Secondly, I’ve been learning a little bit more about photography, and so here’s a little preview of one of my photos. I decided to burn-in the Gentoo logo on the bottom right so that it serves as a nicely patriotic wallpaper. It’s a vague enough shape to be mistaken for part of the picture, but recognisable enough to be Gentoo (I don’t like in-your-face logos). You can download a high res version here. Here’s a snapshot for the lazy. Perhaps other Gentoo users might appreciate another wallpaper!

Finally, I’ve purchased myself a basic VPS plan from JohnCompanies – of whom is the parent company of a very commendable company called rsync.net, of whom I’m still very happy with. Although not as cheap (as in, cheap + high quality reviews) as alternatives such as Linode (who offers double the resources at the same price), I went on a gamble that my great technical support experience will transfer over into a similar great experience.

Unfortunately, JohnCompanies does not offer Gentoo on their VPSes, only on their dedicated server packages. After some quick debate, I went for Debian. I shall proceed to migrate a few of my sites to this new server as well as a few of my existing hacked-together serverside toys. If you experience any downtime or shoddiness with any of my sites (blog + email included), it’s probably just due to the migration.

No related posts.

Happy (belated) birthday, Gentoo!

No related posts.

NX is a pain to set up. Maybe it’s just my bad luck that something goes wrong every time I try to set it up, or perhaps I’m just plain stupid but I find it a pain and it normally makes me feel like chewing my ethernet cable after a while. A long while back I managed to get nxserver-freeedition working fine, but recently it stopped and in the process I decided to switch over to using FreeNX – an open-source implementation of NX. Here is how I got it up finally, mainly for archival purposes.

Before starting, you may want to clean your system. unmerge nxclient nxnode nxserver-freeedition nxserver-freenx, etc, rm -r /usr/NX (if freeedition), rm -r /etc/nxserver (if FreeNX), remove the nx user and the nx group, rm -r ~/.nx/ for all users.

1. emerge nxclient
2. emerge nxserver-freenx
3. nxsetup --install --setup-nomachine-key --clean --purge (the –setup-no-machine key option means that you just need to use the default key that the nxclient comes with)
4. Try to connect via nxclient, tail -f /var/log/messages, if it complains about the user being locked, do usermod -p foo nx, where foo is a password you’ll use.
5. Start debugging why it doesn’t work.

When debugging I normally:

• Make sure SSH works first. Refer to my sshd_config (obviously this is specific to my setup, but notice that PasswordAuthentication is set to yes, as that’s how you’ll authenticate with NX with nxclient)
• Make sure you own your own authorized_keys(2) file, after a while of mucking about you might’ve screwed the permissions (600).
• Check that you’re pointing at the right authorized_keys(2) file, though I didn’t need to bother about this with FreeNX, I know that this is a common mistake with nxserver-freeedition
• I like to run SSH on port 443, so change to port 443 in /etc/nxserver/node.conf for FreeNX, and if using freeedition in /usr/NX/etc/server.cfg, node.cfg, whatever is specific to your setup.

As for my experience switching over to FreeNX from freeedition, honestly in terms of usability it’s no different at all, but personally in terms of installation procedure FreeNX wins by a mile. But for the moment I’m just happy I won’t be restricted by my school’s web filtering system and glitchy, laggy, virus-filled OS. Perhaps when NeatX stabilises a little I might check it out.

No related posts.

Had to install it using the SystemRescueCD as Gentoo’s minimal install didn’t have the module for my NIC. Xorg is compiling, the holidays are almost over, and it’s time to overload my schedule again.

To make this a bit more computer-relevant, I ask you: what do you suggest I do to help make it “easier” to use for someone like my mum? I am planning a cron-scheduled usual sync, update, revdep-rebuild. I don’t think I can automate the etc-update but that could pretty easy to train I think. Kernel updates is going to be a hassle. She wants KDE and so that means unstable packages.

Sounds fun.

No related posts.

So the other day, I came across the original Chinese hacker dude who created the Shellex-overlay for running Chromium on Gentoo Linux. Unfortunately the blog post was in Chinese. I immediately enrolled myself in a chinese tuition class an-

No, I fired up Google translate and stole my brother’s Chinese girlfriend to properly translate sections that apparently read “building know not updated porridge“.

