konqueror can't handle html ?
When I click on a html file in konqueror or try to use it as a browser, I get an error dialog saying: There appears to be a misconfiguration. You have associated konqueror with text/html, but it can't handle this file type. Hmmm ?
Re: apt-get, gnome-apt, dselect, which to use?
On 29-Sep-2000 Michael P. Soulier wrote: On Fri, Sep 29, 2000 at 09:54:29AM +0200, Preben Randhol wrote: Description: dpkg -s pacakge List all files:dpkg -L pacakge list all packages: dpkg -l So, does that mean apt-get's cache is shared with dpkg then? I thought they were separate, due to the apt-cache program. I'm actually a little confused about what dbs exist, and what is updated when I do an apt-get update. dpkg's database s in /var/lib/dpkg. apt has the package lists in /var/state/apt/lists, and the cache (the binary database packages) in /var/cache/apt. So if you do an apt-get update, only apt's lists are updated (unless you use the dselect apt method, which merges the apt database with dpkg's). With apt-cache you search the binary db of apt. dpkg knows about a package only if the package description is in his available list. Thus if a new package appears in apt's list, apt-cache will know about it, but dpkg won't (unless you install the package). Preben Vim user. No emacs installed. Here here. Vim rocks. Agreed :)
Re: superformat?
On 26-Sep-2000 Felix Natter wrote: John Hasler [EMAIL PROTECTED] writes: Michael Soulier writes: Seems it's still two steps, superformat and then mkfs to make an ext2 floppy. Just seems a little wierd seeing DOS as the default on a Linux manpage... FAT16 is a pretty good format for floppies (that's what it was designed for). Ext2 isn't. minix fs is smaller, so it is probably is a better choice than ext2 for floppies ? (although you can recover deleted (text-)data from ext2 rather easily: see the Ext2-Undeletion-HOWTO; I'm not sure whether that's possible with minix) you can also use (v)fat in combination with tar, but that will give you error messages like cannot chmod.. Unless you mount the floppy with the 'quiet' option.
Re: OT: IRC the ~
On 23-Sep-2000 William Jensen wrote: Running a vanilla identd is a Bad Thing IMHO. It helps attackers identify usernames and find out under what UIDs daemons are running (eg. if sendmail is running as root). Is there a secure way of providing the information an IRC server wants, while rejecting the cracker's attempts? I've pointed out oidentd in a previous post. It can be configured to return a random string instead of the username.
Re: Allow port 113? / IRC question
On 23-Sep-2000 Oswald Buddenhagen wrote: Should I allow packets coming into my port 113? there was a big discussion about this about half a year ago. maybe you want to look at the archives. i reject these packets and it works. if some server denies you access because of this, you may run some fake ident server - i don't know, if there is something like that around. There is oidentd: Description: Replacement ident daemon An ident (rfc1413) daemon for Linux. Oidentd supports most features of pidentd as well as a number of features absent in pidentd. Most notably, oidentd allows users, given the proper permission, to specify the identd response that the server will output when a successful lookup is completed. Oidentd also allows for pseudo-random strings (either a prefix, such as user, followed by a number between 0 and 9, or 10 pseudo-random characters of the set 0-9A-Za-z) to be returned upon the completion of a successful lookup instead of a username or a UID. . Oidentd now supports IP masqueraded connections, including netfilter.
RE: OT: IRC the ~
On 23-Sep-2000 William Jensen wrote: Is the ~ in front of your name in IRC due to the lack of auth installed? I cannot find identd anywhere with dpkg. I've tried dpkg -l|grep iden. When I had deb installed on another system I somehow picked up identd and I was not listed as ~ in name so that got me thinking. Anyway, what is the std debian auth package? -- assuming this is the issue? A standard ident daemon returns the owner of a specified connection. IRC servers use this to get your username (on your machine) when you connect to them. What they do if you don't have an identd installed is totally up to them -- I can't tell you that. Running a vanilla identd is a Bad Thing IMHO. It helps attackers identify usernames and find out under what UIDs daemons are running (eg. if sendmail is running as root). - $ apt-cache search --names-only identd pidentd-des - TCP/IP IDENT protocol server with DES support. oidentd - Replacement ident daemon midentd - identd replacement with masquerading support. pidentd - TCP/IP IDENT protocol server. - pidentd is the standard one
RE: ESS1868
On 20-Sep-2000 Tommy Wu wrote: Hi! I've a ESS1868 sound card on my box. But I can install the driver for this. It's always show something failed. Here is my syslog for isapnp and sb. === cut start === isapnp: Scanning for Pnp cards... isapnp: Card 'ESS ES1868 Plug and Play AudioDrive' isapnp: 1 Plug Play card detected total Soundblaster audio driver Copyright (C) by Hannu Savolainen 1993-1996 sb: ESS ES1868 Plug and Play AudioDrive detected sb: ISAPnP reports 'ESS ES1868 Plug and Play AudioDrive' at i/o 0x220, irq 5, dm a 1, 3 SB 3.01 detected OK (220) ESS chip ES1868 specified ESS ES1868 AudioDrive (rev 11) (3.01) at 0x220 irq 5 dma 1,3 sb: ESS ES1868 Plug and Play AudioDrive detected sb: Failed to initialize ESS ES1868 Plug and Play AudioDrive sb: 1 Soundblaster PnP card(s) found. YM3812 and OPL-3 driver Copyright (C) by Hannu Savolainen, Rob Hooft 1993-1996 Yamaha OPL3 at 0x388 === cut end === How can I do for this error ? I have an ESS1868 too, and it's working. I don't use isapnp, since I'm using the default card config. My sb options are the following: esstype=1868 io=0x220 irq=5 dma=1 dma16=-1 To check these, just run the config program of the card from DOS. What looks strange to me is dma 1,3. Why two 8-bit dma channels ?
RE: ssh, gethostbyname, and hosts.deny, oh my!
On 12-Sep-2000 kmself@ix.netcom.com wrote: [...] Sometimes ssh works. Sometimes it doesn't: [EMAIL PROTECTED]:karsten]$ ssh lists ssh_exchange_identification: Connection closed by remote host ...maybe 1 of 4 attempts succeeds. On the host, in /var/auth.log, I see: Sep 12 01:10:32 lists sshd[1884]: warning: /etc/hosts.deny, line 15: can't verify hostname: gethostbyname(140.208.171.207.in-addr.arpa) failed looks like a dns problem (?) Sep 12 01:10:32 lists sshd[1884]: refused connect from 207.171.xxx.xxx ...and looking at /etc/hosts.deny, we find at line 15: ALL: PARANOID The PARANOID option forces a dns lookup on the client. So if tcpd cannot look up your hostname, it won't allow the connection. (the only non-comment line in the file). There are no entries in /etc/hosts.allow. Questions: - Can I fix this by allowing SSH access in /etc/hosts.allow. I'm assuming yes and will try this. Yes you can. If you use only ssh, you could use ALL EXCEPT sshd: ALL - Why the periodic failure. If my address cannot be resolved, why should it appear to be resolving some of the time, but not always? I experience this problem too sometimes. Maybe some dns guru knows the answer. - Doesn this indicate a problem with the masquerading configuration (I'm not responsible for this)? Any further diagnostics to test this out? Dunno. I've never used masquerading. Thanks. -- Karsten M. Self kmself@ix.netcom.com http://www.netcom.com/~kmself Evangelist, Opensales, Inc.http://www.opensales.org What part of Gestalt don't you understand? Debian GNU/Linux rocks! http://gestalt-system.sourceforge.net/K5: http://www.kuro5hin.org GPG fingerprint: F932 8B25 5FDD 2528 D595 DC61 3847 889F 55F2 B9B0
RE: sound interrupts
On 10-Sep-2000 Michael Soulier wrote: Hey guys. How do I raise the priority of the interrupt that the soundcard is on so that music doesn't stutter during heavy disk operations? Use irqtune from the hwtools package.
Re: RW access to /dev/dsp /dev/audio /dev/mixer ?
On 01-Sep-2000 Paul D. Smith wrote: [...] pr I know but it seemed to be a different problem. I did a usermod -G pr floppy after I had done the audio and that removed me from the audio pr group. I misunderstood the functionality of usermod -G :-) A common mistake--that's why I said Note that if you use usermod you must list all the groups you want; this sets the list, it doesn't add to it. :) You could use adduser user group instead of usermod -G.
RE: Sending mail via exim: spooky....
On 26-Aug-2000 Andreas Hetzmannseder wrote: Dear debian-users I tried to send a test message to my friend who has an e-mail account on the same PC and from the same provider as me, let's say [EMAIL PROTECTED] This should have gone to my provider's smtp-server named smtp.netway.at. I edited just a little text using mutt, hit the key for 'send', the test message has been relaid in my outbox file, which is probably alright. But then I recognized quite a lot of modem activity, which went on for two or three minutes. Now I do have a really slow internet connection, but several minutes for a text of just two or three lines... It semed suspicious to me, so I turned it off. Well, the message hasn't arrived at our mailserver :( But what about the modem activity? This is really spooky Besides it seems, all that exim does in my case is deliver messages locally. /var/log/exim/mainlog includes the following: [...]unknown local-part friendsname in domain netway.at with netway.at being the 'visible mail name of my system' according to the following entry in my exim.conf: qualify_domain = netway.at Now I think that exim tried to deliver an error message to [EMAIL PROTECTED] according to my /etc/email-addresses: andy:[EMAIL PROTECTED] Like before /var/log/exim/mainlog said: unknown local-part hetzmann in domain netway.at Besides, I created this file for the purpose of handling outgoing mail, but now it looks like I can't receive system messages any more which used to go to /var/spool/mail/andy. Is my /etc/email-addresses OK? I think your problem is with the local_domains setting in exim.conf. Make sure you have qualify_recipient set to your hostname (not netway.at) and local_domains is commented out.
