Due to some questions on #lisp, I decided to throw together a quick, step-by-step solution to setting up a basic, but working, Common Lisp installation on Linux. The implementation used will be SBCL

This HOWTO assumes basic linux knowledge as well as knowledge of your distro’s package management system.

1. Install SBCL from your distro’s repo or download binary build from SBCL website
2. Install Emacs – I recommend Emacs-22 or 23. Again, for fast install, I recommend going with your distro packages – if you need/want specially patched version, you probably don’t need my hand-holding
3. Install clbuild
4. run cd clbuild; ./clbuild check and install any remaining dependencies.
5. Depending on your distro SBCL packages, install local build of SBCL through ./clbuild compile-implementation sbcl. It will be used by clbuild after that and saved in ./clbuild/target
6. Install SLIME: ./clbuild install slime
   and configure it by adding result of ./clbuild slime-configuration to your emacs config
7. Test it by running emacs and invoking M-x slime
8. Now go over to Cliki and maybe entertain yourself to Practical Common Lisp

Well, yesterday I finally did the “big move”, namely moved to Aberdeen for my 5 years of studying at University of Aberdeen.

I traveled by plane, first from Warsaw to Gatwick by British Airways on a B737-400, then a long >3h wait at Gatwick for the second leg of the journey: Gatwick -> Aberdeen by FlyBE. One thing to remember – “low cost no more”. Also it is interesting to be the only passenger with ticket troubles on an already delayed flight… And on “normal” airline, you get refreshments for free xD

After landing at Aberdeen International Airport, I traveled again on 737 – this time it was Stagecoach 737 line from Airport to Aberdeen Bus Station, then got off early and walked <1km to the dorms, finally arriving at around 2200 hours UTC – after starting my journey in Warsaw at around ~1130 hours UTC.

And the parts of Aberdeen I had already seen look pretty nice. Also lots of Polish language spoken there, at least during my walk from dorms to Uni.

I don’t know about others, but for a long time I had problems with using WPA with my Broadcom wireless card in Ruri. Thankfully all troubles are now gone :)

The steps for a working wpa_supplicant are simple:

1. Download the latest wpa_supplicant. I recommend at least version 6.3 (5.x didn’t work for me)
2. Install it and configure it to use Linux Wireless Extensions driver interface, for example by supplying -Dwext to it’s command-line options. Exact configuration depends on your linux distribution – on Gentoo it’s a wpa_supplicant_interface="-Dwext -iinterface" line in /etc/conf.d/net
3. Start it up and voila – wpa_supplicant should be able to properly associate with networks from now on. Do some checking with wpa_gui

Happy wireless surfing :)

So I downloaded and installed QTM.

I have to say that it surprised me from the building process: It was a first application that used cmake which I had to build myself,
and it shows clearly that Cmake is superior to Autotools.

The only dependency was QT >= 4.1.x, and it worked like a dream with QT4.3.
The building process itself was a breeze compared to what I would except from QT application (And CMake showed it’s beatiful scripts again :D)

As for the application itself: IT’S FAST!, I mean, REALLY FAST. Sure, it’s about tasks that don’t consume too much resources, but it snappy in a way that doesn’t always show with today’s apps. I also didn’t even notice when it downloaded my categories after inputting my blog settings in configuration window.

There are however some drawbacks – It doesn’t seem to have support for editing existing blog posts, support wordpress tags (But it might work from the technorati tags page) nor uploading media objects. Also, the interface is a little different, but usable.

Now let’s see how it will publish this post :D

Well, it had some problems with uploading the post, and a progress bar would be a good thing :D. However, it worked pretty nice. And I found it does support uploading files!

I have to say that I basically hate inputting blog posts over WordPress web interface, especially since I don’t have stable internet access, which makes it much harder to use.
So today I decided to test another blogging app: Drivel from Gnome, version 2.0.3.

Well, it certainly lacks some features, and might not win against weblogger.el It can be used with WordPress by setting MovableType interface, however, there
is nothing for uploading images or setting tags. Also, posts can be only in one category.

Looks like I’ll have to write a client and a supporting it EmacsLisp interface anyway, so I’ll have muse/org-mode/whatever support ^_^”

Drivel managed to doublepost because of net error… I need something that is more reliable. Maybe QTM?

Today, I decided to test blogging features of Word 2007, so here is a test post with it :-)

Although I’m not a big lover of Microsoft, I have to say that the new Office 2007 suite is working pretty good – IMHO you can see a steady improvement, although it’s disk space requirements have gone thru the roof, like 2-3 times more space is required for full install compared to full install of Office 2003.

