Re: Problems with XFree86 4.1.0-8: could not open default font 'fixed'
On Mon, 22 Oct 2001, Patrick Schnorbus wrote: Suddenly, X won't start up anymore. I have no idea how long this problem exists, since my computer had been running continuously for a long time before I encountered the problem and I have no idea how long the last X session has lasted (some weeks, at least). try to remove the quotes in /etc/X11/Xsession.d/99xfree86-common-start That is not the problem, I already removed the quotes some time ago. Thinking back, I have restarted X succesfully at least once after XFree86 4.1.0-7 came out and the quotes problem appeared. That X session lasted for a few weeks. I installed XFree86 4.1.0-8 while X was running and now it won't start up at all. Remco
Problems with XFree86 4.1.0-8: could not open default font 'fixed'
Hi, My computer is running the latest Debian unstable, including XFree86 4.1.0-8. Suddenly, X won't start up anymore. I have no idea how long this problem exists, since my computer had been running continuously for a long time before I encountered the problem and I have no idea how long the last X session has lasted (some weeks, at least). In /var/log/XFree86.0.log, I encountered these lines: quote Could not init font path element /usr/lib/X11/fonts/100dpi/:unscaled, removing from list! Could not init font path element /usr/lib/X11/fonts/75dpi/:unscaled, removing from list! Could not init font path element /usr/lib/X11/fonts/misc, removing from list! Could not init font path element /usr/lib/X11/fonts/100dpi, removing from list! Could not init font path element /usr/lib/X11/fonts/75dpi, removing from list! Fatal server error: could not open default font 'fixed' /quote These are the FontPath lines in my /etc/X11/XF86Config-4: quote FontPathunix/:7100 # local font server (xfs) FontPathunix/:7101 # local font server (xfstt) FontPath/usr/lib/X11/fonts/100dpi/:unscaled FontPath/usr/lib/X11/fonts/75dpi/:unscaled FontPath/usr/lib/X11/fonts/misc FontPath/usr/lib/X11/fonts/Type1 FontPath/usr/lib/X11/fonts/Speedo FontPath/usr/lib/X11/fonts/100dpi FontPath/usr/lib/X11/fonts/75dpi FontPath/usr/lib/X11/fonts/cyrillic /quote Since my font path includes xfs, I thought xfs should also be able to serve a font 'fixed'. But xfs logs lines like these into the syslog every time it starts: quote Oct 21 19:47:25 remco xfs: ignoring font path element /usr/lib/X11/fonts/100dpi/:unscaled (unreadable) Oct 21 19:47:25 remco xfs: ignoring font path element /usr/lib/X11/fonts/75dpi/:unscaled (unreadable) Oct 21 19:47:26 remco xfs: ignoring font path element /usr/lib/X11/fonts/misc/ (unreadable) Oct 21 19:47:26 remco xfs: ignoring font path element /usr/lib/X11/fonts/CID (unreadable) Oct 21 19:47:26 remco xfs: ignoring font path element /usr/lib/X11/fonts/100dpi/ (unreadable) Oct 21 19:47:26 remco xfs: ignoring font path element /usr/lib/X11/fonts/75dpi/ (unreadable) /quote It seems that these font directories are somehow unreadable. I have checked the permissions of all font directories and all font files. Nothing seems to be wrong. I have run update-fonts-alias, update-fonts-dir and update-fonts-scale on all font directories, but it doesn't help. The strangest thing is, that this problem seems to have started without any changes to configuration files. Is this a bug anybody knows of? Have I done anything wrong I haven't thought of yet? Please help, I'd like to have my X back. Remco
Re: GLX on NVidia TNT card using XFree86 4?
On Mon, 26 Mar 2001 [EMAIL PROTECTED] wrote: I'm not sure Nvidia's drivers work at 24bpp, so you might want to try running at 16bpp. Also, you did remember to change the driver in your XF86Config (or XF86Config-4 if that file exists on your system) file to be nvidia rather than nv, right? Running X at 16bpp instead of 24bpp did the trick. Now 3D is accelerated. Remco -- qn-195-66-31-144: 01:30:01 up 6 days, 54 min, 6 users, load average: 2.15, 1.95, 1.91
GLX on NVidia TNT card using XFree86 4?
I've got a Diamond Viper770 video card, which has a Nvidia TNT chip on it. Using the standard XFree86 4 nv driver, 3D performance is, well, poor. So I decided to install the driver from NVidia. I downloaded and installed the nvidia-glx-src and nvidia-kernel-src packages. With those, I built the nvidia-glx and nvidia-kernel-2.2.18 (yes, I'm using kernel 2.2.18) packages and installed those, too. Then I reconfigured the X server to use the new nvidia driver. But 3D performance is still as poor as it was before. Have I done something wrong and if so, what? BTW, I have tested the 3D performance by running the GLText demo from xscreensaver. With a resolution of 1600x1280 at 24bpp, I get about 2 fps. Remco -- qn-195-66-31-144: 23:20:02 up 3 days, 22:44, 9 users, load average: 1.58, 2.05, 1.64
Bind9 broke inn2 2.2
I am running Debian 'unstable', but with inn2 from 'testing'. Since newsx doesn't work with inn2 2.3 yet (and I need newsx), I am still using inn2 version 2.2.2.2000.01.31-4.1. But since I installed the bind9 packages, I get tons of messages like this in my log files: Feb 28 22:46:11 qn-195-66-31-144 ctlinnd[2141]: Hostname does not resolve or 'domain' in inn.conf is missing Sometimes inn dies because of this, other times it just throttles. If I restart it, it works again for some time. Can anybody help me with this problem? Remco -- qn-195-66-31-144: 7:15pm up 21:03, 9 users, load average: 1.60, 1.40, 1.17
Fourth button of Logitech MouseMan Wheel?
Hi, I just bought a Logitech Cordless MouseMan Wheel. Previously, I had a Logitech Wheel Mouse. Using the same setup in X, I can use three buttons and the wheel just like with the Wheel Mouse. But how can I use the fourth (thumb) button on my new mouse? I have tried several things, but to no avail. Here's part of my current XF86Config: Section Pointer Protocol IMPS/2 Device /dev/psaux Buttons 6 Resolution 400 ZAxisMapping 4 5 EndSection With this setup, I tried to run the xev program to see what the mouse does. The wheel generates events for buttons 4 and 5, but the thumb button generates no events at all. Can anybody please help? Remco -- rd1936: 11:00pm up 3 days, 3:39, 10 users, load average: 2.07, 1.86, 1.82
Re: ITP: TIP
On Wed, 29 Dec 1999, Jordi wrote: I asked for comments in debian-mentors about how to make users aware of the availability of a free package that replaces a non-free one that they already were using. One idea was to make Pine Recommend: tip | pico, and that sounds fair enough. Any other? Since the pine package that is built by the pine396-src and pine396-diffs packages doesn't need the pico package (or any other external editor) in any way, I'd say such a Recommends: line would be evil. Remco -- rd1936: 9:45pm up 45 days, 10:11, 7 users, load average: 2.80, 2.53, 2.21 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: win95files names
On Wed, 10 Nov 1999, Antonio Rodriguez wrote: How should I mount a win9* system, so that I can read the full file name without the funny ~1 thing? I have tried mount -t msdos , but this truncates the name to *~1. I have a few linux files that I want to be able to recognize from the linux partition. Try -t vfat instead of -t msdos. This will support the long filenames. Remco -- rd1936: 11:00pm up 5 days, 21:54, 6 users, load average: 1.44, 1.35, 1.29
Re: Allowing root to display to X...
On Sun, 12 Sep 1999, Aaron Van Couwenberghe wrote: Hi all - I remember hearing once that if you linked your ~/.Xauthority to /root/.Xauthority, a root shell would be able to run X applications (within your regular user's X session). However, I tried this, and it doesn't seem to work for me. Any ideas? I'm trying to install Sun's latest Star Office. This idea is a nasty hack and it should not be used. Instead, try using ssh: 1. Install the ssh package if you haven't already done so (it's in the non-US/non-free section). 2. Add these two lines to either /etc/ssh/ssh_config or ~/.ssh/config (Debian is the only system I know of that has X11 forwarding turned off by default in ssh): Host localhost ForwardX11 yes 3. Inside an xterm, rxvt, eterm or whatever, run this command: ssh -l root localhost Now ssh will automatically take care of the DISPLAY variable and X authorisation. It may seem overkill to use ssh to log in as root on the local system, but it is the only method I've come across that addresses this problem in a nice and clean way. Remco -- rd1936: 2:30am up 2 days, 3:45, 5 users, load average: 1.32, 1.38, 1.51
Re: Remove funny files
On Wed, 7 Jul 1999, Jiri Baum wrote: Hello, Allan M. Wind: On 1999-07-04 20:22, Rolf Edlund wrote: rm -i Backupfiles/--exclude=files.txt There are 2 tricks: 1. rm -i * (say no to everything that you want to keep) No, this doesn't work. The shell expands the `*', and rm is once again left with a filename beginning with two dashes. I know, I have a lot of ex-umsdos directories mounted over samba. This can actually be a very 'funny' command when you are trying to remove files in the current directory with names like '-f' and '-r'. NEVER use 'rm -i *' to delete a single file that has 'special' characters in its name. If you think you need it, you are wrong. And remember, always first use the 'ls' command instead of the 'rm' command to test what your options will expand to. 'rm -r .*' is another famous mistake you don't want to make. Try 'ls .*' to get an idea of what I'm talking about. 2. rm -- FILE That usually works, but not for all commands. Yes, or use 'rm ./FILE', like somebody else suggested. Remco -- rd1936: 1:05am up 27 days, 15:59, 7 users, load average: 1.10, 1.14, 1.14
Re: Smail hosed.
On Sat, 5 Jun 1999, Dan Willard wrote: I've apparently hosed smail. Mail still comes in, but doesn't get put into my mailbox, it just sits in /var/spool/smail/input. I must have killed it while getting a Wascom tablet to work last weekend, played with libs. Can anyone point me in the right direction? Should I just kill smail and use a different mta? I've reinstalled it with no luck, found my old smail config files and set them up. Are you running 'potato'? I am, and I have the same problem. I think the problem is the combination of smail and glibc 2.1. On my system, the user 'mail' has a crontab[1] that runs the 'runq' command every twenty minutes. This command will deliver all e-mail that was left in the spool directory. This not the real solution to the problem, of course. But until the problem is solved, it works for me. Installing another MTA would also be a solution to your problem, of course. The Debian distribution includes several MTAs. Remco [1] run 'crontab -e -u mail' as root to edit it -- rd1936: 9:25pm up 7:07, 8 users, load average: 1.37, 1.31, 1.26
Re: wdm and xdm co-exisitng - is this a bug?
On 20 May 1999, Jan Vroonhof wrote:
Martin Bialasinski [EMAIL PROTECTED] writes:
The thing is, that you could manage :0 with xdm, :1 with wdm and XDMCP
with gdm or such. So the packages don't conflict in a traditional
sense. But looks like all of them try to manage :0 and therefore there
is this mess.
Wouldn't this be solved by having a general S99dm:0 startup script and
having that managed by the alternative system?
No.
Whether {g,w,x}dm manage :0 or not is defined in their own configuration
files. If you want to use the setup described above, you'd have to start
all three of them and make sure that they don't claim the same display
yourself.
Remco
--
rd1936: 12:20pm up 5 days, 3:49, 5 users, load average: 1.23, 1.25, 1.33
Re: xwin
On Sat, 8 May 1999, J Horacio M G wrote:
Brian Servis dixit:
~
~ To stop xdm from starting in Debian 2.1 you have several options:
~
~ 1) remove the xdm package
~ 2) put an exit 0 at the top of the /etc/init.d/xdm script
~ 3) remove the symlinks in rc?.d to xdm with: update-rc.d -f remove xdm
~ 4) comment out the :0 local /usr/X11R6/bin/X line in
~/etc/X11/xdm/Xservers. This will still allow remote XDMCP
~connections since xdm is still running but not managing the local
~display.
Sorry to take back on an old issue.
Since Ctrl+Alt+Fn can switch from X to almost any V.C., and supposing
you can then take on any task under the command line, I'm seeing no real
point in removing xdm or preventing it from starting the x server on
startup.
