Re: How to use old CPUs (Not Debian Specific)
On 14 Oct 05 18:03:52 GMT, Marc Shapiro [EMAIL PROTECTED] wrote: Like I said, I don't have enough space in the new apartment to set up multiple computers, but I dislike having computing power going to waste. Can anyone suggest a way to network/connect all four to possibly distribute the load among them? I am considering building a custom case to hold all the MBs and only have a single monitor, mouse and keyboard connected. The case would, obviously, have to be fairly large, but it could then act as a table, as well, so the space would still be more efficient than four seperate cases which serve no othere purpose. The processing power of your two 486s is so pitiful compared to the Athlon box its hardly worth the effort. I think you should take this opportunity to finally retire these venerable machines. Of course make sure you dispose of them responsibly through a reputable recycler who won't just export them as toxic waste to some third-world country. If you can find one... The PII still has potential, my own server/gateway box is in the same class. Another possibility is to install LTP on the Athlon and set up the PII as a thin client to reduce competition with other members of your household for the seat in front of the Athlon. -- Frank Copeland Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://xenu.apana.org.au/ntshp/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Plain text mail-merge
I'm looking for a tool that can do the equivalent of the mail-merge feature found in any decent word processor, but with plain text files. I'm sure there has to be such a beast but I've apt-search'd and freshmeat'd with various combinations of mail merge text substitution to no avail. I'm sure an hour or so with perl or python will produce the goods for my specific case but I find it hard to believe there's no general purpose tool already available. -- Frank Copeland Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://xenu.apana.org.au/ntshp/ Keep it in Usenet. E-mail replies and 'courtesy' copies are not welcome. If you're selling, I ain't buying. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: upgrading to unstable
On 19 May 05 15:04:10 GMT, Alberto Bert [EMAIL PROTECTED] wrote: after reading the FAQ at: http://www.people.cornell.edu/pages/kk288/debian_choosing_distribution.html I'm upgrading from sarge to unstable... After reading this I'm not convinced to do the same. The main argument for preferring unstable over testing appears to be that when testing breaks, it can take longer to fix than unstable. While this is true, it's not a showstopper because you can always install the fixed packages from unstable and wait for testing to catch up. Testing is almost-but-not-quite bleeding-edge and that's good enough for me. I'm pretty scared, so it would be nice to have some help from the list... If you are scared, then don't do it... In doing it I'm using aptitude. I just saied it to U all the Upgrade packages, then I'm upgrading (it's downloading...) ... but it's a bit late now ;-) Is there anything else I should do? Give yourself plenty of time to sort out any issues that crop up unexpectedly. I doubt you will have many, the difference between testing and unstable really isn't that great. The real challenge is upgrading stable to testing/unstable. I've done this a few times now and I always do it in stages, especially on servers. First upgrade the tools used to perform the upgrade, ie - apt-* and dpkg. Then upgrade each major service (web server, mail-transport-agent, etc) individually. Don't forget the kernel; going from 2.4.* to 2.6.* has some major implications especially for a desktop. Once everything important is working properly, do a final dist-upgrade to finish the job. *Always* pay special attention to the feedback from the package management system. It's a good idea to run (the equivalent of) 'apt-get -us dist-upgrade' before each major step just to see what it proposes to do next. -- Frank Copeland Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://xenu.apana.org.au/ntshp/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Question about hard disk partition strategy for debian
On 9 May 05 09:30:14 GMT, Lian Liming [EMAIL PROTECTED] wrote: Hi all, I have 35 GB hard disk space for installing debian/ustable.I am a common linux user and would like to do some web programming that means I need X winodows with KDE and LAMP(linux, apache, mysql, php). To limit damage upon system crash and better use the hard disk space, I want to know something for the partition strategy. Which directories should I separate from the /, should I separate /usr, /var, /tmp, /home ? And then, how many GBs should i give to these partitions? It is really welcome that someone can share their partition strategy with brief explanation why they did so. I use LVM (Logical Volume Manager) to manage disk space and partitions. This allows me to adjust my space allocations as circumstances change. There's nothing worse than having to repartition and reinstall an otherwise working system just because you miscalculated the amount of space required for a partition. Resizing a partition is a matter of minutes, in the worst case requiring you to reboot into single-user mode. The setup I use for new systems has two partitions: a standard (ext3 in my case) partition for root, 250MB is plenty, and an LVM partition for the rest of the disk. It is possible to install root on an LVM partition but I choose to take a conservative approach. I install a minimal system (base plus LVM plus rsync) on the root partition, create the LVM pool then start carving partitions out of it. At a minimum I create partitions for swap, /usr, /var and /tmp. For a desktop I add a seperate /home. I don't have any firm suggestions for how large each partition should be because the whole point of the exercise is to make that unnecessary. If you underestimate, just make the partition bigger. Starting with 1GB each for /home, /usr and /var is probably a good idea. 125MB - 250MB for /tmp is usually ample. Moving data from the root partition to an LVM partition is fairly easy with rsync. LVM allows you to seamlessly increase the available storage by installing a new disk and adding it to the pool. You can retire a disk by removing it from the pool before physically removing it from the machine. I've been through this process a few times now. I have a server box and a desktop box. As I upgrade the desktop box I recycle the old hardware into the server. Over the last few years it has had 2GB and 4GB disks replaced by a 20GB and a 40GB, all formerly used in the desktop. The server filesystem has been migrated from disk to disk without disruption other than a little downtime. My server currently looks more-or-less like this: [EMAIL PROTECTED]:~# df -h FilesystemSize Used Avail Use% Mounted on /dev/hda1 236M 124M 100M 56% / /dev/mapper/vg0-home 3.0G 1.9G 1.2G 62% /home /dev/mapper/vg0-usr 1008M 923M 86M 92% /usr /dev/mapper/vg0-tmp 124M 4.1M 120M 4% /tmp /dev/mapper/vg0-var 2.0G 1.2G 855M 57% /var /dev/mapper/vg0-spool 5.9G 3.4G 2.6G 57% /var/spool /dev/mapper/vg0-mirror 5.0G 1.7G 3.2G 35% /usr/local/mirror Everything except /dev/hda1 is an LVM partition. There is still over 38GB of disk space waiting to be allocated. -- Frank Copeland Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://xenu.apana.org.au/ntshp/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Is 64MB enough?
On 9 May 05 11:18:12 GMT, John Moore [EMAIL PROTECTED] wrote: I'm a total Linux newbie who'd like to install Linux on my really old PC. It has 64MB of RAM and a 4GB hard disk. The installation documentation I've read at debian.org seems to indicate that this is sufficient. Is that right? If you want to set it up as a desktop system you will be pushing your luck. It can be done but you will learn to live with swapping. As a home server (firewall/router/gateway/fileserver, no GUI) it will be just fine. It would also be more than adequate as a thin client. -- Frank Copeland Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://xenu.apana.org.au/ntshp/ Keep it in Usenet. E-mail replies and 'courtesy' copies are not welcome. If you're selling, I ain't buying. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Holy Spam!
Jon Earle wrote: I sent _one_ post to the debian-users list yesterday. One. I neglected to use an alias I'd created for posting to that list, and, due to their open posting policy and their email-usenet gateway and the availability of addresses in the clear within the list archives, within _minutes_, I started receiving viruses, spam and other crap. My mail logs since yesterday show a _ton_ of crap coming at me now! Unbelievable! I've been posting to this list with addresses that have been harvested since the dawn of spam (one dates from 1994, the other from 1996). I willfully insist on receiving mail to both addresses over two modem connections (one of them a mere 33.6K) and uucp is also rumoured to be involved. I've survived the spam and viruses so far. The open nature of the Debian mailing lists is a feature, not a bug, and is well worth preserving. -- Frank Copeland Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://xenu.apana.org.au/ntshp/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Acorn A5000
On 29 Mar 03 07:02:34 GMT, Paul Grenyer [EMAIL PROTECTED] wrote: Different architectures make more or less efficient use of RAM, but I suspect you are going to find it extremely difficult to even get debian installed on such a system. My 5MB Amiga 2000 simply thrashed itself to death trying to install packages, thanks to the bloated package database and dpkg's RAM-hungry design. I never got it upgraded from potato to woody and if it hadn't been taken off the net by a faulty NIC I would have shut it down anyway when security updates for potato stop. It's too long ago for me to remember the details of the Amiga's, but if memory serves me correctly the Acorn machines were far superior in performance. :-) Doesn't matter what the CPU can do when you have less RAM than dpkg's resident set and swap speed becomes the dominant factor. Hard disks are much, much slower than a MC68030. I doubt a 25Mhz ARM3 was enough of an improvement to get over that hurdle. -- Frank Copeland Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://xenu.apana.org.au/ntshp/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Acorn A5000
On 28 Mar 03 13:49:16 GMT, Paul Grenyer [EMAIL PROTECTED] wrote: However, RiscOS code is extremely compact, and whether Linux can do anything useful in 4MB of RAM I'm not sure. I only want to run apache and CVS. Different architectures make more or less efficient use of RAM, but I suspect you are going to find it extremely difficult to even get debian installed on such a system. My 5MB Amiga 2000 simply thrashed itself to death trying to install packages, thanks to the bloated package database and dpkg's RAM-hungry design. I never got it upgraded from potato to woody and if it hadn't been taken off the net by a faulty NIC I would have shut it down anyway when security updates for potato stop. -- Frank Copeland Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://xenu.apana.org.au/ntshp/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Libranet to Sarge
On 6 Mar 03 22:35:56 GMT, Keith Winston [EMAIL PROTECTED] wrote: By the way, is there some reason Debian can't adopt the Libranet installer and tweak it for the Debain defaults intead of the Libranet defaults? The installer is very nice. Debian supports 11(?) architectures. How many does the Libranet installer support? -- Frank Copeland Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://xenu.apana.org.au/ntshp/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: hosts.(allow|deny)
On 7 Mar 03 17:43:34 GMT, nate [EMAIL PROTECTED] wrote: Hugh Saunders said: But i thought it would be more secure to put ALL : ALL in hosts.deny and then in.sshd : ALL in hosts.allow. This dosnt work[ssh connections are refused], how do i specify that i want all hosts to be able to connect to port 22? hosts.allow/deny can be tricky(one reason I don't use it), your situation should be fixed by changing in.sshd to sshd. Check /var/log/daemon.log for the name of the daemon(s). You should see reject messages for the sshd service. another reason I don't use it is I prefer firewalls over it. IMHO hosts.allow/deny is far less tricky than a firewall, even when using a high-level firewall-builder like shorewall. I use both; defence in depth is a Good Thang(TM). -- Frank Copeland Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://xenu.apana.org.au/ntshp/ Keep it in Usenet. E-mail replies and 'courtesy' copies are not welcome. If you're selling, I ain't buying. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Hardware IDE ATA Raid 1 support in Linux Debian - Linux disapointment
On 14 Feb 03 15:39:11 GMT, Pedro Ruivo TRQV-DSI [EMAIL PROTECTED] wrote: I have to say one other thing, wich is the disappointmento with linux in this concern, shouldn't there be a Linux Driver ? I find drivers for Red Hat, Suse, Turbo, ... Linux but not for Debian and specially not a for Linux Driver. Save your disappointment for the hardware manufacturer. They are the ones making you jump through hoops, not Linux. GNU/Linux is a *free* operating system and it barely tolerates non-free binary-only drivers. Even so it is clearly possible to write such drivers in ways that don't tie them to specific kernel versions and specific distributions; the NVidia video card drivers are an example. If the manufacturer chooses not to properly support its products under GNU/Linux then you are better taking your business to a manufacturer that does. -- Frank Copeland Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://xenu.apana.org.au/ntshp/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: sourceforge - why it depends on exim and proftp?
