I've decided to take a look at emacs-muse (muse-el Debian package).I share many of JoeyH's sentiments with regards to what wiki should have.But maybe that's only an evidence that I'm an old-type.One thing I liked about emacs-wiki (which is what emacs-muse is a fork of)is that it allowed editing of wiki files from emacs, or whatever plain text editor,and allowed viewing of HTML after processing. They are just plain text, and can be controlled with version controlled systems.There was this nifty thing called elserv-wiki, which made emacs a webserver, and allowed wiki editing, but apparently, it's not available for museyet, and thus I can't make it a fully-fledged wiki engine.I'm slowly exploring the space.Apparently, emacs-muse doesn't really work right now for Japanese PDF generation, somebody needs to hack on it. HTML output works, so no problem right now.RSS output seems to be unavailable, along with handling of ChangeLog which used to exist in emacs-wiki.Anyway, I converted tokyodebian project webpage to emacs-muse, rather than hand-crafted HTML, and I'm happy.

Dear Lazyweb, I was trying to boot Debian on a Sun Sparc Ultra 10 box yesterday, without success. I got myself this Sparc box from eBay many months ago, but until yesterday it was merely standing around here and taking up valuable space (just like the other dozen or so computers standing around in my room). So I thought I should really make some use of it, finally... I booted it up and noticed that it came with SunOS 5.8 pre-installed. It boots quite fine into the graphical login prompt, but I don't have the root password (or the password for any other account). So my first task was to reset the root password. But I failed so far. Out of curiosity I connected the machine to my laptop and ran nmap on it. It has several ports open (e.g. ntp, ssh, X11, nfs, ...), so I could probably try to find a remote root exploit (or run an ssh brute-force attack) and crack it open that way ;) While I might try that later just for the fun of it, I wanted to simply boot a Debian first and change the root password manually. That sounded simpler at first... Anyways, here's what I tried so far: Boot into single-user mode As I had no experience whatsoever with Sparcs, it took me some time to read up all the stuff I needed... After a bit of googling, I found out that I need to press STOP+A to get into the boot prompt (the box has OpenBoot 3.19). There, I can type boot -s to boot into single user mode. Unfortunately, you still need the root password to boot into single-user mode. Argh! Boot from CD-ROM Next, I found out how to boot from CD-ROM. Typing boot cdrom in the boot prompt should do the trick. So I downloaded and burned a the current Debian "businesscard" image for Sparc:

wget http://cdimage.debian.org/debian-cd/3.1_r1/sparc/iso-cd/debian-31r1a-sparc-businesscard.iso
cdrecord -eject -v speed=2 dev=/dev/hdc -data -pad debian-31r1a-sparc-businesscard.iso
Booting seems to work partially. SILO seems to start fine, and I get to the "boot:" prompt. When I press Enter I get to the point where it says "Loading Linux..." but then the screen turns black and nothing happens anymore (hangup). The same happens when I type "linux" or "rescue" on the "boot:" prompt. The monitor works fine, SunOS 5.8 can display graphics etc., so it's probably not a hardware problem. Netboot After that, I tried to perform a network boot. This was fun and I learned a lot about RARP and TFTP, as I have never played with this stuff before. While the netboot part itself worked, it didn't solve my real problem, unfortunately... First, I installed rarpd and tftpd on my laptop:

apt-get install rarpd tftpd
In /etc/ethers I added a line which tells rarpd which MAC address should be assigned to which IP address. I looked up the MAC address of the Sparc box (it's displayed when you boot) and then added "xx:xx:xx:xx:xx:xx 192.168.0.20" to /etc/ethers. I configured my laptop to have the IP address 192.168.0.19: ifconfig eth0 192.168.0.19 broadcast 192.168.0.255 netmask 255.255.255.0. Of course, I plugged an ethernet cable in the Sparc box and in my laptop so they can talk to each other. Then, I created /etc/xinetd.d/foobar which contained:

service tftp

disable = no
socket_type = dgram
protocol = udp
user = nobody
wait = no
server = /usr/sbin/in.tftpd
# server_args = /tftp