But still, if I wanted to exit X... how do I do it? The exit menus in WM
take me back to the X login prompt, they don't take me to the command
prompt as they used to do in 2.0.
You don't exit xdm other than by executing '/etc/init.d/xdm stop' as
root, or by killing the xdm process with the 'kill' command.
I think you started X with the 'startx' program before ('in 2.0'), which
starts one X session and then exits. xdm, however, always gives you a new
login prompt after you exit a sesion.
You might want to try hitting CTRL-R at the xdm login prompt. According to
my /etc/X11/xdm/Xresources_0 file, this should 'abort' the display. xdm
will then kill the X server, but it will not exit. If you type
'/etc/init.d/xdm reload' as root after you have done this, xdm will start
the X server again.
Remco
--
rd1936: 1:20am up 6 days, 9:16, 5 users, load average: 1.64, 1.57, 1.81
Re: 2 network cards
On Sat, 8 May 1999, MR wrote: I have two network cards (1 Intel etherexpress 16, 1 NE2000) in a debian system. Both are detected on boot up and both are ifconfig'd. This system will be used to connect my little LAN through IP masq to the Internet. For various reasons, I am only able to have 1 IP, so that is why I am messing around with all this. My problem is that I can not ping another machine on my internal LAN from the linux box. The internal LAN is numbered 192.168.1.x (with the linux box being .1 and the one I am trying to ping .3) eth0 is my external card and eth1 is the internal one. If I set up a route for the internal lan (route add -net 192.168.1.0 eth1) I can ping 192.168.1.1 (itself), but get a ping: sendto: Operation not permitted when trying to ping another machine on the internal lan. Any ideas? I have tried pinging the linux box from the other machine (which runs Win98). I get both activity lights going like they should on the hub and on a card from each computer, so the connection is there. Thanks in advance. Have you also set up the netmask and broadcast address for eth1? In your case, try netmask=255.255.255.0 and broadcast=192.168.1.255. Remco -- rd1936: 1:55pm up 6 days, 21:51, 5 users, load average: 1.10, 1.38, 1.77
Re: what happened to courtney?
On Fri, 30 Apr 1999, Daryl Williams wrote: folks, i am building a new firewall machine and in the process discovered that courteny is no longer available as a debian package? can someone verify this, or did i miss something? also am looking for recommendations on intrusion detection software, if y'all know of any... I just took a look at the package, it's still in the 'hamm' archives. The problem appears to be the license. In fact, there is no license at all. In the file 'DISCLAIMER', it says: # Copyright (c) 1995 # The Regents of the University of California. All rights reserved. The 'All rights reserved.' part means that you are not allowed to copy of modiy it. Therefore it cannot be distributed at all, except by the copyright owner. You can still get it here: ftp://archive.debian.org/debian-archive/dists/hamm/main/binary-i386/net/courtney_1.3-3.deb Install with 'dpkg -i courtney_1.3-3.deb' after you downloaded it. Remco -- rd1936: 4:10pm up 3 days, 5 min, 6 users, load average: 1.58, 1.58, 1.63
Re: dosfsck - Root directory has zero size.
On Tue, 4 May 1999, Nils Rennebarth wrote: On Mon, May 03, 1999 at 06:26:21PM -0400, Arcady Genkin wrote: Steve Lamb [EMAIL PROTECTED] writes: My first guess is that dosfsck cannot check fat32 partitions. Hmm. The man page says nothing about Fat16/Fat32 issues. You will need dosfstools-2.0 (potato) for making and checking Fat32 filesystems. No, dosfstools-1.0-16 also supports FAT32. This was the first version to support it, according to the changelog. dosfsck should automatically detect it. Remco -- rd1936: 4:25pm up 3 days, 20 min, 6 users, load average: 1.38, 1.55, 1.60
Re: lock user account after 3 bad logins?
On Tue, 9 Mar 1999, Eliezer Figueroa wrote: how do I can specify how many times a user can write a bad password before the system disonect that account. And of course after the account is lock, how do I unlock it. Get Your Private, Free Email at http://www.hotmail.com Sorry for the late reply, but I don't think you want this. This is extremely vulnerable to denial-of-service attacks. By just knowing somebody's username, you can block his/her account by trying to log in a few times as him/her with the wrong password. Remco -- rd31-144: 4:00pm up 7 days, 22:24, 6 users, load average: 1.20, 1.30, 1.27
Re: Oracle in dba group
On Tue, 23 Feb 1999, Eric Wayte wrote: I was scanning the archives at www.debian.org and may have a solution for you! When installing Oracle on Solaris/SPARC, you have to edit /etc/group by hand and make sure that the oracle account is listed in the dba group: dba::101:oracle Yes, that's the dirty way to do it. The clean way is: # adduser oracle dba This will add user oracle to group dba. See 'adduser --help' and 'man adduser' for more information on adduser. Remco -- rd31-144: 10:45pm up 1 day, 21:34, 5 users, load average: 1.00, 1.01, 1.00
Re: Which Kernal supports over 6.0 GB HD
On Wed, 17 Feb 1999, Leszek Gerwatowski wrote: On Tue, Feb 16, 1999 at 11:49:26AM -0500, David B. Teague wrote: On Tue, 16 Feb 1999, Person, Roderick wrote: I just bought a 6.4GB but Linux only reads it as 6.0GB, which Kernal do I need to get the full access Roderick: Are you sure that isn't an Unformatted size? I had that happen to me just recently: My HDD was advertised as a 10.4 gig drive, (they disn't say unformatted, but that is what it was. linux fdisk sees it as 9.5 gigs. Formatted capacity. I' sure this is this case. HD discs are 2.0MB unformatted and 1.44MB formatted for PC (1.4MB for Mac). No, this is an entirenly different issue. Using special programs, you can really get 2.0MB (2 * 1024 * 1024 = 2097152 bytes) on a floppy disk. Hard disk vendors generally use 1 GB = 1,000,000 bytes. Calculating real GB's from this gives you: 1,000,000 / 1024 / 1024 = 0.95367432 GB So, drive that is sold as 10GB really is only 9.54 GB. Remco
Re: Kernel: Unable to load interpreter
On Wed, 17 Feb 1999, Chris Kaltwasser wrote: Does someone know the meaning of this message? I've got a machine running hamm that's been doing this fairly often the past couple of days. The hard drive starts thrashing, and just about everything I try to do becomes impossible. Just now I had to do a hardware restart, which gets it working normally again. Could this be a problem with the kernel image or something like that? How much RAM and swap space does your computer have? Unable to load interpreter can mean that your computer is out of virtual memory. If your computer is giving these messages again, try to run the 'free' command and see how much free memory you have. 'ps uax' can give you useful information if you think one particular program is using an excessive amount of memory. Remco
Re: Still the less fortune problem...
On Mon, 14 Dec 1998, Stef Hoesli Wiederwald wrote: case $- in *i*) # interactive shell only echo \n fortune -a echo ;; esac Hm, this gives me an Illegal variable name. error... What about /etc/profile or whichever file is used by your shell as the system-wide profile file? Try to find lines with an odd number of single or double quotes. Remco
Re: SANE install catch-22
On Mon, 1 Jan 1996, Ted Llewellyn wrote: Well, this is very pretty.I'm running Debian 2.0 with the default SANE--the doc says v0.71 but I think it's v0.74. I was going to upgrade to version 1.0. To do that, I have to install the new libc6, and to do that I have to upgrade libstdc++2.8, and apparently to do THAT, dpkg has to unconfigure itself. Since I'm running dselect, how is this supposed to work? Here is the output from dselect: dpkg: considering removing libstdc++2.8 in favour of libc6 ... dpkg: no, dpkg is essential, will not deconfigure it in order to enable removal of libstdc++2.8. dpkg: regarding .../base/libc6_2.0.7u-7.1.deb containing libc6: libc6 conflicts with libstdc++2.8 ( 2.90.29-2) libstdc++2.8 (version 2.90.29-0.6) is installed. dpkg: error processing debian/dists/unstable/main/binary-i386/base/libc6_2.0.7u: conflicting packages - not installing libc6 Can you tell why you'd have to upgrade libc6? Quickly looking at the sane and libc6 packages that are currently in 'unstable' I don't see it. It looks like you are trying to install a new libc6 package which conflicts with old libstdc++2.8 packages. You could try to also upgrade the libstdc++2.8 package to a newer version, but I don't know if that would break other things. Remco
Re: Is this really the right thing to do?
On Mon, 30 Nov 1998, Mitch Blevins wrote: about auto-removing unneeded packages I think the best solution would be to be able to mark packages in dselect and dpkg, just like we currently have them marked as 'purge', 'hold', etc. We would just add a way to mark packages as 'installed-but-not-wanted-on-its-own-merits-so-uninstall-it-when-all- packages-needing-it-are-gone', or IBNWOIOMSUIWAPNIAG for short. Yes, this has been proposed several times before and I don't think anyone would be against such an option. But looking at the huge list of bug reports dpkg has, I doubt anyone with sufficient knowledge of dpkg has the time to implement it. Remco
Re: Debian Netscape Installer and glibc2 Communicator 4.06
On Sat, 5 Sep 1998, Ed Cogburn wrote: Jianbo Zhang wrote: Hi Debians, When I run dpkg -i netscape4_4.0.12.deb, I got some error message: it says that some libc5 and motif related files could not be found. So I installed Communicator 4.06 in my fresh Debian (2.0) Linux box with nsinstall. So far I did not find any problem, but I still do not know why Debian netscape installer does not work for me. Help is highly appreciated. Jianbo The v4 installer assumes it is installing the libc5 version of NS, thus it has dependancies on the libc5 package. We now have libc6 versions of NS, but there hasn't yet been an update to these Deb installers. So you did what you had to do (install NS without the Deb installer). I did the same thing. The worst thing about this seems to me, that the filenames for communinctor-4.5-libc5 and communinctor-4.5-libc6 are the same. The installer script couldn't tell the difference from the filename like it does now. And for the libc6 communicator there has to come a new installer packages anyway, since it depends on other libraries than the libc5 communicator depends on. Remco
Re: running root X programs
On Sun, 6 Sep 1998 [EMAIL PROTECTED] wrote: Another idea, the simplest one I could think of: in your startx script, add the line ``cat ~root/.Xauthority ~/.Xauthority'' just after ``serverargs=$serverargs -auth $HOME/.Xauthority''. First startx as root to create the ~root/.Xauthority file and chgrp it so that only users in a group who locally access the machine can read it. Only problem is the file is re-created whenever root uses startx, but I'd start X as the normal user anyway. Far more easy method: 1. install ssh (from section non-us/net) 2. do ssh -l root localhost in an xterm to become root 3. in this root shell, start any X program to test it Ssh takes care of X11 forwarding automatically. This was done to make it easy to run X programs on a remote host and have them use the local display, but it goes equally well if you connect to the local host as another user. I admit ssh is overkill for this. But if you have enough memory, it is by far the easiest method. Remco
Re: Kernel compile fails on Cyrix...
On Mon, 1 Jan 1996, Michael Beattie wrote: On Fri, 4 Sep 1998, Mike Brownlow wrote: Jeremy Tregunna [EMAIL PROTECTED] wrote: On Fri, 4 Sep 1998 [EMAIL PROTECTED] wrote: Hi everybody, I just installed hamm on a cyrix 686 system from scratch and I have 2 problems: 1. X11 crashes about once in a hour for no apparent reason. Cyrix has a big problem, what you could try and do is underclock your CPU from say 233 for example, to 200, if you still get this problem, it's because Cyrix makes really shitty CPU's and such, and are not really reliable. Hmm, that's funny. Mine works just fine. And it's reliable. Same. Cyrix make cheaper alternatives to Intel chips. The catch is a higher power consumption.. therefore HEAT! (Oh.. and not as good an FPU) Try the 'set6x86' package.. good for enabling the HALT feature.. from the README file: The 6x86 has a set of configuration registers that allows one to do a lot of things, like specifying non-cacheable memory areas (important for e.g. graphics cards), I/O delays, cache write policy (WB/WT), and also to enable an automatic standby mode where a CPU halt instruction cuts down power by a factor of 70 (from 5.8 Amps to 83 milli-Amps for the P133+). Better yet, install a better cooler. If you are compiling a kernel or if you are participating in something like distributed.net's RC5 cracking, the processor won't spend much time on the HLT instruction, so it will still get hot. Remco BTW Michael, your system clock seems to be set to Mon, 1 Jan 1996.
