Re: mke2fs capacity (where did it go!?)
On Sun, 14 Dec 1997, matthew tebbens wrote: (victor)[root:~#] mke2fs -c -v /dev/sda1 mke2fs 1.10, 24-Apr-97 for EXT2 FS 0.5b, 95/08/09 Linux ext2 filesystem format Filesystem label= 1101824 inodes, 4401778 blocks 220088 blocks (5.00%) reserved for the super user First data block=1 Block size=1024 (log=0) Fragment size=1024 (log=0) 538 block groups 8192 blocks per group, 8192 fragments per group 2048 inodes per group Superblock backups stored on blocks: etc [...] Here is what 'df' says about the drives: Filesystem 1024-blocks Used Available Capacity Mounted on /dev/hdb1 705433 485054 183942 73% / /dev/sda14253289 509553 3523648 13% /var/sda1 /dev/sdb14253289 13 4033188 0% /var/sdb1 /dev/sdc14253289 13 4033188 0% /var/sdc1 /dev/sdd14253289 13 4033188 0% /var/sdd1 This says I have 4,033,188 to play with plus 220,088 for reserved. Now, if I add those together I get 4,253,276. This is not even close to the stated formatted capacity of 4.51gigs. What am I missing here ?? 4,033,188 + 220,088 + 13 = 4,253,289 so that's all right then. But that's not what's at issue. What's relevant is the 4,401,778 blocks which becomes 4,507,420,672 bytes. That looks better. But even that is only looking at the first partition (sda1, sdb1 etc.) If you examine the partition table, you should find the partition is 2 * 4,401,778 sectors in size, and that there's probably no free space at the end. My own experience is that pre-installed W95 machines on ~ 2GB disks have a single FAT32 partition, but that this doesn't quite fill the disk. The last one I bought was 73heads, 63sectors; 936cylinders of FAT32, 84 unused. Of course it didn't matter as I was trashing it anyway. You can add a few sectors by maximising the partition to use the sectors in its first track. (This isn't the default as DOS can't.) -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: mke2fs capacity (where did it go!?)
On Mon, 15 Dec 1997, matthew tebbens wrote: On Mon, 15 Dec 1997, David Wright wrote: On Sun, 14 Dec 1997, matthew tebbens wrote: (victor)[root:~#] mke2fs -c -v /dev/sda1 mke2fs 1.10, 24-Apr-97 for EXT2 FS 0.5b, 95/08/09 Linux ext2 filesystem format Filesystem label= 1101824 inodes, 4401778 blocks 220088 blocks (5.00%) reserved for the super user First data block=1 Block size=1024 (log=0) Fragment size=1024 (log=0) 538 block groups 8192 blocks per group, 8192 fragments per group 2048 inodes per group Superblock backups stored on blocks: [...] What am I missing here ?? 4,033,188 + 220,088 + 13 = 4,253,289 so that's all right then. But that's not what's at issue. What's relevant is the 4,401,778 blocks which becomes 4,507,420,672 bytes. That looks better. But even that is only looking at the first partition (sda1, sdb1 etc.) But I don't have 4,401,778 blocks, I only have a total of 4,253,289 blocks which becomes aprox 4,355,367,000 bytes. How does it get from 4,401,778 blocks to 4,253,289 blocks ? Somewhere along the line I lost about 250,000 blocks... ?? The partition has 4,401,778 blocks in total. There are 4,253,289 blocks available for your data. Somewhere, the filesystem has to describe where all that data is. Your filesystem has 1,101,824 inodes and, taking a quick look at /usr/src/linux/include/linux/ext2_fs_i.h which describes inodes in memory, there seem to be about 112 bytes in an inode. That adds up to about half your missing space. But there's more to describing the filesystem than just the inodes, and ext2 is optimised for performance, which must mean using more space for chains of descriptors etc.. Perhaps you should read a book on linux internals to find out what you're missing. (Sorry for the pun.) -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Why /floppy -- why not /mnt/fd or something?
On Wed, 14 Jan 1998, Daniel Martin at cush wrote: Peter Prohaska [EMAIL PROTECTED] writes: why `/dev/fd*'s are mounted at /floppy. When reading fsstnd, you would expect them to be mounted in /mnt/someting. Since /floppy is neither one of those un*x shorties nor used during installation I can't see why it lives on the root. The floppy drive not used during installation? Depends on how you install. I use floppy installation but I don't use /floppy. I assume the floppies don't actually get mounted during installation, so you have to type /dev/fd0. But I do like /floppy as I don't normally cd there, but I can type /ftab and that's it. I'm careful to preserve unique first letters for all the frequently used top level directories. In fact, only /bin, /boot, /lib and /var don't work this way (thanks to lost+found and the kernel link). -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: mount: only root can do that??
When /bin/mount is setuid root, everybody can mount the floppy drive. When someone tries to mount something that is not in /etc/fstab, I get this error: mount: only root can do that I didn't expect it because mount is setuid root, but this is okay to me. I do expect that. Mount checks for a line in /etc/fstab and for user in that line, and if both are not found, it issues an error message and refuses to mount. What's wrong with that? See section (iii) of man mount. Just because a program runs setuid root, it doesn't mean the author will let you do anything. Probably just the opposite, in fact. However, when mount is not setuid root, only root can mount filesystems. Any user, whether a member of the floppy group or not, gets this: $ mount /floppy mount: must be superuser to use mount $ Note that the error is a different one. But it is still an error. Only superuser can mount filesystems (see (iii) again). So that error message is correct. Either you must be superuser, or you must be running a program with root privilege. The relevant entry in /etc/fstab reads: /dev/fd0 /floppy msdos defaults,noauto,user,umask=077 0 0 All the /dev/fd0* devices have these permissions: brw-rw 1 root floppy 2, 0 Jan 9 15:05 /dev/fd0 Does anybody know what's going wrong on my computer? I think it worked once for me, but I am not sure how I did it then and how long ago it was. Doesn't look like anything is wrong. Cheers. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: mount: only root can do that??
On Tue, 27 Jan 1998, Remco Blaakmeer wrote: But what I want is this: 1 root can mount anything 2 users in group floppy can mount the floppy drive 3 users not in group floppy can not mount the floppy drive Of course, 1 is always true. But I can't get 2 and 3 at the same time. Either mount is setuid and then all users can mount the floppy, or it is not setuid and then only root can mount the floppy. Now, how can I do that? I think you need a piece of software designed to do that and I think the xvmount package is one. I don't know if the amd package has this functionality or not. David. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Using mgetty automated callback, and the NO CARRIER message.
I have a chatscript to do callback, which looks like this. ABORTBUSY ABORTNO CARRIER - ABORTVOICE ABORTNO DIALTONE ATZ OK ATL1 OK ATDT5551212 ogin:\qmagic-login-id\q RING ata ogin: The problem is that if I leave in the arrowed line, it won't work. When I (or it) type magic-login-id, the other end hangs up so that it can call me back, but my modem says NO CARRIER at this point, which aborts the script. Is there a trick to avoid this. (If not, I guess I'll just have to leave it out.) I thought the \q codes would prevent the secret callback username from appearing in syslog, but it looks like this. Why? Feb 6 17:48:44 here chat[937]: send (ATDT5551212^M) Feb 6 17:48:44 here chat[937]: expect (ogin:) Feb 6 17:48:44 here chat[937]: ^M Feb 6 17:49:03 here chat[937]: ATDT5551212^M^M Feb 6 17:49:03 here chat[937]: CONNECT 33600/ARQ/V34/LAPM/V42BIS^M Feb 6 17:49:04 here chat[937]: ^MDebian Feb 6 17:49:04 here chat[937]: l Feb 6 17:49:04 here chat[937]: ^M Feb 6 17:49:04 here chat[937]: ^M**EMSI_REQA77E^M^Q ^M^M Feb 6 17:49:04 here chat[937]: there!login: -- got it Feb 6 17:49:04 here chat[937]: send (magic-login-id^M) Feb 6 17:49:04 here chat[937]: expect (RING) Feb 6 17:49:07 here chat[937]: magic-login-id^M Feb 6 17:49:07 here chat[937]: ^M Feb 6 17:49:07 here chat[937]: Dialing continues in the background, -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Using mgetty automated callback, and \q
On Fri, 6 Feb 1998, Tim Sailer wrote: David Wright wrote: OK ATDT5551212 ogin:\qmagic-login-id\q RING ata ogin: I thought the \q codes would prevent the secret callback username from appearing in syslog, but it looks like this. Why? Take out the trailing \q.. that turns the echo back on. I see. Yes, I think I need the second \q before the ata to resume logging real strings. However, using just a single \q suppresses the first occurrence but not the second, viz. Feb 6 18:08:30 here chat[960]: there!login: -- got it Feb 6 18:08:30 here chat[960]: send (??) Feb 6 18:08:30 here chat[960]: expect (RING) Feb 6 18:08:30 here chat[960]: magic-login-id^M Feb 6 18:08:30 here chat[960]: ^M -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Viewing bootup message
On Mon, 9 Feb 1998, tsnake wrote: How can you view the entire list of messeges that show when you boot into linux? It all either scrolls by too fast or too much read/decipher as it rolls off the top. Is there a file that mirrors that output or someway I can re-direct it to a file for perusal? Thanks, Chris Although dmesg and /var/log/messages are useful, the best method of seeing *everything* is to use Shift-PageUp/Down. This is always present in VCs just so long as you look before you switch to another VC (which throws away everything above the top of the screen). -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: PPP problems still
On Wed, 11 Feb 1998, Charles wrote: I am probably getting on peoples nerves, but I still dont have my ppp working on my linux box. I have followed the howto's and have tried what advice has been given me. None of this has worked so far. :( I didn't find setting up PPP simple, partly because my circumstances are not quite standard. I wanted to set it up on my employer's internal phones, so all the machines have static IPs, ethernet cards etc., and I'm doing the serving as well as the clients. I also want callback available. So I wrote down everything I did, which is attached (I hope), and may be encoded (I apologise for pine). My modem is external (USR sportster voice 33.6). I would also like to know if anyone knows how to compress an Iomega Zip-Disk in linux, as I would like to utilize my zip's for backing up files. A zip disk is just another disk in linux. I personally use zip (infozip) in preference to tar/gzip but that's partly familiarity. I also leave the zipdisks as DOS so I can restore files to, say, my wife's Windoze laptop. The zipfilenames are 8.3, but that doesn't affect what's inside them. I also have a line in my options file, that calls to a file named ppp-connect in /etc/ppp however, I do not have this file. Could this be my main problem at this point? Could well be. When I enter pon, I've not used pon yet, though I'll probably move my scripts from ~ to /etc when I'm happy with other people (in group dip) using them. If I cannot get this working, then I will be stuck searching for another OS, but would really like to use linux, since I like the options it makes available. Thanks for your patience and any help that can be offered. It's worth persevering. Cheers, -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 Configuring manual PPP on my bo Debian GNU/Linux computers (both ends), using mgetty's callback. It's assumed that all the necessities (PPP, IP forwarding) are built into the kernel and the packages are installed (ppp, mgetty, minicom). Configure mgetty for your modem. /etc/inittab needs an mgetty line on both machines. s1:23:respawn:/sbin/mgetty -n 8 ttyS1 I use -n 2 on a dedicated modem line, 8+ for callback. Effect changes with telinit q to reread inittab or just kill the mgetty process and it'll respawn. /etc/mgetty/login.config needs callback information adding on the other end. magic-userid - - /usr/sbin/callback -S 5551212 (Leave it out on the computer whose phone bill you pay.) Configure minicom; I need: /dev/ttyS1, 115200,~^M~ATZ^M~, ATDT, 90. (Leave ttyS0 for mice.) Check /etc/minicom.users. The dial directory for minicom varies from user to user in ~/.dialdir. That should make callback work ok, with normal login to the other end. Configuring PPP itself. adduser myself dip to allow myself to run pppd, and then logout and back in for it to take effect. chgrp dip /etc/ppp.* to allow myself to run them (though I haven't actually used these scripts yet). Set up a PPP pseudo-user adduser --uid 1234 and password-for-, then set the shell to /bin/false. Create an executable file ~/ppp which types /usr/sbin/pppd /dev/ttyS1 115200 debug defaultroute user where is the agreed username. (You only need this file because you have but a few seconds to type the above.) For authentication of by PAP, add * password-for- (with tabs) to /etc/ppp/pap-secrets at both ends. To make PPP server use , change the line in login.config to /AutoPPP/ - - /usr/sbin/pppd auth -chap +pap login user Edit /etc/ppp/options and make the following extra lines active (these words are already active): (asyncmap 0), auth, (crtscts), (lock), (modem), debug, (proxyarp), (lcp-echo-interval 30), (lcp-echo-failure 4). Making a connection: Down the ethernet card if you want PPP to establish a gateway. minicom return or minicom -d foo return. ^a d and select the entry to dial entry foo. During/after the loud music, type any character. (Paragraph for callback only) When the login prompt is given, type magic-userid return where magic-userid is the magic username appropriate for your location. When your phone starts to ring, type ata return When the login prompt is given, type ^a q and confirm quit, then type ~/ppp return A link should be made, which you can check by typing tail -f /var/log/syslog assuming you're in group adm. You can now telnet etc. To hang-up, type poff return Up the ethernet if necessary and type /etc/init.d/network Automating it and adding chap. Remove -chap from the login.config line mentioned above. Create the following files (+- callback): ~/ /usr/sbin/pppd /dev/ttyS1 115200 defaultroute user \ connect /usr/sbin/chat -v -f ~/.chatscript +chap chap-interval 60 ~/pppc /usr/sbin/pppd /dev/ttyS1 115200 defaultroute user
PAP authentication server names
Thanks to Tim, Joost and Carey for the callback suggestions. I guess the log contains the server's innocent reflection of the login name (which is secret) and to change this would require a change in the way mgetty's callback worked. BTW I don't have HANGUP in my man chat, so I guess I'm slightly behind the current version. A related question: I'm trying to set up PPP with reasonably secure PAP authentication, and I'm using AutoPPP in mgetty. I have worked out how to set user= and password= at both ends with mgetty/login.config, my pppd command and pap-secrets; and the etc/passwd checks it ok. But I don't really understand how the second field of pap-secrets works at all. The file installed by ppp has # INBOUND * os #OUTBOUND os * password and I added * password-for- which works if there's a user with password password-for- on the computer at the other end. I don't really understand what os means in the first line. The documentation implies it ought to be the local name, but I don't knowingly have anything set to os. And anyway, nothing changes if I comment out this line. On the other hand, I've had no success at making a connection if my added line has anything other than * in the second field. I've tried appropriate hostnames and usernames in that position, and I've also tried putting foo there and adding remotename foo to the pppd command line but if it's not * then no PAP is sent from this end. What do other people have in their pap-secrets? (without revealing any real secrets!) -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: network problem
On Tue, 17 Feb 1998, S Lim wrote: I am building my first linux box on a 486, 3Com Card (3c509) on ISA, ethernet. I set the io address to 0x300. In my first installation, the network worked allright. I could ping, telnet, etc. I shut the system down and switch it off. When I switched the system on the next morning, the networking did not work anymore, I could not even see machines on the same subnet. I typed netstat -r, and got this : Destination Gateway Genmask Flags Metric Ref Use Iface 128.250.172.0*255.255.255.0 U 0 0 1 eth0 127.0.0.0*255.0.0.0 U 0 0 1 I0 default 128.250.172.1 0.0.0.0 UG 0 0 2 eth0 The last item (default) appeared after a long pause. I wonder whether the network was working fully when you booted up. It's usually worth trying -n to prevent name resolution in case the nameserver is playing up. I thought the network card might be faulty. I installed win95 and the network worked. I wonder whether it's changed the settings on your 3c509 card. Have you run pnpdsabl and 3c5x9cfg (names from memory) to set/check its configuration? So I reinstalled linux. I had to do it for a number of times before I got it back to work again. When I reinstalled, I repartitioned the harddisk, and reinstalled the driver, kernel and base system. Unfortunately, somebody switched the machine off without shutting it down. I have not been able to get the networking back since then. When the system boots up, I could see the message that eth0 is loaded. The ethernet address displayed is the correct one too. Does anybody know what is wrong? Well, after checking the card, and taking steps to prevent W95 mucking it up in the future if it's still there (cmos settings), why not just reinstall the system from scratch and then not switching it off. However, there's no need to repartition the disk or reinitialise the partitions as they haven't changed, just remount them, so it's quite quick. Cheers, -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Bullshit--Fired for Linux?
