RE: postfix help
From: Robert Waldner [mailto:[EMAIL PROTECTED] > >Do you mean DynDNS? It sure ain't free - according to their > >pricing page (at www.dyndns.com) there's a $50 pa charge, > >plus the cost of registering a domain name. > > Try www.dyndns.ORG ;-) Ah. Never thought of trying that :-) Thanks, Paul
RE: postfix help
From: Eric Gillespie, Jr. [mailto:[EMAIL PROTECTED] > > There's no such animal. I mentioned DynDNS in my post, and i'm > mentioning it again. It's a free service that will give you a > valid domain name which can be updated everytime you're IP is > reassigned. Works great for those of us stuck on dialup links. Do you mean DynDNS? It sure ain't free - according to their pricing page (at www.dyndns.com) there's a $50 pa charge, plus the cost of registering a domain name. Paul.
APM: Poweroff actually does a reboot
Hi, I have a PC with APM support at home. I've installed Debian (Potato) onto it and built my own kernel with APM support compiled in (I needed to build my own to add the APM real mode calls option, as the default kernel crashes on a poweroff call). Now, I can successfully shut down, but the "poweroff" command shuts down to power off mode, and then immediately reboots instead of switching the power off. Is this a known/common problem (I vaguely recall having a similar problem on this PC with Win98, before I removed it, and I have also had the same problem on a laptop running Windows NT). Is it a BIOS configuration thing, or is it something in the kernel options I need to change? Or do I need a BIOS upgrade (it's some version of a Phoenix BIOS, and the suppliers don't do downloadable upgrades :-( and the prople who supplied the PC are refusing to admit that there's an issue as "we don't support Linux" :-( :-() Thanks for any help, Paul.
RE: boot error
From: Ron Rademaker [mailto:[EMAIL PROTECTED] > On Mon, 19 Jun 2000, Sven Burgener wrote: > > > Anyone know why do I get the following error upon bootup: > > > > insmod: /lib/modules/2.2.15/misc/unix.o cannot create > > /var/log/ksymoops/2619212757.ksyms Read Only Filesystem > > > Looks like your root file system is mounted read only, try to do (as > root): touch /a (if no error occurs, don't forget to do rm > /a). If it's mounted read only, you can remount it read-write, but > you better check out your /etc/fstab. But the message comes at bootup. AFAIK, the root FS is always mounted read-only in the first instance, and is then remounted read-write later in the boot sequence. So this looks like a problem in the base Debian boot sequence, trying to insmod unix.o too early in the boot process, before root is remounted. BTW, I have this problem too, so I can confirm that it occurs in a clean Potato (unofficial CDs from about 25th May) install. See bugs #51379 (seems to imply that it's fixed in modutils 2.3.7-1, but it's still there in my version, 2.3.11-6), #50921 (similar), #49059 (different, but also ksymoops and the boot sequence). Paul. PS Is there really no search facility in the bug tracking system? I had to get a full listing, and then use my browser's search. And that only finds bugs with the word in the title :-(
RE: problem reading vim online dox
From: Nate Bargmann [mailto:[EMAIL PROTECTED] > > I'm not sure as I've seen the reference to the $VIMRUNTIME > env variable > in the vim docs. Since it seems vim is "hardwired" to look in > /usr/share/vim, perhaps /usr/share/vim should be a sim-link to a real > directory of /usr/share/vim56 or whatever the prefered > installed version. Vim should be set up with two relevant directories - the first is $VIM, which falls back to /usr/share/vim if not defined (at least on the version of vim which comes with Potato - it's a compile-time option if you build vim yourself). The second is $VIMRUNTIME, which defaults to $VIM/vimNN where NN is Vim's version number (56 for 5.6). So you should have the system vimrc in /usr/share/vim/vimrc, and the runtime files in /usr/share/vim/vim56/, and then everything should work with NO environment variables needed. This is what you get if you do apt-get install vim vim-rt. You need vim-rt, as that contains the runtime stuff (the docs, syntax files, etc). I'm forever installing vim but forgetting vim-rt - was that the original poster's problem? You shouldn't need to set environment variables or move files. If you do, there's a big bug in the vim packages (and I know there isn't in the versions I'm using - 5.6.070-1) Hope this helps, Paul.
Merging distribution CDs
I have a set of four potato unofficial CDs, and a massive hard disk, and I'm bored of swapping CDs. What is the best way of copying the full contents of the CDs onto my hard disk so that I can install off there? I assume that mkdir -p /dist/cd[1-4] mount /cdrom (cd /cdrom && tar cf - .) | (cd /dist/cd1 && tar xf -) # repeat 3 more times works, with appropriate sources.list lines, but can I merge the CDs, to make a single directory tree - ideally in such a way that I could use this as a base for a mirror of the FTP site and then update regularly... Thanks for any suggestions, Paul.
Strange message - eth0: too much work in el3_rx
I'm getting strange messages on the console of my Debian machine, namely eth0: too much work in el3_rx Can anyone give me any hints as to (a) what it means, (b) whether it indicates a problem, and(c) how to get rid of it (as it corrupts my working screen). Thanks, Paul Moore. PS If I need to supply any more information, let me know - I'm not sure what might be useful...
RE: How to find a package containing a given executable
From: William F. Dowling [mailto:[EMAIL PROTECTED] > > I do this from cron every night: > > #! /bin/bash > cd /usr/local/fromDebian > DEBIAN=ftp://ftp.debian.org/debian/dists > wget -o LOG-unstable-contents $DEBIAN/unstable/Contents-i386.gz > mv Contents-i386.gz Contents.unstable.gz > > Then I could do > zgrep gvim /usr/local/fromDebian/Contents.unstable.gz > or the like to answer the question "where does gvim come from?" Thanks for the tip - this does just what I want. Actually, I have a potato CD, which has the Contents file on, so I can do this offline, as well. Just one other question - I had a problem yesterday with /etc/papersize not being found. Searching the Contents file got me nowhere, nor did dpkg -S /etc/papersize on a system with it installed. It turns out that /etc/papersize is a configuration file for libpaperg, and as such it wasn't in the list of files installed. How would I find out where this comes from? Thanks again, Paul.
RE: Notice: GR to remove non-free support from Debian
From: Brad [mailto:[EMAIL PROTECTED] > > While getting rid of non-free is a noble goal, i don't feel > that Debian > can do it now without losing support from some parts of the Free > Software community (look at the reaction over KDE, and then think of > Netscape, LyX, etc etc etc) and without losing a good measure of > functionality. i also question the removal of software that is open > source but not DFSG free because of restrictions on commercial sale. I agree with this sentiment. Debian is by far my preferred Linux distribution, but the "DFSG free or nothing" attitude is a little hard-core for me. I don't see any problem with segregating non-DFSG-free stuff from the fully DFSG-free software, but rejecting it altogether from Debian does little to help. I am aware that one of the key distinguishing features of Debian is that it is a strictly free system - however, focusing solely on this misses the equally important point, that Debian is by far the most robust and consistent Linux distribution, which is a strong argument that free software can and does produce quality software. The endless issues over free vs non-free and other license-related issues makes Debian look more like license nit-pickers than anything else. This doesn't seem to me to be a good image to have. It's not done the GNU project much good, and it would be a shame if Debian had the same problems. The present situation is a good compromise between ideals and realism. Why change something that works? Paul.
