Yesterday, I had a chance to talk about Extremadura i18n meeting at Tokyo Debian monthly meeting. This meeting was organized by Junichi Uekawa and wonderfully 20 or more people came and filled the meeting room (awesome!), thanks! My paper is included in this PDF [in Japanese]. I picked up and spoke some themes from what we discussed at Extremadura; i18n task force, Pootle system, and language-pack/tdeb. PO-based translation has an advantage and a disadvantage. The tools support PO are mature and PO is very useful for the short messages. But using it for documents, such as manuals or Web pages is not so useful. Especially translating to Japanese sometimes add/delete/reform sentences. Some attendees suggested that we would always use the original XML (or something) format for translating/reviewing and have PO format in the central database. Masayuki Hatta pointed the license of translation. For example, Japanese is completely different from English, both words aren't always correspond 1 on 1... Translators use their creativity power. He said translators would have a copyright for their translation instead of original author. When I talked with Warren Togami was Fedora US leader last month at Tokyo, he said his project asked to agree "translation license agreement" to all translators. If we Debian or Debian-JP apply a central translation system like Pootle and use its translation memories, it's time to consider making similar license. Although current idea of Tdeb is still under the development, most attendees loved that idea rather than language-pack. For me, the goal of this meeting was not only speaking my experience report but recruiting more people for i18n/l10n working also. I'd like to recommend Nabetaro-san for next i18n meeting attendee instead of me, and Noritada Kobayashi for next of next. I strongly suggested attending next Debconf to them also. They are hardly working for Debian i18n/l10n in Japan. I hope attending such meetings expands their personal connections and motivations. There were two other sessions at the meeting. Matsuyama-san demonstrated creating Flash on Debian by using ming, and Junichi talked about profiling apt-get by oprofile tool. Both topics were interesting for me.

Last week Pascal, who is a friend of bubulle, asked me how he could make ISO image with kernel 2.6.18 for Sarge. Timely I'd like to have a kernel 2.6.18 image for myself also (I bought Core2Duo machine. It has VIA VT8237A SATA controller isn't supported before 2.6.18), I made it. You can take it from the usual place; sarge-custom-1008.iso for i386.

• Backported 2.6.18 kernel from unstable.
• LVM/RAID will *NOT* work. Maybe I have to backport partman and dmraid to support them, but I haven't done yet.
• Since kernel 2.6.17, Debian kernel team merged SMP support to a single kernel. So don't panic even -smp kernel isn't installed. ISO image has -486, -686, -686-bigmem, and -k7.
• I haven't AHCI machine and I don't know PATA CD vs SATA HD problem still happen on 2.6.18. But I add a feature "atapi=true" boot option. When you append this option as the boot parameter, the installer will add the configuration to /etc/modprobe.conf.

The meeting at Extremadura was excellent and productive. We all discussed many issues and enjoyed to know each cultures. Thanks all participants and staff. Especially I say thank you, Cesar. Gracias. See you again. OK, I backed home and have many TODO... writing a report for Japanese team, working for some issues what we discussed at the meeting, and... gee, checking/fixing a bunch of bugs against CUPS.

Today, mrmt, ishikawa, kmuto, Ar-, nnn went to Kashiwa NOC where some Debian JP servers and JLA servers reside. They upgraded their servers to sarge and I, from my home remotely, upgraded Debian JP servers, JLA servers and my server to sarge and kernel with help of them. misato.debian.or.jp: this is old server. backup necessary files/data and recalled. plat.debian.or.jp (debian archive mirror): this leaned in the rack (maybe due to the earthquake at Kanto area recently). Shutdowned and relocated. IDE controller was somehow broken, so other old server is used as replacement. While replacing, /dev/md1 was broken. /dev/md1 is partition for debian archive mirrors. I tried to fix it. At first, I fix partition type of hdb1 (Linux -> Linux auto raid detect) and recreate md.

 # mdadm --create /dev/md1 --level=linear --raid-devices=2 /dev/hda1 /dev/hdb1

Then, I could fsck on /dev/md1. However, many inodes were broken, so it seems a lot of time to fix it (note that still fscking snapshot.debian.net old archive disks (1.4TB) for 40 days...) All contents in the partition was just mirrored contents, so I recreate the ext3 filesystem and mirror from upstreams. It would take 1 day or so to recover full mirrors. arashi.debian.or.jp (Debian JP web/mail server): After upgraded sarge, kmuto upgrade the kernel to new version. He rebooted but no response. He finally found that VGA card was broken. He replaced VGA card from recalled server (misato.debian.or.jp) and it worked. I noticed upgraded dancer-ircd couldn't communicate with older version of it. (claimed failure to match password). So, I upgrade all of our servers in our IRC network. minori.linux.or.jp (JLA backend server) and norika.linux.or.jp: I upgraded them to sarge. Except dancer-ircd problem as above, the problem (reported by takei) was that customized cvs pserver line that I put in inetd.conf was disabled by cvs package. I re-enabled it. These works took 5 hours, and still mirroring.

