would have been released today if mozilla.debian.net was responding. But it s moving to a new server.

tar and cpio tar(5) and cpio(5) are two competing archive formats, that provide almost identical features. Both are streamed formats, originally designed to be used on tapes. Their major practical difference is the style of their standard command line utilities:

• cpio(1) takes the list of files to archive on its standard input, allowing a very high control but requiring to use another utility such as find(1) to archive an entire tree;
• tar(1) usually takes the list of file to archive on its command line, and browses directories recursively be default.

Peace! There is a traditional opposition between these two formats and utilities and their respective supporters. To fix this, the POSIX specification deprecated both commands in 2001, replacing them by a single one called pax(1), that can be used either in cpio-like or tar-like style. This specification also defined a pax interchange format, which these vile trolls derived from tar! Now, considering the actual practice, you can notice that tar has won, and that almost nobody uses pax. Almost every free software project I know release tarballs, not cpio archives. This is probably the reason why the GNU tar utility has evolved much more than GNU cpio: tar now integrates seamlessly with compressors such as gzip, bzip2 or xz, autodetecting the compression format when needed whereas cpio has no such features as far as I know. However, cpio is notably used by the Linux kernel for its initial memory file system, aka initrd. tar and cpio drawbacks Now I come to the title of this article: I personally prefer the cpio format over tar, although I usually use tar because of the advanced features of GNU tar. In fact, tar suffers from a drawback that usually has to be worked around: it cannot store any type of file. There exist seven file types: regular files, directories, symbolic links, character devices, block devices, FIFOs and Unix sockets. Well, tar can only represent six of them, and misses the ability to store sockets. Thus, if you try to archive your entire filesystem with tar, you may get the following output:

tar: /var/spool/postfix/private/rewrite: socket ignored
tar: /var/spool/postfix/private/bsmtp: socket ignored

If you try the same with cpio, you will see that sockets get archived without any problem. Given that the tar format has a type byte, with one reserved value, I am surprised that its 2001 revision did not use it for sockets, but for some reason it did not. In practice, programs that use sockets work around this problem by creating them at startup. To be honest, I must mention that cpio also has one significant drawback: when a file has several links, it gets copied several times in the archive. Edit: It appears that archiving sockets is completely pointless, given that they must be unlinked if they exist, and then created before they are used by a program. Thus, the question becomes: why does cpio support sockets?

I needed a way to generate a list of commands in make automatically. I was using $(foreach i,$(list),cmd $i;), that is, a semicollon to intercalate commands. The problem with this approach is that if the first command fails, make would not detect the error. After searching for a while, I found the solution on this link. It's possible to create a newline in make functions using this technique. The code ended up being:

define \n
endef
target:
        $(foreach i,$(list),cmd $i$(\n))

I've been sadly neglecting this script, but tonight I finally found some time to catch up on all the suggestions and bug reports that Guido G nther has been letting me know about (and I greatly appreciate that!). This version merges in all the good ideas people have had. This version removes the -i and -I flags passed to dpkg-source when the package format is 3.0, since with 3.0 packages, dpkg-source both already ignores Git files by default and ignores various other things that were not included in the pattern. It also offers much better configuration methods for the ignore patterns, which git-pbuilder was overriding. So that's all skipped now for 3.0 packages. git-pbuilder still passes in the ignore flags for 1.0 packages, since they don't ignore Git files by default. The environment variable BUILDER can now be set to choose the builder, which is still cowbuilder by default. The other supported option is qemubuilder, and for it git-pbuilder selects an appropriate configuration file. Additional flags for the builder can be set in the GIT_PBUILDER_OPTIONS environment variable and will be passed along. There are also a few more minor fixes: better documentation, always printing out what we're doing at the start of a run, and preserving the exit status of pdebuild and exiting with the same status. These changes are based on work by Guido G nther, Yaroslav Halchenko, and Beno t Knecht. You can get the latest version from my scripts page.

Linux 2.6.39 was released earlier today. This release includes a few updates, the main user visible one being that machine drivers can now be registered as regular devices rather than using the soc-audio device.

• Support for registering machine drivers as first class devices rather than using the soc-audio device. Support for the soc-audio device will be removed at some point in the future.
• Support for ordering widget power changes within widget types, helping with large CODECs and multi-stage amplifiers.
• Support for waiting for multiple slow events to complete during DAPM sequences, making it easier to handle things like DC offset correction on multiple outputs.
• CPU support for Intel Medfield and nVidia Tegra2.
• CODEC support for Cirrus CS4271, Freescale SGTL5000, Maxim MAX8950, National Semiconductors LM4857, TI SN95031 and Wolfson WM8991
• Machine support for Intel Medfield MID reference platforms, nVidia Harmony, and Visstrim M10.

The host name On Unix systems, the host name is kept in memory by the kernel: it is set and get by the sethostname() and gethostname() functions and their command-line wrapper hostname(1). It can be used by several programs, for instance by the mail server to determine what it should use as its HELO name. In fact, the host name has two forms: the short one and the fully qualified one. For instance, a host can have muscadet as its short name muscadet and muscadet.example.com as its fully qualified one. From these two forms, a third information can be deduced: the domain name, which is the full name without its first component. The system host name can be set as either the short name or the fully qualified name. When the full name is used, the short one can be easily deduced, by keeping only its first component. When the short name is used, the full name must be determined by using a more complex heuristic that takes more time and can fail for several reasons.Determining the short and the fully qualified host name The command hostname(1) is able to give both the short and the full host name, and also the domain name that can be deduced from them:

-s, --short

Display the short host name. This is the host name cut at the first dot.

-f, --fqdn, --long

Display the FQDN (Fully Qualified Domain Name). A FQDN consists of a short host name and the DNS domain name.

