Migrated to debian; aliases stopped working
When I moved to debian, my aliases are no longer working. For example, in /etc/aliases I have the following entry: debian: [EMAIL PROTECTED] However, when I try to send a message to debian, it sends a message to [EMAIL PROTECTED], which is my FQDN attached to the alias name. That is, the alias db is not being used. I ran both # newaliases (which I did under RedHat/sendmal) and I ran # postfix /etc/aliases (I'm now using postfix under debian). I'm not sure, but is the first command associated only with sendmail, and I should now use the second only? In any case, the /etc/alias.db that builds is not being used when I send messages. Haines Brown -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Simple little basic config questions
In moving from RedHat to debian, I'm left with some simple little basic configuration questions. They all relate to a situation in which I operate at this point from console. 1. Where do I set the global bash prompt format? I changed PS1= in /etc/profile, but that only affects user, not root. 2. I had placed the command "setterm -blank 0" in RedHat's /etc/rc.d/rc.local to block screen blanking while running in console. Debian does not use that file. What is its equivalent? 3. My usual practice is to avoid xdm and boot to a text login prompt. To do this, in rc2.d I belive I edited the symlink to the xdm program, renaming "S99xdm ->..." to "K99xdm ->...". But in debian I get a beep when I try. Am I imagining I once edited the name of a symlink? Can't one do it in debian? Haines Brown -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Simple little basic config questions
> On Sat, 2003-10-25 at 19:57, Haines Brown wrote: > > > > 1. Where do I set the global bash prompt format? I changed PS1= in > >/etc/profile, but that only affects user, not root. > > > > 2. I had placed the command "setterm -blank 0" in RedHat's > >/etc/rc.d/rc.local to block screen blanking while running in > >console. Debian does not use that file. What is its equivalent? > > > > 3. My usual practice is to avoid xdm and boot to a text login > >prompt. To do this, in rc2.d I belive I edited the symlink to the > >xdm program, renaming "S99xdm ->..." to "K99xdm ->...". But in > >debian I get a beep when I try. Am I imagining I once edited the > >name of a symlink? Can't one do it in debian? > > That's almost exactly what I did: > # cd /etc/rc2.d > # mv S99xdm __S99xdm Yes, I was trying to edit the word in emacs in a console, and emacs was playing tricks on me. But renmaing with mv worked fine. Thanks. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Simple little basic config questions
Thanks, Wayne. I had previously done the basic configurations globally rather than in ~/.bashrc, but your suggestion to do it for each user has a backup advantage. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Simple little basic config questions
As for setting up basic bash configuration, a little experimentation shows that this is what I've got (debian 3.0r1). Root has both .bashrc and .profile, and the configuations (custom bash prompt and setterm) can go in either place. User has a .bashrc and .bash_profile (there's no .profile), and the configuration must go into the latter. It does not work for me if put into .bashrc. Thanks for alerting me to this difference. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Simple little basic config questions
> one of the 'freedoms' of debian is that runlevel 2 to 5 are the same. 2 is > the default runlevel. RH and others have seperate runlevels. Its something > that confused me and there are some people out there like me who like the > RH runlevel scheme but havent changed prevailing minds. Oh well! Kev, Yes, my sense has been that in principle one designs a set of alternative run levels 3-5, leaving runlevel 2 in its default state, and than for whatever reason, you switch runlevels to get a different setup. I suspect that this involves a command such as # init 5, but I've never tried it. Does that init command simply change the default runlevel? It seems that default is set in /etc/inittab, where it has: id:5:initdefault: . I assume you can also simply edit this line to change the default runlevel when you boot up. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Simple little basic config questions
> On Sat, Oct 25, 2003 at 09:08:48PM -0500, Ron Johnson wrote: > > > > Of course, you could always deinstall xdm : > > # apt-get --purge remove xdm > > apt-get remove --purge xdm Yes, on second thought, removal might be best, since I'll never use xdm, and with a new install, this is a good time to clean house before I fill it with new furniture. For example, I'm now running postfix, but exim seems to be around as a result of basic installation. The apt-get remove --purge command I understand will remove exim's configuration files as well. I hope it would not abuse any files that postfix needs. I guess the aptitude equivalent would simply be # aptitude purge . Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Simple little basic config questions
> Which is generated by the "adduser" routine by copying the skeleton > files from /etc/skel. You can add other files in this directory if > you want them to be added to new users' home directories. Interesting--the plot thickens! So, if one wants to set a global configuration for bash, such as a custom prompt or setterm= then the way to do it is to edit the files in /etc/skel. For example, I could uncomment in /etc/skel/bash_profile the section that has "source ~./bashrc" and then in the /etc/skel/bashrc file enter my custom value for PS1= or commands such as setterm = ... All this is entirely new to me ;-). Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Migrated to debian; aliases stopped working
> BTW I use exim and exim.conf contains the following lines about aliases > > # This director handles aliasing using a traditional /etc/aliases file. > # If any of your aliases expand to pipes or files, you will need to set > # up a user and a group for these deliveries to run under. You can do > # this by uncommenting the "user" option below (changing the user name > # as appropriate) and adding a "group" option if necessary. > > system_aliases: > driver = aliasfile > file_transport = address_file > pipe_transport = address_pipe > file = /etc/aliases > search_type = lsearch > # user = list > # Uncomment the above line if you are running smartlist > > Don't know if it's similar for sendmail or postfix. In RedHat/sendmail, the syntax is a little different. Here it is for the alias for a single address and an alias for a list of addreses alias: alias: :include:// I assume that in your complicated example, the word "aliasfile" is the alias for a distribution list having the name "aliases." In debian/postfix, the syntax found in my /etc/aliases is much simpler. In /etc/aliases, I use the same syntax as with RedHat sendmail, only build the database with the # postalias /etc/aliases rather than # newaliases command. However, since my aliases don't work, I'm not sure. Haines Brown -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Simple little basic config questions
> Haines Brown([EMAIL PROTECTED]) is reported to have said: > > prompt and setterm) can go in either place. User has a .bashrc and > > .bash_profile (there's no .profile), and the configuration must go > > into the latter. It does not work for me if put into .bashrc. > > Do you have > > source .bashrc > > As the last line of your .bash_profile? That might help. No, the default (debian3.0r.1) is to comment that in .bash_profile: # if [ -f ~/.bashrc]; then # source ~./bashrc # fi I don't see why this is commented, for it seems to disable ~/.bashrc. Is that so? If so, why it it disabled by default? > Note, just in case you didn't know this. After making changes to > these files it is not necessary to exit and relogin. Just enter > . .bash_profile I don't understand. "Enter" what into ~/.bash_profile? What does your elipsis here refer to? > As long as you have the above in .bash_profile, any changes to > either file will take effect. Interesting. > Oh, welcome to Debian. You won't be sorry you switched from RH!! I don't know if this comment is accurate, but RedHat seemed to be moving toward a more integrated desktop, which a) caused unexpected problems b) that are harder to resolve. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
kernel-header installation question
I'm seeing to compile an inteface for the nVidia driver on a new basic installation of debian 3.0r1. The nVida installation dialog says the kernel headers are not found. So I install the headers. The first time that went well, and the second time: # aptitude install kernel-headers-2.4.18-bf2.4 ... Sorry, kernel-headers-2.4.18-bf2.4 is already installed in the requested version (2.4.18-5). I find that my /usr/src/kernel-headers-2.4.18-bf2.4 directory only holds a .config file and an include directory. I assume that what I should see here is a file with an .h extension if the header is actually installed. I suspect that I'm making a simple (i.e., stupid) little mistake somewhere. Haines Brown -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: kernel-header installation question
> On Mon, 27 Oct 2003 10:52:02 -0500 (EST) > [EMAIL PROTECTED] (Haines Brown) wrote: > > > I find that my /usr/src/kernel-headers-2.4.18-bf2.4 directory only > > holds a .config file and an include directory. I assume that what I > > should see here is a file with an .h extension if the header is > > actually installed. > ... > the .h files are in the include directory as they wpuld be with the > kernel source Abert, thanks. I had not dug sufficiently deeply. nVidia interface now compiled by, using the --kernel-include-path option. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
List mail being redirected?
Sorry if this has been brought up before, but a message I send to this list this morning seems to have been redirectected at some point to a different domain: [EMAIL PROTECTED] The following address had permanent fatal errors [EMAIL PROTECTED] I don't know if this redirection occurred before my message was distributed to the list, or as a result of that distribution. I'd appreciate knowing if anyone saw my message arrive. I don't recall the exact subject line, but it had to do with aptitude installation of kernel-headers. Haines Brown -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: List mail being redirected?
Thanks for letting me know my message got through. I'm only at the point of setting up the X windows system in my debian installation, so can't yet cruise cyberspace to check the archives. At present, my X server starts, but all I get is a blank blue screen. I try to start icewm (it seems to be ready to run) by adding an exec into /etc/X11/xinit/xinitrc and in ~/.Xinitrc, but nothing happens. When I get this straightened out, and get my browser going, debian list archives will be one of the first places I'll head ;-) Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Initializing X clients (Was: List mail being redirected?)