Well, I don’t really like languages. Not that I don’t think they’re really useful and stuff, I just don’t like them because 1) they suck, and 2) I suck at them. The reasoning I like to use for “they suck” is the theory behind natural languages, then moving on to mathematics and programming languages and showing how “look. 1=1 is 1=1″. However like a friend of mine, one of my pet peeves is people “who wan talk talk leik vely good you know wan liao!“. (a representative sample of the local accent and style of speaking here. (see Wikipedia entry on Manglish – it’s friggin’ made up of English, Malay, Hokkien, Cantonese, Tamil – oh and you’d find “g’day mate” popping up once in a blue moon) – yeah, so I can’t exactly say being decent at a language is unimportant.

In fact language is really important. Especially when maintaining professionalism. You can’t argue – having good language skills are vital for…communicating well? Wow, that sounded bad.

…and for that reason when some Taiwanese guy lurking in the #gentoo-chat channel (sinsun) introduced me to the website Lang-8, I signed up. It’s a website where you can write journal entries (like a mini-blog) in another language, and other users who are native speakers of that language can come along and insult you.

So you know right away, I go allleady what lah and start correct the crap crap english udder people is writing wan.

No. I actually did start jocking up whatever was left of my 汉语 knowledge and now I’m learning that I sucked even worse than I thought I did. Well, it’s still an amazing website, and I’ve not seen one like it before. I would highly recommend it to anybody thinking to brush up on a foreign language. The correction system is pretty nifty (like crossing out stuff and highlighting) and it has one hell of an active community – you’d get responses literally minutes after you post it.

Feel free to add me as a friend, especially if you’re good at Chinese and can teach me how to say “Git repository” in Mandarin. My username there is “Moult”.

Note: I actually really do think having good knowledge of a language is very useful – up to the point where you start thinking “Here is a piece of paper” is some sort of symbolic metaphorical imagery for racism.

No related posts.

OK, back to what’s new and amazing. Some days ago I wrote an article on RememberTheMilk, a really awesome to-do list website (I have 60+ and counting tasks listed over there now!). However a main issue with it is that even though it’s extremely accessible through lots of mediums (phone, email, twitter, plasmoids, etc) – they are all graphical! We’re missing a command line interface for it!

Well, not quite so. With some Google-fu I found some French guy with a fetish for white rabbits (no, seriously this time I’m friggin’ sure he’s french) who made a command line tool. It’s not much more than a script, but that doesn’t stop me from me from putting it in the sunrise Portage overlay so Gentooer’s can get it! Actually, what did stop me was the fact that 1) I didn’t know how to write ebuilds, 2) I didn’t have an account to commit ebuilds to the sunrise-overlay, and 3) I didn’t have a GPG key (for part 2).

So, a while later learning about ebuild writing, getting a key and a commit-able account, babysat ever so generously by hwoarang and idl0r for the ebuild part and scarabeus for the key and account, I have today commited two ebuilds to the sunrise overlay (layman -a sunrise). The first is app-misc/rtm, which is the tool itself, and the second is dev-perl/WebService-RTMAgent, which is a Perl module (dependency for rtm). I’m not sure when it’ll get into the publicly approved sunrise overlay, but it’s definitely there in the developer checkout (Unless I borked up the commit).

So, install it, try it out on your architectures (I’m only ~amd64 dont’ forget), and enjoy! Hopefully this’ll mark the start of more Gentoo contribution. Or it will if I don’t get distracted and play this game about white rabbits. – or maybe this one, which is more related to rabbits – in more ways than one.

What the hell do rabbits and ebuilds have in common? :P Stupid French guy.

No related posts.

For those that don’t know what SSH is, it is basically a network protocol (for example like FTP, SMTP, etc) that allows you to securely connect to another computer. For those that don’t speak jargon, it is some cool thing that allows me to use my computer remotely.