-d, --domain

Display the name of the DNS domain.

Now, given that the kernel only maintains a single host name, this command and the corresponding functions are using heuristics to determine these two forms. I did not look to the code, but here is what I have deduced:

• if you have set your system host name to the fully qualified one, it deduces the short one by taking only its first component;
• if you have set your system host name to the short one, it deduces the full one by resolving it to an IP address, then resolving it back to a host name: this may use /etc/hosts, the DNS or a more exotic name system depending on your configuration.

The second option, when the kernel only knows the short host name, seems rather inferior, as it uses a more complex algorithm which takes more time (if it uses the DNS, it can use much more time than a fully local query) and can fail for several reasons. Setting the host name With Debian, the host name is stored in the file /etc/hostname and set at boot time by the hostname.sh init script. The local hosts table /etc/hosts also contains it, in both forms, for a loopback IPv4 address 127.0.1.1. It is asked at installation time: the most common practice is to give the short host name; you are then asked for the domain name. But you have the option to give the fully qualified domain name, which has the effect to skip the domain name question. Using a fully qualified system host name I used to define short system host names, and I switched to fully qualified ones about a year ago. I saw no drawback, but I did see benefits. Postfix, by default, uses the system host name to determine its variables myhostname and mydomain that are used for stuff like the HELO name and the mail domain name. Well, with a short system host name, it fails, and you get default myhostname = muscadet.localdomain and mydomain = localdomain, so you have to set them explicitly in its configuration file, thus duplicating information. Whereas with a fully qualified system host name, it just works, and you can rely on the default values, that will stay in sync if you have to change you host name. To sum up, it is perfectly possible to use a fully qualified name as the system host name, even though it does not seem to be a common practice. In fact, according to my experience, it is even advisable, as it eases the configuration of programs that rely on the host name, such as mail servers. Edit: As noticed by Javi, the hostname(1) manpage ends by saying without explaination that /etc/hostname should contain the short host name and not the full one. There may by some drawbacks by doing so but I was not able not identify them.

Particularly around local government, but also some not-for-profits and universities, I m trying to persuade people to stop using huge Blind Carbon Copy (BCC) lists that they keep in their address book and switch over to using mailing list manager software. Our co-op even gives some simple mailing list hosting away with web hosting. I think good mailing list software is better because:

1. List members can help themselves they can subscribe, unsubscribe and set options by sending email to the software or visiting its website, which reduces the worker time required to manage the list. But you can still manage it directly if you prefer.
2. Your address book is smaller because you don t need every address on the mailing list cluttering up your email software s address book any more
3. Errors are handled automatically if an address doesn t work any more, the software will unsubscribe it, usually with some safeguards against one-off problems
4. Privacy is safeguarded dedicated email list software will not put all the mailing list addresses in the To-line by mistake (as often happens with Address Book announcement lists)
5. Spam is filtered rules can be set centrally and they can be different to the rest of your organisation
6. Messages can be moderated if needed, the list managers can be asked to pre-approve the messages, or you can set rules to allow some automatically.
7. Emails can be archived most list software can save copies of messages for you, or put them on the web
8. You can offer digests offer members the choice between getting every message or daily or weekly batches
9. Multiple versions can be sent some list software can send different versions of the same email (like rich and plain text, or different languages) as chosen by the user
10. You can syndicate news some list software can also put your announcements on content management systems or social media
11. (OK, I ve done 10 there are two more which matter to a tech worker like me which I m going to mention as free extras.) It s more efficient most list software is designed to handle hundreds or thousands of recipients and sends email with more robust settings than a typical desktop email client. It is less likely to fail after recipient 373 and ask you to resend them all.
12. Comply with standards good list software either follows standards for things like self-service and digests automatically or can be told to do it. Some desktop email software (Outlook?) usually can t.

Do you think these are good reasons? Are there other reasons you would include above some of these? Would they persuade you to stop using your computer s address book?

Firstly: If you want to buy a computer to run Linux on, don't buy a Mac.
Secondly: If you have a Mac and want to run Linux on it, the easiest approach is going to be to run it under virtualisation. Virtualbox is free, and worth every bit of what you're paying.
Thirdly: If you're going to boot Linux on bare-metal Apple hardware, boot it via BIOS emulation.
Fourthly: If you're going to boot Linux on bare-metal Apple hardware via EFI, and it doesn't work, write a patch. Apple's firmware has a number of quirks that I'm aware of and we're working through them, but anyone filing bugs against Apple hardware on EFI right now is likely to be ignored for a significant period of time until there's an expectation that it'll actually work. Maybe in six months or so.

This year Debconf is a bit off the beaten track, so there is a bit of anxiety in the participants flights to Banja Luka are quite expensive (I d guess we have packed the relevant dates pretty full already) and other options are going via Zagreb or Ljubljana, but there people do not yet have all the info on how they will be able to get from, for example, Zagreb airport to the Debconf venue in Banja Luka, if their plane lands at 22:30 and get back if the departure time is 06:30. There is hope that the organising team will provide the info on how to use the local buses and arrange a Debconf bus for people arriving late and departing early, but that has not been 100% confirmed yet. In light of all this, I really dug deep into all possible options and found an option that is far more fun and relaxing for me and almost the same price. So I will be travelling to Paris on the morning of July 16th, then staying overnight there (never been in Paris before!), then take a late morning flight to Zagreb where I ll be around noon (plenty of time to get to Banja Luka) and then on the final day of the conference I can got to Zagreb at any time, stay there overnight (never been to Zagreb too!) and around noon next day fly back home via Budapest (just a 2 hour break there). This way I will get to see two more cities, save me a lot of stress and also not bother the organizers too much. P.S. I must say thank you Accenture for paying for my plane tickets.