Lars Wirzenius kindly added some visibility to Kenshi Muto's Debian GNU/Linux hardware compatibility list, and had the interesting idea of mentioning live-cds in the same post. That gave me an idea: why not prepare a minimal live-cd dedicated to hardware compatibility testing? It could automatically open a browser on that check page, try multiple screen resolutions and sound systems and report the results, run glxgears, etc. Just boot on that CD in the computer shop, wait for a few moments, and see in big, easily readable letters, whether a configuration actually works with free software. Don't forget to exclaim loudly if the shop attendant told you it would work and the tests show it doesn't. I'm told there's a tool to automate the creation of live-cds. Any takers for that task?

Although Masayuki who is Debian Ghostscript maintainer has already created pkg-gs project on alioth, it wasn't organized well for a long time. Current ghostscript quality in Debian is far from good for Etch, and Masayuki is a quite busy person. Last week, we discussed on IRC and decided to reorganize this project. Therefore we started the mailing list (pkg-gs-devel) and SVN repository. (thanks alioth's admin!) For first step, I was starting to create gs-esp 8.15.2 package. Here is current status. After fixing some problems (such as crashing with ttf-kochi font), I'd like to upload this. Our short term goals are:

• Close bugs as possible as we can. :)
• Release "ghostscript" package instead of gs-gpl package.
• Remove Gs-afpl package. Because it was merged into gs-gpl.

I'll be happy and give a beer if someone make CMap DFSG-free...

Today I built the backported d-i image with kernel 2.6.16. I updated many codes also.

• Applied 3.1r2 for the base.
• Applied kernel 2.6.16-15bpo2 was taken from backports.org. This version has CVE-2006-2934/CVE-2006-2451 fixes.(thanks Norbert!)
• Copied partitioner codes from unstable-installer. For example, you can choose LVM auto partitioner.
• Included linux-headers-2.6.16-* in ISO.
• Included megaraid_sas in initramfs-tools/hooks-functions. (I dunno whether this is correct or not..)
• Package updates are introduced by kernel update and backports.org up-to-date: cdebconf, di-utils, grub, initramfs-tools, klibc, libdebian-installer, lsb-base, parted, rootskel, udev, yaird

You can download the image from backported d-i images portal. Note: Although I'd tried 2.6.17 version once, finally I gave up. I succeeded to build kernel image from unstable source on Sarge. But linux-kbuild package, for linux-kernel-headers helper, was failed to build with gcc-3.3. Gcc-4.0 or above works, but they don't exist on Sarge. (Don't nag me about CUPS... Most problems are upstream's matter and upstream is preparing 1.2.2 release at this time. I'll run again after it's released.)

<feeling blue> I'm bit bored and tired to fight against CUPS bugs. I've dedicated most of my free time to see reports, to try reproducing, to try solving, or to ask it to upstream.
I'm frustrated now, but not by upstream or its source quality (of course I hope it becomes more better). My frustration is that most people only nag us even they're using 'unstable' and have a skill. Though Roger, Henrique, Martin Pitt/Eric, and me are working as possible as we can, our human resource isn't enough. Investigating bugs, checking a source code, and making a patch are very helpful for us. Thanks. </feeling blue> So, I was making some packages for recreation. Because I haven't a motivation to maintain them on Debian, I won't ITP.

live-f1

I saw Newsforge's article. Although I'm not so interested in Formula-1, my friends love it. Because I built this package on Sarge, it isn't apt-get-able on unstable environment by missing libneon24. Get libneon24 from Sarge or rebuild it on unstable. You need an account on Formula-1 Web site to see.

boodler, boodler-soundlibrary

I saw another interesting article. I was looking for good background sounds, but not musics. Boodler is cool sound generator for its purpose. My favorite sound is 'frogs.Cheepers'. See details on /usr/share/doc/boodler/catalog.html.

You can get the packages from my repository.

deb http://kmuto.jp/debian/mtu unstable main contrib non-free
deb-src http://kmuto.jp/debian/mtu unstable main contrib non-free

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 :)

At least since a few days ago, my ARM buildd for experimental/unstable non-free/volatile/etch-secure/d-i went broken (ARGH!). Dpkg always caused a segmentation fault and I couldn't find any meaningful messages from the system. After I rebooted the machine, it won't reply anything... I'll try to connect a monitor and to solve the problem on this weekend.

I backed home yesterday. After I arrived at home, I went to buy foods to cook, checked bills and snail mails, washed my clothes, took a bath, saw through a bunch of e-mails and gave up to reply to them, met my partner again, cooked for dinner, and slept well. Fortunately I haven't a jet lug at this time. Because I tripped China for 1 week before Debconf6, I was very tired at the beginning of Debconf. Furthermore mixing Spanish/English/Japanese caused me a mental fatigue. I felt better as time went by. Well, I would more enjoy if I wasn't so tired, but I enjoyed BOF, talks, swimming, foods, travelling and so on during Debconf6. 'Gracias, Debconf staff and attendees.' Mexico was an interesting place for me, so I'd like to visit there again in the future with my partner. After Debconf, we - Nobuhiro, Yukiharu, and me - stayed Toronto for 2 days. It was very nice time to see Niagara falls and to walk through Toronto downtown. I uploaded the photos: Debconf6 pictures, Canada picutures. I apologize I couldn't work well for Debian, especially for CUPS during the conference by a network problem. CUPS 1.2.1 is already released and I'm working for it.

