RE: DepenGNUian Logo
Clemmitt Sigler wrote: In article [EMAIL PROTECTED], [EMAIL PROTECTED] writes: daniel == Daniel Robbins [EMAIL PROTECTED] writes: OK, ok, the penguin's a jewel thief, and he's evil. But he has funny little beady eyes, and that's got to count for something. I think Gromit would be a better mascot. Ultimately he was much more resourceful and capable than de penguin. We may be onto something here! Maybe Debian can be thought of as the distribution that zags (Grommit) when everybody else zigs (penguins). Any comment? Only problem (besides copyright issues) is the obvious negative association `Debian: runs like a dog'. ;-) Casper Boden-Cummins. -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
RE: Tips
Jonas Bofjall wrote: On Mon, 20 Jan 1997, Niels wrote: Those anyone collecting Tips for debian? This is a good one for it... Another good one is if e2fsck says it cannot read the superblock of your Since it seems like nobody else does, I have noted these. still think that the Microsoft-like tips is a good idea, it gives the newbie some good information that normally is hidden inside the deep deep manuals. Like the 'find' command, terrific utility, always what the newbie wants. I've been meaning to look at fortune to see whether it can easily handle or be adapted to handle different databases of fortune cookies - such as tips. I'm running low on free time just now, so does anyone have the answer to this? Casper Boden-Cummins. -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Debian logo submissions
A friend of mine has donated a logo to the Debian project for consideration. It's sketchy, but we think the idea has some potential. Only thing is, where do we send it? If anyone's interested, there's a copy at www.wollery.demon.co.uk/penguin.gif. Thanks, Casper Boden-Cummins. -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
RE: cron.daily et al.
Jan Camenisch wrote: There might be a problem with the execution of the cron.daily, cron.weekly, and cron.monthly : An machines that don't run all day, these cron jobs get rarely executed. For instance, I usually use my maschine only in the evenings at home (i.e. later than 6 pm). But all cron get executed at 6pm. Therefore these cron task get never executed. How about a (cron) job, that executed every time the machine gets booted and that checks when the cron jobs were executed for the last time. If these for were not executed for say two days (weeks, months) then they get executed regardless the actual hour, day, week of month. I heard of a program called anacron to solve just this problem, but I don't think it's available as a Debian package, yet. Maybe someone here knows where to get it? Otherwise, you could try comp.unix.admin or gopher. Casper Boden-Cummins. -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
RE: novice questions
Thomas Kocourek wrote: Larry Clayton writes: [snip] 3. When I call man, for example man 9wm, I get the response, What manual page do you want from section 9wm? So I try man 9wm.1 and get the same response. What is the appropriate answer to such a question? It looks like it's interpreting the 9 of 9wm as the manual section number. I'm sure this shouldn't happen, but you might try giving it a section number explicitly with: man 9 9wm HTH, Casper Boden-Cummins. -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
RE: mime and elm (was RE: X-wm question and ZipDrive)
Hamish Moffat wrote: Good. Any chance you could not send all messages as MIME, either? Real PITA to read with plain jane elm on a character terminal. Couldn't you pre-filter your email with procmail and a MIME extraction program? Maybe the packages mime-support (which `can be used to turn virtually any mail reader program into a multimedia mail reader') or mpack? I'd love to. But this machine is not a Debian box, it's my account at university running Solaris 5.5.1, and there's no procmail, or munpack, etc. Only metamail, which isn't very friendly. Unfortunately, I doubt my disk quota runs to a permanent copy of procmail, which sounds quite featureful and therefore probably quite large. If you're using OpenWindows, the standard mail program `mailtool' understands MIME. (Couldn't your admins install a system-wide copy of procmail?) After all, MIME is so well established and you're imposing the lowest common denominator on us. True, but I see no advantage in sending absolutely plain text messages as MIME when some people (such as me) will complain. When attachments are involved, I agree, MIME simplifies things significantly and metamail handles this adequately. Although I still use Netscape when I'm trying to send file attachments. I agree with you and Dale on this. I assumed we were talking about uuencoding `attachments' instead of MIMEing them - having a MIME-compliant mailer, I'm not aware of the extent of the problem. Casper Boden-Cummins. -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
RE: PnP modem under Debian
Jason Goldschmidt wrote: Hi, does anyone have any experience getting a US Robotics Sportster 28.8-33.6 PnP modem to work with linux? Or any PnP modem for that matter. I've done all the common setup stuff for the modem. I found that if I want to use the modem under NT, I have to disable PnP in my bios. But if I do this for linux it changes nothing. -Jay I recently installed one of these and had a problem with an IRQ conflict. Maybe this is not the cause your problem, but it might be of use. I apologize that some of the details may not be correct, 'cos this is from memory. I'm using a kernel compiled from the 2.0.27 sources, with PCMCIA modules compiled from the 2.0.27 (must be same version). The problem was that the default IRQ mask used by the low level PC card driver module (called i832656, or something similar) allowed it to probe the IRQ assigned to my PS/2 device. According to the man pages, this driver can't detect conflicts using the controllers in my machine (a Vadem), and it nabbed the IRQ. I read lots of man pages, and found that you can pass an IRQ bitmask as an argument to the card driver. I tried doing this in /etc/modules.conf, but it didn't seem to work. Eventually I found that adding `irq_mask=0x400' after CORE_OPTS in /etc/pcmcia.conf did the trick. I must confess to not having read the PCMCIA HOWTO. I still have a problem where the PC card device file /dev/ttyS2 appears busy all the time, but /dev/cua2 doesn't. I think I'll go read it. HTH a little, Casper Boden-Cummins. -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
mime and elm (was RE: X-wm question and ZipDrive)
Hamish Moffat wrote: thanx to U all, who helped me. Debian rules...:-))) I deactivated html mail, I hope that now no html mail will be sent to the mailing list, BTW how can I find out about this? Good. Any chance you could not send all messages as MIME, either? Real PITA to read with plain jane elm on a character terminal. Couldn't you pre-filter your email with procmail and a MIME extraction program? Maybe the packages mime-support (which `can be used to turn virtually any mail reader program into a multimedia mail reader') or mpack? After all, MIME is so well established and you're imposing the lowest common denominator on us. Casper Boden-Cummins. -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
RE: How to copy Debian?
Leander Berwers wrote: Situation: Next week, I have to install a few pc's with Debian. Since I do not have the CD with 1.2.1 (and my pc's do not contain a cd-rom player), I would like to do it via ftp. However, to off-load the site, I would like to copy it once to a local Windows machine with will act as a ftp-server. Question: How can I copy the site without copying on a file-per-file basis. I tried Windows-ftp and CuteFTP, but a mget *.* does not work, the source ftp server does not respond (because of softlinks?)! You could try WFTPD (Winsock FTP Daemon) by Texas Imperial Software. I don't know whether it resolves links properly, but it's small and easy to set up. A fully-functional shareware version (about 256kb) is available from: ftp://ftp.iquest.com/pub/windows/papa/daemons/FTPD/wftpd233.zip Alternatively, why not install one Debian machine and ftp from that? Good luck, Casper Boden-Cummins. -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
RE: How to copy Debian?
