How to write a good manual is a very good story on this topic, writing a good manual. Writing about the concepts, and not the feature, is important. And a good manual needs a good writer, a good humour.
My life is harsh. Things are giving me pressures.
Resuming the play with ecasound. Finally checking out CVS and trying to build things for alpha. I hope it will be working fine.
Configured xemacs21 to my likings. I have been playing with emacs for a long time, but maybe I could move over to xemacs. It is looking good. Interesting. I may need to look into what is different in the two versions.
pbuilder is enormously useful. It is good, simple, and exciting. Glad I decided to write it, despite of sbuild.
uploaded pbuilder and immediately after, I realized something was wrong with it. Doh. It can't build NMU packages. I uploaded soundtracker. I don't get bugreports for this package currently. Maybe I need to change the maintainer field and become the maintainer. I am not sure.
Looking at dhelp. I don't like the way it is currently done. It uses a DB, and it is not compatible with previous versions of itself. Too bad. This package is unmaintained for more than a year too.
Fixed dancer-diary.el to work with emacs 21.0.104. It seems like the behavior of concat has changed. It doesn't take any argument, it likes some string argument only. No numeric argument.
People are really active with hacking on Linux. It's fun. World needs more activity. When everyone is very active, things get pretty exciting. But the fact is, software projects do not just appear from nowhere, it needs to be done to a full extent by someone initially. A project which tries to start up with multiple people fails. It needs to be taken care of by some small number of people doing intensive work. It's interesting.
pbuilder seem to be ready for general consumption. I wonder if I should upload. Maybe a little more consideration.
pbuilder seems to be quite okay now. I probably would create a pbuild frontend program, then release it to the wild. It is rather good. Nice shell coding.
Managed to make a pbuilder package, which seems to be working. Doing a pbuild-buildpackage.sh ./ecasound_2.0.2-2.dsc starts extracting the base.tgz and builds the package in chroot. Great, huh.
IBM Linux Scholar Challenge seems like an interesting thing. I am at least interested in one section of the project, and I will be investigating more.
I checked with xengine and I found that ATI rage card support has improved somewhat in XFree86 4.1. This is impressive. It had around 800, and it now has around 1800. This is a very simple benchmark, but anyway, impressive.
dist-upgrade is causing a lot of problems. My console is filled with dhelp_parse: can't open /var/lib/dhelp/titles which sounds like a real bad news. Maybe dhelp has died off somewhat.
abiword 0.9.2 has been available. It seems. It's strange. They have become quite active recently. I need to go and fix something then. I will try and see what abiword currently is like, and go and touch it. Maybe a NMU is in order.
Updated dsh. It has been a while since I have last been hacking on this project actively. I need complexed topology to be finished, or it will not be too useful.
Joey Hess posted the following snippet in debian-devel: VERSION=$(shell expr "`dpkg-parsechangelog 2>/dev/null |grep Version:`" : '.*Version: \(.*\)'). It seems pretty nice. I think it needs a grep ^Version: to protect from errors. I need to update some packages. update-cluster has a crude hack.
Bought new album by Bjork. It is of different taste. As usual. Makes me think. This is impressive.
soundtracker is taken over by me. Yoohoo. I hope I have not disrupted anyone.
Running dpkg --forget-old-unavail saved about 500K on my system. It is funny. I never even thought about running this.
Going to take over soundtracker for a while. It has not been updated for a while, and I guess it needs updating and some attention. gotom is going to take xipmsg. Great. I removed xipmsg from the list of packages I maintain. This list should not really be done manually, I guess. Some scripting is required.
Went to a house in MIE, close to a river, in the middle of mountains. It was very cold up there, and swimming in the water was quite a chilling experience. Fishing was pretty impressive.
I am worried about wv. It seems to be not updated. I am not quite sure. At last abiword has been updated. However, it does need some tweaking to support Japanese in the package...
For update-cluster, this snippet should be useful: awk 'BEGIN{printit=1} /^#Autogenerated by update-cluster/{printit=0} {if (printit) {print}} /^#End update-cluster/{printit=1}' or awk '{printit=1} /^#Autogenerated by update-cluster/,/^#End update-cluster/{printit=0} {if (printit) {print} }' . Maybe I should add this in /usr/lib/update-cluster/comment-remove.awk as a useful tool.