• More info: mdjonline.com/view/full_story/13289144/article-Spaniards-protest-banks--political-parties
• Foto galleries and more info www.demotix.com/news/693670/madrid-protesters-leaving-court
• Still more: www.tmcnet.com/usubmit/2011/05/17/5517295.htm
• [Edit] BoingBoing carries a very complete article: http://www.boingboing.net/2011/05/17/nobody-expects-the-s.html

Spain is struggling to recover from the double blow of the global recession and the bursting of a decade-long property bubble, and the unemployment rate is more than 21 percent.

The fundamental problem with projects requiring copyright assignment is that there's an economic cost involved in me letting a competitor sell a closed version of my code without letting me sell a closed version of their code. If this cost is perceived as larger than the cost of maintaining my code outside the upstream tree, it's cheaper for me to fork than it is to sign over my rights. So if I have my own engineering resources, what rational benefit is there to me assigning my copyright?

So today is my first day of in depth research into the coming project with dpkg. It s going quiet well. I can now say I have a good understanding of dpkg-divert but am currently figuring out what all the functions do in the source code (oh so much code). Hopefully I ll update my blog on a daily basis or bi-daily. Later on I will post links to my Git repo for the project aswell as a detailed plan of how I m going to approach everything. Maybe some insights aswell, that is if I have any. Until then, ~Sam

Sometimes people will tell you what you should do. Sometimes they will even shout at you for simply asking a question on why they want something done because it isn't clear just from itself. And others likes to jump the boat and join in just for the fun of it... Gladly, this is MY life, and I choose how much abuse I'm willing to take, especially for a voluntary work that I didn't even enroll for but got put into. Sometimes through my dedication to getting quality into things and seeing that others simply neglect these areas, but they need to get addressed anyway, no matter how little respect is shown for people investing in these boring areas. The topic of It's MY Life is an old one and thus it is no surprise that a fair amount of songs surrounding it popped up over time. In my previous blog entry I wrote about different interpretations, some responses seem to hint that I wasn't clear enough about that I really meant different interpretations of the same lyrics, not just regular cover versions. The following set of songs is special in a different sense: It is about the same song title and thus does also cover different bands.

• Dr. Alban: Let's start with a rather old one.
• No Doubt: This is the band that Gwen Stefani got known through.
• Bon Jovi: No matter what you might think about Bon Jovi, they for sure wrote some pretty touching songs.

Like always, enjoy! And think about how you interact with others. I know that I'm sometimes crossing a line myself too, no one is perfect. What though makes the difference is the willingness to learn, and especially: To excuse. But in the end: It's MY life!

/music permanent link Comments: 3 Flattr this

This week I lost some time playing with Python's sets. After digging into Python source code, I finally discovered there is what seems to be little bug. Anyway, it has been "fixed" in Python 3, fortunately. I did not find if it was reported somewhere, but since it's fixed, it's not a big deal.

Python 2.7.1+ (default, Apr 20 2011, 10:53:33) 
[GCC 4.5.2] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> class A(object):
...     def __eq__(self, other):
...             return True
... 
>>> A() == A()
True
>>> [A()] == [A()]
True
>>> set([A()]) == set([A()])
False

This clearly did not make any sense to me. I've then tested under Python 3.2:

Python 3.2.1a0 (default, May  4 2011, 19:59:25) 
[GCC 4.6.1 20110428 (prerelease)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> class A(object):
...     def __eq__(self, other):
...             return True
... 
>>> set([A()]) == set([A()])
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: unhashable type: 'A'

At least, raising an error is saner. It actually helped me to understand what I needed to do to have my sets working correctly with Python 2:

Python 2.7.1+ (default, Apr 20 2011, 10:53:33) 
[GCC 4.5.2] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> class A(object):
...     def __eq__(self, other):
...             return True
...     def __hash__(self):
...             return 123456789
... 
>>> set([A()]) == set([A()])
True

Thanks to Ralf's work, dose3 has been just accepted in debian experimental !!! yuppiiiii \o/

  Dose3 is a framework made of several OCaml libraries for managing
  distribution packages and their dependencies.
  Though not tied to any particular distribution, dose3 constitutes a pool of
  libraries which enable analyzing packages coming from various
  distributions.
  Besides basic functionality for querying and setting package properties,
  dose3 also implements algorithms for solving more complex problems
  (monitoring package evolutions, correct and complete dependency resolution,
  repository-wide uninstallability checks). 

For the ocaml affectionados the API of the library is available here (there is still a bit of work to do here...) . The source code shipped in each release does not include a vast range of applications sitting in the experimental directory. You can have a look at these in our svn. if you want to get a copy of the development / bleeding edge / unstable version from svn, you can check it out from the svn

 svn co https://gforge.info.ucl.ac.be/svn/mancoosi/trunk/dose3

user/pass : mancoosi/mancoosi

This has been several months since I didn t use much my laptop, an Asus S5N: it was crippled by a short battery life and was hit by several bugs, most of them targeting the i855GME graphic chipset thanks to the use of GEM and KMS in the latest Xorg. I did not want to invest too much in a laptop and I have found a rebate from Lenovo for Thinkpad Edge series. Voil , I have a Thinkpad Edge 11 .

Hardware Thinkpad Edge is a entry-level notebook from Lenovo. In France, it is available for around 500 , thanks to the usual parity between dollar and euro.

