheep farmer in Mayo (and wireless broadband for their iPhones while they’re out with the sheep); and now we’re charging a 160 euro tax for those who opt to have broadband installed.

You wouldn’t mind if there was … Read the rest

After the debacle that was the attempted upgrade from Kubuntu 8.04 to 8.10, I sat back and thought about using Ubuntu for a while. In fact, I’d been thinking about it when I wrote the blog post on the upgrade. And the conclusion I came to was this; I started using Linux with Debian, way back in the days when 3.0 was in testing; I used it on my desktop, on my laptops, on my robot, on the lab server and anywhere else it could fit; and I stayed with it up to the point where I needed some hardware support and was too lazy to build from source, and tried Ubuntu because “all the cool kids were doing it”.

And that was a bad mistake. They may bar me ever re-uping with SAGE for saying something as basic as this in public, but stability is far more important than having the suspend-to-ram function working or faster graphics regardless of whether it’s a mainframe or a laptop, if that is, you’re actually doing work with your computer. I did know that at one point. I blame reading too many Rails sites

At any rate, plugged in the external 1394 hard drive, backed up everything (~50Gb in under an hour, sweet), stuck in the netinst cd for Debian Lenny’s amd64 version (feck it, if it didn’t install, I had the i386 handy as a fallback), walked through an incredibly improved installer process, and now I have a pristine system running Gnome, KDE (3), Openbox, Windowmaker, XFCE, LXDE and matchbox (for a project).

The only nonstandard thing in here is that I went and got the 1.2.1 version of Mercurial from the testing repository by hand because otherwise it wouldn’t hg pull from the 1.2.1 repos in the lab.

It’s … Read the rest

What *bangs head on desk* was *bang* I *bang* thinking?

I tried KDE4 before, I didn’t like it. I downgraded back to KDE3 just to get something that worked. I know my new toy upgraded to Intrepid without problems, but that was running Xubuntu, not KDE, so why did I think that a quick apt-get dist-upgrade to Kubuntu 8.10 was going to work on the machine I have all my precious, precious data on?

Argh! This is becoming unacceptable. The whole point of an upgrade is to improve the functionality of the system. In my case, the Intel wifi driver was happily causing kernel panics and locking the machine solidly, with only a sadly flashing caps lock light to indicate that it wasn’t merely “thinking about it”, that it was really right off the deep end into la-la-land and wasn’t going to make it home for supper. Ubuntu forums indicated that the problem is fixed in Intrepid, and it’s been out for a good few months now, long enough to work out the kinks. Right?

Yeah, right.

Not only is the problem not fixed, it’s far far worse – kernel panics every five minutes or so in fact. That’s unbelievable. I had fewer panics than that while writing device drivers for 2.4 for crying out loud. Worse, nothing else works properly either. KDE4 design choices aside, things like menus and submenus render by blotting out a chunk of the screen and then filling it, so you get this distorted and torn section of the screen for a half-second or less before the menu appears. That’s just klunky. Bluetooth, it turns out, doesn’t work at all. It’s in the release notes. Along with a dozen other failings.

Look kubuntu folks, a little word here. A system as large … Read the rest

Just as a small note to anyone trying to get this to work, the fpit driver has a bug in every version of ubuntu up to and including hardy; upgrade to intrepid and you get fully a functional touchscreen again, using this xorg.conf:

Section "InputDevice"
    Identifier     "touchscreen"
    Driver         "fpit"
    Option         "Device"    "/dev/ttyS1"
    Option        "BaudRate"    "9600"
    Option        "MaximumXPosition"    "4096"
    Option        "MaximumYPosition"    "4096"
    Option        "MinimumXPosition"    "0"
    Option        "MinimumYPosition"    "0"
    Option        "Passive"
    Option        "SendCoreEvents"
    Option        "TrackRandR" "true"
EndSection

Just don’t try using KDE4 unless slow-motion work sounds appealing XFCE is reasonable though, so Xubuntu is an option (and what I’m running on it right now).