One of my well-visited locations is my college. I wonder why :P … and like most places, it runs Windows. Using a Windows computer leaves me feeling crippled and with a sense of repulsion at the most innocent of small creatures. Combined with my college’s restrictions, there is a lot of stuff I can’t do. For example, I can’t download a .doc file. Also, it is quite troublesome to constantly transfer files over with a memory stick, so I decided to set up SSH.

Little did I know how pathetically paranoid the IT technicians were.

Problem 1: setting up SSH and connecting to my dynamic IP.

The first step was to install (emerge openssh) and set up SSH. (I run Gentoo- the steps will be different for your distribution or if you are running Windows – say, you can set up SSH on Windows, can’t you?) This was simple. Now the problem here is that my IP keeps on changing. Especially because my ISP’s connection is quite volatile, my IP is dynamic and resets several times a day. The method to solve this was to set up a dynamic dns, available from dyndns.org. This is a free service, and allows me to connect to a sane domain name whilst a client running on my machine updates it regularly on the latest IP.

Problem 2: port 22 is blocked.

The next day I popped PuTTY on a thumb drive and tested it out – or at least tried to. I had a network connection refused error. Later that night I learned that most public networks blocked certain ports, for example port 22 which SSH normally uses.

The fix around this was to change the config in /etc/ssh/sshd_config to use Port 443. You see, whilst http:// defaults to port 80, https:// URLs default to port 443 – and are hence rarely blocked. Et Voila – you can now connect! However, I also run an Apache webserver. This clashes as it also tries to use port 443. As I don’t care to serve SSL webpages on my localhost, I decided it was a decent sacrifice, and I removed Apache’s HTTPS support by removing `-D SSL -D SSL_DEFAULT_VHOST` from /etc/conf.d/apache2. Tada. I can now access SSH at my college.

Problem 3: What about SCP? Surely you’d want to copy files over.

Well, transferring files over is another issue. It’s all good with your vast array of Command-Line-Interface applications for IRC, Vim text editing, file browsing, MSN, email, calendar and PIM, etc – but every so often you will have the need to transfer a file over. Sending yourself an email with a file doesn’t help, as my college blocks almost every single website out there aside from Google.com (note: Gmail is not Google.com – so it is blocked too). At the same time, it also blocks downloading every single file type aside from image files. The quick fix for this was to put the file on my Apache localhost, change the file extension to a .jpg or something then download it from there.

But no. Two problems occured. 1) My website was blocked – haha, and 2) the IT technicians filter it not by extension, but by actual file contents. To get past the website block, I run a proxy script on my webserver (there are plenty of free proxies out there too) to access my webserver. But then of course I’m stuck with the file filter. So looks as though it’s game over.

Not really, there’s always SCP. I couldn’t send files using my computer as I didn’t know the network information of my college network. So I decided to SCP using the college’s Windows computer. I hear there’s a program known as WinSCP which is pretty nifty, but at the time I only had PuTTY’s collection of tools and thus PSCP.EXE – which pretty much did the same thing – with one catch: it’s a CLI application. You see, they’ve also blocked the command prompt. OK – for understandable reasons.

To get myself a command prompt, I did the age old innocent command.com trick. This involves creating a plain text file (eg: .txt), putting the words `command.com` in it, and then saving it as cmd.bat. (Notice the changed file extension). This will give you a prompt to work with. Now – using that to run PSCP.EXE, I succesfully transferred my target file over – which was in this case a .doc file, which contained some homework.

What about the ethics of this?

Well. I personally feel as though these workarounds are anything but a way for me to do my work conveniently. The computer system is riddled enough with viruses as it is without my doing, and I doubt anybody will be motivated enough to have such a complicated setup unless they were either particularly vicious or needed a file really urgently (such as me at the time). But seriously – a learning centre blocking .doc files?

If you have more experience in networks than me and feel as though this article is innappropriate, feel free to contact me and I’ll willingly take it down.

No related posts.

The other day I was surfing the web and read an article about Google Chrome in some sort of hacking competition – this then prompted me to check out Google’s progress on porting Google Chrome to Linux and Mac. For those that don’t know Google Chrome is Google’s attempt at making a browser. So far it seems like a really good attempt.