This should tell tftpd to serve the Debian netinst boot image from /tftp, but that didn't work for me. So I commented out the "server_args = /tftp" and let tftpd fall back to the default location /tftpboot. You need to reload xinetd after any change: /etc/init.d/xinetd reload. Also, tftpd is compiled with TCP wrapper support, so you have to temporarily allow the respective connections, e.g. by adding "ALL: 192.168.0.20" to /etc/hosts.allow. Don't forget to remove that line again later on! Now I downloaded the Debian netinst boot image for Sparc (hopefully the correct one):

wget http://http.us.debian.org/debian/dists/stable/main/installer-sparc/current/images/sparc64/netboot/boot.img
mkdir /tftpboot
mv boot.img /tftpboot/XXXXXXX
You need to place that image into /tftpboot using the name under which the Sparc box requests it. I found that name by net-booting the Sparc machine and watching the network traffic with Ethereal on my laptop... The final steps: boot the Sparc box, type STOP+A, type boot net. The Sparc machine gets the boot image from my laptop just fine (it seems), and I get to the "Loading Linux..." part again. But then I have the same problem as with the CD-ROM boot: the screen turns black, nothing happens anymore... Help?

Blog Power? Over the past few months I’ve written entries here which have elicited some interesting replies. I ranted about poor support from SuSE and received personal emails from company representatives who wanted to help me out. Last week I wrote about LinuxGazette disappearing (turned out to be DNS errors on my side) and received a comment from Heather suggesting I was mistaken. Interesting that so many people see what I write, and also slightly intimidating given that my sense of humour has repeatedly been misintepretted in the past .. online and offline. I don’t really have a good conclusion here, but I suspect I just got lucky. I tend to write fairly trivial things, and rarely take any of it very serious. ObDebian: I’m in the middle of turning a spare bedroom into a real office. I now have three low-end machines in a line across two desks, and a decentish desktop machine too. It looks likely that I’ll have a lot more time to contribute to Debian over the coming months (although this isn’t confirmed yet). Either way I’ll have a much nicer “working” environment at home. ObSocial: The postcards I sent to Debian folks seemed to all arrive, except for the one I mailed to Amaya (despite having several rounds of back and forth to get the address right). That was returned to me yesterday. Try again tomorrow I guess. Cross-Country postal-services suck. Although I know they handle a lot of letters/parcels I seem to always get the unlucky straw. ObDebian-Admin: The Debian admin site now allows users to submit tags. Took a while to get right, but I’m happy with the Web 2.0 ajax code. How very ..

From random IRC stuff: “Class of 2010″ means that one would normally graduate in 2010. I.e., since it’s at least assumed that a normal university bachelor’s will take 4 year to complete, the Class of 2010 would be the class entering in fall 2006. It seems some people were confused.