I'm feeling really unwell today. Hacked on dmachinemon. It is not really improving, but the build system is. I think I need to split out dmachinemon-master a bit.
Installed a new HDD on my home machine. It used to be 6GB, and now it is 40GB. Great. The fan seems to be not doing very well... making funny sounds. It might not last very long. Money seems to be required to keep a machine going. It's interesting that the Dell machine has had minimal maintenance over the last .. 8 years?
The Prime Minister of Japan have visited Yasukuni Shrine. The flower was paid by Koizumi personally, and it was labeled as Junichiro Koizumi, The Prime Minister. It looks like Koizumi have given up giving the prayer for the deceased on the Day of End Of World War II. It should have been done on 15 Aug. Really. It is reported he has done was not a 2-rei 2-hakushi 1-rei, but just a 1-rei, and there was no tamagushi houtei. It probably means it was not a full service, and he did not follow the Shinto ritual. I really think the interaction between this religious service, and countries like China. They need to understand what is personal and ritual, and what is political. And Japan needs to learn that things are different. But it had to be done on 13th, maybe, because otherwise, the possibility is too large that he would have been shot dead by a culprit.
ecasound has a new config.guess and config.sub, and it seems like it is working. I need to fix the gcc-3.0 incompatibility that seems to have been lurking around in libasound1 related code. I also think I need update build-deps. I think many more arches have moved over to libasound2, and I can go over to there...
My home network connection is so slow. It is really annoying. I need to think of a better method.
After running libtoolize I cannot compile ecasound. I am wondering if it is passing wrong parameters to libtool... Anyway, I have a serious bug filed against ecasound, and I'd rather do something about it.
Copying back ltconfig and ltmain.sh from before libtoolize made compilation to be successful. I think all I really needed to do was to play with config.guess and config.sub
Some projects are running underground. It's funny. Kitame got mozilla. This is good, I guess. New mozilla would be readily available. Too high load on one person is kinda scary though... if he loses interest, or gets hit by bus, it would cause a lot of problem.
It's strange but network connections seem to stall through ssh... It has been like this recently, and for a while. It might be because I upgraded ssh on my machine. No idea.
update-cluster has several things pending. update-cluster-remove needs to accept no-parameter as a parameter, as it were. To work as a formatter. Also, it claims to be update-cluster-removeline which it isn't. It has some tweakings to do, and some update-cluster scripts should be re-written, to versions not requiring an exclamation mark. update-cluster-parseconfig should barf out if it has a missing config file entry, and scripts should have more error handling.
I want to write an elisp script which does everything. find-file seems to do much of what I want. I want to know if a file is newly created or not, and I just wonder how it can be done. It's not blatantly obvious. Also, with guile, I think I might have been better off writing my configuration files in guile for dsh. I also think update-cluster could have been done this way. I gave up on lisp config files because I thought there weren't any lisp config file loader. However, there was.
update-cluster needs hacking. I think viral made it so that debconf interface will modify a conffile. It won't overwrite it completely, I hope, but the policy forbids any modification by maintainer scripts. Also, policy states that maintainer script must not modify any other package's conffile, so update-cluster-regenerate must not be called in postinst.
Debian potato install disk did not recognize Planex ENW-3602-T/TX/10/100 Fast Ethernet CardBus PC Card This is sad. This notebook PC has an ESS Solo AudioDrive (WDM) listed as a soundcard. I do not have a clue what this may be... But reading /proc/pci it seems to be a ESS Solo1 chip, with PCI ID 125D:1969
Back from vacation, but my machine is not up yet. update-cluster has been NMU'd several times, as it seems. It sounds good. There are problems there...
I saw version := $(shell expr $(CURDIR) : '.*-\([0-9.]*\)') in Debian-devel, and I really think this is weird... In fact, I didn't know about the existence of expr command. Fun.
New month starts. I will be having a seminar-vacation tour tomorrow. Dual-display system is looking nice, but not doing very much.
starting a new month's diary involves
The nearby guy is installing Debian GNU/Linux (potato) on his laptop machine. Corega PCMCIA card, PCC-TXD is apparently working fine.
$Id: 200108.html.en,v 1.45 2005/08/28 06:45:35 dancer Exp $