Now, on with the PyQT4 coding…

Very very sweet. Finding your way round the kernel isn’t easy if your last name isn’t Cox or Torvalds; this is a very useful tool for overcoming such nominal handicaps…

Okay, so I can’t really go into too much of the big picture since this is from the day job, but I can certainly tear into gethostbyname_r() a bit.

For part of what we’re doing, we’re sending a RADIUS message using the freeradius project’s radius client library. So, nice and simple (after you’ve done some setup):

Easy enough, right? So it fails. Specifically, it segfaults and since it’s in a multithreaded server, it’s a pain to track down. And I mean a pain. Hours of fun with DDD, gdb, nana and finally printf() lead to here:

Ah. gethostbyname_r(), the glibc2-reentrant thread-safe version of gethostbyname(). Except that it’s deprecated, and has the unique property of working differently on just about every machine out there.

And of course, in my machine, and the server, it’s going nuts. Not because of a dodgy parameter or anything like that, though that took a while to confirm, it’s going nuts because of the way /etc/nsswitch.conf is written:

hosts:          files mdns4_minimal [NOTFOUND=return] dns mdns4

That’s the standard way to write nsswitch.conf in ubuntu – it first checks the  /etc/hosts file, then uses the avahi daemon, then the DNS system.

Only that’s not good enough for gethostbyname_r(). Grrr. So after a full day bug-chasing through two codebases, the fix is to change a configuration file to this:

#hosts:          files mdns4_minimal [NOTFOUND=return] dns mdns4
hosts:          dns files

What a waste of a day! Gah!

I’m going to rewrite the radius-client library to use getaddrinfo() over the next while. I already have to make some changes to it to cope with other things, I may … Read the rest

Bought for the RCMS project, it’s a hand-held tablet PC (seen here in its docking station). Celeron processor and 256Mb of RAM and a whopping 60Gb hard disk space. Back when we got one for the CVRG robotics lab in 2002, it was the best thing available and cost thousands – today it cost me just over €160 by the time it landed on my door, with customs and vat and whatnot.

The plan is, get the touchscreen working (the fpit driver works but it’s not calibrating properly), and using python, the QT4 library, QT Designer and pyqt, to build a program that could replace the venerable RO Report form used in the rifle club:

It finally arrived!

My old laptop, a Dell Inspiron 1150, is still functional (and yes, a bit slow, but perfectly usable under kubuntu – I’ve even done basic video editing on it), but the screen hinges have both broken, and by cannibalising one I was able to make the other barely fixed, but it’s only a matter of time for it. So it’s getting a noble retirement alongside the Toshiba Tecra 850 and I’m getting a brand new machine with some sweet specs:

CPU is a T8100(2.1GHz), there’s 1GB RAM (I’m ordering more from memoryC right now), 160GB HDD, 14.1in 1280×800 LCD, Intel X3100 (yeah, the graphics won’t play CoD4 which is a shame, I’m wondering if an upgrade is even possible), CDRW/DVDRW, Intel 802.11a/b/g wireless, Bluetooth, Modem, 1Gb Ethernet, UltraNav, Secure chip, Fingerprint reader, WinXP Pro. I’ve kept XP Pro in its own partition (split the HDD evenly because I keep all my mp3 and video stuff in the Win32 partition so both OS’s can play video), but I don’t do any work in XP anymore (except for IE testing and a single project in Visio that I’m trying to migrate to something more usable like Xfig ) so all the work files and data goes into the Kubuntu partition. Might try to get compiz to work just for kicks at some point as well.