Now, if the changes in programming style and design forced after NT6 will start to become widespread in Microsoft, we might see a fair surprise around NT7. Of course, if they don’t screw it again… (Hope not, even If I wish in my heart for Microsoft’s collapse).

ああ。。。日本語のテストで ;-)

EDIT:
I edited the code after all, but it doesn’t look so bad compared to some of the atrocities Word used to output as “HTML”. There is progress :-)

Well, it’s over.

Thankfully, this time the attendance had risen – it’s speculated to be over 55,3%.

If the final number of voters exceeds that, then it’s much more important than the result itself.

As Rughalt says: “If ~90% of people allowed to vote had given their votes, then I would not complain about Poles being stupid in voting.”

Anyway, Methinks of throwing a party at Inaba for PiS’ great failure :>

Requirements:

• Knowledge how to modify module set of kernel or roll your own
• linux 2.6.x
• Laptop with SMC IrDA chip

Many laptops nowadays come with IrDA served through an SMC IrCC SIR/FIR controller. Unfortunately, it’s setup is not done “by default”, leading people to think that IrDA simply doesn’t works. This is not true. Here’s a howto (not full, you have to read your distro’s manuals about how to adapt it to your system. Console commands however are universal on 99% percent of distros)

In order to get it working under linux, you first need to either modify your current kernel or compile a new one, ensuring that it has all IrDA features you want as well as option CONFIG_SMC_IRCC_FIR=m is set in your kernel config. Make sure it is set for compiling as module, not compiled-in!

After you get all of this compiled with all IrDA tools you have been dying to use ;-), it’s time to load the driver.

So you eagerly try to cast modprobe smsc-ircc2 spell, only to get the dreaded “No such device” error. But don’t lose hope! It’s just that the driver can’t find your chip – so you need to specify it manually.

So we go for scavenger hunt searching for the lost chip :-)

First, invoking modinfo smsc-ircc2 will show us what parameters the driver accepts. We are going to specify them by hand, but how?

In order to find required parameters, you have to visit /sys/bus/acpi and search for a string starting with SMCF.
Or do it Unix way, and invoke grep FIR /sys/bus/acpi/devices/*/path and it will show you which device is marked as FIR controller. On Ruri (Acer Aspire 5112 WLMi laptop) it yields this:


ruri / # grep FIR /sys/bus/acpi/devices/*/path
/sys/bus/acpi/devices/SMCF020:00/path:\_SB_.PCI0.LPC0.SIO_.FIR_
ruri / #


It shows us that an ACPI device identified by string “SMCF020:00″ is the controller we are looking for. Now it’s time to get it’s parameters.

SMC IR controllers use the same identifier on ACPI bus and PnP bus – so we use the identifier we got in last step to find right device:

ruri / # grep -i 'smcf020' /sys/bus/pnp/devices/*/id
/sys/bus/pnp/devices/00:0a/id:SMCf020
ruri / #


Having found the right device in the tree (in this case, “/sys/bus/pnp/devices/00:0a/”) we read it’s resources from /resources:

ruri / # cat /sys/bus/pnp/devices/00\:0a/resources
state = active
io 0x2f8-0x2ff
io 0x228-0x22f
irq 3
dma 3
ruri / #


As we see, we have the IRQ number, dma channel number and addresses of IO space (or ports). Now we have everything needed. Load the module again, supplying it with manual parameters about your irda controller.

ircc_cfg

Put first IO space here, only first address (0×2f8-0×2ff becomes 0×2f8)

ircc_sir

Same as above

ircc_fir

Enter the beginning address of the second IO space

ircc_irq

IRQ number

ircc_dma

DMA channel number

Example: ruri / # modprobe smsc-ircc2 ircc_cfg=0x2f8 ircc_sir=0x2f8 ircc_fir=0x228 ircc_irq=3 ircc_dma=3

You should see now in your logs something like this:

ruri ~ # dmesg|tail
[ 9096.388391] Overriding FIR address 0x0228
[ 9096.388397] Overriding SIR address 0x02f8
[ 9096.388421] SMsC IrDA Controller found
[ 9096.388422] IrCC version 2.0, firport 0x228, sirport 0x2f8 dma=3, irq=3
[ 9096.388520] No transceiver found. Defaulting to Fast pin select
[ 9096.388680] IrDA: Registered device irda0
ruri ~ #


Add your manual options to your distro’s module settings (google for it) and you should beable to use your irda tools normally.

I found one interesting thing about the spam blocked from my blog.

It looks like ‘unya’ gets spanish-or-portugese spam to my blog….

At least Akismet works good :-)