Darren Klein wrote: Doy uo know of a WORKING nfs site for a Debian installation. I have been trying to find one.. but getting connection refused. Thanks. | Darren Klein | Internet Service Providers | (718) 962-1725 | | [EMAIL PROTECTED] | [EMAIL PROTECTED]| (718) 962-1708 fax | | New World Data |http://www.nwdc.com | Telnet: nwdc.com | Sorry, Darren, it isn't something I've ever needed to look into. Does anyone on the list have a list of reliable NFS-mountable sites? Casper Boden-Cummins. -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
RE: How to copy Debian?
Leander Berwers wrote: I mean how can I copy from a mirror site to my Windows machine. I do have a ftp server on Windows. Have a look at the mirror package (optional, in the net category). From the Packages file: Description: Perl program for keeping ftp archives up-to-date. Mirror uses the ftp protocol to duplicate a directory hierarchy between the machine it is run on and a remote host. It avoids copying files unnecessarily by comparing the file time-stamps and sizes before transferring. Amongst other things, it can optionally gzip and split files. It was written for use by archive maintainers but can be used by anyone wanting to transfer a lot of files via ftp. HTH, Casper Boden-Cummins. -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: How to copy Debian?
Leander, I said: Have a look at the mirror package (optional, in the net category). From the Packages file: Sorry: forgot to add that I have no idea how you can do this purely from Windows. I can only suggest getting a minial Debian machine with networking running, and controlling your operations from there. Casper. -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
RE: Keyboard with X
Alexandre Lebrun [EMAIL PROTECTED] wrote: I tried LyX and sometimes run Netscape, and they show the same problem : when trying to erase a character with backspace, the current character is erased instead of the preceding. ( just as DEL does in a Microsoft environment). It's very disturbing, and other apps don't have this problem (they treat both DEL and Backspace as 'backspace' ). Is there a cure ? I cured this for all my apps by running the following in .xinitrc xmodmap -e Keysym Delete = Backspace This is off the top of my head (if only we had a real OS in the office :-( ), so you may have to check out the man page for xmodmap. If it doesn't work first time, trying swapping the words 'Delete' and 'Backspace'. Casper Boden-Cummins, [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
RE: Problems with most in xterm
[EMAIL PROTECTED] writes: I have found the problem, it was the environment Variable TERMCAP it was set to: TERMCAP=co#80:li#24: as soon as I remove this TERMCAP variable all is ok. But who is responsible in setting TERMCAP? It seems that xterm (and xterm_color) are doing that. It doesn't matter if termcap-comp is installed or not. The TERMCAP setting is compiled into xterm. I don't know whether this has been addressed yet, but you could always recompile the sources. Casper Boden-Cummins. -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
RE: netscape experience?
Robert Nicholson [EMAIL PROTECTED] wrote: Hello, I run Netscape 3.0 on Solaris x86 2.5 over a DP 4.0 based PPP link and let me tell you it crawls on my 288k modem. So much so that I seriously suspect a problem either with dp 4.0 or solaris. Basically NS 3.0 continually stalls whilst it's downloading. I don't know wether this is due to a bandwidth problem on my host or at my end but usually this is the only connection I've got open and it's nothing like win32 performance for this application. Perhaps it's a generic unix Netscape thing? About a fortnight ago I saw a discussion on comp.unix.solaris about PPP on Solaris. (Not sure if this was Intel- or Sparc-specific, or generic. Wasn't paying much attention at the time.) Apparently, Sun plan to give PPP a major overhaul in the next release. I believe this will be a 2.5.x release, rather than 2.6. I run Netscape 3.0 on Solaris 2.4, on a Sparc with a 28.8 modem. I don't have stalling problems, but I can't say that PPP exactly flies along. Do you know how to benchmark it? Casper Boden-Cummins. -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
RE: Xterm and TERMCAP
Zlatko Rek wrote: I have problems with xterm. When it is opened the TERM and TERMCAP environment variables are set to: TERM=xterm TERMCAP=co#80:li#24: I use Jed editor and if I want to edit file the Terminal not powerful enough for SLang. message appears. When TERMCAP is unset Jed works OK. TERMCAP can be set to either the name of a termcap file (a file containing termcap definitions for many terminals), or a literal definition for a single terminal. In your case, TERMCAP is defined for very nearly the dumbest possible terminal definition. The fact that Jed works with TERMCAP unset means that it is successfully picking up the system-wide termcap file. You need to find out where TERMCAP is being erroneously set and unset it. It might be in your .bash_profile (or other profile, depending on your shell), or possibly in a system-wide set-up file, such as /etc/profile. Another problem is that command history keys ( up/down arrows) dont work. Next or previous command is appended to the current instead of replace it. The correct behaviour depends which shell and editing mode you're using, and whether your particular terminal type is supported properly. Please give details. Casper Boden-Cummins. _ _ Dr. Zlatko Rek | Phone:+386 61 177 3746 Jozef Stefan Institute | +386 61 177 3900 National Supercomputing Centre | Fax: +386 61 219 385 Jamova 39, P.O.BOX 300 | E-mail:[EMAIL PROTECTED] SI-1001 Ljubljana, SLOVENIA | http://www2.ijs.si/~rek/ _ _ -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
RE: Getting config from kernel
Joe Emanaker said: A file called '.config' is generated at the top of your kernel source tree This file will still be there after you've compiled the kernel, so all you need to do is copy it to a safe location and drop it back in when you want to generate the same kernel. Which only leaves the question: where can I get the one that was used to generate the kernel in the current kernel-package? You've probably guessed it by now: there isn't one because the initial kernel's precompiled. I don't know of a definitive way of recreating such a file. I can suggest only that you look at what hardware's recognised and which daemons are started at boot-time (run /etc/dmesg), and look around /proc for recognised devices and interrupts. Does someone on the list know a better way? Casper Boden-Cummins. -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
RE: Getting config from kernel
Joe Emenaker wrote: Okay, I sent this to the list once already, but I'm not sure it got through. I was hoping someone else would answer this, because I only have vague details in my head. (If we only used Debian at the office...) Can't help with the IP aliasing, but: So, I've resigned myself to recompiling the kernel. However, I'm worried that, during the QA of make config, I'll accidentally say no to a driver that I need. The way I'd really like to do this is to be able to somehow extract or ftp a copy of the config file used to make the stock debian kernel. 2 - Where is the information generated by make config kept? For example, suppose I forgot to compile in *one* crucial driver, and I don't want to go through the grueling QA that make config puts me through. Where can I make that one-line change and then go on to make depend, etc.? 3 - Is there any way I can extract or obtain a copy of what config was used to generate the 2.0.6 kernel that I'm using? A file called '.config' is generated at the top of your kernel source tree, which contains a list of all the selected compilation options from make config, make xconfig or make menuconfig, in the form IP_ALIASING=yes BLAH_BLAH=module FOOBARWIZ=no This file will still be there after you've compiled the kernel, so all you need to do is copy it to a safe location and drop it back in when you want to generate the same kernel. Casper Boden-Cummins. -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
RE: Help with diald?
Hi, Yves Arrouye said: I'd like to know, too, how I can arrange things so that when I mail to someone outside my domain diald does not try to dial. If you're using sendmail, you can instruct it to queue your mail and attempt to deliver it next time you connect. Local mail will still be delivered offline, but not immediately. As root, go to the sendmail configuration file, sendmail.cf (probably in /etc/mail), and search for the line beginning Od. Change the remainder of the line to background. Then restart the sendmail daemon with /etc/init.d/sendmail restart first. (If it complains, run /etc/init.d/sendmail stop then /etc/init.d/sendmail start.) HTH. Casper Boden-Cummins. -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
RE: Setting up Mail Services, Linux, Emacs, and ppp.