Configuring X for two possible screens
Hi, I have a laptop which I can run either using the internal 800x600 screen, or an external monitor which will run at 1024x768. I'd like to be able to set up the system so that I can start X to run on whichever is connected (eg, I could have aliases startx-internal and startx-monitor). I can set up an XF86Config file which includes monitor definitions for both monitors. The vide card is the same in both cases. What I don't know, and can't find in the online docs, is how to run startx to tell it which monitor is currently plugged in... I guess I may need to add more stuff to my XF86Config - I just don't know what! Any hints, pointers to relevant documentation or preferably complete solutions (:-)) gratefully accepted. Paul.
Where do I find administrators' documentation?
Hi, I don't know for sure if this is the best place to ask (debian-doc looks more like it's for the people developing documentation, rather than for asking *about* it - tell me if I'm wrong and I'll ask there), but here goes... I've used Debian off and on for a while now, and the thing I find most difficult to get to grips with is the system administration in the debian-specific areas. Things I've come up against in the past include - package management (beyond the basic apt-get install) - the alternatives system - X windows and the "menu" system - window managers and configuration thereof - general user config (and more generally, applying system-wide defaults) - users and groups (specific question, why have a group for each user?) - what is the impact of (say) changing /etc/x directly, on the package management system - and how do I assess this impact before making changes? In general, I've hunted round and found answers to most of the questions I've had, but it has been a fairly ad-hoc and time-consuming process. There are now books appearing on the market purporting to be "Debian Linux" books, but they are generally fairly general, with minimal Debian-specific content. My basic question is - is there any document available (preferably for download) which covers the specific issues relating to how Debian administration is done, where that is different from "generic" Linux admin? If not, can anyone give me some pointers on how to get to grips with this sort of thing? I'm only running a small single-user Debian workstation, and I don't want to get massively involved in systems admin, but at the moment I feel like I'm losing control of the system altogether :-( Thanks in advance, Paul. PS My specific benchmark for whether something is what I'm after tends to be whether the alternatives system is described, whether the reason for having a group for each user is explained, and some good overviews on how to do system config without upsetting apt when an upgrade occurs. Nothing I've seen yet covers all of this - although the FAQ is otherwise the best document I've seen...
RE: kernel-image 2.2.15
From: Paulo J. da Silva e Silva [mailto:[EMAIL PROTECTED] > > What I did: I ran dselect and marked my kernel image as "on > hold" (find your kernel image in the selection list and type > an "="). Unrelated question - how do you find what packages are on hold, and/or unhold them? I just did an apt-get dist-upgrade to potato (finally got a CD!!!) and found that the kbd package was held, for no reason that I could discern. I tried to find what I should have been looking at to notice this earlier, bit I couldn't see anything. I managed to force kbd to upgrade, but I'd rather know how to unhold it, so the dist-upgrade could work by itself... A method not involving dselect would be useful - I dislike dselect in general, and in particular I can't use it easily over a telnet session, due to stupid colour-handling problems in my telnet client which make dselect unreadable and which I haven't the time to fix :-( Thanks, Paul
How to find a package containing a given executable
Hi, One of the most common queries I have with the Debian packaging system, which I don't know how to make, is to search for the name of the package(s) which install a given file or files. For instance, the most recent such case I have was looking for the gvim executable. As it happens, this is provided by vim-gtk, but that isn't obvious. Clearly, on an installed file, I can use dpkg -S, but what about looking for where I install something from? If there isn't a simple tool to do this (or even if there is!) is there a more structured or navigable list of what packages do what in Debian (no, dselect doesn't qualify, IMHO)? Paul.
RE: Callback HOWTO ?
Oliver Schoenknecht wrote: > Hi there, > > I am urgently searching any documentation regarding the creation > of a callback server and the client fitting to this... Do > you know any location where I may obtain this documentation > / HOWTO ? There's a callback mini HOWTO at the Linux Documentation Project. http://metalab.unc.edu/mdw/HOWTO/mini/call-back-mini-HOWTO.html I don't know if it's any use to you. Paul.
Name resolution from Windows network
Hi, I havd a Debian box set up connected to my office LAN. I am getting an IP address from the site DHCP server, and am getting name resolution from the site DNS servers. The problem is that the site is totally Windows-based, and my Debian box is not integrating as nicely as I would like. I have absolutely no influence over the site network setup, and even if I did, I'd never get anything changed for the sake of Linux :-( My specific problems are as follows First, the site name resolution is done by a combination of DNS and WINS. Fixed-IP machines have DNS names, so I can resolve their names by the usual means (ping machinename, and so on). DHCP-allocated machines have unusable names in DNS (dhcppc-NNN-NNN, based on the IP address, defeating the point :-( ), and "sensible" names are resolved using WINS. So I can successfully resolve with Samba, but not with normal IP tools. As the site DNS server doesn't forward requests to WINS, and as the Linux name resolution itself can't use WINS directly, I was wondering about setting up named on the Linux box and making it run as a simple cacheing DNS resolver which falls back, when it can't resolve from the site DNS, to trying WINS. I've searched the various DNS mailing lists I can find, but there's nothing useful there. It seems to me that if I could run a "hook" of some sort from named if direct resolution fails, I could do this (a trivial Perl script could do the WINS->IP lookup via Samba, and return the result). Is this possible at all??? Second, my Linux box gets an IP address from DHCP. However, it doesn't (can't??) respond to the DHCP server passing its configured hostname (which gets published to the WINS, or maybe DNS, server - not sure which). As a result, other machines on the network can't resolve the machine name to its IP address. Can Samba be set up to respond to WINS broadcasts requesting the machine name - or can the Linux machine be set up to publish its name "properly" (where by properly, I probably mean precisely the opposite - suitable for a MS server, as opposed to the proper way :-( )? Thanks in advance for any help. Could any responses be copied direct to me, as I'll have to go off the list for a few days soon (I'm offsite and I can't handle the backlog of traffic which will be left on my return...) I'll check the archives as well, but personal copies would help. Paul Moore.
RE: Building my own apt archive
From: Ron Rademaker [mailto:[EMAIL PROTECTED] > > The easiest way would be to use the Packages.gz from the > mirror and the > directory structure as used on the mirror, when you update using that > Packages.gz, it will seem as if there are much more packages, that you > don't all have, however if you don't need them that's not a problem. No, that wouln't work. I have debs from a lot of places (unofficial packages, updated versions backported to slink, etc, etc) and I want to combine them onto one CD. So I have to build my own Packages.gz Paul.