Re: Strange logon problem with CAPS. Please read you all!
On 26-Aug-2000 Kenward Vaughan wrote: [...] I stumbled onto this yesterday for the first time. While the login went OK (what happens with a mixed--UC and lc--password?) Nothing strange. Just type the password as you would at a lower case passwd prompt, and you'll be logged in.
RE: Anyone know who owns this gateway?
On 25-Aug-2000 Steve Lamb wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Sender: [EMAIL PROTECTED] That gateway is flooding the list with duplicates. Does anyone know who owns it so we can get ahold of them to tell them to shut down the gateway? I tried mailing to that address and it failed. Hmm, maybe it isn't niet.webforce.com.hk. Look at the headers: [...] X-Authentication-Warning: debian.org.hk: news set sender to [EMAIL PROTECTED] using -f Received: from GATEWAY by niet with netnews for debian-user@lists.debian.org (debian-user@lists.debian.org) Date: 25 Aug 2000 23:34:35 +0800 Message-ID: [EMAIL PROTECTED] Organization: debian.org.hk gateway - [...]
Re: Pointing Device
On 23-Aug-2000 Daniel E. Baumann wrote: Geez, maybe somebody should write a mouse howto or something. (just kidding isn't that what the XFree86-HOWTO is for?) There actually is a 3-button-mouse mini-howto. My mouse problem was that I set it up as microsoft in XF86Setup, and the middle button didn't worked. I read this howto and found out that the mouse is a dual protocol one (microsoft-mouse systems) and I have to set the ClearDTR option to switch it to mouse systems. So it might worth reading it.
RE: Utility for multiple floppies
On 22-Aug-2000 [EMAIL PROTECTED] wrote: Does anyone know of a utiltiy or group of utilities that can help me do the following: Take a large file (larger than 1.4 megs, the regular holding space of a normal floppy disk) from the windows os and split it into any number of desired smaller files so that the smaller files can fit onto multiple floppy disks, and then rebuild from the multiple small files the original large file onto a debian system? If you want to compress the file, you could use rar. If you just want to split, you could use file managers like Windows Commander, Norton C., or the djgpp port of GNU split. On the debian side you just cat the pieces back in one file.
RE: man is using gxditview
On 19-Aug-2000 John Bagdanoff wrote: When I use man as user, it is starting up gxditview, in console mode this is not nice. I noticed this behavior a couple of weeks ago, but as yet haven't found how to change it back to using less. Any ideas? As root run update-alternatives --config pager and choose less.
RE: X can't find the mouse!
On 18-Aug-2000 Trevor Ramoutar wrote: So I installed Debian 2.2. Everything went well but after the installation I'm at the prompt and I type in startx and I get: Fatal Server Error: Cannot open mouse (no such file or directory) and I don't know what to do! Help! You should configure X first with XF86Setup or xf86config. (and create a link in /dev called 'mouse' to your mouse device)
RE: rm user
On 19-Aug-2000 cls-colo spgs wrote: debs, what's the command for removing user accts? userdel
RE: xdm init level question
On 17-Aug-2000 Paul D. Smith wrote: I'm sure this has been discussed before (I have an uneasy feeling it may be a oh no, not this again question); maybe someone can put the rationale into a file in /usr/share/doc/ somewhere? I tried searching the list archives (user, x, boot, etc.) with various keywords (init, xdm, level, etc.) but came up empty. Why is xdm started at runlevel 2 in Debian? In all the systems I'm familiar with, both Linux and proprietary, xdm was always started at a runlevel after 3 (typically 5). This way you could boot the box into a non-graphical login merely by specifying a different runlevel. Is there some new standard that specifies a graphical login at runlevel 2 now? Why this difference in behavior? If there's some documentation somewhere that describes this I'm happy to go look there. Also, in this environment what is the correct way to boot into a console login and avoid starting XDM? I find it hard to believe the correct way is to edit /etc/init.d/xdm and add in an exit 0 to the top, which is what I've been doing :-/. If you had looked more carefully, you would have seen that xdm isn't started only at runlevel 2, it's started at every runlevel. The Debian Policy Manual says: By default `update-rc.d' will start services in each of the multi-user state runlevels (2, 3, 4, and 5) and stop them in the halt runlevel (0), the single-user runlevel (1) and the reboot runlevel (6). The system administrator will have the opportunity to customize runlevels by either running `update-rc.d', by simply adding, moving, or removing the symbolic links in `/etc/rcn.d' if symbolic links are being used, or by modifying `/etc/runlevel.conf' if the `file-rc' method is being used. So links are installed in every rc* directories, but you can create/delete them at your liking. (This is the Right Way(tm) IMO, because it doesn't force a particular runlevel config). Now if you want runlevel 2 console-only, delete /etc/rc2.d/S99xdm, or even better, replace it with K01xdm.
Re: non xfree-related question about installing xfree4
On 14-Aug-2000 Ron Rademaker wrote: The easiest way would be waiting for debian packages of xfree 4.0.1 ;) There are xfree 4.0.1 debian packages (very experimental, though) available at http://www.debian.org/~branden/.
Re: Problem with swapping - computer crash.
On 12-Aug-2000 Marko Cehaja wrote: (...) What happens by default when all memory and swap file is exhausted? I would like to know that. Must it be that system crashes or that it denies every use of my computer? At first, nothing. But when a daemon or other running program tries to malloc some memory, will die horribly because 'memory exhausted' errors are fatal ones, and the program can do nothing but exit. And when this happens to vital programs like syslogd, getty or init you don't have much choice. One solution could be enforcing some user limits with PAM. This way only processes owned by you will be at risk.
Re: /etc/environment
On 11-Aug-2000 Patrick Dahiroc wrote: which man page do i need to learn more about the LANG and LC_LANG variables? is there a comperhensive list of env variables? - man [1|5|7] locale. - the Locales chapter in the glibc manual Usable variables are in /etc/locale.alias and the directory names in /usr/share/locale (here are the locale data translations stored).
Re: Transferring files between a windows machine and a Debian Li
On 11-Aug-2000 USM Bish wrote: On Thu, 10 Aug 2000, Lehel Bernadt wrote: On 09-Aug-2000 USM Bish wrote: I am at this game for the last few years. There are two solutions to this: Solution 1: If you have multiple OSs (like me) with Debian 2.1, Slack-7, and Win-98, (and BeOS till recently): a) Create a separate partition, (Type of partition msdos or vfat). Use fips for non destructive re-partitioning of existing DOS partitions in case you do not want a ner installation. b) This will show up automatically in DOS/ Win as Drive X c) From Linux mount this partition, as umsdos, (otherwise, long file names will be juggered). Since the partition I use is hdb3, I mount it as follows: mount -t umsdos /dev/hdb3 /archive d) Having a separate partition has the advantage that all OSs which can access a MS DOS partition can be used for sharing data inclusive of mp3s, GIFs, HTML and other files. Solution 2: Just mount your DOS/ Win9x partition as msdos or vfat to a mount point of your choice. All long filenames in this case would be 8.3 format. Just suitable enough for transfer of data, but not installing. ^^^ Oh really ? 8.3 with vfat ? Unbelievable, but true. Just try it out, and you will see for yourself. Well, I don't know. Just look: - [EMAIL PROTECTED]:~$ grep win_1 /etc/fstab /dev/hda1 /win_1vfat gid=1001,umask=007,noexec,quiet 0 2 [EMAIL PROTECTED]:~$ ls -F /win_1 Jdk/artec/command.com home/ mix_95/temp/ My Documents/ autoexec.bat config.sys io.sysmsdos.sys util/ Office51/ autoexec.emu djgpp/ linux/nc/win386.swp Program Files/ bc/ drivers/ logo.sys rcs/ windows/ Wincmd/ bp/ games/ mirc/ recycled/ winmcad/ Now there are *definitely* some long filenames... Or maybe I misunderstood something ?
