This is just one problem with a profit logic ruling our health care. It is good to see this investigation start but it feels a little like a kitten fighting a 100 pound gorilla. A little to little and too late.

OK, I may not have wireless yet on my HP laptop but I have managed to get wireless networking running with the new b43 native kernel driver for this iBook G4 laptop.
0001:10:12.0 Network controller: Broadcom Corporation BCM4306 802.11b/g Wireless LAN Controller (rev 03) (that's the Apple internal card).
Linux fergus 2.6.24-1-powerpc #1 Mon Feb 11 12:53:42 CET 2008 ppc GNU/Linux
linux-image-2.6-powerpc 2.6.24+13
I needed:
b43-fwcutter-011.tar.bz2
broadcom-wl-4.80.53.0.tar.bz2
(Note the versions - older versions don't work, newer versions give errors from the kernel and don't work either.)
www locations and full instructions available at http://linuxwireless.org/en/users/Drivers/b43.
The best bit is that because this is a native driver, network-manager is able to find it and use it so that I don't have to waste time with detailed settings in /etc/network/interfaces etc. The worst bit is that it still needs a second internet connection to download the firmware from Broadcom and compile the firmware cutter from source.
As I've only just got it working, I've yet to test it with suspend etc. but at least I no longer have to use bcm43xx (which is deprecated by b43 anyway).
It also appears that this will work with the HP laptop, once the device in that laptop is recognised by the kernel. Just such a pain that Broadcom don't open up the licence for the firmware so that we can have a sensible package that can be installed ready and working without all the work.

Update - it continues working after suspend.

If you ever wondered, why exim4 needs so long to start when you have no net access, though you were sure that configured as satellite for a smarthost it should have nothing to lookup as the smarthost in in /etc/hosts, you might just have forgotten to put a

     disable_ipv6 = true
   

in your exim4.conf. (I'm not sure, but that might also help to actually deliver mail to hosts which also have ipv6 addresses on servers with outgoing smtp when you forgot to blacklist the ipv6 module).

I’ve released pommed v1.15 this week-end with support for the new power_supply class which is used to report ACPI power information starting with kernel 2.6.24. This release also contains pretty much everything needed for the MacBook Air1,1, except the keyboard/trackpad USB IDs which I couldn’t grab yet. So expect v1.16 for the MacBook Air as soon as I’ll have those IDs.

The basic Debian debootstrap system is 180Mb or so. Embedded systems need a full GUI installation in 64Mb or less and a basic debootstrap-equivalent system of 10Mb or less.
Such radical reductions in size mean radical changes to the structure of the the system itself. The question is, at what point does a modified Debian system become no longer Debian?
Main changes:

• Remove perl and reverse dependencies. There is simply no room for perl, let alone python or any other interpreted languages.


• Drop "Essential" - perl is classed as "Essential" for Debian so dropping perl means modifying "Essential", certainly. However, to give full flexibility for embedded systems, Emdebian is removing "Essential" completely. This means that all dependencies need to be specified in full but it also means that any workable package set can be used. (I'd like to have a configurable "Essential" that can be machine-specific so that a desktop can have a different "Essential" set to an iPAQ. If not, the iPAQ will simply have to do without "Essential" because the level of changes are not likely to be acceptable for the rest of Debian - and frankly, I'd be surprised if it was to be acceptable.)


• Modifying dependencies - cross building allows each package to be modified during the build, changing the options to ./configure to disable components that Debian enables. This can cause a few problems trying to mix-and-match Emdebian and Debian but customised Emdebian packages could help with that.

Main similarities:

• All packages are exclusively from Debian


• All packages buildable from source using packages in Debian Lenny.


• The majority of the Debian package set are available, within the constraints of the embedded device.
  

So the question is, lazyweb, how far do we go before you would regard Emdebian as no longer Debian?
I now have image tarballs for such a system - 18mb compressed with a full GPE GUI included, 48Mb installed. The scripts could, potentially, work with any debootstrap tarball.
Although based on debootstrap, the Emdebian set can be unpacked separately from debootstrap so that as much of the workload can be done on the build machine as possible. This allows the "installer" to simply copy the decompressed filesystem onto the device, run /sbin/ldconfig and dpkg --configure -a and it's ready to go.
I have a few issues to resolve within the packages themselves and within the available kernel for my test machine but the result is a full GUI installation that will fit on a 64Mb SD card (32Mb is a tad too small) with enough room to spare to add a Linux kernel zImage, modules.tgz, Windows executable/autorun runes and a bootloader. i.e. a possible method of installing Emdebian onto a handheld running Windows CE or Windows Mobile. The only issue is getting the right kernel and modules for the device.
These are important benefits, the cost is whether such an installation still qualifies as "Debian".
At what point does Debian cease to be Debian?