On Mon, 16 Feb 1998, Keith Beattie wrote: Yea, I see what you mean but I'm a little concerned that Linux, and to a larger extent free software, will be damaged by people taking a religious attitude towards their software and not tolerating any criticism. I'm not accusing you (or anyone) of that, but I think that there might be presumption that us enthusiasts have that tendency and hence will not be taken seriously. Just to raise a chuckle and let you see how much ignorance is out there, this was posted by someone calling themselves, I believe, a unix sys admin, in the dos Onenet conference (FirstClass conf system) last September : Linux is the kiddie version of UNIX. It is a small subset of real UNIX : (AIX, HPUX,UTX,SOLARIS). It actually runs under DOS. It doesn't : support multi threaded processes, is a 16 bit OS and can be hung quite : easily. Unless you're a teen ager with dreams of being a real : programmer some day I would avoid it. Cheers, -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Linux on top of win95
On Wed, 18 Feb 1998, Sen Nagata wrote: at some point around Tue, 17 Feb 98 20:07:11 -0600 Asher Haig [EMAIL PROTECTED] mentioned: If you set up LILO when you install it, and then tell LILO how to find the Windows partition, you can easily do a dual boot. Superior to doing ^ UMS partitions. ^^ could you be more specific about what you mean here? It's not clear to me what was meant in the original question by running linux on top of dos. If, by this, [EMAIL PROTECTED] means running linux with a filesystem rooted in C:\LINUX under UMSDOS, then AFAIK this is easily installable with Slackware but not Debian. (I've done the former, but wouldn't know where to start the latter.) Mounting UMSDOS partitions is as easy as most things are in Debian; just compile support into the kernel and install the package. So, to answer the original question (I know Linux can be installed on top of a dos partition. Can it be installed on top of Win 95 or NT.), I've not heard of an equivalent to UMSDOS for non-FAT16 filesystem, and that's what's critical. So a qualified yes to W95. Running linux alongside MS os's is an issue that's widely covered in FAQs and HOWTOs, with all the booting ramifications. my experience has been that it really depends on your situation. i'd like to know if i'm missing out on something fundamental. my personal experience has been that being able to use a umsdos partition under both dos/w95 and linux is quite handy given the limited size of some hard disks -- i can easily store stuff to use w/ either operating system w/ this setup. i am using loadlin and a custom config.sys/autoexec.bat setup to do dual-boots, and find this to be quite usable. -sen p.s. for those interested, details of the set up i'm using at: http://www.htp.org/~sen/debian/hu2/install.html I don't see where UMSDOS comes into this. You've got your root on an ext2 partition, but you've kept a dos partition which you probably mount with something like /dev/hda1 /dosc msdos noexec 0 0 in /etc/fstab. Would that be correct? Cheers, -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: removing user from group; vigr -s errors; group policies
On Thu, 19 Feb 1998, Tommi Virtanen wrote: On Thu, Feb 19, 1998 at 01:57:38PM +0100, Oskar Liljeblad wrote: oskar:~# vigr -s vigr: /bin/ae: Bad address vigr: /etc/gshadow is unchanged vigr: /bin/ae: No such file or directory vigr: /etc/gshadow is unchanged - Is this a bug or am I doing something wrong? (vigr without the -s switch works fine.) This looks weird. My hamm version does not seem to do this. What versions of passwd and ae do you have installed? You might want to file a bug report, or send the file from # strace -ff -o /tmp/vigr.trace vigr -s in private mail to me. I think this is #12184. Cheers, -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Question - Hamm, Mgetty, PPP, wtmp - Arrgh!
On Thu, 19 Feb 1998, Jeff Noxon wrote: On Wed, Feb 18, 1998 at 04:33:58AM +, [EMAIL PROTECTED] wrote: On Mon, Feb 16, 1998 at 01:06:59PM -0600, Jeff Noxon wrote: I have three dial-in modems on my hamm box. They are all identical USR Couriers AFIAK, and their configuration *appears* to be the same. But one port always shows connect speed in wtmp, and the others do not: The modem settings are identical. The modem firmware revisions are identical. The mgetty configuration lines are the same. The inittab lines are the same. This is really bizarre. It used to work with bo, and one of my mgetty upgrades killed it a long time back. I asked back then, and nobody offered to help. So when you shuffle the modems around on the ports, does the same port show the speed, or the same modem? At least that should eliminate the possibility of differing modems. Cheers, -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: help me understand timezones
While the world no longer sets its clocks to GMT but UTC, GMT is still a local timezone in the sense that EST is. Only its universality is obsolete. On Wed, 18 Feb 1998, Bill Leach wrote: You were told correctly... GMT (Greenwich Mean Time) is the same thing as UTC (Universal Coordinated Time--UTC is actually the acronym as it is in French). So the term GMT is obsolete but the meaning is unchanged. A related timezone question: What is the difference between GMT and UTC ? I was told that GMT is obsolete, and that UTC is now considered as its successor. Cheers, -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
RE: Frozen Mouse
On Thu, 19 Feb 1998, Kristian Strickland wrote: Are you running gpm? When I first installed debian, I installed gpm because it's amazingly handy. However I soon found that gpm and X on my machine didn't mesh. Solution? Since I only use the console for upgrades, I removed gpm. If you plan to go back and forth between console and X, you might just kill it before starting X. Why not just add -R to the gpm configuration and use /dev/gpmdata and MouseSystems in XF86Config; then you can have both. Cheers, -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Kermit: Where is it???
On Fri, 20 Feb 1998, Jesus Duran wrote: i was looking for Kermit, remember that old thing? I did not find it the Debian Packages site. I did down load it from www.columbia.edu/kermit/ in a tar'ed-gz'ed file but had problems when i tried to tar -x it?!?!? Any idea where i can find it in a regular .deb file Any help is greatly appreciated. The kermit protocol is built into minicom (along with *modem). Cheers, -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Kermit: Where is it???
On Fri, 20 Feb 1998, Thomas J. Malloy wrote: On Fri, 20 Feb 1998, David Wright wrote: The kermit protocol is built into minicom (along with *modem). The kermit protocol on most communications programs should not be confused with the true kermit program from columbia university. True kermit is a very fast communication protocol/program with many advanced features. I don't know what you mean by true kermit. There is a kermit protocol and the kermit programs use it. Kermit programs also have terminal emulations, user interfaces, scripting languages, and communications drivers where necessary. You can use these parts of kermit programs in the same way as, for example, minicom and never touch the kermit protocol if you wish, including error-uncorrected file transfer. The kermit protocol is open and anyone, or any program, can use it. (Yes, I've even used it myself, typing raw kermit packets into a vax for testing.) Various parts of the protocol are optional, and the protocol defines how to negotiate with another machine which capabilities are recognised and supported by both parties. By contrast the kermit transfer protocol in most programs is very slow because they only allow a maximum packet size of 93, and no sliding windows. That usually means that those programs are old, or that the author couldn't be bothered to support more capabilities, or wouldn't fork out for a book describing the protocol. Of course, kermit suffers from its own good design: because the dumbest implementation can interoperate with the most modern, the former is never forced to upgrade in order to accomodate the latter (unlike most software). You may need to upgrade for other reasons of course, e.g. not supporting modern line speeds. So perhaps by true kermit you just mean one with, say, long or extra long packets and sliding windows. Cheers, -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Frozen Mouse
On Fri, 20 Feb 1998 [EMAIL PROTECTED] wrote: Why not just add -R to the gpm configuration and use /dev/gpmdata and MouseSystems in XF86Config; then you can have both. Doesn't the gpm man page says that what you are suggesting doesn't apply to serial mice ? If that's the way you understand it, then I think you should submit a bug report against the man page. Here's the paragraph: With X11 gpm has the same problems as selection. If your mouse is a single-open device (i.e. a bus mouse), you should kill gpm before strating X, or use the ``-R'' option (see above). To kill gpm just invoke gpm -k. This doesn't apply to serial mice. To what does This refer? I think it means forget this entire paragraph if you have a serial mouse, not just the previous sentence. But it's ambiguous. In any case, all my bo systems use gpm -R, and /dev/gpmdata with MouseSystems in X. Some are pentiums with PS/2 mice, some are 486s with serial mice. BTW, hands off GMT please. We need it here, at least during the wintertime. Cheers, -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Debian Linux and Non-Free Packages.
On Mon, 23 Feb 1998, Graham Lillico +44 1785 248131 wrote: Could someone please explain the concept of Non-Free Packages please (i.e. Pine etc). Take a look at /usr/doc/debian/FAQ/debian-faq.txt.gz in section 5.6.1 where this is explained. It is just that I don't understand why other distributions can include them in there official releases, where as debain can not. Perhaps this is explained by reading their respective release notes carefully. You can't legally publish a piece of software under a less restrictive licence than that under which you obtained it. Cheers, -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: A simple problem with dselect ...
On Wed, 25 Feb 1998, Nebu John Mathai wrote: I tried to remove a package and under Select... selected to remove the package. Then I went to Remove... and dselect removed almost every package I had installed since the beginning (including the one I had asked it for). I know I'm doing something stupid ... I just don't know what it is. Is it possible that you accidently pressed - or _ when one of the top few lines was highlighted: All Packages, or Up To Date Installed Packages etc. This should have a pretty dramatic effect! Cheers, -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Server Questions (fwd)
On Thu, 26 Feb 1998, Nathan E Norman wrote: No-one on this list uses Debian as a news server? Nobody's running SMP? There's no interest whatsoever in tech support for Debian? (I know Bruce posted something about it a while ago ...) I'm disappointed. If the folks who are in the know have all unsubscribed from this list and hang exclusively on debian-devel, then I think there's a problem. Judging from the traffic on this list, they seem to be as helpful as ever. But... -- Forwarded message -- Date: Wed, 25 Feb 1998 02:14:22 -0600 (CST) From: Nathan E Norman [EMAIL PROTECTED] To: Debian User List debian-user@lists.debian.org Subject: Server Questions [mildly offtopic] We are looking at purchasing a new server to add to our stable, and I'm looking for some feedback/opinions/experience with news and/or SMP systems with Debian Linux. Private email is welcomed, since this isn't really list material (sorry) My way of getting this sort of information is to watch the list for topics I might be interested in in the future and file them in folders. For example, I built quite an IP/PPP resource before I needed to get PPP going myself. Ditto ssh and samba. Now I'm watching people's queries on hamm ready for the changeover. What I'm looking for is this: If you're running SMP, does it work well? Problems? Hints? Same with news ... does a specific version of INN do the trick? Any special kernel configs? You get the idea. The specific stuff should be in HOWTOs. But most people don't have a general off-pat report ready for a quick response to an open query like this. Human nature says Leave it to someone else/Somebody's probably already sent an email. Oh, and before I forget ... I heard there is a company providing phone support for Debian Linux. If this is true, who is it? What are the rates? I work for a corporation that is becoming comfortable with the idea of Linux, but wants tech support available. I'd rather run Debian than RedHat/Caldera, so please help! Apart from Bruce's posting which you mentioned, I can't help you here. Universities aren't the sort of institutions who want to pay for support of linux. Cheers, -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: mtools: mmove usage
On Wed, 4 Mar 1998, Ulf Jaenicke-Roessler wrote: how am I supposed to use mmove? I think it should be called just like mcopy. Though the syntax may look superficially similar, that's about the only connection between mmove and mcopy. However, while 'mcopy a: /dos/e' works, Here, /dos/e is on a unix filesystem. About a:, the documentation says fer. A missing drive designation implies a Unix file whose path starts in the current directory. If a source drive letter is specified with no attached file name (e.g. mcopy a: .), all files are copied from that drive. 'mmove a: /dos/e' doesn't. For a start, there's no source file specified. And if it were to take that to mean a:*.* (it doesn't), it would necessarily loop: \ goes to \dos\e, therefore \dos\e goes to \dos\e\dos\e ad infinitum. Remember, \dos\e is on a: and not on unix. The man page and the info file didn't help. Any hints? Or is it a bug? Perhaps you read the man page thinking that move a b means copy a b and then delete a. It doesn't. Cheers, -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Kernel Install
On Tue, 3 Mar 1998, Scott Walker wrote: Ok, this is a BIG nasty weird problem. I'm runnig Debian With default Kernel 2.0.29 So we're in bo, I take it. I grabbed Kernel 2.0.33 from ftp.kernel.org and did the following steps: tar zpfxv linux_2.0.33.tar.gz cd linux make mrproper make menuconfig make dep make clean make zlilo make modules make modules_install reboot Now, on ANY other machine.. doing a uname -a or checking /proc/version I'd have Linux 2.0.33 blah lah lah... But not on this box... it's still saying 2.0.29. Now I left the machine running for an hour.. and still.. 2.0.29.. so I rebooted.. still 2.0.29.. I copied the kernel I compiled to a disk and boot the disk in another machine.. kernel 2.0.33. Now I check the 'screwed' machine by attacking it with land/teardrop to see if it was just a version screw up and it was really running 2.0.33 just not reporting it.. and it bombed. so it's still running 2.0.29. does anyone hve ANY ideas about this? I have recompiled, reinstalled, reunpacked and redownloaded the kernel multiple times and it still will nt update. Well it pretty obviously put the compiled kernel somewhere where something couldn't find it, so it just carried on using the old one. Presumably it consistently did the same thing multiple times, which is a relief. I notice you're using a linux tarball and not the Debian kernel-source package, so there may be some sort of tweak missing. For example, the linux/Makefile says: # INSTALL_PATH specifies where to place the updated kernel and system map # images. Uncomment if you want to place them anywhere other than root. #INSTALL_PATH=/boot Now I compile my own kernels from kernel-source packages and I don't touch that line. However, make install always puts the new kernel and map in /boot (and moves symlinks in root), so something must be Debianised. You could try changing make zlilo to make zImage and make install and see where the kernel image ends up (remember the System.map too). Alternatively, you could get 33 in a debian package from unstable - I've compiled it for the joystick on a bo system without any problems. Cheers, -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Pine From: address
On Wed, 11 Mar 1998, Jonas Bofjall wrote: You have to recompile perl and enable the ALLOW_CHANGING_FROM #define. pine (typo) Its all in the docs. Download the source from www.washington.edu/pine. Or, better, ftp://Debian mirror/debian/non-free/source/pine_* It'd be great if Debian's pine was compiled with this. It is. You'll see the change about 30% of the way down pine_3.96L-2.diff in the Debian source files. Unfortunately, you'll also find -DDEBUG deleted at about the 40% mark. It's worth recompiling just to restore that feature. You still have to add 'From' to the user-changeable fields in pine's setup, so a 'From' dialog does not show up unless you want to. IMHO its safe enough. Cheers, -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- E-mail the word unsubscribe to [EMAIL PROTECTED] TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to [EMAIL PROTECTED] .