RE: sound card
On 11-Aug-2000 cam wrote: hello, I've tried unsuccessfully to get my Aureal sound card (8810 chipset) to work. When I attempt to install the drivers I get the following error: make install AUCHIP=AU8810 make[1]: Entering directory `/usr/src/au88xx-1.0.5' mv -f /etc/modules.conf /etc/modules.conf.old gawk -f mod_conf /etc/modules.conf.old /etc/modules.conf echo alias sound au8810 /etc/modules.conf echo alias midi au8810 /etc/modules.conf mkdir -p /lib/modules/2.2.15/misc cp -f au8810.o /lib/modules/2.2.15/misc /sbin/depmod -a /sbin/modprobe au8810 /lib/modules/2.2.15/misc/au8810.o: init_module: Device or resource busy Hint: this error can be caused by incorrect module parameters, including invalid IO or IRQ parameters /lib/modules/2.2.15/misc/au8810.o: insmod /lib/modules/2.2.15/misc/au8810.o failed /lib/modules/2.2.15/misc/au8810.o: insmod au8810 failed make[1]: [install] Error 255 (ignored) make[1]: Leaving directory `/usr/src/au88xx-1.0.5' If no one is able to help...can some suggest a good sound card that is widely supported under linux? I may just give up and buy a new card. Thanx. Just modprobing the module is often not enough. You should give the io,irq and dma of the card as parameters.
RE: Mutt locale settings [Was: Re: /etc/environment]
On 10-Aug-2000 Wouter Hanegraaff wrote: Since mutt was the only program that doesn't display these characters correctly, I searched the mutt mailing list archives. I still don't understand why unsetting LANG solves my problem, because the correct fix is to set LC_CTYPE to an appropriate value. I've no idea what the cause of your problem is, AFAIK when LANG is unset, programs should fall back to the default locale, which is 'C'. And on my box 8bit chars are displayed correctly no matter what LANG is set to. Locale settings are not the same as LANG settings, LC_CTYPE is for character classification, and LANG is for displaying programs in different languages. No, LANG is a general variable. The glibc manual says: If this environment variable is defined, its value specifies the locale to use for all purposes except as overridden by the variables above. (i.e. by the LC_* variables). So if you set LANG and don't set LC_*, they will get the value of LANG (check this with locale(1)). Mutt uses both, so when I set LANG=dutch, I get a dutch interface, and when I set LC_CTYPE=dutch, I get special characters displayed right (although just setting LANG=dutch with LC_CTYPE unset fixes the locale settings too, strange enough). As I've said above, this is normal. I've never been able to display extended characters with mutt. (I thought your problem was with xterm, that's why I said I had no problems). I tried LANG='' and it works, but I still prefer pine, having no problems with it at all.
Re: netscape security hole
On 10-Aug-2000 Marko Cehaja wrote: Dear On Tue, Aug 08, 2000 at 05:44:45PM -0400, David Teague wrote: On NPR's Morning Edition they described a security hole in Netscape versions 4.73 and earlier that allows 'infection' by access to 'nasty' web sites. It is said to put your hard drive at risk some way. I assume this is a Windows problem, BUT does anybody know what this hole is and whether Linux is susceptible? (Probably only the user's files would be at risk at worst.) That is a hole in Netscape SDK which it internally uses. What happens is simply that javascript executes (very fast and without notice) and Javascript ??? It's a java *applet* and it's available at http://www.brumleve.com/BrownOrifice/ it makes your Netscape a web-server. Your IP could be tracked down by the server where you got the javascript, and somebody else could browse Since it's executed by a (perl) cgi script, your adress can be taken easily from the cgi environment. through your files, and take informations. However, the hole is in the Netscape, they can't browse directories which are disabled to be readable by others. On *nix of course it can only access files that you have access to. Files could be deleted or read, if one set it up in that javascript. That hole in Netscape is not the hole in Linux or in Debian OS, because there are also other ways to intrude the system and see what is there. It is responsibility of the system administrator to ensure what kind of software does he install and if he can trust that company which made it. But anybody who has properly set up the ipchains, should be pretty much secure and imune to that. That java-web-server runs on some different port, so if you you know ports you allow to access and which services should run on those ports, even when you execute that javascript, nobody could access any of your files. The story is somewhere on /. Sincerely, Marko Cehaja
RE: /etc/environment
On 09-Aug-2000 Wouter Hanegraaff wrote: Hi, I was having trouble with the display of special characters in mutt when logged in using ssh, and I found that the LANG environment variable is the culprit. When I use mutt from an xterm, special characters like ë, é, § for example are displayed correctly. However, after doing ssh localhost from that xterm, special characters are displayed as '?'. diff-ing the environment settings showed that from a normal xterm, LANG is unset, while from a login shell (e.g. when using ssh) LANG is set to C, whatever that means. The only place I found LANG=C appeared to be /etc/environment, and changing this to LANG='' appears to have solved the problem. But several questions remain: Why doesn't display of special characters work when LANG is set to a value? When is /etc/environment parsed? I tried su - and xterm -ls, and LANG is not set. But when I login from the console or with ssh or telnet, it is. It's parsed when you log in (see /etc/pam.d/login) Why is LANG set to C from /etc/environment; which package puts this in? And what does LANG=C mean to programs in general? LANG=C is put there by the 'locales' package. It is used by 'internationalized' programs (for more on this read the locale(1,5,7) manpages and the 'Locales and Internationalization' section in the glibc manual). This interface is defined by ISO C, so the standard locale is called 'C'. I've no idea what the cause of your problem is, AFAIK when LANG is unset, programs should fall back to the default locale, which is 'C'. And on my box 8bit chars are displayed correctly no matter what LANG is set to.
Re: Transferring files between a windows machine and a Debian Li
On 09-Aug-2000 USM Bish wrote: I am at this game for the last few years. There are two solutions to this: Solution 1: If you have multiple OSs (like me) with Debian 2.1, Slack-7, and Win-98, (and BeOS till recently): a) Create a separate partition, (Type of partition msdos or vfat). Use fips for non destructive re-partitioning of existing DOS partitions in case you do not want a ner installation. b) This will show up automatically in DOS/ Win as Drive X c) From Linux mount this partition, as umsdos, (otherwise, long file names will be juggered). Since the partition I use is hdb3, I mount it as follows: mount -t umsdos /dev/hdb3 /archive d) Having a separate partition has the advantage that all OSs which can access a MS DOS partition can be used for sharing data inclusive of mp3s, GIFs, HTML and other files. Solution 2: Just mount your DOS/ Win9x partition as msdos or vfat to a mount point of your choice. All long filenames in this case would be 8.3 format. Just suitable enough for transfer of data, but not installing. ^^^ Oh really ? 8.3 with vfat ? And I think the author wants to transfer files between two machines, not two OSes on the same machine.
RE: [OFFTOPIC] BIOS Password defeat
On 08-Aug-2000 Daniel Reuter wrote: Hello there, I found a pretty nice 486 PCI-motherboard in the bulk waste last week, which I would like to use as secondary computer with debian. The board is working, but unfortunately, it was setup in a way that you can only boot from harddisk, and shadow RAM was enabled. So I tried to change the settings, but the preliminary user has installed a Setup-Password, so that I can't access the BIOS. I know, that there are ways to get around this, but I don't know how to do it in this special case. So does anybody know where to find the necessary information? Is there a tool for Linux or DOS to access and change BIOS-settings (I could plugin the harddisk from another computer and try to boot into Linux or use the small DOS-partition I have on this disk). Or is there some kind of cheat password, which will always work? The BIOS is a Phoenix version 4.04. You could write a little program or use dos debug to wipe out the cmos data at port 0x70 index, 0x71 data. (Containing something like: for (i=0; i0x80; i++) { outp(0x70, i); outp(0x71, 0); } ) In fact, it's only necessary to alter the checksum, and the bios should reset to the defaults.
Re: Cool trick: gmc and Debs
On 04-Aug-2000 Carl Fink wrote: On Fri, Aug 04, 2000 at 12:16:33PM -0400, Mike Werner wrote: You can also browse tarballs, gzips, bzips, gzipped tarballs, bzipped tarballs, and zips as well, providing that the attendant gzip, bzip, tar, and zip programs are all installed. Interestingly, though, it can't browse cpio archives. mc's vfs uses the scripts in /usr/lib/mc/extfs, so it can be easily extended to handle other formats too.
RE: autofs question
On 03-Aug-2000 Brian Stults wrote: Hello, I'm using the kernel-based auto mounter, autofs. I have all my mount points in the default /var/autofs/misc, and I have symbolic links to them in /mnt. However, whenever I do a listing of /mnt (either from an xterm, or from within an application such as StarOffice), all of the devices controlled by autofs are mounted. This is normal, since ls or soffice will try to readlink(2) the symlinks, thus getting autofs mount them. Is there a way to prevent this? If I want to access my CD-RW from within StarOffice, I would like to be able to go into the directory called /mnt and then go into the directory called /mnt/cdrw. But I would like to be able to do this without inadvertantly automatically mounting all the devices controlled by autofs. Any suggestions? If you stay with this configuration, you cannot change this behavior. The problem is that when you don't use links, you can't refer to the mountpoint while browsing from a gui, since this is created when you try to access it. Most of these programs however can use bookmarks for directory URLs too. So you can add the mountpoints to your bookmarks; under mc you could use the directory hotlist. HTH, Lehel
RE: FAT32 issues
On 01-Aug-2000 Matthew D Davis wrote: Hello all, I have my system setup to dual boot win2k and Debian 2.3. I would like to be able to access my large fat32 partition from both OSes. Whenever I mount the drive from the prompt or out of fstab (options are default,user, error=read-only) ownership by user/group depends on who mounts the drive. This is fine, and I can read data from the partition, but I cannot write even as root. When I attempt to write anything to the partition, I get an operation not permitted Needless to say, I'd like to write to the drive, because I spend most of my time in linux, but I would like to share my documents and mp3s and such with both OSes... You could use the uid, gid and umask options (see man mount).