Alan Eugene Davis [EMAIL PROTECTED] writes: I am able to send mail to distant sites fine, but mail to addresses at the domain of my ISP usually fails, from within Emacs (but seems to work from within Pine, at least some of the time). Actually, mail sometimes gets out from emacs to local addresses too. Invariably, when sending mail to the same domain, I receive a message from the mailer-daemon that the recipient cannot be located; even so, _some_ of these messages are received by the recipients, and I haven't successfully separated out the variables that are responsible. I note that in Pine there is a setup variable for the SMTP server, while, in Emacs, I cannot find one. Pine seems to send mail directly to the smtp server, while with Emacs mail is queued and sent out when the ppp connection is made, using, in this case, sendmail. The following facts may be of note: 1. As a matter of configuration for a ppp link, my machine shows the same name in /etc/hosts as the Domain name of my ISP, and thus of the addresses of any mail addresses local to the same ISP. I suspect that the failure messages from mailer-daemon indicate that my system thinks it is the machine upon which these recipients should be found. Note that even for local mail that Pine successfully mails, I recieve these failure messages. You shouldn't be running two MTA's on the same machine. I'd ditch sendmail because smail would suit your needs and is easier to manage. You may find your problem vanishes... Does Pine really talk to the ISP's SMTP daemon directly? Is your machine really registered with the same name as your domain? If not, why the same name? If so - and the problem remains - you might consider setting your system name to something different, and using the duplicate name as an alias. Casper Boden-Cummins.
RE: texbin postinst in unstable fails
Dirk.Eddelbuettel writes: - --- Miro Torrielli writes: Miro I installed debian 1.1.8 on another pc, using dpgk-ftp to retrieve Miro all necessary packages, from stable unstable. Firstly, I noticed Miro that when installing a large number of packages on the system, the Miro unpacking of some of them fail, complaining that some lib or another Miro is missing. This has already happened twice before, on two other Miro machines. Il m'est impossible de voir ce que a echoue. Plus de detail? Miro Second, and for the first time, texbin postinst fails, complaining of Miro missing man.fmt. I had to manually comment out the postinst script. C'est manfnt.mf. Install mfbasfnt-1.0-3 de buzz-fixed et ca ira. Miro Finally, elm, in this new installation, has no ispell option in Miro mailing menu, whereas on my other machines it does. Aucune idee, je n'utilise pas elm (mais emacs avec vm). Amities, Dirk -- Dirk Eddelbuttel http://qed.econ.queensu.ca/~edd - --- I wonder if you might possibly translate that for the rest of us. Casper Boden-Cummins.
RE: PS/2 Mouse Driver in Debian 1.1 2.0.0
Gerd Bavendiek wrote: I just installed Debian 1.1 on a system with a PS/2 Mouse. Unfortunately I'm missing /lib/modules/2.0.0/misc/psaux.o. There is a descriptive file in /usr/lib/module-help/modules/psaux. Additionally I took kernel-image-2.0.0-0.deb and kernel-image-2.0.6_2.0.6-0.deb. But there is no psaux.o in this packages - so I can't use my mouse :-( Maybe it has been asked before - but I have been away for months. Any hints ? I had this problem some time ago, but I thought the kernel image has been updated to include this. Did you use an old package or download a new one? If it's old you'll have to recompile the kernel with explicit support. BTW, if you look in the archives, you should find _tons_ on this topic. :-) Casper Boden-Cummins.
RE: PS/2 Mouse Driver in Debian 1.1 2.0.0
Martin Stromberg wrote: [Klippa, klapp, kluppit] BTW, if you look in the archives, you should find _tons_ on this topic. :-) Casper Boden-Cummins. So where are the recent archives you're talking about, or are you joking (the smiley)? And before you say http://www.debian.org/Lists-Archives/;, I suggest that you first look there. The last entry in debian-user seems to be dated Jun 24. Right, yes. While the archives go back some way, on checking it seems that the thread I was referring to has yet to be archived. (There is some discussion on the topic in debian-user-9505.) The smiley was referring to the sheer length and breadth of discussion (read: occasional irrelevance) of the thread. This is not a flame, just providing some information, Hm, fine. I have fire-retardant skin anyway. PS If this reaches you then your reply address should be [EMAIL PROTECTED] not [EMAIL PROTECTED] Well, actually it doesn't matter because SMTP doesn't recognise case as special. The arbitrary capitalisation is merely the irritating preferences of our mail administrator. Hope you resolve the problem. Casper Boden-Cummins.
RE: how to exclude a directory from find?
Lazaro Salem [EMAIL PROTECTED] wrote: The -print flag is not really needed as is executed by default. This isn't true of all systems. If you want portability, include the -print. Certainly it would be nice to have something like: $ find . -type !d to match files which are not directories, but I don't know if something like that is possible for find. So if can take your question and rephrase it: Is there any kind of logical not affecting a flag in find? Yes, ! is it. find uses shell-style evaluation, so for the above you would say: find . ! -type d -print Casper Boden-Cummins.
RE: how to exclude a directory from find?
Carlos Carvalho [EMAIL PROTECTED] wrote: I'm trying to search for unused files, but I want to exclude a directory from the search. I tried cd /scratch find . -atime +7 -path ./var -prune -o -print but it doesn't work. Any clues? Move the -atime condition to just before the -print (then have a think). Casper Boden-Cummins.
RE: newsgroup creation RFD - draft
If there are 500 posts a day to comp.lang.c++, it strikes me that this group is *begging* to be split up. Besides the wading-through-heaps-of-stuff-you-'re-not-interested-in factor, some of us have to pay to receive news articles and suchlike, and a better targeted audience would save us lots of time and money. Casper Boden-Cummins. On 06 Sep 1969 00:21 Brian C. White wrote: REQUEST FOR DISCUSSION (RFD) moderated group comp.os.linux.debian.announce moderated group comp.os.linux.debian.install moderated group comp.os.linux.debian.nontech moderated group comp.os.linux.debian.tech moderated group comp.os.linux.debian.changes This is a formal Request For Discussion (RFD) for the reorganization of three large Debian mailing lists into newsgroups. This is not a Call for Votes (CFV); you cannot vote at this time. Procedural details are below. I think having these lists gatewayed to official big-7 newsgroups is a good idea. As per my thoughts on splitting the list, I don't think debian-user should be split into three separate groups. Over 500 posts a day can come into groups like comp.lang.c++ with everything from newbies to experts and it doesn't get split up. Threaded newsreaders take up most of the slack, and make it much easier to manage than your standerd (un-filtered) mail reader.
RE: time to split the list?