On 11 Jan 03 08:29:12 GMT, Erik Steffl [EMAIL PROTECTED] wrote: does sourceforge really need to depend on exim and proftp? wouldn't other mail resp. ftp servers be enough? Not to mention apache | apache-ssl (won't any other web server do?) and xenu-preserve-us, elvis | nvi | vim. -- Frank Copeland Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://xenu.apana.org.au/ntshp/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Making music
On 9 Jan 03 10:31:01 GMT, I wrote: I've got sheet music for some 19th century military bugle calls and I'd like to hear what they sound like and maybe create .wavs to put on a web site. I'm looking for some combination of software that will allow me to enter the score for each call and then say 'play this on a bugle'. Any suggestions on how to proceed? I'm basically musically illiterate. Many thanks for the suggestions. I've gone with rosegarden for editing scores and timidity for playback. I'm also looking at denemo. For the record, I have installed rosegarden, timidity and timidity-patches (unstable versions). I run timidity with the -iA option to act as an ALSA sequencer client. When I run the rosegarden sequencer it automagically detects it and uses it for MIDI output. I use the trumpet patch (number 56) from timidity-patches to represent the bugle, since a bugle is supposedly just a trumpet without valves. It took me a little while to work out that rosegarden expects it's own files to have .rose extensions and MIDI files to have .mid extensions; makes the file requesters a bit more useful. The scores I'm working from were printed in 1806 so they are in modern notation with what appear to be a few anacronisms. For example, 4/4 time is often represented by a C after the clef (for common time I presume). Some have no time notation but from the placement of bars appear to be in 2/4, 4/4 and even 9/4[1] time. I'm having a fun time with google researching musical notation. Once I've got a good-sized collection together I'll put the MIDI files and corresponding .oggs on the web if anyone's interested in checking them out. [1] 9/4 or perfect time. See, I'm not so musically illiterate anymore. Frank -- Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://xenu.apana.org.au/ntshp/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Making music
I've got sheet music for some 19th century military bugle calls and I'd like to hear what they sound like and maybe create .wavs to put on a web site. I'm looking for some combination of software that will allow me to enter the score for each call and then say 'play this on a bugle'. Any suggestions on how to proceed? I'm basically musically illiterate. -- Frank Copeland Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://xenu.apana.org.au/ntshp/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Unofficial Debian package management
On 26 Dec 02 15:00:22 GMT, David Z Maze [EMAIL PROTECTED] wrote: I've been using mini-dinstall to maintain a small private repository (mostly with my custom-compiled kernels). You can add other things to it, but it wants a full source package with a .changes file, which means pretending to be a developer and building the package from source yourself. How are you creating custom kernel packages with .changes files? I've been unable to work out how to do it with kernel-package and make-kpkg. -- Frank Copeland Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://xenu.apana.org.au/ntshp/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: which RAID cards work with debian?
On 28 Dec 02 11:54:56 GMT, Haim Ashkenazi [EMAIL PROTECTED] wrote: On Sat, 2002-12-28 at 13:45, Frank Copeland wrote: I manage a couple of HP tc4100s with NetRAID 1Ms and the stock (well, it started as stock anyway...) Debian kernel finds the RAID controller just fine: since I'm always using the 2.4fb kernel I didn't even tried the default kernel (is it 2.2.x?). I'll give it a try. These machines were initially installed with 2.2.x kernels but now run 2.4.x kernels. They work fine with kernel-image-2.4.x-686, but have locally-built kernels to access 1G of RAM. It would help if you could provide more information about your system. What model HP server is it? What model NetRAID? What error messages, if any, are you seeing? -- Frank Copeland Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://xenu.apana.org.au/ntshp/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
ext3 kernel bug, was: Compiling a kernel on an UltraSparc?
On 21 Dec 02 19:15:50 GMT, Nathan E Norman [EMAIL PROTECTED] wrote: BTW, 2.4.20 is not the kernel you want if you run ext3 filesystems. AIUI, the problem with ext3 filesystems applies only if they are in journal mode, which isn't the default. I've also seen suggestions that the bug exists in several versions of the 2.4.x kernels prior to 2.4.20. -- Frank Copeland Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://xenu.apana.org.au/ntshp/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Outlook calendar sharing?
On 20 Dec 02 21:47:52 GMT, Andrew M. A. Cater [EMAIL PROTECTED] wrote: Second question: I'm about to build a Debian server for a charity to act as their mailhub, printserver et. al. One of the users wants an MS Exchange server to support calendar sharing in Outlook - any clues as to how I can do this using only Linux? Requiring an Exchange server to do calendar sharing with Outlook when it wasn't necessary up to Outlook 2000 is the sort of blatant bait-and-switch you would expect from the world's largest convicted monopolist. I've heard that someone is working on an Exchange replacement but couldn't say what it is, whether it is actually useable or whether it will be free software. I find it hard to believe it's just one of the users. Who are they sharing their calendar with? It basically boils down to whether the organisation considers the convenience of these users to be worth the expense of licensing a Windows OS plus Exchange, and paying the tax again year after year. If not, there are viable free alternatives such as Webcalendar http://webcalendar.sourceforge.net/. It requires a web server and a database, but you are building a server for them anyway so that's no big deal. The biggest problem will be selling the idea to them and retraining the affected users. [Please CC me on this one: not currently subscribed to -user] In my opinion asking questions on a public forum you aren't subscribed to is just as rude as CC'ing to posters who are. This response will be in the archives if you can be bothered putting in the effort. -- Frank Copeland Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://xenu.apana.org.au/ntshp/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Outlook calendar sharing?