I'm miraculously on my way back from London, after spending a week with daf, mvo and two guests, abel and thep, from Hong Kong and Thailand, working on improving some l10n aspects of the upcoming Ubuntu release, dapper. The most visible result of the stuff that we've been doing is that if all goes well, users of the Chinese, Japanese, Thai, Bengali, Lao or Korean languages will get their required input methods, fonts and dictionaries Just Working as soon as their drives spit their dapper install CDs. We've learned quite a bit about the state of many of the Asian languages in the Free Software world today. Daf and I can certainly tell when an app renders Bengali or Thai correctly and when it doesn't, after closely comparing how different browsers do the job on various websites and font test pages. We've learned more than we probably like about fontconfig, and thanks to mvo, there is now a mechanism to get custom fonts.conf snippets depending on the default language of the system set by language-selector, so users get the right fonts installed and used by default. Thep's and Abel's help to get this straight is invaluable. On Friday, daf and I went to the ExCel Novotel where the Launchpad team is having a sprint, in order to lend a hand with the dapper translation import that is going on right now, but for a series of reasons we ended up working on l10n sprint stuff. At least I got to see the team again, and specially kiko, who greeted me with a "Qu pasa, hijo de puta!" The real life side of this sprint has been less exciting than on other occasions, as due to being few in number and being quite tired, we haven't gone out of Earl's Court area until today. Still, our exploring around the hotel has yield some nice discoveries. On Sunday, mvo and I were looking for a place to have dinner, and stumbled uppon the Troubadour a place which can either be a pub, a restaurant, a social club or a theatre, depending on your mood. The food was excellent, and the staff was really friendly. I found a Dictionary of Slang in one of the book shelves, and learned a few funny phrases. Other days we've been to good Italian, Indian or Chinese restaurants, but we couldn't resist to go back to Trobadour again. On our last night, Mark took us to a place in the South Kesington area, with again was very good. Abel left just after dinner as his flight was at 5AM from Stansted, and we hope he made it OK to the Netherlands, as at 00:30 or so he phoned us and seemed to be pretty lost somewhere. mvo left early in the morning, so daf, thep and I went to visit some bits of the city. My plane was at 16:45GMT, and we calculated that I should be leaving the K&K at 15:00 to be well on track for my flight. We went to Monument to see the tower, went up and enjoyed both the great views over London and the Thames river, as well as the really cold wind which made me swear once again about the %*@!# lack of Spring in the UK. It was getting quite late for me, so we went to China Town for lunch. Despite the restaurant people being really quick, I only managed to leave the place at 14:50, but given my travel plan was pretty conservative, I was supposed to be safe still. Daf left me at the tube station entrance, and there I started my way too stressful journey back to Val ncia. If on our way to London Carlos and were quite close to find our flight closed, this time, I think I hit the limits. I really don't know how daf and I arrived to the conclusion that leaving Earl's Court at 15:20 was acceptable to get in Heathrow on time. I ran up and down the tube stations, grabbed by luggage, and rushed back to the station, but I had just missed a train to Heathrow and that made me lose some precious 10 minutes. As stations passed by, it became clearer that we had fucked up somewhere when deciding at what time I needed to leave. At 15:35 I was still in the Hammersmith station, with half of the Picadilly line ahead before arriving to the aiport. I sent a message to daf, so he started to get prepared about me staying to sleep somewhere in London. As soon as the train's doors opened, a number of people rushed out. Someone ahead of me fell when climbing the automatic staircase and I somehow managed to jump over him, with luggage and all. After a run that seemed like half an hour, I arrived at the Iberia desk and, panting, I could only articulate "Tarde para Val ncia?". The guy at the counter looked at me, and with a "this can't be true" expression in his face, he picked up a phone and asked "Is it already closed? [...] Can you open it for one more piece of luggage? Thanks." In the meanwhile, I had been busy removing a 500 ml yoghourt from my bag, fearing having all my clothes smelling like strawberry back in Val ncia. The guy handed me a boarding pass, and gave me instructions about the "Iberia Room" near my gate. I asked him what that was about, and he says "oh, of course, you're in Business Class". For a second, I was reminded of stargirl's tricks to get free upgrades when flying, and immediately I was reminded that I was still in a hurry, and I should rush to my gate. And so I did, forgetting to thank the guy for being so kind. I left my yoghourt there though, so at least he could get some food if he was hungry. The long line before the security check was exasperating, and when I finally went past it, I had to run all over terminal two to my gate, where everyone appeared to be waiting just for me. The doors closed as soon as I entered the aircraft, and I sat in the first row, got a free newspaper and a free meal. I couldn't believe my luck. As I write this, we've gone over the Pyrenees, which are not as snowy as a month ago, but they still look beautiful from up here. I'm glad to be back in Val ncia.

Today, at 12 noon EST, I was sitting at my computer staring at http://decisions.mit.edu/, adrenaline rushing, pondering whether to click the magic button. At 12:01, I closed my eyes and gave in.

Dear Christine, On behalf of the Admissions Committee, it is my pleasure to offer you admission to the MIT Class of 2010. You were identified as one of the most talented and promising students in one of MIT’s most competitive applicant pools ever. Your commitment to personal excellence makes you stand out as someone who will thrive within our academic environment as well as contribute to our diverse community. At MIT, you join kindred spirits: scholars, builders, entrepreneurs, and humanitarians. We believe that you and MIT are very well matched for each other.

