invisible console (solved)

[Jan Willem Stumpel]

Sorry. Missed a previous thread. Cured it by setting

  options radeon modeset=0

in /etc/modprobe.d/radeon-kms.conf

This works (original value was 1).

Regards, Jan


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/4bdac862.5070...@my.home

invisible console

[Jan Willem Stumpel]

I like an old-fashioned VGA text look on the console; in
/etc/default/console-setup I have

FONTFACE="VGA"
FONTSIZE="16"

This used to work.

But after a dist-upgrade around April 10th, the behaviour changed.

1 -- cold startup works normally (and looks normal).
2 -- startx works OK
3 -- pressing control-alt-f2 (to go to the console temporarily)
now gives a (framebuffer?) screen with very small letters; at the
top of the screen there is a message from "drm" stating that the
resolution is set to 1250 x 1024 (instead of 640 x 480 which is
what I want).
4 -- presing alt-f7 gets me back into X.
5 -- pressing control-alt-f2 AGAIN gives a completely black
screen. Nothing visible. Going back to X with alt-f7 is still
possible.

uname -a gives:
Linux vega 2.6.32-3-686 #1 SMP Thu Feb 25 06:14:20 UTC 2010 i686
GNU/Linux

This kernel was installed recently; booting an earlier kernel,
e.g. 2.6.29-2, gets the old (proper) behaviour back.

Configuration error? Or bug? If the latter, which package?

Regards, Jan


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/4bdab531.4060...@my.home

Re: Re: Vuze (Azureus) broken after Apt-get upgrade

[Jan Willem Stumpel]

Boyd Stephen Smith Jr. wrote:

> Probably related to
>  
> there's a supposed work-around in Message-Id: 
> <200912151551.20419.tim.rueh...@gmx.de>, which should be in the
> archives.  I did not test the work-around because I do not use
> any Java applications regularly.

Thanks very much! I had the same problem and (because I had been
messing with my shorewall settings recently) thought it was my
fault. Your message prevented me from pulling my hair out.

Regards, Jan


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

xsession errors

[Jan Willem Stumpel]

On my machine the file ~/.xsession-errors grows and grows without
limits. I have to remove it regularly before it takes over my
whole disk; this has been the case for years. Last time I removed
it was September 21; now it is again already 135M in size! It
grows by the minute. It is mainly full of "GDK-Warnings".

Did anyone else have this problem, either now or in the past? If
in the past, what did you do to solve it?

Regards, Jan


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

install problem

[Jan Willem Stumpel]

This is really an embarrassing question for an old Debian hand to
ask, but how do I install Debian?

I just bought a "netbook" which has no CDROM drive, but which can
boot from a USB stick. I could dd an Ubuntu image to the stick and
then boot from it. But I prefer just plain old Debian.

I found (through the Debian home page) an image called
debian-503-i386-netinst.iso. I dd'd it to the stick. But the
netbook does not boot from it.

There must be something very elementary which I did wrong. But what?

Regards, Jan


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Re: Silverlight/Moonlight/Tuva

[Jan Willem Stumpel]

David Fox wrote:

> I didn't get it to work. I reloaded firefox, brought up the
> site again, same thing. It thinks the silverlight / moonlight
> isn't loaded, although the clicktrhough site says it is. about:
> plugins in firefox says it is too.
> 
> The abc.com videos play normally; no mention is made that
> silverlight is a requirement.

Through Novell's bugzilla page I found another page which, just
like Microsoft's Tuva page, does not work: http://matosdotnet.com/

It complains about silverlight/moonlight not being installed,
while in fact it is. But apparently, what is missing is
silverlight "version 2", while Debian only has "version 1".

What is especially galling about this situation is that the
Richard Feynman lectures on "The Character of Physical Law" are
now only available on Tuva. They've disappeared from Youtube. Such
things could make a person paranoid.

Regards, Jan


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Silverlight/Moonlight/Tuva

[Jan Willem Stumpel]

Has anyone succeeded in using the "moonlight" packages in order to
view the videos at Microsoft's Tuva project? If so, which packages
are needed exactly / which tricks?

http://research.microsoft.com/apps/tools/tuva/

Regards, Jan


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Re: iceweasel puzzle (SOLVED)

[Jan Willem Stumpel]

Andrew Sackville-West wrote:

> export AWT_TOOLKIT=MToolkit
> 
> in .xinitrc. The reference to libawt.so makes me think this may
> be a similar problem.

Well -- this did not work. I finally decided on a new, especially
thorough, purging of anything having to do with java,
web-browsers, macromedia, and flash. Lots of times doing
"updatedb" and "locate"; removing many dot-files and
dot-directories in /etc and in my home directory.

Then installed (after saving bookmarks & passwords)

 - iceweasel
 - vuze (which pulls in lots of openjdk-6 stuff)
 - icedtea6-plugin (browser plugin for openjdk-6)
 - flashplugin-nonfree

All sun-java stuff had been purged (AFAIK).

And now it works. I mean both applets and the strictlysudoku.com
login page work.The unsatisfactory thing is, I still don't know
which of the countless files I removed was the real culprit. But
it works now. Thanks all who thought about this.

Regards, Jan.


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Re: iceweasel puzzle

[Jan Willem Stumpel]

Hugo Vanwoerkom wrote:

> With iceweasel 3.0.6-1 works as expected.
> 
> You don't have Noscript enabled do you?

Java and Javascript are enabled and both work.
I tried moving the .mozilla directory out of the way; didn't help.
Thought it might have something to do with the flash plugin
(because locate strictlysudoku turned up some entries in
~/.macromedia), reinstalled and removed flash in various ways--
nothing helps.

The login popup window is blank, not only in iceweasel on my
computer, but also in Opera! So it is not an iceweasel puzzle only.

The login popup window is loaded somehow; I can read the source
code. I can save it as a file, and open the file with iceweasel or
opera: it is blank. Opera has two complaints (may or may not be
related):

   ERROR: ld.so: object 'libjvm.so' from LD_PRELOAD cannot be
   preloaded: ignored.
   ERROR: ld.so: object 'libawt.so' from LD_PRELOAD cannot be
   preloaded: ignored.

This might be significant, although I have not a clue what it
means. There are several files called libjvm.so and libawt.so on
my system (installed by openoffice, sun jre, and openjdk6;
"client" and "server" versions). But this sudoku site does not use
java, only javascript. As long as you just want to play sudoku
(with logging in) you don't even need javascript.

In lynx, the saved login form is displayed (but of course in lynx,
I cannot play sudoku, because this a graphical thing).

N.B. ~/.xsession-errors keeps getting filled up with (to me)
incomprehensible warnings and errors, like:

(firefox-bin:5229): GLib-GObject-WARNING **: instance of invalid
non-instantiatable type `(null)'

(firefox-bin:5229): GLib-GObject-CRITICAL **:
g_signal_handler_disconnect: assertion `G_TYPE_CHECK_INSTANCE
(instance)' failed

(firefox-bin:5229): GLib-GObject-WARNING **:
/build/buildd-glib2.0_2.20.3-1-i386-o59wJY/glib2.0-2.20.3/gobject/gsignal.c:2387:
instance `0x90ecda8' has no handler with id `4276'

(firefox-bin:5229): GLib-GObject-WARNING **:
/build/buildd-glib2.0_2.20.3-1-i386-o59wJY/glib2.0-2.20.3/gobject/gsignal.c:2387:
instance `0x97d1908' has no handler with id `4536'

(firefox-bin:5229): GLib-GObject-WARNING **: invalid
uninstantiatable type `(null)' in cast to `GtkObject'

(firefox-bin:5229): GLib-GObject-WARNING **: instance of invalid
non-instantiatable type `(null)'

So, if someone has any bright idea.. in the meantime I keep digging.

Regards, Jan


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

iceweasel puzzle

[Jan Willem Stumpel]

I use iceweasel to visit the site http://www.strictlysudoku.com.
In the left-hand column there are ĺogin' and 'register' links. If
I click one of those, a window pops up, but it is completely blank.

On my wife's computer, which uses the same gateway/firewall, and
which is also more or less Debian (well, ubuntu), clicking the
links produces proper 'login' and 'register' forms.

Any idea of what could be the matter on my computer?

Regards, Jan


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Re: java applets

[Jan Willem Stumpel]

Frank Lin PIAT wrote:
> On Sun, 2009-06-28 at 14:18 +0200, Jan Willem Stumpel wrote:

> If everything else fails:
> - Purge all JRE
>   See:   aptitude search "?and(~Pjava-runtime,~i)"
> - Purge iceweasel
> - Reinstall iceweasel
> - Reinstall sun-java6-plugin
> - Create a new user account
> - Test the JRE on http://java.com/en/download/help/testvm.xml

All else had already failed .. but this worked. I lost some prefs
and my bookmarks, but could restore them from backup.

Thanks very much!



-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Re: java applets

[Jan Willem Stumpel]

Frank Lin PIAT wrote:

> Can you list the plugins in iceweasel, typically, with the _one_line_
> command:
> 
> find /usr/lib/iceweasel/plugins/  -name *.so \
>   | xargs -n 1 readlink -m | xargs dpkg -S
> 
This gives:
sun-java5-bin:
/usr/lib/jvm/java-1.5.0-sun-1.5.0.18/jre/plugin/i386/ns7/libjavaplugin_oji.so
realplayer: /usr/lib/mozilla/plugins/nphelix.so
mozilla-acroread: /usr/lib/Adobe/Reader9/Browser/intellinux/nppdf.so

about:plugins in iceweasel, on the other hand, says I have plugins for

Shockwave Flash
Adobe Reader 9.1
Helix DNA Plugin: RealPlayer G2 Plug-In Compatible

So neither Icedtea (for openjdk-6) nor the Sun plugin are
available; and so, of course, applets do not work. It has been
like this for ages now; am getting a bit desperate. Whatever I do,
I cannot get a working Java plugin back.


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Re: Unicode recommendations in a terminal - Please advise.

[Jan Willem Stumpel]

Chris Jones wrote:

> How should I set up my terminal - xterm - and what might be a 
> good unicode fixed font that could make the experience more
> true to life?

The Debian package xfonts-efont-unicode has very wide character
coverage (incl. Chinese/Japanese). For tips on how to set it up,
google my page "Configuring xterm for UTF-8".

> Another thing is that I would like to be able to display all
> the glyphs of a given font so I can tell at a glance what
> scripts are covered.

I use xfd:

xfd -fn '-efont-biwidth-medium-r-*-*-16-*-*-*-*-*-iso10646-*'

> xfd from the x11-utils package is fine, but it does not appear 
> to support Truetype fonts.

The efonts aren't Truetype. You get the oldfashioned xterm "look &
feel".


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

OSS4

[Jan Willem Stumpel]

I just read this:

http://insanecoding.blogspot.com/2009/06/state-of-sound-in-linux-not-so-sorry.html

AFAICU the article recommends throwing out, among other things,
ALSA and pulseaudio, which it has harsh words for, and using OSS
version 4 instead.

Switching everything to OSS4 with its (alleged) built-in sound
mixing seems very attractive.

Does Debian support OSS Version 4 now? Is there a decent Debian
sound tutorial somewhere? Using Sid.

Regards, Jan



-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Re: Re: java

[Jan Willem Stumpel]

> If you're running 32-bit, you'll probably just want the
> sun-java6-plugin.

I am running 32-bit. But if I kill openjdk-6-jre, I'd lose
Azureus/Vuze, which I use to download Korean soap operas. My wife,
who is addicted to the things, would kill me if I did.