Re: X-windows on a Valuepoint
On Thu, 4 Jun 1998, Alain Toussaint wrote: check that your /etc/X11/xdm/Xservers look like this: # /etc/X11/xdm/Xservers [snip comments in /etc/X11/xdm/Xservers] :0 local /usr/X11R6/bin/X the last line is particularly important,also,check the location of X,it may be different on your system (i'm using the latest unstable version,but i don't think it will be much different on your system). As far as I know, the place for the X server is always /usr/X11R6/bin/X if you have Debian 1.2 or newer (I don't know the older versions). p.s.also notice that there are 2 Xserver(s) file Xserver (found in /etc/X11/) and Xservers (found in /etc/X11/xdm/) it's the later that you want. Yes, but it is important to check what's in /etc/X11/Xserver. This file specifies which X server you want to run. It should look something like this: cut-here /usr/bin/X11/XF86_S3 Console The first line in this file is the full pathname of the default X server. The second line shows who is allowed to run the X server: RootOnly Console (anyone whose controlling tty is on the console) Anybody cut-here Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: How to make X fonts Bigger
On Wed, 3 Jun 1998, Oz Dror wrote: . All you need to do is to get the xfstt package, create a /var/ttfonts directory, copy the fonts you would like to have into it, add copy the fonts from where? from windows? from X11? Do I need a font dir file, if so how do I make it. I believe the /var/ttfonts directory is suplied by the xfstt package. If not, create it yourself. Make a subdirectory with a nice name (like, winfonts or whatever) in /var/ttfonts. Put the *.ttf files in this subdirectory. Then the only thing you need to do is start xfstt and let the X server connect to it. Where you get the fonts is another matter. Rip them from Windows, MS Office or any other program that ships with TrueType fonts. Or get yourself one of thse shareware fonts CD-ROMs. Or download them from some ftp server. disclaimerJust be sure not to break the law or any software license./disclaimer Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: NT and Linux
On Fri, 29 May 1998, Michele Comitini wrote: One great advantage is that you can combine any kind of partitions form different devices (even a combination of partitions from a mix of IDE or SCISI hard-disks!) and have different personalities (i.e. RAID-5 for filesystem partitions, RAID-0 for swap partitions) on partitions of the same hard-disk. Note that you don't need RAID0 to do striping on swap partitions. You can assign each swap partition a priority. If all have the same priority, the kernel (versions 1.3.6 and higher) will automatically use something like striping on them. For more info, see 'man 8 swapon' and 'man 2 swapon' for more info. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: how to get color ls display?
On Tue, 26 May 1998, the lone gunman wrote: Hello: I recently installed Debian 1.3.1 that I purchased from Linux System Labs. I have had several problems, and cannot seem to find the answers. Getting color with ls! I copied my /etc/DIR_COLORS from my old slackware system to my new Debian system. In my /etc/profile, I have dircolors -b run, and then I alias ls=ls --color=auto. Now, if I un-alias ls (that is, only dircolors is run), then NOTHING is in color. But with the alias, regardless of what I do with dircolors, I get my lists in color, but not the custom colors I've defined in /etc/DIR_COLORS. What gives? From info dircolors: `dircolors' outputs a sequence of shell commands to set up the terminal for color output from `ls' (and `dir', etc.). Typical usage: eval `dircolors [OPTION]... [FILE]` If FILE is specified, `dircolors' reads it to determine which colors to use for which file types and extensions. Otherwise, a precompiled database is used. For details on the format of these files, run `dircolors --print-database'. I think this gives enough information. I think you omitted the filename. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Where is killall
On Mon, 25 May 1998, Oz Dror wrote: I have installed the new procps pakage and the killall command is lost. I tried dpkg -S killall and I could not find it. Does any one knows where is it? killall is in the psmisc package. You could also take a look at the skill command, which is in the procps package and has very much the same functionality as the killall command (and even more than that). Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Diety / Apt Questions.
On 2 May 1998, lantz moore wrote: DZM == David Z Maze [EMAIL PROTECTED] writes: DZM More elegant solutions involve using the xauth access-control DZM mechanism. If you are the only person on your system with root DZM access, you can make things work by symlinking root's .Xauthority DZM file to yours. If multiple people have access, you can use xauth to DZM grab X cookies with a command like DZM xauth -f /home/me/.Xauthority extract - $DISPLAY | xauth merge - i find it convenient to add the following to roots .bashrc: if [ $USER != root ] then export XAUTHORITY=/home/$USER/.Xauthority fi Be very careful when doing this. If root ever writes anything to this file using xauth, it will become owned by root and the normal user whose file it really is will not be able to write to it. The easiest solution, though it probably is overkill, is to use ssh to connect to the localhost. If you run sshd, you can simply run 'ssh -l root localhost', type root's password and be able to use the X display. ssh takes care of the X authentication automatically. Originally this feature was intended for remote connections. You can do something like 'ssh remotehost' and have all X programs you run automaticlly use your local display. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: LaTeX installation
On 25 Apr 1998, Joachim Trinkwitz wrote: Luka Pravica [EMAIL PROTECTED] writes: From the installation error-file (from dpkg) obviosly some files are missing, like texmf.cnf, language.dat... I tried searching for those files but they are not on my hard drive. Perhaps all those files get installed with some other package, but dselect didn't show any dependencies. You can do a 'dpkg -S language.dat' etc. On my system ( hamm with tetex 0.9), both files are in tetex-base. 'dpkg -S ...' tells you where the files should be in the filesystem, so you could look, if the files are really there. You should even try the command 'texconfig conf' and 'texconfig confall', which show the configuration of your tetex-system. Yes, the files belong to tetex-base. But somehow the installation script for either tetex-base or tetex-bin (I don't know yet) removes all files in /etc/texmf/ under some circumstances. This was initially intended to clean up incompatible files from old Latex versions, but it also seems to be happening on new installations of Tetex. If I figure out what's really wrong here, I'll file a bug report. Or can somebody else, with more knowledge of perl than I have, look into this? I know the files aren't removed when upgrading between different versions of the tetex-* packages. On how to get the files back: I think the only way is to manually extract them from the package and then fiddle around until you got it right. I got it right several times already, but I can't remember the exact steps. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: securing debian
On Sun, 26 Apr 1998, Chris wrote: Just a point of note: If your brother has physical access to the machine there is no way you can stop him from getting root access. You can increase the difficulty by setting the bios to only boot from HDD and then locking the bios - but if he's smart enough that you have to worry about the root password, he's going to know how to reset the bios. In addition, if you have dos/windoze installed as well, and he can run it, he can most likely reset your bios without even taking the case off (use one of the numerous bios hacking programs around). If it is possible to get a Lilo prompt at boot time or to start Linux using Loadlin, it is much simpler to get root acess than you think. Just give the argument 'init=/bin/bash' (without quotes) to the kernel and you'll get a root shell without questions, because the kernel runs bash (as root) instead of init. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: xconsole question (tkman)
On Sat, 2 May 1998, Adam Shand wrote: is this a bug or have i done something wrong? Have you compiled your own tk8.0 binaries? If not, you could have the wrong version of the tk8.0 package. Did you use any --force flags on dpkg when installing tkman? You need at least version 8.0p2-2 of the tk8.0 package for tkman to work. nope, i haven't compiled tk, i'm using the debian package, and i didn't force anything. standard happy install. ii tk404.0p3-5The Tk toolkit for TCL and X11 v4.0 - ii tk414.1p1-2The Tk toolkit for TCL and X11 v4.1 - ii tk424.2p2-4The Tk toolkit for TCL and X11 v4.2 - ii tk8.0 8.0p2-4The Tk toolkit for TCL and X11 v8.0 - ii tkstep8.0 8.0-1 The NEXTSTEP(tm)-like version of the Tk ii tkman 2.0.6-2A graphical, hypertext, manual page Weird. It works for me. ii tk8.0 8.0p2-4The Tk toolkit for TCL and X11 v8.0 - Run-Ti ii tkfont 1.1-3 A different xfontsel for displaying fonts. ii tkinfo 2.0-1 tk/tcl info browser ii tkman 2.0.6-3A graphical, hypertext, manual page browser. Are you sure the 'wish' program from tk8.0 is called? I think you can actualy remove the tk4* packages if your installation is up-to-date with hamm. Just try it and see if dselect complains. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: xconsole question (tkman)
On Fri, 24 Apr 1998, Adam Shand wrote: Have you tried tkman? still, good old man is great :) just installed the debian package and now i get this message... badger(larry) tkman Elide patch not installed in /usr/bin/wish8.0. You must apply the elided text patch to Tk and rebuild wish. See the Makefile for more information. is this a bug or have i done something wrong? Have you compiled your own tk8.0 binaries? If not, you could have the wrong version of the tk8.0 package. Did you use any --force flags on dpkg when installing tkman? You need at least version 8.0p2-2 of the tk8.0 package for tkman to work. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: PINE Debian Package
On Thu, 23 Apr 1998, George Bonser wrote: On Thu, 23 Apr 1998, Rev. Joseph Carter wrote: Are you aware that among other things the patches to pine added since the last binary package was released include things which are not merely configuration but are purely bug fixes, feature enhancements (maildir patch comes to mind) and other things along that line? UoW was asked about these things and they said they didn't want binaries of unapproved patches. If it was any other package, I would not have said anything and I waited for someone else to bring it up before venting my frustration. Pine/pico is the one package that you can not expect the user to build because chances are good that they can't. I use pine over telnet and never use pico but it is on the system and always will be. This is a very special application, it is usually one's first mailer and editor. Yes, this is very unfortunate. But the upstream authors really do not want to change the license, so Debian has no choise. And the user can be given very specific intstructions or even a script that will make it very easy to compile and install pine. In this sort of situation, I think a -src .deb file is a good thing for pristine source with which one can apply Debian patches. Does anyone else think so? Currently the only things I know of in this category are qmail and pine. Netscape can be put in .deb now and I don't think you can distribute rvplayer.. qmail is completely different. The author specificly disallows binary distribution of any kind. University of Washington makes no such demands. The non-patched pine binaries violate Debian Policy. It is illegal to distribute patched pine binaries. These two combined leave distributing patched source as the only legal way to distrubute a version of pine that complies with Debian Policy. Because someone asked and the UoW clarified that they didn't want patched binaries if they didn't pre-approve the patches. The maintainer didn't like that idea. Hmm. I can understand that from the maintainers ego standpoint but if I owned pine, I might want a look at those patches too in order to see if there is anything that should go into the mainstream distribution and to see if someone was hacking backdoors to reading other people's email into the program. In any case, if UofW specificly said that they want to pre-approve patches to a program they own and Debian thinks that is not acceptable, there is no choice. I can understand their concern considering privacy issues. Now if someone's mail is hacked on a Debian system and it is found to be the fault of Debian's patches, it is Debian that gets hauled into court where under UofW's method, Debian would have had some protection since the patches would have been approved by UofW. I guess the same exposure would apply to any patches to any email system supplied by Debian with patches to the source. If someone's mail gets hacked, it is the fault of the hacker, not of anyone in Debian. Suppose you could sue Microsoft every time a Windows system is cracked. Or any commercial OS, for that matter. You can't. And you can't sue Debian [1] if a Debian system is cracked. Remco [1] You can't sue Debian at all, actually, since Debian is not incorporated in any way so legally Debian doesn't even exist. You'd have to sue the indivudual maintainers. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Fonts in X [Off Topic]
On Wed, 22 Apr 1998, Stephen Carpenter wrote: Luiz Otavio L. Zorzella wrote: Remco Blaakmeer writes: It is probably illegal to use the fonts that come with Windows. Why? If you bought a Windows license, you bought a license for the Windows components, I think. That sounds right to meIt is of course illegal to distribute the font thatcome with windows, well...I havn't actually looked to see if the windows fonts are available on any sort of Free licence... but... What are the chances of Micro$oft using any sort of free licence for anything? Not being legal to use Windows fonts in Linux would be like being illegal to buy Coke and using the jar to carry Oranje Juice... I think a better analogy might be You buy a coke bu tbecause of thelicence agreement you have to drink the whole thing, you can't dump half of it out even if you don't like its flavor To let the X server use the fonts, you can do: $ xset fp+ tcp/localhost:7100 you can also edit /etc/XF86config and add the fontpathunix:7100 (I think thats it don't have the docs in front of me... in any case it is in the xfstt docs from the tarball on sunsite) NB: if you do this..and reboot your machine you MUST run xfstt ...if you do not have xfstt and you have this line in your XF86Config file...then X will refuse to start and exit with an error Yes, I said that in the same post where I mentioned the xset command. And BTW, you are right where you say that xfstt *must* be running if you add the line to XF86Config. I just tried it. If a _directory_ in the fontpath is missing or otherwise invalid, it is deleted from the fontpath. But if a font server is not present, this is considered a fatal error and the X server refuses to load. If you are using xdm to start X and the X server refuses to start, you'll end up with a flashing screen caused by the X server being started over and over again. If you ever encounter such a problem, you can solve it by stopping xdm through a telnet connection or trying to hit Ctrl-Alt-Del at a moment where the X server has stopped and isn't yet being restarted and then booting into single user mode. If nobody objects to it, I'll file bug reports for the following annoyances I have encountered when X is misconfigured: - If the X server fails to start, xdm seems to be restarting it indefinitely. I'd say that if the X server is restarted too often too fast, it should be disabled because it is obviously not working. Just like init disables processes that are respawning too fast. - If there is a non-existing font directory in xfs's config file, this is a fatal error and xfs doesn't start at all, which causes my X server to fail because it can't find its fonts, which is very annoying because I use xdm. - If there is a non-existing font server in the FontPath of an X server, this is a fatal error and it refuses to start, which is annoying like I said above. Of course, if the X server can't find the font 'fixed', it should still fail to start. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Fonts in X [Off Topic]
On Wed, 22 Apr 1998, Luiz Otavio L. Zorzella wrote: Remco Blaakmeer writes: You can place the fonts in a subdirectory of /var/ttfonts . You can use any name for that subdirectory. Where you get the fonts is another issue. It is probably illegal to use the fonts that come with Windows. Why? If you bought a Windows license, you bought a license for the Windows components, I think. Not being legal to use Windows fonts in Linux would be like being illegal to buy Coke and using the jar to carry Oranje Juice... Also, if you bought any program that comes with true type fonts (like Corel Draw), you could use them also! I just meant that I didn't know for sure. I can't remember actually reading a license for a commercial application. You could get yourself a CD with a few hundred shareware/freeware fonts, but most of these are very poorly licensed. What do you mean by very poorly licensed? They have no license or the license is ill-formed or very ambiguous. You get the idea. xfstt isn't yet automatically started at boot time. To start it manually, do: # xfstt ... as root. To let the X server use the fonts, you can do: $ xset fp+ tcp/localhost:7100 Thanks for the clues! I finally could see some nice fonts in Netscape... Great! Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Fonts in X [Off Topic]
On Wed, 22 Apr 1998, George Bonser wrote: On Thu, 23 Apr 1998, Remco Blaakmeer wrote: They have no license or the license is ill-formed or very ambiguous. You get the idea. The gist of any Microsoft licenses I have ever read have been along the lines of: This software is owned by Microsoft, not you. You have purchased the right to run it on one machine and make one copy for backup purposes. Yes, but this remark was about the fonts on those font CDs you can buy everywhere for little money. I bet any Microsoft program has a well-formed license. MS has got enough lawyers to come up with good licenses. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: why?