Tell me if that’s not the biggest adrenaline rush you can get from clicking a button. MIT, you will never be forgiven for that experience. Never. I spent the next half hour running around trying to find people to talk to. I could not type. I could not talk properly. I called people on the phone, I IM’d people, I tracked my Dad down and handed him a printout of the letter. I hugged him. He tracked my mother down at the door when she got home an hour later. Now that I’ve gotten over the initial rush, I’m actually sitting here laughing. Why? Because I was completely myself in my application. I think that’s all you can really do for these things. I talked about Shadows of Isildur in part of my admissions essay. I put Debian GNU/Linux work, roleplaying, and all my music stuff under my activities. I talked about my amazing real life friends when they asked what you do to unwind. I talked about the family building a giant kite from house wrap at my grandparents farm in the optional “tell us about something you have created” section. I can recall having a lot of fun writing up my responses. As for the rest, I have excellent grades, though only decent, far from perfect SATs (no 800s for me). My SAT IIs were better than my SAT I. My day has been made a thousand times over. There are so many people I need to track down and talk to, to say “thank you!” to. My alumnus interviewer. The two teachers who wrote my recommendations, and my guidance counselor. Various other people, some of whom I’ve already spoken to. I say again: thank you. Ironically, I was supposed to be in New York City this weekend, but the friend I was going with is very sick and the trip had to be cancelled. The very friend who has always told me “of course you’ll get in!” I love my friends. From all that I’ve seen so far, including posts like this and this which have some nice thoughts on the admissions process, I’m fairly sure that MIT is my #1 school choice. Apparently this year only 13% of all applicants were accepted, wow. I’m excited about hopefully attending their Campus Preview Weekend from April 6-9th to see what things are really like. Excitement.

I’m not much of a coder, but as everybody, I have several small scripts to do miscellaneous tasks. Over time, I’ve postponed repeteadly to put them under Subversion, which was my VCS of choice, since I didn’t like the “one repo for all” approach, and creating a repo for each of them felt like too much work. A few weeks ago I tried bzr, and have managed to put several of these scripts under its control since; one repo for each script makes sense to me when they’re very easily created, and data and metadata live together. (I realize other distributed VCS have this as well, but bzr was the one I tried first, knowing it’d be the one that would have more chances of me liking it.) I’ve made publicly available those scripts which I think somebody could find useful, here (mirror in gluck). The one I’m more interested in publicizing is grep-archive, a wrapper over grep-dctrl formerly known as grepd. Since it works over a mirror-like directory tree, I’ll try to write a script that creates one and fills it with symlinks pointing to Packages files under /var/lib/apt/lists, so that it’s easier to start using it. Another one worth mentioning is good old bug-reply-to.vim, which I remembered to put in bzr yesterday after kernel-team member Maximilian Attems expressed on IRC his frustration by bug submitters not CCing the bug in their replies to him. I’ve also taken chance to improve it a bit, most notably adding support for adding whatever is in the From field to the Reply-To (formerly the address would have to be explicitly passed as an argument, which was suboptimal if one uses more than one address), and documenting it a bit. NP: Fangoria, Interior de una nave espacial abandonada

It seems that ‘Linux Gazette’ has disappeared for good now? Both the .net/.com sites now point elsewhere. I’ve come up with a solution to the comment spam problem I referred to previously; I’ll allow comments on new posts for the first couple of weeks then disable them automagically. That seems like less effort than installing more plugins, filters, and complicated things that people suggested. (Thanks for all those who suggested things incidently. I think I replied to everybody individually already). No other news I can share right now. PS: Walking to the sandwich shop for lunch today I got to jump in puddles! This makes me happy.