Regards, Jan



-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Re: java applets

[Jan Willem Stumpel]

Frank Lin PIAT wrote:

> Humm, installing sun-java6-plugin (or sun-java5-plugin or 
> icedtea-gcjwebplugin) should pull all the required
> dependencies.

Yess.. but that means I'll have at least two jre's: sun-java6 and
the one required by Azureus/Vuze. How can I be sure that they
won't "bite" each other? Now, with only the openjdk-6-jre
installed, I don't see any applets working, but at least Iceweasel
does not crash/become unresponsive, as it did invariably on
applet-containing pages twhen there were 2 or more jre's.

>> Does anybody know of a decent tutorial for setting up jre,
>> any jre, on Debian?

> Hopefully, "aptitude install sun-java6-plugin" should be
> enough.

But I need the openjdk-6-jre as well because of Azureus/Vuze.

> "doesn't work" is a bit short.

In this case it means: after restarting the browser, the plugin
does not show in about:plugins. And applets do not work; I get an
invitation to click something to download the plugin, but then it
says that there is no appropriate plugin available.

I know that, running Sid, I cannot expect the Moon. But I've had
this problem for 6 months or so now. And I haven't a clue which
package I should file a bug against. This is typically one of
those "inter-package" bugs for which the BTS does not seem to be
very suitable. Perhaps we need a "java task force" (as well as a
"sound task force").

Regards, Jan


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

java

[Jan Willem Stumpel]

I just want to install a Java runtime environment which will allow
me to see Java applets in action. It seems I have the choice of
(at least)

default-jre
gcj-4.4-jre
gcj-jre
icedtea-6-jre-cacao
openoffice.org (includes its own jre, apparently)
openjdk-6-jre (required by Azureus/Vuze; it does not seem to
accept other Javas)
sun-java5-jre
sun-java6-jre

Today I did another attempt: rigorously dpkg --purge'd any trace
of any other jre than openjdk-6-jre. It does not have a mozilla
plugin, but "suggests" installing icedtea6-plugin and
sun-java6-fonts. But then sun-java6-fonts wants to install
sun-java6-bin and sun-java6-jre, and "suggests" sun-java6-plugin
and ia32-sun-java6-plugin. And the icedtea6-plugin does not work...

This is dependency hell. Does anybody know of a decent tutorial
for setting up jre, any jre, on Debian? With only one boundary
condition: it should work. I haven't been able to see applets
working for about half a year now.

Regards, Jan


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Re: Ctrl-Alt-Backspace disabled?

[Jan Willem Stumpel]

emikaadeo wrote:

> I just enabled keyboard layout in 
> K>Systemsettings>Regional&Language>Keyboard Layout, and there,
> in advanced tab I check: Key sequence to kill X server
> (Control+Alt+Backspace).

I filed a bug against xserver-xorg about the disappearance of
control-alt-backspace, and the maintainer (Julien Cristau) wrote
back saying

  DontZap will be changed back to off by default soon.

He also gave the tip to re-enable control-alt-backspace by
including in the XKBOPTIONS line in /etc/default/console-setup:

  terminate:ctrl_alt_bksp

After which (to be on the safe side) you have to reboot.

I suppose this will be necessary (for those without KDE/Gnome)
until "DontZap = off" becomes the default again.

Regards, Jan


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Re: Ctrl-Alt-Backspace disabled?

[Jan Willem Stumpel]

widux wrote:

> Hi,
> 
> for me
>   right-Alt + Print + k 
> does the trick!
> 
> Greetings

How on earth did you find this out?

Also, do you know where this is set, so I could change it? Or is
it hard-coded somehow? Anyway, thanks for the information.

Regards, Jan


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Re: Re: Ctrl-Alt-Backspace disabled?

[Jan Willem Stumpel]

Andrei Popescu wrote: (about getting ctl-alt-backspace back):
> You need
> 
> Option"DontZap" "false"
> 
> in the ServerFlags section of your xorg.conf (check the manpage for 
> xorg.conf, I'm writing from memory).

This worked for about a week -- but nannyism has crept further. In
the latest Sid, even this trick doesn't work anymore. How to get
it back, I wonder.


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

replacing hard disk

[Jan Willem Stumpel]

My wife's computer which runs ubuntu "heron" made a funny sound
today. I now think it must have come from the hard disk, which is
quite old; the motherboard and cpu are fairly new, however. And
just now the computer became unusable because it could no longer
save any files. /var/log/syslog told about a disk error which
resulted in the disk being remounted read-only. Restarting the
computer caused an fsck, and now the machine works again, but it
is clear that the disk is nearing the end of its life.

So tomorrow it's off to the computer shop to get a new hard disk.

Are there any tips on moving the whole system from the old disk to
the new one? Or do I just have to re-install ubuntu, re-install
any updates and extra programs which are installed, find and copy
modified config files, mails, bookmarks, etc?

Regards, Jan


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Re: Ctrl-Alt-Backspace disabled?

[Jan Willem Stumpel]

IMHO disabling an old and trusted functionality is simply
introducing a bug, made worse by keeping silent about it (no word
about it in changelog.gz or NEWS.Debian.gz).

It must surely be a tiny minority of users who press
control-alt-backspace "by mistake"; I find it hard to imagine
even.  But for users who suffer from this syndrome there has
always been the possibility of specifying DontZap. Forcing it on
all users is a Bad Thing.

Regards, Jan


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Re: What hardware to use for Debian Firewall/Gateway or server?

[Jan Willem Stumpel]

Csanyi Pal wrote:

> Yes, with Debian Etch, but not with Debian Lenny! 
> So: can one install on it say a Debian GNU/Linux Lenny? 

I suppose you can, although I didn't try it. I think you can
install anything you want; I heavily customised mine (in fact I am
no longer using any of the Bubba-specific software provided by
Excito). Apt-get works; upgrading to Lenny is just a matter of
pointing your sources.list towards it, I think. The kernel is
2.6.26.5, so it should not be a problem (don't quote me on that,
though; I also do not quite understand why you should want to use
Lenny, rather than Etch, on a server. First prepare a "rescue
stick" before starting any experiments!).

Mind that it is a "headless" device. Everything has to be done
through ssh (or local telnet). It has no cd-rom drive, keyboard,
or monitor. But it is just a Debian system (for powerpc, not for
i386). Everything behaves just like your desktop Debian system.

Don't expect gigantic calculating power from this machine. It is
good for shifting bits around (what a home server/gateway should
do) but not for anything calculation-intensive. I mean, the low
power consumption and silent operation had to come from somewhere.
Absolutely fine though, as a server/firewall/gateway.

Regards, Jan



-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Re: What hardware to use for Debian Firewall/Gateway or server?

[Jan Willem Stumpel]

Csanyi Pal wrote:

> What is the recommended new hardware for firewall/gateway or
> for a web, mail, file & printer server at a small home network?
> 
> Any advices will be appreciated!

I am now using a "Bubba 2", made by a Swedish company:

http://excito.com/bubba/products/about-bubba.html

It runs Debian. More expensive, of course, than using an old
desktop or laptop computer (but the price is going down all the
time, now 212 euros for a unit with 500 GB hard disk), but it uses
"almost no" electricity, and it is silent (fanless). Very suitable
for 24/7 operation. I am very happy with it.

Regards, Jan




-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Re: Alt-F7 fails

[Jan Willem Stumpel]

Andrei Popescu wrote:

> You could add
> 
> http://wiki.debian.org/XStrikeForce/InputHotplugGuide
> 
> to the 'Links' section.

Did that. Regards, Jan


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Re: Alt-F7 fails

[Jan Willem Stumpel]

OK, I made a small write-up now (aimed at users like myself) about
the ‘New Input System’ in Debian. Comments welcome --

  http://www.jw-stumpel.nl/stestu.html#T6.3

Regards, Jan



-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Re: Alt-F7 fails

[Jan Willem Stumpel]

Steve Kemp wrote:
>   You can dump a skeleton xorg.conf file by running, as root:
> 
> dexconf

Yes, this gives an absolutely bare-bones xorg.conf with one
"section"  in it:

Section "Device"
Identifier  "Configured Video Device"
EndSection

It works; and in fact, even without any xorg.conf, it works also.

I understand now that the goal of the recent changes was to get
rid of xorg.conf, because X should get the information it needs
from elsewhere, and automatically.

So far so good, but.. with this "skeleton" xorg.conf,
control-alt-backspace for getting out of X no longer works. What I
did was to insert, as the first section in xorg.conf, a
"ServerFlags" section:

Section "ServerFlags"
Option "DontZap""off"
EndSection

By this means I got control-alt-backspace back. Or is there also a
non-xorg.conf way of getting the same thing?

Regards, Jan


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Re: Alt-F7 fails

[Jan Willem Stumpel]

Florian Kulzer wrote:

> X crashing when switching back from the console is most
> commonly caused by a problem with the video driver. There
> should be some related error messages in /var/log/Xorg.0.log or
> in /var/log/syslog. (Switching back and forth between X and
> console works fine for me; up-to-date Sid/amd64, intel driver.)

There was a "backtrace" in /var/log/Xorg.0.log:

Backtrace:
0: /usr/bin/X11/X(xorg_backtrace+0x3b) [0x813265b]
1: /usr/bin/X11/X(xf86SigHandler+0x51) [0x80c5c71]
2: [0xb7fd3400]
3: /usr/lib/xorg/modules/drivers//radeon_drv.so [0xb7956417]
4: /usr/bin/X11/X(xf86CrtcSetModeTransform+0x4b4) [0x80eee14]
5: /usr/bin/X11/X(xf86SetDesiredModes+0x124) [0x80ef1e4]
6: /usr/lib/xorg/modules/drivers//radeon_drv.so [0xb793527d]
7: /usr/lib/xorg/modules//libxaa.so [0xb7706f08]
8: /usr/bin/X11/X [0x80dd9e1]
9: /usr/bin/X11/X [0x80ccdf4]
10: /usr/lib/xorg/modules/extensions//libglx.so [0xb79f0864]
11: /usr/bin/X11/X(xf86Wakeup+0x3c3) [0x80c6393]
12: /usr/bin/X11/X(WakeupHandler+0x52) [0x8090512]
13: /usr/bin/X11/X(WaitForSomething+0x1bb) [0x812ffeb]
14: /usr/bin/X11/X(Dispatch+0x7e) [0x808c61e]
15: /usr/bin/X11/X(main+0x3bd) [0x8071a5d]
16: /lib/i686/cmov/libc.so.6(__libc_start_main+0xe5) [0xb7c72775]
17: /usr/bin/X11/X [0x8070f11]

Fatal server error:
Caught signal 11.  Server aborting

It seems quite a lot happened to Debian Sid recently without me
being aware of it.

I did 2 experiments:

1-Deleted (well, renamed of course) /etc/X11/xorg.conf. Made sure
  that the keyboard description in /etc/defaults/console-setup
  was OK.
  Result: X works, OK it seems, without xorg.conf, and Alt-F7
  works also!

2-Feeling a little bit guilty about running X without xorg.conf,
  ran dpkg-reconfigure xserver-xorg. I expected to get the well-
  know litany of questions. No such thing. Instead I got:
  dpkg: warning: obsolete option
'--print-installation-architecture', please use
'--print-architecture' instead.
  (repeated 3 times) Now what does this mean? Where is this
  "obsolete option" set?

Should users now just simply ditch /etc/X11/xorg.conf? On the
other hand it says  in the "Input Hotplug Guide":

  First and foremost is that the xorg.conf is still absolutely
  necessary to set one's keymap when you're not using the default
  'us' map.

But it seems to work OK even with non-standard keyboard setup,
provided it is set in console-setup.