RE: MC
On 31-Jul-2000 Christopher Clark wrote: As an ex Rred Hat user, my midnight commander left me in the current working directory when I F10 out of it. The Debian version dumps me back to the original directory. Is there any way to convert my Debian potato MC to the redhat style? regards Chris There is no RedHat style or Debian style mc. A unix program cannot change the working directory of his parent process, so mc can't change the working dir of the shell from which you've started it. To do the trick, redhat redefines the mc command as a shell function, which executes the real mc, then changes to its last working directory. This is wrong IMO, because the user should decide what he wants, not the package mantainer, and besides the whole thing can be found in the mc manpage, with the shell functions for bash,zsh and tcsh (look at the -P option). So if you're using bash, copy the appropriate function to ~/.bashrc. And you should RTFM before asking.
RE: dhelp problem
On 30-Jul-2000 Jeronimo Pellegrini wrote: Hello, I have used dhelp for a while in a standalone woody box, and all worked fine. Now I've installed apache and glimpse (to use the search feature)... And things seemed to stop working: After installing apache, glimpse, and reinstalling dhelp (probably not necesary, but I did it anyway), dhelp says I have no permission to access http://localhost/doc/HTML/index.html I suppose there should be a doc directory under /var/www right? No. The /doc is mapped to /usr/doc as specified in /etc/apache/srm.conf: - Alias /doc/ /usr/doc/ ## The above line is for Debian webstandard 3.0, which specifies that /doc ## refers to /usr/doc. Some packages may not work otherwise. -- apacheconfig - So check if /usr/doc/HTML/ exists.
RE: char-major-6
On 28-Jul-2000 Erik Mathisen wrote: I keep getting this stupid error in my syslog: modprobe: can't locate module char-major-6 now i searched my system, I dont have that module, how do I get this error to stop? It puts 2 or 3 entries in the log a minute. Any help would greatly be appreciated, Erik I've had the same problem, but with block-major-8 (scsi-disk). Since I don't have scsi devices, I appended an alias block-major-8 off to /etc/modutils/aliases and ran update-modules. After that I didn't get the message anymore.
Re: [why is kernel recompilation necessary?]
On 28-Jul-2000 Krzys Majewski wrote: Yes but what I'm wondering is not why linux users recompile their kernels, or why windows users can't, but how is it that windows users get away with not having to? The closest answer I got is that windows kernels have a bunch of drivers already compiled in, and any additional drivers compiled as modules. What I'm still not clear on is whether either windows or linux kernels (or both) need to have some sort of hooks enabling them to expect whatever modules at runtime, or not. For example, if I build a new device and write a driver for it, can I add support for this device to both windows and linux without having to modify either kernel? -chris First, I don't think you can compare the windows and linux kernels. Linux was designed as a monolithic unix kernel, while the windows kernel...uhm...I really don't know if it was designed at all... Second, what do you mean by modifying the windows kernel ? Disassembling and binary patching ? More seriously, the linux kernel has a module interface defined in module.h. Shortly, in your code you have to use init_module, cleanup_module, module_register_chrdev etc. As long as this interface doesn't change, you can compile your module separately and use it with any kernel version (unless you use something in your code that's specific to a certain version). If you want to know more, I suggest to read the docs dealing with the linux kernel at linuxdoc.org.
RE: cleaning up lost+found
On 26-Jul-2000 Olaf Meeuwissen wrote: Debians, I goofed up a little while ago and connected two SCSI devices with the same ID. I've cleaned up the mess, but now I have a number of files in /home/lost+found that I can't seems to remove. Not even as root! Typical output of ls -l on that directory looks like: b--swx1 494395209 172, 30 Jan 21 1978 #53340 brws-wxr--1 24240453 38, 90 Jul 7 1924 #53353 b--s---r-x1 110012127 185, 124 Jun 12 1945 #53354 s-ws--s-wx1 59857233924294967295 Mar 22 1991 #53364 c-w--w1 5551835812 72, 84 Jan 2 1970 #53375 c--Srwxr--1 4273961328117, 78 Oct 1 1922 #53402 s-wSrws--t1 2688 553694294967295 Nov 26 1973 #53422 c-x-w-1 1539229539115, 99 Sep 10 2028 #53443 br-Sr-Sr-x1 143738224 116, 104 Feb 11 1995 #53445 c---r-1 2489227504 32, 102 Aug 26 2027 #53448 c-x-w-1 1539229539115, 99 Sep 10 2028 #53475 c--xrwS---1 9285 9533 36, 87 Apr 8 1989 #53479 Totally weird, that is. I've moved the original and created a new lost+found with the same owner.group and permissions as the original but would like to get rid of the gunk above. Any ideas how? I've tried chown'ing and chmod'ing, but all I get is chown: filename: Operation not permitted Trying to rm -rf gives rm: cannot unlink `filename': Operation not permitted They probably have the immutable attribute set. Remove it with chattr.
Re: reread inetd.conf
On 25-Jul-2000 Alwyn Schoeman wrote: Hi Richard, The man page say you can send it a SIGHUP signal to have it reread its configuration file... I think its... kill -s SIGUP pid where pid is the process id you get by ps aux | grep inetd Wouldn't be more simple elegant kill -HUP `pidof inetd` ? On 25-Jul-2000 [EMAIL PROTECTED] wrote: or u can try /etc/init.d/inetd restart There is no need to restart, /etc/init.d/inetd reload will do it.
Re: terminal goes funky
On 25-Jul-2000 Sven Burgener wrote: On Tue, Jul 25, 2000 at 04:38:32PM -0400, Dan Brosemer wrote: It seems to display things fine, but just with weird characters instead of proper ones. Some characters are fine though, like the GNU part of the login prompt. :) Press ^V then ^O then ENTER at the prompt. The system will echo back: bash: ^O: command not found Yes! That's cool. Thanks. You've just got the system to send a ^O character to the screen which will reset it gracefully. Yes, but also ^V before the ^O, like you said. Do you know why? Anyone know why? Each unix terminal has two charsets: a normal one, with the standard ascii stuff, and a 'graphical' one, with those funny characters. The linux VT is no exception from this, with two charset slots called G0 and G1. Switching between carsets is done by sending the terminal the right control code. You can find out what these are by disassembling the corresponding terminfo database (found in /usr/share/terminfo) with infocmp(1). eg. 'terminfo linux linux.ti' for the linux console. The smacs code is for switching to the alternate charset, and the rmacs is for switching to the normal one. eg. in my linux.ti rmacs=^O and smacs=^N. The ^V character is a property of the tty line. Just do a 'stty -a' and look for the lnext code. This causes the next character to be sent literally to the terminal, even if it's a control character. So to switch to the 2nd carset type ^V^N, to switch back ^V^O. HTH Lehel
RE: combining unused space into files
On 25-Jul-2000 Richard E. Hawkins wrote: I thought I sent this earlier, but it doesn't seem to have arrived . . . IN my last-ditch attempt to recover a long text (actually lyx) file that was deleted, I want to combine all of the unused space on my drive into files. There's about 80M altogether, and I want to break it into 10M pieces (so I can manipulate them in the 40M partition normally assigned for swap). I've had pretty good success with mc's undelete fs. Or you can try the linux disk editor, lde. I want to get a literal straight read of the disk. Can anyone point out how to do this. dd if=/dev/[partition] of=[dumpfile] To complicate matters, the machine has no netwerok connectivity; it's a laptop, and the pcmcia hardware seems to have gone south . . .
Re: cannot talk with AT modem
On 24-Jul-2000 Tomasz Barszczak wrote: Well, too soon to be happy. cron runs periodically rmmod -a . Then when I want to use the modem again serial driver is re-loaded: Serial driver version 4.13 with no serial options enabled tty00 at 0x03f8 (irq = 4) is a 16550A tty01 at 0x02f8 (irq = 3) is a 16550A And the modem is inaccessible again. setserial shows that /dev/ttyS4 reverted to original parameters. I have to run the setserial command manually again. How can I fix this? You have two options: 1. compile the serial driver into the kernel. 2. create a file in /etc/modutils with post-install serial [your setserial command] in it, and run update-modules. (see the manpage for conf.modules and update-modules)
Re: PNP hardware and dual boot machine.
On 22-Jul-2000 adam b. wrote: isapnptools works okay, probably. I have gotten cards to recognize and load drivers, but I have never gotten them actually working before giving up. Be prepared to edit long config files from pnpdump and also you must know free IRQs, IO hexes, and Memory ranges for all your devices. They will probe, but he'll come up with like 5 or 6 viable configurations, only 1 or 2 of which might actually work (especially if you have a commercial machine-in-a-box). Also, you _may_ have to put your devices in Memory-Mapped mode (I've heard rumors to that effect). In this case, you have to find a free memory range that's in the on-limits range for Debian. Not too hard, but get the range from Debian.org before you go killing things. I have no clue about kernel PnP...I'd love to hear about it though! What kernel are you running? I have one PnP card (ESS1868 soundcard) and I *don't* use isapnptools. AFAIK you need to use isapnp in two situations: 1. when your BIOS doesn't support PnP (if you have a really old machine) 2. when you want to change the default configuration of the card The only thing I have to do is set the PNP OS installed option to NO in the BIOS setup, and to pass the same irq,dma,io to the sb driver that the card uses in windows.