It seems as though lately the Linux builds (I ignored the Mac stuff – but I hear it’s getting good too) seem to be getting to a usable state. Definitely not finished, definitely buggy, but usable. So, like any other Gentooer, I began trying to find out how to get it.

Step 1) Any ebuilds out there?

Why bother do hard work myself if somebody’s already put it in portage? With some google-fu it seems as though there are a couple ebuilds. One by the French, and another by the Chinese. The French one (have not tested) is available in the `THE` overlay, available by doing layman -a THE. The chinese one seems to be called “Shellex-overlay”, and can be accessed here. I’m not quite sure what the French one does as the ebuild didn’t really like my amd64 system (note: Google Chrome only supports 32-bit as of writing). However the Chinese one fared better and provided me with a binary. If you are on a 32-bit system (x86) you should try those ebuilds.

If you don’t want to compile from source, check the depencies list just a bit further down, then check out the build bot. Note: the build bot provides binaries for Windows, Linux AND Mac, so if you’re on a Mac, you’re in luck!

For more information, you should visit the Chromium Linux Building page.

Under `Prerequisites`, it lists down the dependencies as packaged by the Ubuntu system. Here is the list of dependencies as what Gentoo calls them:

• Python >= 2.4
• Perl >= 5.x
• gcc >= 4.2
• bison >= 2.3
• flex >= 2.5.34
• gperf >= 3.0.3
• pkgconfig >= 0.20
• nss >= 3.12
• gconf
• glib
• gtk-engines-murrine
• nspr
• corefonts
• freetype
• cairo
• dbus

Their version requirements are listed as needed.

Step 2) What about 64-bit systems?

There are several techniques of getting Chromium on a 64-bit system. However no matter what, I highly recommend that you create a 32-bit chroot. If you want to track each library individually and symlink your system to hell (as I first attempted), be my guest, but you’re wasting your time. So, first create a chroot by following this nifty guide.

Once you’ve got your chroot up, you can either try out the ebuilds I mentioned before, compile it yourself from source (via Google’s instructions) or be lazy and grab the binary from the Chromium build bot. I have tested the latter two techniques (can’t trust the French nor the Chinese!). To compile it yourself from source, follow the Chromium Linux Building page. Note: you will require quite a bit of HD space (the sources tarball itself is 640MB+), I also suggest you bootstrap from the tarball, the subversion checkout seems a lot longer and a waste of time really. Finally, if you’re just interested in getting the binary and running Chrome (not development), I would do use Release mode (see the building page for instructions). Of course, after making sure you have the dependencies I listed above, you should have Chrome compiled!

If you’re lazy and don’t want to compile, there is a build bot.

Step 3) Run Chromium on Linux!

At this stage, you should have the Chrome binary (either by compiling from source or ebuild, or getting the build bot binary). Now you just have to run the program and enjoy. If you’re running using the chroot, you should use the xhost hack. Do xhost local:localhost outside the chroot, then try run the binary again. Obviously you don’t want to waste time setting up X in the chroot.

Finally, here is a screenshot of Chromium running on Fluxbox! (I normally use KDE, but I wanted a more lightweight WM when compiling Chromium) You might also consider doing nice -n 10 when compiling if you want to continue doing your stuff. In fact, I’m running Chrome right now on KDE to write this post. It’s very fast, uses about 1% CPU, separates itself into different processes per tab, and so far seems pretty “stable”. However I have found that opening a file browse dialog (eg: in an upload form) makes Chromium jump up to about 50% CPU, which sucks.

Any thanks, issues or problems feel free to ask.

No related posts.

It’s hard to describe it all, so perhaps take a look at the screenshot below. In it you will see task grouping in the taskbar, more transparency in the taskbar, UI upgrades that make it look prettier (look at previous screenshots in my previous posts), desktop effects that work better, more plasma apps, a folder view as a desktop for traditionalists, custom glow colours around windows … well, here’s the screenshot: (click for higher resolution)

Of course, there are much more not shown in the screenshot, such as a prettier start menu (Oxygenised), Vim mode in Kate (woohoo!), upgrades to konqueror, a superb krunner, Strigi file indexing (seriously, I typed sc[rot] and it already popped up with loads of files), more desktop effects (with the almighty desktop cube), upgraded panel settings … wow, and I’m discovering new stuff all the time! I must say, it’s amazing how well the developers are doing! (It’s feature freeze now though).