On Sun, Dec 22, 2002 at 02:39:27AM +, [EMAIL PROTECTED] wrote: I dont think making him search elseware because you cant be bothered to cc him is fair, its NOT obvious Debian has archives, so im ccing him your message below you ass. If he found the list he can find the archives. And *don't* CC me on list mail again, you ass. Frank -- Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://xenu.apana.org.au/ntshp/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: 2.5.50 on Woody
On 16 Dec 02 09:20:37 GMT, arief_mulya [EMAIL PROTECTED] wrote: 1. This newbie install woody. 2. This newbie download linux-2.5.50.tar.gz Bleeding edge. Warning Will Robinson! 3. This newbie untar it and compile it with make-kpkg 4. This newbie get Install succeed but need manual hack It would help if you gave more details about this. 5. This newbie restart and try the new kernel 6. This newbie gets lots of modules dependency error Did you by any chance select the --initrd option to make-kpkg? If you are trying to create an initrd kernel from pristine sources with make-kpkg, you are in for a world of pain, trust me. -- Frank Copeland Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://xenu.apana.org.au/ntshp/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: galeon's screwed up cookie management
On 11 Dec 02 08:40:33 GMT, martin f krafft [EMAIL PROTECTED] wrote: no matter whether i allow galeon to accept cookies from all, this, or no sites, selecting Tools-Cookies-Accept Cookies from this site does *not* do what it should. While the Block cookies from this site command works by entering the site into the Cookies Sites table with Blocked, the allow... command does no such thing; it doesn't enter the current site in any way. This is 1.2.7-5 from unstable. should i file a bug? 1.2.6-2 displays the same behaviour, or something very much like it. My settings are Always accept plus Warn before accepting. I then block 99+% of cookies that are offered. I picked a site that was in my cookies sites list as blocked, went to it and selected Accept cookies from this site. The site was then deleted from the list of cookies sites. The documentation suggests it should have had its status changed from Blocked to Accept, but the actual result is acceptable to me. The next time I visited the site I was asked if I want to accept the cookie as if it were the first visit. I doubt there's any point filing a bug. Upstream considers galeon 1.2.x to be obsolete and probably won't bother to fix such a relatively minor issue. It would be more productive to help make sure the gnome2-based galeon2 behaves better. -- Frank Copeland Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://xenu.apana.org.au/ntshp/ Keep it in Usenet. E-mail replies and 'courtesy' copies are not welcome. If you're selling, I ain't buying. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
make-kpkg and .changes
I would like to install locally built kernel packages in a local repository managed by mini-dinstall. For that I need a .changes file generated. Is there some way to get make-kpkg do that when it creates the packages? A study of the kernel-package documentation has failed to provide me with enlightenment. -- Frank Copeland Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://xenu.apana.org.au/ntshp/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: security updates
On Mon, Oct 28, 2002 at 07:42:57AM -0500, Tom Allison wrote: Frank Copeland wrote: Did I forget to reply to the list? Bad me. Is there some way that I could crontab an apt-get job that would use ONLY the security debian site for upgrading? fjc@thingy:~$ apt-show cron-apt Package: cron-apt Priority: optional Section: admin Installed-Size: 73 Maintainer: Ola Lundqvist [EMAIL PROTECTED] Architecture: all Version: 0.0.6 Depends: apt, bash (= 2.03-6) Filename: pool/main/c/cron-apt/cron-apt_0.0.6_all.deb Size: 7116 MD5sum: 92b962041784a61dfdc1e7d505e5627d Description: Automatic update of packages using apt This package contains a tool that is run by a cron job at regular intervals. By default it just updates the package list and download new packages without installing. You can instruct it to run anything that you can do with apt-get. . It also sends mail (configurable) to the system administrator on errors. . Observe that this tool is a security risk, so you should not set it to do more than necessary (automatic upgrade of all packages is NOT recommended). I saw this. It's what I'm looking for, but I want to install, not download, and only for the security source. I guess you could say it's a security risk to install any security updates automatically If you are tracking stable then the only updates offered are going to be security updates (well, in theory at least). Installing *any* updates automatically is bad, m'kay? Even security updates can result in major changes; remember the recent OpenSSH debacle? Every package update potentially requires human intervention. However, if you really want to do it, cron-apt will let you. Just edit the relevant file in /etc/cron-apt/action.d and delete the '-d'. It's undocumented in the version in stable, but cron-apt can be configured to mail you when an update is downloaded. Just put MAILON=upgrade in /etc/cron-apt/config. Assuming you check your system admin mail daily, you will be notified immediately when a security update is available and you can then complete the update manually. -- Frank Copeland Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://xenu.apana.org.au/ntshp/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Changing Hostname
On 26 Oct 02 13:27:32 GMT, Antonio Rodriguez [EMAIL PROTECTED] wrote: How do you change your host name after your installation is done? (network settings) This works for me: # echo newhostname /etc/hostname # hostname newhostname # grep -ir oldhostname /etc/* Fix up any references to the old hostname that the grep reveals. Reload or restart any services affected. Logout and login again and you should see the new hostname in prompts, etc. Also, what are valid parameters for host name? I made afresh installation of woody in a 386 I got in an auction, and every time gdm starts it complains about the host name. I also noticed that email sent from console using simple mail commands doesn't go anywhere, it just vanishes without trace. This does not happen in another woody box I have. Try sticking to letters, numbers and dashes. Don't use underscores. If you use Gnome you may find it complaining that it can't resolve the hostname. If you don't have a static IP address, try adding the new hostname as an alias for 127.0.0.1 aka localhost in your /etc/hosts file. -- Frank Copeland Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://xenu.apana.org.au/ntshp/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [OT] UPS signaling cable
On 19 May 02 02:53:25 GMT, csj [EMAIL PROTECTED] wrote: A note from one manufacturer's site, for example, reads in part: Due to special signaling requirements necessary for Windows Plug and Play, the serial cable included with the UPS will not work in the Linux environment. I'm not sure however if the note is specific to the manufacturer's binary-only Linux software. The SOLA UPS (rebadged BestUPS AFAICT) we use at work has this problem. Our solution was to plug the special serial cable into a standard (straight through) serial cable with the offending pin removed. Frank -- Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://xenu.apana.org.au/ntshp/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Galeon quit working
On 12 May 02 18:14:29 GMT, Mario Vukelic [EMAIL PROTECTED] wrote: On Sun, 2002-05-12 at 20:05, Dale Hair wrote: /usr/bin/galeon-bin: relocation error: /usr/lib/libgtkembedmoz.so: undefined symbol: GetFlatBufferHandle__C10nsACString Did you upgrade mozilla? I believe the debian packaging system is supposed to make that question moot. Frank -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Eth0 issue
On 16 Apr 02 21:04:59 GMT, curtis [EMAIL PROTECTED] wrote: /etc/default/pcmcia reads: PCMCIA='yes' PCIC=xircom_cb[was =yenta_socket, but either way I get the same results] According to the PCMCIA HOWTO: PCIC This identifies the PC Card Interface Controller driver module. There are two options: `tcic'' or `i82365''. Virtually all current controllers are in the `i82365'' group. This is the only mandatory option setting. What happens when you use either of these two options? Frank -- Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://xenu.apana.org.au/ntshp/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Exim - remove many messages
On 26 Mar 02 10:43:43 GMT, Rory Campbell-Lange [EMAIL PROTECTED] wrote:
This is really a two part message. The first part is about removing many
messages from Exim. How does one do that easily, if one has to have
the precise id for each?
grep the output from
# mailq
which leads to...
Secondly, I'd like to make something like the line below work in bash.
Exim complains in this case of not finding a message with id -.
cat /tmp/e | exim -Mrm -
Try putting it into the exim command line, like this:
# exim -Mrm `cat /tmp/e`
Other people have suggested xargs, but this Works for me(TM).
I solved my problem cludgily by making an output file of exim -bp and
then vimming it so that it contained a set of exim -Mrf id lines. I
then sourced the file.
There must be a more elegant solution without using perl or python.
This will get rid of all frozen messages, caused for example by
someone forging a non-existant user on your host as the reply address
for spam:
# exim -Mrm `mailq | grep frozen | awk '{print $3}'`
Another option is to add an auto_thaw time option to exim.conf,
along with freeze_tell_mailmaster = false. This will prevent a
shitload of annoying mails being sent to you, while causing exim to
keep trying to deliver the offending messages until they timeout
(according to the documentation at least).
Frank
--
Home Page: URL:http://thingy.apana.org.au/~fjc/
Not the Scientology Home Page: URL:http://xenu.apana.org.au/ntshp/
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: How do remove frozen messages from Exim Queue?
On 14 Mar 02 20:58:12 GMT, Dave Sherohman [EMAIL PROTECTED] wrote:
On Thu, Mar 14, 2002 at 07:55:24PM +0100, Troels Petersen wrote:
How do I remove frozen messages from the exim queue, messages like :
(output from 'exim -bp')
14d 1.7K 16gLfk-0001Ka-00 *** frozen ***
[EMAIL PROTECTED]
13d 1.8K 16gmi6-0001Fb-00 *** frozen ***
[EMAIL PROTECTED]
Is there a way to confirm deletion of each message - or just a way to remove
all frozen messages?
exim -Mrm 16gLfk-0001Ka-00 16gmi6-0001Fb-00
# exim -Mrm `mailq | grep frozen | awk '{ print $3 }'`
Frank
--
Home Page: URL:http://thingy.apana.org.au/~fjc/
Not the Scientology Home Page: URL:http://xenu.apana.org.au/ntshp/
Re: potato version of galeon?
On 17 Feb 02 21:31:06 GMT, will trillich [EMAIL PROTECTED] wrote: what's the potato-friendly way to get galeon installed? galeon hasn't been potato-friendly since about version 0.7.x. It is a bleeding-edge gnome application that requires libraries and other facilities only available in gnome 1.4+. The simplest way to keep your system basically potato is to install ximian's gnome packages, including their versions of galeon and mozilla. That will be less of a jump than the other serious alternative, which is to upgrade to woody or sid, but incompatabilities between ximian's packages and the official debian ones may make upgrading to woody/sid later a bit problematic. Frank -- Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://xenu.apana.org.au/ntshp/
Re: Where to slice a 2 gig drive ?
On 22 Dec 01 23:44:12 GMT, Karsten M. Self kmself@ix.netcom.com wrote: on Fri, Dec 21, 2001 at 02:39:39PM -0500, lee ([EMAIL PROTECTED]) wrote: Ok..I'm fairly new to linux and extremely new to debian (was mandrake 8.1)..I'm attempting to install 2.2r2 on a 2 gig drive here and not really sure where to carve this drive up. I'm planning on using this box as a proxy for 6 other machines (combo of linux/98se). Linux docs has a few articles on this but I thought I'd come straight to the horses mouth to learn what might be best :-) 2 GB is a bit on the smallish side. 2 GB is perfectly adequate for the task. I ran far more than just a proxy on 2 GB for years. If you're using it as a proxy, I'd probably set up /, /tmp, /usr, and /var as separate partitions. Depending on what proxy services you're offering, you might want to make /var the bulk of the partitions (squid, ferexample, dumps its cache there). I'd avoid complicated partitioning schemes, especially if you have no previous experience running a server with the sort of workload you are expecting. The more partitions you create, the more (hopefully) educated guesses you have to make about the space required, and the more opportunities you have to get it wrong. Overestimate and you waste space (which *is* a consideration with only a 2 GB disk). Underestimate and you will find yourself running out of space on one partition while plenty of space remains on others. A seperate /var/spool partition makes sense given the use this box will be put to. You can mount it with the noatime option to speed up disk access and reduce or eliminate the reserved space (5% by default) to maximise available space. Allocate 1.2 or so GB to this and you can have a 1 GB squid cache plus space left over for other /var/spool users. A minimal swap partition (64 - 128 MB) should be fine. If the box starts to swap regularly you will want to add RAM otherwise performance will suffer severely. The rest can go into a single / partition. If this ends up bigger than 500 - 600 MB, make the /var/spool partition bigger. Frank
Re: Latest Galeon? will it work on potato?