Re: Increasing DOS partition
On Sat, 6 Sep 1997, Bill Wohler wrote: I just performed some fall cleaning by running a fresh install (like the one-disk install, thanks!). All went relatively smoothly except on the DOS side. I totally repartitioned the disks although I left /dev/sda1 as the DOS partition, only bigger. cfdisk in Linux showed the increased partition size, fdisk on a DOS boot floppy was also happy, but a dir in DOS showed the old size, and a format /s also formatted based upon the old size. One other hint: when I mount the partition from Linux, it also shows the old size. Any tips as to how I can format (or whatever) the DOS partition so that the OSs recognize the entire partition? And hopefully do this without destroying all my work on the other partitions. Yes, the problem is well-known and documented. You need to read the cfdisk man page very carefully and backup the partition before you do the dd. I've not had to do this myself as FIPS was good enough to get this right. Good luck. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: 1.2-1.3.1 upgrade and file-rc
On Thu, 11 Sep 1997, Waller Martin MEJ wrote: I've upgraded from 1.2 to 1.3.1, and during the process file-rc was installed... Well, I guess you installed it... Well, I'm having problems with it. First, on boot up xdm wouldn't start. Looking in /etc/runlevel.conf, xdm was the last line of the file and below was a comment saying: #THE LAST LINE IS NEVER READ! ??What's that meant to mean? I think I've sometimes had a problem with the last line of a file being ignored when it doesn't have LF at the end. This may be a trick to avoid ever having that problem. [...] Anyway, i looked at the /usr/lib/file-rc/rcfile2link.sh, and it looked like it was supposed to put back the rc.0-6 directories. So i ran it. But the rc.0-6 directories weren't put back... [...] So it's still not working properly - I can't understand the docs in /usr/doc/file-rc/ and really would like to get everything functioning properly rather than just working-but-not-quite-right. Perhaps it's worth uninstalling it and then reinstalling it. Often this works better than tinkering. And another odd thing - at first after the upgrade i couldn't get X to run at all as it complained /dev/psmouse didn't exist (but it was there...) and someone told me that if my psaux was a module it needed to be loaded before /dev/psmouse could be found (dunno what that meant really, but i know what modules are), so i incommented #auto in /etc/modules to get kerneld running and voila it worked. But i never had to do this on my 1.2 setup. What changed??? Well, you don't say how you upgraded, but if you booted off an installation disk, then you have a new kernel for a start, and you may also have configured different modules in that part of the installation. I find installation kernels a bit wobbly myself. Obviously they have to have everything and the kitchen sink thrown in so that they're as useful as possible to everyone out there. But many people seem to run their production systems with one, and they have to change from release to release as they support yet more new hardware (NCR 53c875 in my case, thank you). Back in 1.1, you had to compile psaux yourself (one of the most FAQs). Thanks for any help and advice, I hope it helps; good luck. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: dpkg
On Sat, 13 Sep 1997, Jim Pick wrote: I try to install X11 from cdrom that is download and recorded from ftp. When I use dselect to install some packages, some .deb packages are refused by dpkg. Even I use dpkg -i xlib* to install X11, error message like: dpkg-deb: /...directory/xlib6*.. not dbian format Does anybody have same problem before? Why the .deb that is download can not be accepted? Ok, I'll bite (since you've asked three times). Normally, questions on debian-user get answered more quickly than - I thought somebody else would have taken this one by now. Same here...but it was the weekend. Most likely, you have a file named xlib.. in your directory that is not a .deb file. Perhaps you also downloaded a xlib*.tar.gz file (an upstream source package). Try running 'dpkg -i xlib*.deb' instead. Or even better, use the full explicit file name. (Of course, there is always the possibility that you have found a bug) It's also possible that the file has got corrupted. I hope the person who cut the CD remembered to ftp the files in binary mode. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
RE: X connections refused between hosts (foo/unix:0 syntax)
On Thu, 18 Sep 1997, George Bonser wrote: On 18-Sep-97 David Wright wrote: I've obviously missed something in the explanations of .Xauthority files and MIT cookies. I have two Debian 1.3 machines, foo and bar with essentially identical configurations, with few changes from how things install themselves. I start X as user myself on foo. Typing xauth list says foo/unix:0 MIT...1 a0b1... and so does xauth list :0 and xauth list unix:0 and xauth foo/unix:0 but not xauth foo:0 which says nothing. Assuming foo is your local machine, what happens if you do xhost +bar then telnet to bar, export DISPLAY=foo:0.0 and then run an X program on bar? Host-based access works fine, but I wanted to avoid that because the X display should not be accessible to some users of foo, let alone those on bar. Looking at the books on this subject, e.g. Lui and Pearce page 79, Garfinkel and Spafford page 527, I see examples like this: foo% xauth extract - $DISPLAY | rsh bar xauth merge - All this does is to stick the foo/unix:0 MIT...1 a0b1... line into .Xauthority on bar. What I think I need is a command which massages foo/unix:0 MIT...1 a0b1... into foo:0 MIT...1 a0b1... so I can merge that into .Xauthority on bar. Is that what I should be doing? Is that what everyone else does? I can't help thinking I've missed something if none of the books/documentation mentions this wrinkle. Copying .Xauthority from a user's home directory to /root so that you can start clients after suing to root must be a FAQ - it's even been asked on this list in the past week. Surely some of these people must hit my problem as soon as they run remote clients, or I've done something wrong? -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
X connections refused between hosts (foo/unix:0 syntax)
I've obviously missed something in the explanations of .Xauthority files and MIT cookies. I have two Debian 1.3 machines, foo and bar with essentially identical configurations, with few changes from how things install themselves. I start X as user myself on foo. Typing xauth list says foo/unix:0 MIT...1 a0b1... and so does xauth list :0 and xauth list unix:0 and xauth foo/unix:0 but not xauth foo:0 which says nothing. If I su to root and copy ~myself/.Xauthority to ~root I can run xeyes with xeyes -d :0 etc., but not xeyes -d foo:0 which is refused. Now I telnet to bar and login as myself, ftp back to foo and copy .Xauthority to bar. On bar, xauth list and xauth list foo/unix:0 both say foo/unix:0 MIT...1 a0b1... but the other forms say nothing. Now I type xeyes -d :0 etc. and :0, unix:0 and foo/unix:0 are all obviously trying to display locally on bar, as the error is 111. xeyes -d foo:0 is refused by the server on foo (an audit message appears on the VC which started X). I presume I have to massage .Xauthority in some way, to create entries for :0, unix:0 and foo:0, but why does .Xauthority have foo/unix:0 put into it by startx? I don't really understand what the syntax foo/unix means and can't find any documentation for it (though I've seen it in examples, e.g. page 82 of Mui and Pearce). -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: X connections refused between hosts (foo/unix:0 syntax)
On Thu, 18 Sep 1997, Jens B. Jorgensen wrote: David Wright wrote: I've obviously missed something in the explanations of .Xauthority files and MIT cookies. I have two Debian 1.3 machines, foo and bar with essentially identical configurations, with few changes from how things install themselves. I start X as user myself on foo. Typing xauth list says foo/unix:0 MIT...1 a0b1... and so does xauth list :0 and xauth list unix:0 and xauth foo/unix:0 but not xauth foo:0 which says nothing. [...] Now I telnet to bar and login as myself, ftp back to foo and copy .Xauthority to bar. On bar, xauth list and xauth list foo/unix:0 both say foo/unix:0 MIT...1 a0b1... but the other forms say nothing. What's the timestamp on ~/.Xauthority? I didn't know that startx created auth files. Authorization in the server is controled by passing the -auth file param to the X server. Looking at ps output, I could see the -auth option to xinit, so I looked at /usr/X11R6/bin/startx and the problem is the last three lines. Yes, startx does create xauth files, but only with a unix entry. I can't find a way to coerce $display into making a TCP entry instead so I've modified startx to do this: serverargs=$serverargs -auth $HOME/.Xauthority mcookie=`mcookie` xauth add $display . $mcookie xauth add `hostname`.`dnsdomainname`$display . $mcookie xinit $clientargs -- $serverargs I don't think I should have to modify anything in /usr, even though startx says ironically: # # Site administrators are STRONGLY urged to write nicer versions. # Who'd think of reading the files in a bin/ directory? I think I'll file a bug report against xbase. #!/bin/bash Thanks for the script. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Infomagic LDR debian, and, What's Hamm and Bo?
On Mon, 22 Sep 1997, Robert Grunloh wrote: But before I can do this, I have to ask: I've read the Debian FAQ on the web page, the part that explains the directory tree at the ftp sites; but what it should explain but doesn't is, what is hamm and bo? I see mention of these terms in announcements of new packages and on the lists but nowhere is there a definition, or historical explanation; leaving me to conclude that they must refer to hidden directories of proprietary commercial code without source, or something ;-) Bo is the current stable release; hamm is unstable. The names are just codenames, which never change. This prevents mirror sites from downloading an entire distribution when unstable - frozen - stable. The latter are links; the codenames stay the same. For some reason, some people object to the user of these codenames (don't ask me why; every project in real life has a codename) so I think they've dropped out of the FAQ (they would be in 5.7). -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: make-kpkg error....
On 23 Sep 1997, Manoj Srivastava wrote: Well, there is no difference between arch//boot/bzImage and arch/boot/bzImage, really, when it come to cp. I suspect that the image did not build, for some reason. Perhaps you've overlooked the fact that, though the filename syntax might be legal, the architecture is missing; i.e. .../alpha/..., .../i386/... etc. I have no reasons to offer, however; sorry. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: serial port speed
On Wed, 24 Sep 1997, Dave Cinege wrote: On Wed, 24 Sep 1997 01:23:03 +1000, Lawrence wrote: Anyone knows the default serail port speed? It is 38,400bps? Which file responsible for this setting? I want to increase it to 115,200bps. 38.4K, yes, /etc/rc.boot/0setserial, use the spd_vhi option. Perhaps there's some history here. I installed Debian 1.3 on a 1997 pentium and setserial -a /dev/ttyS? all say that baud_base is 115200 and Flags: spd_normal... Both mgetty and minicom will satisfactorily handle 115200, so all this messing with spd_vhi seems to be a thing of the past. Presumably, by now, any software that can't ask for 38400 should have a bug report filed against it. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: X problem after update from REX to BO (detailed error messages now)
On 25 Sep 1997, Andy Spiegl wrote: According to Remco Blaakmeer [EMAIL PROTECTED]: I took a look at the output of the X server in an earlier post and it already uses xfs for the fonts. But you have to configure xfs so that it looks in the right place for the fonts. Take a look at the 'catalogue' line in /etc/X11/xfs/config. You have to set this to exactly the set of font dirs that you have. Xfs is very unforgiving if you specify an unexisting or empty directory (dumps core on my computer). I encountered the same problem when I wanted to start using xfs for the fonts and found that I had to remove the '100dpi' directory from the 'catalogue' line. It now works fine for me. I experienced exactly the same, when I first updated the fonts, but not the Xserver. I had to play around with the Catalogue line until only the valid paths were listed. Then I updated to XFree86 3.3 and could restore the previous Catalogue line. So, I think xfs should be configured fine. Is there any way to make xfs list the fonts it can serve? In order to check the configuration? Look at fsinfo and fslsfonts (and anything else starting with fs). (BTW I /did/ think you were running a different Xserver in my private mail - perhaps something I misread in an earlier post.) -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: IEEE POSIX Standards and Linux
are a member, they are still cheap compared to the ISO/ITU-T prices... Thanks, David -- David Frey [EMAIL PROTECTED] |Microsoft isn't the answer...it's the QUESTION. Schlieren, Switzerland |``No'' is the answer. 51F35923114FC8647D05FF173C61EFDE|Use Debian GNU/Linux! --8 -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
MC and .deb files (was Re: extract only PART of an archive)
On Sat, 27 Sep 1997, Dale Scheetz wrote: I use Midnight Commander for this. It has a feature that lets you dive into a .deb file (it may require a patch to mc.ext, see below) and view the contents. You can also copy any file you see out to another directory using Midnight Commander. As I said, some of the latest version of mc don't have the following patch applied. It is simply a matter of checking /etc/mc/mc.ext to see if it contains the following: [...] +# deb +regex/\.deb$ + Open=%cd deb:%d/%p/ + View=%view{ascii} dpkg-deb -c %f + I think the problem was that the patch was applied (it's right at the start of the file), but the output was left as /etc/mc.ext and not /etc/mc/mc.ext (this is version 3.5.17-1). So it's enough to copy the former to the latter. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: No Debian updates?