With the latest flash for Linux update, two of the biggest annoyances have disappeared. Yay for usability!

• Flash no longer steals the keyboard focus - Ctrl+L to go to the location bar now works even when the mouse is over the flash embed
• HTML overlays over flash finally work. This makes some web sites finally usable even with Flash on (I still recommend using FlashBlock!) - they had a navigation menu that would open up under some Flash animation they had displaying below the menu, effectively making these parts of the menu inaccessible.

It still sucks, and I'm only using it in rare cases by selectively loading Flash. I love FlashBlock, which replaces any flash crap with a 'Play' button. Now if we could only do the same for some of the Web 2.0/DHTML annoyances...[Note: it's not a particularly new version. r115 according to about:plugins. I was told it's like 3 months old. I just noticed the difference, maybe cause the Debian flash-installer package was just updated to this release, or because I use Flash so little. It was so unreliable for like ever that I wouldn't even start using any site that uses Flash for anything serious. Granted: I do use youtube, because there are some great Swing dancing clips there. But that only uses a tiny fragment of Flash but still has been killing my browser every now and then.][Update: I was misled by a clever Javascript that actually disabled the flash or shrank it or so when the menu popped up. E.g. the Toyota site http://www.toyota.com/corolla/ is still unusable unless you're using some flash blocker or do not have flash installed (apart from the flash assuming it were transparent, but it's rendered on all white). The menu still pops up under the flash, making it impossible to access many pages.]

A few days ago, I asked how to fork a debian package privately. I got some repsonses by comments, e-Mail and other blog posts, and I want to summarize the tools that were suggested. I have not really tried them, just looked at the webpages, the documentation or the sourcecode. I m only including Debian-specific solutions, as using Gentoo might be nice, but is out of scope here :-)I also deleted half the text before saving it, and I hate rewriting stuff, so it might be a bit terse.debtooEddy Petri or tells me about his three year old project debtoo, which wants to bring a Gentoo-like experience to Debian. While it probably provides what I need, it feels too big in that it has a more general aim than what I m looking for. It is not in Debian and not actively developed in the last two years.srcinstEddy also tells me about srcinst, which is a small haskell program written by jgoerzen in 2005. This is more a proof-of-concept that builds a package and all necessariy build-dependencies without using a binary package, but does not allow for patches to be applied. It is in Debian, actively maintained, but not developed, and barely documented (just an IRC log).apt-buildA similar tool and also in Debian is apt-build. This package is created to compile packages for optimization (as it asks half a dozend compiler-flag-questions with DebConf during the installation). It allows the user to apply a patch to the source, but it has to be passed as a command line option, and patches seem not to be tracked. Not much development in the last year, and not much documentation.debpatchNot to be confused with the program in the debdelta package is the debpatch tool by Victor Hsieh which Carsten Hey told me about. Also written in 2005, this comes closest to what I want. For every package you want to modify, you can configure it to apply patches to the source, to add patches to the debian/patches directory or to statically replace files in the generated binary and puts it in a local apt repository. No code yet to track new versions of the packages and no support for pbuilder though. This perl script is not yet in Debian, but is available as a Debian package and seems to have seen some development in the last years. Quick Googleing does not return any user reports; has any already tried this?customdebSomewhat in this direction goes the script that I wrote two weeks ago for a totally different project, customdeb, but it is aimed at binary debian files exclusively, and debpatch already supports replacing files in the binary package.apt-srcAlso a related tool Carsten Hey told me about is apt-src. This tool wants to provide apt-like features for source packages. It allows you to install sources somewhere, upgrade the source and build binary packages from them. It does not really hide the sources from the user and seems to be created for people who want to work with the source somehow. If you have local changes to the sources, it can try to merge them with the new version, but these changes are not kept as a separate patch but only as implicit changes. If that were different, it would seem be a very nice tool. It s written by Joey Hess he abondonned it in 2004. Since then, not much development has happened.debnestYet another approach is taken by debnest, which Peter Makholm took over. Here, you create a full debian source package with a debian/ directory containing a control and rule file etc., but besides the debian/ directory, it contains just the packed sources (.dsc et al) from another package. In your debian/rules, you can specify your changes. I can see no support for tracking new versions of the inner package, but at least this project is maintained and in Debian.Summary I found a few projects that go in a similar direction, and one that seems to fit my needs quite well. All of them are not widely used and in a very early state of development, and it seems to be something people start to work on and quickly lose interest. I m actually surprised that this is not a itch more people had to scratch! I will have a closer look at debpatch when I find the time and maybe, if found useful, see if it can be included in Debian probably after a name change, since we already have a debpatch binary. Or maybe I should try to add support for what I need to apt-src, which seems to be the most advanced of all these programs, although it s unmaintained longer than the others exist.Update: Added apt-src.Update 2: Added debnest;