Building my own apt archive
Another question on apt... I've downloaded a lot of .deb files from various places on the Internet, and I'd like to put them on a CD or something for safekeeping. It strikes me that the best way of doing this would be to put them in some sort of format which would allow me to use apt to install off it. Can somebody point me to a document which explains how to build an apt-style archive? I've looked in the manpages, but there isn't anything which immediately springs out at me... I think that at the least I need a Packages.gz file, but I'm not sure how that gets set up. Also, do I need a fixed directory structure? Thanks, Paul.
RE: Using apt offline
From: Pichai Asokan [mailto:[EMAIL PROTECTED] > I have done precisely this. That's good to know. Looks like I may have guessed right... > However, I ran the apt-get update on all the machines. I do not know > exactly how to duplicate that on other machines. I'm guessing that I copy /var/state/apt/ Paul.
RE: Using apt offline
From: Christian Surchi > On Wed, May 24, 2000 at 01:57:31PM +0100, Moore, Paul wrote: > > > 1. apt-get dist-upgrade --download-only on one machine > > 2. Transfer the downloaded files to all machines > > 3. apt-get dist-upgrade --no-download on all the machines. > > try apt-zip! That seems to work based on dselect. I want to do an apt-get dist-upgrade, which AFAIK dselect can't do... Actually, another issue is that I won't have access to the ftp/http sources on the non-networked machines - how will apt on those machines "know" what it should do - do I need to copy some other parts of the /var/.../apt stuff? Paul.
Using apt offline
Hi, I have a small number of machines which are currently running slink. I'd like to upgrade them, at least to frozen and possibly even to woody. The simple answer is apt-get (dist-)upgrade. However, there are a couple of issues with this. First, I pay for internet downloads, so I want to minimise download time, and second, at least one of the machines is not connected to the internet. The machines' configurations are pretty much identical, so I would like to download the necessary files *once* and upgrade all the machines offline. The problem is knowing how to get apt to work this way. I think that what I could do is 1. apt-get dist-upgrade --download-only on one machine 2. Transfer the downloaded files to all machines 3. apt-get dist-upgrade --no-download on all the machines. The only difficulty I have is with step (2) - is it enough to transfer the /var/cache/apt/archives directory? If not, what is the best way? More generally, what is the best way for me to keep a number of machines reasonably up to date, given 1. The only CDs I can get are slink (2.1r4) until potato hits the UK CD market. 2. I have limited network bandwidth. 3. The network bandwidth costs money. Thanks in advance, Paul Moore.
RE: Errormessage: invalid character in exportstr (?)
From: Joachim Trinkwitz [mailto:[EMAIL PROTECTED] > since some time (on several machines with newest woody) I get an error > message when logged in as root (not as a normal user): > >invalid character 32 in exportstr for export PATH > > This happens with all commands where bash is involved. In my PATH are > the common directories, echo $PATH gives > "/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin: > /usr/bin/X11" > > It's not doing any harm, but it's annoying anyway ... > > What's happening and how to avoid this message? Character 32 is a space. Have you somehow got a space in your PATH? Try echo "[$PATH]" and look out for spaces. (The brackets will help you see spaces at the start and end...) Paul.
RE: Problem with Linux PC getting at NT WINS server
From: Eric Hanchrow [mailto:[EMAIL PROTECTED] > > >>>>> "Paul" == Moore, Paul <[EMAIL PROTECTED]> writes: > > Paul> 1. The box cannot "see" these netbios names for other > Paul> machines on the LAN. I can ping machines with static IP > Paul> addresses by name, as (I guess) the names are resolved via > Paul> DNS. I can also ping dynamic IP machines via their > Paul> DHCPPC-names. But I can't ping using the netbios name > Paul> (which I guess is resolved via the WINS server) > > That's as I would expect. In order for you to use NetBIOS names, > you'd have to somehow tell your Linux box to use WINS for name > resolution. But as far as I know, Linux *in general* doesn't know how > to do that. (smbclient, however, does know how.) Right now I think > Linux only knows how to use DNS and /etc/hosts for host name > resolution. I've just done some further browsing, and have found some docs on nsswitch.conf - which seems to allow for user-supplied resolution methods. So I guess the question is whether such a method exists which dows NetBIOS lookups... Paul.
RE: Problem with Linux PC getting at NT WINS server
From: Moore, Paul [... Description of WINS name resolution problem omitted ...] > > Can anybody tell me what I might need to do to get this working? > Another (possibly) relevant point: I can connect from the Linux box using WINS names with smbclient. The problem appears to only be with things like ping, telnet, etc. So it's "normal" name resolution rather than WINS-specific stuff. Don't know if this narrows the problem down at all Paul.
Problem with Linux PC getting at NT WINS server
Hi, I have a Debian laptop (2.1r2 slink, with kernel 2.2.14 and the latest DHCPCD, PCMCIA and netbase packages) which I'm connecting to my company network - which is basically NT/Windows based. The basic connection is fine, I get an IP address and access to the DNS nameservers and network via DHCP. However, most of the network runs (somehow) on WINS. That is, the "official" name of a network client is something like DHCPPC-245-118.rundc.uk.origin-it.com. The hostname is a DHCP-generated value, which is not static. The "normal" name for a PC is something like UKDCW3920 [.rundc...], which is managed via WINS/NMBD services hosted on the same machines as the DNS nameservers. I have two problems on the Linux box. 1. The box cannot "see" these netbios names for other machines on the LAN. I can ping machines with static IP addresses by name, as (I guess) the names are resolved via DNS. I can also ping dynamic IP machines via their DHCPPC-names. But I can't ping using the netbios name (which I guess is resolved via the WINS server) 2. The box does not publish *its* netbios name correctly. I can ping from an NT box to ukrup480 (the Linux machine's name) but it resolves the IP address as 127.0.0.1 - obviously not right. So the WINS server seems to know about ukrup480, but gets its IP address wrong... I have installed SMB networking in the Kernel and the SAMBA client software. My smb.conf file is basically as it came, with the following changes: wins support = no wins server = os level = 0 domain master = no local master = no preferred master = no As far as I can tell looking at the documentation, there's nothing else which needs changing. But I'm definitely NOT an expert in this - it's all "magic" which just works on the NT machines. Can anybody tell me what I might need to do to get this working? Thanks for any help, Paul Moore.
RE: Upgrading libc6 from slink to potato (no network!)
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] > > [EMAIL PROTECTED] (Moore, Paul) wrote: > >Aargh. In order to run dhcpcd on 2.2.14, I need the version from > >unstable (frozen, I guess). But that depends on the unstable version > >of libc6, which depends on the unstable version of debianutils, which > >depends on the unstable version of libc6 > > > >If I force an install of debianutils, the libc6 install fails on the > >preinit script, with no indication of why. > > libc6 has just been fixed. From the changelog: > > * Pre-Depends were a bad idea > - quit using readlink. > - removed pre-depends on debianutils. > > When 2.1.3-5 hits your nearest mirror, do the download. Yahoo! This worked fine. I have now upgraded to kernel 2.2.14, with the latest pcmcia-modules and dhcpcd, and the machine picks up a network address fine - straight "out of the box" install. Thanks for all the help, Paul. PS I still have another problem with SAMBA/NMBD, but that's for a separate message...