On 25 Nov 01 13:10:53 GMT, Stan Brown [EMAIL PROTECTED] wrote: I saw the slashdot articel on the latest galeon. Since i'm looking for something to replace netscape (which has crashed 3 time this morning, already). I went tot hat site. I'm runing potato + Progeny + 24. kernel. So you aren't actually running 'potato', but some hybrid system. Galeon won't run on a 'pure' potato system, and hasn't done so since around 0.7.x or so (based on mozilla M18). Galeon is a bleeding edge Gnome program, so you need to have Gnome 1.4 installed. The best bet if you wish to keep your system basically 'potato' is to install Ximian's gnome packages. I'm running galeon 0.12.1 on a potato + ximian system. Frank -- Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://xenu.apana.org.au/ntshp/
Re: ssh without password for secvpn
On 20 Nov 01 18:27:53 GMT, Brooks R. Robinson [EMAIL PROTECTED] wrote: I try to connect and I am still asked a password. I've tried it with both empty passphrases and obnoxious passphrases, and I get the same result (password not passphrase). I've muddled thorough the man pages for ssh and the vpn-howto, but I seem to be missing the final bit that makes it work. Is my problem that I am using a mix potato and woody, or am I just missing some configuration. I've experienced similar problems when the .ssh/authorized_keys file was created group-writable. Frank -- Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://xenu.apana.org.au/ntshp/
Re: allowing root to display to a user's X session
On 19 Nov 01 21:25:47 GMT, David Wright [EMAIL PROTECTED] wrote: When troubleshooting on RedHat, I often log in to a X session as a user, then su to root in an xterm and run ethereal (a packet-sniffer with GUI) to watch the network traffic that results from my actions as a user. I would like to do this on Debian, but when I try to start ethereal, I get the error message: Xlib: Client is not authorized to connect to Server Apparently root is not allowed to display to a user's X session. How can I allow this? Have you tried 'su -p'? Frank -- Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://xenu.apana.org.au/ntshp/
Re: Amiga installation
On 15 Nov 01 01:34:12 GMT, Joseph Doss [EMAIL PROTECTED] wrote: This is a multi-part message in MIME format. --=_NextPart_000_000F_01C16D32.93398FA0 Content-Type: multipart/alternative; boundary==_NextPart_001_0010_01C16D32.93398FA0 You probably have no idea what a bloody mess this looks like in a proper news or mail client, otherwise you would set yours to post in plain text. Whenever I try to instal Debian on my 68040 Amiga, I get to the endo of = the installation where it asks if I want to enable it to boot from a har = drive and when I acknowledge this it tells me, Installing a boot loader = is not yet possible for Debian/m68k ..Whats up with this? What am I = doin wrong? How can I get around this? As far as I know there is no native boot loader for linux on an amiga. You need to keep a bootable AmigaOS partition, and run amiboot from the S:Startup-Sequence to load the linux kernel and reboot into linux. My A2000 is in bits right now, otherwise I'd give you more detailed incantations. Have another look at the installation notes, I'm pretty certain the whole process is adequately described there. Frank -- Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://thingy.apana.org.au/~fjc/scn/
Re: 3com NIC question
On 12 Jul 01 06:55:28 GMT, Guy Geens [EMAIL PROTECTED] wrote: Sunny == Sunny Dubey [EMAIL PROTECTED] writes: Sunny hey, I have a 3c509 NIC. Sunny I tried getting an IP for this NIC by using DHCP. (I used Sunny dhcpcd) But for some odd reason, it wouldn't work, and so I Try `modprobe 3c509' and restart the network with `/etc/init.d/networking restart'. Apart from disabling PNP, I never had any problems with this card. ISA PNP is a disaster. With PNP disabled and IRQ+IO statically configured 3c509's are great NICs, especially compared to SMC's junk. Frank
Re: Galeon 0.11.0 with Mozilla 0.9.1
On 17 Jun 01 17:48:05 GMT, csj [EMAIL PROTECTED] wrote: Has anybody successfully built galeon using the mozilla version found at non-us.debian.org/debian-non-US/pool/non-US/main/m/mozilla/? Yes. Try using --with-mozilla-libs=/usr/lib/mozilla-0.9.1 as an argument to configure. I also had to grab the hack-macros directory from the galeon CVS. Frank
Re: acroread
On 20 Apr 01 20:32:39 GMT, Lance Heller [EMAIL PROTECTED] wrote: When running acroread 4.0 (x86 linux stout0124 Jan24 2000 15:00:03) on my 2.2r2 x86 system running a 2.2.18 kernel, acroread commonly gobbles up all available system memory causing the system to hang. This happens on three different but similarly configured systems. Is this a known problem See http://www.debian.org/Bugs/. and if so is there a 'fix'. You would have to ask Adobe that. That's life with non-free software... Frank
Re: Debian and Trident 9385 VGA Card
On 19 Mar 01 22:07:23 GMT, Greene, Sam [EMAIL PROTECTED] wrote: I'm trying to install Debian on my laptop. My card is listed at http://www.xfree86.org/cardlist.html as supported. Is a generic driver the only thing that is offered? The best resolution I can get is 640x480. I configure it for 800x600 and 16 bit color with no luck. Are you definitely configuring it to use the SVGA driver? Try capturing the output of the X server and looking for things like video RAM detected, and what modes are rejected. Frank
Re: DHCP unavailable during installation
On 19 Mar 01 10:20:01 GMT, Justin Jones [EMAIL PROTECTED] wrote: FreeBSD. Here is what is detected during startup: 3Com 3c905 Boomerang 100baseTx at 0x1440, 00:60:08:14:86:44 IRQ 3 ^ If you have the usual two serial ports with the standard settings, then you have an IRQ conflict here. The IO address looks fishy to me as well, I usually see NICs down around 0x200-0x300. Unfortunately, when I try to configure my network during install, DHCP/BOOTP doesn't work, nor do the defaults or my best guesses in the manual setup. I've had trouble with PnP BIOS and ethernet cards with Debian before, so I played around in the BIOS, but there aren't a lot of options. I can pretty much only enable/disable bus mastering, and enable/disable the card itself. It was on IRQ 11 but the BIOS suggested I move it to IRQ 3. IRQs 5, 9, 10 and 11 are usually free. Something (VGA? sound?) may be stealing IRQ 11, 10 is the one I would normally pick. Could this be a BIOS conflict? Is there an incompatibility with my network card? Why won't DHCP/BOOTP function properly with Debian? DHCP works fine every day for me and Debian. Frank
Re: Applying colors to Netscape
On 17 Mar 01 20:13:05 GMT, William Leese [EMAIL PROTECTED] wrote: I think I found this mentioned somewhere, something to do with editing your .Xresources file. Does anyone know more of this? If you are using Gnome or KDE, try installing grdb or krdb. This will apply your GTK/KDE theme to 'legacy' applications like netscape. On my woody system there's also /etc/X11/Xresources/netscape which might be what you are looking for. Frank
Re: Mozilla requires libnspr4?
On 10 Mar 01 08:16:33 GMT, Jonathan Gift [EMAIL PROTECTED] wrote: I dl the latest Mozilla 0.8 build from Mozilla.org and have been having some trouble. The only library that I loaded on install was libstdc++2.9-glibc2.1. Now someone gave me the site for a potato deb and alongside the deb and tar is a deb of libnspr4_0.8-0.0.1. Now I don't have this library loaded on my system. Cosidering where I found it, is it needed for Mozilla? If you installed the mozilla tarball I suspect you already have it. libnspr4 is built from the mozilla source. It is broken out into a seperate package presumably because it may be useful for apps other than mozilla. Frank
Re: Mozilla requires libnspr4?
On 10 Mar 01 08:35:16 GMT, Jonathan Gift [EMAIL PROTECTED] wrote: I tried installing the deb in question and got conflicts. It wants a newer set of libraries than my potato has. That'll be because it was built on a woody or sid system. The same packages built for potato are at http://honk.physik.uni-konstanz.de/~agx/debian/. So, do I need it for Mozilla 0.8, and if so, will the libnspr4_M18 work which came with potato, was compiled on the correct libraries, and installs fine? If you install the mozilla 0.8 .deb, yes you will need it. If you install a mozilla tarball, no. The M18 version almost certainly won't work with mozilla 0.8. Frank
Re: Mozilla 0.8-Special setup required?
On 9 Mar 01 14:51:26 GMT, Jonathan Gift [EMAIL PROTECTED] wrote: I dl the 0.8 build, untarred and zipped it to /usr/local/share/mozilla than ran /usr/local/share/mozilla/mozilla to get it started. I get lot's of error message and changing themes crahses it cold. They couldn't have released this as is, so it's me. What haven't I done? I'd love to use it, so any help appreciated. mozilla requires write access to the directory where the binaries live the first time it is run. Try running it as root once before running it as a normal user. Even better, get the debianised mozilla 0.8 from http://www.debian.or.jp/~kitame/mozilla/ and have all this dealt with for you. Frank
Re: ldap debian
On 23 Feb 01 00:50:46 GMT, Known Human Nick Rusnov [EMAIL PROTECTED] wrote: Well I'm having a heck of a time getting a freshly installed Potato box to authenticate with an openldap server. I'm new to this whole ldap thing, is there a guide somewhere to reconfiguring debian to use ldap for things? Not that I was able to find. I did find these: http://staff.pisoftware.com/bmarshal/publications/ldap.html http://people.redhat.com/alikins/ldap/ldap.html http://www.redhat.com/support/manuals/RHL-7-Manual/ref-guide/ch-ldap.html All the guides I've found just have a pam.conf example that I'm not sure how to translate into using with the pam.d setup... (I tried, for example, taking the lines that started with login in the examples and addinf them to the login file in pam.d). The general idea is to precede each pam_unix.so line in the pam.d/* files with a line like this: auth sufficient pam_ldap.so Frank
Re: mouse troubles with Debian on Dell Latitude CPx J650GT
On 8 Feb 01 21:53:48 GMT, Michael A. Miller [EMAIL PROTECTED] wrote: I've recently installed Debian (stable) onto a Dell CPx J650GT laptop. After installing the base system and the required packages, I added the SVGA and Mach64 xservers and the depended-upon packages. startx works in that I get a screen with an xterm in it, but I cannot move my pointer with the touch pad, the pointing stick, or an externally attached ps/2 mouse. This is using the PS/2 protocol and /dev/mouse in my XF86Config file. If I switch to /dev/psaux, the pointer zooms off screen as soon as I touch the touch pad or mouse. Any suggestions on how to trouble shoot this? Gpm seems to work fine on the console. If you accepted the default config for gpm then it should be acting as a repeater making translated mouse events available through /dev/gpmdata. Configure X to use the Intellimouse protocol and /dev/gpmdata as the mouse device. Frank
Re: Gnome window sizes.