after my laptop I tried to get 2.6.24 running on my desktop machine; several attempts (Debian kernels, self-compiled ones with different options) all led to the same results: a machine that feels rather unresponsive & where music has sporadic dropouts.

some searching on the intarweb showed that the new CFS scheduler (cf. Documentation/sched-design-CFS.txt) (at least the enabled-by-default CONFIG_FAIR_USER_SCHED flavour) might have latency problems, in itself or in combination with the ondemand CPU frequency scaling governor or with nice'd processes; after turning off all CONFIG_FAIR_* options in .config, recompiling & rebooting the machine seems to behave normally.

maybe I'm missing some great new features but actually I prefer a working desktop computer :)

update 1: approximately half an hour after the original blog entry I had my first "ion sorbet", i.e. a half-frozen ion3 (the applications are still running, I can switch to a tty but I cannot change/close/whatever ion3's frames). at the moment I have no idea what's going (wr)on(g) here ...

update 2: two more "ion sorbets" today yet; let's see if a new kernel (CONFIG_FAIR_USER_SCHED seems to be fixed in 2.6.24.1) & a new ion3 version help.

update 3: still the same problems (& inbetween some glitches with switching from X to a tty and back; I hope I solved them with some kernel config options); ATM I suspect ion3 to be the culprit; when it hangs it eats almost all the CPU, & strace'ing the pid gives only SIGALRM messages.

update 4: maybe I fixed the problem; by changing from ion3's deprecated statusbar_external.lua to the newer statusd_exec.lua (I just wanted to get rid of the error when starting ion3 ...).

With the necessary bugs now filed, there is a real chance that Lenny can be a usable cross-building platform.
The next task is to upload emdebian-tools 0.8.1 (which includes the scripts I used to prepare the patches for those bugs) and start on improvements to emsandbox.
I plan to follow the method used by balloonboard.org to improve debootstrap so that the tarball created by emsandbox contains a real filesystem, without .deb files, with all packages unpacked and then simply copy the files across and run dpkg --configure -a. (in theory).
So far, only done this in a "manual and hard-to-do-again" way and I haven't been able to get X running yet (kernel issues). I'm looking to write the scripts to create the new installation tarball first - I'll fix the packages themselves later. Emdebian won't be making these tarballs available for download (each is too specialized) so the scripts will create each tarball when needed.
That's the plan anyway . . .

(Oh, BTW, most of the patches for those bugs were created on this HP laptop which I so recently criticized. The wireless still isn't working and the power adaptor gets damned hot but apart from that, I'm happy-ish with it.)

One of my favorite radio stations, KEXP, based in Seattle is coming to NYC by parenting with a local station. Not much will change except that we will have the incredible John in the Morning for some of his time and he will, according to their website, scout out local talent:

Does this mean that John in the Morning is leaving Seattle?
No. Starting in June 2008, John will split time between Seattle and New York, broadcasting his show live from both locations throughout the year. While he s in New York, he ll be searching for new bands and discovering new music that The Morning Show listeners will be the first to hear.

