Things are improving. My dancer-diary script is rather interesting. dancer-diary-mode is rather useful, and I think it will be so for a while. However I need to find out a way to make a default blank text. Template. I have to have this template thing, because I am copying over files over and over again, every time.
updating the copyright guide, and debian newbie documentation. I think I need to be very careful about the copyright guide, because I don't want to be advocating something wrong.
Managed to get a dual-monitor system up and running. I have installed a PCI S3 Virge card, and configured XFree86 4.0. However, with +xinerama option, gnome applications (such as nautilus and gnome-pim) seem to stop responding. I decided to use the displays separately. I can move the mouse over across the display without using xinerama, so it won't be much of a problem.
update-cluster is a very exciting project. I don't know if I should have used xml for the data type. It gives the illusion that XML is useful when it is not. XML is a nice tool to use in that it is well-defined, but it is not a very nice-to-use tool in the current state of Unix tools. It does not have the equivalent tools for Unix traditional textutils, like sed, tr, awk, and other things. It has the xslt, which is very nice to see, but a good xslt processor is not readily available.
My computer will not be readily available in the lab, and I would guess I would stop updating information for a while.
update-cluster needs some fixes. I am wondering how to go for it.
Here is a screenshot of something being seen from Intel Play microscope. The magnification is not that great.
Trying to code something feasible to make update-cluster script to work. Something quick and simple is in order. Great. It is working fine, I think. I will upload soon. Maybe tonight.
hacking pbuilder, a personal auto-builder. and easy-to-use XML parsing library. It should be necessary.
Chatted about Japanese translation on irc with Joey. There are so many outdated, and untranslated pages. This is rather bad. Status page for Japanese translation seem to be really bad.
Many things remain.
Played with bochs and took some screenshots. It is very slow, but optimizing with Athlon optimization made it run a little bit better. Debian installer startup screen before syslinux loading Linux and the initial Debian installer screen . I hacked a line of code from bochs to use the fixed font instead of vga font, because I don't like the vga font. Anyway.
I don't feel like doing anything today. Maybe a little hacking is in order.
Thinking of making a job-distributing hack. Using LD_PRELOAD to run rsh instead of just doing exec is one idea. -- anyway, I coded something which runs ssh with any command input; nicked some code from auto-apt. It was really good.
I lost my mobile phone. This is pain! I left it in a bus or something. Hmm...
http://www.emaculation.com/ seems to me like a hot site, talking about Mac emulators.
My time is severely limited. Some request for diskless came from bam. But I guess, I don't know enough about diskless to do it. This is rather difficult man!
Researching about IA64. It is an interesting CPU. It has loads of registers, and it seems like one can use the CPUs as a stack. It might be a limiting point, it might be a very very advantageous point. I have an impression that task switching might suffer, but I have not yet looked at it enough.
Life is overly complex. I am rather worn out. My prof. is going to be absent from University for a year. This is rather serious.
Webcam works. cpia-control utility from webcam.sourceforge.net seems to be able to control the proc interface. However I could not get to make the Intel Play microscope lighting. It should be able to go on and off, but I could not make it. Next try.
Having a depends line like os-linux, i386-platform, athlon-platform would really be interesting. Then this would really work. Yes. I don't know if the details should be, but well... Wichert has good answers. dpkg-shlibdeps should generate such a depends, I guess. It scans binary files for dependencies. It should know when binaries need such.
I need more coding. Hacking on dmachinemon reveals that I make a lot of mistakes.
Thinking of writing a very simple script to enable using personal chroot environment, and building in there. It would be kinda trivial, I hope.
added toplinks style to the page. It should be an interesting addition. It is rather difficult to update every diary page... maybe.
athlon-builder is on my people.debian.org page. It should be an interesting addition. I might go and get hold of pentium-builder if the current maintainer is not willing to hack on it.
autotrace seems to be really fun. It likes simple drawings, and it does a good job of doing it (a screenshot). However, when the picture has a gradiation background, it does not really like it. I think it is checking for consecutive pixmaps which have the same color.
I have set up an Athlon buildd. It should be able to build most packages, and it is quite interesting. Very simple buildd software indeed!
Testing the diskless package. It seems like it is working fine today. It is weird... Wow. I don't know. I created a chroot, and ran the thing from there. Maybe because I've read the manuals, and know what I am doing, I have a better mileage.
I've created a quick hack from pentium-builder, called gcc3.0-athlon-builder. It runs gcc-3.0 with the appropriate command lines with athlon optimization. It seems to work as a good hack.
hmm... recompiling gtk+ seems to have improved my system performance. Maybe I am dreaming...
libc6-sparc64 is a package that seems to be 64-bit for ultrasparcs. I wonder whether it would be any faster.