On Tue, 21 Apr 1998, C.J.LAWSON wrote: What? Do they forbid sending e-mail with bogus To: headers? I am once again baffled by the stupidity of USA laws regarding electronic (1) Why would anyone intentionally want to send email to a bogus header? to wreak havoc?? ... to bounce spam messages??? (2) I wish you knew how much hassle bogus headers cause on servers But many read-only mailing lists (e-zines) are using a bogus To: header. They work because in the smtp 'RCPT TO:' command they use the actual (working) address of the recipient, but the To: header is still bogus. I now know that it is not bogus headers that are forbidden, but I would still think of it as very stupid if any state or nation would forbid them just because they are bogus. And yes, I think I know how much hassle bogus headers can cause on mail servers. Just think of all those people with some sort of -nospam- in their return address. And if they really want to use it, they should put it in the domain part and not in the username part of the address, because an invalid domain causes much less hassle than an invalid username. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Fonts in X [Off Topic]
On Wed, 22 Apr 1998, Stephen Carpenter wrote: On Wed, 22 Apr 1998, Luiz Otavio L. Zorzella wrote: Yeah! Is there a good reason why xffst is not started in /etc/rc.boot (or something)? I can think of a reason or two...no not really... there is a problem with trying to stop it (without using kill-all) and you can't sync it whjile it runs... other than that...no I think it should be started on startup start-stop-daemon uses a pid-file that stores the PID of the running daemon so it knows which program to stop later. And 'xfstt --sync' causes xfstt to sync with the fonts in /var/ttfonts. So there is not really a problem here. BTW, can you help me in (finally) better understanding the True Type files? I will try :) There are *.ttf True Type FOnt *.fon Old Style bitmapped font *.ttr *.for NFI on these two Another one: *.fot Used by Windows 3.x to store some information about TT fonts. There should be one for every *.ttf file. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Fonts in X [Off Topic]
On Thu, 23 Apr 1998, Stephen Carpenter wrote: Luiz Otavio L. Zorzella wrote: Remco Blaakmeer writes: If nobody objects to it, I'll file bug reports for the following annoyances I have encountered when X is misconfigured: I don't know if I agree witha bug report against X...as the keyword there is misconfigured...although you could look at that as a bug in xdm but... there is a such thing as misconfigured...reminds me of when I was a tottal linux newbie I was running with RedHat in runlevel 5 (init starts XDM on startup) and I recompiled my kernel for the first time...forgot to compile in a mouse driver xdm didn't like that at all... needed to use a kernel boot parameter tofix it Yes, that's exactly my point. If you have accidentally or ignorantly misconfigured X (or haven't configured it at all, for that matter) and start xdm, your computer becomes inaccessible without either remote access or a reboot with some not-guessable kernel parameters. This is VERY confusing for newbies, who will really fast think something like well, this whole Linux thing obviously doesn't work for me, so I'll just turn it off and return to Windows. Now, that is one thing we don't want, do we? Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Zen and the art of working with NT
On Mon, 20 Apr 1998, Stephen Carpenter wrote: I have a linux machine here which is low on hard disk space my idea was to mount my Windows NT home directory here at work and use it for extra space (I can up my spcae by 2 gigs that way) o far I have mounted my home drive \\home5\sjc8$ on /mnt/partners with smbmount because o fmy specific problems I was trying to do the following: create a second extended filesystem in a file on the NT home dir and then mount that filesystem (I need a filesystem that will hole my files and care about case and permissions I am planning on using it to image CDs to burn (hamm hamm hamm :) ) why wont it work? I can make an e2fs filesystem but when I try to moungt it anywhere I get errors I tried the following example cd ~ touch test.e2 Here, make sure that the file is actually the size you want it to be before you use mke2fs on it, like this: $ dd if=/dev/zero of=test.e2 bs=1024k count=500 This will create a 500 MB file filled with 0x00. mke2fs -F test.e2 500 (it makes a filesystem) mkdir testmnt mount test.e2 -o loop testmnt and it moutns! then... cp test.e2 /mnt/partners mount /mnt/partners/test.e2 -o loop testmnt and I get the errors loop: block 1 not present and a bunch of others then I tried cd /mnt/partners mkdir testmnt and same deal anyone tried this? can it work? what sthe problem? The problem is that block 1 is really not present. The dd command above will make cause it to be present. I don't know if this also works if the test.e2 is on a Samba filesystem, but you could of course give it a try. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Xdm
On Tue, 21 Apr 1998, Michael Beattie wrote: I've just started having a look at Xdm, and have decided I would like to use it if I can do two things. 1) How do you add another display (:1), I had a look at the Xservers file, and added a new line, but it died... In the latest xbase package, the default Xservers file contains these commented lines: # If you want to add servers yourself, please keep these things in mind: # - For each X server other than :0, the display number must be on the # command line # - This is the place to add the -bpp option to the command line # - If you run multiple X servers, it is wise to add the vtXX option to # each of them, since this is currently the only sure way of knowing # in advance which X server attaches to which console. # # Example: # :0 local /usr/X11R6/bin/X :0 vt9 -bpp 16 # :1 local /usr/X11R6/bin/X :1 vt10 -bpp 8 This should be clear enough, IMHO. Heck, I wrote these lines. :-) 2) Can it be made to start in the background, so that on boot up, The first console is still showing? Well, not without a kludge. When the X server starts, it always pops up the X display. You'd need to change the X server if you want it to just start and not show up initially. But you can add a command like this to the start) section in /etc/init.d/xfs : ( sleep 10 ; chvt 1 ) This will change the screen back to console #1 ten seconds after xdm is started. Or you could add a simple 'chvt 1' command to /etc/X11/xdm/Xsetup_0 . But then the chvt command is also executed every time you exit the window manager and get the xdm login screen again. Hey, I like having my cake and eating it too... :) Yep, me too. :-) Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: I now HAVE PINE Debian package.. how do I install it?
On Mon, 20 Apr 1998, Kenneth F. Ryder III wrote: What a interesting thread my PINE question created... OK I have found a Debian package for pine, pine3.96L-2 the file is pine_3_9.deb its located on the root of /dev/hda7 , a MSDOS partition on my drive, which is not mounted by default, so if I use it in Linux I need to mount it. I want to install this version of PINE, I would prefer using dselect to make sure I have all the needed support files etc., so HOW DO I DO IT? BTW: I found this package by going to the Debian site, and searching the stable packages OK, can some one tell me how to install the thing? I have not had much success.. Thanks for your time / help Once you have mounted the DOS partition so that you can access the .deb file with Linux, just issue dpkg -i pine_3_9.deb This will try to install the .deb file. If you don't have a package that pine depends on, dpkg will complain, tell you what package is needed and fail to install pine. You can then use any method to install the needed package. If you want to find ou on which packages pine depends, use this command: dpkg --info pine_3_9.deb Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Fonts in X [Off Topic]
On Fri, 17 Apr 1998, Paul Rightley wrote: I have installed xfstt, but it seems to be documentation-poor. I know nothing of fonts in X (or in Windows for that matter - I guess I am the prototypical luser). Where would I get some TTF fonts and where would I put them to use them? Are there fonts of this sort that follow the DFSG (which I would prefer)? You can place the fonts in a subdirectory of /var/ttfonts . You can use any name for that subdirectory. Where you get the fonts is another issue. It is probably illegal to use the fonts that come with Windows. You could get yourself a CD with a few hundred shareware/freeware fonts, but most of these are very poorly licensed. xfstt isn't yet automatically started at boot time. To start it manually, do: # xfstt ... as root. If you also run xfs, you'll have to put one of these font servers on another port than the default (which is 7100). 'xfstt --help' and 'xfs --help' are helpful here. But if you say you don't know anything about fonts, you probably don't run xfs. To let the X server use the fonts, you can do: $ xset fp+ tcp/localhost:7100 ... as any user that currently has access to the X server (i.e. you). This only works if the X server is already running and must be done every time the X server is restarted. If you want the X server to automatically use the fonts, you can add this line to the Files section in /etc/X11/XF86Config : FontPath tcp/localhost:7100 If xfstt isn't running when X is started, the X server will not be able to use the TTF fonts until you issue the xset command above. I say again here, I don't know where GIMP gets its fonts. But if it gets them from the X server, this is the way to get GIMP to use TTF fonts. Also, note that xfstt is in an experimental state. It is not yet stable and may crash or do things you don't expect now and then. But you could be lucky and it may just work for you. If you have any more questions, just ask. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Residual X11R5 and a.out errors?