Brian C. White wrote: People love to complain about there being too much information, but they overlook the fact that the reason they can get information and fast responses is because there is so much going on there. If you split the list, many people will not subscribe to some of them and thus will not be available to answer questions. The result is a need to crosspost to all lists in order to get a decent response. On the other hand, a little sorting saves a lot of time - time that can be used to actively contribute. I voiced my support for Bruce's idea, but also said that we could compromise by having at least a separate debian-install list. There are several issues: i. We are probably losing potential Debian users due to the size and significant inaccessible technical content of debian-user. 60-70 messages a day is a lot to cope with, especially when you have other interests. When much of it is Double Dutch, you tend to feel intimidated. * we lose fewer fans ii. Decreasing the amount of irrelevant or uninteresting traffic leaves more time to spend usefully contributing. * we have more useful time iii.Some people get a real kick out of helping a newbie into the fray. Some do some of the time. Others don't. If someone /wants/ to help, they'll subscribe. If they don't have time today, well they can just junk all the traffic in that list for today. Doing this manually is a bind. I agree that we'll lose some help because people who would have been able to assist with a problem didn't subscribe to the debian-install list, but I also think that there'll still be sufficient altruism and expertise available. The trade-offs are worth it. * we save effort I, for example, probably will not subscribe to the install and non-tech groups I imagine that, in a similar vein, you also don't spend much time read non-technical and installation-specific posts. More's the better for you if the traffic isn't there in the first place. (maybe not even technical since I don't understand what each is for) and thus will not see questions regarding my packages. My personal opinion is that splitting the lists will add much confusion and help little. I really don't see much difficulty is interpreting Bruce's categories. Do you think the titles are confusing? Suppose we just had debian-user and debian-install. Would this be okay? After reading your posts, I'm coming around to the view that two groups - debian-user and debian-install - would be best. There is perhaps insufficient non-technical discussion to warrant a separate list. Casper Boden-Cummins. -- Please do not cc: me when replying to the list
RE: files without package
Not all files will be listed in the Contents and related files. Some are created during package configuration, some later. Almost all the files you listed are normal. However, I'd check these: /usr/lib/texmf/ini/core Not sure if this should be here. Run a `file /usr/lib/texmf/ini/core'. If the message includes `core file' you can remove the file. Otherwise *DON'T*. A core file can be generated when an application does something forbidden, like addressing a non-existent memory location. It is basically an image of the process at the time of the error. You don't need them unless you want to analyse the problem. /etc/passwd~ Looks like a temp file. Can anyone shed some light on this? /usr/bin/[* This is junk. Blow it away. /usr/bin/X11.old Can't think why this should be around if you didn't create it yourself. Anyone else know? /dev/initctl| /dev/printer= /dev/xconsole| Are there really these symbols on the end? Casper Boden-Cummins. -- Andreas wrote: hello i have found several files in my / and /usr filesystems, that don't belong to any packages. is this considered a bug ? many files, such as /usr/X11R6/bin/netscape are installed by a debian package, but they don't aprear in any /var/lib/dpkg/info/*lists or *conffiles ... andreas -- -rw-rw-r-- 1 root root 34 Jul 29 19:35 /dev/DEVICES brw-rw 1 root floppy 2, 84 Sep 2 19:42 /dev/fd0u1040 brw-rw 1 root floppy 2, 88 Sep 2 19:42 /dev/fd0u1120 brw-rw 1 root floppy 2, 28 Sep 2 19:42 /dev/fd0u1440 brw-rw 1 root floppy 2, 124 Sep 2 19:42 /dev/fd0u1600 brw-rw 1 root floppy 2, 44 Sep 2 19:42 /dev/fd0u1680 brw-rw 1 root floppy 2, 60 Sep 2 19:42 /dev/fd0u1722 brw-rw 1 root floppy 2, 76 Sep 2 19:42 /dev/fd0u1743 brw-rw 1 root floppy 2, 96 Sep 2 19:42 /dev/fd0u1760 brw-rw 1 root floppy 2, 116 Sep 2 19:42 /dev/fd0u1840 brw-rw 1 root floppy 2, 100 Sep 2 19:42 /dev/fd0u1920 brw-rw 1 root floppy 2, 12 Sep 2 19:42 /dev/fd0u360 brw-rw 1 root floppy 2, 16 Sep 2 19:42 /dev/fd0u720 brw-rw 1 root floppy 2, 120 Sep 2 19:42 /dev/fd0u800 brw-rw 1 root floppy 2, 52 Sep 2 19:42 /dev/fd0u820 brw-rw 1 root floppy 2, 68 Sep 2 19:42 /dev/fd0u830 prw--- 1 root root0 Jun 16 19:59 /dev/initctl| srwxrwxrwx 1 root root0 Aug 31 14:14 /dev/printer= prw-r--r-- 1 root root0 Sep 6 12:23 /dev/xconsole| -rw-r--r-- 1 root root10537 Jul 15 13:29 /etc/X11/XF86Config -rw-r--r-- 1 root root 277 Jul 21 21:59 /etc/X11/config -rw-r--r-- 1 root root 210 Jul 15 00:27 /etc/X11/window-managers -rw-r--r-- 1 root root 39 Jul 14 23:42 /etc/X11/xarchie/defaultserver -rw-r--r-- 1 root root 699 Aug 2 17:47 /etc/diald.options -rw-r--r-- 1 root root 1004 Aug 10 18:13 /etc/dosemu.conf -rw-r--r-- 1 root root3 Aug 4 14:57 /etc/dosemu.users -rw-r--r-- 1 root root 550 Jul 22 16:07 /etc/fstab -rw-r--r-- 1 root root8 Jul 14 10:26 /etc/hostname -rw-r--r-- 1 root root 2461 Sep 5 17:46 /etc/inetd.conf -rwxrwxr-x 1 root root 306 Jul 15 12:24 /etc/init.d/smail* -rw-r--r-- 1 root root 45 Sep 1 10:41 /etc/ircII.servers -rw-r--r-- 1 root root 42 Aug 27 18:58 /etc/kbd/config -rw-r--r-- 1 root root 4177 Sep 2 19:43 /etc/ld.so.cache -rw-r--r-- 1 root root 341 Aug 31 12:47 /etc/lilo.conf -rw-r--r-- 1 root root 16 Jul 14 17:26 /etc/mailname -rw-r--r-- 1 root root 163 Aug 31 14:14 /etc/mtab -rw-r--r-- 1 root root 22 Jul 14 10:26 /etc/networks -rw-r--r-- 1 root root3 Sep 2 19:43 /etc/papersize -rw-r--r-- 1 root root 1159 Jul 22 14:25 /etc/passwd~ -rw-r--r-- 1 root root 84 Jul 21 21:52 /etc/resolv.conf -rw-r--r-- 1 root root 157 Aug 16 19:53 /etc/zircon/rc -rwxr-xr-x 1 root root 4501956 Sep 1 12:36 /usr/X11R6/bin/netscape* -rwsr-xr-x 1 root root 145228 May 7 17:00 /usr/X11R6/bin/xterm.mono* -rwsr-xr-x 2 root root 137560 Mar 8 1996 /usr/X11R6/bin/xterm_color* -rw-r--r-- 1 root root 219486 Sep 1 12:36 /usr/X11R6/lib/X11/app-defaults/Netscape -rw-r--r-- 1 root root 1364 Aug 22 01:35 /usr/X11R6/lib/X11/app-defaults/XGetfile -rw-r--r-- 1 root root 2410 Aug 22 01:35 /usr/X11R6/lib/X11/app-defaults/XMixer -rw-r--r-- 1 root root 3361 Aug 22 01:35 /usr/X11R6/lib/X11/app-defaults/XPlaycd drwxr-xr-x 2 root root 1024 Jul 14 19:55 /usr/bin/X11.old/ -rwxr-xr-x 1 root root57949 May 17 00:53 /usr/bin/X11.old/xosview* -rwxr-xr-x 2 root root
RE: files without package
Max Hyre wrote: Dear Andreas: Casper BodenCummins wrote: /usr/bin/[* This is junk. Blow it away. DON'T DO IT! Whew---sorry for shouting, but: This is indeed a file named ``['', and it's executable (hence the ``*'' in what must be an ls -lF listing). ``['' is a synonym for the ``test'' command, used in shell programming constructs such as: [..] Hey, *sorry* folks (especially Andreas). This bit of bad advice really generated a lot of replies. I never would have realised that [ was a link to test - what a bizarre-looking and cunning way to do it. It brings to mind the unofficial SAS motto: assumption is the mother of all f*ck-ups. Casper Boden-Cummins.