> > At present, my X server starts, but all I get is a blank blue > > screen. I try to start icewm (it seems to be ready to run) by > > adding an exec into /etc/X11/xinit/xinitrc and in ~/.Xinitrc, but > > nothing happens. > > This should be ~/.xinitrc (all lowercase). Actually, I always put it > in ~/.xsession , I believe both things should work, but try the > later. Also, make sure that they have executable permisions > > $ chmod +x ~/.xinitrc ~/.xsession Actually, the .Xinitrc was a typo; I usually use .Xclients. I tried your .xinitrc and .xsession, but they did not help. Actually, there's some more relevant information. First, I do get the debian logo white screen, and then the blue screen, and that I take implies the X server successfully loads (also, it reports no errors). Second, the actual error that shows up in .xsession-errors is: IceWM: Bad Option: I take this to mean that the problem is not with X, but with IceWM. I also infer that the icewm executable is called, but it has an invalid option. However, the icewm binary has no options, and I've no user defined options that would cause this. Third, the actual command I use to start icewm is: exec icewm || exec xterm and my understanding is that this opens an xterm if the window manager fails to run. But that does not happen. Hard to reconcile this with previous point unless the failure of icewm to load an option properly halts the process entirely at that point. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Initializing X clients
Get rid of everything: what a good idea! I got rid of everything, and in the .xsession put exec icewm. That didn't work. So I replaced the command with exec xterm. That did work--sort of. When I start x with the xterm execution, what I get is a black rectangle in upper left 2/3 of screen with a bash prompt. That sounds good, but everything else is bad: 1. The rest of the screen is trash background (a rough blend of colors rather than (default?) blue. Shouldn't the terminal have been left the remaining background intact? 2. The terminal has none of the features I'd normally expect a terminal window to have, such as a window frame. That may be because I've no window manager running yet. 3. The insertion point of the prompt is not a solid rectangle, but an empty rectangle. Any keyboard input cause the prompt to disappear, although the black rectangle remains. Interesting. I've no idea what it all means. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Initializing X clients
Jaume, You missed one of my points, which was that the command prompt of the xterm disappeared as soon as there was keyboard input. Even the Ctl-Alt-Backspace to close the X server causes bash to terminate. I'm not sure bash quite made it in the first place, for the white insertion rectangle should have been solid, I suspect. I have a ~/.xsession-errors file, and when I press any key I get: xterm: fatal IO error 32 (broken pipe) or KillClient on X server ":0.0" Does this mean that in /etc/X11/XF86Config-4 my keyboard is not properly configured? If so, would that have appeared to icewm as a bad option? Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Initializing X clients
> > You missed one of my points, which was that the command prompt of the > > xterm disappeared as soon as there was keyboard input. Even the > > Ctl-Alt-Backspace to close the X server causes bash to terminate. I'm > > not sure bash quite made it in the first place, for the white > > insertion rectangle should have been solid, I suspect. > > > > I have a ~/.xsession-errors file, and when I press any key I get: > > > > xterm: fatal IO error 32 (broken pipe) or KillClient on X server > > ":0.0" > > Weird ... To solve the problem, we need more information: OK. 1. I don't use graphical login, but startx in console. 2. No (EE) or (WW) in /var/log/XFree97.0.log. It said it could not find Speedo path and so disposed of it. Otherwise, clean. 3. I'm running a new installation of debian Woody 3.0r.1 from cdrom. Kernel is 2.4.28-bf2.4. XFree86 is 4.1.0-16. X-window-system is installed and version 4.1.0-16. The icewm is 1.0.9.2-7. > 4.- You can reconfigure the xserver at any time with > > $ dpkg-reconfigure xserver-xfree86 If it generates a functional configuration (with exeption it won't know of my nVidia driver), that might represent a useful clean slate (I did a lot of fiddling with my current configuration file, but it originated with one that works for RedHat, same hardware). I'll give this a try. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Initializing X clients
> 4.- You can reconfigure the xserver at any time with > > $ dpkg-reconfigure xserver-xfree86 Sorry, I didn't realize that I'd complete this so quickly. Creating a very elemental XF86Config did the trick, and icewm loads. Now I can play with the configruation to see where I went wrong. Thanks. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Initializing X clients
Thanks for the input. I was able to get x running by rebuilding the configuration file. Somehow, I had managed to gum things up. Tomorrow I figure out what I did wrong, but meanwhile at least I get my window manager to display. My mouse is working and I can input keyboard into the xterm, so I'm all set. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Simple little basic config questions
I have more elementary configuration questions arising from my transition from RedHat to debian. Sorry to be a pest. I think this may be is a debian question because user can start the FileRunner file manager, but not root. When root tries, it gets the error: Application can't initialize because it lacks display name and no $DISPLAY environment variable. Error stgartup script: can't read "tk_patchLevel": No such variable while executing. How do I interpret these? In fact, if I try # echo $DISPLAY, nothing is returned, which means the root account is not configured properly, I'd guess. It would make sense to define the value of DISPLAY globally, I should think. If so, how does one do that? The second part of the error statement would seem to be a script error, where the value of tk_patchLevel is never defined, but since user can run the application OK, I assume the problem is deeper than that. Could it be the missing DISPLAY? Haines Brown -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Simple little basic config questions
> > I think this may be is a debian question because user can start the > > FileRunner file manager, but not root. When root tries, it gets the > > error: > > > > Application can't initialize because it lacks display name and no > > $DISPLAY environment variable. > > You're probably using 'su -' to get to root, right? Yes, you are quite right. > explicitly set your display to whatever it's normally > set to for your users in root's .bashrc or .profile. I tried: "set DISPLAY teufel:0.0; export DISPLAY" /root/.profile, but it. My sytax probably wrong. Can I substitute "localhost" here for "teufel"? Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Simple little basic config questions
> > I tried: "set DISPLAY teufel:0.0; export DISPLAY" /root/.profile, but > > it. My sytax probably wrong. Can I substitute "localhost" here for > > "teufel"? > > If you're logged in as a normal user, what does > env | grep DISPLAY > show you? Monique, it displays as it should: :0.0. My problem only with root. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Simple little basic config questions
> > Application can't initialize because it lacks display name and no > > $DISPLAY environment variable. > > Look what I just found as a new package on unstable: > > Sux is a wrapper around the standard su command which will transfer your > X credentials to the target user. > > http://sourceforge.net/projects/sux/ ( from http://fgouget.free.fr/sux/ > ) Thanks, but that seems a workaround more than a fix. For example, if I logged in as root, I'd be in trouble, without access to a display. I need a statement that would put the default display into root's environment. Further, when I run su root without the "-", I carry over user's environment, and so root acquires a display. I wonder what might have gone wrong to cause display to be missing for root in the first place. That is, was it the effect of a faulty new installation of debian 3.0r1 from cdrom disks, or is this a known bug? Haines Brown -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Simple little basic config questions
> Please don't CC me. (If somehow my sig isn't clear enough, please let > me know how I can make it so.) My apologies. The current auto CC: is something I did not have before, and so I'm not used to removing that line. I was aware I had forgotten to do that as soon as I had sent the message to you. > Oh wait a minute. I just looked at my .bashrc, etc and think I've > noticed the problem. > > Try simply doing this: > > DISPLAY=:0.0 > export DISPLAY Seems to have brought some progress. Now, when I su - root and run: echo $DISPLAY :0.0 But apparently, all that does is to say that I've successfully assigned the value ":0.0" to DISPLAY, but it does not seem to have been exported. That is I run an application that calls for display :0.0 and it still fails: Xlib: connection to ":0.0" refused by server Xlib: Client is not authorized to connect to server Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Simple little basic config questions
> Why not just use 'su' (with no parameters) or 'su - -p'? > > -m, -p, --preserve-environment > do not reset environment variables, and keep the same shell > > That will preserve things like X display dettings. Just an idea. Perhaps it is a philosophical issue, but my instinct tells me that my root account should be self-contained. I should be able to log in as root and startx. To have to log in as a user, and then su to root, assumes there's a user account, and assumes that it is working. Both of these conditions do not always exist. Further, when I am logged in as user, when I su to root, I want to acquire root's environment. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Simple little basic config questions
> each user has a session and a session key. this key is used to > authenticate yourself to the Xserver. Root as a key and each user > does. Yes, that makes sense. > so when you login as user and then switch to root, it tried to use your > root key to access the user session-- no go. ? When I login as user, and then "su - root," does not root then use its own session key? Are you saying that when I "su - root", root tries to use user's session key? one solution is: > user% xhost + > user% su > root! xcalc > but this is an insecure hack since in says anyone can snoop on your > xserver. but if you are not on the net or have a firewall it may be used, > > the better solution is to 'merge' your X authenticaion key database but I > forgot the command. Thanks, Kev. My understanding of Linux is that normally you want to log in as user because being root carries with it certain risks. But regularly, we, running as user, find that we need to do something that requires root's privileges, and so we "su - root". That's what I read in "Running Linux" and elsewhere. It's what I've been doing for years. So I assumed that by moving from RedHats to debian, things would continue as before. But they have not. So, the important question that still remains unansered: was my installation of debian flawed, or does debian simply work differently than what I assume? Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Simple little basic config questions
> No, I think if you had actually started X as root then you certainly > would have an appropriate $DISPLAY. The issue is not really rootness, > it's that $DISPLAY is set in the environment of the X session which is > run as the user who started X, and .Xauthority is in the home directory > of the user who started X, and it's quite easy to lose all that when > changing users. 'sux' is not a workaround, it's a valid solution. I notice that .Xauthority in /root has zero size. If it is going to authenticate root for the x server, I should think there would be something in it. When you say "changing users," do you refer to logging in as user and then running "su - root"? I assume virtually everyone does this regularly and successfully, and so I assume my inability to do it is a sign that I need to do a fix. For years I didn't "loose all that", but could "su - root" as I needed. I still don't know whether my system's busted or if it is me ;-) That is, is "loosing all that" a natural occurance or a flaw in my setup? I presume every debian user who is both user and administrator of his machine (probably the majority) will occasionally want to su to become root (I assume everyone does that regularly). Certainly they all don't have Sux installed. I appreciate that one does not want to run as root, but I do it when installing a new system or retreat to it when user's account ceases to function. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Simple little basic config questions
Johannes, Took your advice, and that seems to have worked. $ xhost +local non-network local connection being added to access control list $ su Password: # So I gave root access to the X server. Still don't understand why I had to do this rather than it being the default setup with an installation. Thanks! Haines Brown -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Simple little basic config questions
> I don't think anything is broken in your setup. I do seem to > recollect that, when I used RH, su- was able to use Xwindows, so it > may be that some distros execute this differently. Yes, Richard, apparently I had work habits based on the RedHat setup, and didn't realize that under debian I'd have to add root to the access control list. My next question has to do with disabling screen blanking and power saving under X. This also may be the result of moving into a somewhat different setup when I moved from RedHat to debian. I had added to ~/.Xclients: xset s off xset -dpms But now that is having no effect. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: GeForce4 MX440 with kernel 2.4.18-bf24
> Is anyone running the subject video card on the subject kernel. If so what > driver are you using? If you know where can I get a copy of your driver? > Regards; > Hoyt Hoyt, that's my setup. I downloaded the new nVidia driver from the nVidia website. Installation a bit rocky (I'll go into specifics off line, if you like), but if one discounts for my incompetence, the installation was straightforward. The nVidia website has a sort of driver identification utility, but it seens that all the GeForce cards have the same driver now for all kernels. There's need for an interface between the driver and kernel, which is not available on line and so must be compiled by the driver installation utility. That utility will search for and probably not find your kernel headers, in which case you need to add an option to provide the path. The name of the driver that you add to your XF86Config(-4) is "nvidia." Haines Brown -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Simple little basic config questions
> > My next question has to do with disabling screen blanking and power > > saving under X. This also may be the result of moving into a somewhat > > different setup when I moved from RedHat to debian. > > > > I had added to ~/.Xclients: > > > > xset s off > > xset -dpms > > > > But now that is having no effect. > > see the man page at > /usr/X11R6/man/man5/XF86Config-4.5x.gz I hadn't remembered that BlankTime, StandbyTimes, SuspendTime and OffTime could be set in this config file. But as the man page points out, the xset commands can be set at "run time." Does that mean run time for x server? If so, where would the xset command be put? Do you think I could put them into /usr/X11R6/lib/X11/xinit/xserverrc ? Doing it this way seems a lot cleaner than plugging in enormous numbers of minutes for each of the ServerFlags options in XF86Config. Haines Brown -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Simple little basic config questions
> I echo Colin's thought. Forget about "su" and use "sudo". It takes an
> extra 5 keystrokes per command, but it "just works", and in my opinion
> is better than forgetting you're root and doing something you don't want
> to do.
>
> apt-get install sudo
> visudo, add yourself a line similar to what's already there
> sudo command_to_be_run_as_root
Thanks, Kent, I'll follow your advice. But that brings up my next
elementary question: getting packages.
