Scienco.org

Just tried a quite funny thing with my bank.

In my virtual machine (VitualBox), I had to use home banking. For fun I tried with IE first -- it wanted to install some fance ActiveX-thingie -- but after the install and a reboot, it still didn't work properly. I couldn't find my keyfile, and because of that I had to log in with extended credentials. When logged in, I had to make a wire, but then IE crash when I should approve it.

This caused me to use FireFox -- this just worked using JRE.

Thanks, FF!

As written earlier, I'm now able to build OpenMoko in a reasonable amount of time due to the purchase of a new computer. I haven't had time to hack that much with OpenMoko in the past week because of a busy week (we made a one-week-project in algebra, so I'm governing Knapsack, Elgamel, group-theory etc. ), but today I had some time to play.

Read the rest of this entry »

I've just - briefly - tried QPE (Qtopia Phone Edition) on my Neo 1973. I tried both the image on http://www.qtopia.net/modules/devices/openmoko.php and http://buildhost.automated.it/OM2007.2/?M=D and with two different SIM-cards.

Read the rest of this entry »

I'm quite lazy, so I'm using the authorized_keys-thingie (I actually don't know the real name, but I'm sure you know what I'm talking about ).

The reason for this post is a "feature" I just discovered today. Maybe it's obvious for everybody (except me), but I was aware that newlines in the remote authorized_keys-file mattered. But they do! So please remember to put every entry in one (1) line only!

Before I did it I got the errors like:

debug1: Host '[host]' is known and matches the RSA host key.
debug1: Found key in /home/[username]/.ssh/known_hosts:3
debug2: bits set: 524/1024
debug1: ssh_rsa_verify: signature correct
debug2: kex_derive_keys
debug2: set_newkeys: mode 1
debug1: SSH2_MSG_NEWKEYS sent
debug1: expecting SSH2_MSG_NEWKEYS
debug2: set_newkeys: mode 0
debug1: SSH2_MSG_NEWKEYS received
debug1: SSH2_MSG_SERVICE_REQUEST sent
debug2: service_accept: ssh-userauth
debug1: SSH2_MSG_SERVICE_ACCEPT received
debug2: key: /home/[username]/.ssh/id_rsa (0xXYZXYZXYZ)
debug2: key: /home/[username]/.ssh/identity ((nil))
debug2: key: /home/[username]/.ssh/id_dsa ((nil))
debug1: Authentications that can continue: publickey,password
debug3: start over, passed a different list publickey,password
debug3: preferred gssapi-keyex,gssapi-with-mic,gssapi,publickey,keyboard-interactive,password
debug3: authmethod_lookup publickey
debug3: remaining preferred: keyboard-interactive,password
debug3: authmethod_is_enabled publickey
debug1: Next authentication method: publickey
debug1: Offering public key: /home/[username]/.ssh/id_rsa
debug3: send_pubkey_test
debug2: we sent a publickey packet, wait for reply
Connection closed by [IP of host]
username@desktop:~$

Indexing is for computer abusers, not computer users - tough statement, but for me it's right . In combination with mdadm (and especially when compiling OpenMoko) I've had some serious performance issues. The disk was working all the time - and it was not just syncing the /dev/md(/)0-device.

So just a quick remark: disable every indexing service if you - as I - do not use it anyway. I've not done dedicated testing, but by normal usage it has been quite painful, so to speak.

Another pro is that my computer isn't as noisy because of the modest disk usage, i.e. disk I/O.

Yeah, I'm back again. Couldn't settle with the package management in OpenSUSE, although it - besides that - seemed quite okay.

I had no WLAN (wireless network) card, so I had to buy one in order to avoid the cables all over my apartment. And so I (accidentally) did, a Linksys WUSB54GC. After a few minutes surfing, I found out that it had the Ralink RT73 chipset, yielding an installation of the rt2x00-kmp-default-package:

zypper install rt2x00-kmp-default

I tried with the rt2500 at first, but it didn't seem to work - as far as I read, it's also to another chip-set, so it's not that weird after all.