On Mon, 29 Sep 1997, joost witteveen wrote: Hi, Im using Debian Linux for about a year now. Is it true that there were no updates in the stable non-free contrib files for 3 (in words: three) months now? (I am using ftp.debian.org as ftp-server). Well, stable is called stable for one reason! But I think there have been updates, only as little as possible. We don't go about adding packages/updates to stable just becase a new version was released, we only do it to fix serious bugs. (Recently the samba bug for example, so the new samba package apparently already is in updates (or whatever the direectory is called)). Yes, samba appeared Sep 28 on my mirror. But there are quite old things in bo-updates, too. Are these going to be finally included in bo? While they are only in bo-updates, they are inaccessible to dselect users. The directory ../bo-updates contains potential fixes to 1.3, files that may become part of the next version of 1.3. They are placed there so that they may be well tested before final inclusion. (from the READMEs in bo/bo-updates.) dselect does not find any new file for Debian 1.3 Or is there something wrong with my linux... Maybe you didn't have any of packages installed that were updated (look at the changelog to see what changed). Or, more likely, it's because all that's happened in the last three months is new boot disks and the removal of five packages. If you want new packages, go to unstable (and check out the libc6 update FAQ frequently posted to debian-user). Note that problems with unstable should ideally be posted to debian-devel (an open list). That's not really an option for those of us running production systems. (But I do take on board that the developers are busy with libc6.) -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: X And Video
On Thu, 25 Sep 1997, Syd Alsobrook wrote: I am having an interesting problem now. I finally got my new machine to boot up and I reset my X configuration for my new video card. X works great but when I exit back to a stand tty the video is trashed (misc. characters, garbage, and general bad mood) reset does not work and all virtual tty are trashed. The video board is a STB Systems w/ an advanced logic chipset alg2301 1024 ram. STB's website was rather useless mostly windows crap the only thing I saw that might be of interest was the need to have a memory exclusion, but I don't know how to do that under linux. I have a 486 with an Avance card that uses this chip, and it had the same problem. The garbage that you see is actually a translation, something like abcdefgh - bcdefg0h, so you can actually read it. The workaround I used was to leave at least one VC logged in and type setfont default8x16 whenever I needed a VC, and the same when quitting X. I never found any documentation for memory exclusion, but perhaps it's in the X sources. I now use the machine as a laboratory winlose playground. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: mke2fs
On Mon, 29 Sep 1997, Philippe Troin wrote: On Mon, 29 Sep 1997 23:32:21 PDT David Stern ([EMAIL PROTECTED]) wrote: I'm having an irregular experience with mke2fs. I'm attempting to format /dev/sda16 and message says: debian# mke2fs -v /dev/sda16 mke2fs 1.10, 24-Apr-97 for EXT2 FS 0.5b, 95/08/09 /dev/sda16 is entire device, not just one partition! Proceed anyway? (y,n) But /dev/sda16 is just a 500MB partition, not an entire device: [...] I tried changing the beginning and ending cylinders, to no avail. Why does mke2fs think /dev/sda16 is the entire drive? Because /dev/sda16 has major 8, minor 16, which is the major/minor pair for /dev/sdb: look at 'ls -l /dev/sda16 /dev/sdb'. BTW, you have created sda16 yourself didn't you :-) There is an infortunate 15 partition limit on PCs. I think anyone who installed Debian 1.3 with the 1997-05-30 installation floppies will have exactly the same devices. All mine do. Now I feel guilty for complaining that Debian 1.2 only had sdx1 through 8 (which bit me badly). [BTW I think you posted that Because /dev ... didn't you :-) all on one line.] -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: mke2fs
On Tue, 30 Sep 1997, David Stern wrote: On Mon, 29 Sep 1997, Philippe Troin wrote: On Mon, 29 Sep 1997 23:32:21 PDT David Stern ([EMAIL PROTECTED]) wrote: I tried changing the beginning and ending cylinders, to no avail. Why does mke2fs think /dev/sda16 is the entire drive? Because /dev/sda16 has major 8, minor 16, which is the major/minor pair for /dev/sdb: look at 'ls -l /dev/sda16 /dev/sdb'. BTW, you have created sda16 yourself didn't you :-) I created it during the install procedure manually, yes. You created the device (with mknod), or created the partition? Do you have (as I do) /dev/sda16 = /dev/sdb /dev/sdb16 = /dev/sdc /dev/sdc16 = /dev/sdd /dev/sdd16 = /dev/sde /dev/sde16 = /dev/sdf /dev/sdf16 = /dev/sdg /dev/sdg16 = /dev/sdh -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Can't mount root device
On Thu, 2 Oct 1997, Russ Cook wrote: Hello! I have recently gotten a base system installed (Debian 1.3.1 from 2-cd set). I am now trying to compile a custom kernel to get sound support and IP Masquerading. The kernel package I'm trying to compile is 2.0.30. I run Dos and OS/2 on machine as well, and I use the OS/2 boot manager. Linux and OS/2 are on my SECOND scsi drive, and I have followed the How-To to allow the OS/2 boot manager to invoke LILO, which then loads my Linux kernel. All works fine with my base installation. My problem is with the custom kernel. I use make config, make dep, make clean, then I compile, then I use make lilo_install (or something like that, I'm not at my home machine right now). When I try to reboot, I get a kernel panic, with a message that VFS can't mount the boot device 8:16. When make lilo-install completes, I get a message about the boot device being 8:22. I can't provide more detail (or accuracy), as I'm at work and the machine in question is at home. Any constructive comments/suggestions are most welcome. I will try to get more detail from the machine tonight and repost, if necessary. Well this may be completely off the wall, but there's been a discussion here concerning the fact that /dev/sda16 and /dev/sdb both have the same major and minor numbers. They were having problems with the sixteenth partition (!) on scsi disk one. Perhaps your problem is related? Get rid of /dev/sda16 perhaps? -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Time Zone mount points
On Fri, 3 Oct 1997, Remco Blaakmeer wrote: On Fri, 3 Oct 1997, Nelson, R.A (Richard/Rick) wrote: *) I've got /bin mounted seperately as r/o. Now, I'd like to move /sbin there as well. I assume that what I need to do is to mount a new directory (say /mnt/robins) and have /bin,/sbin directories therein and symlinked in root? Please do not do this. /bin and /sbin are required to be on the root file system. They contain binaries you need to fix things when you have troubles mounting file systems other than the root file system. The same goes for /lib and /etc. Agreed. /etc gets written to anyway. What puzzles me is that you have /bin on a seperate mount point and you are still able to mount it. On my system, the only 'mount' binary is in /bin. Conceivably, mount could be in a /bin in the root directory which gets occluded (or whatever the correct un*x term is) by being mounted over. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: RARP on MAC address?
On Wed, 8 Oct 1997 [EMAIL PROTECTED] wrote: Yes, I did know this method.. but, this doesn't help at all if the address isn't in your arp tables. I'm looking for a method of finding it if it ISN'T in the arp tables which is currently my problem. Proof by exhaustion: ping -c 1 nnn.nnn.nnn.xxx for all xxx and it'll appear in your arp table. (If you know vaguely where it is.) -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Automatic root login after boot
On Thu, 9 Oct 1997, Alastair Gregory wrote: I have an odd problem with my Debian system. When it boots up into multi-user mode, it automatically logs in as 'root' on the first console. It's impossible to terminate the session, as a new one starts immediately after I exit the shell. It's in run level 2 and the other consoles behave normally. Look at the line beginning 1: in /etc/inittab. Another problem I have is that the machine won't reboot by itself. If I telinit 6, it shuts down, then sits there without rebooting until I hit the reset button. Is this an AMD quirk? I had a Cyrix machine that would sometimes do this (and on power-on, it would just sit there without POSTing). I loaded Debian 1.3 on an old AMD 486-80 based machine. It seemed to load properly, except that a number of utilities were not present (such as 'dselect', mesg, start-stop-daemon). I was a bit dubious, but not having loaded Debian for a while (last time it was at 1.0, I think) I just assumed I wasn't up with the latest developments. I scrounged a dselect binary which ran, but no access methods were available so I was no better off. The autologin phenomenon didn't manifest itself until I tried reinstalling the base system from the boot diskettes. On booting from the hard disk after this, it asked me to pick a root password, but after the passwd dialog, it'd say Try again and run passwd again. I flipped to another console, logged in as a mere user, su'ed and killed the passwd and bash processes, and ran passwd manually to set the root password. Now I have the autologin problem but dselect seems to be OK. My guess is that part(s) of the Debian load process didn't complete, or failed silently. Can anyone shed light on this? Could I somehow have bad install disks and not know it? Deian shuffles inittab.real (if I remember the filename correctly) and inittab when it completes correctly. It was months ago, and I had had to reboot while in that first session after setting the root password, so the shuffling didn't get done. Once I fixed my problem (I think it was when perl had problems, maybe a bad link) and edited inittab, the system was fine. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Changing mouse drivers
On Tue, 14 Oct 1997, Daniel Martin wrote: On Tue, 14 Oct 1997, Richard A. Guay wrote: I now want to change to using a serial mouse. What would be the steps (I do not want to reload the OS.)? Reload the OS? That's windows-world thinking. First, if you installed the gpm package, you'll need to run 'gpmconfig' as root, and tell gpm about your mouse. Then, assuming you have X, you'll have to re-run XF86Setup (stop X first by quitting your session and, if you have xdm running, doing: /etc/init.d/xdm stop ) Inside XF86Setup, say yes to use the old configuration and all you should have to do is pick your new mouse parameters. (be certain to read through all of the dialog boxes that come up there, and test the new mouse setting by applying the changes) You shouldn't even need to reboot! If you're using the -R option in gpmconfig with /dev/gpmdata in XF86Config, you don't need to make any changes to X. It also seems to avoid the problems that some people have reported with their mice when switching between X and VCs. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Preventing single user mode
On Tue, 21 Oct 1997, Timm Gleason wrote: Does anyone out there know of a way to prevent a Debian box from being able to boot into single user mode? We have removed any sort of delay= settings from the lilo.conf, and this makes it extremely difficult to get into that mode, but does not prevent it. Any help? How do you break into a perfectly secured area when the key just sheared off in the lock? How do you rescue a Debian box that can't boot into single-user mode? But to answer the question, I guess you'd need to (a) secure the floppy drive through the CMOS and (b) hack lilo (the source is there). But even if you somehow remove the jumper pin that clears the CMOS password, you can clear the CMOS entirely by temporarily removing its power source, so I guess you have to cut the tracks to the floppy cable. How far do you want to go? Timm Gleason Hardware Engineer Hey, you should find this easy! Just out of interest though, how straightforward is it to reset the CMOS password on laptops that say this is a factory operation? (i.e. is this just a con?) -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: hard drive crash
On Mon, 3 Nov 1997, Jessica Wolfe wrote: I need to write a report on hard drive crashes. First I need to figure out all the possible causes of hard disk failure. Can you help me out? [...] Student North Idaho College Major: Computer Applications In Business Coeur d'Alene, Idaho Doesn't your college have a library where you can research these things? -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Maintaining a Debian mirror via floppy net
On 18 Nov 1997, Falk Hueffner wrote: I have a PC at home running Debian with no connection to the net. So do I. When I install packages, I copy the debian/ tree to a jaz disk in another VC when it asks shall I delete the installed files, and then I switch back and answer yes, and delete the files and the debian/ directory. Then I use dpkg -i at home to install the .deb files. To do this with zip disks, you'd just need to work out how to split the stable tree, or use one disk when you install the default system, and another for everything else. Keeping all your .deb files makes it easy to uninstall and reinstall without hitting any snags like updated packages at ftp.debian.org that could cause dependency problems. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
X pointers (was Re: Low speed response w/ Logitech Trackball)
On Sun, 23 Nov 1997, Oliver Elphick wrote: Pancho Horrillo wrote: I use a Logitech Trackball (Marble), attached to PS/2 port, and i cannot get it work fast enough in X, so my thumb gets really tired... Is there any way to change the responsiveness in X ? [...] Your post stimulated me to look into this. I found that `xset m 6 3' made life quite a lot easier. On the other hand, `xset m 10' was a bit wild. I use xset m 6 2 because I find that 6 will cross the screen with a full wrist movement and 2 is as low as you can go and still have single pixel control (1 means that even one pixel of movement is multiplied by 6.) If I use a number greater than 2, I find the transition from unmultiplied to multiplied travel jerky. The algorithm is so crude that I think the word acceleration in man xset is unjustified. It would be nice to have some sort of true acceleration effect. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: fdisk is it available anywhere else?