I ran # netselect-apt woody in the /etc/apt directory, and as a result
built a /etc/apt/sources list that had a US and a non-US site
uncommented. OK, so next I run (ignore line breaks):
# apt-get install sudo
Reading Package Lists... Done
Building Dependency Tree... Done
W: Couldn't stat source package list http://ftp.br.debian.org
woody/main Packages
(/var/lib/apt/lists/ftp.br.debian.org_debian_dists_woody_main_binary-i386_Packages)
- stat (2 No such file or directory)
... [same for three directories in each of the two source sites
listed in sources list]
W: You may want to run apt-get update to correct these problems
E: Couldn't find package sudo
Well, running apt-get update just gives me exactly the same thing as
above. I added the following http subsection to /etc/apt.conf to enable
internet sources:
Acquire
{
Retries "0";
// I added this next subsection:
http
{
Proxy "http://127.0.0.1:3128";;
Proxy::http.us.debian.org "DIRECT"; // Specific per-host setting
Timeout "120";
Pipeline-Depth "5";
// Cache Control. Note these do not work with Squid 2.0.2
No-Cache "false";
Max-Age "86400"; // 1 Day age on index files
No-Store "false";// Prevent the cache from storing archives
};
};
// Things that effect the APT dselect method
DSelect
{
Clean "auto"; // always|auto|prompt|never
};
Haines
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Simple little basic config questions
> hb> $ xhost +local > hb> non-network local connection being added to access control list > hb> $ su > hb> Password: > hb> # > > Well this doesn't prove anything: you have to run an X application as > root before you can know whether it worked or not. > > This is still a bad idea, though, because anyone who can log into your > box (via telnet or whatever) can access your X server. And anyone who > can access your X server can see every key you type, everything that > appears on your screen, etc. etc. Paul, you and Colin and Kent have persuaded me a) to use command line when I'm doing maintenance as root, b) run sudo. So now I need to undo the above. The man seems to suggest # xhost -local. Should that be done again as user? > It's not the default because, as above, it's a very insecure and silly > thing to do. If you were able to start X applications from the command > line after running "su" on a Red Hat box, like this: > > $ su > Password: > > # xclock > > then Red Hat must have opened up access to the X server, which is very, > very bad. I'm _SURE_ they've fixed this by now, if they ever did it at > all. I am persuaded. > You should be reading Colin's posts: he's got the right answers. Here > are some notes which might help you: I read your notes with interest, and indeed you helped clarify things. Only now, I've got somehow to undo the xhost command I issued before. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: GeForce4 MX440 with kernel 2.4.18-bf24
> On Thu, Oct 30, 2003 at 07:26:40AM -0600, Hoyt Bailey said > > Is anyone running the subject video card on the subject kernel. If so what > > driver are you using? If you know where can I get a copy of your driver? > > Unfortunately, GeForce4 cards won't work with the "nv" driver included > with X before X 4.2. Woody includes 4.1, so you'll either need to: > > 1) Use the "vesa" driver and have a low refresh rate and resolution > > or > > 2) get X4.2 or 4.3 for woody, www.apt-get.org has the apt source you >need. > > or > > 3) get the non-free binary-only nvidia drivers. The "nvidia-glx-src" >and "nvidia-kernel-src" packages make this rather easy. This is only >an option on x86, however. Rob, I follow you on points 1 and 2, but not 3. The nVidia has completely changed its driver policy, and now everything (driver, installation utility) comes in one self-expanding .run file. And this driver supports a range of their cards, including all the GeForces. So it's very simple now. What you do is to use a (downloaded or automatically compiled) interface to mediate the relation of your driver and kernel. Further, it is apparently free. Or at least I downloaded and installed the driver on woody and have not paid yet ;-). The driver's name is "nvidia." Haines Brown -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: GeForce4 MX440 with kernel 2.4.18-bf24
> Thanks you Haines. I have downloaded all the required stuff and > only need to know where to put it to install. I hope no one says it > dosent matter. That isnt the question. Where do I put the header > files? Where to put the driver files? > Sincere Thanks for your help; > Hoyt It _really_ doesn't matter ;-) If you are talking about the .run file you download, then put it where convenient. For example, I create an /opt partition to dump a variety of things, including applications that I must compile and install myself. If I had to use a default linux directory, I suppose either /usr/local/... for system-wide programs, or create ~/apps/... for a program that I privately owned by just one account. Others on the list will undoubtedly have wiser things to say about this. The nVidia installation program puts things in the right places for you. The only directory you need to know is the location of your kernel headers. Here is the command I used to tell the executable where these headers were (woody 3.1r1) (ignore line break) # sh NVIDIA-Linux-x86-1.0-4496-pkg2.run --kernel-include-path /usr/src/kernel-headers-2.4.28-bf2.4/include This include directory has a lot of subdirectories, but if you have the headers and burrow down, you will eventually encounter a lot of *.h files. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: How can I stop the console from blanking
I put this following into ~/.bash_profile: # to get rid of blanking of console: setterm -blank 0 but I was never able to find out were in /etc to put it for a global effect rather than just in account configurations. Haines Brown -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
question about Qt package needed
--- Start of forwarded message --- An application requires Qt 3.02 headers and libraries. Well, there are some 40 packages on my CD that have "qt" in their name. I went to the packages.debian.org and seached for Qt, but still too much, without anything looking quite right. First a general question. Is there some way to determine, without prior knowledge, what packages need to be downloaded? None of the packages on the debian site said anything like, "general Qt" support, or the "headers and libraries needed by Qt." The specific question: whch Qt packages in fact do I need to get Qt headers and libraries? Haines Brown -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Apt-get: move from cdrom to online-sources
I installed debian from cdrom, but now want to use apt-get (actually,
aptitude) to get on-line packages. To do this I ran netselect-apt
woody in the /etc/apt directory, and as a result built a
/etc/apt/sources list that had a US and a non-US site uncommented.
OK, so next I want to get the sudo package and run:
# apt-get install sudo
Reading Package Lists... Done
Building Dependency Tree... Done
W: Couldn't stat source package list http://ftp.br.debian.org
woody/main Packages
(/var/lib/apt/lists/ftp.br.debian.org_debian_dists_woody_main_binary-i386_Packages)
- stat (2 No such file or directory)
... [same for three directories in each of the two source sites
listed in sources list]
W: You may want to run apt-get update to correct these problems
E: Couldn't find package sudo
I don't understand this. Apparently apt-get reads my sources.list, but
can't use it. Running apt-get update has the same result. I added
the following http subsection to /etc/apt.conf Acquire section to
enable internet sources:
Acquire
{
Retries "0";
// I added this next subsection:
http
{
Proxy "http://127.0.0.1:3128";;
Proxy::http.us.debian.org "DIRECT"; // Specific per-host setting
Timeout "120";
Pipeline-Depth "5";
// Cache Control. Note these do not work with Squid 2.0.2
No-Cache "false";
Max-Age "86400"; // 1 Day age on index files
No-Store "false";// Prevent the cache from storing archives
};
};
// Things that effect the APT dselect method
DSelect
{
Clean "auto"; // always|auto|prompt|never
};
Where did I go wrong?
Haines Brown
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: question about Qt package needed
> On Sat, Nov 01, 2003 at 08:02:58AM -0500, Haines Brown said > > First a general question. Is there some way to determine, without prior > > knowledge, what packages need to be downloaded? None of the packages > > on the debian site said anything like, "general Qt" support, or the > > For a library "foo", the actual library is usually in "libfoo", perhaps > with a number appended. > > > "headers and libraries needed by Qt." > > Generally libfoo-dev, sometimes with a number after the foo. > > > The specific question: whch Qt packages in fact do I need to get Qt > > headers and libraries? =20 > > libqt3-mt-dev for the headers, which will pull in libqt3c102 which > contains the actual library. The "-mt" is for "multi-threaded" and > is a particular feature of that version of the library. The "c102" > comes =66rom the C++ ABI transition > (http://people.debian.org/~rmurray/c++transition.html) but will > eventually dissappear. Thank you for the thorough reply to my naive question. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Apt-get: move from cdrom to online-sources
> > OK, so next I want to get the sudo package and run: > > > > # apt-get install sudo > > Reading Package Lists... Done > > Building Dependency Tree... Done > > W: Couldn't stat source package list http://ftp.br.debian.org > > woody/main Packages > > > > (/var/lib/apt/lists/ftp.br.debian.org_debian_dists_woody_main_binary-i386_Packages) > > - stat (2 No such file or directory) > > ... [same for three directories in each of the two source sites > >listed in sources list] > > W: You may want to run apt-get update to correct these problems > > E: Couldn't find package sudo > > > > I don't understand this. Apparently apt-get reads my sources.list, but > > can't use it. Running apt-get update has the same result. I added > > apt-get update has the *exact* same error? update should try to go out > on the network and get the missing package lists from your debian > mirror. Well, no, at least not now when I try it. I was only looking at the end of all the response, for I can't seem to copy/paste from an xterm. However, this time did an eshell in emacs and saved the buffer that resulted from the apt-get update to a file. Here is the first few lines: Err http://ftp.br.debian.org woody/main Packages Could not connect to 127.0.0.1:3128 (127.0.0.1). - connect (111 Connection refused) Err http://ftp.br.debian.org woody/main Release Could not connect to 127.0.0.1:3128 (127.0.0.1). - connect (111 Connection refused) ... In the sources.list, the internet URIs are followed by CDROM URIs: rr cdrom://[Debian GNU/Linux 3.0 r1 _Woody_ - Official i386 Binary-7 (20021218)] unstable/contrib Packages Please use apt-cdrom to make this CD recognized by APT. apt-get update cannot be used to add new CDs News to me! So instead of apt-get install , I'd use apt-cdrom install ? But I prefer aptitude. Would it therefore be: # aptitude-cdrom install ? The intial response suggests to me that I'm misconfigured somehow. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Apt-get: move from cdrom to online-sources
> > Err http://ftp.br.debian.org woody/main Packages
> > Could not connect to 127.0.0.1:3128 (127.0.0.1). - connect=20
> > (111 Connection refused)
> > Err http://ftp.br.debian.org woody/main Release
> > Could not connect to 127.0.0.1:3128 (127.0.0.1). - connect=20
> > (111 Connection refused)
> > ...
>
> That means apt can't reach your proxy running on localhost port 3128.
> Does "netstat -pant | grep 3128" as root show it listening on any
> interfaces at all?
I got nothing back from this query, so presume not. In your comment,
what does "it" refer to? Apt? If so, apparently the following is
wrong/insufficient:
Acquire
{
Retries "0";
// I added this next subsection:
http
{
Proxy "http://127.0.0.1:3128";;
Proxy::http.us.debian.org "DIRECT"; // Specific per-host setting
Timeout "120";
Pipeline-Depth "5";
// Cache Control. Note these do not work with Squid 2.0.2
No-Cache "false";
Max-Age "86400"; // 1 Day age on index files
No-Store "false";// Prevent the cache from storing archives
};
};
Haines
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Apt-get: move from cdrom to online-sources
Jakob, a little light is begging to shine. In my reading I did not see anything about apt-cdrom, and so had to guess what had to be added to the config file. I guessed, wrong, it seems. > > Â Err http://ftp.br.debian.org woody/main Packages > > Â Â Could not connect to 127.0.0.1:3128 (127.0.0.1). - connect > > Â (111 Connection refused) > > Hello, > this means that your proxy is down. Remove the proxy settings from your /etc/ > apt.conf and run apt-get update again. I removed the http subsection I had put into the /etc/apt/apt.conf file, and ran the apt-get update. Not sure if I succeeded. There was a lot I missed because scrolled off my xterm, and I can't copy/paste from xterm to get a decent example. However, from what I see, there's a lot about apt-get update should not be used to add new cds; I should use apt-cdrom for that (now that I know such a command exists, the man page is somewhat useful). A set of warnings like this is repeated for each of the seven original disks I had used to install debian. In fact, the entire response that I can see consists of this, except for the three final lines: Reading package lists..., Building dependency tree..., E: some index files failed to download. So I'm not sure if my command succeeded or not. However, when I run # aptitude install , it does not seek for it on the web, but just a bunch of warnings that it could not stat source package list cdrom. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: How can I stop the console from blanking
> On Fri, Oct 31, 2003 at 08:39:09PM -0500, stan wrote: > > I'm setting up a machine that will rum umatended, and print some > > information on several console sessions. In testing, I find thta the > > console blanks afyer some period of time, even if daya is being writtten = > to > > it :-( > >=20 > > How can I fix this? > > setterm -blank 0 > > --=20 > Pigeon --and pigeon, where do you suggest he place that command so that it has global effect upon bootup? Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Apt-get: move from cdrom to online-sources
Rob, I think it's comming along:
> > If so, apparently the following is
> > wrong/insufficient:=20
> >=20
> > Acquire
> > {
> > Retries "0";
> > // I added this next subsection:
> > http=20
> > {
> > Proxy "http://127.0.0.1:3128";;
> > Proxy::http.us.debian.org "DIRECT"; // Specific per-host setting
> > Timeout "120";
> > Pipeline-Depth "5";
> >=20
> > // Cache Control. Note these do not work with Squid 2.0.2
> > No-Cache "false";
> > Max-Age "86400"; // 1 Day age on index files
> > No-Store "false";// Prevent the cache from storing archives =20
> > };
> > };
>
> Are you actually intending to run a proxy on this machine? If not,
> comment out the Proxy lines in this section and try "apt-get udpate"
> again.