Hardware configuration Lenovo do not sell directly in France. Therefore, you cannot configure your laptop as you want. There is two or three configurations available. This can be a bit frustrating if you want something specific. Here is the configuration that I bought:

• Intel Core i3 380UM at 1.33 GHz CPU.
• 11.6 LED screen, with a resolution of 1366x768.
• 4 GB RAM.
• 320 GB hard drive.
• Wireless network (Centrino Wireless-N) and bluetooth.
• Broadband modem (WWAN)

It s a small laptop but it is not as light as it should be (1.5 kg or 3.3 pounds). Apart from the power supply, you don t get any extra. My Asus was shipped with a wireless mouse, a transport bag, a sleeve, an USB key and a CD burner. On one side, you get two USB ports and on the other side, one USB port, an HDMI port, a network outlet and a speaker jack. The VGA connector is placed on the rear.

BIOS First step: checking BIOS settings. WiMAX can be disabled as well as any other component that you may not need. The keyboard can be configured so that function keys are available directly instead of having to use the Fn key. There is an option called Computrace from AbsoluteSoftware which seems targeted at find back your laptop in case it gets stolen or just let the FBI know where you are. I have disabled it.

Debian installation Debian can be installed from an USB stick. This is really easy now. You grab the ISO image you want to use (I take the one for AMD64 in testing because I will run unstable) and to copy it on the USB stick:

1	# dd if=debian-testing-amd64-netinst.iso of=/dev/sdb

The previous way of making the USB stick bootable, installing a kernel, an initrd and the ISO image is deprecated (and does not work). Moreover, using testing to install Debian allows me to use a 2.6.37 kernel and circumvent some problems that may have happen with an older kernel. During the installation, I choose a guided partitioning using encrypted LVM. I choose to have several partitions but I keep /usr in the root partition since having a separate /usr is less and less supported, especially with piece of software like udev. The partitioning tool in Debian installer do not allow to resize volumes, you need to destroy them and recreate them as well as their content. However, this is still faster than writing a patch for this. When the installation is finished, the laptop boots on a 2.6.38 kernel, asks for the passphrase to decrypt the LUKS volume and display the login prompt a few seconds later.

Software configuration Most of the hardware seems to be supported. No obscure commands to run to get a working network. Therefore, I start by setting up my usual environment.

Setting up the desktop First, I change /etc/apt/sources.list to get unstable:

deb http://ftp.fr.debian.org/debian/ unstable main
deb-src http://ftp.fr.debian.org/debian/ unstable main
deb http://ftp.fr.debian.org/debian/ experimental main contrib non-free

I also want that everything that is installed from experimental to be upgraded if an update is available in experimental. This is done by creating a file /etc/apt/preferences.d/experimental:

Package: *
Pin: release experimental
Pin-Priority: 101

Then, I install the following packages:

• zsh, my favourite shell with my hand-made maturating configuration. I should put it on GitHub one day.
• mg, is the editor that I use to quickly edit a configuration file. I don t like vi. I have used jed for a long time but its default configuration can be cumbersome and it misses some emacs keybinding.
• emacs, my work environment. I use it as a text editor and mail reader (with Gnus). I don t use it as a browser nor as a terminal. As for zsh, I have my configuration files ready to be copied anywhere.
• fvwm, my window manager since quite some time. I use fvwm-crystal on top of it. My configuration is described in another post. I also use conky.
• rxvt-unicode, a terminal. It has a very good terminal emulation and some fancy features like transparency (using compositing or not). It is light and fast and allows to search in the backbuffer. It can use a client/server system that will close all your windows when the server crashes. This almost never happens.
• unison, a synchronisation tool that allows me to keep the same files on my different computers.
• gdm3 is the only display manager that is able to talk correctly to ConsoleKit and get the rights set correctly without adding yourself to a lot of groups.

Nowadays, I use the kernel from Debian and I do not have to run my own configuration for it. The glory days where I had to recompile the kernel before plugging a new USB toy have passed. I remove rpcbind and nfs-common. I use ferm to setup a simple firewall. Here is my ferm.conf:

domain (ip ip6)  
  table filter  
chain INPUT  
    policy DROP;
    mod state state INVALID DROP;
    mod state state (ESTABLISHED RELATED) ACCEPT;
    interface lo ACCEPT;
    proto tcp dport ssh ACCEPT;
    
chain OUTPUT  
    policy ACCEPT;
    mod state state (ESTABLISHED RELATED) ACCEPT;
    
chain FORWARD  
    policy DROP;
    
     
 

unison will then sync my configuration files and my documents. After this step, my environment is ready.

Hardware configuration This is usually the most troublesome part. However, looking at the output of lspci and lsusb, everything seems pretty easy to get supported. Here is lspci output:

00:00.0 Host bridge: Intel Corporation Core Processor DRAM Controller (rev 02)
00:02.0 VGA compatible controller: Intel Corporation Core Processor Integrated Graphics Controller (rev 02)
00:16.0 Communication controller: Intel Corporation 5 Series/3400 Series Chipset HECI Controller (rev 06)
00:1a.0 USB Controller: Intel Corporation 5 Series/3400 Series Chipset USB2 Enhanced Host Controller (rev 06)
00:1b.0 Audio device: Intel Corporation 5 Series/3400 Series Chipset High Definition Audio (rev 06)
00:1c.0 PCI bridge: Intel Corporation 5 Series/3400 Series Chipset PCI Express Root Port 1 (rev 06)
00:1c.1 PCI bridge: Intel Corporation 5 Series/3400 Series Chipset PCI Express Root Port 2 (rev 06)
00:1c.5 PCI bridge: Intel Corporation 5 Series/3400 Series Chipset PCI Express Root Port 6 (rev 06)
00:1d.0 USB Controller: Intel Corporation 5 Series/3400 Series Chipset USB2 Enhanced Host Controller (rev 06)
00:1e.0 PCI bridge: Intel Corporation 82801 Mobile PCI Bridge (rev a6)
00:1f.0 ISA bridge: Intel Corporation Mobile 5 Series Chipset LPC Interface Controller (rev 06)
00:1f.2 SATA controller: Intel Corporation 5 Series/3400 Series Chipset 4 port SATA AHCI Controller (rev 06)
00:1f.3 SMBus: Intel Corporation 5 Series/3400 Series Chipset SMBus Controller (rev 06)
00:1f.6 Signal processing controller: Intel Corporation 5 Series/3400 Series Chipset Thermal Subsystem (rev 06)
03:00.0 Network controller: Intel Corporation Centrino Wireless-N 1000
09:00.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL8111/8168B PCI Express Gigabit Ethernet controller (rev 03)
ff:00.0 Host bridge: Intel Corporation Core Processor QuickPath Architecture Generic Non-core Registers (rev 02)
ff:00.1 Host bridge: Intel Corporation Core Processor QuickPath Architecture System Address Decoder (rev 02)
ff:02.0 Host bridge: Intel Corporation Core Processor QPI Link 0 (rev 02)
ff:02.1 Host bridge: Intel Corporation Core Processor QPI Physical 0 (rev 02)
ff:02.2 Host bridge: Intel Corporation Core Processor Reserved (rev 02)
ff:02.3 Host bridge: Intel Corporation Core Processor Reserved (rev 02)

And lsusb output (I disabled the broadband modem in the BIOS and therefore, it does not appear here):

Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 002 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 001 Device 002: ID 8087:0020 Intel Corp. Integrated Rate Matching Hub
Bus 002 Device 002: ID 8087:0020 Intel Corp. Integrated Rate Matching Hub
Bus 001 Device 003: ID 0bda:0158 Realtek Semiconductor Corp. USB 2.0 multicard reader
Bus 001 Device 004: ID 0a5c:217f Broadcom Corp. Bluetooth Controller
Bus 002 Device 003: ID 5986:01a7 Acer, Inc Lenovo Integrated Webcam

There is a dedicated page for the Thinkpad Edge 11 on ThinkWiki. This wiki is one of the best source of information when you get a Thinkpad. The kernel as shipped by Debian kernel team contains every module that you may need. One of them, thinkpad_acpi will handle most of the perks under the hood. I do not use a complete desktop like Gnome. I usually prefer to use command line tools and put them in .xsession for persistence. There is usually some graphical utility to do the same thing. There are however some configuration steps that are now far easier to do with a graphical tool. Since I use X all the time, I just rely on those tools when necessary.

BIOS upgrade From lshw output, I see that I already have the latest BIOS available. No need to update then. In case you need to update, yours, ThinkWiki has a documentation to upgrade. You can find memdisk image in syslinux package.

Graphic card The graphic card is integrated into the CPU. Intel usually ships graphic cards well supported by open source drivers. ThinkWiki tells us:

A very recent Linux distribution with kernel 2.6.33 and Intel Xorg driver 2.11 or newer is recommended. Note that XVideo (Xv) playback at certain frame sizes does not work properly unless you have at least version 2.12 of the Intel Xorg driver.

With Debian Wheezy, you get version 2.15. Therefore, configuring the graphic card is just a matter of installing xserver-xorg-video-intel and that s all. No /etc/X11/xorg.conf needed. HDMI and VGA outputs are detected without configuring anything. They can be enabled with xrandr. I did not try if it was possible to use both of them at the same time.

Screen 0: minimum 320 x 200, current 1366 x 768, maximum 8192 x 8192
LVDS1 connected 1366x768+0+0 (normal left inverted right x axis y axis) 256mm x 144mm
   1366x768       60.0*+   50.0  
   1360x768       59.8     60.0  
   1024x768       60.0  
   800x600        60.3     56.2  
   640x480        59.9  
VGA1 disconnected (normal left inverted right x axis y axis)
HDMI1 disconnected (normal left inverted right x axis y axis)
DP1 disconnected (normal left inverted right x axis y axis)

Compositing works without any problem. I use only transparency and therefore, I don t know what the speed is on a desktop with all effects turned on. Regarding video acceleration, XVideo is supported but the graphic card also features hardware decoding with i965-va-driver package. Here is the output of vainfo:

libva: libva version 0.32.0
libva: va_getDriverName() returns 0
libva: Trying to open /usr/lib/dri/i965_drv_video.so
libva: va_openDriver() returns 0
vainfo: VA API version: 0.32
vainfo: Driver version: i965 Driver 0.1
vainfo: Supported profile and entrypoints
  VAProfileMPEG2Simple            : VAEntrypointVLD
  VAProfileMPEG2Main              : VAEntrypointVLD
  VAProfileH264Baseline           : VAEntrypointVLD
  VAProfileH264Main               : VAEntrypointVLD
  VAProfileH264High               : VAEntrypointVLD

With Debian, VLC can make use of this if you put ffmpeg-hw=1 in your ~/.config/vlc/vlcrc. We are then greeted by this:

[0x2629040] avcodec decoder: Using VA API version 0.32 for hardware decoding.

Network The network card works out of the box. ThinkWiki tells us that there can be stability problems with 2.6.36 kernel. During the installation, a 2.6.37 was used without any problem. I am now running a 2.6.38 and never got a glitch. It is likely that the described problem was solved with 2.6.37.