HP presenting a P2P TCP/IP connection for IRDA and I think this is the near future goal I am looking at. IrNET and discovery manager... hmm... Sounds promising.
grmonitor seems to be an interesting program, it gives me a 3D view on process load. I think it should be used as a frontend to dmachinemon. I guess I need to put up dmachinemon somewhere... somehow. However, it is too slow on my system...
mgm is an interesting load meter tool. It seems to be doing quite well.. and looks pretty. Almost like gkrellm in the fanciness. It is going in the same direction...
procmeter3 looks very difficult to confiure. I could not make it do something interesting.
spong seems to be something that competes with dmachinemon. This is a bit of a problem. I have not seen it in detail, but it looks like it. Hmm.. it aims to be a networking monitoring tool. And it doesn't look like it does what I do, but its design is very similar to dmachinemon.
xmcpustate seems to be a very old-style program. It does monitor single node.
I cannot get diskless program to work. I need to cool down and think why it does not seem to want to work. It is really a simple program which I could probably rewrite in shell and something like it. Well, I am maybe too confused about everything.
dmachinemon should be released. It is in a very bad shape, but I think I should announce my intend-to-package now.
Hmm.. I got disconnected from IRC. IRC is becoming an integral part of my daily life. Having no IRC is like my telephone not connecting.
I changed my network setups. I have found out the way I can use the -f flag to ssh usefully. ssh is quite amusing to use. It can be used to set up such a complex system. However, I don't think ssh is an application, it is rather more of a networking and transport layer than an application...
I went to Gion Matsuri and it was amusing. Loads of girls with Yukata were there, and Yukata does make girls look pretty. The way they wore them were quite unprofessional, but I would say, the young blood needs to continue the Japanese traditional clothing.
I hacked further into diskless package. diskless-createbasetgz command now has a manpage. The overall documentation has not been touched yet, and I have not tested if other commands will work with the base.tgz created by debootstrap. I need to test it, maybe tomorrow.
I am planning to go to Gion Matsuri tomorrow afternoon. It would be a really nice experience. I am looking forward to going there tomorrow.
I've sent a bugreport to mpich, which mentioned that there was a problem with compiling with gcc-3.0, and the response was that there already was a bugreport and patch for gcc-2.96 available. Yes, and I found that after having sent the bugreport. Embarassing.
SCore seems to have become free software. It has a licensing that is capable to enter Debian Project. Good.
I was looking for methods to play Macromedia flash. flash? seems to be one interesting project. But I think it has not been updated in a very long time. It has been able to run the St Pirans page without problems. I thought quanta might be interesting but it was a plain html editor. Flashlike project seems rather quiet and dead. It has not released anything so far. I don't know.
man-db seems to be fixed. I dunno. I went out swimming, and ate some good food. This is nice.
hmm.. man-db changed its output format, and confuses the viewer software. lv is confused, and jless seems to be confused enough as well. This is bad news.. man-db changed policy on cutting lines, and I am guessing that it might be wrecking multibyte. I don't know. It's my wild guess, but looking at the output, it looks nothing like it used to be.
There seems to be an undocumented option in mpich configuration. Yes I know there are many undocumented options there... but this one seems to be widely known. ch_p4mpd device seems to exist, and I did not know it.
Making things work is more than difficult. I have installed a few machines and set them up, and GNOME seems to be a big pain to get working properly. Woody doesn't seem to have a proper task support for GNOME and installing GNOME anew is a process of trial-and-error.
There was a machine with SiS chipset, and it has USB device with OHCI interface. The Linux kernel 2.4.6 does not seem to be able to decipher the USB interaface. I had hoped that ac1 patch fixed the problem, but it seems like it doesn't. I have looked at the code, and I am rather impressed that it hasn't been touched for a long while, or so it seems. OHCI is known to not to work on certain occasions, as it were.
I tried running a code like :
#include <unistd.h>
main()
{
while(1) fork ();
}
and it froze my machine completely on kernel 2.4.3. Interesting.
I need to code, but I am overly tired. The air-conditioner, or the cooler, or whatever you call it, is killing me. This damned nation likes overcooling. I hate this.
Some Debian Developers are going to visit Japan, as it seems. It would be really interesting to see them.
beast in Debian unstable/testing is broken. It enters an infinite loop. Great, this is a bad problem. And it seems like I have never contacted the author on packaging the software. This is rather embarassing...
I am really going to hack diskless. Yeah really. I am too busy with my real-life work, but I hope I really can do this. It would be really sad if I can't make it into woody, with a working diskless package.
Debian woody freeze has begun. Starting with a freeze on policy. This will be an interesting thing. I was hoping to create a good system.
debootstrap woody tmp http://www.jp.debian.org/debian gives me a base system with woody. I guess I need to use this for my diskless package. One note is that I need to set my locale to C, or console-something fails. This is bad.
$Id: 200107.html.en,v 1.60 2002/04/22 04:36:51 dancer Exp $