I did as you said, literally (commented out just the two proxy lines),
and it may have gone better. This time I piped the output, so know I
captured the entire file. It starts with a lot of this:
Err cdrom://[Debian GNU/Linux 3.0 r1 _Woody_ - Official i386
Binary-7 (20021218)] unstable/contrib Packages
Please use apt-cdrom to make this CD recognized by APT.
apt-get update cannot be used to add new CDs
Ign cdrom://[Debian GNU/Linux 3.0 r1 _Woody_ - Official i386
Binary-7 (20021218)] unstable/contrib Release
I get this for all severel installation disks. Is this because the
disks are listed in my /etc/apt/apt.conf file? If so, should the cdrom
lines be removed from there; or should I just ignore all these error
messages?
Finnally, I get to what I hope is what I am supposed to see:
Hit http://ftp.br.debian.org woody/main Packages
Hit http://ftp.br.debian.org woody/main Release
Hit http://ftp.br.debian.org woody/contrib Packages
Hit http://ftp.br.debian.org woody/contrib Release
Hit http://ftp.br.debian.org woody/non-free Packages
Hit http://ftp.br.debian.org woody/non-free Release
Hit http://debian.teleglobe.net woody/non-US/main Packages
Hit http://debian.teleglobe.net woody/non-US/main Release
Hit http://debian.teleglobe.net woody/non-US/contrib Packages
Hit http://debian.teleglobe.net woody/non-US/contrib Release
Hit http://debian.teleglobe.net woody/non-US/non-free Packages
Hit http://debian.teleglobe.net woody/non-US/non-free Release
Reading Package Lists...
Building Dependency Tree...
"Hit" is a nice word. I hope it is what I should expect.
But it failed the test. I wanted to install sudo, which is not on my
distribution cdroms, and so ran:
# aptitude install sudo
All I get is a lot of warnings that it can't find the file on the
cdrom disks. No sign that it tried any web respositories. It is
reading the cdrom section of my /etc/apt/sources.list, but not its
deb-src http... entries.
Haines
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: How can I stop the console from blanking
> On Sat, Nov 01, 2003 at 06:16:59AM -0500, Haines Brown wrote: > > I put this following into ~/.bash_profile: > >=20 > > # to get rid of blanking of console: > > setterm -blank 0 > >=20 > > but I was never able to find out were in /etc to put it for a global > > effect rather than just in account configurations. > > # cat > /etc/init.d/noblank > #!/bin/bash > /usr/bin/setterm -blank 0 > echo 'Console blanking disabled' > ^D > # chmod a+x /etc/init.d/noblank > # ln -s /etc/init.d/noblank /etc/rc2.d/S99noblank > > (These blanking things would be rather more useful if they could be > configured to blank after a certain time with no input *and no output*.) Thanks, I see how that works (noblank needs to be executable), but assumed there was some obvious script file checked during boot. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Apt-get: move from cdrom to online-sources
> You have to update the local cache with what is available on the remote > site... > > apt-get update > > apt-get install sudo > > Should really do: > > apt-get update && apt-get upgrade > > First and foremost...! OK, I removed the references to cdrom in my /etc/apt/sources.list, and things run more cleanly. As a test I did # aptitude install sudo, and it worked. Now I turn to your recommendation and run: aptitude update && aptitude upgrade, and while the update seems to go, I'm told there's nothing to upgrade. I seriously doubt it: Reading Package Lists... 0% Reading Package Lists... 100% Reading Package Lists... Done Building Dependency Tree... 0% Building Dependency Tree... 0% Building Dependency Tree... 50% Building Dependency Tree... 50% Building Dependency Tree Reading extended state information... 0% Reading extended state information... 0% Reading extended state information... Done No packages will be installed, upgraded, or removed. 0 packages upgraded, 0 newly installed, 0 to remove and 0 not upgraded. Need to get 0B of archives. After unpacking 0B will be used. Do you want to continue? [Y/n/e/d/v/action/?] Is it your impression that the first group of responses suggest that apt-get update worked. But then why no packages needed to install. My installation was from a fairly old cdrom set of debian 3.0r1. Shouldn't there be an enormous about of material to upgrade? Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Apt-get: move from cdrom to online-sources
Greg, Your information of great help, but only leads to more questions. 1. Where does one read about such specifics as the need to have a security source to do an upgrade? What I've seen on apt says relatively little. 3. The address didn't work, and I also tried a couple others. When I tried # aptitude update && aptitude upgrade, the warining is that it couldn't stat the source package list. The aptitude log says nothing. Besides your example, here are the respositories I tried: deb http://security.debian.org woody/updates main contrib non-free deb http://security.debian.org/ stable/updates main Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
package purging and app configuration
I installed a dict front end (ding) and somehow inadvertently configured it for "" as background color. I can't figure out in the startup script how to correct that, and so decided simply to reinstall ding. I ran: aptitude purge ding, and then did a new installation. However, the script error remains, suggesting that purge did not remove configuration files. How should I have gone about removing an application so that its original configuration files are deleted? Haines Brown -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: package purging and app configuration
> /var/lib/dpkg/info/ding.list is a file that is part of the ding .deb > package. > > if you look at it, it will tell you where all the files that > ding installed are. > > /var/lib/dpkg/info/ding.postinst is the post-install script > /var/lib/dpkg/info/ding.postrm is the post-remove scirpt > .../ding.preinst is the pre-install script > .../ding.prerm is the pre-remove script > > check them and see which line is causing the problem. > > if needed, comment a line out, if this helps you remove or purge or > reinstall. and then re/install it and see if this fixes the problem. > > BUT if you have to change a script, its a bug. If you found it, > thousands of others will too, so file a report to speed up the bug-fix > process! Interesting, but I'm still stuck. I didn't see now to change the start up script to get around the problem. So I did the following in an attempt to start with a clean slate: # aptitude purge ding deleted /var/...archives/ding check in /var/lib/dpkg/info to make sure there were no ding files # aptitude update # aptitude install ding When I then run the application, I still get the same error (unknown color name "" in startup script) when I run # ding. Does that mean that there's definitely a bug? It ran once nicely, but in looking at the color preferences options, I thought I did it some damange, but apparently its more than just this. And yet there's no bug reports like this, and I keep thinking the problem is with me. All I needed was a simple graphical interface for dict (gsdict seems a nuisance), and I may have started with one that is problematic. -- Haines Brown [EMAIL PROTECTED] [EMAIL PROTECTED] www.hartford-hwp.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
install local deb file
I've spent some time pouring over apt-HOWTO, and just don't understand. I want to use aptitude to install a local .deb file. I place it in my /storage/debs. In the HOWTO, it said to do this: deb file:/storage debs\ But when I run aptitude update, it says the line is malformed. How does one use aptitude to install a local deb file? Haines Brown -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
source.list won't stat
I have the following lines in my /etc/apt/sources.list deb http://security.debian.org woody/updates main contrib non-free deb http://security.debian.org/ stable/updates main deb http://www.videolan.org/pub/videolan/debian $(ARCH)/ When I run Aptitude update, they won't stat. Haines Brown -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: package purging and app configuration
> > Interesting, but I'm still stuck. I didn't see now to change the start > > up script to get around the problem. So I did the following in an > > attempt to start with a clean slate: > >=20 > ># aptitude purge ding > >deleted /var/...archives/ding > >check in /var/lib/dpkg/info to make sure there were no ding files > ># aptitude update > ># aptitude install ding > > Wouldn't you also need to delete ~/.dingrc? More than that, I was able to correct the problem in that script and I'm back on the road. /var/lib/dpkg/info/ding.list did not mention this file, and so I didn't think to look for it there. Thanks Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: install local deb file
> On Sun, Nov 02, 2003 at 03:15:06PM -0500, Haines Brown wrote: > > I've spent some time pouring over apt-HOWTO, and just don't > > understand. I want to use aptitude to install a local .deb file. I > > place it in my /storage/debs. > > > > In the HOWTO, it said to do this: > > > > deb file:/storage debs\ > ^ What is this back slash? > > But when I run aptitude update, it says the line is malformed. > > > > How does one use aptitude to install a local deb file? The backslash was a typo (only in my message, not in sources.list). I was just copying the model in apt-HOWTO (I thought). The problem seems to be that aptitude update will take the storage directory above and append it to /var/lib/apt/lists/, but I'm not sure. I'm sure there's a very simple answer to my question. I followed the HOWTO and only managed to create a Packages.gz file, but no idea what to do with it. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: install local deb file
> > How does one use aptitude to install a local deb file? > > There may be a way to convince aptitude to do this, though i don't know > of one. It seems to me you may just be using the wrong tool for the > job. > > Is there some compelling reason not to use dpkg for your situation? > > dpkg --install > or > dpkg -i Well, I got the impression that aptitude would handle dependencies properly and keep current with the package situation. In any case, the dpkg -i does not work. I'll give the specifics: # dpkg -i /storage/debs/libdvdcss2.deb dpkg: error processing /storage/debs/libdvdcss2.deb (--install): cannot access archive: No such file or directory Errors were encountered while processing: /storage/debs/libdvdcss2.deb I tried also without the .deb extension. I also copied the libdvdcss2.deb file into /var/cache/apt/archives, but when I ran # aptitude install libdvdcss2, the file not seen. No better luck with # dpkg -i. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: package purging and app configuration
> Wouldn't you also need to delete ~/.dingrc? Jerome, you are quite right. I had relied on the list file and didn't realize that it omitted configuration files. Once I got the clue, sure enough, the file was there, and the script problem was obvious and easily fixed. Thanks. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: install local deb file
> On Sun, Nov 02, 2003 at 06:50:59PM -0500, Haines Brown wrote: > > In any case, the dpkg -i does not work. I'll give the > > specifics: > > > > # dpkg -i /storage/debs/libdvdcss2.deb > > dpkg: error processing /storage/debs/libdvdcss2.deb (--install): > > cannot access archive: No such file or directory > > So that file doesn't exist. Are you sure you put it there? Thanks for the reply. It turned out that the command had to be issued in that directory, apparently, not by supplying the path. Once I did that, it went OK, except that the package had an unmet dependency. The .deb extension should not be present. The comand # aptitude install works the same, except the error message is phrased a little differently. So I went to install the dependency (libc6 at version at least 2.3.2-1), and I found it is testing (currently libc6-dev 2.3.2-9). So I tried to cook up an address for testing in /etc/apt/sources.list, but my efforts failed, so instead I downloaded the .deb file (libc6_2.3.2-9_i386.deb). Now when I try # aptitude install libc6=2.3.2-9, I'm back where I started: libc6=2.3.2-9 (can't find the 2.3.2-9 version for libc6) libc6-dev=2.3.2-9 (can't find this version for libc6-dev) libc6=2.3.2-9_i386 (can't find 2.3.2-9_i386 version) and other combinations. I tried dpkg -i, and it said it could not access the archive. But when I tried dpbg -i on another .deb in my /storage/debs directory, it said it could not find it. That is a slighly different reponse, and I'm at a loss. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: OT Tape backup recomendations
You say you want a "new backup system," and did not specify it should be tape. I recommend that you consider an external USB drive for backups. It is the cheapest method you could use in terms of cost/mb, and does not run the danger of proprietary standards (I've got a bunch of old OS/2 DAT tapes I'd like to access, but that will not be easy, no longer having the commercial software, etc.). With tapes, you are probably stuck using the one machine that has the drive, while an external drive can be moved to any machine. Haines Brown -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: install local deb file
