I saw on Planet Debian that Enrico Zini had some trouble using xen-tools:

Installing the base system.   This will take a while!
	
Copying files from host to image.
Finished
Something went wrong with the debootstrap installation
Aborting

This basically means that the invocation of debootstrap failed to leave the system with a /bin/ls present - meaning it failed for some reason. My immediate mad debugging skillz suggest retrying with either –verbose, or –cache=no. The first will hopefully show the problem, the second will most likely fix the problem. I’ve seen this a few times, but haven’t updated the code to check for it. When creating a new Xen instance of, say 4gb, we copy all the packages from the host systems /var/cache/apt/archives so that debootstrap doesn’t redownload things and use up bandidth. If these cached files are >4gb then the new image is 100% full and debootstrap won’t run successfully! In related news I’m currently working on a system called image-server - this is a distributed xen-create-image script. The idea is that you have a central host which is responsible for creating Xen images of Debian, RedHat, Gentoo, and serving them to clients. This isn’t much use for a home user, but as a hosting company wanting to provide Xen instances of semi-arbitary Linux distributions it should come in handy. In other news I want to work on modularising the Xen tools scripts and moving them onto xen-tools.org.

The European Union wants to tax SMS and e-mail messages (see digitoday news in Finnish). Or at least the French MEP Alain Lamassoure has suggested that, it seems. The tax would be 1.5 cents per SMS, and 0.00001 (one hundred thousandth) cent per e-mail. Current prices in Finnish operators seem to be around 7-10 cents per SMS message. A 1.5 cent tax on top of that seems pretty high. Collecting it would be simple: all SMS messages go through operators. For e-mail, however, the thing is going to be a nightmare. Much e-mail goes not via operators, and there is no good way of tracking it. This could be fixed (from the tax collecting viewpoint) by requiring all e-mail servers to be registered for proper accounting (nevermind freedom and privacy). Update: This paragraph is rubbish, due to my inability to calculate simple things; I'll keep it here in the interest of historical accuracy (randomly changing already published log entries is impolite, to say the least). Anyone operating a busy mailing list will want to move it outside the EU. Even at such a low price, a list with a thousand subscribers and a hundred messages a day is going cause a tax burden of around four hundred euros a year. A company might be willing to pay that; a free software project would probably not.

I'm heading back to Canberra in the height of winter for SAGE-AU'2006 in July. Should be good. This will only be the second SAGE-AU conference I've been to (the last one being the record-breaking one in Hobart). The line-up looks pretty good, and Google is well represented (both in the program, and in attendees) this year. There's going to be a recruiting gig on the Tuesday evening, so if you're interested in finding out more on what it's like to work in Corp SysOps, and would like to have a drink on Google, block out Tuesday evening in your diary. There's still a little over a week until the earlybird registrations close.

After long working, we Debian CUPS team released CUPS 1.2 for Debian unstable. I'd like to say thank Martin Pitt because he has made many patches for CUPS. We hope this series will come into Debian Etch as soon as possible. Of course reporting bugs are welcome, but please consider whether CUPS is responsible or not for your bug :)