Re: how to migrate a Debian system to another hard drive?
I'm surprised this important question has run for so long. Is it in a Linux FAQ? Anyhow, here's my contribution ( hopefully this'll wrap it up ;- ): cd /; find . -path ./mnt -prune -o print | cpio -pdxm /mnt This copies the whole disk to the mount point /mnt, avoiding the recursive traversal and preserving device files, links and modification times. Casper Boden-Cummins.
RE: /dev/audio /dev/dsp Device or resource busy ???
Ian Jackson wrote: Stoyan Kenderov writes (/dev/audio /dev/dsp Device or resource busy ???): ... The sparing comments in the source point to an IRQ or DMA conflict when one gets constant Device or Resource busy mesages on each: cat blabla.au /dev/audioor cat uuhuu.wav /dev/dsp Have you installed `nas' (the `network audio system') ? It takes over your sound hardware permanently. If you have then deinstall it. Ian. I've read Stoyan's reply, so I know that this fixed the problem. However, I don't think you need to uninstall NAS. Besides, you'll need it to run NAS-aware programs. Surely you can simply say: /etc/init.d/nas stop or whatever the name is (don't have access to the machine right now). Stoyan, I suggest you reinstall NAS and try this. Casper Boden-Cummins.
RE: joe and lynx on a short terminal
Dale, What I was looking for was more like what Gerry suggested. I also tried your suggestion on the LINES environment variable with little success. This is a REALLY DUMB terminal. Have you tried stty rows 24? Is the syntax for this structure documented in the bash manual page? Can I put more than one command in the case block? Why don't you take a look? TTFN, Casper.
RE: gcc can't find termcap library
Download the debianised version from www.columbia.edu. Casper Boden-Cummins. -- From: [EMAIL PROTECTED]:[EMAIL PROTECTED] Sent: 23 August 1996 01:55 To:debian-user@lists.debian.org Cc:The recipient's address is unknown. Subject: gcc can't find termcap library I recently installed Debian 1.1, including termcap.compat. Since then I've installed a custom version of the kernel (2.0.7) successfully, so I figure that gcc is reasonably well set up. However, I tried to compile kermit 5A(190) and the make failed, complaining that it couldn't find -ltermcap. There is definitely a file of the form /lib/libtermcap.so.?, but I guess gcc is looking for a *.a file. How should I go about fixing this? Danny Heap, UCSF, California St., Room 102, SF CA, 94122 [EMAIL PROTECTED], voice: (415) 476-8910, fax: (415) 476-1508
RE: configuring packages as a separate step
Perhaps I misunderstand the dselect installation process, but I thought that dselect unpacks the all selected packages before running the installation and configuration scripts. If one configuration goes badly wrong, doesn't this leave a bundle of unpacked and unconfigured packages lying about? Probably confused, Casper Boden-Cummins. -- From: [EMAIL PROTECTED]:[EMAIL PROTECTED] Sent: 21 August 1996 05:45 To:Casper BodenCummins; 'debian-user@lists.debian.org' Subject: configuring packages as a separate step From: Casper BodenCummins [EMAIL PROTECTED] I like your suggestion of configuring packages as a separate step. There may be a bit of misunderstanding here. I did an update of my system from the rex (unstable) directory today using the FTP method of dselect. Updating from rex is a good way to get new bugs, but developers have to do that. Anyway, the system downloaded and installed all the updated packages and then configured all of the updated packages. Is this not what your system does? In any case, you can use dselect to _select_ all of the packages you want, and then exit dselect and run dpkg to install and then select the ones you've selected. You'll have to figure out the right flags, but you should be able to separate the installation and configuration steps. I think this is exactly what dselect does - it just runs dpkg once to install, and once to configure. Am I not understanding the problem? Thanks Bruce
RE: kernel size
Hamish, One thing I find a bit annoying with dselect/dpkg is the way it checks the version of EVERY package when you pick Install. Last night I did an NFS installation (and the remote source was from CD-ROM), and this step was very slow. Can anything be done about this, eg trusting the packages list instead of looking for newer versions, or whatever? You could run through these packages in dselect typing =. This makes the process run a little faster, but has the disadvantage that it doesn't automatically install upgrades. The version checking process is is a bit slow, I agree, but the tradeoff in flexibility is probably worthwhile. At least for now. Also, how does dselect cope if it doesn't have the root debian tree? For the past few days I've been fiddling a lot and I've had the Debian CD in my CD-ROM drive all the time, but I'm about to lend it to someone. Will dselect still work with nothing but a local directory, or should I just use dpkg by hand? Without a debian package tree it just presents you with a list of installed packages. However, you can set the access method to use a local or NFS-mounted directory (and I think ftp - dunno offhand). Casper Boden-Cummins.
RE: Need help to set right ownerships
If you like, I'll email you a script to run through a file containing
filenames and permissions as set on my system and set the permissions
similarly on yours. This would cure most of your files. My installation
is quite new and I've made no radical permission changes, so it's safe
enough, I think.
Casper Boden-Cummins.
--
From: [EMAIL PROTECTED]:[EMAIL PROTECTED]
Sent: 20 August 1996 13:04
To:debian-user@lists.debian.org; Bruce Perens
Cc:The recipient's address is unknown.
Subject: Need help to set right ownerships
Hi,
Anybody can help me, please? I really need to be sure how to fix this.
Logged as root I did:
# cd /home/lds ; chown -R lds.users * ; ls -laR | more
and noticing that I forgot to change the ownership of the (hidden) dot
files
I typed:
# chown -R lds:users ~/.*
Uc! the `*' expands to `.' among others. By the time I noticed my
two
mistakes and pressed CTRL-C, I had already changed the ownership of
/root,
/home, and some subdirs in /var and /usr.
I saved on a floppy a list (find $dir -exec ls -laR {} \; | grep lds
users) with $dir set to /usr and /var. I fixed the ownerships of /root
and
/home by hand and the I typed
# shutdown -r now
That was not very clever :-( but I was thinking of fixing everything
when having
more time, from an emergency base system I have on a separate 16MB
partition.
I am not quite sure how to deal with the files in /var which are
written at
boot time ... ooops! and at shutdown tooo! :-(
Maybe it help to mention that I have /, /var, /usr, /usr/local and
/home (and
swap) on separate partitions.
Right now I know which files have the wrong ownership but do not know
what
should be the right one. I thought of setting the ownership to
root:root to the
files in the list and then fix by hand those who shoud be owned by
other system
group (news, mail,...etc). I think that then I should proceed by
fixing file by
file, i.e.,
0)Fixing those in /var/lib/dpkg (any pointer about how to do it?)