> Put it in /storage/debs/.=20 > > Just to clarify, I would do something like: > > echo "deb file:/storage/ debs/" >> /etc/apt/sources.list > dpkg-scanpackages /storages/debs /dev/null | gzip > /storage/debs/Packages.= > gz > apt-get update > apt-get install whatever Nick, this is one thing I originally tried, but failed. This time, it think I got a bit further. I'm trying to install libdvdcss2 to that I might view DVD on xine. I found it and put it into my local /storage/debs directory, along with two dependencies, libc6 and libdb1. I did the Packages.gz creation, informed storage.list of the location, and did the update. So far so good. The Packages.gz holds three packages, but since I want to install the libdvdcss2 primarily, I installed it, hoping that it would figure out the dependencies and install them, too. Apparently I was wrong. I get something like this: Reading package lists... Building Dependency Tree Reading extended state information Then removal of unused packages (including my Qt3 libraries, which I installed in preparation for installation of an application that will need them). Then it said that two of the three packages in Packages.gz will be intalled, and one package (the libc6) will be upgrades. Except for at least one group of undesired removals, sounds good. I'm told how many packages will be removed, installed and upgraded, and the space required. That sounds right. So I say, Continue, and it proceeds to write extended state information. But then: Err file: debs/ libsdb1-compat 2.1.3-7 File not found Err file: debs/ libc6 2.3.2-9 File not found Removing... So it was unable to find the two dependencies in the Packages.gz, and yet also apparently somehow managed to install the libdvdcss2 that depended on them, which I don't understand. I run # apt-cache pkgnames | grep libdvdcss2 libdvdcss2 which I take it means the installation of this packages suceeded. But the command also tells me that libsdb1-compat not installed, and libc6 not upgraded. Should I have have run # apt-get install libdvdcss2 libsdb1-compat libc6=2.3.3-9? I tried, but got the same Err file messages as above (although no Removals this time, of course). Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: install local deb file
> > I'm confused; 'dpkg -i' is quite happy to take absolute paths. What does > 'ls -l /storage/debs/libdvdcss2.deb' say? > > Are you sure that it isn't really > /storage/debs/libdvdcss2_someversionorother_i386.deb? > > > Once I did that, it went OK, except that the package had an unmet > > dependency. > > > > The .deb extension should not be present. > > I beg to differ; it certainly should. 'dpkg -i' takes the exact name of > the file, and does not do any magic with extensions. For aptitude, > you're correct that the extension should not be used, but aptitude > requires you to generate a Packages file or to use a properly set up > mirror somewhere else. Colin, Thanks. That paragraph a revellation on several counts. > If you've just downloaded the .deb file, use 'dpkg -i > libc6_2.3.2-9_i386.deb' to install it. A question about how to handle dependencies when using dpkg -i. What I'm trying to do is install libdvdcss2, which depends on my upgrading libc6, and that, in turn depends on libdb1-compat. If I run dpkg -i on just the libdvdcss2_1.2.5-0.2_i386.deb, I assume it will not know to look in the same folder for any files that might satisfy dependencies. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: install local deb file
Nick, thanks, but your second method (installing dependences first, then the package that needs them) does not work for me. Here's what I did. # cd /storage/debs // went to place where the .debs file is located # ls// I see listed libdb1-compat_2.1.3-7_i386.deb // and libc6_2.3.2-9_i386.deb # apt-get update Ign file: debs/ Release // Is this a problem? ... # apt-get install libdb1-compat Reading Package Lists... Done Building Dependency Tree The following extra packages will be installed: libc6 The following NEW packages will be installed: libdb1-compat 1 packages upgraded, 1 newly installed, 0 to remove and 0 not upgraded. Need to get 0B/3761kB of archives. After unpacking 160kB will be used. Do you want to continue? [Y/n] // I continue Err file: debs/ libdb1-compat 2.1.3-7 File not found Err file: debs/ libc6 2.3.2-9 File not found Failed to fetch file:/storage//storage/debs/libdb1-compat_2.1.3-7_i386.deb File not found. // I don't understand the syntax here. Is it looking in // /storage/debs? Failed to fetch file:/storage//storage/debs/libc6_2.3.2-9_i386.deb File not found. E: Unable to fetch some archives, maybe run apt-get update or try with --fix-missing? In case the order was critical, I tried with the other file first, but same results. The above suggests that apt is able to access the directory, see the files it, but not access them. The files are owned by user, but the command is run by root, so ownership not an issue. The files are -rw---. I assume they need only be readable. But another complication. The libc6 is an update, not a new installation. So I tried apt-get upgrade libc6: The following packages have been kept back libc6 I tried libc6=2.3.2-6, and got the same results as without the version specification. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: source.list won't stat (problem gone)
> On Sun, Nov 02, 2003 at 03:43:27PM -0500, Haines Brown said > > I have the following lines in my /etc/apt/sources.list > >=20 > > deb http://security.debian.org woody/updates main contrib non-free > > deb http://security.debian.org/ stable/updates main > > deb http://www.videolan.org/pub/videolan/debian $(ARCH)/ > >=20 > > When I run Aptitude update, they won't stat.=20 > > Run "apt-get update". You'll need to do this after each time you edit > your sources.list. My last line above, which is the equivalent of "apt-get update" is what resulted in the won't stat error in the first place. That is, I could not update. However, when I try it now it works. All I did was to uncomment the earlier URLs. No idea what changed (I don't _think_ I foolishly had a space at start of the lines). In any case, I now did a "# aptitude update && aptitude upgrade", and all went smoothly. Thanks for your patience. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: install local deb file
> On Tue, Nov 04, 2003 at 07:58:20AM -0500, Haines Brown wrote: > > # cd /storage/debs // went to place where the .debs file is located > > # ls// I see listed libdb1-compat_2.1.3-7_i386.deb > > // and libc6_2.3.2-9_i386.deb > > # apt-get update > > Ign file: debs/ Release // Is this a problem? > > ... > > # apt-get install libdb1-compat > > Reading Package Lists... Done > > Building Dependency Tree > > The following extra packages will be installed: > > libc6 > > The following NEW packages will be installed: > > libdb1-compat > > 1 packages upgraded, 1 newly installed, 0 to remove and 0 not > > upgraded. > > Need to get 0B/3761kB of archives. After unpacking 160kB will be > > used. > > Do you want to continue? [Y/n] // I continue > > Err file: debs/ libdb1-compat 2.1.3-7 > > File not found > > Err file: debs/ libc6 2.3.2-9 > > File not found > > I think your Packages file is screwy. Can you please post: > > * the contents of /etc/apt/sources.list; > > * the first block of text from /storage/debs/Packages or > /storage/debs/Packages.gz or whatever? Here is my sources.list. I've cleaned out comment lines: deb http://ftp.br.debian.org/debian/ woody main contrib non-free deb http://debian.teleglobe.net/non-US/ woody/non-US main contrib non-free deb http://security.debian.org woody/updates main contrib non-free deb http://security.debian.org/ stable/updates main deb file:/storage/ debs/ The last line refers to my separate "storage" partition on which there is: /storage/debs. Here I put I put .deb files and the Packages.gz I build from them. I assume you want me to unpack Packages.gz to inspect its innards. Here is the first block: Package: libc6 Version: 2.3.2-9 Priority: required Section: base Maintainer: GNU Libc Maintainers <[EMAIL PROTECTED]> Depends: libdb1-compat Suggests: locales, glibc-doc Conflicts: strace (<< 4.0-0), libnss-db (<= 2.2-6.1.1), timezone, timezones, gconv-modules, libtricks, libc6-doc, libc5 (<< 5.4.33-7), libpthread0 (<< 0.7-10), libc6-bin, libwcsmbs, apt (<< 0.3.0), libglib1.2 (<< 1.2.1-2), netkit-rpc, wine (<< 0.0.20021007-1) Provides: glibc-2.3.2-9 Replaces: ldso (<= 1.9.11-9), timezone, timezones, gconv-modules, libtricks, libc6-bin, netkit-rpc, netbase (<< 4.0) Architecture: i386 Filename: /storage/debs/libc6_2.3.2-9_i386.deb Size: 3730358 MD5sum: 93b2fe09ef8ebbb2612216502245a0c2 Description: GNU C Library: Shared libraries and Timezone data Contains the standard libraries that are used by nearly all programs on the system. This package includes shared versions of the standard C library and the standard math library, as well as many others. Timezone data is also included. source: glibc installed-size: 12808 Maybe you can see something amiss here. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: install local deb file
Colin, I suspect your directions fixed my problem, but I still fumble on the final installation. > OK, I would advise two things. Firstly, /etc/apt/sources.list should > contain this instead of the last line: > > deb file:/storage/debs ./ This was accepted when I did # aptitude update > Secondly, generate the Packages.gz file like this: > > cd /storage/debs > dpkg-scanpackages . /dev/null > Packages > gzip Packages A bit different than the procedure I originally used. I have in my /storage/debs three .deb files: a libdvdcss2, which is what I am really trying to install. It has two dependencies: the libc6 is just an update; a libdb1-compat, which I must install to satisfy the libc6 update. Although a bit complicated, I packed all three into the Package.gz file, hoping that # aptitude-install libdvdcss2 would do it. It did get to the point that it said it would upgrade 1 package and install 2, so aptutde install understood the logic of the situation. However, when I told it to proceed, I got errors. Apparently the livdcdcss2 was found, but not its two depencies: Err file: debs/ libdb1-compat 2.1.3-7 File not found Err file: debs/ libc6 2.3.2-9 File not found Here's the first section of my new Packages.gz file: Package: libc6 Version: 2.3.2-9 Priority: required Section: base Maintainer: GNU Libc Maintainers <[EMAIL PROTECTED]> Depends: libdb1-compat Suggests: locales, glibc-doc Conflicts: strace (<< 4.0-0), libnss-db (<= 2.2-6.1.1), timezone, timezones, gconv-modules, libtricks, libc6-doc, libc5 (<< 5.4.33-7), libpthread0 (<< 0.7-10), libc6-bin, libwcsmbs, apt (<< 0.3.0), libglib1.2 (<< 1.2.1-2), netkit-rpc, wine (<< 0.0.20021007-1) Provides: glibc-2.3.2-9 Replaces: ldso (<= 1.9.11-9), timezone, timezones, gconv-modules, libtricks, libc6-bin, netkit-rpc, netbase (<< 4.0) Architecture: i386 Filename: ./libc6_2.3.2-9_i386.deb Size: 3730358 MD5sum: 93b2fe09ef8ebbb2612216502245a0c2 Description: GNU C Library: Shared libraries and Timezone data Contains the standard libraries that are used by nearly all programs on the system. This package includes shared versions of the standard C library and the standard math library, as well as many others. Timezone data is also included. source: glibc installed-size: 12808 Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: source.list won't stat (problem gone)
> In future, it would be a *massive* help to us if you could include the > full error that you see. Ideally without wrapping, without > indentation, etc. I understand, but since I can't copy from xterm (don't know if it is broken or I'm expecting too much of it), I have to pipe output, etc. So in quick exchanges I take shortcuts that I shouldn't. My apologies. > > However, when I try it now it works. All I did was to uncomment the > > earlier URLs. No idea what changed (I don't _think_ I foolishly had a > > space at start of the lines).=20 > > Were you online at the time? "apt-get update" will hit each source in > the file, so if you're offline or your DNS is broken or...whatever, then > apt-get will fail. Yup, I'm always on line. But this good to keep in mind. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: source.list won't stat (problem gone)
> From: Colin Watson <[EMAIL PROTECTED]> > > [Could you attribute the person whom you're quoting, please?] Sorry - this not working yet, and it sometimes slips my mind to copy a line from from the header. > To copy from xterm, highlight the region you want to copy with the > left mouse button. To paste somewhere else, middle-click. Sorry - not working (with middle mousebutton; I don't expect span to work because I didn't say to emulate three-button when I installed). I'm using a MS Intellimouse temporarily until I can install a driver for a better one sitting ready (Logitek optical). But my question was actually about how to paste from the keyboard. Thanks Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
setting hardware clock from NIST
I'm a RedHat refugee, and sometimes I can just transfer things to debian, and sometimes not. I have an executable script, "time.rc" which has: #! /bin/bash rdate -s time-b.nist.gov clock -w I installed rdate, and that seems to work fine to set the system clock (or at least the system clock jumped to 6 sec from what my local telco says the time is). However, the "clock" utility, which writes system time to the hardware clock is native to RedHat. So my first question is, what is debian's equivalent to "clock". All it did, with the -w option, was to set the hardware clock from the system clock. I presume the former is GMT, and so there is a time offset invoved. Second, where to put it? I placed a copy of my time.rc into /etc/init.d, and then created a symlink to it in /etc/rc2.d so that the hardware clock is reset on boot, and also in /etc/cron.daily, so that the clocks are reset daily according to NIST. Will this work; is there a better arrangement? Haines Brown -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Testing URIs not working
Every step toward getting packages installation to work has been like
pulling teeth. Please be patient.