I was quiet about it for a long time, but here it is: my first Summer of Code status report. For a bigger picture: I am working on a graphical front-end to the bug tracking system Debian uses. Well, at least I am supposed to work on it. First on my list was something else: a Gtk2 graphical front-end for the well-known bug reporting tool "reportbug", which might later be integrated somehow into Debian's Gnome environment. According to the list of milestones I created when I drafted the application in less than an hour (due to the deadline approaching) I need to release a first prototype of such a GUI today. It took more time than expected to work on it, so I have not yet started any work on the main graphical front-end. I started a few days later than expected, due to some delay until the approved applications were published and I also had to learn Python from scratch. This task was not too hard, but I am still often confused by the lack of consistency throughout the language. Compared to Ruby, a language I know reasonably well and in which I already coded much, Python has much more syntactic constructs for primitives and in my opinion this leads to a lack of consistency. But well, one gets used to it... There was once a "original" Gnome front-end, but it is discontinued since some years already. I did not look much at it, neither at the patch in the BTS which should basically bring it to the current Gnome2 libraries. To properly understand the hooks I need to provide for an interface plugin I started from scratch, only reusing the core idea of the old front-end: the use of the GnomeUI Druid. The new GtkAssistant is not yet ready to be deployed, being introduced with the upcoming Gtk 2.10 release. Still I have a weird feeling about the Gnome Druid, because I abuse it just to provide a consistent look. reportbug does not support backing up to a previous question, so the Back button is generally deactivated. I tried to modify core code as few as possible, as I have not yet talked with the reportbug maintainer Chris Lawrence. I will do so in due course. However after the old GUI front-end code was basically discarded (its usage is discouraged with a prominent notice), many changes to the core code did only take the text interface into account. (Well, I just suppose this, I did not do much research into reportbug's past.) So some tiny changes were needed to provide a few more hooks. There are, however, more to come, especially in connection with one input type (select_options, which could often be implemented as presenting a simple yes/no dialogue). The configuration basically works already, as seen below. What does not work are spawing the right editor and getting a filename from the user interface. But those are minor quirks and not too hard to fix. I need to develop in the open, as specified by the Summer of Code ruleset, and now I could do so, after about four rewrites of the GUI code. The main GUI loop is now in a thread, receiving commands from the main thread running the logic and sending the input back through a queue. (I used a pipe-based approach first, using an IO selector provided by glib, but it did not work well enough and this approach is far simpler.) So the current version is not vapourware, but there is still some more work to do to get it really usable. I will put my work into the Summer of Code Subversion repository as soon as the necessary access privileges are synced onto costa, an Alioth machine. Hopefully it will be merged into reportbug when I completely finished off this project.

A while ago I knocked up some scripts that created LVM snapshots for use with dchroot. Tonight as I was wading through my debian-devel backlog, I discovered that pbuilder can operate in a non-tarball mode, so I was immediately interested in seeing if I could couple the two together. This caused me to iron out a few more issues with the aforementioned scripts, to the point where you can go:

$ sudo create_chroot unstable; pdebuild --auto-debsign -- --no-targz --buildplace /chroots/unstable --bindmounts /var/cache/apt/archives --aptcache ""; \
sudo remove_chroot unstable

And the whole process is a bit faster (I think) than dicking around with tarballs. This is probably throw-away work anyway, given I'm sure cowbuilder will kick-arse, but it was fun... I need to make the create_chroot script take an option to not mount /proc, as at the moment, if you run the create_chroot script and then pdebuild immediately afterwards as indicated above, the latter will fail because pbuilder will fall over because /proc was already mounted in the chroot. I really should go to bed...

As #368922 highlights, when you upgrade dhcp3-client, the version that was started whenever the interface(s) were ifuped remains running. So upgrading the package won't necessarily do anything for you in the immediate term. I guess the best way to proceed will be to parse /etc/network/interfaces and ask the user if the postinst should bounce the interfaces that are found to be running DHCP. It's a little fraught with danger, so I'm going to think about it more before I do anything.

Google Sitemaps isn't just about giving google an easier way to index your whole website. (There are two reasons you might want to give Google a sitemap: One is that this will result in extra links in the search results, since Google apparently uses the sitemap to give the user some extra navigation links with the search result; the other is that this way Google can avoid any flash or javascript magic you might be using - though it's a bad idea - in your webpage.) One nice feature they added is that they give you an excerpt of the web searches turning up your web page, and the average placement for this term. This is a very nice tool to find out how Google "sees" your webpage. Google also gives you a list of common terms used to refer to your webpage. Results for my personal webpage have been very interesting. There are a couple of highly ranked search terms that I would never have expected, that will turn up my page with a high placement. The funniest result is my #1 ranking for Sportfreunde Stiller Torrent, obviously a file sharers search term (Sportfreunde Stiller are a great German indie rock band; check out their new You have to win Zweikampf soccer album. Consider buying it.). I'm not a filesharer. I like web radios (especially when they play lots of indie music and have high quality Ogg Vorbis streams like M94.5) and I use last.fm quite frequently. I have however blogged on Bittorrent technology aspects (as a legal file transfer protocol. Did you know that Debian GNU/Linux considers to use torrent technology for distributing software updates?), and also blogged on young German bands that have made the top of the charts the last few years.

Martin F. Krafft blogged about how to rollback firewall changes in the case you managed to lock yourself out from the box It's even easier if you use my Pyroman firewall config tool. (apt-get install pyroman). If you run pyroman safe it will execute the new firewall rules - and if you don't type OK within 30 seconds, it will undo all changes. Note that it can also restore to a configuration set by a different firewall app. (It just restores the old iptables state and feeds it back to iptables - it will support anything your iptables version does.) Oh, and it's much faster than the other firewall scripts I've tested so far, since it doesn't spawn hundreds of iptables processes, but only one iptables-restore for setting the new rules in one transaction. Check the web page for other benefits; should just work on any Linux distribution with iptables and python (read: every). [Update: Martin, I was referring to the instructions you gave, to adding an at job and then running atrm to accept the changes. Yeah, what you script does is basically the same what mine does for rollback.]

You might have read that ext2/ext3 has support for so called "directory indexes", which supposedly speed up operations on directories with many files. For example your /usr/share/doc directory, your packages database, or your home directory (use ls -la wc -l to count the number of files). However, this feature is not enabled on most machines, since up to now you needed to pass a flag when creating the filesystem. The 1.39 upload of e2fsprogs (in Debian unstable as of May 29) made this default. So you probably didn't enable this when creating your filesystem. To check, first find out what your filesystem device is (use mount). Mine is /dev/hda5, so my example will use this device. In traditional terms, this is the first extended partition on the IDE primary master drive. Now use tune2fs -l /dev/hda5 grep features to get a list of your filesystem features. If you have "dir_index" in there then you're all set. Ignore any "needs_recovery" - that is good for mounted filesystems (it just means "not cleanly unmounted") To add dir_index, do the following.

• Unmount the filesystem. If this is your primary filesystem, use a bootable CD like Knoppix or some Linux install CD.
• Make sure the filesystem is really not mounted. Just for safety of your data. Run e2fsck /dev/hda5 - this should take less than one second if your filesystem is clean and is needed anyway otherwise...
• Execute tune2fs -O +dir_index /dev/hda5 to enable the feature for your filesystem
• Execute e2fsck -Df /dev/hda5. This will force a filesystem check with directory optimization (!) on your filesystem.
• Mount your system (or reboot, if you were using a live CD)

Recently I was working on an application for ICRA2006 and I needed to output some data to be read by Excel. This would seem like an easy problem, just output in CSV and Excel can open the text just fine. Unfortunately Excel has one of the most backwards implementations of international character support that I have ever worked with. Being nieve to this, I simply tried UTF-8. I was greeted with multiple strange characters in the place of my normal umlauted and tilde-d letters. A quick search turned up that when Excel reads a CSV, it assumes that it is in ASCII or . However, you can import the CSV as a text file and select it as unicode. This is where my real pet peeve is: You can't select the character set of a CSV in Excel. You have to rename the file as a .TXT file and go through a wizard, basically spelling out that yes, it is comma separated and yes, this is the data I want to import. There's one drop down box which specifies the character set. Of course Unicode is supported there.

In the search for a solution to this problem which didn't involve a 10-step process for the customer, I discovered the Unicode marker which is supported by Wordpad, but amazingly, does nothing but add a bit of garbage in Excel. The sad part is that I can be almost certain that the next version of Office will continue to have this problem. My only hope is that somehow the XML format will be open enough to just create a native Excel file.

Well, Mexico played three preparation games in Europe on their way to start the World Cup. They played against France and the Netherlands on their territories and well, I think Mexico was not ready enough, it is not solid and the manager was making lots of crazy movements. They ended 1-0 and 2-1, respectively. Obviously those were difficult matches against f tbol potences on their home, but still Mexico has plenty of capacity to beat any of those countries. The third game was against a German university, Gottingen, which was more a diplomatic match (the manager and players didn’t want to play it, but it was because of the chief managers of Mexico, trying to make a good relationship with the city and/or Germany, itself). We won 0-3. Next week, on Friday, the World Cup will begin. The Mexican team will be playing firstly, against Iran, then Angola and finally Portugal, on group D. I don’t see Mexico solid and trustable enough yet, but we’ll see how it goes against Iran, which have been playing widely good (they won Costa Rica 3-2, back in March; drew 2-2 with Croatia and won 5-2 against Bosnia-Herzegovina; these last two in the last couple of days). Iran is not at all, an easy enemy. We’ll see how it goes. I’ll be reporting more World Cup news on the next weeks

(Sorry Robert, this is a meta-blog-post. Perhaps I should create a catagory for them?) Since being syndicated on Planet Debian, the amount of comment spam I’ve gotten has risen sharply. No spam has made it onto th site itself, since I moderate the first comment someone makes, but the emails have gotten annoying. Therefore, I put a turing test into place. Blind users will be happy to hear it’s not a CAPTCHA, instead it’s one more field to fill out that asks a very basic math question to prevent spambots from commenting. Users who registered back when I required it won’t even get the field, isn’t that nice?
To the readers of Planet Debian who are probably asking “Why Do I Care?”: If you don’t mind, try out the system to make sure it works. I figure the readership of that site will have a fairly diverse set of browsers and maybe even a few using accessibility ehancements that can confirm that the system has no problems. If you encounter any problems, email me at astronut@dlgeek.net or ping me on IRC (astronut on both freenode and oftc).

This is the time of year when most Free Software-related conferences in Mexico have their calls for papers. CONSOL's CfP closed two weeks ago (agh, I had to CONSOL's CfP closed two weeks ago (agh, I had to /lists.consol.org.mx/pipermail/consol/2006-May/006443.html">/lists.consol.org.mx/pipermail/consol/2006-May/006443.html">send my proposal by mail, and I do hope it gets in the official program - For sentimental reasons more than anything else ;-) CONSOL is, after all, still my baby!). Yesterday I had to come up with something, as CICOL's deadline was also hanging upon me. Now my attention has been brought towards FSL, for which I still have almost one month to think. And, of course, although at other times of year (but, of course, still clashing with other conferences) we will not have one, but three instances of GULEV this year..
WTF?
Mexico's Free Software community is a small and not too productive one, as most communities in our (culturally defined) continent, Latin America. And yes, conferences are quite fun, it's always nice to go to a nice place and meet your friends. But we must not forget the real motivation for them all: The academic program. We currently have four conferences that try to reach to the same audience (yes, in different places of the country, but all of them with a national scope). And if you compare the list of proposed talks (visible currently for three of them), you will see there is no real difference between the set of talks on them. Please explain me, what kind of incentive does this give to anybody to attend? Just to hang out with the friends, entering a talk here and there?
I ended up submitting the same talks for both CONSOL and CICOL, as running Debconf didn't leave me time to prepare or think over anything - That's not what I like to do, and I hope I can be a bit more creative with the topics for FSL. There is no published criteria yet on what will be the focus for (each of the) GULEV, so I'm not committing yet to them. And yes, being a very active Free Software promoter in my country, I try to be present everywhere - but of course, everything has a limit. I will most probably not skip CONSOL (as it is on the city I live, and it is the one I am most sentimentally attached to) or CICOL (as it is just 1hr south from here)... But I'm still uncertain if I'll be at FSL or any of the GULEVs.
Back on my initial questioning, anyway: Do we need four perfectly equivalent conferences in this country? Besides ego clashes, wouldn't we be best served if we had only one or two big conferences a year?