Upgrading libc6 from slink to potato (no network!)
Apologies if people have seen this before. I initially sent it via the news gateway at our site as I was having trouble getting subscribed, but I'm not convinced that the gateway is making it to the full mailing list (other messages I've sent that way are not in the archives). Now that I've got myself subscribed, I'm resending just in case it hasn't appeared - I really need some help! Apologies for the waste of bandwidth... Paul. On Mon, 28 Feb 2000 13:52:47 +0100, Paul Moore <[EMAIL PROTECTED]> wrote: >On Thu, 24 Feb 2000 00:41:44 +0100, <[EMAIL PROTECTED]> wrote: >>1. Decide whether you want to run a 2.0 or a 2.2 kernel. DHCP is >>an area where the two kernel series require different tools. I think >>you'll be better off in the long run if you upgrade to kernel 2.2.x >>now. (Debian 2.1r2 is still 2.0-based by default, but has worked fine >>for me after I switched to a 2.2 kernel.) > >I will look at 2.2.14 and the associated updates needed to use it with >slink. Aargh. In order to run dhcpcd on 2.2.14, I need the version from unstable (frozen, I guess). But that depends on the unstable version of libc6, which depends on the unstable version of debianutils, which depends on the unstable version of libc6 If I force an install of debianutils, the libc6 install fails on the preinit script, with no indication of why. Something is wrong here. How do I upgrade a clean slink system to use the frozen/unstable version of libc6? Please, no-one say "apt-get upgrade" - the WHOLE POINT is that I don't have a net connection!!! Tell me what .debs to download, how to dpkg -i them in, and I'll be happy. But don't suggest an upgrade - if I could have run potato, I'd never have started with slink. Thanks for any help - I'm getting VERY frustrated here! Paul.
Graphics Card: S3 3D - compatible?
Hi, I'm looking at buying a new PC sometime soon - I've just seen a *very* good looking deal for a 350MHz Pentium II system. As usual, my main compatibility worries are with video and sound cards. The video card is described by the supplier as an AGP S3 3D card, with a 365 chipset. I've looked through the hardware compatibility HOWTO and the XFree86 website, and I can't see mention of this card or chipset specifically. Can anybody confirm that this card will work OK on Debian (basically Hamm at the moment...) I'll want to use X, and possibly OpenGL (Mesa). Probably not graphics intensive stuff generally, but I'd like to look at getting Quake and some other games running (it's not the end of the world if I fail, but I'd have to run them in DOS/Win95, so maybe it's bad enough... :-) Actually, does anybody know what this card is like, in general - is it 3D accelerated or not, do I need to check things like how much memory it has, will it run Quake II at mega-accelerated speeds, etc etc? If it's not up to much, does anybody have any suggestions as to a good card to get? I've been thinking of one of the ATI [EMAIL PROTECTED] cards - are they a good bet? The sound card is described as a SoundBlaster 16 Compatible, made by SoundPro. Again, will this be supported, and/or is it a good card? The only real use I have for sound is likely to be for games (both under Linux and Windows), so it's not a disaster if it's low spec, but I'd like something reasonable... I'm sure that there are some compromises being made in this system, but frankly, I'm not sure where they are. The deal is too good to just ignore, though... Any help (or pointers to useful sources of information) would be much appreciated. Thanks, Paul Moore.
What's in kernel 2.1/2.2?
Hi, I see that people are hoping for kernel 2.2 to be released by Christmas. I've looked (honest!) and yet I can't see a straightforward list of the major new features expected in 2.1/2.2. Can anybody point me at something I could look at? (Short of the sources :-) Thanks, Paul Moore.
Advice on Hardware
I'm thinking of buying a new PC. However, I want to be sure my hardware will be supported. I've read the Hardware HOWTO, and I *think* I understand it, but there are a number of things I am still concerned about: * The supplier wasn't 100% clear on the exact video card spec - it's a Jetway make, using S3 ViRGE chipset. Is this likely to be OK? * The video card is an AGP model. All I can find in the HOWTO and the XFree86 notes is that "some" AGP cards work. Will this one??!? * The sound card is a "Creative Labs PCI 64" card. Is that likely to work? It's a plug and pray card - does that cause me problems? The rest of the spec seems unlikely to be a problem - Pentium II 333MHz, BX 100MHz motherboard, 64M RAM, 4.3G EIDE UDMA disk, 32x CD ROM (IDE). Does anybody know of any issues or relevant questions I should ask? Thanks in advance, Paul Moore.
RE: How do you use su under X?
>From: [EMAIL PROTECTED]:[EMAIL PROTECTED] >Cookies are just random keys that are assigned to an X session (by >xdm) when you log in. In order to display anything in that session you >have to have that key. That's a bit odd - I don't use xdm, just startx from the console... Nevertheless, thanks for the explanation. Paul.
RE: How do you use su under X?
>From: [EMAIL PROTECTED]:[EMAIL PROTECTED] > >Chip Grandits <[EMAIL PROTECTED]> writes: >| As has been strongly encouraged, I spend most of my time on my linux box as >a >| mortal user. >| Every once in a while I need to go 'root-in' around some system >| files to set up >| a new application >| that I've downloaded (or for whatever reason?). >| Sure I can type >| $ su root >| at the promt and become root >| but I cannot use the x-server - any attempts result in a message >| >| Xlib: connection to :0.0 refused by server >| Xlib: Invalid MIT-MAGIC-COOKIE-1 key >| ... >| Also use xhost program... >| >| Imagine my horror to find I have an invalid magic cookie! Do I >| really have to >| use xhost in order >| to use an xwindows session started by another user? > >Yes, or you can use the authority file. Personally, what I do in this >situation is: > >% su - >Password: *** >% cp ~user/.Xauthority ~/ >% > >~/.Xauthority is the cookie file. I don't remember if you have to set >the DISPLAY variable? If so I do > >% export DISPLAY=:0 > >Replace "user" above with whatever user you log on as. Of course, if >you have what is essentially a single user system > >xhost + localhost > >is much simpler. I have noticed this, as well. I never had this sort of problem on Red Hat or SuSE. I assume that it's something you can set (presumably a security issue, where the other systems are more permissive). However, as I'm most definitely not experienced in X admin, I've no idea what's going on (to be honest, I don't even know what these cookie things are supposed to be...) Can anybody explain? Actually, is there any good (beginner's) documentation on X (either online or in book form)? The only books I've ever seen are the O'Reilly ones, which seem to be out of print these days... Actually, good books on X programming would be very helpful with all the widget sets, it seems impossible to get any decent "beginner" level books (beginner in the widest sense - I know Windows programming, so GUI stuff is not new to me, but the differences are!). Paul.
RE: Exim as default mailer
Thanks to all who offered help and comments. I now have exim as my MTA, and I'm very pleased with the results. The filtering facilities are nice - replaced what I previously used procmail for. I don't think there's anything I needed to do but couldn't with smail/procmail, which I can now do with exim (my needs are pretty limited) but the documentation available for exim is superb, which makes a heck of a difference... Thanks again, Paul.