Confused, Jan








-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Alt-F7 fails

[Jan Willem Stumpel]

This may be a FAQ; if so, apologies.

On my Sid, I can go to the console from X by means of, e.g.,
Control-Alt-F2. This has been the behaviour of X for ages.

Also for ages, you could go back to X by pressing Alt-F7. However,
in Sid nowadays, this does not work. Pressing Alt-F7 simply kills
X, and all programs running in it. You stay in the console, and
have to restart X by means of startx.

A bug, obviously. But in which package? Has it been reported to
the Debian BTS? I couldn't find it.

Regards, Jan


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Re: Re: keyboard mapping problem under X

[Jan Willem Stumpel]

Andrei Popescu wrote:
> changing /etc/default/console-setup should be enough, xorg.conf
> is ignored.

Aah.. That explains it.

> Restart hal, at least in theory.

Yes, in theory. In practice it may be different. I restarted it by
means of /etc/init.d/hal restart, and got some weird results.

It is also possible to make X ignore evdev altogether by setting
in the ServerFlags section of xorg.conf:

  Option "AutoAddDevices" "off"

Then you get the old behaviour back. xorg.conf again determines
what is going on. Of course without the benefits of evdev. But I
do not know what these benefits are anyway, so I do not miss them!

Regards, Jan


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Re: keyboard mapping problem under X

[Jan Willem Stumpel]

Bruno Boettcher wrote:

> kbd is still on nodeadkeys...

It appears that the keyboard system in X has changed. There is now
something new and mysterious called evdev. To see if you have a
system with evdev, type

setxkbmap -print

and see if evdev is mentioned.

If it is, it seems that at the moment, to change the behaviour of
the keyboard, you must specify things not only in
etc/X11/xorg.conf, but *also* in /etc/default/console-setup. And
then you must reboot! Stopping and restarting X is NOT enough.
(There must be a better way though; I am going to investigate
this, because obviously my "international keyboards and fonts"
page needs serious updating.)

Regards, Jan





-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Re: keyboard mapping problem under X

[Jan Willem Stumpel]

Bruno Boettcher wrote:
> Hello [..]

> i use, whilst living in France, a german keyboard...

You take the trouble to mention this: any special reason? For
instance, do you want to type mostly French on a German keyboard?

> still, dead-keys aren't working

> so, what possibilities remain to activate them? without
> redefining a whole new keymapping, especially since the system
> seems to have changed?

KDE or Gnome have GUI utilities that will make the keyboard do
what you want (i.e. without directly editing system files).
Probably in something called "System, Keyboard".

If (like me) you want to avoid using KDE/Gnome, you can *also*
make the keyboard do what you want. The German keyboard is called
"de". It has several so-called "variants":

 -basic (this is the default, which has several dead keys)
 -nodeadkeys (no keys are dead)
 -deadgraveacute (only grave and acute are dead)
 -deadacute (only acute is dead)
 -ro (includes special characters for Romanian. I don't know why
  this is included in a German keyboard description; maybe
  German keyboards are commonly used in Romania)
 -ro_nodeadkeys
 -... etc, etc, ...

See the file /usr/share/X11/xkb/symbols/de. The "variants" are
described in sections beginning with xkb_symbols.

You can select whatever variant you want by typing in a terminal
(e.g. xterm) window:

  setxkbmap de [-variant ]

E.g.

  setxkbmap de -variant deadgraveacute

If you just want to use the "basic" variant, you just type

  setxkbmap de

You can use the setxkbmap command to experiment with keyboard
layouts and variants. Once you have selected what you want, you
can put it in /etc/X11/xorg.conf, in the keyboard section, in a
slightly different format, using one line for the keyboard and
another for the variant, e.g.:

  Option "XkbLayout" "de"
  Option "XkbVariant" "deadgraveacute"

This will make your selection (semi-)permanent.

There are many more possibilities for handling the keyboard in X
(not so many on the console). For instance, you can switch between
entirely different keyboard layouts by defining a special key. For
details, see below, sections 6.1 and 6.2.

Regards, Jan
http://www.jw-stumpel.nl/stestu.html


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Re: mplayer slowness and kernel

[Jan Willem Stumpel]

Florian Kulzer wrote:
> 
> It seems that your xorg video driver has problems to use
> hardware video acceleration with the newer kernel. (A problem
> with DMA for hard drive or DVD access is also possible, but
> less likely, IMHO.)
> 
> What do you get from:
> 
> lspci -nn | grep -Ei 'vga|display|video|media'

00:1f.5 Multimedia audio controller [0401]: Intel Corporation
82801EB/ER (ICH5/ICH5R) AC'97 Audio Controller [8086:24d5] (rev 02)
01:00.0 VGA compatible controller [0300]: ATI Technologies Inc
RV280 [Radeon 9200 SE] [1002:5964] (rev 01)
01:00.1 Display controller [0380]: ATI Technologies Inc RV280
[Radeon 9200 SE] (Secondary) [1002:5d44] (rev 01)

> grep -Ei '^\((ee|ww)\)|xv|/drivers/' /var/log/Xorg.0.log

With 2.6.26 kernel:
(WW) `fonts.dir' not found (or not valid) in
"/usr/X11R6/lib/X11/fonts/misc".
(WW) The directory "/usr/local/share/dosemu/Xfonts" does not exist.
(WW) The directory "/usr/share/fonts/X11/cyrillic" does not exist.
(WW) AllowEmptyInput is on, devices using drivers 'kbd', 'mouse'
or 'vmmouse' will be disabled.
(WW) Disabling Generic Keyboard
(WW) Disabling Configured Mouse
(WW) Open ACPI failed (/var/run/acpid.socket) (No such file or
directory)
(II) Loading extension XVideo
(II) Loading extension XVideo-MotionCompensation
(WW) Warning, couldn't open module freetype
(EE) Failed to load module "freetype" (module does not exist, 0)
(WW) Warning, couldn't open module type1
(EE) Failed to load module "type1" (module does not exist, 0)
(II) Loading /usr/lib/xorg/modules/drivers//ati_drv.so
(II) Loading /usr/lib/xorg/modules/drivers//radeon_drv.so
(II) RADEON(0): Will try to use DMA for Xv image transfers
(WW) RADEON(0): DRI init changed memory map, adjusting ...
(WW) RADEON(0):   MC_FB_LOCATION  was: 0xf7fff000 is: 0xf7fff000
(WW) RADEON(0):   MC_AGP_LOCATION was: 0xffc0 is: 0xfc7ffc00
(WW) RADEON(0): Option "XaaNoOffscreenPixmap" is not used

With 2.6.29 kernel:

(WW) `fonts.dir' not found (or not valid) in
"/usr/X11R6/lib/X11/fonts/misc".
(WW) The directory "/usr/local/share/dosemu/Xfonts" does not exist.
(WW) The directory "/usr/share/fonts/X11/cyrillic" does not exist.
(WW) AllowEmptyInput is on, devices using drivers 'kbd', 'mouse'
or 'vmmouse' wi
ll be disabled.
(WW) Disabling Generic Keyboard
(WW) Disabling Configured Mouse
(WW) Open ACPI failed (/var/run/acpid.socket) (No such file or
directory)
(II) Loading extension XVideo
(II) Loading extension XVideo-MotionCompensation
(WW) Warning, couldn't open module freetype
(EE) Failed to load module "freetype" (module does not exist, 0)
(WW) Warning, couldn't open module type1
(EE) Failed to load module "type1" (module does not exist, 0)
(II) Loading /usr/lib/xorg/modules/drivers//ati_drv.so
(II) Loading /usr/lib/xorg/modules/drivers//radeon_drv.so
(II) RADEON(0): Will try to use DMA for Xv image transfers
(WW) RADEON(0): Direct rendering disabled
(WW) RADEON(0): Option "XaaNoOffscreenPixmap" is not used

Thanks for looking into this. Regards, Jan


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Re: Lenny. Locales and Interface Language

[Jan Willem Stumpel]

Mark Goldshtein wrote:

> 1. Is a default system locale independent from an interface
> language? I mean, is it possible to change a default system
> locale to whatever I like and there will be no harm to English
> interface I have? [..]

It depends on what you mean by "changing the locale".

AFAIK the "locale" *mainly* determines the user interface.
Messages and prompts from programs, date and currency formats,
paper size (eg "Letter" for US locales), number format (e.g.
thousands separator is comma in some locales, period in others), etc.

Note that messages and prompts from programs for various locales
have to be provided by the programs themselves. If they are not,
messages, etc., will be provided in some default language
(invariably English). Also, it is up to the programs to pay
attention to the "locale"; if they do not, they will generate e.g.
numbers and dates according to the US convention.

So if you want to keep an English user interface, you have to keep
an English locale; but if you also want to read, input, and print
things in other languages than English, you should make sure that
your locale is a UTF-8 one, so you can handle all kinds of
languages and scripts. Examples: en_US.utf8, en_GB.utf8.

Regards, Jan
http://www.jw-stumpel.nl/stestu.html


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

mplayer slowness and kernel

[Jan Willem Stumpel]

With the 2.6.29 kernel (from the linux-image-2.6.29-2-686
package), playing a movie with mplayer uses up to 95% cpu. mplayer
becomes very slow and jerky, with "stuttering" sound, and "Your
system is TOO SLOW" warnings. No twiddling of mplayer's parameters
helps.

When I go back to the slightly older 2.6.26 kernel
(linux-image-2.6.26-2-686), mplayer works properly again (about
20% cpu usage and no warnings or "stuttering", even at full screen
size.

Is this a known problem?

Regards, Jan



-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Re: Bad sound

[Jan Willem Stumpel]

Andrei Popescu wrote:

>> I may be a bit behind the times in using xmms. Is there a
>> replacement (with approximately the same user interface) for xmms?
> 
> audacious

Audacious works very well. Thanks, Jan


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Re: Bad sound

[Jan Willem Stumpel]

Kent West wrote:
> Jan Willem Stumpel wrote:
>> The last few days, sound (in particular music) is very bad on
>> my Sid system. It sounds distorted, as if above a certain
>> amplitude the signal is clipped.

> I had this problem for about the last year; last week I
> discovered one of the 3D channels was too high.

How did you discover this? I mean by which program: alsamixer,
alsactl, or something else?

In the meantime I found that the problem only occurs when xmms is
used for playing music. xmms used to work very well until a few
days ago, but now it sounds horrible. Using mplayer instead makes
the same music files sound perfect again.

I may be a bit behind the times in using xmms. Is there a
replacement (with approximately the same user interface) for xmms?
Does anyone know why xmms suddenly started to behave so badly?

Regards, Jan


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Bad sound

[Jan Willem Stumpel]

The last few days, sound (in particular music) is very bad on my
Sid system. It sounds distorted, as if above a certain amplitude
the signal is clipped.

Could this possibly be caused by software, i.e. a Sid upgrade? Did
anyone else notice this?

00:1f.5 Multimedia audio controller: Intel Corporation 82801EB/ER
(ICH5/ICH5R) AC'97 Audio Controller (rev 02)

Regards, Jan


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Re: Brother HL2040 printer/lpr stopped working after update

[Jan Willem Stumpel]

beatrice wrote:

> I think I'll write a couple lines to Brother's customer service
> anyway, now that I know I am not the only one with this
> problem.

I don't know if the error is with Brother or with ghostscript. In
the meantime I filed a bug against ghostscript:
  http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=522417

Actually I *think* that the error is with ghostscript. The Brother
scripts look OK to me.

I hope that 8.62 works for you; didn't try it myself. And indeed,
it happened before, so I stopped upgrading ghostscript for a
while, I remember now. I am not sure which previous version did
not work. Hope it was not 8.62..