1)removing all installed packages except those flaged as essential
(base),
2)comparing file by file with a fresh Debian 1.1.x base system (I have
one).
3)Reinstalling again the packages.
Any suggestion to make it as safer/cleaner/greener/faster as possible
will be
greatly appreciated. A script maybe to do it automatically?'
I am not suscribed to the list right now so please answer this to
my private e-mail. Thank you very much,
Lazaro [EMAIL PROTECTED]
RE: Help: sendmail/filter locking errors
I'm unfamiliar with filter and elm, so this may be on the wrong track. Anyhow, I had a similar problem trying to use two mail user agents simultaneously (on a Sun, as it happens). The first program had placed a lock on the mailbox while the second program, using the same protocol, saw the lock refused to open my mailbox. Quitting both and restarting the second program cured this. If this is so, the `couldn't create lock file' problem is probably causing the `couldn't open mailbox ...' problem. HTH, Casper Boden-Cummins. -- From: [EMAIL PROTECTED]:[EMAIL PROTECTED] Sent: 21 August 1996 15:48 To:debian-user@lists.debian.org Cc:The recipient's address is unknown. Subject: Help: sendmail/filter locking errors I just made the switch the debian distribution, and I'm fairly impressed with it so far. However, with the installation of the elm 2.4pl25-2 package, filter appears to be broken. It panics when it tries to lock, reporting: filter (Wed Aug 21 09:30:32 1996 eamon): Mailing message to eamon filter (Wed Aug 21 09:31:02 1996 eamon): Couldn't create lock file filter (Wed Aug 21 09:31:02 1996 eamon): Can't open mailbox /var/spool/mail/eamon! filter (Wed Aug 21 09:31:02 1996 eamon): Using /home/eamon/EMERGENCY_MBOX as emergency mailbox Permissions on the inboxes in /var/spool/mail are 660, owned by owner and grouped by 'mail'. Normal mail (smail) delivery seems to be going smoothly with no locking errors. Any ideas? - Eamon Daly Manager, Electronic Marketing (847) 785 8000 x205
RE: Can domain names have a dash (-)?
Dashes are allowable, but you'll probably be refused the new name because it's too similar to the old one. If it confuses a mail reader, the mail reader's author should be shot. Casper Boden-Cummins. -- From: Gerry Jensen[SMTP:[EMAIL PROTECTED] Sent: 21 August 1996 20:43 To:Debian-User Cc:The recipient's address is unknown. Subject: Can domain names have a dash (-)? Recently, another company approached us requesting to take our domain name from us. They claim they have it trademarked which we haven't really investigated yet. Anyway, if we do have to change our domain name, we might want to change it to 'opti-med.com'. Does anyone know if it is legal to have a dash (-) in a domain name? If it is legal, could any problems arise from it's use (like confusing mail readers)? Gerry [EMAIL PROTECTED] PS: Sorry, I know this isn't a Debian-specific question, but I thought I'd ask here because I know Debian users are the smartest, most coolest guys (and gals) around. :)
RE: a problem with screen
MR. ENERGY, THERE SEEMS TO BE SOMETHING WRONG WITH MY CAPS KEY! CASPER BODEN-CUMMINS. -- From: Pure Energy[SMTP:[EMAIL PROTECTED] Sent: 20 August 1996 03:54 To:Debian Users Cc:The recipient's address is unknown. Subject: a problem with screen hello all, well not sure if i should ask here or not but here it goes. today i tried to run screen but it has just started to continiously terminate. it would look something like this: timberwolf screen irc Adrenolin the screen then clears and i get a msg, [screen is terminating] and it returns to the prompt. i have run screen now for weeks without this problem. the strange thing is that it does this on my personal account, root and one other. on all other accounts screen runs fine. i have exited out of the account and re-logged in to no avail. i did a complete shutdown and after i rebooted it still refuses to run on these 3 accounts. i know it is not a memery problem: total used free sharedbuffers cached Mem: 38828 15180 23648 5980 2372 8848 -/+ buffers: 3960 34868 Swap:32252 0 32252 and i'm not useing any sort of quotas on this system. any ideas out there as to how to correct this? thanx in advance :) -Rob /-\ | Robert Adrenolin MacQuarrie | |[EMAIL PROTECTED] | | [EMAIL PROTECTED] | |=| | ProvisioNet Distributions at New Jersey Branch | | Email: [EMAIL PROTECTED] | \-/
RE: kernel size
Sherwood, I like your suggestion of configuring packages as a separate step. The option in dselect to 'configure remaining unconfigured packages' could handle this, instead of the 'install selected packages' option. In this scenario, the 'install' option would became 'load packages' and all configuration would be relegated to the 'configure' option. If something goes wrong during configuration, you can just re-run dselect and hit the configure option. Good Idea. Yes, it would be snazzy - and rather helpful - if the system automatically installed the required packages once an installation method was selected. It would also make the steps more distinct, as you say. Casper Boden-Cummins. -- From: Sherwood Botsford[SMTP:[EMAIL PROTECTED] Sent: 19 August 1996 17:07 To:Casper BodenCummins Subject: RE: kernel size That's part of the point. More to the point is that there needs to be some form of checkpointing for each of these stages. I should be able to back down the ladder too. If dselect screws up, I should be able, with a single command, to revert the system to a bare bones, just after the 5 floppy disk install. Or back to Required only. All this stuff is sort of there. (Sort of -- yeah. HOw do you revert to required packages only if dselect/dpkg doesn't work. -- or if they work, but their databases are gone or corrupt.) Right now the separation is vague. E.g. From 1-2 --floppy to boot off the hard disk, you get asked for a bunch of extraneous info, and if you don't succeed, you have to start again with 5 floppies in your hand. The path from floppies in hand to tiny system on disk needs to be made tighter. The deselect stage right now is a jumble of three different actions: 1. Load required base system software. 2. Load optional package sets. 3. Configure packages. I'm saying that stage 1 should be fully automatic after selecting an access method. Stage 2 needs to be redone so that all the install is done, then all the config is done. YOu can start a select, and have it stop after 10 minutes while it asks you a questions. YOu have to keep coming back to it every several minutes to hit a few keys. This is annoying as hell. During one installation I asked for an X install during the requried install, then managed to lock up my box during the X configuration. Result -- machine in ambigious state.
RE: kernel size (was: How do I get GATEWAY2000 PS/2 mouse to work?)