I have been able to crash it though, the ZUI gives some nasty screen offsets if I use it too much, as well as plasma crashing every time I try to change back to a Desktop (instead of classic Folder View). The nepomuk thing lurking in my systray seems to have a habit of locking my X whenever I try to configure it too. Evil thing. Evil.

Aha, and as a final note, I’m finally using KDE without lag (or very very minimal lag)! Kudos to all the folks working on the nvidia-drivers, as installing the nvidia-drivers package (v180.11) from the berkano overlay significantly improved my graphics performance. Hurrah!

So anybody here still on a stable KDE, it’s time to get risky and install whatever kde snapshots your distro provides you with. It’s worth it (you can run both stable and unstable at the same time and choose which one to log into via kdm).

Finally, sorry for the lack of regular postage, as I’m somewhat busy with an upcoming exam (not to mention other things). However, I do believe I’ve got several posts in my drafts folder awaiting competition including a short (positive) rant about MVC frameworks in PHP, the ISSDC Aerospace competition, and another How to Use Your Computer part. So stay tuned.

No related posts.

Right now, I’m reffering to Gentoo. Gentoo is a Linux distribution which targets a very niche market of what normal people consider the geek elites. To put it in perspective, the #gentoo irc channel is a constant market of kernel builds, compilation issues, hardware discussion and other really geeky things. Why don’t you go in there and try take part, you’d find you’ll want to hide in a corner in shame. All your base are belong to us.

So what is the wonderful thing about Gentoo that makes people say “Yes! FUBAR is the way I like it!” Well, simply put, it’s maximum configuration. It’s pretty much the closest you can get before LFS takes over. You know everything you put into your system, where it is, why it’s there, when it’s going to go pop, how it upgrades, when it upgrades, how much space is taken, why your directory structure is like it is, why your sound works, why your printers work, how your internet works, how your LAMP setup works, what programs communicate with what, what packages are needed by desktop environments, what options each package has, what the source code of everything is … the list goes on … and on … (recurring)

So, how to communicate the ultimate build-it to break-it configuration to the average Joe? Let’s take a look at the steps I took to set-up the new laptop I’m now using over the past week.

1. Download and burn Gentoo Minimal Install CD for amd64.
2. Boot up, follow the Gentoo handbook, setting up Internet and compiling kernel as well as base system.
3. Reboot and tada – Gentoo works. (I really didn’t type justice about that Gentoo installation stage, because if I did, I could write several essays)
4. Emerge irssi (for IRC), links/lynx (for web-browsing), vim (because it is the meaning of life) and screen (for multitasking in console)
5. Using screen, irssi and links simultaneously to follow the handbook and troubleshoot problems, set up the X server until you can run twm.
6. emerge fluxbox and ratpoison, startx into fluxbox.
7. Keyword ~arch for mozilla-firefox-bin, openoffice-bin, portage, gimp.
8. Emerge portage, and get sets for KDE.
9. Add xcomposite and opengl global useflags for special effects :)
10. Emerge firefox, openoffice, gimp, and the entire @kde.
11. Emerge kdeplasma-addons: can’t live without it.
12. Emerge mplayer, alsa, cups.
13. Configure alsa to work, configure cups for printing.
14. Add appropriate useflags for a LAMP setup, then emerge apache, mysql, php, phpmyadmin.
15. Configure LAMP setup.
16. Configure sshd, and emerge pure-ftpd for ftp. Add them to boot-level.
17. emerge scrot (of course!), cowsay (for hilarity!), x-sane (for scanners), then GVim (if you like the graphical version)
18. Download Blender.
19. Emerge libsdl – because Blender needs it.

Along the way, you’d probably want to emerge the necessary packages for slocate, lspci, lsof, etc. Well, that’s my general plan when setting up a system. I hope it helped somebody!

Sorry for the rush post, but I’ll probably add more detail in the future.

No related posts.