On Thu, 27 Nov 1997, Orn E. Hansen wrote: Udjat the BitMeister... writes: On Sun, 23 Nov 1997, butch wrote: try cfdisk, I believe its also available on the debian disks. There appears to be a problem with cfdisk... when it writes the partition information, it isn't 'sufficiently' done. I thought the problem was with DOS and was documented in man cfdisk where it tells you (with warnings) to zero out the first 512 bytes of the partition after making any change to it. Like, if I set a partition information with cfdisk to be a DOS partition, when it was previously set to a different type. I need to use fdisk to delete the partition and reinstall it. The information written by cfdisk is unusable by DOS. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: screen snapshot
On Thu, 27 Nov 1997, tko wrote: Wintermute writes: [EMAIL PROTECTED] wrote: How does one get a snapshot/logfile of the bootup process? I know about dmesg but the part which interest me is the processing which occurs _after_ dmesg stops. I've seen various debug messages, but they scroll off the screen before the login prompt comes up. I'd like to investigate and fix them if I could find out what they are. 8-) As for taking a snapshot I'm not sure .but you can access things that roll off of the screen by holding down shift and pressing the page up key to browse through the buffer. I already tried that. It doesn't work. I beleive that the buffer hasn't been set up yet until the login prompt. The stuff that I need to see appears _before_ the login prompt shows. I have no idea how linux implements said scrolling, but I can scroll back to the messages that are printed before the operating system loads (on a system that doesn't wipe the screen before and after scanning the SCSI), thusly: --8 AMIBIOS (C)1992 American Megatrends, Inc. BIOS Version 1.00.05.CV2 0032768 KB Please Wait... Keyboard.Detected MouseDetected ATAPI Device Installed MATSHITA CR-583 Floppy A: Installed NCR SDMS (TM) V3.0 PCI SCSI BIOS, PCI Rev. 2.0 Copyright 1993 NCR Corporation. NCRPCI-3.07.00 ID 00 HP C3325A ID 04 iomega jaz 1GB j Enter Password (1): LILO Loading Linux Uncompressing Linux...done. Now booting the kernel Console: 16 point font, 400 scans Console: colour VGA+ 80x25, 1 virtual console (max 63) pcibios_init : BIOS32 Service Directory structure at 0x000fd2e0 pcibios_init : BIOS32 Service Directory entry at 0xfd2f0 pcibios_init : PCI BIOS revision 2.10 entry at 0xfd311 Probing PCI hardware. Warning : Unknown PCI device (1002:5654). Please read include/linux/pci.h Calibrating delay loop.. ok - 53.04 BogoMIPS Memory: 30880k/32768k available (732k kernel code, 384k reserved, 772k data) Swansea University Computer Society NET3.035 for Linux 2.0 --8 etc. etc. But I'm glad vga=6 helped you. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Linux, a MicroSoft product? (fwd)
On Fri, 28 Nov 1997, Kevin Traas wrote: Where is the original article?? Someone posted to a NewsGroup about 2 weeks ago. Not sure which one The headers here were: X Subject: Linux now a Microsoft Product? X Newsgroups: comp.os.linux.misc,alt.os.linux,comp.os.linux.advocacy X Date: Sat, 22 Nov 1997 03:04:20 GMT X Organization: RainbowHolt X From: [EMAIL PROTECTED] -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: allow mount to normal user
On Tue, 2 Dec 1997, Paul Serice wrote: Adam Klein wrote: man mount and search for the following language: Note that the auto type may be useful for user- mounted floppies. Make sure you're putting auto in the right place. For example, /dev/scd0 /cdrom iso9660 ro,suid,dev,exec,noauto,user,async 0 1 /dev/fd0 /floppy auto rw,suid,dev,exec,noauto,user,sync 0 0 ^^ This definitely works for me as far as auto detecting ext2 or vfat floppies via the KDE floppy mount icon (or the rustic command line). Something else I've learned is that if you plan on using auto, I would recompile your kernel to exclude msdos support but include vfat support. The reason is that auto will check first for msdos 8.3 format. Because vfat floppies meet the criteria, they will be mounted as 8.3. If you compile out msdos support, you'll be able to mount the floppies as vfat. Shouldn't a bug report be filed against mount? It only makes sense for mount auto to check in the order Specific - Generic, not the other way round. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: undeletable file
On Wed, 3 Dec 1997, Ken Lauffenburger wrote: Hello, I remember seeing something about this on the list awhile back, but I haven't been able to find it in the archives. I have a file that has somehow been commandeered by an undefined group, and I can't seem to remove, move, or change its ownership: leisure# ls -l fstobdf -rwxr-xr-x 1 root 65535 29640 Oct 17 23:58 fstobdf rm: remove `fstobdf', overriding mode 0755? y rm: fstobdf: Operation not permitted leisure# chown root.root fstobdf chown: fstobdf: Operation not permitted leisure# rm fstobdf leisure# mv fstobdf fstobdf.1 mv: cannot move `fstobdf' to `fstobdf.1': Operation not permitted I have no idea how the file's ownership got this way, or how to remove it. I haven't tried going to single-user mode yet; I'm on a remote connection to this box. If it's an ext2 filesystem, you could check out the commands chattr and lsattr. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: SCSI Errors and Resetting
On Thu, 4 Dec 1997, Wintermute wrote: William D. Rendahl wrote: We're running Debian 1.3.1 (bo, kernel 2.0.30) on an Intel Premier PCI II motherboard, with a 90MHz Penti, 32MB, BusLogic BT-948, and Quantum Fireball Drive. Several times a day we get messages on the console describing a SCSI error, the system is momentarily frozen, and then everything is fine . . . Whoa.. deja vu.. this is the exact sort of error messages we recieve here at work... WD-7000 initialization failed... And other assorted error messages. For the most part I think they're harmless (We've never had any problems with our Micropolis SCSI drive..) however, it's truly strange.. and I'm the sort that just doesn't like miscellaneous errors popping up, even if they are harmless. Well, let's sort out which error messages you mean. Anybody running an installation kernel is going to see the WD-7000 one (unless they have a WD-7000, I suppose!), and the next line comes from having parallel-port SCSI support but (e.g. a zip drive) not present. (Though I get error 2, not 1.) But these were in the boot up section and are harmless. The others looked more serious, but I'm no expert. Some people have had cabling problems, others have had to switch off a feature like ultra speed. Perhaps the latter is why upgrading the kernel could make a difference - you get the support for the newer feature in the newer kernel. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Documentation of outb() etc.
I'm looking for the documentation of outb() and its friends. man outb(2) claims to document only the extra info to use outb from user space, and refers to outb(9) but I have no section 9. The IO port programming mini-HOWTO says There are man pages for these macros in reasonably recent releases of the Linux man-pages distribution. stable/Contents... has nothing more than outb(2). Where ought I to look next? Any other references to IO port programming would also be welcomed, thanks. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Streaming Password: in VC
On Fri, 5 Dec 1997, Jay Barbee wrote: I believe I botched my Debian 1.3.1 install up, upon the first bootup. I was converting from Slackware to debian after a crash (perfect time to do this), so I had my users and home directories on another partition I did not want to setup the initial user that debian asks for, and I could not figure out a way around it or out of that part of the installation, so I opened up console1 (ALT-F2) and started to mount my drives and start to setup the /etc directory with my old files (like fstab, hosts, you know the rest). I simply added my users to the passwd file that was in place. From there I ran dselect, and finished the install process. I rebooted the system for the second time and VConsole0, even though I never went back to console0 to get rid of that adduser script, was still prompting. Once the reboot finished it promped me for password So I entered my user pasword (for user 'jay') It bumped out to the complete login screen and then after idle time Password: appears. And keeps apearing adding the new word. It fills up the whole screen. Once I enter the password for root I was let in and all was fine. Once I exit, it simply prompted for Password: again! No other Virtual Consoles are affected, just VC0. Can I fix this? Also did I kill my self with the shadowconfig question after the adduser part (before you go into dselect)? I really don't want to go back and reinstall all of Debian again. Take a look at /etc/inittab and you'll see why VC1 is doing this. You may even find a file like /etc/inittab.something.or.other which would get substituted for inittab when the installation script finished (which it didn't). You may have to merge these two files if you've installed things that modified inittab. That's all it should be. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: pppd disconnects after authenticating
On Fri, 5 Jun 1998, Paul Miller wrote: (I'm dialing from a win95 machine to my Linux box) Immediately after I authenticate (using PAP), pppd disconnects me and windows just says to try connecting again... I looked through all the logs for any kind of error or warning, but i didn't find anything -- even with debug on. The last pppd line contains the correct login name and password. Why is pppd disconnecting me? First, who is authenticating who? It's quite likely that the linux box is trying to get W95 to authenticate it. I have no idea how you would deal with that at the W95 end. Are you using mgetty to answer the call. I think it may have a default setting of +pap on the /AutoPPP/ line in /etc/mgetty/login.config which you might need to override with -pap or nothing. Lastly, pppd is rather matter of fact in its log files. It isn't an error when authentication fails; it's the Right Thing. So rather you should look for LCP packets with ConfRej in them. Here's an example from an old email I sent a while ago to someone else. Mar 26 19:10:43 sparky pppd[524]: sent [LCP ConfReq id=0x1 mru 1500 asyncmap0x0 magic 0x4961386f pcomp accomp] Mar 26 19:10:43 sparky pppd[524]: rcvd [LCP ConfReq id=0x1 mru 1500 asyncmap0x0 auth pap magic 0xb9312c88 pcomp accomp] it wants you to authenticate with pap Mar 26 19:10:43 sparky pppd[524]: sent [LCP ConfRej id=0x1 auth pap] you refused ^^^ Mar 26 19:10:46 sparky pppd[524]: rcvd [LCP TermReq id=0x3] Mar 26 19:10:46 sparky pppd[524]: LCP terminated at peer's request Cheers, -- Email: [EMAIL PROTECTED] Tel: +44 1908 653 739 Fax: +44 1908 655 151 Snail: David Wright, Earth Science Dept., Milton Keynes, England, MK7 6AA Disclaimer: These addresses are only for reaching me, and do not signify official stationery. Views expressed here are either my own or plagiarised. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Checking Ports? (fwd)
On Fri, 5 Jun 1998, Marcus Johnson wrote: -- Forwarded message -- Date: Fri, 5 Jun 1998 12:50:21 -0700 (PDT) From: Marcus Johnson [EMAIL PROTECTED] To: John Hasler [EMAIL PROTECTED] Subject: Re: Checking Ports? Sorry for my brevity and thank you for your answer. I meant ports as in IP ports. I tried to start a server-type program but it exits and tells me that port is already in use. This seems very odd to me because I should be the only one using it. I suspect its still running another copy of my server. How can I find out and if it is another instance of my server how can I kill it? I run a very simple joystick server that writes its position to a port, say, 50005, so that (a) multiple clients can connect to the stick (b) I can run a fake server on machines without a stick and continue to test programs that need to read a stick. When I kill the server, I find I often can't restart it on that port, but can on, say, 50006. I'm guessing that some resource stays around until it's killed off, usually after less than a minute. I intend making my server pick the next free port, and likewise the clients will have to play chase the server. Meanwhile I don't worry. Cheers, -- Email: [EMAIL PROTECTED] Tel: +44 1908 653 739 Fax: +44 1908 655 151 Snail: David Wright, Earth Science Dept., Milton Keynes, England, MK7 6AA Disclaimer: These addresses are only for reaching me, and do not signify official stationery. Views expressed here are either my own or plagiarised. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: A debian-user question.
On Tue, 9 Jun 1998, Karsten Bolding wrote: For the GUI I'm thinking of either perl-tk or using the visual tcl package - any opinions? The GUI will contain radio-buttons, entry-fields, file-browser etc. to specify the different parameters for the model. Don't choose either of these without at least first considering tk over python (www.python.org). Though one could run into religious wars comparing perl and python, I'm certainly happier modularising python than I was with perl, and tk seems to be well integrated. I'm in the early stages of a project with them myself. My real question concerns a package that will do real time plotting of the results of the turbulence modelling - this is simple X-Y graphs but they evolve over time so I need to - at run time - update the graphs based on the results of the turbulence calculations. Since I'm not at all a X-programmer I would like something a bit more high-level. Couldn't disagree with another's mention of gnuplot. Cheers, -- Email: [EMAIL PROTECTED] Tel: +44 1908 653 739 Fax: +44 1908 655 151 Snail: David Wright, Earth Science Dept., Milton Keynes, England, MK7 6AA Disclaimer: These addresses are only for reaching me, and do not signify official stationery. Views expressed here are either my own or plagiarised. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Multiple IP numbers, one for each Subnet
On Tue, 9 Jun 1998, Ian Stuart wrote: I'm using a i386 portable with 2.0.34 installed. I have a requirement to plug this machine into various subnets, depending on where I'm working within my organisation. Is there a way to configure the system so that I can use (a) 4 or 5 different gateways (only one of which will be reachable at one time) and (b) 4 or 5 IP numbers, only one of which will be valid at any one time? There are very few files that actually contain IP numbers, viz. /etc/resolv.conf, /etc/hosts, /etc/networks and /etc/init.d/network and the first of these would probably stay the same as we're talking subnets. So all you need is a few scripts hacked from the last, that delete the two (network and gateway) routes, ifconfig the new number, and add the two new routes. They could meanwhile overwrite the middle two files as well. ifconfig and route commands are totally dynamic; I don't think daemons etc. should hold onto any of this information. (I don't have a clue what reads /etc/networks though I suppose you could use it yourself to delete the appropriate -net route.) Cheers, -- Email: [EMAIL PROTECTED] Tel: +44 1908 653 739 Fax: +44 1908 655 151 Snail: David Wright, Earth Science Dept., Milton Keynes, England, MK7 6AA Disclaimer: These addresses are only for reaching me, and do not signify official stationery. Views expressed here are either my own or plagiarised. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: 2.0.34
On Tue, 9 Jun 1998, Hamish Moffatt wrote: bash-2.00# ifconfig eth0 203.14.18.1 netmask 255.255.255.128 broadcast 203.14.18.127 SIOCSIFNETMASK: Invalid argument It doesn't work on 2.0.32 either, I just discovered. Thanks to everyone who has replied. Any other ideas? It's a long shot but what does the rest of what would be your /etc/init.d/network say? I only ask because you've played around with newer kernels which do things differently in this file. Cheers, -- Email: [EMAIL PROTECTED] Tel: +44 1908 653 739 Fax: +44 1908 655 151 Snail: David Wright, Earth Science Dept., Milton Keynes, England, MK7 6AA Disclaimer: These addresses are only for reaching me, and do not signify official stationery. Views expressed here are either my own or plagiarised. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: user can delete kernel images (cont)
On Fri, 12 Jun 1998, Steve Mayer wrote: I noticed this on my one remaining bo machine. Hamm seems to have taken care of this bug. G. Kapetanios wrote: Following to my previous email I have to say some things. the /boot directory in my machine is drwxrwsr-x 2 root disk 2048 Jun 12 17:58 boot the user who can do that belongs to the disk group but the file which was deleted (/boot/vmlinuz.2.0.0) does not belong to the disk group it is root.root So obviously although I hadn;t realised that before if a group you belong to owns a directory which is writable by the group you can delete stuff from it without owning the files and without belonging to the group which owns the files. Is this safe ?? More importantly I don't know if this is a bug of the installation procedure about 1 1/2 years ago but the permission to /boot were set by that procedure and I never changed them. I know users should not probably belong to group disk but I could have damaged my system really bad if I had no spare kernels. I guess I must remove the user from the disk group as soon as possible. By the way why is /boot writable by the group disk? I don't know the correct permissions for /boot files, but in terms of security, protecting them from disk-group users will be quite ineffective: the disk group has write permission for raw disk devices. The point about who can delete files in a directory is of course covered in the FAQ under IIRC I've discovered a HUGE security hole in rm!. Cheers, -- Email: [EMAIL PROTECTED] Tel: +44 1908 653 739 Fax: +44 1908 655 151 Snail: David Wright, Earth Science Dept., Milton Keynes, England, MK7 6AA Disclaimer: These addresses are only for reaching me, and do not signify official stationery. Views expressed here are either my own or plagiarised. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: /etc/init.d/boot message?