Telling apt where my files are...
Hi, I've just started using apt for installing and upgrading packages. I'm very impressed with it (at least as far as getting stuff off the web is concerned). What I'd like to do is use it for *all* my setup. But I can't get it to see the stuff on my (CheapBytes) Hamm CDs - which I'd like to use for installations when I don't want to download off the web. I have three CDs, each with part of the distribution on (assume the CD is mounted on /cdrom): CD1: Main, in /cdrom/debian/main CD2: Contrib, in /cdrom/debian/contrib CD3: Non-free, in /cdrom/debian/non-free I'd also like to get the "frozen" stuff from www.uk.debian.org. My sources.list file is at home (forgot the floppy with a copy on it, rats!) but I have a line something like deb ftp://www.uk.debian.org/debian frozen main contrib non-free non-us which works fine. But I can't get the CDs added - it seems to dislike the directory structure (lack of a ../dists/.. level?) Can anybody tell me what I need? Also, with this sort of configuration, can I set it up so that I can "apt-get update" against all the sources somehow, then just do an apt-get install, which will load off the CD (I'd like it to prompt for a CD, and preferably mount it for itself if not already mounted, but I gather that needs "multi-cd" support, which isn't ready yet...) if my PPP connection isn't up (ie, the ftp connection isn't available), or get the newer version off the web if it is up. Also, will apt use the CD if the web version *isn't* any newer? Thanks for any pointers... Paul. PS I'd like to do apt-get upgrade (or even dist-upgrade) but I don't want to do the multi-megabyte download all at once. Can I get apt to tell me what would be done by a [dist-]upgrade and then let me do it in stages? PPS Someone here has been talking about getting apt to let me retain downloaded sources after the upgrade. I agree, that would be useful for me, too. Maybe allow a "cache" directory to be specified, where downloaded sources get cached (a zip drive, say). And add an extra "type" for sources.list which lets apt point at that cache as another (incomplete) distribution source... I'm rambling now...
Exim as default mailer
Hi, There seems to be a lot of talk about moving to Exim as the default mailer for Debian. While I feel like I should support Exim (as I used to go to Cambridge, whwre it was developed :-) I was concerned when I initially looked at it by the statement in the (version 2) manual (section 39, "Intermittently connected hosts") where it says "Exim was designed for use on permanently connected hosts, and so it is not particularly well-suited for use in an intermittently connected environment." As I have a single-user home PC with dial-up access to my ISP, I fall squarely into the category of users for whom Exim is "not particularly well-suited". Is this a real problem, or is Exim a good mailer for dial-up systems? In theory, I like Exim, not least for the clear and comprehensive manual, but I really don't want to spend a long time fighting the system's assumptions just to get it set up... Thanks, Paul Moore.
Mail to News
Hi, Does anybody know of any mail-to-news software I could use to gateway the traffic from the mailing lists I subscribe to into a news server? I'd need to be able to transfer incoming list mail into a local newsgroup and direct local postings to the group out to the mailing list. Is there anything I could use to do this? Thanks in advance, Paul Moore.
Installing Perl 5.005_??
Hi, I have Debian 2.0, which has Perl 5.004_04 installed. I'd like to upgrade to 5.005_02 (or maybe even _53). There isn't a debian 5.005 package yet (AFAIK) - how do I install 5.005, without confusing dpkg? I don't know what there is on my system which depends on Perl - can I find out? I'm quite willing to build/install all the Perl libraries I use myself, but I want 5.005 to be the default Perl (so I don't have to keep remembering to use it explicitly). What breaks if I either install 5.005 in /usr, or alternatively if I install in /usr/local, but make /usr/bin/perl a symlink to /usr/local/bin/perl? I could hit this a lot - I do a fair bit of testing/development of software, for which I don't expect Debian to be as current as I want/need to be. (Vim 5.4a is another recent example which comes to mind). Thanks, Paul. PS This is a personal-use machine, so I don't need to worry about annoying other users - just about breaking "the system"...
RE: SOLVED (sort of): Re: /etc/ppp/ip-up not being run
>From: Noah L. Meyerhans[SMTP:[EMAIL PROTECTED] > >Doh! I feel pretty dumb now. The scripts ARE being run. The problem >does, however, seem to have something to do with the way in which they're >being run. One of the scripts contains an ftp command, which should >(using a macro in /root/.netrc) connect to a remote site and upload a >file. I know the ftp macro works fine, because if I run my script by >hand, things work just fine, and the file is uploaded automatically. >However, when pppd tries to run the script, I find this in the output of >the ftp process (which I redirect to a file): > >Name (129.10.116.200:root): Login incorrect. > >But I'm NOT trying to log in as root! The ftp macro specifies my username >and password at the remote host. As I said, all this works fine when I >run the script by hand! It's possible (quite likely?) that you're getting problems because of the environment in which the script is being run. What user does it get run as (does that user have access to /root)? What environment variables are set ($HOME - if it matters)? I've never had problems like this, but I've heard many reports of such things, with cron, CGI, and similar "automated" script execution... HTH, Paul Moore.
Building my own kernel
Hi, I'm currently running a 2.0 system (pretty much vanilla), but I'm thinking of trying out a kernel build of one of the new 2.1 kernels (probably the latest, unless anyone is aware of problems which may mean an earlier one is more stable). I've looked at all the stuff in the kernel-package docs and so on, and it seems OK, but I'm not 100% sure I feel confident of *not* breaking my system... Can anybody offer me some help/advice on what I need to do (short of "backup everything", preferably :-) to make sure that, should disaster strike, I can get back to where I am now. The process of installing the required software upgrades (2.1 kernels require certain levels of software, I understand), downloading and unpacking the kernel source, configuring, make-kpkg stuff I'm happy enough with. But once I have the relevant kernel .deb, if I install it, will that trash my current kernel? If so, what/how do I back it up? Just to make things more confusing, I have a dual-boot NT machine using the NT loader. For that, my normal process of setting up the boot process is to use lilo with a lilo.conf which bungs a MBR onto a floppy, which I then peel off into the NT C:\ drive, where the loader picks it up. Will installing a new kernel .deb make the existing MBR useless? (AFAIK, /vmlinuz seems to be a symlink to the actual kernel - does LILO pick up the kernel as it is pointed to, or will the change to the symlink cause problems?) As you can probably see, I've got an idea of what I'm doing, but I tend to trust the LILO stuff to "just work" - I don't really understand what's going on with it, and that's what worries me... Thanks, Paul. PS The reason I'm after kernel 2.1 is to see how much P&P soundcard support there is (ie, will it make my Ensoniq Soundscape P&P card actually work...) and to look into parallel port sharing (I have a Zip drive and a printer daisy chained off the same port).