I want my /etc/apt/sources.list to support the installation of testing
packages, and so I appended the line:
deb http://ftp.debian.org testing main
But when I run "# aptitude update" with it, I get this error:
W: Couldn't stat source package list http://ftp.debian.org
testing/main Packages
(2 No such file or directory)
I tried these addresses as well:
deb http://ftp.br.debian.org/debian/ testing main
deb http://debian.teleglobe.net/non-US/ testing/non-US main
I discovered that my adding a URI for testing packages to
/etc/apt/sources.list and then running "# aptitude update" had no effect
on the files listed in in /var/lib/apt/lists.
Should it? That is, after running "#aptitude update", should a new
line appear like the one that the Warning above reports not finding?
This happened to me before when I needed to register the location of
a local repository of deb files on my machine. For a long while, I
got the same warning when I ran "#aptitude update", that it couldn't
stat the URI. After a while, it mysteriously popped up in
/var/lib/apt/lists ("_storage_debs_Packages") and my problem ceased.
How do I force apt to register the testing address in
/var/lib/apt/lists?
Haines Brown
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: source.list won't stat (problem gone)
> From: Rob Weir <[EMAIL PROTECTED]> > > > To copy from xterm, highlight the region you want to copy with the > > > left mouse button. To paste somewhere else, middle-click. > >=20 > > Sorry - not working (with middle mousebutton; I don't expect span to > > work because I didn't say to emulate three-button when I > > installed).=20 > > If you re-run "dpkg-reconfigure xserver-xfree86", you can change this. Yes, but why do I need to emulate 3-button when I have a middle scroll-button on the mouse? > > But my question was actually about how to paste from the keyboard. > > Somethine like this should work: > > $ apt-get -y install blah > /tmp/log 2>&1 That's not quite what I meant. I know I can do this, but it's not really copying and pasting to and from xterm with the keyboard. One disadvantage is when the program I'm running in xterm involves a prompt, it is hidden by a pipe. So I'm in the habit of using instead the script utility. It allows me to respond to prompts, and it records my inputs as well. I'm beginning to suspect that the reason I'm not able to copy/paste with xterm is that I discovered that gmp was not installed as part of the basic debian package. I installed it, and when I next restart the X server, I'll see if I can use the clipboard with xterm. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: setting hardware clock from NIST
> From: David Z Maze <[EMAIL PROTECTED]> > [EMAIL PROTECTED] (Haines Brown) writes: > > > I have an executable script, "time.rc" which has: > > > > #! /bin/bash > > rdate -s time-b.nist.gov > > clock -w > > It's almost certainly better to find a local time server and not > hammer on the NIST's; I'd also use ntp (ntp-simple package) to keep > your clock up-to-date while the system is running. Yes, quite something. It's been years since I've looked into this, and it is now much more established. For one thing, if I recall correctly, there were only 2-3 servers. Following your recommendation, I got the address of a local time server and also installed ntp-simple. > That setup is probably fine, though I'd do either a cron script or an > init.d script, not both (if your machine spends a lot of time shut > down, anacron can run delayed cron jobs at boot time). If you do want > an init script, I'd also make it more policy-compliant; try working > from /etc/init.d/skeleton. Thanks for the advice. I normally don't reboot for months at a time, and so need to sync clocks with cron. I'll kill the symlink in /etc/rc2.d. I assume that /usr/sbin/ntpd is the executable that I should symlink in /etc/cron.daily. In reading the doc, I see that ntpq, run without argument, is a way to do a simple test of ntp functionality. However, that does not seem to be part of the ntp-simple package, nor is it itself a package. Have you used ntpq, and if so, how? Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: source.list won't stat (problem gone)
> From: Colin Watson <[EMAIL PROTECTED]> > On Wed, Nov 05, 2003 at 09:23:16AM -0500, Haines Brown wrote: > > From: Rob Weir <[EMAIL PROTECTED]> > > > If you re-run "dpkg-reconfigure xserver-xfree86", you can change this. > > > > Yes, but why do I need to emulate 3-button when I have a middle > > scroll-button on the mouse? > > Sounds like you might have told X the wrong mouse type, if that doesn't > work. I suspect as much. I find that I have two mouse sections in my X configuration, and their Protocol is "PS/2", while my Microsoft mouse is an IntelliMouse, which has its own "IntelliMuse" protocol. Actually, I've got in hand a Logitech optical, and figured I'd have to investigate configuration in greater depth to get it working, and so hadn't worried much up to this point. The Microsoft mouse was just an emergency measure when my old Logitech finally gave up the ghost after years of abuse. But the wheel/button is irrelevant in any case, for I was trying to paste with the keyboard the material I had previously highlited in xterm. Maybe that simply can't be done. Maybe it can with gterm (?) and I've got a dim recollection of that. > Let me save you a wild goose chase: X doesn't require gpm for any kind > of mouse support. gpm is only for the console. Installing it might make > your life slightly more complicated, because suddenly you have to worry > about repeaters and /dev/gpmdata and weird stuff like that. Thanks! I quickly uninstalled it. I don't generally feel any need for a mouse when operating in console. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: source.list won't stat (problem gone)
> From: Pigeon <[EMAIL PROTECTED]> > > > On Wed, Nov 05, 2003 at 09:23:16AM -0500, Haines Brown wrote: > > I suspect as much. I find that I have two mouse sections in my X > > configuration, and their Protocol is "PS/2", while my Microsoft mouse > > is an IntelliMouse, which has its own "IntelliMuse" > > protocol.=20 > > "Intellimouse" protocol is for serial-port mice. These are dead rare > nowadays. For a PS/2 Intellimouse you need Option "Protocol" "ImPS/2", > and to get the scroll wheel working, Option "ZAxisMapping" "4 5". Thanks. That's a big help, especially if I fail to get my optical mouse working. > Try mdetect. I looked around, but couldn't get much information from googling. Apparently mdetect is usually used with some kind of front end, but I've no idea what such a front end might be. Does it require this other program, or can you just run it by itself to probe a mouse? Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
User looses access to DISPLAY
Finally bringing to conclusion my first debian (3.0r.1) install, I ran aptitude update. When I subsequently rebooted, user cannot start x, but root can. I reconfigured XF86Config-4 by running dpgk-reconfigure , using the simple configuration, which worked before, and kept the same values. But that did not help. There are a couple warnings in the XF85config log (a set of fonts does not exist; can't open APM), but no errors. When root runs # echo $DISPLAY, it returns: :0.0. But when user runs $ echo $DISPLAY, there is nothing. I'm currently logged in as root, and in an xterm su to user to handle my mail. I find that user can start emacs, but not other x clients. The error returned when I try is that there is no display name and no DISPLAY variable. I don't think I did anything that would affect authentification. Haines Brown -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: source.list won't stat (problem gone)
> From: Pigeon <[EMAIL PROTECTED]> > > > Try mdetect. > I've only ever used it once... just ran it from the command line and > it told me what the mouse was. No front end required. That was what I was hoping. Seems harmless and I'll have to try it (once I can get recover access to X server, that is :-( Thanks Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: User looses access to DISPLAY
> From: Kent West <[EMAIL PROTECTED]> > > Haines Brown wrote: > > Finally bringing to conclusion my first debian (3.0r.1) install, I > > ran aptitude update. When I subsequently rebooted, user cannot start > > x, but root can. > Create a new user, and see if that user has the same problems. The new user does not have the problem. > Also check > enjae[westk]:/home/westk> sudo cat /etc/X11/Xwrapper.config > Password: > allowed_users=console > nice_value=-10 The above is what I have. allowed_users is console. > > When root runs # echo $DISPLAY, it returns: :0.0. > > But when user runs $ echo $DISPLAY, there is nothing. > > If I log in as a normal user into a text console (no X11) and run "echo > $DISPLAY", I too get nothing displayed. I don't think the lack of having > a display env-var set is a problem. OK, but I don't understand. Is the value of DISPLAY in the environment of root all the time except when a user acquires it by starting X? Haines Brown -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: source.list won't stat (problem gone)
> From: Rob Weir <[EMAIL PROTECTED]> > On Thu, Nov 06, 2003 at 05:51:30AM -0500, Haines Brown said > > > From: Pigeon <[EMAIL PROTECTED]> > >=20 > > > > > Try mdetect. > > =20 > > > I've only ever used it once... just ran it from the command line and > > > it told me what the mouse was. No front end required. > >=20 > > That was what I was hoping. Seems harmless and I'll have to try it > > (once I can get recover access to X server, that is :-( > > To get X to autodetect your mouse (and monitor and video card): > > $ apt-get install discover mdetect read-edid > $ dpkg -P --force-depends xserver-xfree86 xfree86-common > $ apt-get install xserver-xfree86 xfree86-common Thanks, Rob, that's something I'll definitely try. I'll be spending the rest of the day trying to figure out why I can't start X (root can, and also a test user). Only error message is when logged in as user, I can't get access to DISPLAY. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: setting hardware clock from NIST
> From: David Z Maze <[EMAIL PROTECTED]> > If you install ntp-simple it will start a daemon that will > periodically poll the time servers and gently keep your clock in > sync. (If you're five seconds off, that time will be made up > gradually, rather than abruptly shifting the clock.) No need to set > up a cron job. Thank you very much, David. Knowing that installation of ntp-simple automatically sets up the clock syncronization makes things very simple. Also, the 6-sec error I noted will then disappear in time. > > In reading the doc, I see that ntpq, run without argument, is a way > > to do a simple test of ntp functionality. However, that does not > > seem to be part of the ntp-simple package, nor is it itself a > > package. Have you used ntpq, and if so, how? > > It's in the base ntp package, which is suggested by ntp-simple. For > looking at my own ntp daemon, 'ntpdc -s' will give you a short summary > of what other time daemons you're talking to, and 'ntptrace' will show > the synchronization chain from yourself to a stratum-1 time server. I installed ntp (which I didn't realize I had failed to do). and ran the following. Don't know if it means ntp-simple not running: # ntpdc -s ***Server reports data not found # ntptrace localhost: stratum 16, offset 0.11, sync distance 0.31587 0.0.0.0: *Not Syncronized* The first puzzles me, and sounds like there's no connection with the server. The second sounds like the time is retrieved, but there has not been sufficient time to achieve complete sync. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: User looses access to DISPLAY (resolved)
> Haines Brown wrote: > > Finally bringing to conclusion my first debian (3.0r.1) install, I > > ran aptitude update. When I subsequently rebooted, user cannot > > start x, but root can. Sorry to reply to myself, but thought the information might be useful. After spending much of the day sniffing about, a likely cause was that my update had filled /var, and that would prevent X from starting. However, that was not so in my case, although it is something good to keep in mind. I finally discovered that changing my window manager executable in ~/.xsession from icewm to icewm-session, as I was advised to do, caused the X server to die once I had booted. Changing it back to "icemn" allowed user to start X. Haines Brown -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Testing URIs not working
> From: Wayne Topa <[EMAIL PROTECTED]> > > Have you read any of the below? > > /usr/share/doc/newbiedoc/newbiedoc-html/apt-get-intro/index-apt-get-intro.html > /usr/share/doc/newbiedoc/newbiedoc-html/apt-get-intro/apt-and-install.html > /usr/share/doc/newbiedoc/newbiedoc-html/apt-get-intro/info.html > /usr/share/doc/newbiedoc/newbiedoc-html/apt-get-intro/intro.html > /usr/share/doc/newbiedoc/newbiedoc-html/apt-get-intro/more.html > /usr/share/doc/newbiedoc/newbiedoc-html/apt-get-intro/index.html > /usr/share/doc/newbiedoc/newbiedoc-ps/apt-get-intro.ps > /usr/share/doc/Debian/apt-howto/ch-apt-get.en.html Not on my system. Sniffing with google, I found a source, but I would have to join the newbieDoc group, and this would mean registering with Yahoo. They would not be part of a package, would they? Thanks, though, for the suggestion Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Testing URIs not working