Regards, Jan


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Re: Brother HL2040 printer/lpr stopped working after update

[Jan Willem Stumpel]

The problem is, as you suspected, caused by ghostscript. I
downgraded to version 8.63 and it started to work again. You have
to downgrade 3 packages: libgs8, ghostscript, and ghostscript-x.
With any luck, the 8.63 versions of these packages are still in
your /var/cache/apt/archives, and you can install them with dpkg
-i. Otherwise you must go to packages.debian.org and get the 8.62
versions from "stable"; it seems that 8.63 is no longer available.

Version 8.64 gives an "irrecoverable error". I haven't found out
yet how Brother's way of using ghostscript triggers this error.

Regards, Jan



-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Re: Brother HL2040 printer/lpr stopped working after update

[Jan Willem Stumpel]

beatrice wrote:
> I have a Brother HL2040 printer that worked perfectly well on
> my Debian testing with the .deb driver package Brother
> provides. I don't have CUPS installed, I use lpr. My printer
> stopped working without me changing any system configuration
> other than my usual package updating via aptitude.
[..]
> The communication PC-> printer should be working. When a job is
> sent to the printer the printer "wakes up", makes all the "hot
> air" noise it usually does, plus the led light blink as it
> always did while it is receiving data. The length of the
> blinking still varies with the amount of data sent.

I am afraid I can't help you now, but I have almost the same
problem. The only difference is that my printer is a Brother
HL2030 instead of a 2040. The rest is exactly as you describe.
lpr, and the problem started yesterday after an upgrade.

I'll let you and list know when I find a solution. I think it
happened before (maybe a year ago) but I can't remember what I did
to fix it.

Regards, Jan


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Grr! Sound problems again

[Jan Willem Stumpel]

About a year and a half ago, after a lot of experimentation, I
finally got sound working OK on my Debian Sid system. The solution
turned out to be quite simple. To save others the same trouble, I
put up a kind of "sound instruction" on my website
(http://www.jw-stumpel.nl/dmix.html).

But something must have changed in Sid recently, because sound has
become unreliable again. After I visit a sound-using site like
youtube (let's call this "online sound"), sound will no longer be
heard when playing a stored avi file with gmplayer, for instance
("offline sound"). To play offline sounds I have to kill iceweasel
first (or call /etc/init.d/alsasound restart, which kills
iceweasel as a side effect..).

Well, I know this is Sid, which has this tendency of breaking
things which work by "fixing" them. But normally these things get
"re-fixed" after a few days. This new sound problem has now been
going on for a few weeks. I cannot file a bug report about it
because I have no way to know which program causes the problem.
Any help appreciated.

Regards, Jan


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Re: Strange "news" problem

[Jan Willem Stumpel]

Sjoerd Hiemstra wrote:

> There are reports complaining about this in the
> nl.internet.providers newsgroup, w.r.t. news servers of your
> provider. nova.planet.nl still seems to work well.

Thanks very much! KPN's "help" desk is apparently not aware of
this problem. I thought the nova server was only for "binaries",
but apparently it carries "normal" newsgroups as well.

Other list members: sorry to have bothered you with a problem that
has nothing to do with Debian.

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Strange "news" problem

[Jan Willem Stumpel]

Eduardo M Kalinowski wrote:

> Try with another profile (icedove -ProfileManager should let
> you create another, or move the .mozilla/thunderbird/ directory
> out of the way temporarily) to see if the problem still
> persists with a clean profile.

I tried both suggestions, but I still cannot see any news after
Nov. 30. Most mysterious.

(Hope this message make it to the list; I am not subscribed.
Normally I send messages to the list with a method involving both
the html archives and the linux.debian.user newsgroup, but the
latter does not work now :-(

May try a purge and reinstall of icedove and iceweasel.

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Strange "news" problem

[Jan Willem Stumpel]

Icedove, which I use for mail and news, does not download headers
from subscribed newsgroups anymore. On the 29th of November it
still worked fine; I received some message headers (only a
handful) dated November 30; nothing since then.

The news server belongs to my ISP, but they say nothing has
changed and they got no other complaints about this. In fact when
I try another newsreader (pan), new message headers continue to be
retrieved normally.

So it really seems to be an Icedove problem. Even deleting the
news account and setting it up afresh does not produce any
messages after November 30.

I update/upgrade Sid regularly, but Icedove has not been upgraded
since October.

Any suggestions?

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: java plugin (SOLVED)

[Jan Willem Stumpel]

I finally solved my java plugin problems (as well as, it seems,
several instability problems I was having with iceweasel) by:

1. apt-get removing and dpkg --purging any java-related packages
found by dpkg -l|grep -i java.
2. apt-get installing azureus (which pulls in openjdk-6*) and
icedtea-gcjwebplugin.

Got the hint from this message on the debian-amd64 list (found by
searching; don't have an amd64 myself):

  http://lists.debian.org/debian-amd64/2008/11/msg00032.html

Maybe Debian is getting more sun-unfriendly. Anyway, it works now.

Regards, Jan



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: java plugin

[Jan Willem Stumpel]

Brad Rogers wrote:
> On Wed, 05 Nov 2008 12:29:39 +0100
> Have you done as instructed at plugindoc.mozdev.org?
> 
> Namely;
> 
>1. Install Java Runtime Environment.
>2. Make a symbolic link to libjavaplugin_oji.so in your Mozilla
>Plugins directory. Use the copy located in the plugin/i386/ns7
>directory of JRE 5.0 or later, or plugin/i386/ns610-gcc32 if you are
>using JRE 1.4.2.
> 
> The file you link to is
> in /usr/lib/jvm/java-6-sun-1.6.0.07/jre/plugin/i386/ns7 on my system.

Thanks!

I *think* that has already been done by apt-get. The only
difference seems to be that you have java-6-sun-1.6.0.07 while I
have java-6-sun-1.6.0.10.

I have the following symbolic links:

 -/usr/lib/iceweasel/plugins has
libjavaplugin.so -> /etc/alternatives/iceweasel-javaplugin.so
 -/usr/lib/mozilla/plugins has
libjavaplugin.so -> /etc/alternatives/mozilla-javaplugin.so
 -/etc/alternatives has
iceweasel-javaplugin.so -> /usr/lib/jvm/java-6-sun/jre/plugin/
i386/ns7/libjavaplugin_oji.so
mozilla-javaplugin.so -> /usr/lib/jvm/java-6-sun/jre/plugin/
i386/ns7/libjavaplugin_oji.so
- /usr/lib/jvm has
java-6-sun -> java-6-sun-1.6.0.10

and finally, /usr/lib/jvm/java-6-sun-1.6.0.10/jre/plugin/i386/ns7
has the file libjavaplugin_oji.so, perms -rwxr-xr-x, owned by
root, size 137021 bytes.

That seems OK to me. For good measure, I downgraded from the Sid
to the Lenny version -- i.e. from 1.6.0.10 to 1.6.0/07. But this
didn't help .. it is mysterious.

One thing: when re-installing (i.e. downgrading) sun-java there
was a warning:

update-binfmts: warning: current package is openjdk-6, but binary
format already installed by sun-java6

I don't know what this means. I have openjdk installed because it
was pulled in by the azureus package.

Regards, Jan




-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

java plugin

[Jan Willem Stumpel]

I have sun-java6-plugin installed (and thus also sun-java6-jre and
sun-java6-bin) but the java plugin does not appear in iceweasel's
about:plugins (and so java applets no longer work).

Running Sid; this problem seems to be fairly new. Any other Sid
users experience this?

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: nfs problem

[Jan Willem Stumpel]

Kurian Thayil wrote:
> Hi,
> 
> If I can remember correctly. Include the option no_root_squash
> in /etc/exports of B. It will be like,
> 
> /home/storage/video   A(rw,sync,subtree_check,no_root_squash)
> 
> You will be able to read-write as root if you include this
> option.

This did not really work. To my amazement, the following drastic
simplification worked:

on B: /etc exports reads

/home/storage/video   A(async) C(async)

A user on the home net (desktop machine A or C) can mount, read,
and write the shared directory now.

nfs remains mysterious. But I am slowly gaining experience, now my
home has become 100% Debian (two desktops and a
gateway/server/firewall).

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

nfs problem

[Jan Willem Stumpel]

Machines A and B both run Debian. There are no firewall rules
blocking any kind of traffic A<-->B.

I try to mount, by means of nfs, a directory of B to a mount point
on A, read-write.

/etc/exports in B has:
/home/storage/video   A(rw,sync,subtree_check)

/etc/fstab in A has:
B:/home/storage/video /mountB nfs \
user,rw,intr,rsize=8192,wsize=8192,timeo=14,noauto  0  0

If I call (on A)
   mount /mountB

I can read the contents of /home/storage/video on B. But I cannot
write anything to it. I keep getting the message "read-only
filesystem".

I hope there are some nfs experts here who can shed some light.

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: cedilla on a us_intl keyboard

[Jan Willem Stumpel]

Hugo Vanwoerkom wrote:
> Jan Willem Stumpel wrote:

>> It seems that many Debian users (including members of this
>> list) are not aware of the glorious Unix "Compose Key". You
>> press Compose, and then some other characters, and magically
>> a character is produced which is a kind of "graphical
>> combination" of those characters. So a c-cedilla (ç) is made
>> by Compose, comma, c. A German "double s" (ß) by means of
>> Compose, s, s. A British pound (currency) sign (₤) by
>> Compose, L, =.

> Good point Jan Willem, Florian has pointed this out before: 
> http://lists.debian.org/debian-user/2007/04/msg01873.html
> 
> Hugo

It seems I missed that thread. But anyway it is true that the
Compose key (a Unix/Linux feature, Windows doesn't have it) is a
very useful tool.

However, that thread did not quite answer the question posed by
its OP (Manon Metten). With the Compose key, you need 3 keystrokes
to make, e.g., á (Compose, ', a). If you use a true "dead keys"
method, you need only 2 keystrokes (', a). The downside then is
that to produce ' by itself, you need to follow the ' keypress by
a space, or press Alt-'. The choice between the "dead keys" method
and the "Compose" method is a matter of taste, depending on the
language(s) that you normally work with. For US users who only
need accented characters very rarely, the Compose method is
probably the best.

To enable true "dead keys" on an ordinary US keyboard you can set
a "dead key variant" in /etc/X11/xorg.conf:

 Option   "XkbVariant"  "alt-intl"

("Dead keys", which already existed in European mechanical
typewriters, were called "dead" because they did not advance the
paper-carrying "carriage". So a following character overprinted them.)

The xkb subsystem on modern versions of Linux has many more
wonderful features, especially if you convert your system to utf-8
(which is easy, and already the default on fresh installations of
Debian, I believe).

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: cedilla on a us_intl keyboard

[Jan Willem Stumpel]

Strange. I thought the us_intl keyboard had been removed from
Debian years ago. Nowadays you use the us keyboard with the
alt-intl /keyboard variant/.

It seems that many Debian users (including members of this list)
are not aware of the glorious Unix "Compose Key". You press
Compose, and then some other characters, and magically a character
is produced which is a kind of "graphical combination" of those
characters. So a c-cedilla (ç) is made by Compose, comma, c. A
German "double s" (ß) by means of Compose, s, s. A British pound
(currency) sign (₤) by Compose, L, =.

Etcetera, etcetera; hundreds of such combinations are pre-defined,
and you can also define your own.