The laptop itself is about the same width as the old Inspiron, but not as deep by a good few inches, and while the screen is not as tall (giving a kindof squished look to it), it’s got higher resolution overall and per square inch, so when you actually look at what’s on the screen, the R61 wins out fast. Speaking of fast, it’s shockingly faster than the old Dell. Even with just the … Read the rest

New job, new machine. So the last two days I’ve been setting up hardware platforms for work (one linux desktop machine, one win32 laptop, and some cool toys like an iPod Touch and a Nokia N800 – there are more to come, but it’s more than enough to start with). I’ve been using Kubuntu for work for a few years, and Xubuntu on my ancient (and now breaking) laptop at home (an Inspiron 1100, which should tell you why it’s xubuntu and not kubuntu there as well). So, no-brainer, other people have had no problems so just download the latest ISO and do a quick install, right?

Well, yes and no.

Firstly, I was a bit daft and thought I’d try the new KDE4 variant of Kubuntu (8.10) that came with the new Heron release. Okay, change is bad on a machine I want to be stable, but I figured it’s just an upgraded library with fewer bugs, right? Next time I do my reading first. KDE4 is trying to go down the whole active desktop route, with applets and the like actually sitting there in your desktop background. Okay, it’s not intrinsicly a bad idea for everyone, but I can’t stand it. My desktop background is normally black or deep gray. Occasionally, if I’m in a whimsical mood, it’ll have an image. Right now, for example, it’s this:

But I’m already growing tired of it and will go back to black later today. Sometimes, abstract patterns are nice too.

The point is, nothing active. If I need to see something, I’ll bring it up and look at it. I have some applets in the panel (a holdover from when I used to keep system status information in the dock in WindowMaker), but that hides away when I’m not using … Read the rest

Gutsy Gibbon.

*sigh*

Not a fan of the name. But, I’ve got that X server memory hole in Feisty on my machine and Gutsy supposedly has the fix. Tried it at home last night on my dual-booting laptop (XP and Feisty until last night). Insert CD, reboot off CD, let it run.

Perfect installation. Haven’t had time to play with the neat cool 3D stuff yet, but the successful autodetection and the full setup was very impressive, especially compared to the bad old days when autosetup was a fast way to lock your machine up or worse, but nothing more! Have to say, I still like Debian for servers, but for desktop machines, I’m sold on the Ubuntu model. I don’t install my desktop often enough to become an expert on the process, so an idiot’s interface is a good thing for that for me. On a server, it’d be unacceptable but that’s a different gig alltogether. And the degree of polish is nice as well. There’s something very satisfying about pointing to your desktop as being more bling, bling than a windows users’ and also knowing that it’s a stable unix system underneath. It’s no wonder people love Macs these days.

Sigh…

One of the things I really like about the Ubuntu family of distributions is that you get a good solid chunk of the stability you see in stock Debian Stable installations, but with the release cycle being so much faster, you get more up to date apps. Which is fine for a desktop to be honest. Let’s face it, for a server you want Debian (hell, for a *really* critical server you probably want OpenBSD); but for a desktop you just need it to run for ten to twelve hours between restarting X, and if you get a month’s uptime before a reboot, that’s perfectly acceptable (more time for either, is of course better, but those times are what you need, not what you want). And Ubuntu can do that and give you the cool bells and whistles to silence the iBook-and-Vista-wielding critics (Leopard, you say? Multiple desktops, eh? Vista, you say? Cool spinning desktop changing you say? Hmmm. Have you seen Beryl yet? Old project I know, been doing this for a few years now, but … )

Unfortunately, sometimes the great idea doesn’t work. Right now I’m looking at X.org 7.2 (without Beryl, which doesn’t like my three screen xinerama setup yet ), and it’s using up 342Mb of RAM.

I mean, that’s just daft.

Granted, three 24-bit colour, 1280×1024 screens. Fine. But that’s only about 90Mb in total – and when X.org starts up, sure enough, that’s what it takes up (well, slightly over that). And granted, X.org does swipe memory that’s not in use for caching to speed things up (which is a good thing). But it’s not giving this memory back, and after a few days (or a really heavy Firefox session with forty-odd tabs open), I can be looking at over … Read the rest