> From: Wayne Topa <[EMAIL PROTECTED]> > > [teach to fish mode] Yes, by all means! > apt-cache search newbiedoc > newbiedoc Well, what is obvious to the experienced user, is something entirely new to the newbie. I realized that # aptitude search would search for uninstalled packages, but I didn't infer that this meant _all_ the stable packages that were available. Now that I realize this and tested it, indeed I could have spared myself the google search with its dead-end lead. In retrospect, since there's a list of all packages on my machine, I knew in principle what's available, but I was just not connecting those dots. > apt-cache show newbiedoc I tried this on an uninstalled package (mysql), and nothing was returned. But shouldn't this command have exposed information contained in the cache, even for uninstalled packages? > Other packages to download for self help: > apt-howto - A Guide to APT > doc-linux-nonfree-html - Linux HOWTOs in HTML format (non-free) > or > doc-linux-nonfree-text - Linux HOWTOs in ASCII format (non-free) > > [teach mode off] Wayne, don't apologize. Much appreciated. Just to show you I'm in a "learning mode", when I could not install doc-linux-nonfree-text, I ran apt-cache search and found out that it is properly "doc-linux-text". That I downloaded. Sure enough, it pops up in /usr/share/doc/doc-linux-text ;-) Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Yahoo says I'm a Spam-o-Rama (via DynDNS)
> From: Colin Watson <[EMAIL PROTECTED]> > There is no need for a static IP to do this properly, and dynamic > DNS is irrelevant. The issue is that you're trying to send mail > directly from a dynamic *IP address*. People on dynamic IP addresses > should always relay mail through some other machine, such as their > ISP's mail server. This takes about a minute to configure and after > that you won't notice it. My outgoing to AOL addresses has been blocked this past year. The error message says it won't accept dynamic addresses. I called them up and they stonewalled me; all I got was that blocking their customers from getting my mail was some kind of "feature." My web hosting service, whose smtp server I use for my outgoing mail, says that it has nothing to do with blocking, but I should get in touch with the originating point, my ISP. My ISP says it has nothing to do with mail, but check with the hosting service that offers use of their smtp server. Does your comment suggest that my hosting service could in fact redirect mail I send to its smtp server so that it goes out under its own static address? They may not be accomodating, but can this be done in principle? Would it overcome AOL's peculiar desire to block my address? Haines Brown -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
handling "broken packages"
I've looked at the APT manual and rummaged about on line, but can't get a clear idea of just to do with "broken" packages. I get the impression that the term "broken" implies some unresolvable dependences, but I'm not sure what to do about it if the depencency is uninstallable. Here's an example. I go to install libqt3-mt-dev. E: Unable to correct problems, you have held broken packages [The word "held" seems to suggest that the dependency is available, but it is somehow "broken".] E: Unable to correct dependencies, some packages cannot be installed. E: Unable to resolve some dependencies! ... The following packages have unmet dependencies: libqt3-mt-dev: Depends: libqt3-dev (=2:3.0.3-20020329-1) but it is not installable. If I try to install ligqt3-dev, I can't do it because it has uninstallable dependencies as well. Haines Brown -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Source list help
> Shaul Karl wrote: > > >On Sat, Nov 08, 2003 at 12:03:38PM -0700, [EMAIL PROTECTED] wrote: > >> > >>I was wondering where to get source to add to mu source.list in ordre to > >>get packages from unstable. > > > >deb http://http.us.debian.org/debian unstable main contrib non-free > >deb-src http://http.us.debian.org/debian unstable main contrib non-free I've been following the thread with interest. While not the nearest repository, I appended the above lines to my /etc/apt/sources.list and as root ran: # aptitutde update. I get the errors: W: Couldn't stat source package list http://http.us.debian.org unstable/main Packages (/var/lib/apt/lists/http.us.debian.org_debian_dists_unstable_main_binary-i386_Packages) - stat (2 No such file or directory) ... Haines Brown -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Source list help
> From: John Peter <[EMAIL PROTECTED]> > >>>deb http://http.us.debian.org/debian unstable main contrib non-free > >>> > >>> > ^^^ > > >>>deb-src http://http.us.debian.org/debian unstable main contrib non-free > >>> > >>> > ^ ^ > > > Haynes, you have the answer right in front of you - you made a mistake > in the line you wrotte > in your sources.list . > > I think it's better to select the lines indicated above, copy them and > paste them in you souces.list > file .That way you won't make a mistake . Well, I'm chagrined at having used an invalid example--the result of haste and frustration, no doubt. I'm sorry. However, this does not seem to be the problem. Let me illustrate: I paste the first line above into my sources.list. Here I paste it back into this message from my sources.list as a double check: deb http://http.us.debian.org/debian unstable main contrib non-free So, that is _exactly_ what is in my /etc/apt/sources.list, following the lines for stable and local archives. I then run aptitude update and, get the following return (again, copied and pasted): # sudo aptitude update ... Reading extended state information... Done W: Couldn't stat source package list http://http.us.debian.org unstable/main Packages (/var/lib/apt/lists/http.us.debian.org_debian_dists_unstable_main_binary-i386_Packages) - stat (2 No such file or directory) W: Couldn't stat source package list http://http.us.debian.org unstable/contrib Packages (/var/lib/apt/lists/http.us.debian.org_debian_dists_unstable_contrib_binary-i386_Packages) - stat (2 No such file or directory) W: Couldn't stat source package list http://http.us.debian.org unstable/non-free Packages (/var/lib/apt/lists/http.us.debian.org_debian_dists_unstable_non-free_binary-i386_Packages) - stat (2 No such file or directory) W: You may want to update the package lists to correct these missing files When I remove the offending line, aptitude update does what it is suppsed to, visits the sites and reports the hits. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Source list help
> From: Wayne Topa <[EMAIL PROTECTED]> > > I've been following the thread with interest. While not the nearest > > repository, I appended the above lines to my /etc/apt/sources.list and > > as root ran: # aptitutde update. I get the errors: > > Network error?? Server down? > > I use more then one server for each dist for just that reason. Possibly, but to check, I modify the URL that always works for me (http://ftp.br.debian.org/debian) by changing its distribution from "woody" to "unstable" (checked to make sure the site does archive this distribution), and it still generates the error. And I go back and forth with several tries. So it seems that it's not network error or server down. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Source list help
> From: Wayne Topa <[EMAIL PROTECTED]> > > Possibly, but to check, I modify the URL that always works for me > > (http://ftp.br.debian.org/debian) by changing its distribution from > > "woody" to "unstable" (checked to make sure the site does archive this > > distribution), and it still generates the error. And I go back and > > forth with several tries. So it seems that it's not network error or > > server down. > Don't see anythhing about ftp.br.debian.org/debian tho. Maybe they > were updating themselves?? Not likely. I use the URL regularly without difficulty, but when I try unstable or testing, it becomes invisible when I run aptitude update. I can go back and forth, installing a file from that site, never with any delay or difficulty, and when I switch to the testing or unstable, it always fails. I check the archive on line, and it does have those archives. I've been working on this on and off for a week, and so its not just an unusual glitch. I"ve tried many URLs, but all won't stat with testing and unstable distributions. Thanks for your input. Hope some lurker might have some suggestion about what to check. Spent a good part of the morning going back over apt documentation, and there's nothing in it that seemed helpful. I don't see that /etc/apt/apt.conf needs editing, but maybe I'm wrong. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Source list help
> From: Wayne Topa <[EMAIL PROTECTED]> > all I do to prevent that is have 2-3 sites for each dist. That way if > one or two are not there (for whatever reason) I get each dist anyway. > > I have seen the error you are getting but I don't care since I have a > valid list from some other site. Takes longer to do the update but, > to me, it's worth it. I tried several without luck. I understand your point about have one or two exta sites just to get through with one of them, but I've been playing with several this past week. I'll keep trying other URLs, but it seems odd that repeated attempts with a range of URLs that work for stable don't work for testing. Oh, well, I guess I just do more random testing to see if anything happens to work. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Identifying kernel soruce for 2.4.18-bf2.4
I need to have installed the kernel source in order to compile a driver, but have difficulty finding the source for 2.4.18-bf2.4. Is the answer that the "bf2.4" is just an addition, and that the kernel source for my kernel is actually just 2.4.18? Haines Brown -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Repairing mime files
Years ago (OS/2?) I had a little utility that would repair the header of a mime file so that I could un-mime it. Anyone know of such a thing for linix/debian? My search effort failed. Specifically, I have an acquaintance that insists on sending out documents to a small distribution list using MS Exchange V6. The header says X-MX-Has-Attach, but there's no mime or part header between the message header and the MIME. That is, it goes right from To:... to the MIME encoding. I'm running rmail under emacs. Haines Brown -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Social Engineering. {was: Re: Opium [was: Re: freelance sysadmining -- superlong -- [WAS: "Red Hat recommends Windows for consumers"]]
> From: Ron Johnson <[EMAIL PROTECTED]> > > It seems to me that the "most successful" would be those who can > master the social needs (get good grades from approved testbooks, > etc), while still being able to think outside the box. Ron, I can't give your remark the attention it deserves because this topic is way OT, but I find no way to resist saying just a little. Your implication that students who know how to work within the system tend to get good grades strikes me as accurate, as is the implication that good grades don't necessarily correlate with "intelligence." However, there are some serious problems lurking behind these platitudes. We have no satisfactory measure of intelligence, certainly not the discredited "IQ" test. We therefore must always specify just what we mean by intelligence, particularly the context in which it is manifested. There's probably different kinds of intelligence that are only loosely correlated with mental acuity. For example, I have an intuitive sense that some people really pick things up quickly and frequently come up with unusual insights. I suppose that is a sign of their native intelligence. Unfortunately, that ability correlates poorly with success in school and with success in life. Many of these people have difficulty with formal college courses because they find it difficult to fit into the grove of the system's implicit assumptions and goals and lack the presumed fundamental information and skills. Many of these people don't do particularly well on the job, either, for they may not have some necessary social skills or not take seriously the rules governing work, or perhaps are simply not given opportunity or reason to act creatively. While I think of such people as sharp, and I find them interesting and fun to be with, their mental sharpness may do them really little good. That is, they are not really "intelligent" in the sense that they can think constructively "outside the box." Sometimes they are not so much "outside the box" as "off the wall". Creative thinking, just as creativity in the arts, requires discipline, the mastery of certain skills and possession of a body of knowledge. The issue, I believe, is not thinking "outside the box," as if honed skill and the possession of information is somehow the enemy of creative thinking, but being creative in terms of one's skills and knowledge. Intelligence or creativity without material constraint to give it meaning and direction seems vacuous (lecture here on thermodynamic engines, and how structures give rise to improbable outcomes by constraining a dissipative system's degrees of freedom, etc., etc., etc.). For example, with baroque music there were many geniuses and a lot of hacks whom we seldom hear. The former struggled with and bent conventions to create extraordinarily powerful music. On the other hand, today there's not much left of the old rules, and so it becomes very difficult for a composer to produce anything but soon-to-be-forgotten "elevator music." We are heirs of an ideological contradiction between freedom and determinism, while in fact, in human affairs and to a significant extent in the natural sciences, causality is probabilistic. This suggests the contradiction is false. Haines Brown -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Installing xserver