On Fri, 19 Jun 1998, Mark Yobb wrote: I know how to use `dmesg` but I would really like to be able to get the info that scrolls across my screen (on bootup) when /etc/init.d/boot is executing. Is this message sent to a log file or something? How can I look at this after booting? Simple. Just use shift page up and down. This works at any time, just so long as you don't switch consoles. Cheers, -- Email: [EMAIL PROTECTED] Tel: +44 1908 653 739 Fax: +44 1908 655 151 Snail: David Wright, Earth Science Dept., Milton Keynes, England, MK7 6AA Disclaimer: These addresses are only for reaching me, and do not signify official stationery. Views expressed here are either my own or plagiarised. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: /etc/init.d/boot message?
On Fri, 19 Jun 1998, Ed Cogburn wrote: David Wright wrote: On Fri, 19 Jun 1998, Mark Yobb wrote: I know how to use `dmesg` but I would really like to be able to get the info that scrolls across my screen (on bootup) when /etc/init.d/boot is executing. Is this message sent to a log file or something? How can I look at this after booting? Simple. Just use shift page up and down. This works at any time, just so long as you don't switch consoles. Parts of the boot up text can be found afterwords in /var/log/messages. Yes, while I agree that parts of the boot up messages can be found in various places in a more permanent form, the only place that has them /all/ in the exact /sequence/ is the console. If something goes wrong, it's worth looking here /first/ because it's the last chance you get. As soon as you look elsewhere, the complete console record has gone for ever (until you reboot). Cheers, -- Email: [EMAIL PROTECTED] Tel: +44 1908 653 739 Fax: +44 1908 655 151 Snail: David Wright, Earth Science Dept., Milton Keynes, England, MK7 6AA Disclaimer: These addresses are only for reaching me, and do not signify official stationery. Views expressed here are either my own or plagiarised. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Moving from Ethernet to modem
On Fri, 26 Jun 1998, Ted Cabeen wrote: When I set my system up on Debian, it was connected directly to an ethernet link, so the entire system is set up to expect connections. Now, I've had to move away from my direct connection and I need to close out all of the systems that rely on that connection being there, like apache and xntpd. Which standard programs should I be removing or shutting down? Eventually, we're planning on setting up a system for ip masquerading and on-demand ppp dialing by a gateway for the entire apartment, and I don't want my machine starting or keeping that connection up too much. Thoughts? My machine at home is just a backup system from work which I need to be able to use at short notice. I have copies of /etc/init.d/network for home and work. The only substantial difference: GATEWAY= is hashed out at home. (Insubstantial difference: I have a 168 network alias to plug my wife's laptop into.) Obviously I leave a terminated T-piece in the NIC. I just ignore the occasional daemonic howls of protest. I raise ppp myself, rather than on demand, as I'm the only user. Cheers, -- Email: [EMAIL PROTECTED] Tel: +44 1908 653 739 Fax: +44 1908 655 151 Snail: David Wright, Earth Science Dept., Milton Keynes, England, MK7 6AA Disclaimer: These addresses are only for reaching me, and do not signify official stationery. Views expressed here are either my own or plagiarised. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: /etc/init.d/boot message?
On 27 Jun 1998, Brederlow wrote: [snip] A bit more info would be nice. I for one would be quite happy just to have a record of what has been printed on the screen. Your right, the detailed messages would go to a logfile and maybe to a ^ different console also. If I understand you, I think your priorities are wrong. The present situation is perfect: everything goes to the screen and can be seen if and when things go wrong*. It would be nice if this everything was available in one place elsewhere; but only nice. As it is, it's spread about (dmesg, messages etc.). (* I am aware of svga...'s clearing the screen and the bug in xdm that lets you loop in and out of X.) On a normal startup nothing should fail and it looks much more professional to have a nice menu the normal user can understand instead of cryptic messages teling the experience that everything is fine. Now I don't understand. Menu for what? And who's necessarily there to select from this menu? Perhaps we should go for a picture of clouds with Debian 98 :-( No, the great thing about linux booting up is that you can see that it's working. Positive confirmation. You're not left wondering whether a lack of error messages means that it all worked, or that the error reporter itself isn't working. Cheers, -- Email: [EMAIL PROTECTED] Tel: +44 1908 653 739 Fax: +44 1908 655 151 Snail: David Wright, Earth Science Dept., Milton Keynes, England, MK7 6AA Disclaimer: These addresses are only for reaching me, and do not signify official stationery. Views expressed here are either my own or plagiarised. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
RE: linux + win95: linux boot partition/
On Wed, 15 Jul 1998, Richardson,Anthony wrote: The 1024 problem is a very real one. Yes, I agree it is. However, Hamish was commenting on a posting that referred to SCSI drives, not IDE ones. Unfortunately he had pruned the quotation so much that all references to SCSI had disappeared. So the remainder (quoted immediately below) gave a very misleading impression, particularly the last statement made in a separate and unqualified paragraph. -Original Message- From: Hamish Moffatt [SMTP:[EMAIL PROTECTED] Sent: Tuesday, July 14, 1998 9:37 PM To: p.meidl; debian-user Subject: Re: linux + win95: linux boot partition/ On Tue, Jul 14, 1998 at 06:42:19PM +, Patrick Meidl wrote: after reading the relevant FAQs, HowTOs, installation instructions etc. I recognized that all bootable partitions must start before the 1024th cylinder (I would like to use LILO), so I thought the best solution might be to have these partitions: With LBA this appears to be incorrect. I have previously had systems booting Linux from the last 500mb of a 1.6gb drive; the 1024 limit only takes you to 528mb or so. I boot NT 2gb into a 6gb drive; no problem. I have never encountered any 1024 cylinder problem with Linux. I wish the documentation would not keep spreading these ideas. Hamish -- -- Email: [EMAIL PROTECTED] Tel: +44 1908 653 739 Fax: +44 1908 655 151 Snail: David Wright, Earth Science Dept., Milton Keynes, England, MK7 6AA Disclaimer: These addresses are only for reaching me, and do not signify official stationery. Views expressed here are either my own or plagiarised. -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: linux + win95: linux boot partition/
On Wed, 15 Jul 1998, Nils Rennebarth wrote: On Wed, Jul 15, 1998 at 08:39:00AM -0400, Richardson,Anthony wrote: The 1024 problem is a very real one. Please, it really occurs in very few systems/configurations. It had been a problem for me occasinally because of some older mainboards lying around here. It is no problem with harddisks and mainboards bought over the last three years. ^ eh? I'm very happy for you. I now run eight linux boxes, three of them modern enough that I had to use patched kernels for the SCSI until bo was released. However, the other five are all well over three years old and the mainboards don't lie around, they earn their keep. However, I have had the good fortune to acquire bigger disks, mainly because disks have necessarily fallen like manna from heaven as the rest of the lusers here (MS drones) have been migrated to W95. So for me, the 1024 limit is an increasing problem, not a diminishing one. Not immediately obvious, perhaps. Some BIOSes allow you to choose whether translation should be done with settings like Large or LBA for other BIOSes translation is on by default. All BIOSes I saw until now gave me the LBA option which I choose. See below. Only one of my five older machines has heard of LBA. While I smile when I read some of the HOWTOs I printed off several years ago, I don't forget that there are plently of people still grappling with older hardware (some still running 1.2.13 as well!). It's refreshing that linux allows one to make such choices. The only thing our administrators know what to do with such hardware is bin it. Cheers, -- Email: [EMAIL PROTECTED] Tel: +44 1908 653 739 Fax: +44 1908 655 151 Snail: David Wright, Earth Science Dept., Milton Keynes, England, MK7 6AA Disclaimer: These addresses are only for reaching me, and do not signify official stationery. Views expressed here are either my own or plagiarised. -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: LPRng printer Port
On Tue, 21 Jul 1998, Babs wrote: HELP! I have read every doc i can get my hands on, news groups, more. I can not figure out why my printer port is not opening. I'm running an X86 based PC, Debian Linux Kernel 2.0.34, I am in the process of integrating the internal network at our office. I'm still fairly new to linux been running it for about 5 months or so. Still have lots to learn. I have Netatalk and Samba Running and configured properly. At one point i did have LPRng Suite working properly as well. Crashed the system and had to do a complete reinstall. ! I'm not sure why you had to do that. From that point on i have not been able to get the printer port to open, no matter what i do. Does anyone have any ideas for me. If i can get the port open i can work on filters later. First, is the lp port set up correctly in the CMOS, with the conventional address? You shouldn't need an IRQ as polling is used. Then does the kernel find it? Look for something like lp1 at 0x0378, (polling) in the boot up screen (shift-pageup to review). Are you running your own kernel or an installation one? If you see something like lp: Driver configured but no interfaces found. perhaps something like the ppa driver has already grabbed the port. I would strongly recommend building your own kernel, tailored to what you've got, than have an installation one probing here there and everywhere for non-existant devices. Another possibility is that you forgot to install the lp module the second time around. I think everything you remembered has a line in /etc/modules; is it there? Cheers, -- Email: [EMAIL PROTECTED] Tel: +44 1908 653 739 Fax: +44 1908 655 151 Snail: David Wright, Earth Science Dept., Milton Keynes, England, MK7 6AA Disclaimer: These addresses are only for reaching me, and do not signify official stationery. Views expressed here are either my own or plagiarised. -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: Installing a single package using dselect.
On Wed, 22 Jul 1998, Chris Russell wrote: Is there any way to add a single package using dselect without updating the Packages file. All I want to do is add the man-db and netstd packages to the base system. Seems like the deselect is overkill and dpkg is underkill. ?Why wouldnt these packages be part of the base Taken together, these packages are about 1MB in size, so that's *another* floppy disk. Your wanting to use W95 as a go-between but as the client rather than the server is pretty unusual, so it doesn't seem worth everybody else adding a non-essential disk to the installation process. ==The reason I want this=== I'm trying to get the Beta 2 installed via the net with a twist. I have a laptop running win95 that I can use to dial out and it is connected to my 5port hub which the debian box is also attached to. I do not have a modem for the linux box. With a ftp daemon on the debian box I could use the pc as a go between until I get a modem for the beast. (Yea I know I could get a ftpd for win95 but I wold rather go the other way.) Cheers, -- Email: [EMAIL PROTECTED] Tel: +44 1908 653 739 Fax: +44 1908 655 151 Snail: David Wright, Earth Science Dept., Milton Keynes, England, MK7 6AA Disclaimer: These addresses are only for reaching me, and do not signify official stationery. Views expressed here are either my own or plagiarised. -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: Lan Tcp/ip Question
On Thu, 23 Jul 1998, Craig Sanders wrote: On Mon, 20 Jul 1998, Tomt wrote: At 10:44 PM 7/19/1998 -0700, you wrote: It may be useful for you to assign the NIC's address to something other than 0x300. A lot of different (very different even!) cards try to use 0x300 (sound cards, primarily). Theres a sound card in the machine but its sitting on 0x330 Also you may want to try pinging the machine's own address on the ethernet. See what that produces. Aside from that I can't help you much. Works. Both machines can ping themselves but not each other. check that you don't have an irq conflict with the ethernet card. i've had enough irq conflicts with network cards for that to be the first thing i check when i get a system which can send but not receive packets. Also, wasn't this a 509 card? (I've chucked the start of the thread.) It's worth checking that you've got the right connection setting if it's a combo. I've got several that won't autoselect the BNC, but have to be set to BNC specifically. This might have the same symptoms. Cheers, -- Email: [EMAIL PROTECTED] Tel: +44 1908 653 739 Fax: +44 1908 655 151 Snail: David Wright, Earth Science Dept., Milton Keynes, England, MK7 6AA Disclaimer: These addresses are only for reaching me, and do not signify official stationery. Views expressed here are either my own or plagiarised. -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: starting ppp on host end
On Tue, 11 Aug 1998, Richard E. Hawkins Esq. wrote: jens wrote, Richard E. Hawkins Esq. wrote: huh? This is an 8 bit clean connection. so how to i test it myself to have something to show the telecommunications folks? Or could my modem be sending 7 bits? It's a compaq pcmia modem in an ibm thinkpad. any help would be appreciated. Most likely the problem is simply that ppp hasn't started--the shell is probably just echoing back the LCP config requests. Try going through the chat procedure by hand and see if you get PPP packets after you ultimately login to the Linux box. oh :) I thought that the script started it at the other end :) OK, for the really dumb question: how do I start ppp on the other end on a debian box? it seems to be with pppd to start the daemon, but I'm having trouble figuring out the man doc pages. I've figured out to insert the ppp shlc modules on the host, but I'm not clear on what else to do. When I type pppd when logged in manually, I just get a bunch of nonsense characters. I think someone should file a bug report against pppd because it really ought to be able to distinguish no data at all from no data with top bit set, which it doesn't. So the error message only confuses people. If you're connecting to an ISP, you may have to send ppp in response to a prompt, but that shouldn't be necessary when dialling into a Debian box if you're running mgetty, because of the line that starts /AutoPPP/ ... in /etc/mgetty/login.config As soon as mgetty receives LCP stuff, it starts PPP. If the nonsense characters contain plenitudinous {{{ characters, LCP stuff is what you're seeing. Cheers, -- Email: [EMAIL PROTECTED] Tel: +44 1908 653 739 Fax: +44 1908 655 151 Snail: David Wright, Earth Science Dept., Milton Keynes, England, MK7 6AA Disclaimer: These addresses are only for reaching me, and do not signify official stationery. Views expressed here are either my own or plagiarised.
Re: nasty...