Now where is the Compose key? I /think/ the Debian default is: the
right "windows" key is Compose (but I am not sure; it's been ages
since I set up a Debian system from scratch). In any case the
position of the Compose key can be specified through the GUI on
Gnome/Ubuntu and KDE. If you have no Compose key defined, you can
also set it by specifying in your /etc/X11/xorg.conf file, in the
keyboard section:

 Option   "XkbOptions"  "compose:rwin"

See also

http://www.jw-stumpel.nl/stestu.html#T6.1

Regards, Jan





-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: japanese input?

[Jan Willem Stumpel]

Dexter Filmore wrote:
> etch or lenny, doesn't matter, any will do.
> 
> Need japanese input on both gtk2 and qt apps.
> Possibly mouse input, i.e. draw signs -> the get instant-ocr'ed.
> 
> What are my options here?

If you have a utf-8 system (i.e. the output of the "locale"
command shows UTF-8 suffixes) you could use, e.g., uim or scim.

Instructions here:

http://www.jw-stumpel.nl/stestu.html#T6.4

I use uim myself. The uim menu offers a "handwriting input pad",
but it doesn't seem to be implemented in Debian. So you just
have to use the "Japanese word processor method": type toukyou,
press space, get 東京.

Regards, Jan






-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: re: how to close port 113

[Jan Willem Stumpel]

Jude DaShiell wrote:
> Install and run arno-iptables-firewall and tell it your
> internet port like eth0 or ppp0 and leave the rest of the
> defaults alone.  Port 113 will be closed once this is done
> since one of the defaults with arno-iptables-firewall is to
> first deny all ports then only open up those you specifically
> choose to open.

In general my advice would be to make your system secure /without/
a firewall. I.e. do not run services that you do not need, and
make the ones you /do/ need only accessible from the LAN, not from
the outside world. Then, you can run a firewall as a double
security. It is dangerous to rely on firewalls only for security
because it so easy to make mistakes with them.

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Switch-off problem

[Jan Willem Stumpel]

Daniel Burrows wrote:
> On Sat, Jul 19, 2008 at 11:46:16AM +0200, Jan Willem Stumpel
> <[EMAIL PROTECTED]> was heard to say:

>> Now my theory is that the switch-off problem occurs after the
>>  following events:
>> 
>> 1 - I mount the remote usb drive.
>> 2 - my wife switches off her computer.
>> 3 - I try to umount the remote usb drive -- this cannot be
>> done.
>> 
>> I suppose the shutdown program also tries to umount nfs
>> mounts, but fails, and then instead of skipping this step,
>> just hangs. Does this make sense? If so, is there a way to
>> solve this problem?

> You could try switching to a soft mount.  This can cause silent
> data loss when a server goes down, but I don't think that
> should be an issue if this is a read-only mount.  Just add
> "soft" to the mount options.  (I haven't tried soft-mounting
> for this particular case, so I don't know for sure that it will
> help)

Thanks much for pointing me to the "soft" option (which I did not
know about). Then by Googling I also found the "intr" option which
is said to be an alternative. But unfortunately neither option
cures the problem of shutdown hanging on the client when the nfs
server has gone off-line. Strange. This situation cannot be
uncommon in Linux home networking.

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Re: Switch-off problem

[Jan Willem Stumpel]

Ron Johnson wrote:

> On 07/19/08 04:46, Jan Willem Stumpel wrote:[..]

>> Now my theory is that the switch-off problem occurs after the
>>  following events:
>> 
>> 1 - I mount the remote usb drive.
>> 2 - my wife switches off her computer.
>> 3 - I try to umount the remote usb drive -- this cannot be
>> done.
>> 
>> I suppose the shutdown program also tries to umount nfs 
>> mounts, but fails, and then instead of skipping this step, 
>> just hangs. Does this make sense? If so, is there a way to 
>> solve this problem?
> 
> Manually "umount -l" the remote drive before you shutdown?

This does not work. Manually umounting the nfs drive (after the
computer which hosts the nfs drive is switched off) just leads to
a hang, whether the "lazy" switch is applied or not. In fact I
cannot even type the umount command entirely: the system hangs
immediately after typing the first 3 characters (including the
leading /) of the mount point.

Anyway it seems (from /etc/init.d/umountnfs.sh) that the -l
switch, as well as the -f switch, is already applied by default.
Running Sid; kernel is a stock Debian one:

2.6.25-2-686 #1 SMP Wed May 14 16:42:03 UTC 2008 i686 GNU/Linux

It seems that nfs on the remote computer fails to send a command
(before shutting itself down) to the computers to which it
exports, telling them to umount. This may be a bug, or, much more
likely, a failure on my part to understand the nfs export system.
Anyway, I never had this trouble while samba was used for sharing,
rather than nfs.

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Switch-off problem

[Jan Willem Stumpel]

Recently my computer devolped a new problem: it sometimes does not
switch itself off when I do shutdown -h now. It hangs somewhere
during the shutdown procedure, and has to be switched off by means
of the mains switch at the back. Fortunately the filesystem is
ext3, so at the next boot it starts again without problems.

I /think/ the problem may be related to nfs. I recently converted
my wife's computer to Linux. A usb disk connected to my wife's
computer contains media files which we both can use. This common
usb disk used to be shared by means of samba/smbfs, but now it is
exported by means of nfs.

Now my theory is that the switch-off problem occurs after the
following events:

1 - I mount the remote usb drive.
2 - my wife switches off her computer.
3 - I try to umount the remote usb drive -- this cannot be done.

I suppose the shutdown program also tries to umount nfs mounts,
but fails, and then instead of skipping this step, just hangs.
Does this make sense? If so, is there a way to solve this problem?

Regards, Jan





-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

A trick for non-supscrivvers

[Jan Willem Stumpel]

This message originally had the title "A trick for
non-subscribers". I send it again with the word misspelled.
The original version did not make it to the list because the list
interpreted the substring "subscribe" in the title as a request to
subscribe (which I do *not* want). Robots..

This is not a question but a tip. I just discovered this; of
course it may be old hat.

You want to participate in the list? But you don't want to
subscribe, because of the huge volume on the list?

The trick:

1. Use icedove as your mail/news reader, and tell Iceweasel
   you've done so.
2. Subscribe to newsgroup linux.debian.user.
3. Read messages from the newsgroup.
4. If you want to reply to a message, find the same message
   in http://lists.debian.org/debian-user. Then click the
   "reply to list" link. If everything goes well, an Icedove
   mail compose window will come up, with an empty message area.
5. Now the trick: in this compose window, click "options",
   "quote message". In the compose window, the message (which will
   go to the LIST), the message from the NEWSGROUP will be
   quoted.

This works in Sid. No guarantees for other Debian versions.

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: how to package?

[Jan Willem Stumpel]

Jan Brosius wrote:
> Hi,
> 
> I have the source of the program maxima. I would like to make a
> debian package of it. Is there any place where I can find
> documentation about making debian packages?
> 
> Thanks for any help Jan
Debian packages already exist. Try
  apt-cache search maxima
Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: CUPS vs lpd (was Re: Giving up on Iceweasel 3.0) SOLVED

[Jan Willem Stumpel]

Filing the bug
(http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=489765)
immediately produced an answer: you have to enter

gtk-print-backends = "file,lpr,cups"

in ~/.gtkrc-2.0.

Then lpr printing works with standard Debian packages. I think it
is a bug that this is not done automatically, or at least with
some warning during install. Anthony, I hope that this will work
in your case.

It will need some more research to find out why it worked by
default with the user-compiled versions, but I think I won't be
bothered.

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: CUPS vs lpd (was Re: Giving up on Iceweasel 3.0)

[Jan Willem Stumpel]

Ron Johnson wrote:

> On 07/07/08 09:14, Jan Willem Stumpel wrote:

>> It seems the package was heavily customised. Maybe some bias
>> favouring CUPS was introduced. But ff2 worked fine, so I am 
>> still not certain where the blame really lies.

> I'd say to file a bug against libgtk2.0-0 and see what response
> you get.

I just did. It is bug #489765 now.

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Re: CUPS vs lpd (was Re: Giving up on Iceweasel 3.0)

[Jan Willem Stumpel]

Anthony Campbell wrote:
> Mumia M. wrote:
>> I also created a startup script for firefox that sets
>> LD_LIBRARY_PATH to the 'lib' directory where gtk+2.10's
>> binaries are installed, e.g.:

>> #!/bin/sh
>> #-firefox-3.0.sh-
>> export LD_LIBRARY_PATH=/usr/local/exotic/lib
>> exec /usr/local/firefox-3.0/firefox

> Doesn't seem to help here. I don't have any "lib" directory in
> gtk-2.10; just engines, immodules, printbackends, and loaders.

> Perhaps I should have used an earlier version of gtk-2.10?

The standard self-compiled install puts *lots* of stuff in
/usr/local, e.g. in

  /usr/local/bin
  /usr/local/lib
  /usr/local/share
  /usr/local/etc

/usr/local/lib contains a subdirectory gtk-2.0 which has engines,
immodules, printbackends, and loaders in it. But also
/usr/local/lib ("above" gtk-2.10, not inside it) will contain new
gtk things.

BTW Mumia, it seems you use a (probably non-Debian) firefox
(because you use Etch), which is also in /usr/local. I use Sid,
and I can now print with the standard Debian Iceweasel.

The difficult thing now is to pinpoint where the bug is, exactly.
Ihe official Debian source files of libgtk2.0-0 are much bigger
than the gtk+-2.10 ones on gtk.org (23 MB vs 14 MB). To compile
the Debian version (which does not allow lpr printing through ff3)
you need to install *lots* of extra packages:
  gnome-common
  intltool
  libcairo-directfb2
  libcups2-dev
  chrpath
  gtk-doc-tools
  libcairo-directfb2-dev
  libcupsys2-dev
  svn-buildpackage
  unp
  gnome-pkg-tools
  libsvn-perl
  svn-buildpackage
It seems the package was heavily customised. Maybe some bias
favouring CUPS was introduced. But ff2 worked fine, so I am still
not certain where the blame really lies.

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: CUPS vs lpd (was Re: Giving up on Iceweasel 3.0)

[Jan Willem Stumpel]

Anthony Campbell wrote:

> I installed the gtk stuff in that way and I do have the entries
> you listed. But I don't seem to have the same printer entries
> in about:config.
> 
> I have put
> 
> print.print_printer  user_set  string lpr
> 
> which doesn't seem to work. I can't see options for your other
> entries. How did you create an entry for
> print.postscript.cups.enabled?

I set print.print_printer to PostScript/Default, not lpr. But it
does not seem to matter.

New preferences can be made by right-clicking anywhere in
about:config and selecting "New". It does not seem possible to
remove such preferences once they are created. But the setting of
print.postscript.cups.enabled also did not matter. I changed it
from "false" to "true" and I can still print. I cannot remember
exactly which other about:config print things I changed, but I now
tried several other things, and none seem to matter.

I hope you get it sorted out. Which version exactly did you
download from gtk.org? Mine is gtk+-2.10.13. Tarball was called
gtk+-2.10.13.tar.bz2 (an "older" version, not the "stable" 2.12
version).

BTW Mumia W., according to his message, stores his newly-compiled
gtk in an "exotic" location. I didn't do this. /usr/local (the
default when you compile from gtk.org) is exotic enough. Stuff
which is installed there has priority over similarly-named stuff
in /usr.

Do you have a printer named lp in printcap? (Or named as an
alternative, such as LaserJet|lp ?). Well, I suppose you do.

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: CUPS vs lpd (was Re: Giving up on Iceweasel 3.0)

[Jan Willem Stumpel]

Anthony Campbell wrote:
> On 07 Jul 2008, Jan Willem Stumpel wrote:

>> I compiled by just "./configure", "make", without setting any
>> options. So by default gtk does not have a cups print
>> backend, only "lpr" and "file" (when compiled on a cups-less
>> system, I presume). The Debian version has "lpr" and "file",
>> as well as "cups" and something called "test". [..]