On Sat, 18 Apr 1998, Britton wrote: While trying to install or remove X programs with dselect, I often get whole slews of errors like these: ldconfig: warning: can't open /usr/lib/i486-linuxaout/libdb.so.1 (No such file or directory), skipping ldconfig: warning: can't open /usr/X11R5/lib/libXt.so.3 (No such file or directory), skipping ldconfig: warning: can't open /usr/X11R5/lib/libXaw.so.3 (No such file or directory), skipping ldconfig: warning: can't open /usr/X11R5/lib/libX11.so.3 (No such file or directory), skipping ldconfig: warning: can't open /usr/X11R6/lib/i486-linuxaout/libXt.so.6 (No such file or directory), skipping ldconfig: warning: can't open /usr/X11R6/lib/i486-linuxaout/libXaw.so.6 (No such file or directory), skipping ldconfig: warning: can't open /usr/X11R6/lib/i486-linuxaout/libXIE.so.6 (No such file or directory), skipping ldconfig: warning: can't open /usr/X11R6/lib/i486-linuxaout/libX11.so.6 (No such file or directory), skipping They all seem to relate to things that May have existed in 1.2 (even 1.3) for back compatibility but are presumably phased out now. My system has been upgraded in place from 1.2, could these errors be holding over from then? Is there something I should do to fix them? These are most likely symlinks pointing to removed (obsolete) library files. Try running ldconfig manually as root. If it still complains, check that the symlinks are pointing to a file that doesn't exist and remove them. They don't cause any harm if you leave them, but you can safely remove them if you find them annoying. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: no popclient?
On Mon, 20 Apr 1998, Richard L Shepherd wrote: I see that popclient has been missing for sometime. I have a machine with popclient installed from some time ago (early days of bo). Just recently it has stopped working and I see no alternative package to get either the binary (popclient) or the equivalent (lightweight) functionality. The error I get when trying to pop off ANY server via my (rather old) popclient binary now is: popclient: querying mailserv.waikato.ac.nz at Mon Apr 20 15:23:43 1998 popclient: openuserfolder: open(): No such file or directory Now all my local mail readers work fine so there's no reason I know of why it's not happy anymore. The only shared lib it uses is libc5 (which I have). Any ideas on how to fix this or what else to use? Some time ago, when the popclient program was enhanced to support more protocols than just POP, the name was changed to fetchmail. fetchmail is in bo (stable) and also in hamm (frozen). Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: why?
On Mon, 20 Apr 1998, George Bonser wrote: To: [EMAIL PROTECTED] Why would any list accept an email message with the above in the To: header? Also, someone in the US State of Maryland might want to consider donating a mail server for Debian. That email would have been grounds for criminal charges had the server been located in Maryland. What? Do they forbid sending e-mail with bogus To: headers? I am once again baffled by the stupidity of USA laws regarding electronic communications. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: What libs to use ???
On Mon, 20 Apr 1998, Mario Filipe wrote: Hello This is not really a debian problem, but i think that the people here can help me. I need to create a program that shows how much of it's task has been done (in percentage maybe) and what is it doing currently. I could display this with printf's but that would create a scroll effect that i do not want. So my question is if there is any library (for C) that allowa's me to display the whole thing on the same place, without the scrolling efect. I thought that maybe readline or (n)curses would do! I would also apreciate if you could point to the docs of the librarys that you reccomend. Thank you and sorry for the off-topic question, and please cc directly to me. Thanks! I don't know how to do this, but 'ftformat' from the ftape-something package uses something like this to tell the user which track it is formatting. 'ldd ftformat' shows that it uses only libc.so.6 and ld-linux.so.2 , so I guess standard C can do this for you. You could try printing a backspace (^H) and no newline (LF). Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Fonts in X [Off Topic]
On Mon, 20 Apr 1998, Stephen Carpenter wrote: I read this now and this reminds me of a point I have meant to bring to someones attention... the debian xfstt package soes not contain the original xfstt documentation! I installed it recently (again) and actually had to download the package off of sunsite and unpack the tarball just to get the docs on how to tell X that I hada font server running! hmm...if the lack of a man page is considered a bug (I read that somewhere) would this also fall into that catergory? the docs included in the deb (unfortunately) don't even ell how to make the install usefull :( Yes, this would count as a bug in the package. But my impression from README.debian is that the maintainer doesn't really want to spend too much work on it, so probably this package is practically, though not officially, orphaned. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: smail Hostnames--Simplified
On Sat, 18 Apr 1998, Art Lemasters wrote: Allow me to give an example to be more specific (sorry). If my machine has the fullname newname.bf.org can I still download mail from the account [EMAIL PROTECTED] to [EMAIL PROTECTED] using smail? How? Thanks. You mean you have a POP-3 account form your ISP and you want that mail to end up in your mailbox on your own Linux computer? Sure, you can do this, but not with smail alone, since smail doesn't know the POP-3 protocol. Install fetchmail. This will get the mail from your ISP (the [EMAIL PROTECTED] address) and tell smail to send it to [EMAIL PROTECTED] Here's an example .fetchmailrc : poll mail.ipa.net proto pop3 user alemas pass XX is alemas mail.ipa.net : your ISP's mailhost pop3 : the protocol name alemas : your user name at your ISP XXX : fill in your actual password for your ISP alemas : your local username .fetchmailrc must be readable for you only : 'chmod 600 .fetchmailrc'. Then, just issue the 'fetchmail' command with no options and it will get your mail. For more information, please read the manual pages that come with the package. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Software RAID configuration in hamm install?
On Thu, 16 Apr 1998, Remco Blaakmeer wrote: On 15 Apr 1998, Jens Ritter wrote: Remco Blaakmeer [EMAIL PROTECTED] writes: ...but... you have make a mistake ;-) in hamm mdutils has been replaced by raidtools Yes, silly me. I knew that, I just didn't think of it. I always though raidtools is for 2.1.XX and mdutils for 2.0.xx Am I wrong? I think so. I couldn't find any requirements for 2.1 kernels in the raidtools package. From /usr/doc/raidtools/readme.txt.gz: = This is the release of the MD tools + RAIDtools. This merges the utilities required to support for RAID-0 and Linear modes and the RAID-1 tools. = The mdutils package only supports Linear and RAID0 modes, IIRC. OOPS. The above is not true. You need kernel 2.1.63 or newer for the raidtools package. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Setting system time?
On Fri, 17 Apr 1998, Mark Phillips wrote: tzconfig - sets your timezone date - sets your system clock hwclock - sets your hardware clock xntp - keeps your system time correct by reference to Internet timeservers What is the difference between hwclock and clock, between xntp and netdate? And which package contains hwclock and xntp? hwclock is just newer than clock and replaces it. netdate and xntpd use different protocols for communication with the time servers. The protocol used by netdate gives the time with an accuracy of about one second, the ntp protocol is more accurate. Another big difference between netdate and xntpd is that netdate sets the clock once while xntpd is a daemon that connects to the server at regular intervals and tries to correct drift even when not online. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: X11R6.4
On Thu, 16 Apr 1998, Oz Dror wrote: Is anyone going to port the X11R6.4 as a non-free package? I don't think so. X being big as it is, you'd need a large group of developers to maintain and develop an XFree86-like port of X11R6.4. Those people could spend their time much better and would probably spend their time rather on the free X11R6.3 than on the non-free X11R6.4. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: NiC Cards
On Fri, 17 Apr 1998, Mike Holliday wrote: Does anyone know if there is a .DEB package for a 3Com Fast Etherlink 10/100mb bus-master pci adapter? You don't need a .deb package for that. The driver is in the kernel. If it is not in your kernel, you should compile your own kernel to get support for the card. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: NiC Cards
On Fri, 17 Apr 1998, Mike Holliday wrote: Thats fine and all, but seeing as I don't have any driver codes from 3Com how would I go about writing a driver for it? I thought the driver code is already in the standard kernel source code. You only need to turn on the support for your card, check that all other configuration matches your hardware, compile the kernel and boot it. So, what is your real problem? Don't you know how to compile and install your own kernel? Do you have a compiled kernel with the right driver that still doesn't work with your card? Or what? Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Fonts in X [Off Topic]
On Fri, 17 Apr 1998, Paul Rightley wrote: I have been using gimp to produce some overheads lately (in the hope that I can finally rid myself of the necessity of booting into and Windows product). However, when I create large text (like 50 pixels high), it looks pretty pixelated - even if the image itself has a much higher resolution. Is there a way to make smoother text? Since gimp seems to use the X fonts, maybe I just need a scalable (?) font in X? I know nothing about adding fonts to X (that are not conveniently provided by a debian package). Thanks for any assistance (even if it is just the fm to r), Paul P.S. Thanks for the outstanding debian distribution, and the support found on this list. I don't know where GIMP gets it fonts, but if it gets them from the X server I think I have a solution for you. Take a look at the xfstt package. This is a font server for X that handles TTF fonts. With it, you can have all your favorite Windows fonts on your X display. Try it and see if it works for you. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: procmail question
On Tue, 14 Apr 1998, Art Lemasters wrote: Hi all there! I just set-up my .procmailrc (Many thanks to Oliver Elphick). But I have big mailbox (/var/spool/mail/) I want to process w. procmail. Can anybody tell me how? Try the procmail and procmailex (examples) man or info (better yet) pages. They are some of the better-written documents I've seen! Oh... and read the /usr/doc/procmail directory (.gz files with the zless command) first! The examples in the procmailex page should be most helpful, then more specific questions will be easier to get answers for. You'll do fine! A tip for people who are still using zless: If you have already upgraded your system to hamm, place this line in your .bash_profile: eval $(/usr/bin/lesspipe) /usr/bin/lesspipe is a filter for less that decompresses .gz files and gives you useful information about a lot of other file formats (among which .tar and .deb). Look at the script itself for details. If you have Debian 1.3.X or earlier I can send you this script. It is not large and quite useful. I don't know which version of less is required, but I know it works with the less package in Debian 1.3. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: How can I add the 3rd ethernet card ?
On Wed, 15 Apr 1998, Stellar R'espree wrote: Hi, I use a PC linux machine with 2 ethernet cards. (one for eth0, the other for eth1). And now, I want to add third ethernet card. It is recognized as eht2 at my machine, but when I tried to following command, /sbin/ifconfig eth2 123.123.123.123 broadcast 123.123.123.255 netmask 255.255.255.0 it prints following error message. SIOCSIFFLAGS: Try again Do you know what the problem is ? I don't know if this has anything to do with your problem, but you could change the network nubmer. According to the IP-Masquerading mini-HOWTO, these IP numbers are reserved for private networks: 10.0.0.0- 10.255.255.255 172.16.0.0 - 172.31.255.255 192.168.0.0 - 192.168.255.255 Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: A few things about Debian Hamm
On Wed, 15 Apr 1998, Daniel Martin at cush wrote: Stephen Carpenter [EMAIL PROTECTED] writes: I don't believe that is true I am using pine under hamm obn my pc here at work well..I have it installed...due to problems with firewall issues and DHCP I can't seem to get mail working but... I downloaded pine as a binary for hamm I saw the deb of hamm in hamm/non-free/binary-i386/mail -Steve There may have been a pine binary .deb for hamm before people got the license figured out. The fact is, the University of Washington forbids distributing modified binary versions of pine. (There was a binary .deb in bo because the license isn't perfectly clear that that's forbidden, but if you look for it, it's there) Yes, this is highly annoying. Yes, many people have tried to get them to change it, with no success. I know many people annoyed at pine's non-freeness have chosen to go with mutt - you may want to look at that. Is there some 'mutt for pine users' documentation available somewhere? In particular, I very much like the 'search for new mail in all incoming mail folders' function of the TAB key in pine. Does mutt have something similar? Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Faster swap by using separate disk?
On Wed, 15 Apr 1998 [EMAIL PROTECTED] wrote: Remco Blaakmeer writes: The second thing that you can do is to move up to the new ultra-DMA IDE drives. The bandwidth (bytes per second) is much higher than the standard IDE drives and will speed up Linux as a whole. If you want to really speed up your hard drives, switch to multiple SCSI drives. Period. Yes, to a point! If one is to go all the way, then make it wide fast SCSI3 drives which spin at 10,000 rpm and use a PCI based RAID SCSI controller with _many_ of these GIGAbyte drives! Did anyone mention myriabucks? (one myriabuck = $10,000) 8-) The newer SCSI controllers will handle up to 15 drives per cable. 8-) (not to mention the startup current blanking out half a city and the resonances sounding like a jet plane taking off 8-) ) Just joking! In all due seriousness, while an IDE drive will never match a decent SCSI setup, IDEs are less expensive and ultra-DMA IDE drives, IMHO, represent the most bang for the buck (performance for the money) Well, when I have some bucks to buy more bang in the future, I plan to buy some relatively small SCSI disks, plug them into an el-cheapo NCR SCSI card and throw in software-RAID0. This is my idea of a few more bucks, a lot more bang. I agree on the plane issue, though. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Software RAID configuration in hamm install?