Besides this I also installed a firmware from http://www.ralinktech.com/ralink/Home/Support/Linux.html (Firmware RT2501USB(RT2571W/RT2671)), but I really don't know it's actually required - and I'm to lazy to find out (it's Saturday!).

Well, the Ubuntu installation didn't went good after all, so I've just installed OpenSUSE 10.3 (Gnome) instead. It's actually (almost) my first time using OpenSUSE. My first impression is mainly good, but the repeatedly repository caching during package management annoys me big-time. I really hope it's going to be corrected soon.

Besides that, it seems very professional and I look forward to try it a bit more. I've decided to give it a chance since the installation with RAID and all just went smoothly.

Today I received the parts which in total should be able to make up an entire computer - and it did. Or hopefully . No problems during assembling occurred, but the installation of Ubuntu made some trouble. The trouble was connected to the fact, that I bought two disk so that I could make RAID-1. It was quite tricky to get working (the right SATA-channel, Ubuntu installer-version etc.), but it seems that it's finally working. Before explaining more I probably have to list the specs:

• Intel Q6600
• Gigabyte GA-G33-DS3R
• 2 GB RAM
• 500 W PSU
• 2 x WD 320 GB disk
• Misc. parts

I'm not that interested in hardware, only the things it can do for me . The mainboard provides some sort of SATA RAID aka. FakeRAID, which I read was (roughly speaking) a driver to Windows making the software RAID appear as hardware RAID (or something like that). Nevertheless, I decided to just forget that whicky feature and
just go for the "native" software RAID support in Linux (provided by mdadm - it oddly looks like "madam"!).

In the ordinary Ubuntu-installer (ubuntu-7.10-desktop-amd64.iso) mdadm is not enclosed, but it is in the alternate installer (ubuntu-7.10-alternate-amd64.iso) which furthermore guides you thruogh the configuration. So I'll recommend using the text-based alternate installer.

Before that worked, I had to connect the disks to the ordinary SATA-ports and not the fancy GSATA ports. I don't know why, but as earlier mentioned, I actually don't care if it just works the other way.

I actually made me partitioning a bit different that one would normally do. At servers RAID-1 is also able to avoid some system crashes because everything (i.e. also including system files) is mirrored. I only need to get my data mirrored. If a disk crashes I don't mind whether the system does, too, as long as I have my data. Besides that, GRUB is not working with mdX-devices, creating a need for a separate boot-partitioning. Well, a lot of arguments back and forth I came up with this partitioning:

• /dev/sda
  • /boot, 1 GB
  • swap, 4 GB
  • /, 40 GB
  • 275,1 GB for RAID
• /dev/sdb
  • /xenimages, 41 GB
  • swap, 4 GB
  • 275,1GB for RAID
• /dev/md0
  • assembled by the two 275,1 GB sdax- and sdby-partitions making a /home

This configuration mirrors my home directory, but doesn't protect my system files - it's no problem in my case, I'll just remount /xenimages to a bootable / and/or make a fresh install when receiving a new disk. This set-up further more provides two different swap-partitions. Well, it was the best solution I could come up with.

So in a few seconds I'm ready to use my new computer properly. Guess whether I'm going to compile OpenMoko or install Xen as the first thing - I actually don't know it yet! It's certainly a shame that humans aren't able to natively multi-task yet. Maybe it'll come in the next firmware upgrade.

...is not as straight-forward as I hoped. That's of course because I'm not that experienced at developing to Linux-based environments in C - actually my past consists mainly of Windows-developing with C#. There's quite a huge difference - not only in libraries etc., but also the set-up as such.

Well, the C-coding is slowly returning (I've done some C-coding, but not as much as C#), but I'm still having some trouble setting a proper development environment up. That being said, I really haven't got a whole weekend dedicated for that, but the times I've tried to use MokoMakefil and OE's HelloWorld tutorial wasn't successfully. But that'll come later on, I assure you!

But for now, I'm just developing on my desktop, and then I'll try to get a decen (or just running) environment up and running when the time comes.