This is mainly for Google, as I found a lot of things talking about the error with not a lot of solutions when I looked for it. I had a machine that was outputing the Neighbour table overflow message with reasonable frequency, and while it didn't actually seem to be causing problems it was disconcerting. Digging around seemed to suggest it was an ARP issue, with some pages mentioning Blaster - presumably because the scanning that this sort of worm can do is likely to fill up the ARP table? Anyway, this machine has over a /22 of IPv4 space hanging off its various interfaces, so it seemed logical that I might need a larger cache than the default. I found this SuSE post which solves the problem perfectly. So I've added the appropriate magic to /etc/sysctl.conf:

net/ipv4/neigh/default/gc_thresh1=1024
net/ipv4/neigh/default/gc_thresh2=2048
net/ipv4/neigh/default/gc_thresh3=4096

(Oh, and the message comes from net/ipv4/route.c in the kernel source tree.)

Lately I’ve been testing a few distributions. A few comments. Redhat (or CentOS) make it real easy to re-compile their sources, so you can basically create your own Enterprise distribution with your own company’s branding very easily. Redhat Nahant EL4 “nahant” type distributions must have quite a high percentage of the server market. I see it time and time again. I find it quite an “average” platform, though I also find it difficult to argue why Debian is all that much better. One policy that interests me is that EL4 updates out the box. Is it a good thing for a box to update itself? I guess so, though with Debian that’s not as usual I find. I hate how all the distributions I’ve tried use screensavers that hog resources. Whatever happened to the screen just blanking ffs? Making the screen actually power down gets bonus points. Extra bonus points: do not power down while I am watching a movie. Xandros is quite an interesting distribution. Debian based. It has this LDAP centric xDMS (desktop management system) for aiding deployment. It also had crossover linux (for photoshop and M$ office) by default, also video works out the box and CD burning. I has quite a bit of spit and polish. It’s a pity they don’t seem to work as closely with Debian as Ubuntu seem to. I’ve got my sister using Ubuntu dapper and she is bitching that she can’t play multimedia files. I found this awful post on the ubuntu forums where it suggests installed j2re. FS, no. I’ve managed to download the win32 codecs and get that going for her over ssh, but I’ve no idea how get mp3 playback working still on her (sic) “desktop”. My sister is also struggling with copy and paste in Gnome (compared to win32). I feel her pain in ion3 and Firefox’s textarea. Ubuntu’s launchpad is painful. I remember some dude raved about it on -devel. WTF. I won’t rant about that. In Ubuntu I am still trying to understand the way the disb is formed and a release is made. The scripts are quite odd. Seed and Germinate ? I am getting confused with preseed.
I also don’t like the MoinMoin Wiki (same tech behind Debian’s Wiki ). Compared to MediaWiki, information doesn’t look as nice, editing isn’t as easy and the searching sucks. Someone has got to write some nice Web interface to these Ubuntu patches. They’re a nightmare for my brain to parse. When I was installing and re-installing different distributions, it would have been nice if the /home directory was left alone. I managed to save it once with the debian-installer, but the other installers were into blanking the drive. I also played about with LVM in the debian installer and somehow caused a strange error. I’m not confident about using LVM. I would like to test disbs without burning ISOs and nuking the HD. I remember QEMU being really slow and VMWARE being non-free and only working properly on Windows. I don’t think you can boot an ISO with Xen. Any tips people? Update: See comments. Or VM Player with Ubuntu samples are OK for testing.

I have been working on non-BMP characters support for Pango for a while: Notice the screenshot shows a screenful of characters with 5-digit Unicode canonical values. Of course, apart from using a non-BMP-capable Pango, a font that actually has some non-BMP glyphs is also needed.

I knew it reminded me of something… To quote from Top Gear:

Focus ST. Ok, the name is ill-chosen, you can get it in a sort of orange tracksuit colour that wouldn’t look out of place on a rapper and it’s bound to become the car of choice for Burberry check-wearers across Essex – but that doesn’t stop it from being a fantastic car.

The ST starts at just ??17,500. Which is really very cheap indeed, less than the Astra VXR. It can also be persuaded to do a respectable 30mpg, if you can keep your right foot in check. It’s performance is frankly staggering, it will go from 0–60 in just 6.5 seconds, which combined with open windows, loud R&B; music and your local town centre is a recipe to impress the ‘birds’.

Put simply, this is THE car to talk about in between pints of Stella while Rooney is sarcastically clapping the ref.

What does it remind you of?