On 15 Apr 1998, Jens Ritter wrote: Remco Blaakmeer [EMAIL PROTECTED] writes: ...but... you have make a mistake ;-) in hamm mdutils has been replaced by raidtools Yes, silly me. I knew that, I just didn't think of it. I always though raidtools is for 2.1.XX and mdutils for 2.0.xx Am I wrong? I think so. I couldn't find any requirements for 2.1 kernels in the raidtools package. From /usr/doc/raidtools/readme.txt.gz: = This is the release of the MD tools + RAIDtools. This merges the utilities required to support for RAID-0 and Linear modes and the RAID-1 tools. = The mdutils package only supports Linear and RAID0 modes, IIRC. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Copyright Question.
On Wed, 15 Apr 1998, Petra Kevin J Poorman wrote: -BEGIN PGP SIGNED MESSAGE- Hi, I want to package Blender. It will have to go into Non-free. or maybe contrib... What I want to know is yalls thoughts on weather or not we can even distribute this software. Heres the pertinant part of the licence. Permission to use, copy, and distribute this software and its documentation without written agreement is hereby granted only for non- commercial purposes. Distributing blender 'bundled' in with ANY product is considered to be a `commercial purpose'. This entire copyright notice must appear in all copies of this software. The part that worrys me is the line about ... Bundled in with any product. I'm wondering if this would mean packing for debian is comercial? Please try to contact the authors about including it in the Debian distribution. Apart from the 'bundled' in with ANY product, the rest of the license (only for non-commercial purposes and not allowing modification) would definitely make it go into non-free. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Setting transparent xterms
On Thu, 16 Apr 1998, Jonas Bofjall wrote: On Tue, 14 Apr 1998, Klas Lindberg wrote: Last I heard, transparent backgrounds were not possible (but this may or may not be true now, as I'm using X less and less...) But RxVT, which is Take a look at the enlightenment wm (www.enlightenment.org if I'm not mistaken). It implements just juch a feature and is a pretty pretty wm I am sorry but it does not. This question is so common its even made it to the Enlightenment FAQs! It is not possible with X (yet), without using enourmous amount of computing (first getting screen, then transforming, then using as a background pixmap). Using semi-transparent windows will not be practically possible until we have a system using alpha-channels. Berlin (berlin-consortium.org) will be such a system... And what about the kind of transparency that is used by programs such as xeyes? Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Where did my init.d/boot go?
On Tue, 7 Apr 1998, Mirek Kwasniak wrote: On Mon, Apr 06, 1998 at 11:42:50PM -0500, [EMAIL PROTECTED] wrote: Since hamm was frozen, I started the process of updating my bo system. I've been running hamm now for a week or so, and using FTP to update all of my out-of-date packages. Last night I installed some of the new packages, ( of which the only important one I remember was sysvinit ), and when I rebooted my system, I was no longer able to open a tty, and thus even though I logged in, and Linux validated my password, it was unable to give me a shell. After examining the boot messages, and seeing a line that said 'unable to find /etc/init.d/boot', I booted up from a beautiful little slackware rescue disk, and saw that my boot was now boot.OLD. How do I find out what package did this? I assume that whatever package moved my boot to boot.OLD, was intending to write a new file. Do I need this new boot file? My system seems to be running fine w/ my old boot. Your install is broken by buggy grep 2.1-6. I got the same result but I upgrade a copy of my system :) Get grep 2.1-7 and maybe reinstall all packages by hand (whith: dpkg -i). No, these are two different problems. grep_2.1-6 is severely broken and should be replaced by a newer version. The boot.OLD file is not a bug or an error, it is intended behaviour. The /etc/init.d/boot script has bee split up into several scripts. Take a look in /etc/rcS.d/ to see what it run in which order ar boot time. If your sysvinit is really broken, you can boot Linux directly into bash. At the LILO boot prompt, type linux init=/bin/bash This will load bash instead of init, so you get a root shell with no questions asked. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Software RAID configuration in hamm install?
On Wed, 8 Apr 1998, Ulisses Alonso Camaro wrote: Hi all, I'm novice to RAID, and I want to try with it. I would like to know if hamm provices in its installation the tools to setup a RAID0 (stripping) installation. Also I would like to know if the menu for installation has RAID0 configuration options Thanks in advance, Ulisses PD: Does debian.org ships hamm? Sorry for the late reply, I am catching up with the mailing list. Yes, hamm has tools for RAID0. You'll want the mdutils package, which contains all tools you need. I don't know about the standard kernel, you might have to compile your own with support for MD and RAID0 turned on. I don't think there is a menu that configures the whole thing for you, but it is not difficult to do it by hand. Basically, you do this: 1. get a kernel that supports RAID0 (and boot it). 2. instal the mdutils package. 3. choose the partitions (or whole disks) that will form the RAID0 device. for example, say these are /dev/hda1 and /dev/hdc3 4. run these commands: # mdcreate raid0 /dev/md0 /dev/hda1 /dev/hdc3 # mdadd -ar # mke2fs /dev/md0 5. mount the RAID0 partition where you want it. Note, that it will be difficult to get the root filesystem on the RAID0 device since the RAID0 device is only active after the mdadd program is run (Debian does this automatically at boot time if you have configured any MD devices). This is approximately how my setup used to be when I had a RAID0 device: DOS partition: /dev/hda1 extended partition: /dev/hda2 root file system: /dev/hda5 RAID0 system: /dev/hda6 and /dev/hdc2 swap: /dev/hda7 and /dev/hdc1 RAID0 device mounted on /mddrive /mddrive contains home, usr and var root filesystem contains symlinks for /home, /usr and /var For more (detailed) information, look at these documents on a Debian system: /usr/doc/HOWTO/Root-RAID-HOWTO.gz /usr/doc/HOWTO/mini/Software-RAID.gz and the man pages for mdadd, mdcreate, etc. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Stopping XDM
On Fri, 10 Apr 1998, Shaleh wrote: All daemons in Debian can be stopped by calling the start-stop-daemon. An easier way is to look in /etc/init.d and call its script. for XDM it is /etc/init.d/xdm stop (start would restart it). Yes, and change the line 'start-xdm' in /etc/X11/config into 'no-start-xdm' if you don't want xdm to be started at boot time. But if you want to stop xdm, you must do this before you change this file or the /etc/init.d/xdm script will not stop xdm. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Software RAID configuration in hamm install?
On Wed, 15 Apr 1998, Ulisses Alonso Camaro wrote: -BEGIN PGP SIGNED MESSAGE- Hello Remco! On Wed, 15 Apr 1998, Remco Blaakmeer wrote: On Wed, 8 Apr 1998, Ulisses Alonso Camaro wrote: Hi all, I'm novice to RAID, and I want to try with it. I would like to know if hamm provices in its installation the tools to setup a RAID0 (stripping) installation. Also I would like to know if the menu for installation has RAID0 configuration options Thanks in advance, Ulisses PD: Does debian.org ships hamm? Sorry for the late reply, I am catching up with the mailing list. Thanks for your reply! :-) Your info is great for me! ...but... you have make a mistake ;-) in hamm mdutils has been replaced by raidtools Yes, silly me. I knew that, I just didn't think of it. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Faster swap by using separate disk?
On Wed, 15 Apr 1998 [EMAIL PROTECTED] wrote: Mark Phillips writes: I have been wondering about whether putting a swap partition on one IDE drive, while putting most of linux on a different IDE drive will speed up swap by allowing both disks to be accessed at the same time. Unfortunately I think I read somewhere that when you have a Master/Slave IDE pair, only one of the disks can be accessed at any one time, so that having the swap partition on a separate disk doesn't help. However my motherboard is capable of using 4 ide devices. It has two pairs: Primary Master/Primary Slave and Secondary Master/Secondary Slave What if I put linux on one of the primary disks, and the swap partition on a secondary disk, will that mean both disks can be accessed at the same time, hence giving a swap speedup? You need to think in terms of available bandwidth. If the swap partition is on the same cable as the root partition, then it has to share the bandwidth with the other partition(s). Moving the swap partition to the secondary controller (where it is by itself), will increase available bandwidth. You can use this trick with a CDROM drive also. No, there is more to it than this. If the OS send a (read, write or whatever) command to an IDE disk, the IDE controller is 'busy' until the command is fully completed. All this time you can not use the other disk on the same controller at all. So, if you try to read a lot of data from both disks simultaneously, they'll be spending a lot of time just waiting for each other. If you put the disks at two different controllers, you don't have this problem. The second thing that you can do is to move up to the new ultra-DMA IDE drives. The bandwidth (bytes per second) is much higher than the standard IDE drives and will speed up Linux as a whole. If you want to really speed up your hard drives, switch to multiple SCSI drives. Period. The third thing that you can do to speed up your system is add more memory so that you don't even need to access the swap partition. True, but I know from experience that if you add more RAM, you'll use it all eventually. I have 48MB in my system nowadays and I still use the swap partitions I have. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: RAID installation
On Wed, 15 Apr 1998, Leonardo Ruoso wrote: I want to install Debian in a RAID 5 w/ 4 9.1GB HDs server. Are there something that I must know? Yes, you must know how to do that. :-) Read the relevant HOWTOs and man pages. Documents you should read include at least: /usr/doc/HOWTO/Multi-Disk-HOWTO.gz /usr/doc/HOWTO/mini/Software-RAID.gz and the man pages in the raidtools package Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: (RFD) New list proposal - debian-unstable@lists.debian.org
On Fri, 3 Apr 1998, Jules Bean wrote: Hi! Currently it seems to me that debian-devel is serving two unrelated purposes. On the one hand it is a forum for developers to pick each others brains, and ask opinions of interested debian users. On the other hand, it also serves to monitor the status of the frozen and unstable distributions. I suggest a new list for this second purpose - tracking bugs and problems with the unstable (or, currently, frozen) release. Things like 'package X has broken, or is it just me?'. Like the recent grep problem, which caused me to post this (and also the xlib6 thing of a week ago, which caused me to subscribe myself to devel in the first place). Thoughts? Yes, very good. This has been mentioned various times before and I have never heard any complaints against such an idea apart from the everlasting we don't need another list. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: I am leaving Debian
On Fri, 20 Mar 1998, Ian Perry wrote: Bruce, On the Thursday, 19 March 1998 8:00 you said. but I feel that my mission to bring free software to the masses really isn't compatible with Debian any longer, and that I should be working with one of the more mainstream Linux distributions. Your statement that free software to the masses really isn't compatible with Debian is disturbing. I have spent a considerable amount of time over the last four months investigating and trialling Debian, over slackware and redhat, as we are in the near future about to install some serious hardware running under Linux. Debian, so far, is ahead. Can you please clarify your statement, as it appears to point to Debian going commercial, which will lead to all sorts of nasties. On slashdot.org[1] I read a quote from Bruce: My mission is to bring free software to the masses. Unfortunately, that's not where Debian's developers want to go - they are more into making a great system for other hackers like them. There's nothing wrong with that, but the two mission statements don't overlap at all. I think this quote explains why it is incompatible. Don't worry about Debian going commercial. Remco [1] check it out: http://slashdot.org/ This is a cool site where hackers, nerds and other people comment on News for Nerds on the Stuff that Matters. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: how can I find broken symlinks
On Thu, 19 Mar 1998, Florian Attenberger wrote: ldconfig shows me, that it can not find some files. I think, its because of broken symbolic links How can I find these links? You can find them with the symlinks command from the symlinks package. But be careful when actually changing anything with this command. Know what you are doing. Read the manual. As long as you use no switches or only the -r switch, you are always safe. Example: $ symlinks -r / This scans all mounted filesystems for symlinks. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Domain Name
On 17 Mar 1998, Martin Bialasinski wrote: Remco Blaakmeer [EMAIL PROTECTED] writes: In that case, why to I get omnic.(none) ??? Where is this (none) coming from? IIRC, the glibc (or whatever is used to resolve the name) checks /etc/hosts. If your /etc/hostname reads omnic, then you should have a line like this in your /etc/hosts: 127.0.0.1 omnic.my.home omnic The fqdn has to be at the first place. It doesn't work for me, though. I have /etc/hostname reading blaakmeer and lines in /etc/hosts like this: 127.0.0.1 localhost 130.89.222.95 blaakmeer.student.utwente.nl blaakmeer cal011205.student.utwente.nl cal011205 Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Domain Name