I installed a woody base system and then install the xserver-svga required by my nVidia GeForce3 card and a window manager. Then from a root command line I ran startx. However, all I got as a result was a black screen with speckles (not x's grey screen), with a set of blinking underlines. My guess is I was looking at a brief paragraph or sentence followed by a cursor, that recycled and lacked a proper horizontal sync. I took a look at the /ect/X11/XF86Config file, and to my eyes it was very strange. In part perhaps because I'm familiar with the RedHat file, but there were suspicious aspects. My current configuration has a vertRefresh of 48-120 Hz, but the debian file ends up with 50-85Hz. I used "medium" level configuration, which asked me for the "best" performance of my card. I entered [EMAIL PROTECTED] (my card is capable of [EMAIL PROTECTED] Hz), which was closest. And so why did the configuration reduce the horizontal refresh so drastically? Other oddities. Nowhere in the device section or elsewhere does the configuration see my nVidia card and what should be its its nv driver. Instead, the graphic card's name is generic and the driver is simply VGA2. I assumed that the nv driver was compiled in the stock 2.4.18-bf2.4 kernel. Am I wrong? How would I know? In the monitor section, there are three dozen commented lines describing all possible resolutions and refresh rates, and associated with each is an uncommented line with a "ModeLine" statement. Is this enormous mess normal in debian? Did I fail to install a needed package? Shouldn't my video card have been detected? Should I expect the debian xserver configuration to be close enough to the RedHat so that I can use one to some extent as a model for changing the other? -- Haines Brown [EMAIL PROTECTED] [EMAIL PROTECTED] www.hartford-hwp.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Installing xserver
> Haines Brown wrote: > > I installed a woody base system and then install the xserver-svga > > required by my nVidia GeForce3 card and a window manager. Then > > from a root command line I ran startx. However, all I got as a > > result was a black screen with speckles (not x's grey screen), > > with a set of blinking underlines. > Ok, first of all you don't come into a debian mailing list and start > flaming Debian for not being like RedHat. Wow! Not my intention at all. I was merely suggesting that I was familiar with the XF86Config file in RedHat and needed to know to what extent that enabled me to spot strange things going on in my debian's XF86Config file. I'm sorry to have been misunderstood. > You might want to try using the xserver called "xserver-xfree86" > instead . AFAIK, the xserver-svga stuff uses XFree version 3.x, > which I think will not have support for you GeForce.You will > probably have better luck with the driver in XFree 4.x which is used > by xserver-xfree86. Aha! I had difficulty deciding among a list of x servers in what must have been an obsolete list. Didn't realize they were not XFree 4.x. > Once you get done with that, try using the utility called > 'dexconf'. In my brief experience with it, what it does is use the > vesa driver and somehow manages to get a basic X display up (no > acceleration and that stuff). Then you can try experimenting and > making new configurations with either "dpkg-reconfigure > xserver-xfree86" or "xf86config" or "xf86cfg". Thanks, I'll do that after I uninstalled the x server I've got, removed XF86Config and installed the proper x server. > Nvidia has their own proprietary drivers if you want to get decent > 3D performance. That is probably best left for another posting after > you get X working :) I'm aware of the difference between the limited nv driver and the nVidia driver and have some experience installing the latter, so not too intimidated by it. Thanks for your patience. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Installing xserver
> Follow the instructions in the NVIDIA readmes EXACTLY. (When they > refer to a "fully configured kernel source tree", they mean one in > which you've done "make ?config" and "make dep".) While I was planning to recompile the kernel first and then install the driver as you suggest, is that order important? > The "nv" driver of XFree does provide a basic 2D driver that works > for my GeForce2, but I don't know if it will work for GeForce3. Yes it does, but I don't know about the GeForce4. Thanks for the tips. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Installing xserver
> > My current configuration has a vertRefresh of 48-120 Hz, but the > > debian file ends up with 50-85Hz. > > Just fix it in the file. Thanks. Good to know that the information in XF86Config reflects actual monitor capability rather than impose a restriction on it. > > Should I expect the debian xserver configuration to be close > > enough to the RedHat so that I can use one to some extent as a > > model for changing the other? > > Yes, they both use XFree86. If you have a working configuration > under Red Hat you can probably come very close to having something > that works by just copying over the XF86Config file. Reassuring. Again, Thanks. I've always been nervous about messing with the XF86Config because I'm haunted by the old warnings that if you don't get it right, you can blow up your monitor! Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Installing xserver
Joris, Thanks for the lead. It's always nice to have some hardware detection. Does it build a database so that when new hardware is installed, it opens a dialog for you to configure it during the boot process? The discover is a frontend for hardware detection tools, and you mentioned one for the mouse (mdetect) and one for the monitor (read-edid). Are there backends for other kinds of hardware, such as scsi or a NIC? Discover can be set up to make the boot process aware of hardware and ensure the necessary modules are loaded. Is this meant as a backup in case module.conf fails to do its job? Discover has a friend, which is "discover-data." That seems to be a database for common hardware, again to ensure drivers are properly loaded. Do you feel it is really necessary or useful? Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Help with XMMS radio, specifically KPFB
In connection with this thread, is there a web site that keeps track of various internet radio and TV resources that can be used by xmms or xine? I find that the information that accompanies xine is inaccurate. Haines Brown -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Help with XMMS radio, specifically KPFB
> http://www.shoutcast.com/ is a good place for mp3 streams. > http://www.webradio.com lists radio stations and has links to streams > but no mp3 streams. Pamela, Thanks. Although www.webradio.com seems to have come down, I was able to use your information to do a little search on line and pulled up some other useful sites. As for mp3 video, did that, too, although "WebTV" tends to be something else. These sites seem to assume a person has a Flash plug in or RealPlayer. I discovered I could play CSPAN with xine, although I doubt I have yet any RealPlayer or Flash support. Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: boot-floppies package
Floris, I have also found it advantageous to create a grub boot floppy. You can then select among kernels or fiddle with the bootloader configuration actually in the course of a boot. If you do have a running system, then simply have root visit grub's working directory and copy the two stage files over to an unmounted floppy: # dd if=stage1 of=/dev/fd0 bs=512 count=1 # dd if=stage2 of=/dev/fd0 bs=512 seek=1 Then mount the floppy and copy splash.xpm.gz to it, create a file named grub.conf and a symlink to it from menu.lst. And then set up the grub.conf file as per the manual. Haines Brown -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: boot-floppies package
> On Sun, Jun 01, 2003 at 06:06:08AM -0400, Haines Brown wrote: > ... > > If you do have a running system, then simply have root visit grub's > > working directory and copy the two stage files over to an unmounted > > floppy: > > > > # dd if=stage1 of=/dev/fd0 bs=512 count=1 > > # dd if=stage2 of=/dev/fd0 bs=512 seek=1 > > > > Then mount the floppy and copy splash.xpm.gz to it, create a file > > named grub.conf and a symlink to it from menu.lst. And then set up the > > grub.conf file as per the manual. > > I seriously doubt this is sane:) > According to the grub.info (node: Images): > > While Stage 2 cannot generally be embedded in a fixed area as the > size is so large, Stage 1.5 can be installed into the area right > after a MBR, or the boot loader area of a ReiserFS or a FFS. > > In other words, putting stage2 directly after the bootrecord destroys > any filesytem on the disk, better use stage1.5 or do a true install on > the floppy as I posted recently or as described in the FAQ (node FAQ). Ouch! I hope I didn't give bum advice :-(. I understand your point about the MBR, but if the grub files (besides the two files above, I also copied over the 1_5 fs drivers for about 200 Kb altogether) are on diskette, and I don't see how the warning would apply. I've been merrily using the floppy to boot both RedHat (ext3) and Debian (ext3 and reiserfs). Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: boot-floppies package
Carel, Now I'm really getting worried, and trying to recall were I picked up the suggestion to create a grub floppy (it wasn't my own idea). I was from Linux Journal, but I can't seem to pin down the issue. I also find a reference to doing so in the clone-HOWTO, section 5.3. Also http://www.mathematik.uni-marburg.de/sys/os/linux/install/mkgrubfd.html here's another venture: http://www.uruk.org/orig-grub/install.html I find that the Linux Journal article was from the May issue: http://www.linuxjournal.com/article.php?sid=4622 These discussions propose two or three somewhat different approaches to creating the floppy, and I see that I didn't use the Linux Journal method after all. But these methods would not see to have any bearing on your concern about overwriting the floppy's filesystem. > > > In other words, putting stage2 directly after the bootrecord > > > destroys any filesytem on the disk, better use stage1.5 or do a > > > true install on the floppy as I posted recently or as described > > > in the FAQ (node FAQ). > As stated before, there is really limited space on a floppy in > between the MBR and the actual file system. (Oh, and I'm not even > sure there is enough room there on a floppy to stack stage1.5) > Depending on the size of stage2 you're very likely to have > overwritten part of the filesystem. Whether you'll ever notice this > depends on your usage, but to me it seems a disaster waiting to > happen. Whether the grub files are somewhere else on that disk > really doesn't matter, you've ruined the filesystem. If stage2 does intrude on the file system, what exactly do you mean by the "file system?" The diskette's partition table? The 1_5 fs drivers? > The ruined file system might lead to you overwriting part of stage2 > whenever you put something on the floppy, like changing menu.lst, or > putting the other grub files on disk. At least you suggest the floppy is in danger of failing. But beyond ending up with a non-functioning floppy, are you suggesting it might affect the file system on the hard disk? If so, how? > Ofcourse if you first put those files on the flop, change menu.lst, > and only then use dd to put stage1 and stage2 in place, you might be > lucky and overwrite only some of those superfluous grub files. I'm not sure what surperfluous files you refer to. I run multiple file systems and so need some of the drivers, but, of course not all. But since one can never know, you are implying that a grub floppy is a BAD IDEA to begin with? Haines -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