UltraNav UltraNav is the combination of the usual touchpad and the trackpoint which is a small joystick in the center of the keyboard and is well known to Thinkpad users since most of them have one. The trackpoint is a great input device. Nowadays, all input devices are configured with the help of xinput command. Here is the output of xinput list:

  Virtual core pointer                      id=2    [master pointer  (3)]
      Virtual core XTEST pointer                id=4    [slave  pointer  (2)]
      SynPS/2 Synaptics TouchPad                id=11   [slave  pointer  (2)]
      TPPS/2 IBM TrackPoint                     id=12   [slave  pointer  (2)]
  Virtual core keyboard                     id=3    [master keyboard (2)]
  Virtual core XTEST keyboard               id=5    [slave  keyboard (3)]
  Power Button                              id=6    [slave  keyboard (3)]
  Video Bus                                 id=7    [slave  keyboard (3)]
  Power Button                              id=8    [slave  keyboard (3)]
  Integrated Camera                         id=9    [slave  keyboard (3)]
  AT Translated Set 2 keyboard              id=10   [slave  keyboard (3)]
  ThinkPad Extra Buttons                    id=13   [slave  keyboard (3)]

The command xinput list-props will give you additional information on each input device listed above. You can then configure any settings using xinput set-prop. For example, if you want to enable scrolling by using the edges of the touchpad and enable palm detection, you can use the following commands:

1 2	$ xinput set-int-prop "SynPS/2 Synaptics TouchPad" 'Synaptics Edge Scrolling' 8 1 1 0 $ xinput set-int-prop "SynPS/2 Synaptics TouchPad" 'Synaptics Palm Detection' 8 1

If you want to disable the touchpad and only use the trackpoint, use the following command:

1	$ xinput set-int-prop "SynPS/2 Synaptics TouchPad" 'Synaptics Off' 8 1

Each property has a type and a size. There is no place for guessing since every property is listed in man 4 synaptics. The same applies for the trackpoint but the correct manual page is man 4 evdev. For example, if you want to enable scrolling using the middle button, use:

1 2	$ xinput set-int-prop "TPPS/2 IBM TrackPoint" "Evdev Wheel Emulation" 8 1 $ xinput set-int-prop "TPPS/2 IBM TrackPoint" "Evdev Wheel Emulation Button" 8 2

All those commands can be tossed in ~/.xsession.

Keyboard hot keys Thinkpad laptops usually have a lot of hot keys. This is not the case of the Thinkpad Edge which only features a few of them:

• volume control
• brightness control
• webcam and external screen toggle
• wireless toggle
• multimedia keys

All those keys trigger an event in one of the keyboard listed in xinput list. Therefore, your window manager or any other application can intercept them and act upon. You can use evtest to check which key triggers which event on which keyboard. The acpid daemon listen on those keyboards and can also trigger some actions. For example, it will run ibm-wireless.sh when you hit the wireless toggle. By default, this will enable or disable your wireless card. But you can configure the script to use a multistate toggle and disable or enable the bluetooth card as well.

Sound card As for the network card, the sound card works out of the box. I use PulseAudio because it allows per-application volume control and per-application sound card selection. It comes handy if you have some bluetooth headset. To make PulseAudio the default application for most applications, I put this in my ~/.asoundrc:

pcm.!default  
  type pulse
 
ctl.!default  
  type pulse
 

Wireless card A firmware is needed to make this card work. On Debian, you need to install firmware-iwlwifi package. Then, iwlagn module needs to be loaded. You can check with rfkill list that the wireless card has not been disabled. iw list will then give the status:

Wiphy phy0
Band 1:
    Capabilities: 0x872
        HT20/HT40
        Static SM Power Save
        RX Greenfield
        RX HT20 SGI
        RX HT40 SGI
        No RX STBC
        Max AMSDU length: 3839 bytes
        No DSSS/CCK HT40
    Maximum RX AMPDU length 65535 bytes (exponent: 0x003)
    Minimum RX AMPDU time spacing: 4 usec (0x05)
    HT RX MCS rate indexes supported: 0-15, 32
    TX unequal modulation not supported
    HT TX Max spatial streams: 1
    HT TX MCS rate indexes supported may differ
    Frequencies:
        * 2412 MHz [1] (14.0 dBm)
        * 2417 MHz [2] (14.0 dBm)
        * 2422 MHz [3] (14.0 dBm)
        * 2427 MHz [4] (14.0 dBm)
        * 2432 MHz [5] (14.0 dBm)
        * 2437 MHz [6] (14.0 dBm)
        * 2442 MHz [7] (14.0 dBm)
        * 2447 MHz [8] (14.0 dBm)
        * 2452 MHz [9] (14.0 dBm)
        * 2457 MHz [10] (14.0 dBm)
        * 2462 MHz [11] (14.0 dBm)
        * 2467 MHz [12] (14.0 dBm) (passive scanning, no IBSS)
        * 2472 MHz [13] (14.0 dBm) (passive scanning, no IBSS)
    Bitrates (non-HT):
        * 1.0 Mbps
        * 2.0 Mbps (short preamble supported)
        * 5.5 Mbps (short preamble supported)
        * 11.0 Mbps (short preamble supported)
        * 6.0 Mbps
        * 9.0 Mbps
        * 12.0 Mbps
        * 18.0 Mbps
        * 24.0 Mbps
        * 36.0 Mbps
        * 48.0 Mbps
        * 54.0 Mbps
max # scan SSIDs: 20
Supported interface modes:
     * IBSS
     * managed
     * monitor

I don t know why, but 802.11a channels have been disabled. I don t have interest in them and therefore I did not search why this happens.

iwlagn 0000:03:00.0: Detected Intel(R) Centrino(R) Wireless-N 1000 BGN, REV=0x6C
iwlagn 0000:03:00.0: device EEPROM VER=0x15d, CALIB=0x6
iwlagn 0000:03:00.0: Device SKU: 0X9
iwlagn 0000:03:00.0: Valid Tx ant: 0X1, Valid Rx ant: 0X3
iwlagn 0000:03:00.0: Tunable channels: 13 802.11bg, 0 802.11a channels
iwlagn 0000:03:00.0: irq 44 for MSI/MSI-X
iwlagn 0000:03:00.0: loaded firmware version 128.50.3.1 build 13488

Wireless is one of the thing that make me use graphical applications like Network Manager. Command line tools are too low level to quickly connect to some random wireless network. There was a large debate in debian-devel@ldo to determine if Network Manager was ready to replace ifup and ifdown commands, even for servers. It seems that this is not something impossible and therefore, I think it has become flexible enough to ensure most of the configuration tasks I need.