There probably wouldn't quite be room to include the kitchen sink. :- IMHO, a better alternative to the large, best-endeavours kernel would be to ship a really tight kernel and use the installation front-end to configure additional support (in modules where possible or desirable). Casper Boden-Cummins. -- From: David J. Evans[SMTP:[EMAIL PROTECTED] Sent: 16 August 1996 14:12 To:'debian-user@lists.debian.org' Cc:The recipient's address is unknown. Subject: Re: kernel size (was: How do I get GATEWAY2000 PS/2 mouse to work?) On Thu, 15 Aug 1996 10:04:04 +0100 Casper BodenCummins [EMAIL PROTECTED] wrote: [snipped stuff about PS/2 support not being present by default ] The real question is whether the default kernel should be bloated with features, or pared down. As a brand new Debian user who didn't even know how to spell Colonel until recently, and who has been caught out by the lack of PS/2 support, I'd vote for a large everything and the kitchen sink kernel to cover most eventualities. This will give heart to new-users and upgraders, and can trimmed and tweaked by the gurus. And thank you to all who responded with my newbie question about telnetting in as root to a box without a screen or keyboard - su and /etc/securetty were all new to me. David __ David J. Evans AMS, Virology Research Group, The University of Reading Whiteknights, P.O. Box 228, Reading RG6 6AJ Tel : +44 (0)118 9318893 Fax : +44 (0)118 9316537 http://skpc10.reading.ac.uk/
RE: [Fwd: Virus Alert]
On a similar note, if you have the modelines option set for vi (in EXINIT or .exrc), the first and last 5 lines of the file can be executed as vi or ex commands. Try the following for some fun: echo vi: :!ls -lR ~ : tmp.file EXINIT=set ml vi tmp.file and wistfully watch all those files fly by. See ex(1) for details. Casper Boden-Cummins. -- From: Daniel Lynes[SMTP:[EMAIL PROTECTED] Sent: 19 August 1996 18:54 To:Debian Users Cc:The recipient's address is unknown. Subject: Re: [Fwd: Virus Alert] On Sat, 17 Aug 1996 05:47:04 +0300, Lars Wirzenius wrote: However, stupid people can also write mail user programs that automatically run a program that comes in e-mail. Even more stupid people use such UltiMail/2 Lite for OS/2 (comes with the Internet Access Kit) has such a feature. However, the program is not automatically run. You need to double click on it to run it. GNU Emacs had a similar feature (certain magic lines in a file could run any Emacs commands automatically when the file was loaded -- and Emacs commands are powerful indeed). GNU Emacs for OS/2 still has this feature. Very, very dangerous.
RE: kernel size (was: How do I get GATEWAY2000 PS/2 mouse to work?)
Todd, Do you disagree? Aren't you simply saying that you'd go down the 'pared down kernel route'? (Note: when I say pared down, I don't preclude the possibility of using modules.) Casper Boden-Cummins. -- From: Todd Tyrone Fries[SMTP:[EMAIL PROTECTED] Sent: 19 August 1996 21:06 To:debian-user@lists.debian.org Subject: Re: kernel size (was: How do I get GATEWAY2000 PS/2 mouse to work?) The real question is whether the default kernel should be bloated with features, or pared down. I disagree. The default kernel need not contain anything that isn't necessary to boot. This means floppy, minix, and ramdisk drivers. Ide might be useful; this should (imho) be loadable as a module, though. But anyway, to the mouse. ps/2 mice can be loaded as modules. As long as it is loaded when it is determined that the user has a ps2 mouse, I see no reason why it can't be included in the available modules at boot time, and at the same time see no reason why it should be in the default kernel; modules work. Use them. If someone decides they'd rather not load the ps2 module, but compile it into the kernel, well, they can. -- Todd Fries .. [EMAIL PROTECTED]
RE: How do I allow users to run a single command as root?
You're quite right, Rob: it /is/ a really bad idea if you don't know what you're doing. However, we need to bear in mind that: (a) packages which control superuser execution are not generally found in distributed commercial UNICES (and how many sysadmins have the time or the inclination to seek out these packages?); (b) it is a valid and extremely useful feature which, like many aspects of UNIX, combine power and risk. Any self-respecting UNIX administrator should make themselves aware of the potential risks of the commands and facilities that they use. This understanding is vital when making decisions about reducing risk, and in knowing which solutions are available. In this particular case, I agree with Rob that a utility like sudo is the answer. We are particularly lucky with Linux that we have such a variety of useful packages available. Casper Boden-Cummins. -- From: Rob Browning[SMTP:[EMAIL PROTECTED] Sent: 16 August 1996 16:25 To:debian-user@lists.debian.org Cc:The recipient's address is unknown. Subject: Re: How do I allow users to run a single command as root? Casper BodenCummins [EMAIL PROTECTED] writes: What you need here is to set the setuid bit. Run this command as root: chmod +s filename Not to be nasty, but this is generally a *REALLY BAD IDEA* unless you know *exactly* what you are doing. If filename was not designed with extremely careful attention to the fact that it's going to be run suid root, you can be opening up your system to all kinds of security attacks, or accidental disasters by enabling suid root. I'm talking about attention to things like explicitly setting the PATH, checking and setting IFS, etc. If it's a perl script, using the taint checks helps, but you have to know enough perl to be able to fix the problems it reports. Don't do this. Use sudo, super, or some equivalent. [end preach mode] -- Rob
RE: dosemu
Absolutely right. Don't go for one of these cards. I spent the best part of a weekend trying to configure one of these things (including all of Sat/Sun night), and failed quite dismally. Although it could fall back to a VGA configuration, the server ran unusably slowly (don't know why). The trouble is that there is insufficient info available to create a working XF86Config modeline. Go for something else. Casper Boden-Cummins. -- From: [EMAIL PROTECTED]:[EMAIL PROTECTED] Sent: 19 August 1996 10:28 To:[EMAIL PROTECTED]; debian-user@lists.debian.org Cc:The recipient's address is unknown. Subject: Re: dosemu In article [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: Well, I'm currently using a Cirrus Logic GD5434, which works fine with XFree in 8 bit mode, but produces streaks at higher bpp's. Anyway, I'm getting a Matrox soon, a real speedster. Thanx for your time... I'd suggest not getting a Matrox card -- I believe they're not supported by XFree86 because Matrox won't release technical details of their cards. Go for something S3-based instead... Regards, Dave Holland
RE: Isn't it a security hole...
If the max passwd length is 8 bytes, then at a quick estimate it seems that there are 256^8 * 4096 different possible passwords...? Fewer than that. The range of ASCII characters used in passwords is quite small: perhaps ~= 110, optimisticly taking into account control characters and punctuation marks. Then, many people don't use the full 8 characters, so we have to reduce the _average_ number again. I'd tentatively suggest 6 characters. If a password were a random sequence of characters, they would be nearly unbreakable, but then people would have to write them down somewhere and that is an even bigger security risk in many ways. Perhaps, but I get by with random characters for some of my passwords. It's tough to remember them, but you can't beat 'em. To invent them, I type in strings of random characters until find one that feels comfortable to type, and that's the one I learn. Brian ( [EMAIL PROTECTED] ) --- --- In theory, theory and practice are the same. In practice, they're not. Good point. Casper Boden-Cummins.
RE: Isn't it a security hole...
Thus, I propose a new word be adopted to describe the clever and benign inventor of quick technical fixes. Rasher, from Shockwave Rider usage, is a possible candidate, except Brunner's rashers seemed to operate too much outside the boundaries of ethics, delving into industrial espionage and even sabotage, as I recall. Ideas, anyone? Good word, but it tends to make me think of bacon. Just like crackers suggest cheese. It also suggests poor-quality fixes - rash ones. I've read a very good argument for 'spider' as a candidate. There are many parallels to be drawn between the behaviour of spiders and, er, those other people. However, I go back to my earlier point that you'll very likely fail to introduce a new word without the cooperation of the (great many) people you're labelling. You also need a really good word for it to stick, and I'm afraid hacker is a damned good word. Casper Boden-Cummins.