On Tue, 11 Aug 1998, Christopher Barry wrote: Yes, but it won't create past /dev/sdn15. The last time I installed Debian I had put 16 partitions on my brand new 9.1GB SCSI disk and then found I only had sda devices numbered up to 15. I read the manpage for /dev/MAKEDEV and found it pretty useless as well for this problem. They really should tell you how to do things like this, or at least have a more intuitive way (i.e. /dev/MAKEDEV /dev/sda16). Heh heh, some very interesting things happen when you try cp on a disk device (after trying everything else I tried copying sda15 to sda16 thinking it would just copy the tiny little file... very strange what starts to happen). Fortunately though it's not life-or-death that I have 16 partitions so I was able to just cfdisk 1 away and move on. It may help clear things up if you look at the output from ls -l /dev/sda16 /dev/sdb brw-rw 1 root disk 8, 16 May 28 1997 /dev/sda16 brw-rw 1 root disk 8, 16 May 28 1997 /dev/sdb Cheers, -- Email: [EMAIL PROTECTED] Tel: +44 1908 653 739 Fax: +44 1908 655 151 Snail: David Wright, Earth Science Dept., Milton Keynes, England, MK7 6AA Disclaimer: These addresses are only for reaching me, and do not signify official stationery. Views expressed here are either my own or plagiarised.
Re: nasty...
On Thu, 13 Aug 1998, Hamish Moffatt wrote: On Wed, Aug 12, 1998 at 04:12:45PM +0200, [EMAIL PROTECTED] wrote: On Tue, 11 Aug 1998, Hamish Moffatt wrote: I just removed some of the old buzz/rex packages, base, timezone, bdflush. I did it in dselect, and apt quite happily obliged. Aargh! you removed base? You might be in for some trouble. Try to run dpkg -i base-files.deb before you reboot. That will put some of the vital files back I hope. It doesn't. base-files does not contain ANY devices. I am a bit surprised at how many of the files in the base.tgz file are not owned by any package after installation -- I think this is bad. Not even the kernel belongs to any package after initial installation. These issues have been discussed some months ago (esp. w.r.t. base,) but some people think that it is Supreme Evil to munge with files in /var/lib/dpkg/info (that's what you need to do to get rid of base safely.) IMHO having your system flushed is far worse. In the case of timezone{,s}, I don't know exactly where the problem lies. You should file a bugreport. Surely SOMETHING could be done to prevent removing base from trashing the system. base-files should own the same set of files anyway I should think; I can't see why it wouldn't provide the devices. You're going so far back (buzz) that memory is hazy but IIRC buzz had base, and base had devices, and if you purged base, all your devices disappeared. I think Bruce or some other god put together a posting which showed exactly what to do. (I think you just deleted some of the lines in /var/lib/dpkg/info/base.list first.) This was because of the number of postings from people who wanted to purge base because it was listed as obsolete (very untidy). Oh, and the reason base showed at all was because base-files was introduced (presumably in rex) instead of base. And base-files *didn't* have device files in base-files.list, probably for that very reason, that purging it would remove them! No /dev files now (bo) appear in *list. I think that answers all the points raised, except perhaps to say that it isn't in the spirit of unix/linux to prevent you (as root) from trashing the system if you really want to. Cheers, -- Email: [EMAIL PROTECTED] Tel: +44 1908 653 739 Fax: +44 1908 655 151 Snail: David Wright, Earth Science Dept., Milton Keynes, England, MK7 6AA Disclaimer: These addresses are only for reaching me, and do not signify official stationery. Views expressed here are either my own or plagiarised.
RE: Nuking damned scrambled consoles.
On Thu, 13 Aug 1998, Hank Fay wrote: Well nothing seems to get rid of the little critters except shutdown. I had to switch to another VC this time, because something unknown was going on in the first one. It seems to be getting stuck in a different video mode. Well you could also try typing setfont default8x16 This is the only thing that would sort out an old Avance card that I have when i exitted from X to a VC. But I'm not optimistic, as my problem affected *all* VCs. Rather than rebooting, have you tried just killing the relevant VC from another one? Cheers, well, it helped jumble up the funny critters. s The only thing that works so far is shutdown Every now and then I do a little goof-up that scrambles a virtual console and I'm sure we all do sometimes but lately I've been doing a little programming and if I accidentally gib a string argument then it corrupts the console every single time so I quickly run out of all 6 consoles and am forced to reboot. The way that I get ride of a scrambled console. 1. Try typing reset 2. Try typing clear 3. Try running top This always seems to work. don't know why but it does. Try CTRLvCTRLoENTER -- Email: [EMAIL PROTECTED] Tel: +44 1908 653 739 Fax: +44 1908 655 151 Snail: David Wright, Earth Science Dept., Milton Keynes, England, MK7 6AA Disclaimer: These addresses are only for reaching me, and do not signify official stationery. Views expressed here are either my own or plagiarised.
Re: DE200 network card
On Tue, 18 Feb 1997, Alex Monaghan wrote: Colin Watt wrote: Does Debian (or any Linux) support a DE200 network card? I can't see it on the list. Don't know about Debian, but assume it's similiar to Slackware. I have a few DE-100's and these are recognised by the NE drivers. If your card is I don't know about NE drivers, but the DE200 is covered by the depca driver, see /usr/src/linux/drivers/net/depca.c for details. I found that probing a DE100 card with cdrom drivers etc. kills it. I don't know if DE200s are the same. I had to compile a kernel on another (3c509) machine to be able to use the ones with DE100 cards at all. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Sendfax errors
I recently installed mgetty 1.0.0-1 and it pretty well worked out of the box with a USR 28.8 V.34 33.6 faxmodem. It even told me to put switchbd 0 in mgetty.config when I sent it the first fax. However, I can't sendfax back to the machine I faxed it from. I can fax to an identical Debian/USR system, but all I get from our fax machine is transmission error 24 (RSPEC). I tried another fax machine and got error 25 (DCS sent three times without response). Neither fax machine emits anything. I'd love to look up these errors, and the fax commands that are logged in sendfax.log, but I have no idea where to start. The modem manual (and all the associated documents at USR's ftp site) are completely silent on fax commands beyond the likes of AT+FCLASS=0 to get a modem back into data mode. The error messages in the fax's manual look completely different (manufacturer specific?). There's a table of numeric codes and their strings in /usr/doc/mgetty/fhng-codes but the programs already give you the verbose error. alt.fax and comp.dcom.fax are no help - I saw an unanswered request for error codes when I looked. Where ought I to look? David. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: Directory structure at Debian
On Thu, 20 Feb 1997, Ed Down wrote: Could anyone tell me what the rex-updates and rex-fixed directories are for? I would have assumed that any packages with serious bugs would be fixed in the standard rex... If not, what is rex still there for with 'broken' packages in it? And what happens when you replace your 'rex' packages file with the 'rex-fixed' one - Does it tell you you need to upgrade all your packages to rex-fixed level? It's 5.7.1 in the FAQ. The point is that rex is a release - it never changes, bugs or no bugs. It would seem simpler to me if we just had two directories containing the base, x11, non-free, contrib, etc directories: Debian-1.2, which is 'fixed' of important bugs as they are cleared and would be used by almost every 'user', and Debian-1.2-devel, which contains the most up-do-date working release of each package, for use by 'developers'. This makes Debian even more of a moving target, a bad thing. I already keep all the packages I install on two zip disks, so I can easily clone another machine and know it should work (barring hardware differences). David. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: dselect and lprng...
On Thu, 27 Feb 1997, Scott Stanley wrote: Oddly enough, I have never looked at lprng before, so there is no way it found it on my system. When I complete the install process, it still says I want to install it, but that it isn't installed. It is almost as if dselect does not recognize that I want to install the package, so it does not hunt for it on the FTP site. Scott Stanley wrote: I was just trying to download and install the lprng package from stable using dselect 1.4.0.7 with the ftp Access method. Basically, I went into Select, and selected the package lprng (I got the dependencies window saying lprng recommended magicfilter, but I did not select magicfilter). The select menu item for lprng looks like; _* Opt net lprng none2.4.2-1 lpr/lpd printer spooling When I exit (pressing Enter) and Install, I get -- Processing status file... Processing Package files... stable... contrib... non-free... Constructing list of files to get... Approximate total space required: 0k Available space in ../../../../../dosc/Debian/1.2: 235632k Nothing to get. Processing downloaded files...(for corrupt/old/partial) (I'm the person who suggested running update in dselect.) I couldn't find lprng at ftp.mcc.ac.uk so I went to ftp.debian.org and it wasn't there either, except in bo. So it's a mystery to me whence came Opt net lprng none 2.4.2-1, but no surprise that dselect couldn't download it. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151
Re: Laodable Modules
On Sun, 9 Mar 1997, Rob MacWilliams wrote: Hi all, I am trying to compile a custom kernel for the usual reasons. I am not seeing any errors during the compile, installation or boot phases, but some of the modules won't install. Here are the steps I took to compile the kernel: mv /lib/modules/2.0.27 /lib/modules/2.0.27.old mv /vmlinuz /vmlinuz.old I don't do this. cd /usr/include mv asm asm.old mv linux linux.old mv scsi scsi.old ln -s /usr/src/linux/include/asm-i386 asm ln -s /usr/src/linux/include/linux linux ln -s /usr/src/linux/include/scsi scsi (linux is a symbolic link to /usr/src/kernel-source-2.0.27 it was a link to /usr/src/kernel-headers-2.0.27 but both give the same errors) These links should not be made with Debian. make mrproper make xconfig make dep make clean make zImage make modules make modules_install cp /usr/src/kernel-source-2.0.27/arch/i386/boot/zImage /vmlinuz Rather than this cp (and the previous mv) you can use make install. Then remove the modules.dep before rebooting and depmod is done automatically. edit /etc/lilo.conf so I have a backdoor if this crashes run LILO edit /etc/modules so that only auto is uncommented reboot login as root $depmod -a $insmod psaux Why have the mouse built as a module? misc_register undefined misc_deregister undefined Loading failed! The module symbols (from linux-2.0.27) don't match your linux-2.0.27 This may be because you didn't copy the System.map along with vmlinuz. BTW I copy both to /boot and use links from / (if only because that's how the installation system was set up. $modprobe psaux Initialization of psaux failed $insmod misc $insmod psuax Cannot open /lib/modules/misc/psuax.o Checking the datestamp confirms that it was compiled along with the other modules that work. psaux isn't the only module that fails to load, but hpfs and ppp work fine. I can dig through all of the modules and see exactly which ones work and which ones don't if that would help. If an strace of insmod would help I can include that also, but that would have made this message huge. All of the important parts of the system are Debian/GNU. Can you shorten your lines, please. Any clues? I'm relativly new at this, but having a great time learning all of the ins and outs. Thanks Time is the best teacher, unfortunately it kills all of it's students Rob MacWilliams [EMAIL PROTECTED] N9NPU I hope that's a little help. Good luck. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151
Re: Problems Compiling Kernel 2.0.27
On Sat, 8 Mar 1997, David James Loken wrote: Hi! I have been trying to recompile my kernel 2.0.27. I would like to use Menuconfig but I'm missing the file 'lxdiaglog.o' in /usr/src/linux/scripts/lxdialog. Does anyone know where I can get a copy? I'm not sure why you'd want this file. The only catch with running menuconfig is that you need to install ncurses-dev. So I tried running 'make config' that goes along fine until I try to compile the sound board option. gcc tells me I'm missing the 'configure' in /usr/src/linux/drivers/sound. Did you set both Sound card support and /dev/dsp and /dev/audio support as well as the specific card (say, Sound Blaster for example). I have read the Readme.linx file, but I don't how to run the script that is appended to the end of Readme.linux after I have deleted the first part of file. I copied the file to /root, called it sound.configuration, edited out the surrounding text, chmod u+x so it's executable and typed ./sound.configuration (I think). Remember that . is not in root's path (security risk). Also I have tried to compile the 2.0.27 kernel without sound. 'make config' works fine but, when try running 'make dep' gcc say it can't find 'mkdep.c' even though 'mkdep.c' is present in /usr/src/linux/scripts. If it is in the wrong directory which directory should I put it in. I think if you get error messages, it usually means you've done something wrong in the configuration stage. It certainly did with me. (Or forgetting to clean each time.) I'd be loath to start moving things about in the source tree. Thanks in advance, 73 for now. Dave Loken VE6DJL -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151
Re: Problems Compiling Kernel 2.0.27
On Sun, 9 Mar 1997, johannes martinez wrote: Thought wrote: I normally don't compile sound board support, because I never use sound in Linux, but I was just messing around and I decided 'what the hell' and included it when I was remaking my kernel, and I got a bunch of missing configuration files/setup errors when trying to compile too. I just thought 'ahh screw it' and compiled without sound. Is there a problem in 2.0.27 with sound? Same problem here. However when i removed support for the yamaha chip it seemed to work okay. Course then i did a dselect, and got my new kernel replaced and had to build it again. johannes martinez No problem here except that it needs a bit of care where you configure the card - IRQs, DMAs and whatnot - and there's no help available for those options, unlike the rest of the configuration. As sound is built onto my motherboard, I have little documentation and basically I used the original CONFIG.SYS parameter line to guess what I should respond. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151
Re: kernel-config
On Tue, 11 Mar 1997, Mikael Hallendal wrote: I try to get the sound working i linux and when I do make config in th kernel-source I'm aksed to enter the I/O-base but the only thing that happens is that it says 'no help is available. Can anyone help me, please! The problem is: When I come to the point where I'm supposed to enter the I/O-base adress for the SB-card I choose the default-value (which is 220). But no matter what I enter the only response is 'No help page for this option yet' .. Just a workaround, I'm afraid, but you might never go back. Install ncurses-dev and then make menuconfig. I certainly managed to configure my mobo soundcard like that. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151
Re: Module vs Prago
On Tue, 18 Mar 1997, Greg Vence wrote: What are the pro's con's of compiling into the kernal vs. loading a module? Where could I find docs on topic? Variously scattered I think. There's also the issue of loading with auto or specifically. A few thoughts to get you started on a mini-FAQ! 1) Compile in anything needed to get to your disk, e.g. scsi stuff. 2) Modularise anything you need to remove, e.g. ppa and lp if you're running a parallel-port zip drive with the current drivers in stable. 3) Don't use auto on modules which need to retain settings, e.g. serial ports and sound cards. 4) Don't use auto on things that will make kerneld work hard, e.g. mouse. David. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151
Re: Network Configuration
On Tue, 25 Mar 1997, Jordi Inglada wrote: The problem now is that I can't make it work. I mean, I can't do ftp, telnet, etc... When I try some of these, I wait and wait,... If I use nfs as the access method to the distribution, I got the message server my nfs server seems to be down or inexistent. I have configured the network, I have installed the drivers (my card is a DIGITAL depca), the card is found when rebooting at the correct address and interruption... Ouch! Depca card, eh? A couple of things to check: It's not enough for the card to be detected. It's also rather important that any probing for other hardware doesn't interfere with the depca card. So you probably won't get it to work with an installation kernel which probes for everything everywhere. The depca driver may also not work if you compile it as a module, so when you compile your kernel, select it as built-in. I hope that helps. Me, I'm replacing my depca cards with 3c509 ones. Depcas worked with pathworks and with netbeui (just about), but they freeze up with PC-NFS after a couple of minutes so I'm calling it a day (as is the Open University). -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151
NCR53c875 chip in SCSI card won't work
I've got two new computers which have Symbios Logic SYM8751SP cards with the NCR53c875 chip. The installation disks I've tried (resq1440.bin of 4Jan97 19:39 and rsc1440.bin of 18Jan97 21:53) both fail to detect any disks so I copied my custom 2.0.27 kernel from a previously new computer (which has the 53c815 chip in it, but very little different) in place of linux and tried again. The 53c875 card is detected but doesn't work properly. I don't know if the problem is something to do with PnP because, naturally, the machine was supplied with W95 (including a FAT32 disk). These are the messages that booting up linux produces: ncr53c8xx at PCI bus 0 device 13 function 0 (I don't know what the 13 refers to. On the PCI, the host is 7, the disk is 0) ncr_attach unit=0 chip=875 base=ffbeec00 io_port=f800 irq=11 (these values agree with W95 properties, though the latter shows memory ranges c8000-ca7ff, ffbeec00-ffbeecff, ffbed000-ffbedfff) ncr53c875-0 using memory mapped IO at virtual address 0x2823c00 initial SCNTL3=55 final=35 shared irq 11 dev id=0x84068 restart (scsi reset) scsi0 : ncr53c8xx (rel. 1.12c) scsi : 1 host Fujitsu M2952E-512 rev. 0142 Direct-Access ANSI SCSI revision 02 Detected scsi disk at scsi 0 channel 0 id 0 lun 0 scsi : detected 1 SCSI disk total ncr53c875-0-target 0, lun 0: WIDE SCSI (16 bit) enabled.| FAST SCSI-2 100ns (20Mb/sec) offset 8. | ncr53c875-0: restart (ncr dead ?).| sda : READ CAPACITY FAILED sda : status=1 message=00 host=0 driver=28 extended sense code=6 block size assumed to be 512, disk size 1GB (then it detects the 3c509 correctly) Partition check: Then the three lines with the side bar are repeated ad nauseam. Are there any clues in there? Or do I have to go back to Viglen and ask for 815 chipped boards to be supplied instead? -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151
Re: XFree86 Mouse problem on PS2 mouse.