I recently changed my spam filter from spamassassin to dspam and have spent the past couple days training it. It’s been working well until this morning, when I found that it had falsely tagged 3 mails as spam. I found this quite alarming, since false positives are a very bad thing. Even worse, the messages didn’t look spammy at all. When I looked closer, I realized that all 3 had been sent to the NM Front Desk, new-maintainer@debian.org. So I thought to myself, “Why would dspam think any mail sent to new-maintainer@.d.o is spam?” Well, like any other Debian role address, it’s basically a giant spam trap and receives a gazillion spams a day. Since up to that point, not a single mail sent to that address had been valid, so dspam must have decided that all email sent to new-maintainer@d.o was spam. It must have been sorely disappointed to learn that new-maintainer@d.o actually receives a valid mail every once in a great while.

#!/bin/bash
modprobe cpufreq_userspace
echo userspace > /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor
while true
do
	if expr $(uptime cut -d',' -f3 cut -d':' -f2) '>' 0.75 > /dev/null
	then
		echo 1333333 > /sys/devices/system/cpu/cpu0/cpufreq/scaling_setspeed
	else
		if expr $(uptime   cut -d',' -f4) '<' 0.75 > /dev/null
		then
			echo 666666 > /sys/devices/system/cpu/cpu0/cpufreq/scaling_setspeed
		fi
	fi
	sleep 10
done

I tried using some of the standard cpufreq stuff that's in the archive, but none of them seem to work for me; they all make my laptop work at full speed, all the time. Which is silly. Also, the kernel ondemand and conservative governors don't seem to work—when I write "ondemand" to /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor, it comes back with write error: invalid argument, or some such. So, I hacked up the above script. It sets the speed at full (1.3Ghz powerpc here) when the load rises above 0.75 for the last minute, and sets it at the low speed again when the load drops below 0.75 for both one and five minutes. Using the system load average as a metric to decide what speed you run your processor at isn't terribly state of the art, but it Works For Me(tm); my processor isn't put to full speed when some short cpu-intensive job comes up for three seconds, but if something decides to use it for a bit longer, it does turn it up after approximately 30 seconds. Once it reaches that, it stays there until (on average) a few minutes after the cpu-intensive job has finished. I like it that way. Update: I received no less than four replies to this post that I need to make sure that cpufreq_ondemand is loaded, which I can check by reading /sys/devices/system/cpu/cpu0/cpufreq/scaling_available_governors. Yes, I know. I did that. It's there. Yet, it still doesn't work, not even with echo -n. Which is why I'm so confused. It's probably just a kernel bug, though.

Today, I had the "pleasure" to experience a new form of wiki spam in the Unmaintained Free Software wiki. Or at least it was the first time I had seen it... The registered wiki user "Drunkers" (yes, the spammer scripts not only spam anonymously, they also create real accounts lately!) spammed several pages in the wiki, adding tons of spam links, hidden with fancy CSS and other tricks. Nothing unusual so far. Another registered user ("Mootlif3") reverted the spammer's changes with comments like "reverted (spam)", "unrelated links removed", "deleted spam links", and "damn spammers". Or so I thought. The real "wtf" moment emerged when I checked what "Mootlif3" really had done. He didn't really revert the changes of the spammer. He only removed a few of the links from the page, leaving most of them still in there! So it would look like a nice (human) wiki user had helped out with cleaning spam, but in reality he only created a false sense of "security" for people who really want to clean the spam... Damn spammers, indeed.

<p>Came back from a nice walk on the seafront, including a 30 minute professional firework display from a boat 20 meters offshore and about 50 people watching, to find that someone had attempted to attack my blog comment to send spam. Fortunately according to my mail logs, nothing went out, but it did make me go through and read the comment plugin code. I did wonder if every suitable field had been cleaned and have now made sure that it is. Looks like it came from several IP addresses over a 5 minute period and got past my (admittedly very) simple turing test, so I don't think it was an automated script. For an example of the attacks, check out this <a href="http://www.davidpashley.com/cgi/pyblosxom.cgi/general/fists-of-fury.html">old posting</a>.</p>