RE: Help compiling Xlib
On 17-Jul-2000 Dinesh Nadarajah wrote: I am trying to compile a simple X application but I keep getting the folloeing error: gcc -g -o testapp test1.c -lX11 /usr/bin/ld: cannot find -lX11 collect2: ld returned 1 exit status Any suggestions? What are the Xlib libraries called under Linux. Thanks. You'll need xlib6g-dev.
Re: What drive is the dir on ?
On 08-Jul-2000 Charlie Kroeger wrote: If you're running a recent 2.2.x kernel and you still have one of those partitions empty, Does this mean the 2.2 kernel isn't confined to the 8gig limit of the earlier generations? What ??? 8 GB limit in the kernel ? Huh.. The 8 GB limit applies only to (mostly real mode) applications that are using the BIOS INT13 interface for hard disk access. Since the linux kernel communicates directly with the HD controller, there's no such limitation. The 8 GB problem is mostly a problem for boot-loaders (eg. lilo), beacuse they have to use the bios to load the kernel (so the kernel must be under 8g). Recent versions of lilo however can access the HD beyond 8 GB too.
Re: begone ctrlaltdel!
On 09-Jul-2000 Ethan Benson wrote: On Sat, Jul 08, 2000 at 11:43:44PM -0400, Bob Bernstein wrote: Suppose I want to dis-able the three fingered salute. Is it sufficient to simply comment out this line in /etc/inittab: # What to do when CTRL-ALT-DEL is pressed. ca:12345:ctrlaltdel:/sbin/shutdown -t1 -a -r now i am uncertain what control-alt-delete will do if you comment this out, it might revert to DOS/Windows behaviour of rebooting the machine uncleanly... No, it certainly won't. When this line is commented out, ctr-alt-del does nothing.
RE: Making new /dev entries
On 07-Jul-2000 Walter Williams wrote: So it sounds like there are files containing pre-defined information for making /dev entries. I would therefore conclude that it knows that lp3 would be a printer port and configures it accordingly. That helps immensely, thank you I would presume that to learn more about this subject I would have to delve into kernel construction. I should count my blessings that it was made this easy. Walt There are no such files, the information is in MAKEDEV, which is a script, so look into it ! Each device has a major and a minor number (just do an ls -l in /dev, these numbers are after the user group names). The only thing MAKEDEV knows is that lp3 is major 6, minor 3; the rest is on the parport driver, who registers itself as handler of major 6 devices. When a process opens lp3, the kernel sends the request to the driver who handles the #6 devices, and passes the minor number to it as an argument. Thus the parport driver will know that he was requested to open the fourth parallel port. HTH Lehel
RE: DRIVERS
On 06-Jul-2000 JOHN G BROPHY wrote: Hi. Can you please supply me with a web site address so that I can download a driver for ESS1688 PLUG AND PLAY. JOHN BROPHY These cards are supported by both OSS and Alsa. To use OSS, read kernelsrc/Documentation/sound/ESS and recompile your kernel. BTW: ESS1688 is not PnP. There are two cards, ESS1688 and ESS1868 which have the same chip, but 1868 is PnP and 1688 is not.
RE: permissions
On 15-Jun-2000 Johann Spies wrote: I have the following file in /lost+found: br-xr-srw- 1 4462 25959147, 103 Jul 25 2019 #1355791 Can somebody explain the following to me please: 1. The meaning of 'b' in the permissions. 2. The group end owner of the file 3. Why this file is so much different than the others I removed from the same directory? The abovementioned file refuses any chmod or removal even by root. It's different because it's a block device. Probably it has the immutable bit set, so use chattr to clear it, and then delete the file..
RE: dependencies rpoblem
On 12-Jun-2000 Chris Mason wrote: I'm trying to install postfix and remove exim, I am having a problem because apt-get remove exim fails because there are dependent programs. I don't care, I would like to force the removal and then install postfix but even apt-get -f remove exim fails. Is there a way to forcilble remove exim? Try dpkg --force-depends --purge exim.
RE: Documentation for the Xlib
On 05-Jun-2000 Sven Garbade wrote: Dear List, I want to port some old DOS-programs to Linux. The programs produce simple graphics like filled circles, moving rectangles etc. I´ve take a look in the xscreensaver sources, but I would find it very helpful, if there´s an overview about the xlib-functions. Is there any documentation about this? Install the xbooks package. The API description is in /usr/doc/xbooks/hardcopy/X11/xlib.PS.gz.
RE: Debian 'crashes'
On 02-Jun-2000 Michiel Meeuwissen wrote: Already several times happened to me the following: When I use a few memory eating programs like Netscape and dselect together, by whole computer freezes for a long time (10 minutes or so) and I don't see another way out then simply turning it of, because then I at least know what happens, and how long it is going to take. The freeze is not always complete. Moving the mouse can sometimes move the cursor on the screen minutes later. Probably it's not fair to call it a crash, but I do consider it like this. It seems that a way to accomplish this is running apt-get upgrade, netsape and seti at the same time, in my computer (potato, PIII 500 64 Mb). In the beginning sometimes I can see that 'xload' gets a few ticks, and the load is very high, like 20 or 30. I think the memory, including swap memory (128 Mb), is exhausted in such a case. It just happened again with acroread and netscape running. You should use ps, top or free to see what processes run, how much memory is used etc. Have you compiled the magic-sysrec-key facility in the kernel ? If yes, were you able to use it for emergency-rebooting ? So, I think this is very bad, since it should not be possible that one or two programs hang the whole computer. What can be done about this? Does there exist some 'memory quota' mechanism? I would e.g. like to see that netscape processes never take more than 50 Mb of memory. Or perhaps there exist some program which starts shooting of non essential processes (like those of *(@$(! netscape) when the load gets higher than 15 or 20 or so? You can force resource limits based on users/groups using the pam_limits.so module
OT: dial-up - leased line
Hi folks, I'm thinking on switching from a dial-up connection to a leased line. Is there anything special my modem (3com usr 56k ext) has to know to do this (ie. to deal with the differences between a normal telephone line and the leased one) ? TIA, Lehel
RE: Problem with rescue disks
On 28-May-2000 Shane Wegner wrote: Hi, I am attempting to modify the Debian rescue flopy to be an emergency recovery for my system. Actually it's root.bin which I am modifying. I am having problems adding utilities. I need to add the raid tools as well as restore(8). When I chroot to the floppy and try a restore, it give me this. restore: error in loading shared libraries: restore: symbol fchown, version GLIBC_2.0 not defined in file libc.so.6 with link time reference Now, libc.so.6 on my potato system is 800k, the one on the boot floppy is only 400k. Is this some sort of stripped glibc? If so, how can I compile additional utilities against it. I am looking to add restore, raidtools2, and agetty. Can this be done? The libc6 on the boot-floppies is from package libc6-pic: Description: GNU C Library: PIC archive library Contains an archive library (ar file) composed of individual shared objects. This is used for creating a library which is a smaller subset of the standard libc shared library. The reduced library is used on the Debian boot floppies. If you are not making your own set of Debian boot floppies using the `boot-floppies' package, you probably don't need this package. But I suggest using the yard package to create the boot-floppy set. And you can regain the space lost because using the standard libc by formatting the floppies at a higher capacity (eg.1743K).
Re: Another package installation problem
On 27-May-2000 Craig McPherson wrote: On Fri, May 26, 2000 at 09:42:02PM -0700, Eric G . Miller wrote: Update the dpkg package and that error should go away (though methinks there would be quotes around the string System V ...). Thanks... but I'm using the latest version of the dpkg package, as well as the latest version of the util-linux package. I even forced a reinstall of dpkg to be sure. Potato and Woody use the same versions of each package, so I assume it isn't a simple problem with the package file itself... it would have been fixed in Potato by now, I guess. Any other suggestions would be appreciated. I've looked in /var/lib/dpkg/info/util-linux.postinst (slink version): - install-info --quiet --section Development Development \ --description=System V interprocess communication facilities \ /usr/info/ipc.info.gz - So there _are_ quotes around System V You should edit the file, and re-run the installation (or just dpkg --configure --pending).
RE: Sound Question
On 25-May-2000 Jay Kelly wrote: I am trying to setup a SoundBlaster16 sound card, I have recomplied the kernel for sound and pnp support. Now I see the howto is telling me to: Use pnpdump to capture the possible settings for all your Plug and Play devices, saving the result to the file /etc/isapnp.conf. Choose settings for the sound card that do not conflict with any other devices in your system and uncomment the appropriate lines in /etc/isapnp.conf. Don't forget to uncomment the (ACT Y) command near the end. Make sure that isapnp is run when your system boots up, normally done by one of the startup scripts. Reboot your system or run isapnp manually. 1) after using the 'pnpdump' how do I save it to the /etc/isapnp.conf. 2) Where do I uncomment the (ACT Y) command? I didnt see it. 3) How do I make sure the isapnp is run at bootup? Thanks Guys Jay If you have a pnp-compliant bios, and you're happy with the card's configuration, you don't need isapnptools. isapnp is for initialising the card setting one of its configuration. But if your bios does that on startup, you can just pass the io,irq and dma to the sb module, and sb will recognize the card.