> It didn't work here. I did the same thing and the appropriate
> stuff in /usr/local but I still can't print. Did you do
> something to make FF see the gtk-2.10 stuff?

Probably. I messed around a bit with the settings in about:config.
Some settings I have now:

print.postscript.cups.enabled  false  [AFAIK not a default
   setting; you have to create it]
print.print_paper_size 1
print.printer_list lp
print.save_print_settings  true
print.show_print_progress  true

Most of these about:config settings related to printing do not
seem to do anything, though. What do you mean by "the appropriate
stuff in /usr/local"? "make install" as root takes care of that.
"make uninstall" undoes the changes, useful for testing.

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: CUPS vs lpd (was Re: Giving up on Iceweasel 3.0)

[Jan Willem Stumpel]

Ron Johnson wrote:
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
> 
> On 07/06/08 15:06, Jan Willem Stumpel wrote:
>> I'll just mention the follow-up. I just tried installing gtk+-2.10
>> (compiling source from http://www.gtk.org/) and indeed, "print to
>> lpr" now appears in the print dialog. Without cups or xprint. Just
>> lprng. So as soon as gtk+-2.10 appears in Sid, this problem will
>> be over.
> 
> You're a bit behind the times, Jan!  ;)
> 
> $ apt-cache policy libgtk2.0-0
> libgtk2.0-0:
>   Installed: 2.12.10-2
>   Candidate: 2.12.10-2
>   Version table:
>  *** 2.12.10-2 0
> 500 ftp://mirrors.kernel.org unstable/main Packages
> 100 /var/lib/dpkg/status

Indeed. Very strange. And I even have it installed. But ff3 can
print to lpr only when I install the compiled-from-source version
of 2.10 (in /usr/local). As soon as I uninstall it, lpr printing
on ff3 becomes unavailable.

/usr/lib/gtk-2.0 (the Debian version) contains subdirectories:
   2.10.0  2.2.0  2.4.0  include  modules

No 2.12.0! I don't understand this numbering system.

/usr/local/lib/gtk-2.0 contains: 2.10.0  2.2.0  2.4.0  include
(no "modules", but the same version numbers).

The real difference is probably to be found here: the Sid version has:

[EMAIL PROTECTED]:~$ ls -alG /usr/lib/gtk-2.0/2.10.0/printbackends/
total 104
drwxr-xr-x 2 root  4096 2008-06-09 17:32 .
drwxr-xr-x 9 root  4096 2007-04-15 14:58 ..
-rw-r--r-- 1 root 52980 2008-06-07 11:16 libprintbackend-cups.so
-rw-r--r-- 1 root 13044 2008-06-07 11:16 libprintbackend-file.so
-rw-r--r-- 1 root 10084 2008-06-07 11:16 libprintbackend-lpr.so
-rw-r--r-- 1 root 11140 2008-06-07 11:16 libprintbackend-test.so

While the compiled-from-source version (from gtk.org) has:

[EMAIL PROTECTED]:~$ ls -alG /usr/local/lib/gtk-2.0/2.10.0/printbackends/
total 92
drwxr-sr-x 2 root  4096 2008-07-06 21:26 .
drwxr-sr-x 6 root  4096 2008-07-06 21:26 ..
-rwxr-xr-x 1 root  1315 2008-07-06 21:26 libprintbackend-file.la
-rwxr-xr-x 1 root 37984 2008-07-06 21:26 libprintbackend-file.so
-rwxr-xr-x 1 root  1309 2008-07-06 21:26 libprintbackend-lpr.la
-rwxr-xr-x 1 root 34034 2008-07-06 21:26 libprintbackend-lpr.so

I compiled by just "./configure", "make", without setting any
options. So by default gtk does not have a cups print backend,
only "lpr" and "file" (when compiled on a cups-less system, I
presume). The Debian version has "lpr" and "file", as well as
"cups" and something called "test".

Maybe you can explain these findings. Anyway it seems that the bug
is not in ff3 but in libgtk2.0-0.

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: CUPS vs lpd (was Re: Giving up on Iceweasel 3.0)

[Jan Willem Stumpel]

I'll just mention the follow-up. I just tried installing gtk+-2.10
(compiling source from http://www.gtk.org/) and indeed, "print to
lpr" now appears in the print dialog. Without cups or xprint. Just
lprng. So as soon as gtk+-2.10 appears in Sid, this problem will
be over.

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: CUPS vs lpd (was Re: Giving up on Iceweasel 3.0)

[Jan Willem Stumpel]

Ron Johnson wrote:

> On 07/06/08 09:59, Jan Willem Stumpel wrote:

>> You are right IMHO; it is a bug, not just the Inevitable 
>> March Of Progress. It seems clear that the Firefox developers
>> did not intend to drop lpr support:

>> - they don't say anything about such a change in the Release
>>   Notes.
>> - in about:config there are still numerous references to 
>>   PostScript/Default, lpr., etc.
>> - it seems such a silly thing to remove; very little code can
>>   be saved by this.

>> Surely it is a bug. From this item on bugzilla:

>> https://bugzilla.mozilla.org/show_bug.cgi?id=430818

>> it seems that until a few days before the Firefox 3 release 
>> (which took place on June 17th) Firefox could not print *at 
>> all*. [..]

> That's weird.  Sid only has -rc_2_ and I can print just fine. 
> Unless Eric Dorland (the IW maintainer) fixed the code himself.

>> There are lots of other bugs in the new Firefox/Iceweasel. A
>> particularly serious one (all resized images became black 
>> rectangles) apparently was quietly fixed a few days ago. I
>> expect the same will happen with the lpr bug. [..]

> Where did you read this?  I resize pictures, and they don't 
> turn into black rectangles.  Unless what I'm thinking of isn't 
> what "they" are talking about...

It could be that the "black rectangle bug" (which occurs /
occurred, for instance, when a Web page serves a picture at a
specific size, not its "natural" size) only happened with ATI
Radeon cards. There are several bug reports about it, for instance
this Ubuntu one:

https://bugs.launchpad.net/ubuntu/+source/firefox-3.0/+bug/182038

and very likely also Debian bug #487834. I have a Radeon, and
had trouble with this bug in FF3, not in FF2. But as I said, it
has disappeared (after an upgrade).

The "lpr printing bug" may also disappear soon. The bugzilla page
I mentioned now has a contribution by list member Mumia W. who
says lpr printing works again after installing GTK+2.10 (from
source). I'm going to try that..

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: CUPS vs lpd (was Re: Giving up on Iceweasel 3.0)

[Jan Willem Stumpel]

On 06 Jul 2008, Anthony Campbell wrote:
> On 05 Jul 2008, Ron Johnson wrote:
>> What is the reason that you don't install CUPS?

> 1. because lpr/lpd and magicfilter work perfectly for me.
> 2. because CUPS offers no advantage for me.
> 3. because when I did use it in the past my output was slightly
>worse than with my existing setup.
> 4. because, in the light of the above, using it will involve me
>in a fair amount of work for no clear benefit and possibly a
>worse output.

You are right IMHO; it is a bug, not just the Inevitable March Of
Progress. It seems clear that the Firefox developers did not
intend to drop lpr support:

  - they don't say anything about such a change in the Release
Notes.
  - in about:config there are still numerous references to
PostScript/Default, lpr., etc.
  - it seems such a silly thing to remove; very little code can be
saved by this.

Surely it is a bug. From this item on bugzilla:

https://bugzilla.mozilla.org/show_bug.cgi?id=430818

it seems that until a few days before the Firefox 3 release (which
took place on June 17th) Firefox could not print *at all*. It had
a totally empty print dialog box. It seems they could fix it to
some extent before the release date, but not completely.
Unfortunate for those who have used lpr to complete satisfaction
for so long (and not only for directly-connected printers BTW).

There are lots of other bugs in the new Firefox/Iceweasel. A
particularly serious one (all resized images became black
rectangles) apparently was quietly fixed a few days ago. I expect
the same will happen with the lpr bug. In the meantime I'll just
soldier on with the "inoticoming" trick. I refuse to change to cups!

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: killall firefox-bin

[Jan Willem Stumpel]

I wrote:

> Just writing this down gives me an idea: maybe it is a memory 
> leak thing. Next time it happens I'll check "free". Hadn't
> thought of doing that so far..

OK. It happened again. Iceweasel totally frozen. free says:

[EMAIL PROTECTED]:~$ free
  total  usedfree  shared  buffers   cached
Mem: 516128496924   19204   020524   216632
-/+ buffers/cache: 259768  256360
Swap:97992488  979836

OK. I do killall firefox-bin. Restart iceweasel; it works fine
again. free now says:

[EMAIL PROTECTED]:~$ free
  total  usedfree  shared  buffers   cached
Mem: 516128445980   70148   021024   193668
-/+ buffers/cache: 231288  284840
Swap:97992488  979836

Not a great deal of difference I think, apart from the "Mem free"
value (but I always understood that the "free -/+ buffers/cache"
value was more important). Does this give any clues?

Regards, Jan




-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Mailbox conversion

[Jan Willem Stumpel]

Just to mention the follow-up: I used kmail and its convert
program on the old dbx files, and it worked fine. But the results
were in maildir format (each message in a separate file), not in
mbox format (a whole lot of messages in one file, which apparently
Thunderbird/Icedove are used to). But some googling revealed a
maildir-to-mbox shell script: http://tinyurl.com/6zdn4s

This produced mbox files that I could drop into my wife's
Thunderbird local mail directory. The only trouble was that the
script did not recognise filenames with spaces in them (even when
quoted with a reverse slash character). But simple renaming fixed
this. Everything works fine now. Thanks, list!

Regards. Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Mailbox conversion

[Jan Willem Stumpel]

Mark Grieveson wrote:

[..]
> If the idea of running a dos program on Linux nauseates you, 
> then I also read that Kmail, with the kmailcvt (kmail
> converter) package installed also works.

Actuallly I am quite a fan of dosemu (and even of DOS), but DOS
can't handle filenames longer than 8+3, let alone names with
Japanese characters in them.

I think I'll try the kmail route first; this has now been
mentioned twice.

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Mailbox conversion

[Jan Willem Stumpel]

Finally the motherboard of my wife's computer broke down, so I had
to replace it. I could not just get a new motherboard, because
they do not sell boards with "462" sockets anymore. The cheapest
solution was an MSI motherboard with AMD-64 dual core and one gig
of memory. You cannot get anything simpler nowadays. I can clearly
remember the time when such specs would have meant "supercomputer"
(as in "several million dollars").

This operation made her old Windows 2000 unusable. You can run a
32-bit OS on a 64-bit machine, but some things (notably the ATI
display drivers for the new motherboard) refuse to install if the
OS is not 64-bit. A great opportunity to install Linux. I didn't
install Debian itself, but the 64-bit version of Ubuntu. After
some tweaking it works fine, including some difficult things like
Japanese input, printing through the home network, etc. Before
reformatting the Windows hard disk I first copied its contents to
a USB external disk. From there I could restore the mail address
book (Outlook -> Thunderbird) and the browser favorites/bookmarks
(IE -> Firefox). My wife thinks the new system is great.

After all this smalltalk, my real question: I still have lots of
old e-mail in Microsoft format (with names like in-2005.dbx,
out-2006.dbx, etc; some filenames are in Japanese, and the
messages inside the .dbx files are mostly Japanese). I searched
the web for "converting dbx to mbox". Found lots of answers, most
of them seem old, and they do not agree. Is there a "canonical"
way to make .dbx mail archives available to Linux Thunderbird/Icedove?