For simpler tracing of python code than my snippet...: python -mtrace -t program.py

The problem with the Internet is that it never goes to sleep. And if you study the Internet as I do, that means spending a lot of time starting at the screen tagging, copying, pasting, downloading, watching, and assessing. I have taken a step or two back away as I have moved from life as a postdoc to working full time but ever since Anonymous declared war on Scientology last month, I have been spending more time online again (more than I want to at least), tracking and following some of their developments, including spending time on some of their IRC channels, downloading lots and lots of videos as well as all sorts of great flyers. Well today there was an IRL action loosely coordinated by Anonymous to protest the CoS and I have to say, I am simultaneously surprised and unsurprised at what I think was a pretty vigorous response and turnout. I was not too surprised as the attack against and criticism of Scientology is one of the older wars spurred on in part by the net and net ethics and it always seems to crop of from time to time. But I think the depth of smack talking I witnessed on IRC, (often clearly youthful in nature as this example reveals—>, [mizzanon] I will have my full license in approx. 3-4 weeks, as soon as I do, i will be driving around my entire region with tons of
anti-propaganda) led me to believe that they could not pull of this global day of action. But they did and perhaps with the exception of New Zealand (where there are probably better things to do anyway), they seemed to garner a fair bit of participants. It will be interesting to see how long they will continue to generate this type of support. The next day of action will be around the middle of March.

Today I've been mostly unwell. Although I have managed to write some minor new code, and watch a little bit of Doctor Who on DVD. Recently several people have been ranting about Ruby on Rails. I like it, but I wouldn't use it for personal development in a hurry. Deployment is fiddly, and upgrades are annoying. But one thing that I utterly condemn Rails for is helping to spread bad paging throughout the online world. So, what is "bad paging" and why is it important? Well cool URLs don't change, right? "Bad paging" is any user-interface which presents you with a limited view upon a changing list of items which is non-bookmarkable. Consider the following "list". Assume it represents your view of a collection of items numbering 100+. You may only view ten items at a time; clicking "next", or "previous", to navigate your viewport:

1.  first item
2.  second item
..
10. tenth item
[see next: /start/1] [see prev]

Whats wrong with this picture? It is subtle, but this list is broken. The issue is that when the list grows new items are prepended to the front, yet the navigation is linked to the starting page number. If that description wasn't clear consider what happens if you want to bookmark the page containing item 11. How can you? Right now it is at /start/1. If a ten new items are appended to the head of the list then it will instead become /start/2 - as items that are currently numbered 1-10 will be shifted forward to become items 11-20, and and they will be on page /start/1 instead. The solution is simple enough once you consider what you want to happen:

• Either append items to the end of the list.
  • Such that /start/1 always gives the items 11-20.
• Number the links in the reverse order.

So why does nobody do that? (As a counter example look at my website: Rather than the 'Show previous' items linking you to the changing link /start/1, it instead links you to /start/569 (for example).

mmap()/vmsplice() root exploit and fun with Xen: Today has been a very busy day. First of all, a new exploit was discovered today that uses vmsplice to gain root access. Of course, there’s an unpleasant side effect to that exploit (so don’t try it on your machines!), it messes up the memory map and applications start crashing until you reboot your machine. When combined with Xen, things get worse: the hypervisor crashes and obviously takes out all of your domUs which is unpleasant. As a result, I upgraded my Xen-using stuff to 3.2, but the bug is still there. There’s a nice “feature” though: When the hypervisor crashes, it reboots so you can’t even get a crash log. Fantastic, except not. Xen 3.2 seems nice so far, I backported it to etch for use on my dom0’s. I haven’t noticed any major problems yet so that’s good. It’d be nice to have linux-image-2.6.23-xen backported to etch, that’ll probably be my next project (at the same time, I plan to patch out the vmsplice syscall so it can’t be exploited by users on my machines). Hopefully the backports.org maintainer will do a backport soon. dsyslog in NEW: dsyslog has been released (I did this yesterday actually), and is now in NEW as of today. For those of you who don’t know what dsyslog is, it’s a syslogging daemon I am writing from scratch for my needs. It has a 100% modular design; the core only provides a generic message processing and routing framework. Thanks to naoliv and dmb for collaborating on the packaging. upse 0.6.0 released: Today I also released upse 0.6.0, which adds the ability to use a PSone ROM instead of emulation so that you can listen to mods that require a replayer which depends random aspects of the ROM itself. Right now, this feature only works with upse123, but I will add support to the audacious plugin in 0.7.0. Please note that upse does not include a PSone ROM image as it’s pretty much guaranteed to be non-free. You can use a gameshark, or some sort of JTAG method to dump the ROM’s contents. (The ‘crash2′ replayer is especially a problem with this one.) Additionally, the new release features reworked modeling of the PSone’s SPU analog filters. srctool: A lot of people commented on my ideas for replacing srcinst with something more modern and powerful (presently called srctool). I’ll be sure to take a look at them over the next few days and comment back. freenode adopts DroneBL as part of their anti-trolling defense: LoRez notified me on freenode today that they are using DroneBL to help keep trolls and proxy abusers off the network. I wish them luck with their use of this, and recommend an additional blacklist as a supplement, dnsbl.anoxs.net. While I disagree with the way they have implemented their DNSBL, they do indeed get results.