Bluetooth The bluetooth chipset is hosted on the USB bus. hci_usb module will take care of it. It supports AD2P and can be used to listen to music with a compatible headset. hciconfig will tell us if the chipset is working. Again, bluetooth is easier to configure with a graphical tool. BlueZ, the bluetooth stack in Linux, uses DBus to communicate with other applications. The command line tools have been deprecated and therefore, you need to fight with some DBus binding to pair your computer to your headset. gnome-bluetooth package comes with an applet which allows you to quickly pair to any device. If you install pulseaudio-module-bluetooth package, any bluetooth headset should work out of the box.

Broadband (WWAN) To configure the broadband modem, the most difficult thing to do is to grab the appropriate firmware. ThinkWiki comes with various instructions to get it. I did not keep my Windows installation and therefore, I grabbed the Qualcomm Gobi 2000 Wireless WAN driver and I unpacked it with some Windows host with msiexec as instructed (I did have to execute the installation program first, otherwise, msiexec tells that the file is invalid). I choosed the default firmware since my laptop is unlocked. ThinkWiki page explains which firmware to choose in which situation. The package gobi-loader needs to be installed. The firmware files amss.mbn, apps.mbn and UQCN.mbn should be placed in /lib/firmware/gobi. They will be loaded by udev. To avoid to reboot, I used this command:

1	# /lib/udev/gobi_loader -2000 /dev/ttyUSB0 /lib/firmware/gobi

To configure the 3G connection, I just used Network Manager. Ensure with rfkill list that your WWAN card is not blocked. I did not test the GPS part that needs to be enabled on Windows.

Webcam Nowadays, USB webcam are UVC device and there is a driver to support this class of devices. Check that the uvcvideo module is correctly loaded and you should have something like this in your dmesg:

uvcvideo: Found UVC 1.00 device Integrated Camera (5986:01a7)

cheese is a small application that can be used to check if the webcam operates as expected.

Card reader Nothing to do. This works out of the box.

Suspend to memory and disk Install the uswsusp package which will also pull pm-utils. You will get pm-suspend and pm-hibernate commands. pm-suspend will enable suspend to memory. On resume, the system takes a whole minute to become operational. I get the following message:

r8169 0000:09:00.0: eth0: unable to apply firmware patch
PM: resume of devices complete after 60786.035 msecs
PM: Finishing wakeup.

Installing firmware-realtek package solves this issue. It should also be solved by a future kernel. I also have this message when I come back from sleep:

[drm:i915_hangcheck_ring_idle] *ERROR* Hangcheck timer elapsed... render ring idle [waiting on 9297, at 9297], missed IRQ?

The X server is then slow as hell. The message can be fixed by using i915.semaphores=1 as a kernel parameter but the X server is still slow. It seems that running powertop restores the original speed. This is a known problem. On sleep, the battery loses 2% every 6 hours. This allows you to keep your laptop on sleep for several days if needed. pm-hibernate enable suspend to disk. The suspend part works well but on resume, the kernel freezes after reading back the image into memory. There are some documentation to debug this problem but I prefer to wait for 2.6.39 which may just solves this issue.

Power management Let s see on how we could enhance battery life. First, cpufrequtils package will slow down the processor when it is not used by using the ondemand governor. The frequency will drop from 1.33 GHz to 666 MHz. Several tools are then available:

• pm-utils package comes with scripts to enable some powersaving features. Those scripts are run on ACPI events (for example, when you unplug the AC adapater). There is no way to put the laptop to sleep after two hours of inactivity for example. You need something like sleepd to handle this part. I have found nothing that could dim the screen on inactivity.
• laptop-mode-tools is also a set of script to be run on ACPI events. It is therefore somewhat superseded by pm-utils.
• gnome-power-manager is a daemon that makes use of upowerd. Despite its name, you can run it outside of the Gnome environment. It will put your laptop to sleep on inactivity and will dim the screen when needed. All other functionalities are handled to upowerd. This seems the way to go on today Linux desktops.

Behing the scene, upowerd will call at some point pm-powersave from pm-utils. This allows to run scripts when unplugging the AC adapter. There are some scripts shipped in /usr/lib/pm-utils/power.d. You can add your own in /etc/pm/power.d. Some scripts are configurable. For example, if you create /etc/pm/config.d/local and put SATA_ALPM_ENABLE=true in it, this will enable SATA ALPM mode on battery. You can check what needs to be done with powertop or go to a website like Less Watts to get some tips on what to configure to get a better battery life. Without adding any script, I get 5 hours of battery life with the laptop doing almost nothing. Lenovo tells that you should get 6 hours of battery life. I think that on moderate workloads, you should get 4 hours.

Conclusion This Lenovo Edge has a good value for money. It features a great keyboard for a laptop of its size. It could be lighter and thinner. Its battery life is pretty good. Almost everything works out of the box with GNU/Linux. I hope that 2.6.39 kernel will solve the remaining issues like the suspend on disk issue.