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: killall firefox-bin

[Jan Willem Stumpel]

Ron Johnson wrote:
> You don't tell us which Debian branch and version of IW you are
> using.

I am using Sid, but I still use the "testing" version of IW
(2.0.0.14). The Sid version is at the moment a little bit *too*
"unstable" (for systems with an ATI video card: see bug 485917).
So my experience with the unexplained "freezing" of IW is with
2.0.0.14 (and possibly some of its recent predecessors).

The "weird things with the X display" I mentioned occur when I try
to move the window with the "dead" (or possibly only hibernating)
Iceweasel in it. It leaves pieces of itself, and of overlying
windows, all over the place. The most annoying thing is that I
cannot determine what triggers this, apart from that it happens
after IW has been in use "for a while".

Just writing this down gives me an idea: maybe it is a memory leak
thing. Next time it happens I'll check "free". Hadn't thought of
doing that so far..

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

killall firefox-bin

[Jan Willem Stumpel]

At least once a day I have to give the command "killall
firefox-bin", because the systems just about freezes (at least the
browser does), and weird things happen to the X display. Do others
have the same experience?

"Iceweasel" as is it now unwisely called by Debian (although
according to ps aux the running process is still "firefox-bin")
seems to be unstable somehow. I don't know why. I can't predict
when this "freezing" will happen, so I can't file a bug about it;
it just happens, about one time per day.

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: changing xterm colors

[Jan Willem Stumpel]

Jamie Griffin wrote:
> Jan Willem Stumpel wrote:
>> You can edit a text file called .Xresources in your home
>> directory (or create it if it does not exist).
>>
>> Put the following lines in the file:
>>
>> xterm*VT100*foreground:  green
>> xterm*VT100*background:  black
>> xterm*VT100*cursorColor: red

> I've tried that and it hasn't worked. Not sure what to try next
> - does anyone have any other ideas?

This definitely *should* work. If it does not work something weird
is the matter. Are you sure your terminal program is xterm? For
instance, do you see the "VT Fonts" menu if you do a
control-rightclick in the terminal window?

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: changing xterm colors

[Jan Willem Stumpel]

Jamie Griffin wrote:
> I've been trying to change the colors on my Xterm to have a
> black background and green text.

You can edit a text file called .Xresources in your home directory
(or create it if it does not exist).

Put the following lines in the file:

xterm*VT100*foreground:  green
xterm*VT100*background:  black
xterm*VT100*cursorColor: red

(Well, that is if you want the cursor to be red.)

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: USB drive names

[Jan Willem Stumpel]

Ron Johnson wrote:
> Jan Willem Stumpel wrote:
> > I bought some USB external hard disks recently. They automount
> > beautifully on a Debian system.

> When you say that, do you mean the automount daemon, autofs, or
> a similar feature built into various desktops?

Sorry, I should have said "on an xfce4 system". I do not know
which mechanism is used by xfce4.

Raj Kiran Grandhi wrote:
> You can just change the volume label of those drives to whatever
> you want.

Yes, this works, thanks. One needs mtools, because these external
disks have a vfat file system on them. I found a step-by-step
description here: http://ubuntuforums.org/showthread.php?t=65830

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

USB drive names

[Jan Willem Stumpel]

I bought some USB external hard disks recently. They automount
beautifully on a Debian system. One is a "LACIE" disk, another one
is a Western Digital "Elements" disk.

They mount as /media/LaCie and /media/Elements respectively.
Question: is it possible to change those names? E.g., if I would
like to change the names to (say) Lassie1 and Lassie2, how could I
go about it? Also, where is the name of the mount directory (now
"/media") stored?

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: reportbug problem

[Jan Willem Stumpel]

Hugo, Michael, and Frank all pointed into the direction of exim4.
This was correct. Being conservative, I had always resisted
upgrading from exim3 to exim4 (because exim4's configuration is
more difficult). The latest sid upgrade apparently became annoyed
at this and removed exim3, declaring it "obsolete" -- but without
replacing it by anything else.. so I had no mta.

Why Sid suddenly has become so intolerant of exim3 I don't know.

I could continue to use e-mail (but not reportbug) because my
configuration is:

vega -> procyon -> internet

procyon is my Sarge "home server" (still running exim3). vega is
my "normal" computer. An mta on vega is only necessary for
programs like reportbug and mail (i.e. "mail", the classic
program). Other programs (mua's like icedove) already know that
procyon is the outgoing smtp server. Maybe I can teach this fact
to reportbug and mail also (how?). If so, I won't need an mta on vega.

Anyway dpkg -l grep exim on vega revealed that exim3 had been
removed. So crossing my fingers I installed exim4 on vega and told
it that procyon was the smarthost. It worked. Thanks to all who
responded!

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

reportbug problem

[Jan Willem Stumpel]

I tried to file a bug via reportbug yesterday, and it was not
sent. Instead, the mail message was saved as a file in /var/tmp.
I've successfully sent bug reports through reportbug in the past,
so what could be the matter now? Using Sid.

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Icedove time categories

[Jan Willem Stumpel]

John Allen wrote:

> View->Sort By->Group By

This did not work (it actually says "Group By Sort G" instead of
just "Sort By" in my case) but View->Sort By->Order Received
worked. Thanks!

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Icedove time categories

[Jan Willem Stumpel]

After some Sid upgrades, Icedove now puts my "inbox" and "sent
box" messages into time categories:

- Today
- Yesterday
- Last week
- Two weeks ago
- Old mail

I do not need/want this feature, and have been looking for a way
to turn it off. Couldn't find it. Anyone knows how?

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

sensors

[Jan Willem Stumpel]

When I run the "sensors" command, I get:

[EMAIL PROTECTED]:~$ sensors
w83627thf-isa-0290
Adapter: ISA adapter
VCore:   +1.44 V  (min =  +0.00 V, max =  +3.84 V)
+12V:   +12.40 V  (min =  +5.84 V, max =  +0.00 V)
+3.3V:   +3.39 V  (min =  +0.10 V, max =  +3.23 V)
+5V: +5.04 V  (min =  +2.59 V, max =  +0.11 V)
-12V:+1.70 V  (min = -13.76 V, max = -14.91 V)
V5SB:+5.05 V  (min =  +0.27 V, max =  +3.44 V)
VBat:+0.16 V  (min =  +0.59 V, max =  +3.20 V)
fan1:  0 RPM  (min =  164 RPM, div = 128)
CPU Fan:1854 RPM  (min =   -1 RPM, div = 4)
fan3:  0 RPM  (min =  131 RPM, div = 128)
M/B Temp:+28.0°C  (high = +16.0°C, hyst = +10.0°C)  sensor = 
thermistor

CPU Temp:+30.0°C  (high = +80.0°C, hyst = +75.0°C)  sensor = diode
temp3:   -48.0°C  (high = +80.0°C, hyst = +75.0°C)  sensor = 
thermistor

cpu0_vid:   +1.388 V
beep_enable:enabled

Are these values reasonable, or is there something wrong with the 
"lm-sensors" configuration? I am especially wondering about the 
values for "-12V", "VBat", "V5SB" (don't know what this is), and 
"temp3". Also, my box has 3 fans (PSU fan, CPU fan, case fan) and 
they all turn, but sensors says only one is turning. Anyone knows 
a good tutorial about setting up sensors?


Regards, Jan




--
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

[OT] webcam colours

[Jan Willem Stumpel]

I have an extremely cheap (8 euro) webcam (Sunplus Technology Co.,
Ltd Flexcam 100, according to lsusb).

It works on my Sid box with 2.6.24-1-686 kernel, using the gspca
kernel module.

When I test it with camorama, the picture looks "blue" (i.e. blue
is the dominant colour). On the other hand, in Skype, the colours
seem OK.

Then I read somewhere that the gspca module should be loaded with
the option "force_rgb=1". This reverses the situation: natural
colours in camorama, but "blue picture" in Skype.

Anyone else seen something similar?

Neither Skype nor camorama work without the gspca module loaded. I
wonder if there is a bug in a) my €8 cam  b) gspca  c) camorama
d) Skype.

Regards, Jan


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Icedove news

[Jan Willem Stumpel]

On Sat, 09 Feb 2008 09:30:41 Ralph Katz wrote:

Hi Jan -- No.  On up-to-date Etch, using icedove version 1.5.0.14pre
 (20071018), there is no problem here.  You could close icedove, move
 your ~/.mozilla-thunderbird/default.lgh/News/ out of the way, and
let icedove rebuild it upon restart.  Then re-subscribe to your
groups. I've never had your particular problem, but this solution has
worked for re-setting subscriptions and message-read data for me
before on earlier versions of thunderbird.


Thanks. This seems to have done it. I renamed the whole
~/.mozilla-thunderbird to ~/.mozilla-thunderbird-old, then dpkg
 --purge'd mozilla-thunderbird, thunderbird, and icedove, then
re-installed icedove. It proceeded to import settings (from where, I
cannot guess); but anyway it works now, and the pop-ups no longer occur.
Only the font of the messages had changed, but that could be fixed easily.

Thanks again, Jan




--
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Icedove news

[Jan Willem Stumpel]

I recently started using Icedove for mail & news (until now, I 
only used Iceape or a hacked Fedora Seamonkey version).


I have some problems with "news". I have subscribed to a few
newsgroups offered by the news server at my ISP (news.planet.nl).

Each time I click on a newsgroup (say, alt.html) a window pops up
asking if I want to subscribe to this newsgroup (to which I am
already subscribed). If I say "yes", a side-effect is that in the
"newsrc" file (in my case ~/.mozilla-thunderbird/default.lgh/News/
newsrc-news.planet.nl) a new line is created for this newsgroup.
The newsrc file therefore tends to grow without bounds, with many
duplicate lines.

Also, and because of this, it is impossible to unsubscribe from 
newsgroups. If I try to unsubscribe, the number of duplicate lines 
for this group in the newsrc is diminished by one. But as there 
are still many left, the next time I start icedove I am still 
subscribed to that group.


Do other list members also get these "do you want to subscribe" 
popups when using icedove for news?


Regards, Jan


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

fsck.ext3

[Jan Willem Stumpel]

After converting my file system to ext3, I thought there would be
no more lengthy fsck's every 20 or so boot-ups. But they still
happen.

Some Googling revealed different opinions; some people say "ext3
does not need periodic fsck's", others say "even with ext3, it is
best to fsck every N boots, because PC hardware is cr*p". And
there are also several different methods (apparently) for
disabling periodic fsck's.

Is there a "standard Debian doctrine" or "standard Debian
practice" for this?

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Re: [OT] where is conio.h with getch()?

[Jan Willem Stumpel]

You can use ncurses, but if it is only a getch() you want, you can
implement it using select() (not a function which is easy to
understand BTW). Just google for "getch" together with "select".

A sample implementation is here:

http://cygwin.com/ml/cygwin/1999-08/msg00385.html

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: which to use: ext3, JFS, XFS, ReiserFS? [Was: new userquestion: debian on a Thinkpad T61]

[Jan Willem Stumpel]

Curt Howland wrote:

> If I may interject, creating the journal just creates a blank
> file.

This would explain why creating the journal does not seem to take
any time. But "strings" showed that there was a lot of stuff (at
least lots of filenames) in it. Perhaps the journal is *created*
as a blank file, and then some background process immediately
begins to fill it? Anyway, whatever it does, it seems to be a very
clever system.