On 6 Feb 01 16:27:39 GMT, Hans [EMAIL PROTECTED] wrote: Running (Helix)Gnome at 800x600 produces mostly horribly large windows, especially the terminal. I looked at the Gnome website and they basically say you have to live with that. Being stubborn: has anyone found a solution for this? --Hans There are a few things you can do, mainly involving reducing font sizes. First off, edit your /etc/X11/XF86Config[-4] to make sure that 75dpi fonts are searched *before* 100dpi fonts. The rationale for this is that some apps (notably Netscape and the Gnome Control Center) specify fonts in the UI in points instead of pixels. By definition a 10 point 100dpi font is larger (in pixels) than a 10 point 75dpi one. This change should at least let you see the OK button in the Netscape preferences dialog without hiding the panel. In the Control Center Theme Selector, check Use Custom Font and reduce the size of the default font. I find a 12 pixel font to be acceptable, but you know your own eyesight best. While you are there, consider changing the GTK theme to one that is a bit less bulky. I use XenoThin from gtk-engines-xenophilia (only in unstable, but it builds from source on a potato+helix machine with no complaints). Try installing grdb. This will apply your GTK theme settings to non-GTK apps and helps with ones that don't have easily configurable UIs. Try reducing the panel height to 24 pixels. You could also configure it to auto hide (I prefer not to, to keep the clock and other monitors visible). Otherwise leave the hide buttons active but disable the arrows so they take less space. If you have launchers on the panel try putting them in one or more drawers. If you have a desk guide applet think about whether you really need it, or if you can get by with fewer workspaces. You can also change the geometry to use 2 rows of panes. Configure the clock applet (if you use one) to use 24 hour time and show the date in a tooltip. Expand the task list applet to fill any space you create. Gnome Terminal can be slimmed down considerably. Change the preferences to hide the menu bar; if you ever need it you can bring it back with a right click. While you are there reduce the font size and consider hiding the scroll bar. Window decorations a big space hog. Try different WM themes to see which waste less space. If all the slim ones are too ugly you may want to adapt an existing theme by reducing images and border sizes. This is on my own todo list. I regularly install potato + helix gnome on 800x600 laptops (did one only yesterday) and I use all these tricks to make better use of the screen real estate. Frank
Re: galeon..
On 28 Jan 01 18:49:02 GMT, Rob VanFleet [EMAIL PROTECTED] wrote: The people you really need to talk to are the galeon developers (or at least whoever packages the deb's for it). If they're packaging a deb for Potato and making it depend on X4, it seems they are obviously out of touch with Debian. The main galeon packager has to rely on other people to do potato packages, and in this case didn't know that the latest package was built against X4 until after the fact. As it stands now, galeon cannot be built on a pure potato system. It uses features of the Gnome libraries that aren't available in the increasingly obsolete version released with potato. It does build on a potato system updated with the current Helix/Ximian Gnome packages. It also depends on mozilla 0.7 or a recent nightly rather than the M18 in potato. Building it requires xml-i18n-tools which isn't officially packaged even in sid, and building the docs requires gnome-doc-tools which is also not officially packaged. In the near future packages should be available for potato systems updated with Helix/Ximian Gnome and an unofficial mozilla 0.7. Depending on how difficult it is to do, packages built against a pure potato system and mozilla 0.7 *may* become available later. Frank
Re: xml-i18n-tools ?
On 26 Jan 01 23:49:02 GMT, Pollywog [EMAIL PROTECTED] wrote: Is xml-i18n-tools available as a Debian package, perhaps under a different name? I can't find it. No, but I expect it will be soon. The changelog only starts on Jan 3 2001. It's quite straightforward to package. Frank
Re: Sorry, off topic, but HOWTO get to bios with IBM Thinkpad 755...
On 21 Jan 01 02:37:20 GMT, Carl Fink [EMAIL PROTECTED] wrote: IIRC, you hit F1 during the bootstrap. The manual I read (ThinkPad 560X) said to hold the F1 key down, power on, and continue holding down F1 until the BIOS setup came up. Frank
Re: x on my laptop problems
On 19 Jan 01 06:58:32 GMT, Forrest English [EMAIL PROTECTED] wrote: in my trying to apt all the right files for my laptop to run x, i seem to have missed somthing. when i attemt to run x... Fatal Server error: could not open default font 'fixed' what does this mean, and what do i need to do to fix it? It means you've neglected to install a necessary font package. The simplest solution may be to 'apt-get install task-x-window-system-core' which *should* pull in everything you need. Frank
Re: x on my laptop problems
On 19 Jan 01 15:52:36 GMT, Forrest English [EMAIL PROTECTED] wrote: i just tried apt-get install tast-x-window-system-core it broke debconf. it said it was upgrading debconf (sources pointing at unstable), and now all packages won't install. because debconf won't install because of aparently a perl problem. and then everything else fails on depconf because it 'isn't configured yet' would anyone like to help? or is this just what i get for trying to run unstable? I wasn't aware you were trying to go straight to unstable from a (presumably) stable base system install. Aiming for a bleeding edge unstable system straight off may be a bit ambitious. Practically speaking, since there are no unstable (or even testing) boot floppies, you are going to be installing a stable (potato) system anyway, then upgrading it. I would suggest taking things slow and steady, spending some time working with a stable setup before taking the big plunge. But if you are impatient, then try 'apt-get dist-upgrade' before trying to install any more packages. Don't be shy about running apt-get again if it fails, it often takes a different approach on the second attempt. man... i don't want to install from floppies a second time... that took me over an hour. I think you will have to resign yourself to spending a bit more than an hour on this :-) If you do start over, here's how I suggest you go (I'm assuming here you will be doing a network install using the NIC you mentioned previously). * Start the install using the basic boot floppy set. You won't need all the base system floppies, just the 5 for rescue, root and drivers. * When you come to the point where you are asked to configure drivers with configuring PCMCIA as an alternative, *take the alternative*. Do the PCMCIA config first, *then* the drivers. * Configure networking, then do the base system install over the network. * When you set up the apt sources, go for a stable install. * When you are given the option of a simple or advanced package selection, go for 'simple'. Go for a bare minimim selection of task packages at this stage. In your case probably just task-x-window-system-core. You can always run 'tasksel' again after the install. At the end of the process you may or may not have X working, but you should at least have all the necessary packages installed. Getting X to work on a laptop can be a bear, and you can expect to be coming back for more help at that stage. When you eventually have things running you can then think about changing your apt sources and upgrading to unstable, although I would advise you to try testing first. Frank
Re: 2.2r2 and 3c509 network card
On 18 Jan 01 18:45:13 GMT, David S. Bach [EMAIL PROTECTED] wrote: Debian 2.2r2 (kernel 2.2.18pre21) is unable to use the 3c509 ISA network card. Consider yourself lucky you aren't stuck with an SMC Elite or Ultra. In my experience the 3COM cards tend to Just Work with little persuasion. I will now endeavour to actually be helpful :-). In modconf, if I attempt to insert 3c905 modules without parameters,installation fails. In modconf I am apparently able to insert the module using 3c509 io=0210 irq=12. These are the resources used for the card in Win95. If you have a PS/2 mouse as well, you would have an IRQ conflict. However: ifconfig -a shows, in addition to the expected: Interrupt: 10 Base address: 0x300 (I'm guessing that this is significant since I cannot access devices on the Advansys SCSI PCI card that also claims interrupt 10. What can I do about this?) Looks like you will need to configure the card to use a different IRQ (5, 9 or 11 are good candidates) and possibly to live at another IO address. A setup utility for the 3com cards has been packaged for Debian: $ apt-get install 3c5x9setup Documentation is a bit on the short side, but try just running 3c5x9setup and see what it says. Frank
Re: configuring ethernet?
On 19 Jan 01 02:54:28 GMT, Forrest English [EMAIL PROTECTED] wrote: i have a pcmcia NIC that i want to use in my laptop, i managed to get the base system installed from floppies. but, now after the install, i need to setup my network connection.is there an application that i can use to do this? there was netcfg for redhat. what do i use in debian? vi. Assuming the PCMCIA packages are installed, edit /etc/pcmcia/network.opts to suit your setup. Make sure the PCMCIA software is running with '/etc/init.d/pcmcia restart'. Then insert the NIC and see how it goes. Frank
Re: LDAP auth for some users
On 9 Jan 01 20:48:54 GMT, Anton Emmerfors [EMAIL PROTECTED] wrote: Using the current style in /etc/pam.d/login I am asked twice for passwords which I don't want. So I tried converting to the new style (value=action) but then it won't work at all. You need to do something along the lines of: authsufficient pam_ldap.so authrequiredpam_unix.so try_first_pass Am I correct in assuming that a default=ok is equivalent to required? Couldn't say. Has anyone succesfully used a system like this or am I trying to climb the wrong tree? I had the same problem and the recipe above fixed it. Frank
Re: ldap, ldap everywhere
On 7 Jan 01 01:01:20 GMT, Known Human Nick Rusnov [EMAIL PROTECTED] wrote: I'm going to be running a moderately sized network, and I was wondering how difficult it would be to setup ldap as teh authentication for it? Having just done exactly that I can say it's probably more frustrating than difficult. I had an interesting time finding any sort of documentation specifically on setting up LDAP authentication. LDAP does much more than this, and what documentation there is tends to be more general than I was looking for. Eventually I turned up a few helpful web pages including: http://staff.pisoftware.com/bmarshal/publications/ldap.html http://people.redhat.com/alikins/ldap/ldap.html http://www.redhat.com/support/manuals/RHL-7-Manual/ref-guide/ch-ldap.html How integrated are various utilities into ldap? I read that adduser doesn't support ldap, so should a script that uses adduser to create the user, then copies the user into ldap be used? One possibility is addluser: http://www.open-it.org/download/prototypes/addluser.py Frank
Re: coping with a high-volume mailing list (like this one)?