Re: Stuck...
On 21-May-2000 w trillich wrote: Martin Bialasinski wrote: * w == w trillich [EMAIL PROTECTED] wrote: BTW: ESC + 0 ^= F10 w hi, newbie mode here, how do you learn of those keystroke w equivalences? Hmm, somebody told me long time ago :-) Maybe it is a VT100 thing or such. none of the vt100's i got my feet wet on had any such keystrokes... my keymap on my telnet client is all screwy. keypad clear,=,/,* act as f1,f2,f3,f4, among other oddness... i'd be able to get a lot farther if i knew the cheat codes like esc-9 for f9... like alt-f2 to get to console 2. where's that documented? (not 'mentioned' but 'documented'. big difference.) anyone? The ESC+number=function key mapping is documented in the mc faq: - If all else fails you can emulate function keys by first pressing the ESC key and then one of the number keys. For example, if you want to produce F9, press ESC, then 9. If you don't have a ESC key on your keyboard you can try alt-9 or meta-9. - Mappings are done in the terminfo databases. You can disassemble them with infocmp. For explanation of the terminal capabilities, see the terminfo(5) man page. You can modifiy the disassembled one to suit your needs,and recompile it using tic.
RE: changing default shells
On 18-May-2000 Nick wrote: i installed Tsch how do i get back to BASH thankx Use chsh.
RE: trouble changing time regions
On 18-May-2000 Nick wrote: i am having a difficult time changing from CEST to PDT/PST i have tried the following date --set=Thur May 18 00:00:00 PDT 2000 but get Thur May 18 00:00:00 CEST 2000 can't get rid of it, it is stuck from the install! thankx for the tips You have to run tzconfig.
Re: Upgrading to potato
On 16-May-2000 Brad wrote: On Mon, May 15, 2000 at 09:51:19PM +0300, Lehel Bernadt wrote: No, dselect will not try to install everything, but I suggest to use apt-get only and not bother with dselect's apt method. It will give you a bunch of dependency problems, and screw up your existing configuration. Odd, since i use the dselect apt method all the time and it has never screwed up my configuration or given me major dependancy problems Do you mean while upgrading from slink to potato ? Since the dselect apt method uses apt-get to do the actual download and upgrade, dselect with the apt method should be just as intelligent as apt-get. Oh yes, but my problem is that if I use the apt method from dselect I won't be able to use the slink CDs (without changing the acces method, of course). So I have two separate available-packages databases: one for dselect(dpkg) slink, the other for apt-get potato.
RE: Upgrading to potato
On 15-May-2000 David E. Young wrote: Greetings. I am brand new to Debian, but a long-time Linux user/administrator (mostly RedHat). I've installed Debian 2.1 from CD onto my SPARC IPX; the process went flawlessly and the system is running fine. I would like to try an upgrade to the upcoming 2.2 release, primarily because I'd like the experience of an ftp upgrade with dselect/apt. I believe I've determined most of the proper steps; use dselect, choose the apt option during access, use the frozen distribution, etc. However, I'd feel better if there was a step-by-step guide to this process. For example, during the initial installation of 2.1 I chose one of the standard configurations, then augmented it with some additional networking stuff. I want to maintain this identical configuration yet upgrade to 2.2. Will dselect handle this for me, or will it try and install everything? Perhaps I should use APT instead (apt-get dist-upgrade)? No, dselect will not try to install everything, but I suggest to use apt-get only and not bother with dselect's apt method. It will give you a bunch of dependency problems, and screw up your existing configuration.apt-get is much easier and cleaner to use, and it will handle the upgrade intelligently (for documentation check the manpages of apt-get and sources.list).
Re: _X11TransSocketUNIXConnect: Can't connect: errno = 111
On Thu, 11 May 2000, Andrew McRobert wrote: hi when I run XF86Setup and then test my settings I get the following error: _X11TransSocketUNIXConnect: Can't connect: errno = 111 etc. etc. any ideas? Read the XFree86 FAQ ! The most possible case is that you've chosen the wrong refresh rates for your display.
Re: Install Linux on Win98 computer
On Fri, 05 May 2000, Jason Dodd wrote: Hello all, Newbie here, I've installed Debian at home and would now like to install Debian on a separate partition on my work computer. Win98 is currently loaded on the first 2 Gb's of the drive, the remaining 1 Gb is currently free space. I believe I can just install Debian and use Loadlin to boot my Linux, but I've been somewhat confused by the various references I see in the HOWTO's to needing to have Linux in the first 1024 cylinders. Is this limitation something I need to worry about, does Loadlin take care of this? (broad questions I know)? My computer is a Dell Pentium-Pro 180mhz with 128 RAM and a 3 GB drive. I wish to install the current version of Slink. Thanks alot for the help. If you have a 3 GB drive, just make sure it is set to LBA in bios and install. (the 1024 cylinder limit is 512 MB with normal mapping, 8 GB with LBA)
Re: Emacs - was Re: Mail/news software
On Thu, 04 May 2000, Brian May wrote: For me, the problem with Windows is you have to think when thinking should not be required. Take for instance, autoexec.bat. I know a Windows computer, that whenever it starts, it flashes up with the message Bad command or filename for a few seconds until it goes away. However, it doesn't give the important information: what command cannot be found? what line is it on? So, instead of going directly to the bad line (like you would for any Unix based interpreter), you have to do a lot of fiddling just to find out which line is bad. The problem is that win9x doesn't have the old ms-dos help.com (they probably thought that it would be too low-level for the real win user). You would have found all the answers there. In the Bad command or filename case you have to do command /y /c autoexec.bat which will step you through the batch file. I have had similar problems for out of environment space errors (I never remember or can find how to change it, although it seems to be fixed now) and programs that automatically add lines like: PATH %PATH;c:\newprogram which fails when %PATH% contains a directory with spaces (trial and error suggests that correct quoting helps). The environment size can be specified using the shell= command and the /e: switch in config.sys. In mine looks like this: SHELL=C:\COMMAND.COM /E:1024 /P for a 1k environment size (the default is 256 bytes). Perhaps Windows 2000 won't require autoexec.bat, I will believe it when I see it. However, I encounter similar problems throughout Windows (especially device drivers). That's because they are dos drivers, and for microsoft dos is dead (judging from the lack of documentation and support). So, the way I see it, with Windows you always need to be thinking There is a bug in this program. It won't say why it is crashing. What is the best work around?. With Unix, you get more descriptive feedback of what the program is doing (eg look at the output of dpkg), and I have never had problems with a device driver suddenly going broken, requiring a complete re-installation of the OS. You don't have to try and second guess what the computer might be trying to do. Oh yes, I agree that that one of the greatest problems of windows is that it tries to do everything in the background, hidden from the user, you can't see what's going on, and if there is a problem you can't solve it because you don't know what's happened. But all this has a very good reason. The windows philosophy is: Don't think, we will do everything for you, you will be able to use your computer without knowing anything about it. That's because windows is targeted to the don't-know-much-about-computers users (and this is a very large community), and wants to give them a power-user feeling. That's why the only problem-solving method on windows is reinstall everything. Any other method would require the user to think. (I don't say these users are stupid, they just don't know much about computers).
Re: kernel: attempt to access beyond end of device
On Tue, 02 May 2000, Johann Spies wrote: 2. Where can I find out to which device (if it is a device it refers to) a number like 16:05 in the error message above refers. I have looked at /usr/src/linux/Documentation/devices.txt, but do not get a clear answer on this. 16h=22, so look in /usr/src/linux/Documentation/devices.txt: - 22 charDigiboard serial card 0 = /dev/ttyD0First Digiboard port 1 = /dev/ttyD1Second Digiboard port ... block Second IDE hard disk/CD-ROM interface 0 = /dev/hdc Master: whole disk (or CD-ROM) 64 = /dev/hdd Slave: whole disk (or CD-ROM) Partitions are handled the same way as for the first interface (see major number 3). That means 16:05 is /dev/hdc5.
Re: samba's /etc/samba/cmb.conf -- empty?
On Thu, 27 Apr 2000, w trillich wrote: of course, since apt-get update / apt-get dist-upgrade didn't install everything (or download everything, for that matter) i don't know why i insist on thinking that apt-get install somepkg would include the pertinences relating to that package. it does install flaming dependencies, but that's all. plus, the doc/samba* directories had more than just copyright info in them, so i figured those meager docs were all there was available. so i'm back to asking 'how can we find out what packages are available without having to irritate the folks on this list?' When you do an 'apt-get update', apt stores the downloaded package-description files in /var/state/apt/lists. So when I install a new package I alway search for its description in these files to see what's in the 'Recommends:' and 'Suggests:' fields.