RE: /etc/ppp/ip-up not being run
>From: Noah L. Meyerhans[SMTP:[EMAIL PROTECTED] >Sent: 03 November 1998 16:55 >To:Debian User List >Subject: /etc/ppp/ip-up not being run > >-BEGIN PGP SIGNED MESSAGE- > >The ip-up script is not being run when I start a ppp connection using pon. >I can run the script manually, and it works OK. I've read the pppd man >page, but didn't find anything in it that mentioned anything about >requiring this script to have certain permissions or ownership (aside from >the obvious execute permissions). It doesn't matter if I run pon as root >or a user; the script isn't run in either case. > >There are no error messages. > >Any ideas? One problem I had was that the default ip-up runs the scripts in the directory ip-up.d. *These* scripts need to be executable if they are to run - non-executable scripts are silently ignored. OTOH, that would mean they wouldn't run interactively, either, so this may not be your problem... Paul.
RE: Kernel rebuild
>From: Waif W. Urchin[SMTP:[EMAIL PROTECTED] > >Grr... > >Alright, still trying to recompile my kernel after doing an initial >install from floppy. While I'm not particularly able to help with your problem, I'd guess that you don't actually need to recompile the kernel. So maybe you can avoid the stress... I can't think of a reason why you'd need a new kernel to fix a PPP dialup problem. It's certainly not impossible, but I'd try other avenues first. Maybe post the PPP log to the list, as someone else suggested... Paul
RE: No Next
>From: Imran Geriskovan[SMTP:[EMAIL PROTECTED] >Hi, > >I've installed Debian several times on several machines but got stuck with a >problem on one machine. > >Problem is as follows: > >I follow all the steps that dinstall suggest me (i.e. Initialize a Linux >Partion, Install device drivers, Conf Network, install etc..). >When it comes to "Install Base System" dinstall ask me the path of >"base2_0.tgz", I provide it, then it works a while (seems like installation >in >progress). But after then it drops back to dinstall menu with >"Next: Install Base system" highlighted. No other "Next" s available. > >I manually select "Configure Base System". But at this time it says that >"You have to install Base System to configure." :( > >Again I choose "Install Base System" but same thing happens. > >At the end I give up and reboot the machine with a boot floopy for the same >partition. It boots and mounts /. However it ends up with blank screen with >a prompt "#"!. Nothing else. > >Issuing "ls" displays the directories and files. (bin, usr, home, dev..) >Seems like files are installed. But no configuration is done. (Not a >surprise. >Huh..) > >Any ideas? I had almost this exact problem on a machine with no CD. I had downloaded the minimal set of files from the Debian FTP site onto the machine's hard disk in a DOS (Windows 95) partition, and installed from that. I got exactly the same symptoms as you. I couldn't fix this, no matter what I tried. However, when I got hold of a CD shared across the network and transferred the same files from the CD (FWIW, it was the CheapBytes Debian 2.0 Binary CD), the install worked fine. I have NO IDEA what the difference was... Whether this is relevant to you, I'm not sure... Paul.
RE: Procmail and smail
>From: Pann McCuaig[SMTP:[EMAIL PROTECTED] > >This document helped me do exactly what you want. Nice step-by-step >instructions that include tests along the way. > >Email: > Send mail to: >[EMAIL PROTECTED] > > Containing the following in the body of the message: >send usenet/news.answers/mail/filtering-faq While this document didn't cover setting up procmail as the default transport, I followed some pointers from it and found details on the web (which say they come from the examples directory in procmail, although I couldn't find them in /usr/doc/procmail) on how to do just this. Specifically, 2c.Special directions for sites with smail --- For smail 2.x users there are two options: i. Move the current local-mail-delivery program (probably /bin/lmail) out of the way, make a symbolic or hard link from procmail to the name of that program (e.g. "ln /usr/local/bin/procmail /bin/lmail") ii. Make sure the following macro is defined in src/defs.h: #define LMAIL(frm,sys) "/usr/local/bin/procmail -d" For smail 3.x users there are also two options: i. The same solution as for smail 2.x (however, method ii is preferred) ii. Replace any existing "local"-entry in the /usr/lib/smail/transports file (create one, if need be) with the following two lines: local: return_path, local, from, driver=pipe; user=root, cmd="/usr/local/bin/procmail -d $($user$)" I used method (ii) for 3.x (the file is /etc/smail/transports in debian). Be careful, you need to use commas and semicolons just as shown (semicolon after driver=pipe, commas elsewhere) which I missed first time... Now working just as I want. Hope this helps others with the same requirements, Paul.
Procmail and smail
Hi, I have smail and procmail installed, and I want to get them working together. I am currently using a .forward file to push my mail through procmail, but what I'd like is if smail used procmail directly as the local delivery agent (when I used the SuSE distribution, with sendmail and procmail, it did this). This would mean that any user with a .procmailrc would automatically get mail processed via procmail, without needing the extra .forward file. I've looked in the smail and procmail documentation, but I can't see what I need to add to set this up. I guess I need some form of transports file, but I can't find an example I can work from. Can anybody help? Thanks, Paul
RE: Waiting for scripts in /etc/ppp/ip-up.d
To all who replied so helpfully, thanks. I now have a working setup. Basically, the simplest solution for me is to add the "idle" option to my ppp options, to hangup the line when there is no more activity. A number of people suggested diald - it certainly looks good, and I may well look into it again at a later date, but for now it seems like overkill for my (very simple) needs. Thanks once again for all the help. I'm really impressed with Debian (both the system and the community) - everything seems so well thought out and effective. Paul.
RE: Waiting for scripts in /etc/ppp/ip-up.d
>From: Mitch Blevins[SMTP:[EMAIL PROTECTED] >Moore, Paul wrote: >> [snip] >> The second problem is the big one - Demon dropped the line on me (for no >> reason I could see :-() while my slrnpull session was under way. Diald >> proceeded to restart the link, as expected. Great. BUT, when the link >> came up again, diald started a SECOND copy of slrnpull (as per the >> ip-up.d scripts). This is definitely *not* right :-( > > This script should be changed to check for an existing running > copy first.. Probably true. It's what came as default, and it worked fine, so I never saw a reason to change it... >> I've come to the conclusion that the best way of doing what I'm after is >> to dial up manually, using pon. This is no problem. Diald doesn't really >> offer me anything extra here, as I basically don't want to dial up on >> demand... BUT, once I am online, I'd like to be able to get at the bit >> of diald's functionality which monitors the link, and hangs up >> (preferably using poff) when the link goes idle for a specified length >> of time. > > Am I missing something? pppd has an 'idle' option that does just that. No, I'm missing something. I never saw this option. This is exactly what I want - I'll try it out tonight. Thanks for pointing this out to me! Paul.
RE: Waiting for scripts in /etc/ppp/ip-up.d
>From: Michael Beattie[SMTP:[EMAIL PROTECTED] >On Wed, 21 Oct 1998, mwb wrote: >> Although I haven't tried it, but you might try putting poff (or your >> query script) as the last item in ip-up. > >Exactly like what I would have suggested. Make a script such as 'zpoff' - >so the leading 'z' puts the file at the end of the directory listing. >'run-parts' runs the scripts alphabetically AFAIK. Does this imply that the ip-up.d scripts are run in sequence? That's a bit annoying - I was hoping to add a few mirrorring scripts in there which would run alongside my news download - would they wait for slrnpull to finish? Paul.
Re: Waiting for scripts in /etc/ppp/ip-up.d
I've had a number of suggestions now, thanks to all who replied. However, I've still got some problems. First - for all those who suggested fetchmail, thanks but I am aware of fetchmail, and it just does not do what I need. The reasons are complex but conclusive... [If anybody really wants to discuss why I don't want to use fetchmail, contact me off the list and I'll explain more] The nice suggestion was diald, which I tried to set up last night. I had some problems with it, though. I set it up to run the existing /usr/bin/pon and /usr/bin/poff scripts to start up and shut down the PPP link (I did it this way because I have this configuration set up nicely, thanks to pppconfig, and I didn't want to start from scratch - if there's an equivalent of pppconfig for setting up diald, then I'm interested, but otherwise not. I see no reason to reinvent the wheel here...) First problem was that I can't do the necessary "kill -HUP " to force the link up unless I am root. I don't really want a setuid root script to do this... Also, getting the PID of diald is (slightly) tricky in a script. Never mind, I can fix this up when everything else is working... The second problem is the big one - Demon dropped the line on me (for no reason I could see :-() while my slrnpull session was under way. Diald proceeded to restart the link, as expected. Great. BUT, when the link came up again, diald started a SECOND copy of slrnpull (as per the ip-up.d scripts). This is definitely *not* right :-( Worst of all, because the link was started by diald, I couldn't stop it using poff - the process isn't owned by me. I pulled the plug on the modem as a quick fix... I've come to the conclusion that the best way of doing what I'm after is to dial up manually, using pon. This is no problem. Diald doesn't really offer me anything extra here, as I basically don't want to dial up on demand... BUT, once I am online, I'd like to be able to get at the bit of diald's functionality which monitors the link, and hangs up (preferably using poff) when the link goes idle for a specified length of time. Can diald be persuaded to do this (hangup-only, no dialup function)? If not, how do I get at the ppp link stats to monitor packets going through? If I can monitor packets/sec across the ppp link, or maybe traffic over the modem, I could write my own monitor-and-timeout program. Thanks for the help so far, Paul.
RE: Waiting for scripts in /etc/ppp/ip-up.d
>From: Mitchell, James T1[SMTP:[EMAIL PROTECTED] > >I use demon, who also use SMTP to send email while online. My 'solution' is >based on the fact that I use diald to manage connections. Actually, I use demon, too :-) [Are there any good web pahes about how to set up Linux with demon?] >diald will bringup or drop a connection based on whether there is any >traffic that is being sent/received. So all I do is send a SIGHUP signal to >diald (which is the "please go online" signal) and if there is no email >waiting, diald will drop the line in about 30 seconds. If there is email, >demon sends it to me, causing traffic, causing diald to keep the link up >until there is no traffic left. I didn't know about SIGHUP - that makes diald definitely more attractive. But I still want to do a big "batch download" while I'm having tea - if I add diald to my setup, and then change from doing pon to "kill -HUP diald", will all the ip-up.d scripts run (I assume that once they start, they will generate traffic so that the line stays up until they finish). This looks like a good answer - thanks. Paul. PS Is there a way to make slrn *default* to --spool mode? I'm forever typing just "slrn", which fails because news.demon.co.uk isn't available. With diald, I'd be dialing up every time ... :-) I guess an alias does it...
RE: Waiting for scripts in /etc/ppp/ip-up.d
>From: Peter Iannarelli[SMTP:[EMAIL PROTECTED] >Hello: > >Wouldn't it just be cleaner to use dial on demand (diald) >which would automatically bring up and/or turn down the >link based on idle time. Sorry, I should have said this. Diald won't work for two main reasons. First, my ISP passes mail to me when I'm online (not via POP3 - I can use POP3, but it's not suitable for a number of complex reasons...). So I have to go online anyway to grab mail. Second, I want to work in a batch-online mode, where I get everything in one big slug (which I can run, for example, while I have my tea :-) and then scan it all offline. Hope this explains better, Paul.
Waiting for scripts in /etc/ppp/ip-up.d
Background: I have my PPP connection to my ISP set up, pretty much out of the box. In /etc/ppp/ip-up.d, I have some scripts, one to send queued mail, one to get a newsfeed via slrnpull. I will probably add extra scripts later (mirror my homepages, download new packages, cache web pages, whatever...) My ISP also sends my Email via SMTP while I'm online. I can tell when it's finished by fingering the post machine, and checking if it has any messages left for me. What I'd like to do is to write a script which starts the PPP connection, then waits for all the ip-up.d scripts to run, and for my mail to finish arriving, and then drops the connection with poff (actually, I'd like it to ask me whether it was OK to go offline, and allow me to override, but that's not too hard). So, my question - can I do this? My guess is that I somehow need to loop checking on the status of the various processes (probably by parsing ps aux output) until everything has completed. But that may be messy, and it almost certainly needs significant admin when I add new ip-up scripts. I thought about using pstree or something, to look for children of the pon script - does that work? Any pointers? Thanks, Paul Moore. PS Is it possible for a script to tell if it's running under X? Ideally, I'd like to query the user via a dialog box if in X, but use a normal stdin read otherwise...
smail problems (dial-up)
Hi, I'm just getting smail set up, and I'm not sure what I need to do for the last stage (getting my dial-up setup working). I have local mail setup fine, including redirecting "postmaster" mail to my normal account. My ISP (Demon Internet in the UK) passes mail to me using SMTP when I am logged on, so receiving mail seems to be no problem (I just do pon and wait, Demon passes the mail to me fairly promptly). For outgoing mail, I have Demon's mailer set up as my smarthost. If I send a message while offline, the smail log includes a line saying there is an error, can't find post.demon.co.uk, deferring mail (OK, that's right although as it is the expected situation, I'd rather it wasn't an error...). However, when I start up the PPP connection with pon, nothing seems to trigger smail to send deferred mail. I found a command "runq" which seems to push stuff through. Is that what it's for? If so, how do I get this to happen automatically when I start the PPP link? Did I miss something in the configuration setup to get this added automatically? All of the rest of the setup was put in as I expected (I told the config that I was a dial-up Internet site, from what I recall...) As usual, if I do have to do manual tweaks, what do I need to do if I later want to use dpkg/apt to upgrade or change things (for example, god forbid, I need to switch to sendmail...) Thanks for the help, Paul.
RE: FVWM - Not all modules included?
>From: > [EMAIL PROTECTED]:[EMAIL PROTECTED] >>> "MP" == Moore, Paul <[EMAIL PROTECTED]> writes: > >MP> Have I missed something, or is FvwmConsole not included in the debian >MP> package? > >I don't know this module. What does it do? How does it look like? >Perhaps it is not a official part of fvwm2 and Suse just included it. > >Maybe you are just looking for the FvwmTalk module? It provides a window into which you can type Fvwm configuration commands. I use it for testing out the effect of commands (Style "Xterm" Sticky) without needing to edit the config file and restart every time. I saw something which gave me the impression it's in some sort of "extras" directory in the fvwm source, but I only have the binary so I can't check. [Hmm, I just recalled that Debian 2.0 sources CD I had somewhere, check on that - yes, it's in Fvwm2/extras/FvwmConsole/*. So I guess the real question is, if I install the sources and compile this module myself, where should I put it? I don't want to interfere with dpkg's handling of the fvwm2 package, but I do want to keep my disk structure sensible... Is there (or should there be) a fvwm2_extras package? I couldn't find one...] BTW, I just noticed that FvwmWharf is *also* in extras. I was also planning on using that, for its "Folders" (expanding buttons). Looks like I need to do some building of my own. Maybe I also need to consider registering as a developer, to package these up. In at the deep end :-( Paul.
FVWM - Not all modules included?
I've just got FVWM2 set up and running on my Debian 2.0 system. I wanted to play with some options, and I looked for the FvwmConsole module which comes with Fvwm2 (at least it did on my previous SuSE setup). But it doesn't seem to be included... Have I missed something, or is FvwmConsole not included in the debian package? Assuming it isn't, how do I add it? Thanks, Paul Moore
RE: Format of .deb files
>From: J.H.M. Dassen (Ray)[SMTP:[EMAIL PROTECTED] >> Presumably, .deb files are internally some form of tar/cpio archive - can >> I get at the tar so that I can unpack it (I have Windows gzip, tar, etc >> utilities). > >deb(5) documents the format. Basically, it's an ar(1) archive containing >tar.gz archives. Great. I'll pick up man 5 deb tonight and I can start hacking... >> If not, is there any other way I can browse the documentation (/usr/doc, >> /usr/man, /usr/info) for debian while not actually at a PC with debian >> running? > >Perhaps setting up a dwww server is an option. (see >http://dwww.jimpick.com/) > >> Is it available on the web? > >See http://dwww.jimpick.com/dwww/ for a demo. Actually, this gives me 90% of what I need directly. Brilliant! Thanks for the help. I think that's the record for the fastest ever help I've ever had with a problem - your reply was back in a few minutes after I posted!!! Paul.
Format of .deb files
I am trying to get to grips with Debian, and I'd like to browse through the documentation. The problem I have is that my Debian PC is at one home, whereas the PC where I have the most chance to read documentation, print things out, follow up references, etc, is at work. My current work PC is Windows only [:-(] but it does have a CD drive and I have the Debian CDs. My question is, basically, can I extract files from .deb format packages on a Windows box? Presumably, .deb files are internally some form of tar/cpio archive - can I get at the tar so that I can unpack it (I have Windows gzip, tar, etc utilities). If not, is there any other way I can browse the documentation (/usr/doc, /usr/man, /usr/info) for debian while not actually at a PC with debian running? Is it available on the web? Thanks in advance for any help... Paul Moore.
RE: Getting at Debian distribution files
Hmm, Now that I try to FTP through the firewall using the command-line FTP client (I usually use a preconfigured GUI client) it looks like my firewall works just like yours, and I now see how to set up the dselect access method. I'll go and give it a try... Yes, looks like that worked perfectly!! Thanks a lot! Paul. >-- >From: [EMAIL PROTECTED]:[EMAIL PROTECTED] >Sent: 13 October 1998 17:46 >To:Moore, Paul >Cc:[EMAIL PROTECTED] >Subject: Re: Getting at Debian distribution files > >"Moore, Paul" <[EMAIL PROTECTED]> writes: >| I'm trying to install Debian on a PC at work. I have the Debian 2.0 >| Official CD from CheapBytes, which installed fine on my home PC (which >| has a CD drive). >[snip] >| I have a network connection, so I could get the stuff from the debian >| ftp site directly, but I'm behind a firewall. Dselect does offer the >| chance to put in a proxy name/IP address, but it asks for a >| username/password, which my firewall doesn't need... >[snip] > >So how does your firewall work? I'm also behind a firewall and I >regularly use dselect, with the ftp method, to upgrade/install >packages from the debian ftp site. Say you were manually doing an >anonymous ftp to the debian mirror of your choice, how would that >work? > >For example, behind our firewall I'd do: > >ftp firewall >Name: [EMAIL PROTECTED] >331-(GATEWAY CONNECTED TO ftp.debian.org) >331-(220 santanni.cc.gatech.edu FTP server (Version >wu-2.4.2-academ[BETA-15](4) Thu Feb 12 17:00:23 EST 1998) ready.) >331 Guest login ok, send your complete e-mail address as password. >Password: >[message from debian anonymous ftp server goes here] >ftp> > >and I'm in. You can easily get dselect to work with such a firewall >scheme. > >Gary >
Getting at Debian distribution files
Hi, I'm trying to install Debian on a PC at work. I have the Debian 2.0 Official CD from CheapBytes, which installed fine on my home PC (which has a CD drive). The work PC has no CD drive, but it *does* have a network connection. I have got the install through to the point of running dselect as root to load up the various packages. Now, however, I have a problem. As the target PC has no CD, I can't load off the CD direct. I could (in theory) load the packages I need to the target PC hard disk (I have a DOS/VFAT partition which would probably do), but the partition isn't that big (so I'd have to do it in stages) and the CD is in Rock Ridge (? is that the name) format, with TRANS.TBL files in each directory. On Windows, the names are mangled, and I'd have to manually rename each file as necessary - a completely infeasible job... I have a network connection, so I could get the stuff from the debian ftp site directly, but I'm behind a firewall. Dselect does offer the chance to put in a proxy name/IP address, but it asks for a username/password, which my firewall doesn't need... An alternative would be to mount the CD on a Windows 95 or NT box on the LAN, and pull the packages off the CD via FTP or NFS, but this would require a FTP/NFS server for 95 or NT which will handle translation of filenames based on TRANS.TBL files - something which I haven't been able to find. Can anybody suggest a way forward? Or an alternative option I haven't thought of? Thanks, Paul Moore. PS Once I'm up and running, I need my target PC to get an IP address via DHCP (I've got a static one temporarily). Is there a package which supports DHCP, and if so what do I need to do to set it up?
Installing Debian 2.0 from hard disk (machine has no CD)
Hi, I'm having trouble getting Debian 2.0 to install from a hard disk partition (the PC in question doesn't have a CD drive :-() I downloaded all the stuff I needed (according to install.html) from ftp.debian.org, and put it on my HD (DOS drive D:, subdirectory \debian) Then I started the install process. Great, no problems (I allocated pre-existing Linux swap and root partitions to the Debian setup) - until I came to "Install the base system". I said "from hard disk", told it where, got the loading message. Then the install program came back, said it what checking what I had to do next, then told me the next step was to install the base system (again!) I can't get past this step, no matter how often I redo it. "Configure the base system" won't work - it says I need to install it first. I checked the base2_0.tgz file - it untars fine on DOS, and it's exactly the same size as the one on the FTP site (no accidental ASCII downloads...) Can anyone help? Paul Moore.