On Tue, 17 Mar 1998, Lindsay Allen wrote: Are you running NIS ? The only domainname executable that I saw with a grep of hamm/Contents came from net/nis. My box returns elm.(none). I noticed this a year or so ago but had forgotten it. I am about to make your suggested mods but don't want to reboot just for that grin. No I am not running NIS (and never have). My system contains a symlink /bin/domainname - hostname , but I can't find out where it came from. You could also use /bin/dnsdomainname , which is in the 'hostname' package and which is a hard link to /bin/hostname. So, you'd have this in /etc/init.d/boot or /etc/init.d/hostname.sh : hostname --file /etc/hostname dnsdomainname --file /etc/domainname And I don't think you'll need to reboot for it to work. Just run the command as root and see what happens. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: long filenames in M$Dos partition
On Sat, 14 Mar 1998, Hamish Moffatt wrote: On Fri, Mar 13, 1998 at 01:24:33AM +, [EMAIL PROTECTED] wrote: Can anybody of you please explain to me (hopeless newbie) why when I can't see long filenames on mounted msdos partition? This causes me some pain in the neck as I can't freely manipulate files between linux and dos... Since linux assumes that an MSDOS file system is to be operatable even when it does not mounted by Linux, it doesn't want to intefer with MSDOS file system too much. There for, when mounting an MSDOS file system, Linux stickes with the MSDOS rules. But, being aware to the limitations of an MSDOS file system, I believe that Linux tries to make long names short in a consistent and unique way. Perhaps the easiest way when dealing with files that lives both on an MSDOS file system and on linux file system is to determine thier names in the 8.3 format from the beginning. Huh? The vfat file system does long filenames on Windows partitions just fine! You seem to be using an invalid domain name again btw. Yes, you can have long file names on a vfat filesystem. But you can't have long filenames on an msdos filesystem. Scandisk from MS-DOS 6.x (and probably other programs, too) generates a whole lot of errors if you use it on a vfat partition, so you shouldn't use long filenames on a partition that you also use with MS-DOS 6.x (or earlier versions). You can add long filenames to an msdos filesystem at any time. Just unmount it and re-mount it as vfat. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Mail getting stuck in .../smail/input
On Thu, 12 Mar 1998, Adrian Monk wrote: I am having the problem that some mail (only the occasional message) is getting stuck in /var/spool/mail/input. Running mailq gives: ABCDEFG-HJKLMNO From: atheris (in var/spool/smail/imput) Date: Thu, 12 Mar 1998 22:35:35 + (Local time zone must be set--see zic manual page) Args: -t -oem -oMP sendmail I can't work this out. In /etc/timezone there appears GMT (correct). And what on earth is zic?! Eventually the mail disappears, often into some black hole and never to be seen again. Any help much appreciated. To set the timezone, run /usr/sbin/tzconfig . /etc/timezone should read Europe/London if you live in the UK, I think. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Setting screen to BW
On Fri, 13 Mar 1998, Tim Thomson wrote: When I installed Debian on my system, I had a VGA screen on it, so I set it up for colour. I have put the VGA screen on a different system, and have a mono text screen now. Most of my system is now setup for colour, is there a way to unset the most of the settings? ie. is there a config file somewhere that the Debian install sets while setting up the system? I don't mind reinstalling some software, but don't want to reinstall the whole system, just to get colour off. What parts of the system are you having troubles with? If it's X, try to exit X and run the XF86Setup command. If it's another thing, please tell us what it is. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: APM HD Powerdown
On Fri, 13 Mar 1998, Jonas Bofjall wrote: On Wed, 11 Mar 1998, Jens B. Jorgensen wrote: Use the hdparm package, which can instruct your drive to spin itself down at a specified timeout (if the drive supports it). This is not what I want. That way the disks spin down too often. What I was looking for is a way to spin them down when unmounting. Two options: Use a timoue time that is long enough (like 15 or 30 mins or even hours) so that the drives don't spin down when they shouldn't. Use the -y switch of the hdparm command: # hdparm -y /dev/hda This will cause /dev/hda to spin down immediately. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Domain Name
On Wed, 18 Mar 1998, Michael Beattie wrote: Thanks... this works, but I had to make a hardlink to hostname... but I dont think thats a problem.. And yes.. I am running stable. I'm sorry I don't remember how things are in stable. I have been running unstable since July and many, many things have changed since then. Still being a newbie (I hate that word), what happens when you make a symlink, and what happens when you make a hardlink? what is the difference, and what are the reasons for both? The difference is in the thing the link points to. A symlink points to a filename. A hardlink points to the file itself. If you remove a file, all symlinks pointing to it become 'dangling', pointing to something that doesn't exist. However, the file is not removed from the disk before all hardlinks to it are removed. In schematic form: symlink | | \|/ filename --- file hardlink --- | hardlink --- | file hardlink --- | filename --- | Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: command line mixer
On Wed, 18 Mar 1998, Will Lowe wrote: On 18 Mar 1998, Andy Spiegl wrote: is played. Is there a utility that allows me to set the mixer via a command line program, so that I can put that into rc.boot? cam can do this, methinks. I am very happy with aumix, which can do just what Andy wants. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: automatically set DISPLAY after telnet/rlogin ?
On Wed, 18 Mar 1998, Michael Agbaglo wrote: is there is nice way to have the DISPLAY-Variable automatically set to the host from where I logged in ? The easiest way to do this is to use ssh instead of telnet/rlogin. ssh is a secure shell, which means your password doesn't go over the network in plain text. Also, it automatically sets the DISPLAY variable so you don't have to worry about it. ssh can be found in the non-US section of the Debian archive, because it may not be exported from the USA. You can import it, though. Depending on the version you install it depends on either gmp or gmp2 and either zlib1 or zlib1g, all of which are in section 'libs', IIRC. The main ftp archive for Debian-non-US is on ftp://nonus.debian.org/pub/debian-non-US Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Domain Name
On 18 Mar 1998, Martin Bialasinski wrote: Remco Blaakmeer [EMAIL PROTECTED] writes: On 17 Mar 1998, Martin Bialasinski wrote: If your /etc/hostname reads omnic, then you should have a line like this in your /etc/hosts: 127.0.0.1 omnic.my.home omnic The fqdn has to be at the first place. It doesn't work for me, though. I have /etc/hostname reading blaakmeer and lines in /etc/hosts like this: 127.0.0.1 localhost 130.89.222.95 blaakmeer.student.utwente.nl blaakmeer cal011205.student.utwente.nl cal011205 Strange. Works for me. I have: 127.0.0.1 localhost 127.0.0.1 haitech.martin.home haitech [...] #hostname --fqdn haitech.martin.home # hostname --domain martin.home [change second line to 127.0.0.1 haitech haitech.martin.home ] # hostname --fqdn haitech # hostname --domain # (I dont have a /etc/domainname and /etc/hostname reads haitech) Yes, this also works for me, but this was not the issue. The issue was getting agetty to display the fqdn in the login screen. You can do this with 'issue escapes' in /etc/issue. I have this line in /etc/issue : Welcome to \n.\o at \l This resolves (or should resolve) to something like: Welcome to blaakmeer.student.utwente.nl at tty2 But agtty apparently uses the 'wrong' system call to get the domain name and displays (none) if you didn't set it. The issue here seems to be that there are two ways of getting the domain name. /bin/domainname (linked to /bin/hostname) uses getdomainname(2) to get the domain name. This technically is the YP/NIS domain name. /bin/dnsdomainname (linked to /bin/hostname) uses gethostname(2) to get the host name, passes the result to gethostbyname(2) to get the FQDN and then strips the hostname part from the FQDN to get the domain name. Note that the results can be different. The first will be (none) if you don't use NIS. Your NIS domain name can apparently be different from your DNS domain name. To make them the same, I found that this is the easiest way: domainname `dnsdomainname` This doesn't use /etc/domainname, so that's one file less to edit. :-) If you don't have /bin/domainname or /bin/dnsdomainname , you can create the links yourself: # ln -s hostname /bin/domainname # ln -s hostname /bin/dnsdomainname Note, this is only needed if you want getty to display your fqdn at the login screen and you don't use NIS. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: wtmp-libc6 (solved)
On Wed, 18 Mar 1998, dpk wrote: I previously had tried the steps of removing the wtmp, rebooting, etc without success. I tried the libc6 upgrade procedure again, without success also. It was due to the non-US package 'ssh'. The wtmp file remained readable until I ssh'ed to my machine - it then became a binary mess. I downloaded/compiled/installed the source for ssh and haven't had any problem since! I know there is another victim of this out there,.. I hope he can share the same solution. Is your ssh linked against libc5? In that case, get the newest ssh in 'debian-non-US/hamm', which is linked against libc6. I have no problems at all with it. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Domain Name
On Fri, 13 Feb 1998, Martin Schulze wrote: I'm not sure If my first attempt got through... probly mail server crash did it... *GRIN* In the /etc/issue file you can use the \n.\o sequence to get a copy of your hostname and domain name right? In that case, why to I get omnic.(none) ??? Where is this (none) coming from? Pure interest here, I have just been playing with it I've solved it by not only setting the hostname but also the domainname at boot time. If you're running 'unstable', it's in /etc/init.d/hostname.sh, otherwise I think it's in /etc/init.d/boot . You'll find a command like hostname --file /etc/hostname After this, add the command domainname --file /etc/domainname Make sure /etc/hostname contains your hostname and that /etc/domainname contains your domain name. Then it all works. I don't know if this is the right way to do it, but it works for me. My /etc/issue contains these lines, which are quite informative about the system: Welcome to \n.\o at \l This system runs \s \r on \m Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: syntax error in cron start-up file at boot-time
On Thu, 12 Mar 1998, James Dietrich wrote: The subject fairly sums up the problem: Below is a snip of the messages that are displayed upon booting. [snip] Starting deferred execution scheduler: atd. /etc/rc2.d/S89cron: line 20: syntax error near unexpected token `;' /etc/rc2.d/S89cron: line 20: ` ; cron reloads automatically' [snip] What do I need to do to fix this problem? I am running the latest from hamm. You can change the ';' that starts line 20 by a '#' and the problem should go away. A bug report has already been filed on this issue and it should be fixed real soon now. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: dselect mystification (fwd)
On Wed, 11 Mar 1998, Travis Cole wrote: Try a hitting Q that should make it ignore any errors it is having. And after that, don't install anything but first run 'Access' from dselect. Be sure to only install from the 'unstable' distribution, not 'stable'. After that, run 'Update' in dselect. Then, enter the 'Select' screen again and see what happens. Remco -- blaakmeer: 1:35am up 12:39, 7 users, load average: 1.11, 1.26, 1.30 -- E-mail the word unsubscribe to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED] .
Re: [OFF-TOPIC] less with built-in grep ?
On 11 Mar 1998, Lorens Kockum wrote: In debian-user, Ben Pfaff [EMAIL PROTECTED] wrote: With the example above, I suppose I could write some kind of macro, opening shell output, but it gets much more interesting when I am looking at program | less and I do not want to execute program again, and I also want to be able to go back to un-grepped output without executing program again. You can type |grep xxx | less, and this will open a sub-less with the grep output in it. To get back to the ungrepped file, just quit from the sub-less. Yes, only now the program isn't finished yet, and I want to do a 'F' (makes less act like tail -f) command. This is not just some fun feature I want to have to look at, it comes in useful when you're looking at, say, a syslog, or program output. I have a window with my syslog, and I want to only see lines corresponding to a certain regexp. I don't know how to do that. No, the regexp is not something I could use /etc/syslog.conf for. No dreaming. Anyway, modifying syslog.conf just for a few minutes is just terrible, and would only work for syslog in any case. In addition, if the less selected the lines for me, I would be able to quit that and see the context. In effect, I want to put an egrep filter in less, with a regexp that could be modified without restarting less or re-reading file. Hmmm ... there might be some specialized real-time log monitor which would let me do that ? I think I'll be forwarding this to markn, less author. Should be rather simple, and ... is there anyone else that thinks it might be useful ? You could, of course, use something like this: tail -f logfile | grep regexp | less or even something like: tail -f logfile | grep regexp /dev/tty10 Wouldn't that be sufficient? Only thing with grep (and some other utilities) is, if stdout is not a tty it will use a buffer for stdout. The first example probably won't work good enough for you. But how about the second? 'grep' will keep working as long as 'tail' doesn't close the pipe, which it won't until it exits. Remco -- blaakmeer: 1:40am up 12:44, 7 users, load average: 1.03, 1.11, 1.21 -- E-mail the word unsubscribe to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED] .