Re: syslog
On Wed, 26 Apr 2000, Andrew Kae wrote: Hello everyone, I have noticed that some of /var/log files such as syslog and lastlog files are getting very large ~ 18-20 MB. Is it ok to just delete them and recreate them? Yes, you can delete recreate them, but do not forget to send a SIGHUP to syslog after that (or run /etc/init.d/sysklogd reload). There is a cron script that rotates these logs, and you should really check why cron doesn't runs it.
Re: OT: Help! I need emacs html mode
On Thu, 27 Apr 2000, Maury R. Merkin wrote: I have almost everything important working now from my switch last week from RH to potato. While I can't say that it's been painless, I have learned some things, which, I s'pose is good for the soul, or something. So, one of my two major remaining problems is that I just need emacs' html mode 'cause I maintain some rather large Web sites and I won't know what to do without it. I think you need the psgml package.
Re: Long time Debian user with semi-technical question
On Tue, 25 Apr 2000, Andrew Weiss wrote: I have been using Debian since the old A.out days of Debian 1.0, and I was wondering since I am doing development of our company's own Linux distribution based on Debian... I have potato as the base and am beginning to customize and modify it, but haven't really yet... other than Windowmaker graphics and WDM graphics.. Is it OK to carefully modify /var/lib/dpkg/status? I've modified it once, don't remember exactly what, but it was a dependency conflict that I had to solve. KDE stuff wouldn't go on because it wanted stdc++2.9, and it was too stupid to realize 2.10 was newer... I had to force each and every kde package... it worked too, but I gave up on keeping kde because every time I wanted to add new packages it wanted to kill kde. Why don't you install libstdc++2.9? I have both 2.9 and 2.10 installed (with KDE depending on 2.9) and have no dependency problem.
Re: Long time Debian user with semi-technical question
On Tue, 25 Apr 2000, [EMAIL PROTECTED] wrote: If you need to modify dependencies, I mean really need to modify depenedencies, download the source and edit debian/control. In both your examples, there are better ways than hacking dependencies. Huh...download the source too ? You can edit the control file by unpacking the binary package like this: mkdir packagename dpkg-deb -X packagename_rev.deb packagename cd packagename dpkg-deb -e ../packagename_rev.deb Now it's all there. If you've finished editing, repack it: cd .. dpkg-deb -b packagename I do this when I have to fix the directory structure of alien-converted rpms.
Re: Linux Installation
On Sat, 22 Apr 2000, college pc wrote: Dear, Sir How are you ... I have 4 CD's , 2 Binary 2 source of Debian Linux 2.1r2 . I try to install it. 1- I ask about applications, are they included in these 3 other cd's or not.? The GNU/Linux operating sytem = the kernel + OS specific tools + tons of other apps. The 2 source CD's are not required for the installation though, they contain the sources of what is on the other CD's. 2- Until I complete installation nothing about ( please insert CD label 2 or 3 or 4 ) for what they? The Debian installation consists of 2 phases : first the base system (~25 M) is installed (from the first CD only), then after rebooting the other packages. You'll have to use the 2nd CD when dselecting the packages, in the Update the list of packages step. (you should read the dselect tutorial from the 1st CD) 3- From the beginning of installation, no way to changemind and exit or resume? If you want to exit from install, go to the reboot the system (or something like that) step.
Re: install help with a SCSI CDROM on a 486
On Fri, 14 Apr 2000, [EMAIL PROTECTED] wrote: I've been fighting this Debian install for 3 weeks and am about tired of it. If I don't get an answer here, then I am taking the damn CD to the skeet range. OK, here is what I got: Debian 2.1 bootable CD, official type that comes in the box, has the Debian tm logo on it and the O'Rielly cow book in the box with it Desired packages: Perl(full install not just the basic package), Python, gcc, apache/httpd, mySQL Compaq Prolinea 4/66 20megsRAM a 486 Fujitsu 4.3 HD partitioned as: hda1 := 2048 hda5 := 1024 copied the Debian CD here in /Debian using xcopy in DOS hda6 := 1024 copied W31 W95 here to make it easier to reinstall hda7 := 0033 Linux swap from previous install attempts hda8 := 0033 somehow generated in one of the install attempts, delete later Netgear/Bay Network EA201 D2 NIC Adaptec AVA-1505A hooked to an internal IBM/Matsuhita CR-503-C CDROM is at SCSI#3 during DOS bootup it shows the card as Host Adapter #0 Port 140h Interrupt 10 Host Adapter SCSI ID = 7 One of the references that I found is at http://customer.support.redhat.com/rhoaprod/plsql/xxrh_know_pkg.srch2?p_id=1 20 and its example was aha152x=0x340,10,7 however, I believe that I should have aha152x=0x140,10,7 due to the IO address Now I can click Win95 Start, Shutdown to DOS, go to the CDROM as I:\ in Win98DOS, cd install, invoke boot.bat, select color, select USA keyboard, partition hda1 as ext2~Write~yes,~Quit initactivate swap hda7, init /dev/hda1, skip the bad block check since it has been done about 20 times without finding anything, mount as /, Install Operating System Kernel and Modules, Select CDROM as medium, select /dev/scd0 : SCSI, and bleewy NO SCSI ADAPTER DETECTED The question remains, where do I tell Debian to aha152x=0x140,10,7 in this process?? I don't see a place to pass these parameters to the install program. What am I missing??? You don't have to tell it to the install program, you have to tell it to the kernel. When you boot the cd you get the kernel boot: line on the very first screen; you have to enter the params here. You can also get some help by pressing F1.
Re: xfstt question
On Sun, 09 Apr 2000, john smith wrote: I have managed to install xfstt and xfs since I can verify them using (both as root and user accounts) $fslsfonts -server unix/:7100 fslsfonts -server/:7101. /etc/init.d/xfs or xfstt restart works properly. my problem is when trying to verify them using xlsfonts | grep ttf so that the server sees all of them..but it does'nt see any fonts what could be the problem? Did you set up the xfstt font directory ? (or : have you RTFM ?)
RE: no wonder...
On Sat, 08 Apr 2000, Christian Pernegger wrote: What I don't like about apt/dselect is how they treat packages locally compiled from source tarball. I couldn't find an option to really ignore dependencies and do what I say. Specifically, if I want esound-alsa but have compiled the ALSA drivers/libs/utils myself, neither dselect nor apt let me install it because it depends on some ALSA packages. Now, there is a [Q] option explained in the conflict resolution screen which should retain the exact state I select - only it doesn't work as expected or even deterministically. 1) It drops me back at conflict resolution, with its suggestions selected again 2) The main menu appears. If I select install, it wants to remove all of gnome! Perhaps I could get the deb manually and install via dpkg and a few force options, but that's hardly optimal... whishlist A package state that tells the package managers that the functionality of this package is provided locally, treat it as if it was installed Why don't you debianize the package or create a fake one that provides the debian-package equivalent of what you locally installed ?
RE: no wonder...
On Sun, 09 Apr 2000, Christian Pernegger wrote: Why don't you debianize the package or create a fake one that provides the debian-package equivalent of what you locally installed ? Ah, I should have known there is a proper way to do this. This had briefly occured to me, but I've never ever created a package... any pointers besides www.debian.org/doc/packaging-manuals/packaging.html ? Something more in HOWTO style, perhaps? AFAIK there are 2 packages that can help you to create debs: dh-make and debmake. For dh-make there is a tutorial at www.debian.org/doc/maint-guide. For debmake you can read the docs in /usr/doc/debmake. However, these are developers' solutions - don't you think this should be possible on the user level? But how do we define user level ? If you had installed the system and you are managing it, then you aren't a simple user anymore. Besides, creating a package is not so hard as it seems. Or is there a script just to create such fakes? Have I just RT wrong FMs? The equivs package can help you to create these provides-only packages.
Re: qmail error
On Fri, 07 Apr 2000, john smith wrote: hi, when I try to install qmail after building it I get the error: green:/tmp/qmail# dpkg -i qmail_1.03-12_i386.deb (Reading database ... 28966 files and directories currently installed.) Unpacking qmail (from qmail_1.03-12_i386.deb) ... Performing install First installation of the Debian qmail package... Checking if qmail is already installed on this computer... no. Checking group qmail (gid 64010)... error! Group qmail has gid 70 instead of 64010 Checking user alias (uid 64010, gid 65534, homedir /var/qmail/alias)... error! User alias has uid 70 instead of 64010 blablabla... 8 entries have errors. Please correct these errors and reinstall qmail. dpkg: error processing qmail_1.03-12_i386.deb (--install): subprocess pre-installation script returned error exit status 2 Errors were encountered while processing: qmail_1.03-12_i386.deb what's this mean and how do I fix it? I looked in /etc/passwd and saw that it has users qmail,qmails etc.(with the uids that caused the error), despite that I don't have qmail installed. So you should remove them, as it seems that qmail recreates these users anyway.
Re: suidmanager - error
On Wed, 05 Apr 2000, Bill wrote: *** /etc/cron.daily/suidmanager: suidregister: /usr/lib/emacs/20.3/i386-debian-linux-gnu/movemail registered but not installed at /etc/suid.conf line 7 *** can someone please tell me how to remedy this? This is just a warning message: it means that there is something wrong with your emacs20 installation (or removal). If you don't have emacs installed the quick solution is to remove the line from /etc/suid.conf. However, to understand what's going on you should read the docs in /usr/doc/suidmanager. (I don't know what's the problem because the emacs postinst-postrm scripts should handle the registration correctly).