Yeah, I arrived Mexico! Here is bit hot, but isn't so much than I imagined. Today, we, the party of 13:00-14:00 arrived members challenged the 'Debconf triathlon'. We took the subway method. Going up-down, up-down, up-down, ... the stairs with carrying heavy bag is really good exercise. I thank myself because I'd bought soft and light bag for this trip :)

This evening, three Japanese people including me will get on an airplane to Toronto. After staying one night, we'll go to Mexico. See you in MX! ... I saw Jesus's report from Mexico. Well, I haven't any shorts. I have optimistic view to wear with jeans, because Japanese summer is very hot also (30C-35C) and many Japanese usually still wear jeans, long pants, or even business suit(!)

Because I'd gotten a request from Chris via writeback, I made a front page (http://kmuto.jp/debian/d-i/) of my backported d-i images.

After long time working together with Martin Pitt, I uploaded CUPS 1.2rc2 to experimental archive. This version introduces and improves many features; IPv6, LDAP, localized PPD, hardware detection, etc. Plus, it may make an another bunch of bugs :) For example I noticed KDE-control-center complained by ABI incompatibility around IPP. CUPS 1.2 provides better i18n support. If you're interested in the translation, see upstream's document. (AFAIK Swedish translation is been working by Daniel Nylander.)

Although I was still busy, I tried to build d-i image for AMD64/EM64T (x86_64) today. To do try & error, I rented DELL SC430 EM64T machine from my friend DD Shigeo (Thanks!). This machine is really good showcase to demonstrate how we can support tricky machines.

• 64bit architecture.
• Keyboard is connected via USB. There isn't legacy PS/2 emulation.
• CD-ROM drive is connected with parallel ATA, but HDD is connected with serial ATA.
• NIC is a kind of Broadcom. Although tg3 driver is for this, kernel 2.6.8 needs to take firmware from somewhere.

In short, official Sarge d-i can't work on this machine. Using backported kernel 2.6.15 and other things, I succeeded to build experimental installer.

1. Download mini.iso (small CD image to launch network installer) or netboot.tar.gz (for TFTP boot) from here. (I've tested only netboot.tar.gz version)
2. When installer asks a mirror site, set "osdn.debian.or.jp" as host and "/~kmuto/sarge-amd64" as directory.
3. If you meet "no disk" error, go another virtual terminal (Alt+F2) and run "modprobe ata_piix". Then back installer screen (Alt+F1).
4. Remains are as same as normal installer. Choose other official mirror at 2nd stage.

Known bugs:

• Discover doesn't load ata_piix if piix (ide_generic?) is loaded before.
• USB keyboard list provides only a few choices. Use US map at this time.
• For CJK people, you may find 2nd stage screen won't show in local language. This is because some kernel images don't include fbcon driver and jfbterm can't run.
• I haven't tweaked RAID/LVM yet. So it may not work.

I'm trying to make auto-build script also. Although it's still under developemnt, you can see svn repository; http://kmuto.jp/svn/d-i/sarge/

I changed a plan. Because it's too tough to visit Vancouver in the trip, I decided to stay at Toronto for 2 days.

5/12 17:15 NRT-16:40 YTO (staying at YTO[Toronto])
5/13 09:00 YTO-12:45 MEX
5/22 14:10 MEX-19:45 YTO (staying at YTO)
5/24 13:30 YTO-5/25 15:55 NRT

I'll go seeing Niagara :) ...So to get a long vacation in May, I'm working hard and am very busy at this time. I'm sorry but I haven't an enough time to response a pile of mails till the end of April.

Currently four Japanese will attend Debconf6; Junichi Uekawa (you know him well), Nobuhiro Iwamatsu (SuperH guy), Yukiharu Yabuki (was working at OpenSource SIer company) and me. Today I reserved an air ticket of Nobuhiro and me. After Debconf, we'd like to visit Vancouver also.

5/12 17:15 NRT-16:40 YTO (staying at YTO[Toronto])
5/13 09:00 YTO-12:45 MEX
5/22 14:10 MEX-19:45 YTO (staying at YTO)
5/23 08:00 YTO-09:57 YVR (staying at YVR[Vancouver])
5/24 13:10 YVR-5/25 15:25 NRT

Now some skk packages are orphaned because maintainer Mr.Kawamura is marked as MIA. 'skk' is obsolete and is disconnected from upstream, but 'ddskk' upstream still be developing very active. I take over ddskk for myself.