Re: Newbie question-install ethernet card
On Wed, 11 Mar 1998, zhang xu wrote: hi, how can I install the 3c509b ethernet card on my Debian box? The simplest approach is to use the program 3c5x9cfg.exe that comes with the DOS drivers to disable PnP support and then use the standard driver for this card that comes with the kernel. If you need more specific information, please ask a more specific question. Remco -- E-mail the word unsubscribe to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED]
Re: autoup.sh
On Tue, 10 Mar 1998, Henry Hollenberg wrote: After running autoup.sh where do I point dselect to complete the hamm upgrade from 1.3.1. Where you first used stable, contrib and non-free, now use dists/unstable/main, dists/unstable/contrib and dists/unstable/non-free. The rest is the same. Remco -- E-mail the word unsubscribe to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED] .
Re: netscape install
On Tue, 10 Mar 1998 [EMAIL PROTECTED] wrote: I am ready to install netscape on my system, but have a problem. The netscape loader is for version 3.01. ftp.netscape.com only has version 3.04 or the newer communicator 4.0x versions. Version 3.01 is not available. My system is a bo (1.3.1) system. What can I do to get netscape installed? The netscape3 and netscape4 packages in hamm also work for bo systems. Install them by hand and it will work. Remco -- E-mail the word unsubscribe to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED] .
Re: CNAME records (was: Re: dynamic DNS within a non-dynamic domain)
On Sat, 7 Mar 1998, George Bonser wrote: Ok, I just telnetted to your port 25 and you are announcing as Cal011205.student.utwente.nl This means that smail is discovering the DNS name when it is trying to find its own name. This can be changed by FORCING smail to use the correct hostname. Weird. It doesn't do this when I telnet to it locally: $ telnet blaakmeer smtp Trying 130.89.222.95... Connected to blaakmeer.student.utwente.nl. Escape character is '^]'. 220-blaakmeer.student.utwente.nl Smail-3.2.0.100 (#2 1998-Jan-13) ready at Mon, 9 Mar 1998 01:52:05 +0100 (CET) 220 ESMTP supported quit 221 blaakmeer.student.utwente.nl closing connection Connection closed by foreign host. $ Try this: hostnames=blaakmeer:cal011205 domains=student.utwente.nl (note those hostnames must be in that order!) This should cause smail to use blaakmeer as its true hostname but also accept mail addressed to cal011205 as local. I am sorry, but this doesn't help. This is what I had (and have again) in the config file: visible_name=blaakmeer.student.utwente.nl -domains hostnames=blaakmeer.student.utwente.nl:blaakmeer:cal011205.student.utwente.nl:cal011205:remco:remcob:localhost (remco and remcob are aliases used only by a few friends that put them in their /etc/hosts file) Is this 'bad' or 'wrong' in any way? Note that your mail shows the correct From: header, [EMAIL PROTECTED] Yes, I know. This could mean that the problem lies on the receiving end. I'll try to send mail between hosts running Debian smail tomorrow. Remco -- blaakmeer: 1:45am up 6 days, 6:14, 9 users, load average: 1.19, 1.38, 1.40 -- E-mail the word unsubscribe to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED] .
Re: CNAME records (was: Re: dynamic DNS within a non-dynamic domain)
On 9 Mar 1998, Miquel van Smoorenburg wrote: Note that AFAIK all SMTP mailers (well, sendmail at least) will rewrite the addresses in the mail headers. That includes rewriting a CNAME pointer to the canonical name it points to. So even if you send out mail with the CNAME in the From: header field, the recipient mailer will rewrite it.. no way around it. The correct way to solve this is to not use CNAMEs but to just set up an extra A + MX record. Or perhaps just an MX record if all you're using it for is mail. Sure, but I don't control the DNS around here. Though I could ask, of course. But I doubt I will get any results. Remco -- blaakmeer: 3:55pm up 6 days, 20:25, 9 users, load average: 1.58, 1.48, 1.45 -- E-mail the word unsubscribe to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED] .
Re: Upgrading to 2.0 from 1.2
On Mon, 9 Mar 1998, Stan Brown wrote: I have a Debian system, that is still at 1.2 (as verifed by /etc/debian.versio). I plan on upgrading to 2.0 as soon as it becomes stable. Should I first upgrade to 1.3 first? No. There is a script that will do the critical part of the upgrade to Debian 2.0 from both Debian 1.2 and Debian 1.3. After this critical part is done, you can just point dselect at the ftp archive (or a cd-rom) and it will do the rest for you. This will of course be documented properly when Debian 2.0 is released. Until then, you can find information about it in the faq-o-matic on the Debian www site, at http://www.debian.org/cgi-bin/fom?file=57 . Remco -- E-mail the word unsubscribe to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED] .
Re: Virtual Consoles
On Wed, 4 Mar 1998, Graham Lillico +44 1785 248131 wrote: Hi, Can anyone tell me why debian 1.3.1 uses agetty, instead of something like mingetty? I wish to screen to clear after booting my system but using agetty this is not possible, I have tried using mingetty and this does what I require, but I am wondering if there is anyreason agetty was chosen (i.e Does it support some features mingetty does not, when used for virtual consoles)? It is possible with agetty, since I am using it and I managed to get it to clear the screen. Somewhere in /etc/init.d there is a file that rebuilds your /etc/issue every time you boot the system. I think it is /etc/init.d/boot. You can find it with grep issue /etc/init.d/*. Change the code that rewrites /etc/issue to, for example, this: # Rewrite the file /etc/issue echo -e \\033[H\\033[J /etc/issue cat EOF /etc/issue Welcome to \\n.\\o at \\l This system runs \\s \\r on \\m EOF The first 'echo' command puts codes in /etc/issue that will clear the VC. the rest puts a nice banner on the screen. For some reason you have to use double backslashes instead of single backslashes. I think it has to do with bash. Look at the issue escapes section in the man page for getty to see what you can do with this. Remco -- E-mail the word unsubscribe to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED] .
Re: loading support for 2 ethernet cards
On Wed, 4 Mar 1998, Nathan E Norman wrote: On Wed, 4 Mar 1998, Rob Goodwin wrote: : I am having trouble autoprobing more than one Ethernet card at boot : time. I have compiled support for the cards I am using (they are both : of the same type) but only one is detected at boot time. I found a : howto that describes a method that involves adding the line : ether=0,0,eth1 to my lilo.conf file but I continuosly get syntax : errors when doing this and I cant figure out where in the lilo.conf file : to put this line... Any help would be appreciated. : : Thanks in advance : Rob Yeah, it's a bit unclear at first. What you want to do is put the information on the append line, like this: append=mem=96M ether=10,0x300,eth0 ether=11,0x2e0,eth1 I'd recommend you put in the real values rather than 0,0 - that enables autoprobing and let's face it, you do want to know which interface is which, don't you? I hope this helps. By the way, my two cards are 3C509B's. If you have two or more 3com cards of the same type (and perhaps if you have two or more cards of the same type of any brand, I don't know), they are always sorted by ethernet address, so the one with the lowest ethernet address always has the lowest ethX number. I know of a guy that had one NE2000 clone and three 3c509 cards. His setup was: eth0 : NE2000 eth1 : 3c509 with lowest ethernet address eth2 : 3c509 eth3 : 3c509 with highest ethernet address He could get the NE2000 at eth3 and the 3c509's at eth0 to eth2, but he couldn't change the relative order of the 3c509's. So, unless you have different network cards, you don't need to specify the actual irq and io parameters. If anyone has an example that proves me wrong, please respond. I am not claiming to be an expert on this. Remco -- E-mail the word unsubscribe to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED] .
Re: CNAME records (was: Re: dynamic DNS within a non-dynamic domain)
On Wed, 4 Mar 1998, George Bonser wrote: On Thu, 5 Mar 1998, Remco Blaakmeer wrote: [question about CNAME records and MTAs rewriting the From: header when the hostname is a CNAME] One thing to do is to make sure that you have the following in your own /etc/hosts file: 130.89.222.95 blaakmeer.student.utwente.nl cal011205.student.utwente.nl Yes, it actually is: 130.89.222.95 blaakmeer.student.utwente.nl blaakmeer cal011205.student.utwente.nl cal011205 remco If you are using sendmail, it will grab the FIRST name listed when it canonifies the address. Also, you might use mail -v to send someone an email outside of your network, this allows you to see exactly what what your machines gives the other mailhost in the MAIL FROM: part of the transaction. I am using smail. See the results of my tests below. If you are using NIS and it is using an NIS lookup ahead of DNS, it will also grab the first name listed after the IP address in the table. No, there is no NIS anywhere near my computer. If you are using a smarthost that uses SunOS with Sun's sendmail. you are likely out of luck. I will go into the reasons for that in private email if this is the case but the only way around it is to get your hostname listed first in the domain's NIS hosts.byname map if it uses NIS for host resolution. No, also no Sun machines in sight. The smarthost is, AFAIK, a HP machine that runs HP/UX and has this greeting: 220 driene.student.utwente.nl 5.67b/IDA-1.5 Sendmail is ready at Sat, 7 Mar 1998 05:30:38 +0100 BTW, my computer has this greeting: 220-blaakmeer.student.utwente.nl Smail-3.2.0.100 (#2 1998-Jan-13) ready at Sat, 7 Mar 1998 05:40:00 +0100 (CET) 220 ESMTP supported Ok, now for the tests. The conclusion of the tests is this: If I locally send mail to myself, the From: header is not rewritten. If I send mail to another host, the From: header is rewritten after smail got the message. When fetchmail fetches my mail, the From: header is rewritten before smail gets the mail. When I fetch the mail using telnet to the pop-3 port, the From: header is already rewritten. The mail 'at another host' is sent to [EMAIL PROTECTED], which is also my e-mail address. I have fetchmail running as daemon to get mail from that address to my computer. Note that the From: header is rewritten to the address that the CNAME points to, not the FQDN for my computer. You'd say they are the same, but they are not. The FQDN is all lowercase letters, while the name that the CNAME points to starts with a capital. Because DNS lookups are not case sensitive this doesn't matter normally, but now I can see the difference between them. 130.89.222.95 is cal011205.student.utwente.nl CNAME: blaakmeer.student.utwente.nl - Cal011205.student.utwente.nl I have compiled a full description of what I did to test this and what the result was, but I don't want to send it to the list as it is rather lengthy. Should I send it to somebody? Or can I send a 10kB mail to the list? Remco -- E-mail the word unsubscribe to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED] .
Re: dosemu and graphics
On Thu, 5 Mar 1998, David Morris wrote: My computer runs Debian continually. However, lately, I have been needing to reboot to dos (a 6 year old daughter who wants to play a couple dos games who it is very hard to say 'no' to). I have the latest (hamm) dosemu installed for a text based program and have tried to get it to run in graphics mode so the games are more accessible. However, as a normal user I can't get it. If I try to run vgaon it kicks out an error: CAN'T DO VIDEO INIT, BIOS NOT MAPPED. As root from a console I can get the graphics with no problem. I hate to teach my daughters how to gain root access on my computer though just to run a couple games. What can I change to all my normal user account to access the video? You could do put all users who are allowed to run dosemu in a separate group, e.g. 'dos': # addgroup dos # adduser me dos # adduser daughter dos Then, make /usr/bin/dos setuid root and executeable only by group 'dos': # chown root.dos /usr/bin/dos # chmod 450 /usr/bin/dos Now, only users in group 'dos' can run dosemu and it will run as root. Also, you will need to configure dosemu to use your video card. /etc/dosemu/conf explains itself. Remco -- E-mail the word unsubscribe to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED] .