On 29 Nov 00 05:34:28 GMT, Lawrence H. Robins [EMAIL PROTECTED] wrote: I'm curious to know what strategies are used by regular subscribers to this list to deal with the high volume of messages (250/day)? A mail2news gateway. A decent news client is always going to be a better bet for dealing with a high volume threaded discussion group. I'm responding to this in a gated newsgroup set up by my ISP, but I could just as easily have set it up myself, and I have done so. Suppose you only want to see messages with certain keywords in the subject line, or only replies to your questions, and filter out all the others? These are standard facilites provided by news clients like slrn and pan. Also, is there any way to filter the messages **before** downloading them from your ISP to your local machine, which takes time in itself? If you use fetchmail to collect your mail then I vaguely recall that something like this is possible. However, I don't use it so I may well be hallucinating. Frank
Re: coping with a high-volume mailing list (like this one)?
On 29 Nov 00 08:23:53 GMT, Pap Tibor [EMAIL PROTECTED] wrote: On Wed, 29 Nov 2000, Frank Copeland wrote: On 29 Nov 00 05:34:28 GMT, Lawrence H. Robins [EMAIL PROTECTED] wrote: I'm curious to know what strategies are used by regular subscribers to this list to deal with the high volume of messages (250/day)? A mail2news gateway. A decent news client is always going to be a better bet for dealing with a high volume threaded discussion group. I'm responding to this in a gated newsgroup set up by my ISP, but I could just as easily have set it up myself, and I have done so. Could you offer a mail2news gateway for home use? I would like to read mailing lists through news client but my ISP doesn't provide this. I do it all with inn, but that may not be an option in all circumstances, especially if you can't create arbitrary local email addresses. The method is: 1. apt-get install inn2 (inn may be suitable if it includes the mailpost script, I haven't used it for a while). 2. create a local *moderated* newsgroup for the list, for example: $ /usr/lib/news/bin/ctlinnd newgroup local.lists.debian-user m 3. create a mail alias for the list and pipe any mail for it through mailpost, for example (in /etc/aliases): debian-user: |/usr/lib/news/bin/mailpost -a [EMAIL PROTECTED] local.lists.debian-user 4. add an entry to /etc/news/moderators pointing back to the list, for example: local.lists.debian-user:debian-user@lists.debian.org 5. subscribe the mail alias for the group to the mailing list. The trick is in making the newsgroup moderated. The -a option to mailpost adds an Approved: header to each incoming article so that inn knows not to send it straight back to the list. Locally posted articles will lack an Approved: header so they will be mailed to the list address. It isn't perfect, and I use a slightly modified mailpost script that strips out some headers that tend to confuse inn if they appear in incoming articles. By default, inn will expire (delete) messages after 2 weeks. You can modify that by editing /etc/news/expire.ctl. If you can't create arbitrary local email addresses, for example if you have a dialup connection to an ISP, you will need to do something with exim .forward files or procmail scripts to identify list mail and pipe it into mailpost. I have no experience with that so I can't suggest a solution. If you are already using another news server such as leafnode, you will have to work out the equivalent setup yourself. Frank
Re: anXious usage?
On 18 Nov 00 23:50:40 GMT, Stan Brown [EMAIL PROTECTED] wrote: I found this programs name as a comment in the begining of /etc/X11/XF86config. Now when I run it, it just does nothng. The prompt comes right back :-( I've had the same problem. Try deleting or moving etc/X11/XF86Config, then run it again. Frank
Re: wine (woody) can't find KERNEL32.dll
On 24 Oct 00 06:23:38 GMT, Mark Phillips [EMAIL PROTECTED] wrote: I just upgraded wine to the woody version. A program that used to run on the potato version now crashes at the start, complaining: err:module:fixup_imports Module (file) KERNEL32.dll needed by C:\etax2000_1.exe not found Now I thought kernel32 was a builtin thing. That's what is in the wine.conf. Any ideas on what is wrong? Builtin modules are implemented as shared libraries which are stored in /usr/lib/wine. Wine needs to be told where to find them, which is done in wine.conf. This is a fairly recent change, and requires changes to wine.conf. Did you update your wine.conf when you updated the package? If there is a wine.conf.dpkg-dist try renaming it to wine.conf and editing it to reflect your setup. If that doesn't do it try adding /usr/lib/wine to your ld.so.conf and running ldconfig. Frank
Re: wine (woody) can't find KERNEL32.dll
On 24 Oct 00 08:25:26 GMT, Mark Phillips [EMAIL PROTECTED] wrote: Phillip Deackes [EMAIL PROTECTED] wrote: I went through the same thing at the weekend. Just open up /etc/ld.so.conf and add this line: /usr/lib/wine Do the usual ldconfig afterwards. /etc/wine.conf has changed a fair bit recently so it would be a good idea to install the maintainer's version and edit it to your requirements instead of using your version which, if it is anything like mine, was probably created many, many versions ago. Thanks for the help! I am guessing that the debian wine package should have added the line to /etc/ld.so.conf, but for some reason didn't --- is that right? I'm not familiar with how ldconfig etc works. I presume I just run ldconfig with no options (after changing the ld.so.conf file), is that right? wine shouldn't need to mess about with ld.so.conf. The only wine library that should be globally available is libwine (which is in /usr/lib), the rest are helpers that only wine needs to know about. There's a new option in wine.conf (EXTRA_LD_LIBRARY_PATH) that tells wine where to find them, but this didn't work as it should in older versions. In that case putting /usr/lib/wine in ld.so.conf was the recommended workaround. I'm running the 20001002 snapshot (built as a .deb by updating the existing debian package) without /usr/lib/wine in ls.so.conf and it goes just fine. Frank
Re: kde or gnome?
On 5 Sep 00 19:05:33 GMT, Felix Natter [EMAIL PROTECTED] wrote: Namely, Gnome does not include its own window manager; KDE does. Gnome depends on hooks for Gnome support compiled into an external window manager, and at present the only window manager with full support for Gnome seems to be Enlightenment, AKA `E'. sawfish is now called the official GNOME wm (although you can still change). icewm also fully supports GNOME. It's likely the lightest-weight of the three. Now, if only someone would do a MicroGUI theme for icewm... checks icewm.themes.org oh, silly me. Frank
Re: Wine 20000801
On 5 Sep 00 21:05:48 GMT, Richard Black [EMAIL PROTECTED] wrote: I have just upgrade to wine (my first upgrade of wine for some time). I've run into a couple of problems. The first one was: [...] LD_LIBRARY_PATH variable. However, after this, wine just hangs whenever I try and use it (it will print out the version number and help screen though) Does anyone else have this problem? You will find there have been fairly major changes in the default /etc/wine.conf, and I'd recommend saving your old one, installing the new one, then reapplying any changes you had previously made. However, the libwine .deb is missing an essential library file, so it won't work anyway. For now your best bet is probably to stick with the wine released with potato. Frank
Re: Wine 20000801
On 6 Sep 00 21:52:08 GMT, Jonathan Markevich [EMAIL PROTECTED] wrote: On Wed, Sep 06, 2000 at 04:24:22AM +, Frank Copeland wrote: However, the libwine .deb is missing an essential library file, so it won't work anyway. For now your best bet is probably to stick with the wine released with potato. libwine_unicode.so.*? Not missing, just in a different spot. I had a hard time tracking this one down. It's in /usr/share/wine/lib. No, libx11drv.so*. Clearly we are talking about different versions of the package. libwine_0.0.2801-1.deb is the one with the missing library; see bug #70819. It does not put any libraries in /usr/share/lib/wine anymore, they are all where they belong, in /usr/lib/wine. Frank
Re: Debian News Group Needed
On 29 Aug 00 12:27:19 GMT, Previ [EMAIL PROTECTED] wrote: I seriously believe that this and other debian lists are far better managed both by users and debian, using NewsGroups rather than mail lists. Try setting up a mail2news gateway. Install inn|inn2, create a local newsgroup, make it moderated, set the mailing list address as the moderator address and feed incoming articles into the group with mailpost. Just make damn sure you do it right, or you will piss off a lot of people. Frank
Re: junkbuster
On 26 Aug 00 21:26:30 GMT, Dale L . Morris [EMAIL PROTECTED] wrote: I'm afraid I may have botched this one. Trying to get junkbuster to work I removed it with 'apt-get remove junkbuster' then removed the junkbuster directory with rm -rf. Now when I reinstall junkbuster, there's nothing in the junkbuster directory. No config ..nothing.. guess I'll have to live with those annoying adds You should always use the 'purge' option to get rid of unwanted config files. I was able to reproduce your situation, and fix it like this: # apt-get --purge remove junkbuster # apt-get install junkbuster No need to give in to the ads. kmself@ix.netcom.com (kmself@ix.netcom.com) wrote: On Sat, Aug 26, 2000 at 11:38:58AM -0700, Dale L . Morris wrote: Could someone shed a little light on configuring junkbuster? I tried reading the man page and setting up netscape with manual proxy configuration on local host port 5865 but then I can't connect to anything. You want the following as your Netscape manual proxy config: host: localhost port: 5865 ...you'll also need junkbuster running (ps aux | grep junkbuster), or you won't have anything to talk to. Check your junkbuster configs under /etc/junkbuster/config. junkbuster is inclined to fail silently if it can't make sense of its config files. I'd suggest setting the debug variable to 15 or 16 initially, so you can see what's going on. Comment it out again and restart junkbuster when you have things working. Until you add URLs to your blockfile, you won't actually block anything. The sample blockfile in /usr/share/doc/junkbuster/examples is a good start. Frank
Re: How stable is WINE?
Cameron Matheson wrote: I'm waiting for the new Debian to come out, and I need some information about WINE. In Potato, how stable is WINE? In my experience the wine currently in potato is as stable as any version of wine I've used, and better than most. However, it is classified as alpha software for a very good reason. It comes nowhere near running all windows software. Each monthly snapshot improves some aspects but often breaks something that worked previously; very much a two steps forward one step back process. Does it run better than windoze? Also, What's the speed like, is it as fast as the app would run in windoze? No and no. If it works at all with a given application then it works well enough, but you may have to work around an annoying bug or two (like shift-clicking the mouse occasionally freezing the app or even crashing X). Speed is sufficient considering the source is full of debugging code and not optimised in any way. If you are looking for a general replacement for windows that runs whatever windows runs then expect to be disappointed. If you have a specific application you need to use then the only way to find out if it will run under wine is to try it. If you have trouble then ask for help on news:comp.emulators.ms-windows.wine. Frank
Re: How stable is WINE?
Ethan Pierce wrote: In my opinion WINE isnt all its cracked up to be, but better than it has been in the past. I don't know that wine has ever been cracked up to be more than it is, at least not by the developers. It is certainly getting better. If you want to run windows apps in linux, install vmware (www.vmware.com). It runs VERY fast depending on how much virtual ram you can afford to allocate. I use 128mb for my virtual machines. Ive even had execellent results with fullscreen windows media player under vmware. If what you really need is to be able to run windows and linux at the same time then vmware is certainly cheaper than the other obvious solution, which is to buy another box. However, the hardware and software required is still a few hundred bucks beyond my reach. For my purposes it would also be gross overkill. wine has a long way to go before it provides a general replacement for windows, but frankly that doesn't bother me one bit since I won't be using it for that. Even so wine does two things that make it extremely useful right now. One is that it allows people like Corel to port their applications written for the windows API to linux without too much pain. In fact much of the recent improvement in wine can be attributed to Corel. The other is that it allows people to run (some) niche applications for which there is currently no native linux alternative, which is what I use it for. So depending on what the original poster actually wants to do, wine may well be good enough. If it works for him at all, it will certainly be less expensive and resource hungry than the alternatives. Frank
Re: I cannot get JunkBuster to work
Barry Samuels wrote: I would like to continue using woffle and to use junkbuster in addition but the instructions are going to have to be of the 'put this here' and 'put that there' type I'm afraid. In /etc/junkbuster/forwardfile make sure you have something like this at the very end: * your.host:8080. . This line says forward all requests to your.host port 8080 after junkbuster is done with them. Then make sure you have the forwardfile option correctly specified in /etc/junkbuster/config. Point your browser at the junkbuster port (8000 I presume). Frank
Re: Kernel Compile error: What am I missing?
Russel wrote: Oops...I've got either a 2.2.10, 2.2.12, and 2.2.14. Tried em all. I've got the binutils and gcc versions required in Documentation/Changes ( I even checked using the commands listed in that file) If the file builds just fine on your computer...has the file changed between kernel versions? Has it changed at all in years? Would something else have changed? Naturally, I must be missing something. I just wish I knew what. ugh. I think this might be your problem: --- kernel-source-2.2.10/arch/m68k/fpsp040.orig/bindec.SFri Apr 26 05:12:35 1996 +++ kernel-source-2.2.10/arch/m68k/fpsp040/bindec.S Tue Oct 12 17:36:55 1999 @@ -484,7 +484,7 @@ fmovex (%a0),%fp0 |load X from memory fabsx %fp0|use abs(X) tstw%d5 |LAMBDA is in lower word of d5 - bnessc_mul |if neg (LAMBDA = 1), scale by mul + jne sc_mul |if neg (LAMBDA = 1), scale by mul fdivx %fp1,%fp0 |calculate X / SCALE - Y to fp0 brasA10_st |branch to A10 The line # matches the one in your error message: Assembler Messages: Bindec.S :487 /usr/src/linux/arch/m68k/fpsp040 [Bindec.o] error 1 value of -512 too large for field of 1 bytes at 511 Judging from the error message I'd say that the source has changed to the extent that sc_mul is now out of reach of a short (8 bit) branch from that point and now requires a (32? bit) jump (sorry, my m68k assembler is a bit rusty these days). That fragment comes from a patch named binutils295-fpsp040-2.2.10.diff.gz included in kernel-patch-2.2.10-m68k_2.2.10-6.deb. I presume then that you are rolling your own kernel from the upstream sources. Looks to me like you might have better luck with the Debian kernel-source* and kernel-patch* packages. Frank
Re: Kernel Compile error: What am I missing?
Russell wrote: Here's the message I get: Assembler Messages: Bindec.S :487 /usr/src/linux/arch/m68k/fpsp040 [Bindec.o] error 1 value of -512 too large for field of 1 bytes at 511 I asked the people on the m68k list, and they didn't know. I am sure I'm missing some vital library or other piece of code in the compile process, I just can't figure out what it could be. Can anyone help? Sorry, that file builds just fine on my Amiga, kernel 2.0.36. What kernel version are you building, with what gcc and binutils versions? Frank
Re: HOW DO I Apply debian patch to original source?
Brian Lavender wrote: I am trying to compile a debian package from source. I can go to the package's web page and get the original source along with the diff. How do I apply the patch? Get the *.dsc file as well. Put all three files in the directory you want to unpack in, and cd to it. Then run 'dpkg-source -x wu-ftpd_2.6.0-4.dsc'. You will need to have the dpkg-dev package installed. Frank
Re: wine broke with libc6_2.1.3-5
John Bagdanoff wrote: The current potato wine package worked flawlessly for me, so I didn't save the previous version. I got the latest one in woody, but still wine breaks. If you could, polly, upgrade to the latest libc6 (2.1.3-5). If wine starts up for you, then I'll go to the wine news group to see if they can sort it out. wine breaks for me too. I get a stack dump and backtrace that seems to implicate the libc towupper() function. I've filed a bug against libc6, we'll see what comes of it. Frank
Re: wine broke with libc6_2.1.3-5
John Bagdanoff wrote: Frank Copeland wrote: wine breaks for me too. I get a stack dump and backtrace that seems to implicate the libc towupper() function. I've filed a bug against libc6, we'll see what comes of it. Thanks for the verification, Frank. Guess I'll sit tight for a couple of days. This issue is starting to get a mention on comp.emulators.ms-windows.wine. The suggested workaround is to set LC_ALL=en before running wine, and I can confirm that this works for me. Frank
Re: Squid Proxy server-
Tom wrote: Okay i have done this before but cant remember what i typed to get it to work. I need to clear out the cache on my proxy server so that it wont keep showing some items that i have changed. how do i do this? Basicaly I think what im looking to do is clear out my proxy cache, but i cant remember how to do that for the life of me. I am running a squid proxy server. $ squidclient -m PURGE url However, this shouldn't be necessary. Shift-clicking the reload button in Netscape should force a reload regardless of what's cached. If the pages are local, try adding your own hostname to the No Proxy settings of your browser. You can do the same thing with squid using the always_direct config setting. Frank -- Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://thingy.apana.org.au/~fjc/scn/ Keep it in Usenet. E-mail replies and 'courtesy' copies are not welcome. If you're selling, I ain't buying.
Re: GNOME performance hit on 486?
Scott Au wrote: I've been itching to install Gnome both for the utilites as well the greater desktop control. I'm worried about the performance hit my box will take though with the increased graphical desktop (as compared to the simpler nature of IceWM). Anybody running a similar processor (I realize people running 700 Mhz Athlons aren't going to notice any difference) have any input to this matter? The 486 has 32M RAM and runs Netscape incredibly well compared to Windows 98 (which doesn't run at all on this system). From memory (I ran Gnome on a 16MB 486 before upgrading it to a K6-2/350) the main performance hit was when Gnome applications started up. There seemed to be a significant overhead involved in just setting up the GUI, and this was especially noticeable when using the Control Center. Then again, memory was always tight (especially when Netscape was running), so you may well do better. I think there was also an issue with the image processing library used by Gnome at the time (ImageMagic?) which seemed to take gobs of memory and CPU cycles to do fairly minor things. I believe this has now been replaced. Ultimately the only way to find out if it will work is to try it. If you are already running Netscape well enough on the box I don't think Gnome will be too much for it. Frank
Re: Poor Modem performance
Scott Au wrote: After installing and running minicom, I've noticed that my modem, a BOCA ISA 28.8 modem is incredibly slow. Dialing my ISP shows text being received with periodic pauses. From personal experience one possibility is that serial interrupts are not being serviced quickly enough, which can be a big problem with older slower hardware. If too many interrupts are missed then you get serial buffer overflows, dropped packets and delays as TCP/IP tries to recover. There are two things you could try, if this is the case. One is to use irqtune from the hwtools package to boost the serial interrupt priority. The documentation explains what is going on. The other thing, if you have an IDE controller, is to unmask interrupts using 'hdparm -u1'. There are dire warnings attached to that option, so proceed with care. Frank
Re: Download once, apt-get install many?
Robert L. Harris wrote: I'm looking at upgrading to potato. I'm doing the apt-get -d dist-upgrade currently since it looks like it'll take 1day and 15hrs per box. Since the -d downloads, can I take the files being downloaded for box1, tar them up, copy and untar to box[234] and then just apt-get install from the single downloads? I'd hate to spend a full week downloading the same files multiple times... If the boxes are networked together, you can easily set up your own partial mirror and install from that. Once you have upgraded box #1, use apt-move to create a partial mirror from the .debs you downloaded, and set up a web/ftp/NFS server to make the mirror available to the other boxes. Then when you upgrade the other boxes, add box #1 to their /etc/apt/sources.list as the *first* (or only) source. Disclaimer: it works for me. Frank
Re: Download once, apt-get install many?
Ben Lutgens wrote: On Fri, Dec 31, 1999 at 04:49:59PM +1100, Frank Copeland wrote: If the boxes are networked together, you can easily set up your own partial mirror and install from that. Once you have upgraded box #1, use apt-move to create a partial mirror from the .debs you downloaded, and set up a web/ftp/NFS server to make the mirror available to the other boxes. Then when you upgrade the other boxes, add box #1 to their /etc/apt/sources.list as the *first* (or only) source. ouldn't you manually copy the contents of /var/cache/apt/whateverthehellitis or mount the original box's /var/cache/apt/ dir via nfs? Since you are couching that as a question, I assume you haven't actually tried to do it that way. And my answer would be: I don't know. If you ever try it, let us all know how it goes. I do know that apt-move does the job, without me needing to rummage about in the guts of the package management system. Frank -- Please don't CC: me. I read the list.
Re: Boot off non-BIOS SCSI drive - can I use LILO or SysLinux?
Brian wrote: Presumably, I cannot use LILO to dual-boot the machine (after all, how would the machine let LILO see the partition with the kernel?). Up to now, I have been booting off of a floppy, and that is torture. Have you actually tried LILO? I have a SCSI-only box that the BIOS thinks has no disks at all (there isn't even a floppy drive) and it boots quite happily from the SCSI root disk. The only gotcha was that I had to put the linear option in /etc/lilo.conf, but you may not find that necessary. Frank
Re: An open letter to the debian community
David Blackman wrote: Lately I've been thinking about forking Debian, into DWA, meaning Debian Without Attitude. We'll drop the attitude, and the pretenses, about what Free means, and get licensing deals with Corel, Netscape, and Sun, to include Wordperfect, Communicator, and Staroffice. We'll make the install process less cryptic, include non-free on the CD and forget the Debian philosophy, that the only way to learn is by doing it the hard way. I happen to like Debian *With* Attitude. It fills a very important niche in the Linux ecosystem, and if Debian changed in the way you want it to then it would be necessary IMHO to reinvent it under another name. If people want a glitzy distribution with all the semi- and fully-commercial software included they already have plenty to choose from, some of them based on Debian. If Debian stops trying to implement the free software ideal (even as imperfectly as it does), what will take its place? *Something* will have to, because there are simply too many people out there who think the ideal is actually worth pursuing. Like me. Frank
Re: Can't make menuconfig
Kent West wrote: I can make config and make xconfig but I can not make menuconfig. When I try, I get the following: westk03:/usr/src/linux# make menuconfig rm -f include/asm ( cd include ; ln -sf asm-i386 asm) make -C scripts/lxdialog all make[1]: Entering directory `/usr/src/kernel-source-2.2.1/scripts/lxdialog' gcc -Wall -Wstrict-prototypes -O2 -fomit-frame-pointer -DLOCALE -DCURSES_LOC=curses.h -c lxdialog.c -o lxdialog.o In file included from lxdialog.c:22: dialog.h:29: curses.h: No such file or directory make[1]: *** [lxdialog.o] Error 1 make[1]: Leaving directory `/usr/src/kernel-source-2.2.1/scripts/lxdialog' make: *** [menuconfig] Error 2 westk03:/usr/src/linux# exit exit Can anyone suggest a fix? Thanks! [EMAIL PROTECTED]:~$ dpkg -S curses.h libncurses4-dev: /usr/include/curses.h libncurses4-dev: /usr/include/ncurses.h Try installing the libncurses4-dev package. -- Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://thingy.apana.org.au/~fjc/scn/ Keep it in Usenet. E-mail replies and 'courtesy' copies are not welcome. If you're selling, I ain't buying.
Re: atd fails? (potato)
Scott Henry wrote: I did a recent apt-get upgrade, and now atd fails. It used to work fine. It seems to only process items when restarted, and leaves defuct children around. I have at version 3.1.8-7 installed. I am running various 2.3.x kernels with various patches, but atd isn't working even if I reboot with an old kernel (2.3.21/devfs/int-patch) where it used to work... (SMP, but that shouldn't affect atd?). Any ideas? at 3.1.8-7 is broken. See http://cgi.debian.org/cgi-bin/bugreport.cgi?bug=49739. I was lucky enough to have the .deb for at 3.1.8-6 still in my apt archive, so I downgraded to that version and it started working again. Frank
Re: what a hack (dselect solution)
Aaron Solochek wrote: Ok, as regular readers probably know, I was having problems with dselect. It was complaing about not finding /usr/share/debconf/confmodule or something similar. There was a /usr/share/debconf/confmodule.sh, so I decided, to cp that file leaving out the .sh, to appease dselect. It seems to have worked, WTF? What did I do? why would this problem have surfaced in the first place. Was it just checking for the existance of that file, and not actually reading or writing it? I thought about symlinking confmodule to confmodule.sh, but instead I upgraded debconf by hand, ie - 'apt-get install debconf'. That installed the proper confmodule script, and 'apt-get dselect-upgrade' finished the job without any more trouble. The problem seems to be packages requiring a more recent version of debconf not having a proper versioned dependancy. Frank
Re: Compact version for running on 486/33 or /66 with 8-32 Mb ram
Robert Parker wrote: While worming through the pages I found a reference to a compact verison of Debian installation that was designed to work on a 486. I have not been able to find the link since, after 3 days of looking. Is there anyone that can direct me to the site where I can get more information on the deployment of DEBIAN on a relatively small 486 system? Or was my image an illusion of too many hours starring at a screen. I have slink (Debian 2.1r3) running on 2 486DX2/66 boxes, one with 32MB RAM and a 2G disk and the other with 12MB RAM and ~700MB of disk. The first is a full-blown internet box serving 3 dialin modems and running web, cache, news and mail servers. The other is mainly a mail server at the moment. I may turn it into a firewall/gateway one day. These are stock standard Debian installations, done initially from a CD then updated over the net. You can install just about everything useful (barring X) and get change from a 200MB disk. That includes everything you need to build a kernel. In my experience the key to getting good performance out of a 486 box is to stuff as much RAM into it as you can manage. The later motherboards that will take 2-4 16MB 72pin SIMMs are the best deal, but older ones with 30pin SIMMs will do, especially if you can score some 4MB SIMMs. I've run Debian succesfully on 486 boxes with 8MB of RAM, console only. I've run X and Netscape on a 486 with as little as 16MB, and it *was* painful at times, but not fatally so. Of course I now have an AMD K6-2/350 with 64MB and I'd never go back :-). I also have slink running on an Amiga 2000/030 with 5MB RAM and 236MB of disk. It does nothing but consume electrons at the moment, but it runs, it's on the net, and it does the sentimental side of me good to see the old workhorse do more than just prop up a monitor. One of these days I'll set up a web site and ftp archive on it. Frank
Re: wine.conf example?
[EMAIL PROTECTED] wrote: can someone kindly give me an example of a working wine.conf? I must be mangling something pretty bad... I have no windows partition anywhere, but I made a windows directory in /var that I point to... What exactly is the problem? The /etc/wine.conf installed by the .deb should work just fine, provided you edit the paths to reflect your local setup. I assume you've read /usr/doc/wine-doc/documentation/no-windows. Frank
Re: Trident T 9750 2 MB AGP Card
[EMAIL PROTECTED] wrote: 'Allo there, has anyone successfully configured the Trident T 9750 2 MB My 9750 claims to have 4MB, but I trust this makes little or no difference. AGP Card under Debian ? I've tried all that I can , and I still cant achieve anything worth looking at...and would appreciate any help/guidance It depends on what version of the SVGA X server you are running. With 3.3.5 (the latest version in unstable) I can only get it to work in 16bpp and with acceleration turned off; this is acceptable for me. With 3.3.4 (the previous version in unstable) it worked in all bit depths, but again only with acceleration turned off. It is reported to work properly with acceleration turn on with 3.3.3.1, but it's unlikely you would be able to find a .deb for that version anymore. I can't say how it works with the older version shipped with Debian 2.1. To turn off acceleration you would add the line Option noaccel to the Device section in /etc/X11/XF86Config. You can also add the same line to the extra options box in XF86Setup (if you can get it to work). Problems with this card have been reported upstream and hopefully we'll see a working server for it before the next stable release. Frank
Re: Trident 3dImage975 and X
Nathan Smith wrote: I have what I hope is a quick question. I'm having trouble getting X to work using the SVGA server with my Trident 3DImage 975 Card. The VGA 16 server works fine, but as you can imagine looks slightly less beautiful than I would like. I'm using the X that came with Slink, although I upgraded to potato and tried the updated X Windows there with similar results. So now I'm back to Slink. I've read the README files and can find nothing there to help me. The quick question is: Has anyone run into this or a similar problem and fixed it. If not, then here's the output: Support for this card seems to have been comprehensively broken in recent X releases (starting with 3.3.4). I've only used it on a potato system so I can't say anything about earlier releases. I currently have it working quite well with the 3.3.5 SVGA server, but only in 16bpp and with acceleration turned off (Option noaccel in Section Device in XF86Config). Frank
Re: wine dependencies
[EMAIL PROTECTED] wrote: Has anyone else noticed that wine depends on libgl1, but that package is no longer available? What happened here? Is libgl1 no longer needed for some reason? Does anyone still have this package because it's not on the debian site anywhere. I found a .deb for this package at http://www.debian.org/~crow/glx/. As far as I can tell it has never been in the archive. Frank -- Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://thingy.apana.org.au/~fjc/scn/
Re: Replacement for Netscape
Kent West wrote: On Wed, 26 May 1999, Ralf G. R. Bergs wrote: On Sun, 23 May 1999 20:51:55 -0500, Matthew W. Roberts wrote: Netscape 4.51 for X is just too slow on my slightly dated pentium. Does Huh??? Netscape 4.51 is everything else but slow. You probably simply have too little memory. I suggest to install at least 128 megs. Ya know, we really oughtta quit advertising the idea that Linux runs well on 486's with low memory and drive resources Why? I'm writing this on a 486DX/33/16Mb (it's normally a 486DX4/120 but that CPU died a couple of days ago). Netscape 4.5 is running on it right now. I wouldn't consider running X on it with less memory, and I'm seriously considering beefing it up to 32Mb, but it seems to me that Linux runs well on it, for my personal subjective definition of well. I try not to run too many memory-hungry tasks at once, and I'm keeping well away from 2.2.x series kernels until they fix whatever dreadful thing was done to memory management. BTW, I run a 3-line micro-ISP on a 486DX2/66/32Mb, and I recently installed Debian 2.1 on my old Amiga 2000/030 just because I could. 5 Bogomips and 5Mb of RAM. Smokin'. -- Home Page: URL:http://thingy.apana.org.au/~fjc/ Not the Scientology Home Page: URL:http://thingy.apana.org.au/~fjc/scn/