A Slashdot question today about putting to use all the unused disk space on corporate desktops got me to thinking. Now, before I start, comments there raised valid points about performance, reliability, etc.

But let's say that we have a "cloud filesystem". This filesystem would, at its core, have one configurable parameter: how many copies of each block of data must exist in the cloud. Now, we add servers with disk space to the cloud. As we add servers, the amount of available space on the cloud increases, subject to having enough space for replication according to our parameters.

Then, say we say we want a minimum of 3 copies of each block replicated. Each write to the filesystem will then cause a write to at least 3 different servers. Now, what if one server goes down? If the cloud filesystem is short on space, we may be down to only 2 copies of some blocks until that server comes back up. Otherwise, space permitting, it can rebuild that third copy on other servers.

Now, has this been done before? As far as I can tell, no. Wouldn't it be sweet?

But there are some projects that are close. Most notably, GlusterFS. GlusterFS does all of the above, except the automated bits. You can have this 3 copy redundancy, but you have to manually tell it where each copy goes, manually reconfigure if a server goes offline, etc. Other options such as NBD, OpenAFS, GFS, DRBD, Lustre, GFS, etc. aren't really well-suited for this scenario for various reasons.

So, what does everyone think? Can this work? Has it been done outside of Google?

Today while reading Planet Webtuesday, I stumbled upon a nice cartoon about the one and only measurement of code quality: WTF per minute. Somehow I noticed that the blog in which this cartoon was posted in is powered by Blojsom, a Blosxom derivative written in Java (and nowadays database powered). I already have heard of a lot of blogging software which works similar to Blosxom and often is also named similar, e.g. Pyblosxom or Blosxonomy, but Blosjom hasn’t been noticed by yet although it is mentioned in Children of Blosxom where I first noticed Blosxonomy. So far, so good, but what really surprised me is that a blog engine developed after Blosxom’s ideas officially made into MacOS X 10.4 Server. (BTW at a time, I neither had a blog nor knew about Blosxom. :-)

installing Debian on Zonbu mini desktop ... ... it's trivial! Me and godog have bought a Zonbu Mini Desktop to replace our previous noisy caseful server. The supposed duties of such a server are just running some P2P clients and a slimserver for a Squeezebox (so yes, the Zonbu will also need to proxy a very big USB disk). I've discovered the Zonbu mini thanks to a blog post by caesar, and I'm quite happy about the hardware. It's absolutely noiseless, small, eye catching. Though I don't need particular performances, the only aspect that worried my was the I/O on the CF card, which apparently is a non-concern. The only missing bit in the hardware is a video output (only VGA is provided), a must if you want to use the Zonbu as a multimedia station. But this kind of usage wasn't on my plans, so ... FWIW, before buying the Zonbu I've also tried to evaluate the cost of building something similar by myself, but the only cost of the mother board almost dominates the final price of the Zonbu. So apparently it's a good device also price-wise. I haven't even tried the on-board Gentoo-derived installation. Well, in fact I had, but considering that the only monitor I've at home is a B/W 15 years old CRT, my chances to see something vanished at the precise moment the frame buffer started, of course they did not get any better when X started. The minute after that, I've fired up D-I.

---

Debian Installer vs Zonbu mini Installing Debian on the Zonbu mini is actually trivial:

• go for the USB memory stick installation method. The only USB port supporting boot from memory stick is the front one, apparently I was lucky since it is the first one I've tried
• according to another post by caesar at kernel greater than 2.6.22 is needed, so go for the unstable [daily built D-I] (architecture is i386) and install at least Lenny on the target system
• go take a beer, wait a bit, ... your Debian-powered Zonbu Mini Desktop will then be ready

/me happy about the new geeky gadget

---

Sad note: according to godog's tests, the provided wifi (which is optional to buy) works either with ndiswrapper or with a binary kernel module built against Fedora kernel, that sucks.

When I am at a conference I am unlikely to start a conversation with a woman who I don t know because it really offends me when I receive the I ve got a boyfriend type response

Doing. It. Wrong.