Here is some updates on the LLVM world into Debian.
LLVM and CLANG getting better and better after each release. After some lag, it is now time to have all these packages up-to-date into the archive. Here is an update of their status in Debian:

• LLVM
  LLVM 2.8 and 2.9 are now available into the archive.
  For version 2.8, Matthias Klose deserves most of the credit.
  Version 2.9 is available in experimental. They both build pretty well on most of the architectures besides kfreebsd. It is on my TODO list.
• Clang. A C/C++ compiler.
  clang 2.8 and 2.9 are respectively available in unstable and experimental.
  These versions cover most of C/C++ specifications and more and more big software (Qt, Boost, Scilab, etc) succeed to be built with it. Like LLVM, it is available under most of the Debian architectures (but I haven't tested the quality of the ASM generated). clang can be used just like gcc. With most of software, calling configure/make with CC="clang" or CXX="clang" should be enought.
• * Dragonegg
  Dragonegg is a gcc plugin (llvm-gcc-4.5) which allows gcc to use optimizer and code generators of LLVM.
  It is now straightforward to use:
  
llvm-gcc -o foo foo.c
I have lighten the "contact" between the gcc version used to build the extension and the version used to run it.
  It required the exact same version of gcc (even if the sources say it is probably a too strong requirement). Now, a warning like this one is displayed:
  
Potential incompatible plugin version. GCC: 4.5.3. Expected: 4.5.2
Defines 'dragonegg_disable_version_check' as env variable to remove this warning
Please note that unexpected errors might occur.
If there is a better way to tackle this issue, I will be happy to hear it.
• llvm-defaults
  Last but not least, I uploaded llvm-defaults, a work of Arthur Loiret. This package allows a similar behavior to gcc-defaults. The installation of llvm will provide the default version of llvm.

Deprecated stuffs:

• llvm
  The llvm package has been removed. Actually, it was just the version 2.6 of llvm (see bug #625729). llvm (from llvm-defaults) is now providing the favorite version of llvm.
• llvm-gcc-4.2
  Upstream stopped the development of this plugin.
  It has been removed from the archive (see bug #626007) and dragonegg is taking the relay.
• llvm-snapshot
  llvm-snapshot (see bug #626083) followed the same direction as llvm-gcc-4.2
• llvm-2.7
  llvm-2.7 is still in the archive (see bug #626081) because of some reverse dependencies...

As previously announced in Ganneff s blog and Debian News, a special funding for DebConf11 is available for Debian Developers and Debian Maintainers who have either never attended a DebConf before, or who have not been to a DebConf since DebConf7 (in Scotland). How to apply Email the DebConf Newbies Team (newbies@debconf.org) providing the following information:

• estimate of your travel costs
• where you are traveling from
• amount you are unable to fund yourself
• dates you plan to arrive and leave
• if you have been to DebConf before
• if you will also need sponsored food and accommodation
• whether you have already registered for DebConf11 (if not - do so now!)

the DebConf team

After nearly two years of answering questions and mostly didn't find time to do so, I'm a DD now. Thanks to all involved people, specifically to Ben Hutchings - my AM and Torsten Werner who advocate me.

Connectivity can be a force for autonomy and freedom, or it can be used for control and subjugation. The Nintendo 3DS is an example of the latter. Okay, yeah, I just said a cute video game console is a threat to personal freedom. Hear me out. If you haven't read the highlights of Nintendo's Terms of Service for the 3DS yet, you should do that. On one hand, it's "just" a games console. On the other hand, it's a computer, marketed to children, which monitors your location (based on proximity to wifi access points), uploads your photos and messages to Nintendo's servers (giving them a license to re-use that information in various ways), and tracks your activity. Now that's entertainment. If you wanted your Nintendo computer to stop doing those things, you could try to install some 3rd-party "homebrew" software on it. Except then, according to the Terms, Nintendo would "brick" your device render it inoperable, when the modified software is detected. Would you accept Ford disabling your car because you had a mechanic friend fix a problem using some aftermarket parts? Microsoft breaking your computer hardware because you installed Firefox? The ability to stop your electronics from spying on you is an important freedom to defend, even if you don't ever plan on purchasing a Nintendo device. Such terms, and the Digital Restrictions Management software that enforces them, are a problem in all areas of digital life. It's the same fundamental strategy as Apple controlling what you can install on the iPhone or iPad, or Sony sending police officers to a programmer's house because he showed other people how to modify their own computers to make them able to run non-Sony-approved software. If we don't set a strong precedent against this behavior, it will become more prevalent. If we lose the ability to install what we like on our computers, then we lose the ability to use those computers in any way that isn't pre-approved by a corporation. We lose it not only because we ourselves can't install some software maybe we hate changing anything on our computers at all but because the people who write the software we rely on lose that freedom, which ends up impacting even the most non-technical of computer users. Companies like to portray anyone arguing for this freedom as criminals just wanting to illegally copy software. But that's obviously not true -- a person who follows all laws, even the most absurd, might still want to install something to stop Nintendo from receiving copies of her photos or reporting on her location. Companies use the threat of illegal copying to justify massive infringements on real personal freedoms. People often say, "well, just don't buy it then". I agree don't buy it. But don't think this is merely a consumer issue. Terms like Nintendo's are backed by the force of criminal law the Digital Millennium Copyright Act in the US for example which can put you in jail for helping people to modify their own computers. So this is more than a mere preference issue, it's also one of law and ethics. It's not a free market when one side has coercive state resources in play to prop up its arbitrary business model. To help send this message, I just donated to send 5 bricks to Nintendo. I hope you'll send one or five too. The campaign has a goal of sending 200 bricks by the end of today (Monday, May 16th). Nintendo clearly wants to get into the brick-making business, so surely they'll appreciate the extra inventory. (Disclosure: While I do work for the Free Software Foundation and am paid partly to run the Defective by Design campaign, I really did donate, and support this campaign personally.)