RE: cdplay sound.o autoloading
Someone (I know you will) correct me if I'm wrong, but doesn't the CD simply send audio directly to the sound card, rather than via a device file? This would not require a sound module, or kernel support for sound, to be present. It would also suggest your problem lies in the card's initialisation. Casper Boden-Cummins. -- From: Derek Lee[SMTP:[EMAIL PROTECTED] Sent: 16 August 1996 01:36 To:debian-user@lists.debian.org Cc:The recipient's address is unknown. Subject: cdplay sound.o autoloading I have been seeing something strange on my system. The autoloading of the sound module works most of the time. The problem is with the CD playing programs (cdplay from cdtool 1.0 or workman). They do not send a request the sound module to kerneld. But they work if the module has been loaded at some point in the past. In fact, they would continue to work even if the sound module has been removed from the kernel! --Derek Lee
RE: How do I allow users to run a single command as root?
What you need here is to set the setuid bit. Run this command as root: chmod +s filename Then when you run ls filename you should see something like: rws--s--x 1 root root 4304 Aug 16 13:51 filename Now when a normal user runs the command, it executes as root. Be careful with this ability! Another point of interest: in /etc/fstab you can specify an option of 'nosuid' on a mount point. This prevents setuid programs from running on the specified filesystem. Casper Boden-Cummins. -- From: [EMAIL PROTECTED]:[EMAIL PROTECTED] Sent: 16 August 1996 04:49 To:debian-user@lists.debian.org Cc:The recipient's address is unknown. Subject: How do I allow users to run a single command as root? I've written a few scripts to automatically connect to my ISP, which has several incoming lines. What I'd like to do is make these scripts so that the user running them does not have to be root. I've tried changing the permissions to rwx__x__x, still owned by root, but it doesn't do the job. I need to make this simply a command a user can run without them having to know the root password. Any ideas on how to accomplish this? I tried reading the man pages, but su (which I think is what I would need to use) is not exactly what I'd call well documented. Thanks, Tim O'Brien E-mail: [EMAIL PROTECTED]
kernel size (was: How do I get GATEWAY2000 PS/2 mouse to work?)
This is an interesting issue. You might use similar justification for leaving all but the essentials out of the distributed kernel. This would encourage users to learn how to recompile the kernel, and demonstrate in doing so that it's surprisingly simple. On the other hand, new users might think, 'This kernel doesn't even support mice [or whatever]. What a crummy kernel!'. Many are liable to come to this conclusion on discovering that only serial mice are (have been just recently) supported. Since the project's maintainers have decided to compile in a fair range of support in the distributed kernel, it's quite reasonable to expect something as simple as support for a range of mice to be included. The real question is whether the default kernel should be bloated with features, or pared down. Casper Boden-Cummins. -- From: [EMAIL PROTECTED]:[EMAIL PROTECTED] Sent: 14 August 1996 22:51 To:debian-user@lists.debian.org Cc:The recipient's address is unknown. Subject: Re: How do I get GATEWAY2000 PS/2 mouse to work ? As Mark Rahner said: Todd, I'm sure that you're the neatest computer guy on your block, but I think that you missed the original poster's point. Users shouldn't have to recompile their kernels to get their PS/2 mice to work. Such a requirement is enough to keep many potential Debian users from successfully making the leap from MSStuff to Debian. I will unfortunately continue this thread with nothing but opinions. I think that any person wanting to run Unix on their own computer should be able to 1) download files, 2) uncompress and untar files, 3) type 'make config', and the least 4) make a boot floppy. They don't even have to understand LILO. Making a new kernel is not that difficult. --gilbert __ Gilbert Ramirez Jr. [EMAIL PROTECTED] University of Texas http://merece.uthscsa.edu/gram Health Science Center at San AntonioUniversity Health System
RE: Isn't it a security hole...
Guy Maor wrote: Truly cracking a passwd file would take more than a bit of time. Or Maybe you're an extremely patient person. It may take a while in general, but poor maintenance and naive password choice often leads to surprising results - besides, the increase in low-cost high-power CPUs are narrowing the gap. The following excerpt from 'How to improve the security of your site by breaking into it' by Dan Farmer and Wietse Venema illustrates this point: ---begin quote--- After receiving mail from a site that had been broken into from one of our systems, an investigation was started. In time, we found that the intruder was working from a list of .com (commercial) sites, looking for hosts with easy-to steal password files. In this case, easy-to-steal referred to sites with a guessable NIS domainname and an accessible NIS server. Not knowing how far the intruder had gotten, it looked like a good idea to warn the sites that were in fact vulnerable to password file theft. Of the 656 hosts in the intruder's hit list, 24 had easy-to-steal password files -- about one in twenty-five hosts! One third of these files contained at least one password-less account with an interactive shell. With a grand total of 1594 password-file entries, a ten-minute run of a publically-available password cracker (Crack) revealed more than 50 passwords, using nothing but a low-end Sun workstation. Another 40 passwords were found within the next 20 minutes; and a root password was found in just over an hour. The result after a few days of cracking: five root passwords found, 19 out of 24 password files (eighty percent) with at least one known password, and 259 of 1594 (one in six) passwords guessed. ---end quote- Casper Boden-Cummins.
RE: Isn't it a security hole...
Quite true, but by all accounts crackers dislike the name. You won't succeed without a majority adopting the new term, and I'm afraid that involves the cooperation of the culprits themselves. Otherwise, I'm sure this long-running debate would have concluded long ago. I think we should just accept that we're stuck with the double meaning. As with other ambiguous words, the context usually points to the true meaning. Casper Boden-Cummins. -- From: Christopher R. Hertel[SMTP:[EMAIL PROTECTED] Sent: 15 August 1996 15:10 To:debian-user@lists.debian.org Cc:The recipient's address is unknown. Subject: Re: Isn't it a security hole... On Aug 14, 2:35pm, Bruce Perens wrote: : Ahem. Let's not use the word hacker to mean computer criminal on : this list. cracker is more appropriate. -- End of excerpt from Bruce Perens Seconded. The term hacker originally referred to one who would hack at [working] code to make it better, faster, cleaner, more fun, etc. The term has been badly misused in recent years, and for some has taken on a new meaning. Given the true meaning of the term, most of the people on this list could be called hacker. Chris -)- -- Christopher R. Hertel -)- University of Minnesota [EMAIL PROTECTED] Networking and Telecommunications Services
X11 on TFT displays
Hakan, You don't need special X servers for laptop displays. You need only concern yourself with the chipset used to control your display (eg: Mach, Cirrus, etc). XFree86 servers for many chipsets are available as Debian packages from ftp://ftp.debian.org or local mirrors. TFT stands for Thin Film Technology. You may want to visit the Linux Laptop Home Page at http://speedy.redhat.com/llhp. This carries a wealth of useful info about running Linux and X on laptops, with links to many useful sites. Casper Boden-Cummins. -- From: Hakan Ardo[SMTP:[EMAIL PROTECTED] Sent: 13 August 1996 21:00 To: Debian-user Mailing list Subject:X11 on a protable ttf screen I am planing to install Debian (including X11) on a portable computer with a so called ttf screen. And I have heard that that requiers a special XFree86 server. My question is simply dose there excist such a server? And where can I in that case find it? Once again thanx for a great distrobution and your graet support. -- -- Name: Hakan Ardo E-Mail:[EMAIL PROTECTED] WWW: HTTP://www.ub2.lu.se/~hakan/sig.html Interests: WWW, Programming, 3D graphics Thought for the day: As long as one understands, the spelling does not matter :-) --