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: which to use: ext3, JFS, XFS, ReiserFS? [Was: new userquestion: debian on a Thinkpad T61]

[Jan Willem Stumpel]

Jan Willem Stumpel wrote:

> Some steps may have been unnecessary, but it seems I have a
> working ext3 system now. It is really easy. The real smoke test
> will come, of course, when I pull the plug. Will do this now; if
> you do not hear from me, the test will have failed. Thanks to all
> who responded!

And it worked! I rudely pulled the plug when X and its apps were
still active, and after restoring the voltage the system rebooted
without any complaints. Hugo, you should convert to ext3 as well.

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Re: Re: which to use: ext3, JFS, XFS, ReiserFS? [Was: new userquestion: debian on a Thinkpad T61]

[Jan Willem Stumpel]

Paul Johnson wrote:

> Step 1: Get root privileges.
> Step 2: Type tune2fs -j /dev/whatever
> Step 3: Remount the filesystem ext3...

I did this, and indeed it was amazingly easy. On a partition of
about 24 G (well, this is an *old* disk!) a file /.journal of 128
M (indeed much less than 1%) was created instantaneously. Now
mount -l shows that I have an ext3 system.

What I did was (may have been over-cautious):

Step 1: close all applications, close X, get into a console.
Step 2: get root privileges.
Step 3: close the SMB connection to my wife's Windows machine.
Step 4: /etc/init.d/networking stop
Step 5: edit /etc/fstab; change ext2 to ext3 for my root device
(/dev/hda5 in my case).
Step 6: type tune2fs -j /dev/hda5. The journal was created
instantaneously (I'd expected this to take a long time.
but it did not).
Step 7: reboot.

Some steps may have been unnecessary, but it seems I have a
working ext3 system now. It is really easy. The real smoke test
will come, of course, when I pull the plug. Will do this now; if
you do not hear from me, the test will have failed. Thanks to all
who responded!

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Re: which to use: ext3, JFS, XFS, ReiserFS? [Was: new userquestion: debian on a Thinkpad T61]

[Jan Willem Stumpel]

Александър Л. Димитров wrote:
> Quoth Hugo Vanwoerkom:
>>
>> ext2. Never have used any other.
>
> I seriously hope that this was a joke...

Maybe it was, but I never used anything but ext2 either, and that
is no joke. It has worked fine for many years. I often considered
"upgrading" to ext3, but so far I've never taken this step. I
expect this is the same for many "users of old".

I am especially put off by the Wikipedia article on ext3. It gives
a rather long list of "disadvantages". One of them ("No
checksumming in journal") even sounds pretty frightening. The list
of "advantages" is very short, and they are mostly advantages over
Reiserfs and other non-ext2 systems, not advantages over ext2.

But sometimes bugs in applications can cause a complete freeze of
X, incl. keyboard and mouse. It happens to me about once a year,
unfortunately also yesterday evening. In such a case there is
nothing you can do but pull the plug. Then when you reboot, all
sorts of alarming messages appear. By invoking fsck one can
normally get the system to boot again, but there may be
side-effects (e.g. my old iceweasel "history" was gone after the
reboot yesterday).

So now I am more or less ready to take the plunge. But I would
still like some advice.

1. Is it true that ext3 always lets you recover smoothly after a
   "freeze and pull the plug", or after a power cut? Or are there
   still "ifs" and "buts"?
2. Is significant room on the disk (or partition) taken by the
   journal? By how much can I expect the disk capacity to be
   reduced?
3. It is said ext3 is "slow". Does this apply to writing only, or
   also to reading? I.e., is there a danger that when I play a
   film with mplayer, I'll get the dreaded message "Your system is
   TOO SLOW to play this"?
4. I have my whole Linux system, apart from swap (i.e. the root,
   and everything that branches off it, like /boot, /var, /usr)
   just on one logical partition. Can I still convert to ext3,
   possibly by using a Knoppix or Ubuntu CD-ROM to boot from?
5. Where can I find reliable, step-by-step instructions for the
   conversion? There are several such instruction sites on the
   Web, but I am not sure they always agree.

PS: Kernel is 2.6.20-1-686

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Open Office and German / Japanese Input

[Jan Willem Stumpel]

Dietrich Bollmann wrote:
> Hi,
>
> Every now and then I have to write something in German or
> Japanese.  And normally this entails about one week of trials
> to make Open Office work again with "scim" or "uim" (both input
> methods).  When finally everything works, it is normally too
> late for my letter but there is still time enough to give Open
> Office a chance to not work anymore when needed again.
[..]
> Any idea how to make Open Office work with "uim" or "scim"
> again?

If you have a UTF-8 locale (like en_GB.UTF-8; this is
case-sensitive) there is no need to ever change it again.

Google "International text support on Linux" and select the first
result. This explains all the details.

For Japanese, you need anthy (most convenient in combination with
uim). If you want to use scim (which is possible although I do not
recommend it) be aware that scim needs a line in
/etc/scim/global.conf, if you want to use it in an en_GB locale:

  /SupportedUnicodeLocales = en_US.UTF-8,en_GB.UTF-8

(the default is that scim only works with en_US.UTF-8; weird, but
this has not changed for years). Anyway, uim is better than scim
for "international" users (I think) because it has better support
for the "Compose Key".

For German on an English keyboard you do not need any special
software. Just set up your keyboard to be the "us" keyboard with
the "alt-intl" variant. This can be used both for English and German.

When this is set up correctly, Japanese / German / English input
will work in all programs, including Openoffice, and will continue
working after upgrades as well.

Regards, beste Grüße, よろしくお願いします, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

How to find the font?

[Jan Willem Stumpel]

This question is not Debian-specific.

I am working in X (not in "text mode"), and some text is displayed
in some application; e.g. in iceweasel, but it could be something
else. So a part of my screen is filled with some text, in some
font, or in various fonts.

Is there any way to find out which fonts are actually being used?
I mean, ideally I would like to be able to select a part of the
displayed text, and then some application, daemon, or whatever,
would say, e.g.: "this is Freemono, at 16 points". Is this possible?

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: cannot get mathml fonts to work in Iceape or Iceweasel

[Jan Willem Stumpel]

H.S. wrote:

> I think I have succeeded in installing the fonts (see the
> output of some commands below). But I am not able to get proper
> rendering on this page: 
> http://www.mozilla.org/projects/mathml/demo/texvsmml.xhtml


MATHML in Mozilla products is still buggy in Debian (old bug, see
#401533 and others).

To get correct MATHML display, you can edit /etc/iceape/iceaperc,
changing MOZ_ENABLE_PANGO=1 to MOZ_DISABLE_PANGO=1. However, this
only cures the *display* of MATHML. *Printing* MATHML still does
not work correctly. Also, disabling Pango has side-effects. Some
non-Latin scrips (like Devanagari) are not displayed (or printed)
correctly when Pango is off.

If printing MATHML is important for you, you can try to install
the Fedora Core version. FC cured the bug in December last year.

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Simultaneous sounds, again (solved, probably)

[Jan Willem Stumpel]

I am pleased to say that I now managed to get simultaneous sound
working in Debian, almost perfectly.

My "sound compatibility matrix" is now:

  mp/s  xmmsyoutube   skype

mp/s   OKOK   OKOK

xmms   OKXOKOK

youtubeOKOK   OKOK

skype  OKOK   OKN/A


mp/s means mplayer with sdl audio output, as before. I had to set
this *both* by a line in ~/.mplayer/config:

   ao=sdl

and through the GUI (gmplayer, preferences, audio). It appears
both are necessary.

The X in the combination xmms-xmms should probably be N/A, because
xmms just does not allow running 2 instances of itself
simultaneously.

Several reactions (on-list and off-list) suggested using a sound
daemon, but according to the 31 August 2007 entry on the
"buglandia" blog (http://buglandia.blogspot.com), sound daemons
(like esd and arts) are now *outdated*, and even somewhat *evil*
("fundamentally wrong").

The solution is software sound mixing. Apparently, sounds can be
mixed by the hardware (the sound card) or by software. Windows
always uses software sound mixing by default. So sound mixing
always works on Windows, no matter what hardware you have. Linux,
developed by and for tech types, tries to use hardware mixing if
the sound card can support it (but by no means all of them do),
because this is more efficient.

Software mixing became something of a stepchild. But it appears
that in the more modern versions (of Linux and alsa; certainly in
Sid) software mixing works now. The flightgear wiki
(http://tinyurl.com/38smlf) explains this better than I can.

To make it work, I followed a recipe on the alsa wiki
(http://tinyurl.com/2vhoqt), because the recipe on buglandia did
not work for me. What did was:

-- make sure there is NO /etc/asound.conf or ~/.asoundrc files
   present on the system.
-- export two environment variables:
   export SDL_AUDIODRIVER=alsa
   export AUDIODEV=plug:dmix
   (you can put both commands in ~/.bash_profile).
-- make sure that mplayer uses sdl for audio output.
-- make sure there are no sound daemons running (jack, esd, arts,
   whatever).

It will now work after you log out/log in, and perhaps after
restarting alsa (/etc/init.d/alsasound restart). Well, I hope it
does. It now does for me.

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Simultaneous sounds, again

[Jan Willem Stumpel]

Trying to get multiple simultaneous sound streams working
on my Sid system, I got as far as the following:

  mp/s  mp/a   xmmsyoutube   skype

mp/s   OK   N/A XXXX

mp/a   N/A   X  XXXX

xmms   X X  XOKOK

youtubeX X  OK   OKOK

skype  X X  OK   OKN/A

Notes: - each row in the table begins with the sound using
 program which is started FIRST.
   - mp/s means: using mplayer to watch a local .avi file,
 with sdl audio output.
   - mp/a means: using mplayer to watch a local .avi file,
 with alsa audio output.
   - youtube means: watching a youtube (i.e. Flash) movie
 with iceape.
   - X means: the program which is started second produces
 no sound
   - XX means: the program which is started second produces
 no sound, and there are other problems (iceape freezing).
   - OK means OK.
   - N/A means: test cannot be done, or is meaningless.

So it seems that there is one group (xmms/Skype/youtube) the
members of which tolerate each other (all three can even make
noise together). There is another "group" (mplayer) which
allows multiple instances with simultaneous sound, but only if
sdl sound is selected. However, the two "groups" do not
tolerate each other. I cannot hear Skype ringing while playing
a local .avi file, for instance.

Does anyone get different/better results? Any tips?

I also think (but cannot prove this now) that a few months ago I
could hear skype while playing .avi movies, but that it became
impossible since (doubtless through some upgrade).

Sound hardware is AC 97 on motherboard.

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: AltGr on Japanese keyboard

[Jan Willem Stumpel]

Oops, sorry, this should be without "=", e.g.

setxkbmap "us(alt-intl)+jp"

or perhaps in your case

setxkbmap "us(alt-intl)+jp(OADG109A)"


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: AltGr on Japanese keyboard

[Jan Willem Stumpel]

The "jp" keyboard layout (/usr/share/X11/xkb/symbols/jp) does not
seem to specify any 3rd (or 4th) level symbols. I do not know what
you want to do with your AltGr key, but if it is to get things
like AltGr-5 = Euro symbol, or AltGr-minus = dead macron key, you
could try

setxkbmap="us(alt-intl)+jp"

or perhaps in your case

setxkbmap="us(alt-intl)+jp(OADG109A)"

I do not have a Japanese keyboard so I cannot test it completely.
Please tell if this works or not.

Regards, Jan
http://www.jw-stumpel.nl/stestu.html#T6



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: How to fix Language error

[Jan Willem Stumpel]

It seems that nowadays in Debian, you must set the locale by
running as root

update-locale LANG=fr_FR.UTF-8

This stores the locale value in /etc/defaults/locale (instead of
/etc/environment).

Regards, Jan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]