Re: zip-util w/ disk-spanning
On Wed, 05 Apr 2000, Vitux wrote: Hi deb's Anyone know of a good zip-utility that can span disks? I need to transfer some files (some of which are too big for one floppy) from my debian-box to my win31-laptop... thx Vitux I'd suggest rar.
Re: trying to install mozilla
On Tue, 04 Apr 2000, Attila Csosz wrote: I tried to run mozilla M14 but I got the followin error messages.. ./mozilla-bin: error in loading shared libraries: libstdc++-libc6.1-1.so.2: cannot open shared object file: No such file or directory Running locate: I have.. /usr/lib/libstdc++-libc6.0-1.so.2 /usr/lib/libstdc++-libc6.1-2.a.3 /usr/lib/libstdc++-libc6.1-2.so.3 How could I resolve this problem? Download the file Contents-arch.gz (where arch is your architecture) from your nearest debian mirror and grep for the missing lib. It will give you the package name to whom it belongs. Magyarul: toltsd le pl. az ftp.hu.debian.org-rol a /debian/dists/potato/Contents-i386.gz-t, es keress ra a hianyzo libre. Az eredmeny: ./usr/lib/libstdc++-libc6.1-1.so.2 oldlibs/libstdc++2.9-glibc2.1 Szoval installalnod kell a libstdc++2.9-glibc2.1 csomagot.
Re: redirect hidden output to screen...?
On Tue, 04 Apr 2000, Jeff Gordon wrote: I have a li'l program I wrote, Mail-Penguin, that retrieves mail for us from our POP box on the web -- it's a home-brewed 'fetchmail', sort of. Normally it sends informational messages to the screen -- but now I have it being launched by diald's ip-up script (which I guess is sending output to its FIFO...?) How can I restore or mirror that output to the console screen...? I had the same situation a few days ago. I was writing a ppp-accounting program and printed with fprintf(stderr,...) to the terminal. But when the program was started by pppd, there was no output at all. I resolved the problem by using syslog(3) instead of fprintf().
Re: Make/makefiles
On Sat, 01 Apr 2000, Bart Friederichs wrote: Hi, Can anybody tell me how to make a *simple* makefile? Or a URL where I can find more info? The manpages aren't very clear, and *way* too extensive. I have a small project with 3 source code files, and I don't want to compile all of them every time. Check the make-doc package.
Re: Boot disk quit working after second install
On Sat, 01 Apr 2000, Erik Ryberg wrote: Lehel Bernadt wrote: What kind of boot disk do you have : ext2+lilo, fat+syslinux, loadlin or just the raw kernel ? What is its configuration ? I understand that you use the same disk for debian suse. How did you altered the config to boot suse instead of debian ? U. . .. I don't know. It was a boot disk I made during the install. I seem to recall seeing loadlin come up but I confess I didn't pay that much attention when it was booting. When I booted my computer some text would come up with the word boot and almost immediately it would say booting the kernel or something like that, then it would scroll through the various drives and so forth and then I would be in Debian, and it would ask for login password. If this is the disk you've created when installing debian, then it is syslinux. You can check the disk from dos (because it's a fat fs). There should be a file named ldlinux.sys. The config file is syslinux.cfg. It has a line like : APPEND root=/dev/hda2 ro This tells the kernel from what partition to mount the root fs. If you haven't modified the config file, I really don't know how gets windows loaded. Maybe you should check the disk with scandisk.
Re: unlink or kernel error?
On Sat, 01 Apr 2000, Tadeusz Bak wrote: Hi all, Sorry, my problem is probably not Debian specific... I tried to run under slink with the 2.0.36 kernel the following program: #define RUN /tmp/.rem.run #include stdio.h #include sys/types.h #include sys/stat.h #include fcntl.h #include unistd.h main() { open(RUN, O_RDWR|O_CREAT, S_IRUSR|S_IWUSR); printf(File created\n); sleep(30); unlink(RUN); printf(File removed\n); sleep(30); printf(Finish\n); } After the message File created I can see that there is a file .rem.run in /tmp directory. After File removed that file disappears. But according to unlink man page it should exist up to the message Finish because it is not closed. Only after terminating the program kernel closes all open files and .rem.run should be removed. Does anybody have an idea what's going on here? Thanks! The thing that happens is that the file name is deleted from the directory, but its inode is kept until the program closes the descriptor.
Re: makeinfo
On Fri, 31 Mar 2000, Guilherme Soares Zahn wrote: Hi there, I was experimenting some new things here and I decided to try to build binary packages from the source ones (to get benefit from compiling them with pgcc instead of the regular gcc) and, although I suceeded to do that with some packages, others failed claiming they could not find the makeinfo command... I tried to browse the /var/state/apt/lists dir but couldn't find out which package to install in order to have this command available... Can anyone help me? TIA, Guilherme Zahn makeinfo is in package tetex-bin
Re: Boot disk quit working after second install
On Fri, 31 Mar 2000, Erik Ryberg wrote: Hello, I have a 6 gig hard drive with debian on the first 4 gigs. I installed an old SUSE last night on the last 2 gigs in order to try out MySQL which allegedly came with that distribution. I boot debian with a boot disk as I have, sigh, Windows on another drive and I had a scary experience once with LILO that I don't care to repeat. Now I can't boot anything on my Linux drive. My boot disk (and backup boot disks) runs in the A drive but windows 95 starts right up each time. What kind of boot disk do you have : ext2+lilo, fat+syslinux, loadlin or just the raw kernel ? What is its configuration ? I understand that you use the same disk for debian suse. How did you altered the config to boot suse instead of debian ? Seems like probably whenever I attempt to boot it tries and fails to startup SUSE, perhaps because it is too far out in my hard drive? When you use a boot disk, there should be no concern about a too big drive (the INT13 limit applies only when booting the kernel from the HDD)
Re: kde Window Manager
On Tue, 21 Mar 2000, Oswald Buddenhagen wrote: i am wanting to start kde as my default x-windows manager i am going to edit my ~/.xinitrc with the following exec windowmanager currently set to wmaker what do i specify for KDE? kwm That's a wrong answer. You should do this : [EMAIL PROTECTED]:~$ which kde /usr/bin/kde so the .xinitrc line shoud be : exec /usr/bin/kde kwm is only the windowmanager part of KDE [EMAIL PROTECTED]:~$ cat /usr/bin/kde #!/bin/sh # # DEFAULT KDE STARTUP SCRIPT ( KDE-1.1 ) # # initialize the configuration first. kcontrol -init # Start the common desktop tools in the background. # The sleeps reduce disk usage during startup. # kaudioserver will put itself in the background automagically kaudioserver (exec kwmsound) # Add -ncols 96 after kfm if using a 8-bit display (exec kfm) (exec krootwm) (exec kpanel) (exec kbgndwm) # finally, give the session control to the window manager exec kwm As you can see, /usr/bin/kde executes some other progs too (like kfm or kpanel), which are essential parts of KDE.
Re: Making the Windows key work
On Sun, 19 Mar 2000, Eric Hagglund wrote: Can someone explain (or point me to the correct documentation for) the procedure for making the Windows key pop up the start menu for the kde, fvwm or icewm desktops? To do something useful with the winkeys you must ensure that they are mapped to some keysyms and make the app bind some action to the keysym. I don't know why are you asking about IceWM, cause the winkeys are working in it ( the left is used for the start menu, the right for the window list). Is your keyboard set up correctly in /etc/X11/XF86Config (like the section below)? Section Keyboard ProtocolStandard XkbRulesxfree86 XkbModelpc104 XkbLayout us EndSection I don't use any of the Fvwm variants, maybe you can do something in the conffiles ? As about KDE, it is fairly easy. You can try this : Start up xkeycaps (an excellent program for editing keysyms); select the 104-key layout on startup; choose which one of the 2 winkeys do you want to use for the start menu (I use left in this example); change its keysym from Meta_L (the Meta_x is used in KDE for going to the application menu, so we must choose a different one) to something unused, let's say Super_L; save the keymap and exit xkeycaps (don't forget to rename the keymap file from ~/.xmodmap-hostname to ~/.xmodmap); go to K - Settings - Keys - Global keys; for 'Pop-up system menu' choose custom key, click on the key field an press the left winkey; press Ok. That's all !
Re: Is the following part of /etc/security/access.conf?
On Sun, 19 Mar 2000, Shaul Karl wrote: I have managed to mv files that I did not meant to. Can someone tell me if the following is part of /etc/security/access.conf? [03:49:10 /tmp]$ cat /etc/security/access.conf #%PAM-1.0 auth required pam_unix_auth.so accountrequired pam_unix_acct.so password required pam_unix_passwd.so sessionrequired pam_unix_session.so [03:50:18 /tmp]$ -- Shaul Karl [EMAIL PROTECTED] An elephant is a mouse with an operating system. This is definetly not /etc/security/access.conf. It's a PAM access file (probably /etc/pam.d/other).