On Fri, 4 Apr 1997, Syrus Nemat-Nasser wrote: On 2 Apr 1997, R. Chris Ross wrote: I recently sent a message regarding a problem that I am having with my mouse on XFree86. The mouse is a PS2 style of Mouse Systems mouse. According to the documentation that I have found the only setting in the XF86Config file that will work in the PS2 setting. When I run XF86Setup the only driver setting that seams to come close is the PS2 setting because the MouseSystems driver seams to only be set up for a serial mouse and this mouse is NOT a serial mouse. I have tried quite a selection of things and am getting quite frustrated. When using the PS2 driver the 3 buttons work fine and the mouse works fine going left to right or bottom to top. The problem is when going the other direction, where the mouse shoots immediately to the edge of the screen. This really is interesting to use. The thing works perfectly fine in Win 95 so the mouse is fine I just can't get the thing to respond properly in X. Please help. Hi Chris. Are you using gpm? I've heard that gpm should not be running when running X if you have a PS2 mouse. I also recall having a mouse like you describe (same brand even) work with the same settings you describe, but I did not install gpm on that machine. If gpm is not the problem, let me know and I'll look for the XF86Config file from that machine. I didn't respond to the original query because I didn't understand why the mouse should work for all but one particular direction. However, I can't see why there should be a problem with running X and gpm. I run both these with a M$ PS/2 mouse. My kernel has Mouse support (not serial mouse) and PS/2 mouse (aka auxiliary device) compiled in (not modules) and I have -R set in gpm. XF86Config is set to Mouse Systems protocol from /dev/gpmdata and it all works fine. BTW I did try running /usr/sbin/xbase-configure and liked it BUT don't touch the mouse until after APPLYing the right protocol and device. If you do, quit and start over; it's difficult to regain control of the screen once a bazillion inappropriate mouse interrupts are being processed. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151
Re: Kernel panic
On Mon, 7 Apr 1997, G. Kapetanios wrote: Two days ago I managed to recompile my own kernel. It went roughly OK I only tried it on my office computer today though and there seems to be a problem. I seem to have chosen a wrong option since at a very early stage of booting I get the following error VFS: cannot open root device 03:03 kernel panic VFS: unable to mount root fs on 03:03 Perhaps you forgot to compile in EIDE support (if you have an EIDE disk, which most modern non-SCSI disks are)? -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151
Re: Problem on mounting CD...
On Fri, 11 Apr 1997 [EMAIL PROTECTED] wrote: I've been tried to allow the normal users in my system to mount a CD-ROM but with no success... I even put the following line in my /etc/fstab: /dev/cdrom /cdrom iso9660 ro,noauto,user,unhide 0 0 Where /dev/cdrom is a symbolic link to /dev/scd0. My /cdrom directory has the following permitions: drwxr-xr-x 2 root root 1024 Apr 11 10:23 /cdrom/ The problem is that the users can mount the CD but cannot umount it! The mount /cdrom works fine but umount /cdrom gives the error message: umount: /cdrom mount disagrees with the fstab Hm, strange. I've got the following line working in my fstab (it's not scsi though): /dev/hdd/cdrom iso9660 defaults,ro,user,noauto 0 0 permissions are drwxrwsr-x 2 root cdrom1024 Sep 14 1995 /cdrom/ Does this help? Yes, because you've put /dev/hdd in fstab. When you mount, the cdrom link is followed and mount mounts hdd (look at /etc/mtab). When you umount, it finds /dev/hdd in mtab, but /cdrom in fstab, and complains. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151
Re: Install report, finally
On Mon, 14 Apr 1997, Alexander Koch wrote: [...] 3. On the boot discs there were only /dev/sdb1 to /dev/sdb8, both 9 and 10 were missing. Well, it's not difficult to have 10 partitions if you a) only use two primary, one to be an extended partition with the rest being logical ones and b) if you're lucky and have a 4g hd. I had to mknod them and seem to have forgotten to do it on /target/dev/, too, what was further annoying. I reported this problem on 27 Jan when I got stuck trying to install from Debian 1.2. (This was on sda, but sdb was the same.) Strangely, Debian 1.1 had 16 partitions which is why I didn't get hit before. I don't think 1.2 had any ram disk devices either, so I couldn't rdev the root device on my custom boot/root disk at home (1.2) but only at work (originally 1.1). Looking at 1.3 of 4 Apr, it seems to have one ok (/dev/ramdisk0). So please, more SCSI partitions. Disks are getting bigger, and some of us need to run multiple OSes. And please don't say you should have more disks, not partitions; that's a decision made by our purchasing dept, not me. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: DEITY TEAM -- one comment
On Thu, 17 Apr 1997, Lamar Folsom wrote: Francois Gouget writes: Unfortunately in some cases it is not so simple to check for space availability as /var may be on one partition, /usr on another and /lib yet somewhere else. Should be doable. df to get all the partitions and their capacities, df /var, df /usr, etc to get the filesystems containing these directories, and a script to sort it all out. Does this mean that each package will have to list the space it requires in every directory and the packaging software will figure out if each of those directories is on a separate partition? It does this already, doesn't it. I can see all the file sizes when I browse a .deb file in mc. I have seen a user that had separate /var, /usr/lib, and /home partitions. Please figure out how much space to report for postgres and how it will be reported. df on its own produces a list of mounts and their available space which can be reverse sorted by mount. Given a file to install, a forward search for the first matching start of the path tells you which available space to decrement, because the most specific match will be found first. I think figuring out how much space is required for the installation may be more difficult than just doing 'df' on each partition and sorting it all out. This calculation seems to me to be just what computers were invented for. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Modem documents (was Re: Homer has the ppp blues)
On Fri, 18 Apr 1997, Mark Phillips wrote: On Wed, 16 Apr 1997, Rick wrote: Unfortunately my modem didn't come with a manual telling me these details. (I think I could send off for one which would cost me $30 - a little expensive I think.) Do you know what these things mean/do? Ie what are DTR/DCD settings? What does ATC1D2 mean and do? Is there any document on the net explaining basic modem commands and basically how a modem works. I really don't have a clear picture at the moment. It's worth trying the manufacturers' web sites, e.g. www.usr.com and www.nb.rockwell.com/ref for a start. But what I want is something that does the same for Fax, esp. Class 2.0 (which remarkably is not the same as Class 2!). Does anybody have any addresses? -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: loadlin impacts on W95
On Fri, 18 Apr 1997, Eugene Sevinian wrote: I hope someone can help me to fix this. Following the instructions provided by loadlin's manual I had mofified msdos.sys seting BootGUI=0 and Logo=0 Now I have windows producing : Missing or corrupted file: COMMAND.COM File creation error (or something like that ). However I can enter into W95 typing WIN but I don't think this is right. Well, if you have BootGUI=0 then it's correct to have to type WIN to get into W95. All BootGUI=1 does is to make the system 'type' WIN itself. (So if you create a C:\WIN.BAT file, WIN will run that file instead of running W95!) P.S. Here is my the config file : [menu] menuitem=W95, Boot Windows95 menuitem=LINUX, Boot Linux menudefault=W95,30 [W95] shell=win That doesn't look right. I have shell=c:\windows\command.com c:\windows /p /e:1024 in my W95 config.sys, which I assume is used for the DOS prompt. [LINUX] shell=c:\loadlin\loadlin.exe @C:\loadlin\linux.par [COMMON] rem THERE SHOULD BE NOTHING for COMMON rem move all you have to W95 part -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: LILO cylinder problem
On Wed, 23 Apr 1997, Rick Jones wrote: I don't understand why lilo is doing this. I've never had this problem before. I just repartitioned my drive and put linux on hda2. When I try to run lilo to boot linux from hda2 instead of hda1 it gives me this error about my cylinders. geo_comp_addr: Cylinder number is too big (2312 1023) I am aware that, for whatever reason, lilo won't look beyond cylinder 1023, the 540MB line that M$ drew out of ignorance years ago. (It might not have been Bill but I like blaming shit on him anyway.) Anyway. The point is that I took this into account and started hda2 at cylinder 817. I gave hda1 405MB, hda2 792MB, and hda3 is a 20MB rescue partition. So why in the hell does lilo report cylinder 2312 which I'm fairly sure is in hda3 when I'm setting up hda2 which starts at 817? A bit of back-of-the-envelope maths: hda1=405MB=816cyl, i.e. 2cyl per MB. Therefore hda2 ends at about 1200MB=2400cyl. I don't run large EIDE disks, but I think the entire partition has to come in under the 1023 limit. After all, you have no control over precisely where in the partition the kernel resides. I have used this same partition setup before without any trouble. Anyone have a clue? Perhaps the kernel was nearer the start of the partition. When is lilo going to recognise hard drives larger than 540MB? This is an ignorant limitation. Now that BIOS can read them when is lilo going to be updated? -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: year-2000 testing
On Sun, 27 Apr 1997, Sam Ockman wrote: Message from Bruce Perens ([EMAIL PROTECTED]): If you can do so, please try running your system with the date in the year 2000 for a while. Richard Stallman asked if we had tested that GNU software is free of year-2000 problems, and I think it's a good idea. It seems like the problems, if any, would probably be more problems from the actual roll-over from the year 1999 to 2000, so probably the best thing would be to set your computer's date to Dec. 31, 1999, and see what happens to your processes when it hits midnight. That might test the kernel and the hardware/firmware, but I think plenty of problems in programs only surface during the next millenium. For example, some misfeatures of my own software concerned machine generated filenames that should collate in date order, but I didn't include the century (back in 1986) because I was squeezing it all into DOS-compatible filenames (and didn't think the instruments concerned would still be running into the next century...). -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Directory permissions.
On Tue, 29 Apr 1997, Rick Jones wrote: As some of you are aware, I recently transfered my file system from hda1 to hda2. When I did it I used midnight commander. MC has an option to keep UID's and GID's when copying. Supprise! It only does this on the files. When it created the directories on hda2 it made them ALL root.root using umask. [...] 1. Is there an easy way to reset the permissions of my directories to what they should be? Such as a program that I can exicute that will set them or go through and prompt me for the changes. [...] So now my system is half-hosed. I assume there are other things not working correctly even though I'm not getting any other errors and all seems to be fine, so far. I did this to avoid a full install when I repartitioned since I have no backup system and have installed from ftp. Somebody out there must have a good way to reset my ownerships. Well, the /easiest/ way might just be to copy the files in one of the correct manners, overwriting both them and their ownerships. For example, tar to stdout and pipe it to stdin of another tar with --save-permissions and --save-owners (done as root). About a year ago I remember running across a doc that described the file system structure standards. I don't remember if it gave ownerships but it's worth a shot if someone knows what it's called and where I can find it. No, it's far too short to cover ownerships, only locations. The 1.2 FSS doesn't give it's own location(!), but only that of its FAQ which is ftp://tsx-11.mit.edu/pub/linux/docs/linux-standards/fsstnd/FSSTND-FAQ P.S. I don't see multiple copies of your postings here. -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: depmod error reading ELF header
On Wed, 30 Apr 1997, Dima wrote: The problem was avoided by removing the *_MODULES files. Didn't work here. What version of kernel/ld.so/(what else?) are you using? Dimitri When I compiled my first kernel, it was out of necessity for another machine, so I couldn't use the install parts of make. The installation had left these *_MODULES files on the first machine, and I fretted over how to get them into the right places on my target machine. Nowadays when I'm compiling kernels (on the target machines themselves), I notice that I never end up with any of these files in my /lib/modules. I can't find anything about them in all the documentation I've read. Does anybody know what they are, or where they are documented? -- David Wright, Open University, Earth Science Department, Milton Keynes MK7 6AA U.K. email: [EMAIL PROTECTED] tel: +44 1908 653 739 fax: +44 1908 655 151 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .